Studies in Yogācāra-Vijñānavāda idealism I: The interpretation of Vasubandhu’s Viṃśikā

Birgit Kellner and John Taber

Studies in Yogacara-Vijnanavada idealism I: The interpretation of Vasubandhu’s Vimsika

Abstract:

In recent scholarship there has been a persistent tendency, especially among North-American scholars, to deny that Indian Yogacara philosophy is a form of idealism. The discussion has naturally focused on the interpretation of Vasubandhu's Vimsika, a foundational text of the school, as well as one of the most accessible, which other researchers have taken to be denying the existence of a material world external to consciousness.

In this article, after noting some of the points in favor of a non-idealist read­ing of the Vimsika, we shall offer a new reading that supports the old “standard”, but still widespread, interpretation that it indeed intends to deny the existence of physical objects outside of consciousness. We suggest that Vasubandhu develops in the Vimsika an extended argumentum ad ignorantiam where the absence of external objects is derived from the absence of evidence for their existence. This reading is the result of examining argumentation strategy rather than investigat­ing the logical structure of individual proofs in isolation, and it takes cues from Vasubandhu's strategy for refuting the existence of a self in Abhidharma- kosabhasya IX. In addition, our reading looks at the entire Vimsika, rather than isolating a purported argumentative “core” (vv. 11-15), and draws attention to the relevance of some of its subtleties. Finally, we also suggest that Vasubandhu might have opted for a less direct argumentation strategy to prove the non­existence of the external world because of specific soteriological aspects of the doctrine of vijnapti matrata. DOI 10.1515/asia-2014-0060

In recent scholarship there has been a persistent tendency, especially among North-American scholars, to deny that Indian Yogacara philosophy, or what Buescher 2008 calls Yogacara-Vijnanavada, is a form of idealism.

Here we under-

Birgit Kellner: Chair of Buddhist Studies, Cluster of Excellence “Asia and Europe in a Global

Context - The Dynamics of Transculturality”, University of Heidelberg, Karl Jaspers Centre, VoBstraBe 2, Building 4400, 69115 Heidelberg, Germany. E-mail: kellner@asia-europe.uni-heidelberg.de

John Taber: Department of Philosophy, The University of Mexico, MSC 03 2140, 1 University of New Mexico, Albuquerque, New Mexico 87131-0001, U.S.A. E-mail: jataber@unm.edu stand by “idealism” the view that objects cannot exist without being cognized or, as it is sometimes put, that there is no “mind-independent world.”

In the history of Western philosophy this view is sometimes referred to as “subjective idealism”, to distinguish it from other forms of idealism, such as “absolute idealism” and “transcendental idealism.” The classic statement of the position in Western philosophy was by the early modern Irish philosopher George Berkeley, who fa­mously declared that “to be is to be perceived” (esse est percipi), implying thereby that something that is not perceived or cognized cannot exist. (Absolute idealism, on the other hand, is associated with the philosophy of Hegel, while Kant identi­fied his own philosophy as transcendental idealism.) Wayman 1979, Kochu- muttom 1982, Hall 1986, Hayes 1988, Oetke 1992, King 1998, and most recently Lusthaus 2002 can all be seen as denying that Yogacara-Vijnanavada is idealism in the sense defined, though in somewhat different ways. The discussion has naturally focused on the interpretation of Vasubandhu's Vimsika, a foundational text of the school, as well as one of the most accessible, which other researchers, naively or not, have taken to be denying the existence of a material world external to consciousness.

In the present study we shall argue for an idealist interpretation of the Vimsika. We, first, review the non-idealist interpretation of this text as presented in recent literature, then assess what we believe to be the strengths and weak­nesses of such an interpretation. We conclude that, while the non-idealist inter­pretation has certain points in its favor, it is not entirely satisfactory, hence a re­newed attempt to work out an idealist reading is justified. The primary challenge of seeing the Vimsika as an idealist text is to ascertain its logical structure: if it is indeed presenting an argument in favor of the idealist position, then what kind of argument is it? Here we believe it is

crucial to note that idealism is equivalent to a negative thesis, that is, the negation of the statement that there are uncognized objects or objects outside of consciousness. In another of his works Vasubandhu is clearly intent on establishing a negative thesis, namely, the ninth chapter of his Abhidharmakosabhasya, where he attempts to prove that there is no self. We un­dertake a detailed analysis of Vasubandhu's argument there and determine that it has the structure of what is known in informal logic as an argumentum ad

igno- rantiam (argument from ignorance): there is no self, because there is no evidence for one. Then, by carefully comparing the Vimsika with AKBh IX we believe we are able to discern an argumentum ad ignorantiam for the conclusion that there are no uncognized objects in that text as well. Along the way, we draw attention to other applications of the argument from ignorance in classical Indian philoso­phy. anupalabdheh, “because it is not apprehended”, is a reason frequently used to prove the non-existence of something, and came to be recognized as one of the three types of valid inferential reason (hetu) by Dharmakirti. We conclude that the Vimsika is an idealist text, after all, which attempts to establish that there are no uncognized objects by means of reasoning that has the characteristics of an argu­ment from ignorance, though Vasubandhu never identifies his argument as such, nor does he even explicitly state its conclusion but allows the reader to draw it for himself. Finally, we speculate about why Vasubandhu might have preferred a less explicit, indirect approach when it came to defending the central Yogacara thesis of “mere-cognition” (vijnaptimatra).

1 Non-idealist readings of the Vimsika

For a sketch of the non-idealist position we shall refer to Hayes 1988, who follows Hall 1986, and Oetke 1992. In limiting the discussion to these scholars we do not mean to lump together all non-idealist interpretations of Yogacara-Vijnanavada; Wayman and Lusthaus, in particular, are concerned with much more than just the Vimsika. In a section titled “Vasubandhu's Phenomenalism” in his pioneering study Dignaga on the Interpretation of Signs, Richard Hayes presents D. N. Shastri as an example of those who have viewed Yogacara-Vijnanavada as a kind of idealism. Shastri writes,

Subjective idealism consists in the assertion that there are no other things than thinking beings; that the things we believe ourselves to perceive are only the ideas of thinking beings.

In short, the theory holds that there is no objective world independent of the perceiving mind In Indian philosophy it is represented by the Yogacara school of the Buddhists. In Western thought, Berkeley is the chief representative.

By contrast, Hayes maintains that what Vasubandhu, for instance, is concerned to show in his Vimsika is simply that our experience can be completely accounted for by factors within consciousness itself, and that it is not necessary to posit any external objects. That, clearly, is not the same as denying that there are any exter­nal objects - objects that exist independently of being cognized.

Vasubandhu is intentionally questioning the assumption that the correct account of experi­ence is that a passively conscious subject experiences directly something entirely outside consciousness itself and is suggesting instead that what there really is is simply an inte­grated experience onto which we project (or out of which we abstract) the notions of per­ceiving subject and object perceived.

Thus, the position being represented in the Vimsika is no more representational- ism - the view that in perceptual experience we are directly aware of ideas or mental representations caused by external objects - than idealism. Here Hayes relies on Hall.

The term vijnapti signifies a “phenomenon” of consciousness, a “manifestation” to con­sciousness, or a “percept” - so long as one bears in mind that these terms should not be taken in a naively realistic or a naively idealistic sense. [.] To translate vijnapti here by “representation” conveys its public aspect, but seems to imply representation of something, presumably an external object or referent, which suggests a “representationalist” theory of knowledge. On the contrary, the purpose of the argument throughout the Vimsatika 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.

We shall briefly address the interpretation of vijnapti as “percept” or “represen­tation” further below. To be sure, Hayes is aware that at the beginning of the treatise Vasubandhu says that the term matra, “nothing but”, in vijnaptimatra, “nothing-but-vijnapti”, is “for the sake of ruling out objects” (arthapratise- dhartham), but according to Hayes this may only mean that “the objective compo­nent of experience is being excluded from consideration” in working out a theory of experience, not that its existence is being denied altogether. Later in the text, in vv. 11-15, which we shall refer to as the ayatana section (which investigates the ayatanas or “sense-spheres”), Vasubandhu appears to be

arguing that physical objects, whether conceived as aggregates of atoms, as wholes, or as single atoms, are impossible. This is a key passage on which other scholars (e.g., Kapstein 1988) have based interpretations of the Vimsika as advocating idealism. But the point of the discussion according to Hayes is, again, not to deny the existence of physical objects but only to show that “reason” cannot postulate them as the causes of our experiences, since however they are conceived, they are logically incoherent. Thus, all features of our experience are to be explained in terms of the elements of experience itself. He writes, again referring back to Hall: “The motivation behind declaring that all experience is nothing but phenomena is, according to Hall, not to make ‘a metaphysical assertion of a transcendent reality consisting of “mind-only”. It is a practical injunction to suspend judgment: “Stop at the bare percept; no need to posit an entity behind it.”'”

Oetke's formulation of the non-idealist position is more concise than that of Hayes-Hall, but he seems to be making much the same point. Oetke maintains that the thesis Vasubandhu is arguing for in the Vimsika is best expressed as,

(T) There are no entities which become the objects of cognition, which can be formulated more precisely as either,

(T1) There are no concrete particulars which are the objects of our experience, or,

(T2) There are no material bodies which are the objects of our experience.

He supports this interpretation with a brief analysis of the ayatana section, which as noted before, is a crucial passage for the idealist interpretation of the Vimsika. He suggests that what Vasubandhu shows in this passage is just that the things we are experiencing are not aggregates of atoms or wholes, i.e., physical objects. Thus, strictly speaking, Vasubandhu is arguing for either (T1) or (T2). But - and this is Oetke's main point - it follows from neither of these statements that there are no physical objects. “These theorems exhibit the formal scheme: ‘There are no F-things which are G' and since propositions of this form do not logically entail ‘There are no G's' or ‘There are no F-things,' both (T1) and (T2) do not strictly imply [...] that there are no material bodies or other particulars of a world of phys­ical things.”

In order for either (T1) or (T2) to entail that there are no material bodies exter­nal to consciousness one would have to supply another premise to the effect that “there are no non-mental particulars which are not objects of our experience”, and Vasubandhu never articulates any such premise. Therefore, the main idea Vasubandhu is arguing for in the Vimsika does not “entail the impossibility of an external world”, or that things exist only insofar as they are perceived, that is to say, subjective idealism.

2 Revisiting the Vimsika

While one could dispute some of the details of Hayes' and Oetke's analyses of the Vimsika, especially when it comes to the ayatana section, their overall impres­sion of the work, that it stops short of a full-throated denial of the existence of external objects, seems correct. The reticence or ambivalence of the Vimsika in regard to external or uncognized objects becomes particularly evident when one compares it to Western presentations of idealism. To bring this out, we consider briefly the locus classicus for subjective idealism in modern Western philosophy, Berkeley's Treatise Concerning the Principles of Human Knowledge.

There is little doubt that Berkeley in his Principles intends to deny the exis­tence of objects outside of consciousness - what he refers to as “material sub­stances.” As is well known, Berkeley felt that “materialism”, which he under­stood to be the positing of material substances, and which he saw in the dominant philosophies of his day, the systems of Descartes and Locke, had to be refuted because it led directly to skepticism and atheism. In the Principles one is pre­sented immediately with arguments to the effect that material substances are in­conceivable and even contradictory. His main argument, developed in the first few paragraphs of the work, is that objects consist of various qualities: colors, shapes, smells, textures, and so on. But such qualities, often called sensible qual­ities, are of such a nature as to exist only for a perceiving mind.

That neither our thoughts, nor passions, nor ideas formed by the imagination exist without the mind is what everybody will allow. And it seems no less evident that the various sensa­tions or ideas imprinted on the sense, however blended or combined together (that is, what­ever objects they compose), cannot exist otherwise than in a mind perceiving them The table I write on exists, that is, I see and feel it; and if I were out of my study I would say that it existed - meaning thereby that if I was in my study I might perceive it, or that some other spirit [i.e., conscious subject] actually does perceive it. There was an odour, that is, it was smelled; there was a sound, that is to say, it was heard; a color or figure, and it was per­ceived by sight or touch.

Thus, since objects such as “houses, mountains, rivers”, and so on are collections of qualities such as color, shape, and smell, they can only exist insofar as they are perceived. “For what are the aforementioned objects but the things we perceive by sense? And what do we perceive besides our own ideas or sensations? And is it not plainly repugnant [i.e., contradictory] that any one of these, or any combina­tion of them, should exist unperceived?”

Whatever argument for idealism Vasubandhu is making in the Vimsika, if he is making any, it is not as explicit or direct as this. Vasubandhu does not launch an all-out, frontal assault, as it were, on the idea of objects existing outside of consciousness. The Vimsika seems to begin, rather - after citation of a scriptural passage which declares that the world consisting of three realms is mere cogni­tion - with the idea that, since we sometimes experience “non-existent objects”, such as hairs floating before the eyes, we

could always be experiencing non­existent objects! Vasubandhu then goes on to show that essential features of our experience, e.g., the fact that certain cognitions are restricted to certain times and places, or that persons in the same place experience objects in the same way, do not require us to postulate physical objects as, say, the causes of our cognitions (vv. 2-7); nor does anything the Buddha said entail that there are such objects (vv. 8-15); nor, finally, is their existence established by perception (vv. 16-17ab). Thus, the overall trend of the treatise seems to be merely that, considering all the available evidence, our experience seems compatible with the non-existence of external objects. It offers no direct proof that they are in fact absent. Even the arguments of the ayatana section appear not to provide any such proof. For, first of all,

those arguments subserve the assertion that scripture does not establish the existence of objects. In other words, that objects are impossible, no matter how one conceives of them, is not what Vasubandhu is primarily trying to prove in this section. Second, if Vasubandhu really were intent on proving that objects are impossible as a way of showing that external objects do not exist - for impos­sibility implies non-existence - and if this were the main point he wanted to make in his treatise, then why didn't he do so at the very outset, as Berkeley does? Finally, if Vasubandhu thought he proved in this section that objects are impos­sible, then other arguments of the treatise, in particular those of the first section to the effect that an external object needn't be posited in order to account for our experience, would be rendered moot. For such objects being impossible, there would be no question of postulating them for any purpose; their absurdity would immediately rule out their playing any role in causing our perceptions.

Thus, the non-idealist reading of the Vimsika cannot be dismissed so easily. It is perhaps not the text of the Vimsika itself that inclines one to resist it so much as the later development of the tradition to which it belongs and the critique of that tradition by outsiders, Buddhist and Brahmin alike. Both clearly depict Yogacara-Vijnanavada as denying the existence of objects outside of conscious­ness, i.e., “external objects” (bahyartha); and so one would expect that the Vimsika, which stands at the beginning of the development of Yogacara- Vijnanavada as a rigorous, coherent philosophical system, would also be defend­ing that position.

There is, however, another consideration that inclines us to resist the non­idealist interpretation and encourages us to reconsider an idealist reading of the text. What, after all, is Vasubandhu trying to say according to the non-idealist interpretation? We would put it as follows: Our immediate awareness in percep­tion is not, as direct or commonsense realism maintains, of material objects, but of sensa, percepts, representations, sense impressions - whatever one wishes to call them - which are private, transitory, mental entities. Hayes and Hall would add that Vasubandhu also suggests that all the features of our sense experience, in particular, that one has particular perceptual experiences at particular times and locations, can be explained by factors that reside within consciousness itself. But if this is what Vasubandhu really shows in the Vimsika - once again, that we can describe and account for our perceptual experience without referring to ma­terial objects - then the obvious question is: What evidence is there for material objects? If the answer is none, then there is no reason to think that they exist. On the non-idealist reading of the Vimsika, then, Vasubandhu may not explicitly say that there are no material objects, but he doesn't have to. The reader is left to draw that conclusion for himself, and it would not seem far-fetched to suggest that that is precisely what Vasubandhu could be intending him to do.

Thus, even the non-idealist interpretation of the Vimsika depicts the text, wittingly or not, as presenting an incomplete argument for the non-existence of objects outside of consciousness, which the reader simply has to complete for him- or herself in one easy step. But it would be rather odd for Vasubandhu, or any author, to leave the final conclusion of his work to the reader in this way. Perhaps then the Vimsika, after all, is somehow arguing for idealism from the outset, though by a different method than Berkeley's? At the very least, we are left with the sense that we do not fully understand what is going on in this text. It merits another look.

3 Reconsidering an idealist reading of the Vimsika : a new approach

In attempting to reconsider the idealist interpretation of the Vimsika, it is doubt­ful one will make much headway just going back over the text. It is true that the critical edition of the text by Sylvain Levi, published in 1925, can still be improved on the basis of a manuscript now available in photographic reproduction, as Bal­cerowicz and Nowakowska have demonstrated in their recent reedition. Harada and Hanneder have, moreover, independently arrived at the conclusion that what we thought to be the first stanza of the text is most likely a “versification” of an argument originally presented in prose. In addition to revisiting philological evidence for the Vimsika itself - including the recently published Sanskrit frag­ments of a commentary by Vairocanaraksita - we might also look for new evi­dence in other writings of Vasubandhu. Schmithausen considers the following works also to have been written by the author of the Vimsika:

Abhidharmakosabhasya Trimsika Karmasiddhiprakarana Pratityasamutpadavyakhya Pancaskandhaka Vyakhyayukti

Now, the Trimsika, Karmasiddhiprakarana, Pratityasamutpadavyakhya, Pancas- kandhaka, and Vyakhyayukti are also considered Yogacara-Vijnanavada works. One might at first be tempted to examine them for evidence that Vasubandhu be­lieved that objects do not exist outside of consciousness. Yet we are doubtful that philological improvements to the text of the Vimsika as well as a re-examination of these writings will bring us any further. Any evidence one might find in them of Vasubandhu's true position regarding the external world is likely to be just as ambiguous as it is in his Vimsika. For it is likely that any statement to the effect that “we are not aware of external objects”, and possibly even any statement to the effect that “there are no external objects”, will be able to be construed phenomenologically, as pertaining just to our experience, i.e., as meaning that the things we are experiencing are not external, physical objects, and not onto­logically, as denying that there are material objects outside of consciousness or asserting that things only exist if they are perceived (esse est percipi). Semantic considerations about the meaning of such terms as “visible form/matter” (rupa), “thing/object/referent” (artha) or “external” (bahya) that occur in such state­ments are helpful, but ultimately insufficient, as the import of such statements can only be reasonably determined by careful scrutiny of the argument strategies that support them.

We shall therefore adopt the following approach. Taking as our hypothesis that Vasubandhu is denying the existence of objects outside of consciousness in the Vimsika, are there any other writings of his in which he is clearly denying the existence of something? If there are, then an effort should be made to determine whether there are any significant similarities between the argument strategies employed in the Vimsika and in those other writings. But no sooner do we ask this question - are there any other works in which Vasubandhu's main purpose is to deny the existence of something? - than the answer springs to mind: the ninth chapter of the Abhidharmakosabhasya, where Vasubandhu refutes the existence of a self. We shall, then, in what follows, offer an analysis of Abhidharma- kosabhasya IX in order to ascertain exactly what kind of argument Vasubandhu develops against the existence of a self there - we would maintain it has gone unrecognized - with a view, ultimately, to determining whether he might be em­ploying a similar kind of argument to prove the non-existence of objects outside of consciousness in his Vimsika.

4 Vasubandhu’s argument strategy for the non-existence of the self in Abhidharmakosabhasya IX

Abhidharmakosabhasya IX is by no means an uncomplicated text. It is an elabo­rate discussion of all the evidence for the existence of a self, or anything resem­bling one (e.g., a “person” or pudgala as upheld by the Vatslputrlyas/Sammitlyas, who are fellow Buddhists), brought forward by other philosophers. We believe that the most important feature to note about it is that it is primarily, if not exclu­sively, critical in its method: it is mainly devoted to refuting other theories. Vasu- bandhu never presents any direct proof that there is no self in the form of a proper inference or anumana.

Vasubandhu announces what we take to be his main argument at the very beginning of the treatise. The question is posed: How is it understood that the designation “self” refers just to a temporally limited section of a series of ska- ndhas (aggregates) and not to something else? Vasubandhu responds, “Because of the absence of perception and inference” in regard to any such thing (pratyaksanumanabhavat). For, he explains, there is, in the absence of any ob­struction, an immediate, perceptual apprehension of those entities or dharmas which exist, for example, of the six kinds of objects of the senses and of the mind (manas). And we are also able to infer things that exist, as for instance when we

infer the existence of the five external senses from the fact that, when some of the causes of perception are present - an object placed in the light and mental atten­tion (manaskara) - a perceptual apprehension of an object sometimes occurs, but at other times it doesn't occur (e.g., for someone who is deaf or blind). Thus, one postulates a sense-faculty as another cause whose presence or absence in con­junction with the other factors brings about a perception. “But it is not like this for the self; [hence] it does

not exist”, Vasubandhu says. In other words, there are no such considerations that would require us to infer or postulate one; there­fore, we can conclude there isn't one. Thus, Vasubandhu's initial argument against the existence of a self appears to be: there is no self because there is no evidence for one! This is a type of argu­ment that in Western philosophy is identified as an argument from ignorance (argumentum ad ignorantiam), and we shall discuss this type of argument in Indian philosophy, or at least certain varieties of it, more generally further below. Our concern at this point is to show that the rest of Abhidharmakosabhasya IX is but the elaboration of such an argument.

Vasubandhu does not initially consider objections to the assertion that there is no perceptual or inferential evidence for a self. He will, in fact, interestingly enough, never take up a serious challenge to the claim that the self is not per- ceived.27 Later in the text he will weigh a series of considerations that other phi­losophers believed compel us to postulate or infer the existence of a self. But in the first part of the text he is occupied with refuting the doctrine of the person (pudgalavada) of the Vatsiputriyas or Sammitiyas.28 Yet this is quite in keeping with the overall strategy of the text to provide an argument from ignorance. A Buddhist may perhaps readily accept that there is no

perceptual or inferential evidence for a self but will still want to know if there is any scriptural evidence. The Vatsiputriyas indeed put forward the pudgalavada as a teaching they be­lieved is implied by key Buddhist doctrines, especially the doctrines of karman and transmigration. They also maintained that the Buddha refers to a pudgala in some of his statements. In order to provide a compelling argument from igno­rance against the existence of a self, then - i.e., that there is no evidence of any kind for a self, neither perceptual nor inferential nor scriptural - Vasubandhu must show that there is no reason to hold that the Buddha ever accepted the exis­tence of anything even like a self, such as a pudgala.

Vasubandhu does not, however, immediately discuss scriptural passages. Rather, he launches into a lengthy attack against the coherence of the pudgala- vada. The pudgalavadins maintain that there is a pudgala which is neither the same nor different from the skandhas, but, precisely because it is not different from them, it is not an eternal self; at the same time, it is the bearer of karman and the entity that transmigrates and attains Nirvana.29 Vasubandhu begins by posing serting something, the ground for one's assertion. Thus, perception or inference both can count as evidence.

27 Though, as we shall see below, he does argue that a pudgala could not be the object of any of the six vijnanas.

28 Henceforth referred to, for convenience, as just the Vatsiputriyas. Vasubandhu refers to the adherents of the pudgala doctrine as “Vatsiputriyas” (AKBh 461,14). Yasomitra (AKVy 699,3) glosses: vatsiputriya aryasammatiyah. Both schools are recorded as advocating pudgalavada. Sources regarding the relationship between Vatsiputriyas and Sammatiyas/Sammatiyas/ Sammitiyas are presented in Kieffer-Pulz 2000: 296f.

29 For a concise summary of the pudgalavada see Eltschinger 2010: 294-296. See also Priestly 1996.

the dilemma: does such a pudgala exist “substantially” (dravyatah) or “nomi­nally” (prajnaptitah)? If the first, then the pudgala would have to be different from the skandhas after all, for it would have its own distinct nature. If the latter, the pudgalavadin just agrees with Vasubandhu: the pudgala has mere nominal existence (461,14-18). To avoid this dilemma the pudgalavadin proposes that “the pudgala is designated in dependence on the skandhas” (skandhan upadaya pu- dgalah prajnapyate) (461,19-20). Vasubandhu proceeds ruthlessly to decon­struct this statement - what, for instance, does “in dependence on” mean? - showing that however it is interpreted it involves the denial of one of the

pudgalavadin's premises; typically, it will entail that the pudgala is either the same or different from the skandhas (461,21-463,9). The Vatsiputriyas also appar­ently held that the pudgala is perceptible. Vasubandhu is therefore also intent on showing that there is no conceivable way it could be cognized by any of the six vijnanas, and by visual cognition in particular (463,10-465,9). The purpose of this last passage concerning the unknowability of the pudgala, we would submit, is not merely to establish that we don't actually have a cognition of a pudgala but, more broadly, that the Vatsiputriya theory of a cognizable pudgala which satisfies the condition of being neither the same nor different from the skandhas is unten­able, hence it is not something the Buddha ever could have taught.

Next, Vasubandhu adduces a number of “explicit” (mtartham, 465,18) scrip­tural passages that suggest that when it comes to talking about a “person” one has only to do with the skandhas, that terms like sattva, nara, manusya, purusa, and pudgala are “mere names, mere expressions” (sanjnamatrakam vya- vaharamatrakam), and that no self is to be found among the dharmas (entities) (465,10-466,17). And he ridicules the idea that these passages, even the famous saying “All dharmas are without self”, are not pramanas for the

pudgalavadin because they are not found in the Vatsiputriya canon (466,17-24)! Finally, Vasu- bandhu considers a series of other scriptural texts and orthodox doctrines that would seem to imply the existence of a pudgala: e.g., the teaching of the omni­science of the Buddha (467,13-468,1), the text that speaks of “the burden, the taking up of the burden, the laying down of the burden, and the carrier of the burden” (468,1-9), the characterization of the denial of spontaneously arising beings (sattva upapadukah) as a

false view (468,10-14), the Buddha's refusal to say whether the living being is the same as the body or not (469,9-24) and to explicitly deny for Vacchagotta that there is a living being or a self at all (469,25­470,18), his statement that “I have no self” is a false view (471,19-23), the doc­trine of transmigration (471,24-472,3), and the Buddha's claims that he remem­bered being this or that person in a previous existence (472,3-7). None of these statements or teachings, Vasubandhu shows, is to be taken as referring to or im­plying that there is a pudgala in the sense understood by the Vatsiputriyas.

We would suggest that the purpose of this lengthy section on the pudgala- vada, which comprises more than half of the work (461-472), is not to refute the pudgalavada per se, but to show that there is no scriptural basis for belief in a self, even in the guise of a pudgala. The rest of the text, then, which takes up arguments made by the Tirthikas (or Tirthakaras), or that are common to both Tirthikas and the Vatsiputriyas, is to show that other rational considerations do not require us to postulate a self, either. This

of course is not quite the same as saying that there is no anumana that proves the existence of a self. Vasubandhu, interestingly enough, never considers a formal anumana to that effect. Neverthe­less, the reasons or grounds for believing in a self that he criticizes are in many instances the very factors that were cited as “inferential marks” (linga) of the self by Nyaya and Vaisesika philosophers.40

Thus, Vasubandhu argues: a self is not required as the factor that combines mental states in the act of memory (472,16-473,14);41 nor is it the one who cog­nizes (473,15-474,9);42 nor could it be otherwise causally responsible for the aris­ing of cognitions (e.g., as that with which the mind comes into contact) (475,1-12) or be the substratum (asraya) of cognitions and other mental states (475,12- 476,3);43 nor is it the referent of the notion “I” (476,4-16)44 or the one who is happy and unhappy (476,16-18);45 nor, finally, is it necessary or even logical to postulate a self as the agent of karman and the enjoyer of the results of karman (476,19­478,13).46 All of these phenomena can be

explained otherwise, without introduc­ing a self, while the hypothesis of a self also is not free from problems.47 Thus, the implication seems to be, none of the reasons cited by other philosophers for infer­ring a self is conclusive. Hence, there is no inferential evidence for a self. In sum, on our reading Abhidharmakosabhasya IX should be interpreted as a lengthy argument from ignorance: (1) There is no perception of a self. As noted, Vasubandhu seems to take this as a given and doesn't feel compelled to argue for

substance (dravyantaram evatmanam), but [in reality] the defilements arise from the grasping of a self.” 40 Thus, e.g., NS 1.1.10: icchddvesaprayatnasukhaduhkhajndnany atmano lingam; “Desire, aver­sion, effort, pleasure, pain, and cognition are the inferential marks of the self.” VS 3.2.4: prana- pananimesonmesajivanamanogatindriyantaravikarah sukhaduhkhe icchadvesau prayatnas cety atmalingani; “Exhaling, inhaling, closing and opening [the eyes], life, the movement of the mind, the change in another sense [when an object is perceived by one sense], pleasure and pain, desire and aversion, as well as effort are the inferential marks of the self.”

41 The opponent asks: yadi tarhi sarvathapi nasty atma katham ksanikesu cittesu ciranubhuta- syarthasya smaranam bhavati pratyabhijnanam va? (472,16-17). 42 The opponent asks: evam ko vijanati, kasya vijnanam ity evamadisu vaktavyam? (473,13). 43 The opponent asserts: avasyam atmabhyupagantavyah smrtyadinam gunapadarthatvat tasya cavasyam dravyasritatvat tesam canyasrayayogad iti cet (475,22-23, emended according to AKBhL 148). Cf. NBhTh 292,2-3; NVTh 391,9-15. 44 atmany asati kimarthah karmarambhah aham sukhi syam aham duhkhi na syam ity eva- marthah? ko 'sav aham nama yadvisayo 'hamkarah? (476,4-5). 45 yady atma nasti ka esa sukhito duhkhito va? (476,16). 46 asaty atmani ka esam karmanam karta kas ca phalanam bhokta bhavati? (476,19).

47 E.g., it cannot really function as a substratum or an agent. In general, na hi kincid atmanah upalabhyate samarthyam ausadhakaryasiddhav iva kuhakavaidyaphuhsvahanam; “For no ca­pacity of the self is apprehended, any more than [a capacity] of the Phuh!’s and Svaha!'s of quack doctors when it is established that the effect has been brought about by herbs” (475,10-11). it (nevertheless, he shows that the pudgala is not cognized by any of the six per­ceptual cognitions), though it will of course be challenged later by Brahminical philosophers. (2) There is no statement of the Buddha affirming a self - to the contrary, there are many statements by which he appears to deny it - nor is there any orthodox teaching that implies its existence. Finally, (3) there is no basis for inferring a self. Therefore, given the total lack of evidence for a self we may conclude that there is none. Of course, in the process of making these points Vasubandhu presents a rich alternative account of facts that supposedly justify the existence of a self, such as memory. But we would argue that this is secondary to his main purpose, which is simply to prove, indirectly by “non-apprehension” (anupalabdhi), that there is no self.

Note that according to our interpretation Vasubandhu is not saying that the proponents of a self have failed to prove one, therefore we are justified in not believing in one - the burden of proof is on them. He is saying something much stronger than this, namely, since there is no evidence for a self, a self does not exist. The principle behind this argument is that if something exists, it will some­how make its presence known; it will be accessible to one of the pramanas. If there is no evidence for something, if no pramana reveals it, then we may con­clude that it does not exist to be revealed.

We have of course been much more explicit about what is going on in Abhidharmakosabhasya IX than Vasubandhu is himself. He offers no summary at the end of the text like the one we have just provided. Therefore, it would be de­sirable if there were other, independent support for this reading. We believe that such support can be found, first, in Uddyotakara's account of the Buddhist objec­tions to the Nyaya arguments for the existence of the self that he discusses in his Nyayavarttika. At NV 1.1.10, Uddyotakara presents the Buddhist assertion that the various phenomena cited in the sutra as the “inferential marks” (lingas) of the self - desire, aversion, etc.

- all of which presuppose memory, can be explained in terms of causal relations between cognitions without bringing in a self as the ele­ment that binds them together. And he (the Buddhist) concludes, “Thus, this con­necting up [of past and present experiences in memory] being possible otherwise, it is not capable of proving the existence of a self.” The idea here is not that there is another better (or simpler) explanation of memory than the one that appeals to a self, but that, given the existence of alternative explanations, one doesn't have to bring in a self to explain memory. Thus, the various lingas listed in 1.1.10 are not conclusive. Uddyotakara's reading of the type of arguments Vasu- bandhu develops in the latter half of Abhidharmakosabhasya IX against various Tirthika theories - we can't be sure he had Vasubandhu specifically in mind, but given discussions of Vasubandhu's ideas elsewhere in the NV it is not unlikely that he did - is consistent with our reading, namely, they are intended to show that there is no conclusive inferential evidence for a self.50

It is, however, even more evident in his commentary on NS 3.1.1 that Uddyo- takara considered one of the main Buddhist arguments against the existence of a self to be an argument from anupalabdhi.51 There, after dispatching the anumana, “There is no self, because it has not arisen” (nasty atmajatatvat), he takes up the argument, also apparently an anumana for him, “There is no self, because it is not apprehended” (nasty atmanupalabdheh). Uddyotakara argues that the reason is not true: the soul is apprehended, for it is the object of the cognition “I”, which appears to be perceptual in nature. After defending this view (which was not shared by all Naiyayikas) at some length he concludes, Thus, to begin with, the self, being the object of the notion “I”, is perceptible. How the self is also apprehended by means of inference has been explained under the sutra “Desire, etc.” [i.e., NS 1.1.10]. There is also scripture [that proves the existence of a self, viz. the Upanisads]. Thus, those three pramanas unitedly, insofar as they all refer to the same thing, prove the self. And there is no pramana that gives rise to the opposite opinion. Hence, “be­cause [the self] is not apprehended” is an unestablished reason.52

This seems to be directed against precisely the type of argument we analyzed Abhidharmakosabhasya IX as presenting. Therefore, we take it as support for our interpretation of Abhidharmakosabhasya IX.

a consideration in favor of the no-self view, though one can perhaps see him arguing in this way at Vs 7. 50 Cf. NVjj, 389,12-13: The Buddhist says, na maya karyakaranabhavat pratisandhanam sadhyate api tv anyathaiva tad bhavatiti hetor asiddharthatadoso 'bhidhiyate. “It is not proven by me that memory is due to the relation of cause and effect [among only cognitions] but rather, [[[memory]]] is possible otherwise [than by postulating an atman]. In this way, [your] reason is indicated as having the fault of unestablished meaning.” 51 See Taber 2012: 107-111. 52 NV 705,13-16: tad evam ahampratyayavisayatvad atma tavat pratyaksah, anumanenapi yatha- tmoplabhyate tathoktam icchadisutra iti. agamo 'py asty eva. tany etani pramanani triny ekavi- sayataya pratisandhiyamanany atmanam pratipadayanti. na ca pramanantaram vipratipattihetur asti. tasmad anupalabdher ity asiddho hetuh. This finding is strongly corroborated by the Yuktidipika. Prior to the discus­sion of Sankhyakarika 17, which gives the reasons for the existence of the self, the author of the YD presents a Buddhist purvapaksa that explicitly states that “there is no self, because it is not apprehendended by any pramana.” Several (uniden­tified) scriptural texts that also appear in Abhidharmakosabhasya IX are cited. In the course of explaining how a self is not apprehended by perception, infer­ence, or scripture the purvapaksin affirms the general principle, “Here [in this context], that which exists is apprehended by one of the pramanas perception, etc., e.g., visible form”, upon which Vasubandhu's argument from anupalabdhi appears to be based.

5 Arguments from ignorance in Indian philosophy

We have been talking about Vasubandhu's argument against the existence of a self in his Abhidharmakosabhasya IX as an argument ad ignorantiam, an argu­ment which derives the non-existence of some entity from a lack of evidence for it, or, in other words, from its non-apprehension. Given how intuitive it is to at least seriously doubt the existence of something when it is not apprehended, it is not surprising that arguments to that effect left their traces in Indian philosophy already at a relatively early date.

One of the driving forces behind the articulation of arguments from igno­rance, and reflection on their probative force, was surely the controversial status of supersensible entities or phenomena which only some assumed to exist. The Carakasamhita argues with materialists who deny rebirth on the grounds that it is not perceived; it lists in the process eight causes for the non-perception of things that are generally perceptible, including their being too far away or too close, the senses or the mind being damaged, or

things being simply too subtle (ati- sauksmya). In the Carakasamhita the argument seems mainly to be driving at an epistemological point: Because there are many things which cannot be perceived at all, and because even things which can be perceived are sometimes not per­ceived due to obstructing circumstances, it is unreasonable to claim that only that exists which is or can be perceived; hence one should not, as the materialists do, rely only on perception, but also consider other means of knowing the existence of things. Yet, such lists of causes for non-perception also were taken to show why things exist in spite of their (occasional or even universal) imperceptibility. In the Sankhyakarika, challenges to the existence of the purusa and unmanifest prakrti resulted in a list of causes for non-perception corresponding to that of the Carakasamhita (SK 7); one can certainly see the focus on inference and on specific proofs for the existence of supersensible entities as reactions to similar challenges.

But we have so far only cited defensive reactions against arguments from ignorance. Indeed, there seem to be not too many sources in which arguments from ignorance are actually propounded, although the few that we have been able to find might nevertheless have been historically influential. In Dignaga’s Nyayamukha, the challenge to the Sankhya supersensibles is condensed into the inference “primordial matter and the like do not exist because they are not appre- hended.” But Dignaga seems to be only interested in this argument as far as one particular logical problem is concerned, namely that of stipulating something non-existent as the subject or property-bearer (sadhyadharmin) in an inference. Dignaga - and, following him, Dharmakirti - addressed this by conceiving of the inferential subject as conceptualized, as simply the meaning of a word.

In the course of Nyayasutra 2.2.13-38, where the question of the eternality vs. non-eternality of sound is discussed, it is asserted by the Naiyayika that sound does not exist prior to being uttered - therefore it must be brought into existence - “because it is not apprehended and no obstruction, etc., is apprehended [either];” hence it must not be eternal. Vatsyayana introduces the sutra by asking, “Moreover, how indeed is it known ‘This exists', ‘This does not exist'? By apprehension and non-apprehension by means of a pramana.” Here we see not only the formulation of a general principle, that non-existence is known through non-apprehension by a pramana (or by pramanas), but also an awareness

that an argument from ignorance must be qualified, if probative: to show that sound is absent one not only has to point to its non-apprehension, but also demonstrate that this non-apprehension is not caused by obstructions - we can take this to be in some measure a response to lists for causes of non-perception as given in the Carakasamhita or in the Sahkhyakarika. Near the beginning of the Nyayabhasya, Vatsyayana makes even more specific claims about the grounds on which some­thing can be proclaimed as non-existent because it is not apprehended: the absence of a pramana proves the non-existence only of such objects that, if they existed, would inevitably be apprehended - a line of thinking that will become central to Dharmakirti's complex theory of the “non-perception of a perceptible” (drsyanupalabdhi) as a separate type of reason in an inference.

What we have detected in the Abhidharmakosabhasya is, on the other hand, not a simple inference, but rather a more involved procedure along the following lines: the apprehension of an entity by means of all sorts of pramanas is consid­ered one by one; each pramana fails to prove the entity in question, and hence it can be regarded as non-existent. Although this procedure is hardly made explicit, its elements are there. Our point, for now, is that Vasubandhu is certainly not the only, nor the first, Indian philosopher to consider the non-apprehension of a type of object as grounds for denying its existence; nevertheless, philosophers at his time had not yet thematized this kind of argument in the context of their theories of inference, or in their theories of pramanas more generally.

It is intriguing to note that Kumarila, who stipulates absence as a separate pramana, later makes use of the same procedure in what in fact seems to be the only application of absence as pramana: he considers whether any of the other five pramanas establish a particular doctrine or entity, concludes that this is not the case, and from this concludes that it is an object of abhavapramana. Kumarila applies this kind of reasoning to the doctrine that cognition is without an external objective basis (i.e., niralambanavada), the idea of smrti as an author­itative type of scripture, the assumption that the Vedas have an author (kartr), and the idea that human beings can be omniscient.

Dharmakirti's immediate predecessor, Isvarasena, apparently held anupa- labdhi to be a third pramana to stand alongside perception and inference, but given that his works are not preserved, it is uncertain what he thought to be its sphere of application, and whether he also envisioned this kind of procedure to be connected with anupalabdhi. Dharmakirti specifies anupalabdhi as a separate type of logical reason in an inference, but in so doing effectively limits its scope so that it cannot prove the non-existence (that is, unreality) of whole types of things. Understood specifically as the non-occurrence of perception,66 anupalabdhi can only prove that things that are not perceived in a situation where their perception, if they existed, would be inevitable, can be cognitively, linguis­tically and physically treated as non-existent.67 But the non-apprehension of things without any further qualification is not evidence for their absence; non­apprehension in this sense only yields the absence of its treatment as existent, because to treat something as existent presupposes its apprehension.68 Subjec­tively, however, when someone does not apprehend a thing it means that for him the object is as good as non-existent.69

But Dharmaklrti also discusses what basically seems to be an argument from ignorance: one that allegedly proves the non-existence of “remote” objects (viprakrsta) on the ground that they are not apprehended by any of the three in­struments of knowledge, perception, inference, and scripture.70 He does not con­sider this kind of argument to be sound. The absence of a scriptural statement proving the existence of something does not prove the non-existence of that thing, since scripture is of a specific kind: it teaches what is relevant for attaining a particular purpose. There are many things that are irrelevant to that purpose,

cf. Kellner 1997: 107, n. 166) he held that the negative concomitance (vyatireka) in an inference is determined by a mere non-perception (adarsanamatra) of the reason where the inferred property is absent, and he also advocated anupalabdhi as a third pramana and understood it as the “mere absence of apprehension” (upalabdhyabhavamatra). But the connection between these two as­signed views is not entirely clear from the accounts known so far.

66 Or, more precisely according to the Hetubindu, as the occurrence of another perception.

67 Dharmakirti's point that non-perception does not prove a thing's absence, but rather justifies treating it as absent (asadvyavahara), simply means that the absence of a thing no longer needs to be proven when its perception is known to be absent (in a situation where all conditions for its perception are met); cf. Kellner 1999. In the same way, that a Simsapa is a tree is not in need of proof when the Simsapa is seen; it is only corresponding cognitive, linguistic, and physical prac­tice - vyavahara - toward the Simsapa as a tree that is then the subject of an inference based on the reason of essential property (svabhavahetu).

68 PV 1.2-3ab=4-5ab (Kellner 2003: 125ff.). 69 PVSV ad PV 1.198=200. 70 PVSV ad PV 1.199=201; PVin 2, p. 62. This continues the discussion of v. 32cd, which begins on the preceding page. Steinkellner 1979: 82, n. 186 regards this to be a rhetorical position because no tradition of Dharmakirti's time advocated three pramanas. This is not entirely true, for the Sankhya advocates precisely these three pramanas, as do some Yogacara-Vijnanavada thinkers, including Vasubandhu. We are not aware, however, of any Sankhya philosopher to advocate a proof that entities do not exist because none of the three pramanas establish them. If our argu­ment that Vasubandhu did pursue such an argument is judged convincing, then Dharmakirti could here well construe the position he refutes on the basis of Vasubandhu. and these are simply outside the reach of scripture. If they are not mentioned in it, surely this does not prove they don't exist.

Second, as remote objects lack the capacity to produce a cognition of them­selves, they are not of the kind that their effects - cognitions - could be observed as evidence for their existence. One can therefore also not conclude that they do not exist when they are not known by any of the three pramanas, for they might exist without giving rise to a cognition of themselves. Non-apprehension in gen­eral is therefore - strictly speaking - not a pramana at all, since it does not result in ascertainment (niscaya) or certainty (vyavasaya). It can negate treating something as existent, but it does not provide certainty in this regard, only doubt. This means that people can act on its basis - for there can also be action based on doubtful cognitions - but it is still not a pramana. In the Pramana- viniscaya, Dharmakirti adds another argument that concerns the scope of non­apprehension. A general non-apprehension by all persons cannot establish anything because such a general apprehension is not known to oneself, nor is it known to anyone else. It is only one's own non-apprehension of an object that one knows. Such a non-apprehension may apply to remote objects - I do not apprehend a fire that might be burning behind a mountain - but it is not the case that these then do not exist at all.

Dharmakirti's description of an argument from ignorance that allegedly proves the non-existence of remote objects seems to coincide with the under­standing of the Buddhist argument for no-self implied by Uddyotakara's sum­mary of his refutation of it, which we have quoted above: “Thus, it is found that all three pramanas” - i.e., perception, inference, and scripture - “establish the existence of the self [.]. Hence, the premise ‘because the self is not-apprehended' is absolutely untrue.”

The consequences of Dharmakirti's elimination of arguments from ignorance from his theory of inference remain to be assessed. It does not seem to have affected the perception of Buddhist arguments against the existence of a self greatly, for these continue to be interpreted as arguments from anupalabdhi by later Brahminical writers, e.g., Udayana, Bhatta Ramakantha, and Utpalade- va. Even in later Buddhist texts, e.g., the Tattvasangraha, the strategy for argu­ing against the existence of a self is chiefly to show that there is no evidence, either perceptual or inferential, that establishes it. But Dharmakirti himself does not argue along these lines. In the scattered passages where he

discusses Brahminical arguments for the existence of a self he primarily considers them as illustrations of the violation of various logical principles. The Nyaya argument “This living body is not without a self because [if it were,] it would follow that it would not have breath”, e.g., is the kind of fallacy that would be allowed if the vyatireka could be established by “mere non-observation.” In the course of his rejection of Nyaya-Vaisesika attempts to infer the self as the agent of phenomena such as breathing or the substratum of mental states - for being imperceptible, Dharmakirti points out, one could never establish a causal relationship between the self and other things - he does note that the Buddhist (typically?) denies the self on the grounds that it is unperceived. But Dharmakirti himself does not adopt such a strategy. Rather, he develops an interesting pragmatic argument in PV 2.220-256: Because a self is necessarily an object of attachment, any belief in a self (even as “pure” and disassociated from body, mind, senses, etc.) will pre­vent liberation.

With the construction and elaboration of complex systems of pramanas, one might expect arguments from ignorance that make use of a pramana framework to also increase considerably in complexity. They have to our knowledge not been addressed in any one particular study. This may be because studies on Indian logic have focused on the atomic inference structures which form the core of the theory of anumana, leaving aside patterns and strategies of argumentation that, for reasons that remain to be investigated, were not subsumed under the rubric of a particular pramana, or a particular type of reasoning and inference explicitly stipulated within some classification.

But is this type of argument a fallacy? Usually, in modern logic textbooks, this kind of argument is considered a fallacy. Its general form has been analyzed as follows: A is not known (proved) to be true (false), therefore A is false (true). Arguments of this type were used to great effect by Senator Joe McCarthy in the Senate Un-American Activities hearings in the early 1950's: “Mr. X is not known not to be a Communist, therefore Mr. X is a Communist.” Yet they are also used, legitimately it would seem, in scientific research. When scientists systematically conduct experiments to detect a certain phenomenon - e.g., the lumeniferous aether that was once postulated as the medium for the

propagation of light - but do not find it, they conclude that it does not exist. It would seem that the argu- mentum ad ignorantiam is a reasonable argument where it functions as a strong abductive argument, i.e., an “argument to the best explanation”: under certain circumstances, the non-existence of something provides the best explanation why there is no evidence for it. In that case, one may be allowed to presume, though not assert, that it does not exist. In such circumstances, one could maintain that the fact that there is no evidence for P is evidence that not-P. Scientists often refer to such absence of evidence as “negative evidence.”

Philosophers have also used arguments from ignorance convincingly, e.g., to claim that God does not exist. We also often employ such arguments in com­monsense reasoning, e.g., to assert that there are no ghosts, no UFO's, no Santa Claus, etc. But whether such arguments are regarded as convincing depends on whether there are agreed-upon standards of verification. It is very difficult to know what would count as good evidence for or against the existence of such things as UFOs, whereas the affirmation or denial of lumeniferous aether relies on established scientific methods. The argument that UFOs do not exist because we do not know or observe them to exist may be fallacious simply

because UFOs raise specific problems regarding the nature of evidence and verification. Not only in the case of UFOs, but more generally, debates about arguments from igno­rance typically come to focus on the question just what it is that could count as evidence for the existence of objects - after all, one might believe that there is no evidence for something simply because one was looking for the wrong kind of evidence, or because one failed to consider something as evidence that should be counted as such. Clearly, when Vatsyayana or Dharmakirti introduce restrictions to the effect that non-apprehension proves the absence of something only if the thing in question would have to be be apprehended were it to exist, they effec­tively aim to narrow down conditions for what counts as evidence for existence. The upshot of Dharmakirti's train of thought, however, seems to be that argu­ments from ignorance can never be convincing because the only kind of evidence for existence, perception, can be lacking only temporarily.

In the next section, then, we shall offer our new reading of the Vimsika, and in the process also bring out parallels between Abhidharmakosabhasya IX and the Vimsika. We believe that this will show that Vasubandhu is arguing along the same lines in both works, by means of an argument from ignorance, against the existence of a self in the Abhidharmakosabhasya, and against the existence of objects outside of consciousness in the Vimsika.

6 A new reading of the Vimsika

In spite of its importance and availability in modern translations the Vimsika has been a relatively neglected text. There exist a few studies of it, to be sure, which mainly focus on vv. 11-15, seen as presenting Vasubandhu's core arguments. Only Kochumuttom's study attempts to analyze the entire text in depth. There seems to be a tendency on the part of scholars to assume that one knows what it means. Yet the text contains many subtleties that have never been noticed, let alone discussed. We, too, cannot offer a complete, in-depth analysis of the text, but we hope, at the very least, to reawaken an interest in it by calling attention to some of its overlooked nuances.

The Vimsika begins with the assertion, “In Mahayana it is established that the world consisting of the three realms is mere cognition (vijnaptimatra)”, fol­lowed by a citation from the Dasabhumikasutra: “Oh Sons of the Conqueror, mere mind (cittamatra) is this world consisting of the three realms indeed.” Although the term vijnapti can be used both for an epistemic act or event - a cognition that makes something conscious or brings it to mind - as well as for the content of that cognition (what Hayes and Hall refer to as “percept” or “phenomenon”), Vasu- bandhu seems to lean towards the event-aspect in the Vimsika, for he states right after this citation that the word citta which is used in it is synonymous with manas, vijnana, and vijnapti.90 The word matra, “mere” or “only”, is “for the pur­pose of denying an object” (arthapratisedhartham). We then have the following verse in the Levi edition:

This is mere cognition indeed (eva), because of the appearance of non-existent objects, like the seeing of non-existent hairs by someone afflicted by floaters.91 Independently of each other, Harada Waso and Jurgen Hanneder convincingly argued that this verse, which is missing from the Tibetan translations and from Xuanzang's Chinese translation but corresponds to prose passages in both, may actually have been fashioned from a prose statement of the Vrtti when a karika- only text was composed,92 for on that occasion it would have become obvious that otherwise the work would abruptly begin with an objection (as stated by the second verse of Levi's edition). In other words, what is now the second verse of Levi's edition might originally have been the first verse, and what is now the first

Dasabhumikasutra 49,10: cittamatram idam yad idam traidhatukam. Schmithausen 1973: 172f. discusses this passage in connection with the development of the vijnaptimatrata doctrine.

90 Schmithausen 2007: 213, n. 2 suggests the translation “representation”, which was also adopted in Yamabe 1998 (without discussion). In Schmithausen's view, “representation” is not only noncomittal as regards the act-/event- or content-distinction, but it also has the advantage of not precluding subconscious mental processes and their contents that are also at times in­tended by vijnapti. But as Hall (1986: 14; see above p. 712) also pointed out, “representation” risks being associated, in philosophical contexts, with the position of representationalism that as­sumes mental states to be (or contain) representations of an independent, external reality. This is not a position to which Vasubandhu subscribes in the Vimsika. Moreover, there is no occur­rence of vijnapti in Vs(V) which unambiguously refers to cognitive content; most occurrences unmistakably refer to cognitive acts or events, and subconscious mental processes and their con­tents are not topical in the Vs(V). We therefore translate “cognition.” This forces us to translate vijnana, jnana, buddhi, and vijnapti with the same word, but in the Vimsika they indeed seem to express the same concept.

91 v. 1: vijnaptimatram evedam asadarthavabhasanat / yadvat taimirakasyasatkesondukadidar- sanam // This is the text according to the Nepalese karika manuscript Vs-ms-A 3a5. For variant readings cf. the overview in Hanneder 2007: 213, and also the earlier discussions in Funahashi 1986 and Harada 2003. For the identification of the timira disease as floaters (or muscae voli- tantes) see Chu 2004: 131, n. 67 (reporting an idea by Anne Macdonald).

verse of Levi's edition might have been fashioned out of a sentence (or a couple of sentences) of the introductory portion of the commentary. Be all that as it may, whether this first statement was originally a stand-alone verse or, in somewhat different wording, part of the commentary, it does not appear that it intends to present a formal anumana that would establish the char­acter of “this” as “mere cognition” by citing a hetu that consists in some property that is invariably connected with being “mere cognition”, as proper anumanas should. Rather, it simply mentions another fact in support of the claim that “this” is mere cognition, namely, that we sometimes

have cognitions of objects that do not exist. The idea seems to be - given the lack of information provided by the text, we have to speculate - that all of our cognitions are structurally indistin­guishable from ones in which were are presented with non-existent objects. Therefore, we are justified in regarding all cognition in the same way, as mere cognition without an object. Now, since this is so weak an argument as not to be considered really an argument at all, it seems most appropriate to interpret this initial statement not as any kind of proof, but rather simply as a statement of the thesis to be proved in the treatise to follow, together with a prima facie rationale for it. The actual proof of the thesis, as we shall see, will be of a much less direct nature.

But is the Vimsika even intended to establish this thesis, or is it simply a de­fense of it against objections? Frauwallner seems to have taken it as the former, but an influential interpretation of the text sees it as the latter. Thus, S. Levi:

Le premier traite, en vingt vers (Vimsika ou Vimsatika), est une sorte d'introduction au sys- teme, plutot critique que constructive. Vasubandhu, avant d'exposer en detail sa propre doctrine de l'idealisme absolu, s'attache a refuter les objections de principe qu'on peut lui opposer a l'interieur de l'eglise bouddhique elle-meme; puis il s'attaque a la theorie ato- mique des Vaisesikas, l'interpretation physique de l'univers la plus puissante que le genie hindou ait elaboree, et qui s'etait insinuee dans le bouddhisme, jusque chez ces Vaibhasi- kas du Cachemire que Vasubandhu avait longtemps suivis avec sympathie. Sa critique de l'atome, ou s'affirme la vigueur de sa dialectique, est restee classique pendant des siecles. L'atome mis hors de cause, ce n'est plus qu'un jeu pour lui de montrer les insuffisances de la these materialiste en general, tandis que les donnees en apparence les plus refractaires a la these idealiste, la memoire, le reve, la mort, s'integrent sans difficulte dans ce systeme.

And D. Shimaji notes, in his “Historique du systeme vijnaptimatra”, contained in the same volume, that the second patriarch of the Chinese vijnaptimatrata school, Hui Zhao, calls, in his commentary on the Cheng weishi lun, the Vimsika “the trea­tise demolishing the mountain of heresy” and the Trimsika “the treatise raising the banner of the Dharma.” We consider this a plausible reading of the text as far as it goes, which is not inconsistent with ours and which does justice, in particular, to the wealth of references to Buddhist theories and concepts that it contains. Certainly, the Vimsika is not a purely philosophical treatise that can be completely taken out of its religious context. However, we believe that, in light of the understanding of the argumentative structure that we are about to work out, the text can also be seen as taking a much stronger dialectical stand - namely, it is an attempt to establish the thesis of vijnaptimatrata, not merely to defend it against objections - which was probably not lost on its readers, both within and outside the Buddhist community.

The second verse, then, immediately expresses an objection to this thesis, which we now paraphrase with the help of the autocommentary: If cognition does not arise from the object (vijnaptir yadi narthatah), then how is there a “re­striction” (niyama) of cognitions to certain times and places? How is there “a non-restriction of the [[[cognition]]] series” (santanasyaniyamah), that is to say, how is it that some cognitions arise for “everyone situated in those times and places” (taddesakalapratisthitanam sarvesam)? Finally, why do things like food or drink experienced in waking cognitions produce the effects that can be expected of them (krtyakriya), whereas the same objects do not produce these effects when we experience them in dreams? None of these facts about our experience seems possible if we were not cognizing physical objects, which indeed are the sorts of things that would be restricted to certain times and places, yet which would be intersubjectively available to all who are present at those times and places and would produce real effects, unlike dream images.

In the famous passage that follows (vv. 3-7) Vasubandhu explains away these facts about our experience by appealing to the phenomena of dreams and the experiences in the hells that Buddhist doctrine envisions. In dreams, he notes, one sees things that appear to be confined to particular times and places, without an external object (v. 3ab'); in other words, dreams are just as spatially and tem­porally specific as waking experiences. Due to the same maturation of their deeds, the spirits of the dead (preta), plagued by hunger and thirst, all see rivers of water as filled with pus and excrement, overseen by frightening guardians (v. 3b'd). Moreover, in dreams we also experience the effects of the

things we see - he cites the example of a nocturnal pollution (v. 4ab'). But it is by experiences in the hells that all the phenomena which the opponent raises as problems are exemplified (v. 4b'c'). The guardians in the hells, along with horrible birds and dogs and moving mountains of iron, are seen in certain places at certain times - and they are seen by all living beings reborn in the hells, not just by a single one. And these guardians inflict pain on the denizens of the hells, and hence produce effects (v. 4c'd). Yet none of these things, the rivers of pus, the guardians, and the other awful creatures of hell, really exist, but they are experienced by the deni­zens of hell in these particular ways “due to the influence of the same ripening of the[ir individual] deeds” (samanakarmavipakadhipatya). Thus, it is possible for “this” world, too, to consist of objects that are restricted in regard to time and place, intersubjectively available, and causally efficacious, and yet also be “mere cognition.”

There are many things about this first section of the text, consisting of vv. 1-7, that merit discussion. One of the most interesting aspects of the passage for our purposes is how Vasubandhu approaches the existence of the hell guardians. Some Buddhist schools such as the Mahasanghikas and Sammitiyas believed these to be real living beings. But, as Vasubandhu argues at first, taking on the position of the Sarvastivada school, just like the horrible dogs and birds that appear to the hell denizens, the hell guardians cannot be living beings because they do not experience the pain that living beings reborn in the hells inevitably feel. Rather, the past deeds of the hell denizens, when they come to fruition, give rise to particular forms of matter (bhutavisesa) that undergo a transforma­tion (parinama), also under the influence of the denizens' karman, and as a result appear as making threatening gestures. The same mechanism also explains the moving mountains and other forms of frightening movement seen in the hells. Hence, it is not true

that hell guardians and other forms of moving matter do not arise (sambhavanti) at all; it is only that they arise through the ripened karman of those living beings that experience them (VsV v. 5). Vasubandhu next criticizes this (Sarvastivada) position: If it is assumed that the hell guardians arise in the hells through the force of karman, why not simply grant that cognition itself transforms in such a way, that is, into images of hell guardians? Why assume that material elements are being produced? (v. 6) In other words, one might just as well opt for a wholly mind-based explanation; nothing forces us to stipulate a causal influence of mind (mental traces of karman) on matter. Surely, with this argument Vasubandhu is not merely bracketing the physical existence of hell guardians; he flatly denies it, for the same reason that we would argue he denies external, physical objects of cognition: there is no evidence for them.

Taken together, the first section of the Vs (vv. 1-7) has on our interpretation for its main point to show that nothing requires us to postulate external objects in order to account for certain facts about our experience. The non-existence of physical objects is on our interpretation implied by some of Vasubandhu's argu­ments - notably, the analogy with the hells - but it is subservient to the larger issue that reason cannot establish the existence of physical objects. For any proof of physical objects would demand that there is evidence for their existence - and this evidence would be all the more convincing if it made their existence neces­sary, if it could not be explained otherwise with equal cogency. The kind of rea­soning which would serve Vasubandhu best if he really wanted to deny the exis­tence of external objects - short of conclusively proving that they

are impossible - is, in other words, of the nature of (non-deductive) inference to the best expla­nation for the lack of evidence for their existence, and not deductive inference. But we saw precisely this kind of reasoning at work in AKBh IX. The strategy of this section of the Vimsika is in fact identical to the strategy of the last part of AKBh IX (472,16-478,13, discussed above p. 724), where Vasubandhu shows that there are no reasons to postulate a self. Similarly, in Vimsika vv. 1-7, the main point is arguably that there are no reasons to postulate external objects of per­ception, as a number of facts that are usually explained by them can also be accounted for through mere-cognition. Thus, one may see this part of the text as the first stage of an argument from ignorance; namely, none of the three accepted pramanas, inference, scripture, or perception, attests to the existence of such objects, therefore they do not exist and all “this” is mere cognition.

The first section ends with a question: Why would one suppose that the im­pressions (vasana) of karman exist in one place, namely, in the cognition series of the agent of the karman, and its fruit in another, namely, in a place called hell? Why would one not suppose that the fruition of karman occurs where the vasana is located, i.e., in the mind itself (v. 7)? To this question, which puts forward yet another argument against the Sarvastivada account of the hell guardians, Vasu- bandhu has his interlocutor answer:

Scripture is the reason. If mere cognition had the appearance of visible form, etc., then there would be no visible form, and so forth, as object. [And] then the Blessed One would not have declared the existence of the ayatanas visible form, etc.

In other words, it is the Buddha's teaching of the ayatanas or sense-spheres, which includes the sense-faculties and their objects, that establishes the exis­tence of physical objects, including the guardians of hell, outside the series of mental cognitions, i.e., outside the mind. It is the word of the Buddha, scripture, that essentially provides evidence for the existence of objects.

In the following verses, vv. 8-15, Vasubandhu sets aside this objection, and hence also the idea that scripture provides evidence for the existence of objects. This would, then, be the second stage of his argument from ignorance. The Buddha did not affirm the existence of the ayatanas as his final position, Vasu- bandhu maintains. Rather, he mentioned them “on account of an intention (abhipraya) concerning the people to be instructed by that [[[teaching]]].” (v. 8bc) What was his intention? To show that there is no self but just

factors (dharma) and their causes, in a word, to demonstrate that the person is without essence (pudgalanairatmya, v. 10ab’). Specifically, all cognitions involve the appearance (abhasa) of a certain form “due to a seed which has attained a particular trans­formation” (svabijat parinamavisesapraptat) within the mental series itself. That seed and that appearance are referred to as the sense-faculty and the object, respectively (v. 9). Ultimately, however, the Buddha will teach that the factors themselves are without essence (dharmanairatmya): there really aren’t any dharmas that have the nature of visible form, etc. And that is to be accomplished by the teaching that all is “mere cognition.” (v. 10b’)

How does one know that the Buddha “taught the sense-spheres with this intention, and that these things which are individually the objects of vision, etc., do not exist?” Here, in vv. 11-15, Vasubandhu develops his famous proof of the impossibility of physical objects of perception. As this passage has already been analyzed by several scholars, and their disagreements do not affect our main line of interpretation, it is not necessary for us to go into details here. Suffice it to say that in v. 11 Vasubandhu gives what would appear to be an exhaustive enumeration of all the ways in which physical objects might exist and rejects each one of them. The ayatana which serves as the object (visaya)

cannot be “one”, i.e., the whole of the Vaisesika. Nor can it be “many according to atoms” (anekam [...] paramanusah), i.e., many atoms, each of which is perceived individually. Nor, finally, can it be “aggregated” (samhatah) atoms (v. 11). He then proceeds, in the following verses, to show why each of these positions is untenable. After demolishing the idea that the object is “one”, i.e., an undivided whole, in v. 15, Vasubandhu concludes that a division into atoms is necessary (VsV 8,18­20); yet he has already shown (v. 12ab, 14ab) that the notion of an atom “as a single substance” is not established (VsV 7,1-2). Having thus exhausted the last hope of making sense of an object of cognition, Vasubandhu concludes, in his autocommentary to v. 15:

That [single atom] not being established, it is not established that visible form, etc., are ob­jects of vision, etc. And so (iti) it results (bhavati) that mere cognition is established.

It would appear from this statement that Vasubandhu thinks that the argument of the ayatana section against the possibility of objects indeed establishes that there are no physical objects, hence that “this” world is mere cognition. Once again, while the advocate of the phenomenalist interpretation could point out that one need only take the passage to be arguing that there are no “material bodies” or “concrete particulars” that we are directly experiencing (cf. Oetke 1992), we are proposing, for now, as a hypothesis that it is arguing that such objects do not exist. In the end, we believe that this hypothesis makes better sense of the Vimsika overall. Yet Vasubandhu now has his interlocutor object (VsV 8,21-23):

Whether something exists or not is ascertained on account of the pramanas. Since percep­tion is the most authoritative of the pramanas, if the object does not exist, how does the cognition arise that something was perceived?

We interpret the meaning of this interesting transitional passage between what we take to be the second and third sections of the treatise in the following way. As announced in v. 8, the section up to v. 15 refutes the idea that scripture offers evi­dence for objects, by pointing out that the Buddha's mention of the sense-spheres is to be understood in such a way that these do not exist as external and physical objects. The arguments in vv. 11-15 undermine scripture as evidence for objects in essentially the same way Vasubandhu's arguments against the pudgala in his AKBh IX undermine the idea that there is scriptural evidence for the existence of a self, namely, by showing that the Buddha could not have taught such a thing because however one conceives of it, it is absurd.

But Vasubandhu now points out that the evidence of perception, which has greater weight than that of the other pramanas, including inference treated in vv. 1-7, seems to establish the existence of objects directly. Therefore, any attempted demonstration of the impossibility of objects will be inconclusive and trumped by perception. If the preceding section had the “proof” of the non-existence of an external world for its main purpose, that would make the entire rest of the treatise superfluous. In fact, it is in this short transitional statement where Vasubandhu enunciates the principle that we believe governs his entire discussion, namely, “Whether something exists or not is ascertained on account of pramanas.” That is to say, something exists if at least one of the pramanas can provide evidence for it; something does not exist if no pramana provides evidence for it. Only after one has shown that none of the pramanas - inference, scripture, or perception - pro­vides evidence for the existence of objects can we safely conclude that there are none.

Now Vasubandhu turns to perception, to show that it, too, really doesn't provide any evidence for the existence of objects. Notice that Vasubandhu is pro­ceeding in the opposite order from AKBh IX. There, after first summarily dismiss­ing perception, he proceeds to show that there is no scriptural evidence for a self, and then demonstrates that there is no inferential evidence for one, either. Here, he first deals with the alleged inferential evidence for objects, then with scripture, and now finally with perception. Moreover, when it comes to the self he invests far more effort in rejecting scriptural evidence than with respect to external ob­jects. This difference in procedure may have something to do with the fact that, when it comes to the existence of a self, it was scriptural evidence that was of paramount importance for the Buddhists; for there was a large and

influential group of Buddhists who thought that the existence of a pudgala was sanctioned by the Buddha himself. Therefore, scriptural evidence needed to be addressed more extensively, and prior to reason. When it came to the existence of an exter­nal world, on the other hand, the relative importance of the evidence may have seemed different to Vasubandhu. Buddhist scripture, meanwhile, does not make very many clear pronouncements about the existence or non-existence of an ex­ternal world, compared to the scriptural

statements that refer to persons. Indeed, it is rational objections that first come to mind when someone suggests that we are not really experiencing objects: How, then, do you explain the impression that our experience is not self-generated, but dependent on factors outside us, or that specific experiences are restricted to certain places and certain times, etc.? For these reasons, one may speculate, in the Vimsika it is the evidence of reason that Vasubandhu tackles first, then scripture, then perception. Long-standing habits of argumentation within Buddhist doctrinal literature may also have played a role, since examining doctrine through reasoning and scripture, yukti and agama, was a well-established method, whereas the pramanas - including per­ception - were by comparison a more recently developed conceptual framework.

The evidence that perception might offer for external objects is its mere occurrence, as it is subsequently known: if one knows that one just perceived something, surely this would be evidence for the existence of a perceived object. So how is then the cognition that an object was perceived possible without the existence of an object? In v. 16 Vasubandhu declares that this cognition indeed arises in the same way as dream cognitions. Moreover, the idea “this object was perceived by me” (idam me pratyaksam) arises at a

time when no external object is seen, for it is presented by a mental cognition (manovijnana) which occurs after the sensory cognition has already disappeared - and this time-gap becomes even more pertinent if it is assumed that an alleged external object would be momen­tary, as this object would then already have disappeared at the time of its percep­tion. Nor does memory establish a previous experience of the object remembered, according to the principle that one can only remember something one has previ­ously experienced; for Vasubandhu has shown in the course of his discussion (in the ayatana section) that a cognition can arise possessing the appearance of an object even in the absence of an external object. That kind of cognition can serve as the basis of a later memory as if of a previous experience of an external object (v. 17ab).

Through v. 21 Vasubandhu deals with a final series of altogether five objec­tions, voiced as critical questions. The first (VsV 9,8-10) concerns the analogy of ordinary experience to dreams. If we are experiencing objects as in a dream, why don't we realize their non-existence on our own, as we do for dreams? But, he responds, as long as one does not wake up one does not comprehend the non-existence of what one is experiencing (v. 17cd). In the same way, when one becomes awakened (prabuddha) by attaining the highest non-conceptual or transconceptual cognition (nirvikalpakajnana), one then realizes the absence of objects through the “pure mundane insight” (suddhalaukikajnana) that follows after transconceptual cognition (VsV 9,14-16). We note here in passing that this objection and the response to it, as well as the others to follow, are prima facie very difficult to reconcile with a phenomenalist interpretation of the Vimsika.

The second objection (VsV 9,16-19) attaches to the restriction of cognitions according to time and place solely due to a specific transformation of one's own mental series (svasantanaparinamavisesa), to which Vasubandhu has alluded in v. 9. How can there then be any distinction of cognitions due to such causes as the association with sinful or virtuous people, or the instruction in true or false teach­ings, if indeed neither such association nor such instruction exist? Vasubandhu responds (v. 18ab): Such distinctions occur because different mental series influ­ence each other (anyonyadhipatitva), that is, causally affect each other. A specific cognition belonging to one series then arises from a specific cognition belonging to another series - a virtuous person or a sinful person, a true or a false teacher - not from an external object. Third, if both waking and dreaming cognitions are without external objects, why does wholesome or unwholesome conduct not bring about the same effect for those who are asleep and those who are awake? (VsV 9,23-25) Answer: In a dream, the mind is afflicted by torpor (middha), hence the difference from a waking cognition (v. 18cd). Fourth, if all “this” is indeed mere cognition, there is neither body nor speech. Why, then, do sheep die when killed by slaughterers? Or if their death is indeed not caused by the slaughterer, how can the slaughterer then be subject to the fault of taking life (pranatipata)? (VsV 9,23-10,2). Here, too, Vasubandhu invokes influence by another person's cognition (vv.19-20): the death of the sheep is due to a specific cognition in an­other mental series (the slaughterer) which results in an obstruction of the vital faculty (jivitendriya), causing the sheep's mental series to cease reproducing itself. Vasubandhu supports this idea - a fully mind-based account of karmic efficacy and human life - by pointing to the example of how a demon's mental powers might cause a person to lose their memory or be possessed. Another example is drawn from Buddhist scripture, from the Upalisutra, where the Buddha claims - against the Jainas - that acts of violence (danda) by force of the mind entail more serious offences than those carried out by force of body and speech. In supporting his claim, the Buddha recalls how rsis or other holy men enraged by kings destroy large areas through rain of stones and fire, solely by force of their thoughts, thus killing many living beings. Surely, the example works only if the destruction is indeed mentally caused, and not brought about through some other kind of cause. Hence these examples lend scriptural support to the claim that cognitions, mental acts, can result in the taking of life, which, though not a physical matter, is nevertheless real in terms of having actual karmic consequences.

The fifth and last objection asks whether, if “this” is only mere-cognition, one can know other minds (VsV 10,19-21) - the Chinese translations supply a ratio­nale behind this question, absent from the Sanskrit and Tibetan versions: if one can indeed know other minds, then this would invalidate the doctrine of mere-cognition. The underlying assumption seems to be that this would be a case where a cognition knows something external to itself (or existing inde­pendently of the mental series to which it belongs), and that this is in contradic­tion with mere-cognition. Vasubandhu declares in his reply that knowledge of other minds does not truly apprehend them as they are, just like knowledge of one's own mind does not apprehend it as it really is - in its true inexpressible nature, which can solely be known by Buddhas; for ordinary knowledge has not yet abandoned the subject-object-dichotomy (v. 21). Vasubandhu does not deny that this is indeed a case where a cognition knows something “external”, but merely points out that this knowledge is not real knowledge, as far as ordinary people are concerned. This suggests that as far as different mental series are concerned, the idea that something external (a mental event of series A) causes a cognition (in mental series B) having that cause's image is acceptable. What would make this account, which in general form would later become identified as a Sautrantika position, unacceptable for external phys­ical objects is that we have no evidence for their existence, given the arguments Vasubandhu adduced before. Interpreted in this way, in fact, the response to the objection on other minds would lend additional support to the view that the Vimsika is not just about denying the cognition of external reality, but

more spe­cifically about denying the existence of physical objects of experience. This declaration leads over to the concluding verse 22: This establishing of mere-cognition has been carried out by me, according to my abilities. But this [mere-cognition] is not conceivable in all its aspects; it is the domain of the Buddha.

As he explains in the commentary, mere-cognition cannot be conceived (cintya) in all of its aspects by people like himself, “for it is not an object of reasoning” (tarkavisayatvat) - of the kind of reasoning called tarka which in many Buddhist sources is associated with limited cognitive faculties, often characteristic of Tirthikas, of non-Buddhist “outsiders” (bahyaka). The true nature of reality is ultimately inaccessible to such reasoning. In the Trimsika, Vasubandhu states (vv. 26-28): As long as cognition does not abide in mere-cognition the burden of the duality of appre­hension does not cease.

For even through the apprehension, “[All] this is mere cognition”, one does not abide in that [[[cognition]]] alone, because one [still] places something before oneself. If, on the other hand, cognition does not apprehend an object-support, then it stands firm in being mere cognition because, due to the absence of that [[[object]]] which is grasped, there is [also] no grasping of that [[[cognition]] which would grasp the object]. The final argument is reminiscent of a pattern of realization comprising several stages that can be traced in Yogacara-Vijnanavada works ascribed to Asanga or Maitreya, and also in Vasubandhu's own Trisvabhavanirdesa: There is an initial apprehension of mere-cognition, still tainted by duality. Realizing that when cognition does not apprehend external objects, it is also not something which apprehends - a grahaka - a deeper insight arises which Vasubandhu in the Trimsika identifies as cognition “standing firm in being mere cognition.” As he clarifies in Trimsika v. 25, this state for him amounts to the realization of suchness, or tathata. The Trimsika thereby suggests that there are different stages in the (meditative) realization of the vijnaptimatrata, an initial “apprehension” and a fuller and firmer “abiding.” We do not have to explore the complex edifice of Yogacara-Vijnanavada soteriology any further, for the point seems clear enough: there are certain aspects of mere-cognition as “suchness” which are only realized in a higher meditative state, the nirvikalpasamadhi or transconceptual meditation.

This consideration may also provide us with an ulterior rationale for the negative argument strategy Vasubandhu pursues in his Vimsika - to be placed alongside the fact we find appeals to arguments from ignorance elsewhere in Indian philosophy and that Vasubandhu had employed the same strategy pre­viously himself, in AKBh IX. Aware that one's object of proof - mere-cognition - has aspects that are inaccessible to argument, one will be hesitant to try to prove it directly. One will be more confident, rather, in showing that the negation of one's thesis is false simply from the fact there there is no evidence for it. For pre­sumably, if it were true, there would be some evidence for it. Thus, by implica­tion, mere-cognition is established as true. In conclusion, we have attempted to demonstrate that the elements of an ar- gumentum ad ignorantiam are present in the Vimsika, and even more: that these elements provide the treatise with a structure and a strategy. The overall plan of the treatise is negative. There is no clear statement of an anumana establishing mere-cognition at the outset (v. 1 or its prose equivalent). Each of the pramanas, inference, scripture, and perception, is considered in turn in three sections, re­spectively, vv. 1-7, 8-15, and 16-17ab. For each pramana the negative conclusion is reached that it does not provide evidence for the existence of objects. Even the proof of the impossibility of objects within the ayatana section is not meant to stand on its own, but it subserves the point that there is no scriptural evidence for them. The overall strategy of the text is epistemological, we could say, not meta­physical. It considers for each pramana what it can prove; is it powerful enough to establish the existence of things that are causing our cognitions? And in each case it answers, no. While discussing five critical questions in VsV 9,8-10,28 (vv. 17cd-21), Vasubandhu also puts forward arguments that contain a denial of the existence of external physical objects, whereas he regards an account of cogni­tion and karmic retribution which only posits different mental series that causally interact with each other to be acceptable; this lends additional support to an idealist reading of the text.

Careful not to offend against the Mahayana doctrine that the true nature of reality can only be known in nirvikalpasamadhi, Vasubandhu refrains from stating his conclusion himself. He leaves it for the reader to draw the conclusion, in accordance with the principle he has enunciated in the course of his discussion, “Whether something exists or not is ascertained on account of the pramanas”, namely: objects outside of consciousness do not exist. This does not mean, however, that the physical entities we know as visible form, etc., do not exist at all, Vasubandhu is also careful to say (VsV 6,14-21). Dharmanairatmya does not mean that dharmas do not exist. Rather, it means that dharmas are without the nature they are imagined to have by the unenlightened, as grahya, grahaka, etc. Dharmas exist in their inexpressible nature (anabhila- pyenatamana), which is known only by the Buddhas. But it is by the teaching of mere-cognition - that dharmas such as visible form, in particular, do not stand as entities over against, and independently of, cognition - that one realizes the dharmanairatmya, which culminates in the (transconceptual) comprehension of their inexpressible nature. It is, in short, not only possible to detect the pattern of an argumentum ad ignorantiam in the Vimsika - just like such a pattern can be detected in Abhidharmakosabhasya IX - but also to indicate a reason for why Vasubandhu might have chosen this argument as a strategy for the specific case of establishing mere-cognition.

Appendix: the refutation of the pudgala in Asanga’s Mahayanasutralankara

The refutation of the pudgala as a real substance (dravyatah) in Asanga's Mahayanasutralahkara might similarly be viewed to express an argumentum ad ignorantiam: The pudgala must be said to exist as a [mere] designation (prajnaptyastitaya) but not as a [real] substance (dravyatah), because one does not apprehend [it] (nopalambhat) [...] (MSA 18.92ac’).

Vasubandhu, the presumed author of the Mahayanasutralankarabhasya, explains: [[[Pudgalavadin]]:] But how can one know that this [[[pudgala]]] does not exist as a [real] sub­stance? [Answer:] Because one does not apprehend [it] [MSA 18.92c’]. Indeed, contrary to [[[dharmas]]] such as visible [things], this [[[pudgala]]] is not perceived as a [real] substance. [The Pudgalavadin:] But what is called “apprehension” [also consists in] a cognition by the intel­lect (buddhi). Now, it is not the case that [we] Pudgalavadins do not cognize the pudgala through the intellect. Moreover, the Blessed One has said: “In this very life, [the living being] perceives an atman, designates [an atman].”

Although “a cognition by the intellect” could refer to a mental awareness (manovi- jnana), hence a kind of perception, as Sthiramati seems to interpret it, it could also refer to inference, one of the two pramanas capable of establishing an entity as a real substance (*dravyasat), as Sthiramati also acknowledges. Thus, Asanga’s assertion “because one does not apprehend [it]” should perhaps be taken to mean, because one does not apprehend it at all - not only perceptually but also inferentially and, as the passage further suggests, scripturally; for the Pudgalavadin, as Vasubandhu represents him, reacts to Asanga’s assertion by citing both “cognition by the intellect” and scripture as support for his view.

On this background, one might argue that Vasubandhu extends this strategy “because one does not apprehend it”, limited to the refutation of the pudgala in his Mahayanasutralankarabhasya, to the refutation of a self more generally in AKBh IX. Nevertheless, later in his MSA commentary Vasubandhu states: So far (evam tavat), [it is] by resorting to reason(ing) [alone that it has been demonstrated that] the pudgala is not apprehended (nopalabhyate) as a [real] substance. And [this can also be demonstrated by resorting to scripture,] because [the Blessed One has] taught [that] all dharmas are selfless, [that] ultimately [there is nothing but] emptiness and [that] to per­ceive a self is harmful (atmopalambhe dosah) [MSA 18.101].

This does seem to suggest a more direct strategy, namely, the pudgala is disproved by various arguments Asanga has brought forward - it cannot be a real thing if, as the pudgalavadin maintains, it is neither the same as nor different from the skandhas (311-316); it cannot function as the “seer”, nor as an agent (317-322) - and by scriptural passages that imply that there is no self or that it is a pernicious error to believe that there is one (322-325). Still, the arguments in question have the nature of refutations of points typically made in favor of a pudgala, while the scriptural passages are cited in anticipation of references to scripture (e.g., the Bharaharasutra) the pudgalavadin will go on to make in support of his view (325). The overall strategy of the discussion of the pudgala in MSA(Bh) still seems to be primarily indirect or negative; that is to say, it rejects the existence of a pudgala on the grounds that there is no evidence for one.

Acknowledgements: Birgit Kellner's research for this article was conducted within the research project D16, “Reasoning in South Asian and Tibetan Buddhism”, supported by the Cluster of Excellence “Asia and Europe in a Global Context” at the University of Heidelberg.

Bibliography

Primary sources in Sanskrit and in Tibetan translation
AK(Bh) - Vasubandhu's Abhidharmakosabhasya. Abhidharmakosabhasyam of Vasubandhu. Edited by P. Pradhan. 2nd ed., rev. by A. Haldar. Tibetan Sanskrit Works 8. Patna: K. P. Jayaswal Research Institute, 1975.

AKBhL - Vasubandhu's Abhidharmakosabhasya. Abhidharmakosabhasya of Vasubandhu, ChapterIX: Atmavadapratisedha. Edited by Jong Cheol Lee. Bibliotheca Indologica Buddhologica 11. Tokyo: Sankibo Press, 2005.
AKVy - Sphutartha Abhidharmakosavyakhya by Yasomitra. Edited by Unrai Wogihara. Part II: IV-VIII Kosasthanam. Tokyo: The Publishing Association of Abhidharmakosavyakhya, 1932-1936.

ATV - Udayana's Atmatattvaviveka. Atmatattvaviveka, with the Commentaries of Sankara Misra, BhagTratha Thakkura, and Raghunatha TarkasiromanT. Edited by Vindhyesvariprasada Dvivedin and Lakshmana Sastri Dravida. Calcutta: The Asiatic Society, 1986 [1907-1939 (sic)].

Bhavanakrama - KamalasTla's first Bhavanakrama. In: Minor Buddhist Texts, part II. First Bhavanakrama of KamalasTla. Sanskrit and Tibetan Texts with Introduction and English Summary. Edited by Giuseppe Tucci. Roma: Istituto Italiano di Medio e Estremo Oriente, 1958.

Carakasamhita - The Caraka Samhita of Agnivesa Revised by Caraka and Drdhabala With the Ayurveda-DTpika Commentary of Cakrapanidatta And with ‘VidyotinT Hindi Commentary By Pt. KasTnatha SastrT. Edited by Gangasahaya Pandeya. 2nd ed., 2 vols. Kashi Sanskrit Series 194. Varanasi: Chowkhamba Sanskrit Series Office, 1983.
Dasabhumikasutra - Dasabhumikasutra et Bodhisattvabhumi. Chapitres Vihara et Bhumi. Edited by J. Rahder. Paris: P. Geuthner, 1926.

DhP - Dhammapada. Edited by O. von Hinuber and K. R. Norman. Oxford: Pali Text Society, 1994.

Kathavatthu - Kathavatthu. Edited by A. C. Taylor. London: Pali Text Society, 1894-1897. Lankavatarasutra - The Lankavatara sutra. Edited by Nanjo Bun'yu. Kyoto: Otani University Press, 1956.

MBh - The Vyakarana-Mahabhasya of Patanjali's. Edited by F. Kielhorn. 3rd revised ed. by K. V. Abhyankar. Poona: Bhandarkar Oriental Research Institute, 1965 [1883].
MSA - Mahayana-Sutralamkara: expose de la doctrine du grand vehicule, selon le systeme Yogacara. Edited by S. Levi. Vol. I: Texte. Paris: Champion, 1907.
MSABh - Vasubandhu's Mahayanasutralankarabhasya. See MSA.


MV - Madhyantavibhaga-bhasya: a Buddhist philosophical treatise edited for the first time from a Sanskrit manuscript. Edited by G. Nagao. Tokyo: Suzuki Research Foundation, 1964.
NB - DharmakTrti's Nyayabindu. Durvekamisra, DharmottarapradTpa [being a subcommentary on Dharmottara’s NyayabindutTka, a commentary on DharmakTrti’s Nyayabindu]. Edited by Dalsukhbai Malvania. 2nd rev. ed. Tibetan Sanskrit Works 2. Patna: K. P. Jayaswal Research Institute, 1971.

NBh - Vatsyayana's Nyayabhasya. See Nyayadarsana.
NBhTh - Vatsyayana's Nyayabhasya.
Nyayadarsana of Gautama, with the Bhasya of Vatsyayana, the Varttika of Uddyotakara, the TatparyatTka of Vacaspati, and the Parisuddhi of Udayana. Volume 1 Chapter 1. Edited by Anantalal Thakur. Darbhanga: Mithila Institute, 1967.
NS - Nyayasutras. See Nyayadarsana.

NV - Uddyotakara's Nyayavarttika. See Nyayadarsana.

NVTh - Uddyotakara's Nyayavarttika. See NBhTh.
Nyayadarsana - Nyayadarsanam with Vatsyayana's Bhasya, Uddyotakara's Varttika, Vacaspati Misra’s TatparyatTka and Visvanatha’s Vrtti. Edited by Taranatha Nyaya-Tarkatirtha and Amarendramohan Tarkatirtha. Calcutta: Munshiram Manoharlal, 1985 [1936-1944].
PV 1 - DharmakTrti's Pramanavarttika, chapter 1 (svarthanumana). See PVSV.


PV 2 - DharmakTrti's Pramanavarttika, chapter 2 (pramanasiddhi). In: “Pramanavarttika-Karika (Sanskrit and Tibetan)”. Edited by Miyasaka Yusho, Acta Indologica 2 (1971/1972).
PVin 2 - Pramanaviniscaya 2. See Steinkellner 1979.

PVSV - Pramanavarttikasvavrtti. The Pramanavarttikam of DharmakTrti: the First Chapter with the Autocommentary. Edited by Raniero Gnoli. Serie Orientale Roma 23. Rome: Istituto Italiano per il Medio ed Estremo Oriente, 1960.

SK - Sankhyakarika, see YD.
SN - Samyuttanikaya. The Samyutta-Nikaya of the Sutta-Pitaka. Edited by L. Feer. London: PTS, 1884-1898.
Trisvabhavanirdesa - “Vasubandhu's Trisvabhavanirdesa”. Edited by Louis de la Vallee Poussin, Melanges Chinois et Bouddhiques 2 (1932/1933): 147-161.
Trs - Vasubandhu's Trimsika Vijnaptimatratasiddhih. See Vs.

VN - DharmakTrti’s Vadanyaya. TeilISanskrit-Text. Edited by Michael Torsten Much. Vienna: Verlag der Osterreichischen Akademie der Wissenschaften, 1991.
VS - Vaisesikasutra of Kanada, with the commentary of Candrananda. Edited by Muni SrT Jambuvijayaji. Baroda: Oriental Institute, 1961.
Vs - Vasubandhu's Vimsika Vijnaptimatratasiddhih. In: Vijnaptimatratasiddhi: DeuxTraites de Vasubandhu: Vimsatika et Trimsika. Edited by Sylvain Levi. Paris: Libraire Ancienne Honore Champion, 1925.
Vs-ms-A - Vimsika, Nepalese karika manuscript. National Archives, Kathmandu, Nepal, Ms. No. 1-1697. Manuscript A. In: Mimaki et al. 1989.
VsV - Vasubandhu's Vrtti on Vs. See Vs.

YD - YuktidTpika. Edited by Albrecht Wezler and Shujun Motegi. Alt- und Neu-Indische Studien 44. Stuttgart: Franz Steiner Verlag, 1998
Primary sources in English and secondary sources
Ayer, A. J. (1956): The Problem of Knowledge. London: Macmillan.

Balcerowicz, Piotr and Monika Nowakowska (1999): “Wasubandhu: ‘Dowod na wylaczne itsnienie tresci swiadomosci w dwudziestu strofach' (Vasubandhu: Vimsatika - Vijnapti-matrata-siddhi). Opracowanie tekstu i przeklad z sanskrytu”. Studia Indologiczne 6: 5-44.

Berkeley, George (1970): A Treatise Concerning the Principles of Human Knowledge, with Critical Essays. Edited by Colin Murray Turbayne. Indianapolis: Bobbs-Merrill.
Buescher, Hartmut (2008): The Inception of Yogacara-Vijnanavada. Vienna: Verlag der Osterreichischen Akademie der Wissenschafen.
Chu, Junjie (2004): “A Study of Sataimira in Dignaga's Definition of Pseudo-Perception (PS 1.7cd-8ab)”. Wiener Zeitschrift fur die Kunde Sudasiens 48: 113-149.
Deleanu, Florin (2006): The Chapter on the Mundane Path (Laukikamarga) in the Sravakabhumi. A Trilingual Edition (Sanskrit, Tibetan, Chinese), Annotated Translation, and Introductory Study. Tokyo: The International Institute for Buddhist Studies.

Duerlinger, James (2003): Indian Theories of Persons: Vasubandhu's “Refutation of the Theory of a Self”. London/New York: RoultledgeCurzon
Eltschinger, Vincent (2010): “On a Hitherto Neglected Text Against Buddhist Personalism: Mahayanasutralankara 18.92-103 and its Bhasya”. Asiatische Studien/Etudes Asatiques 64/2: 291-340.

Eltschinger, Vincent / Ratie, Isabelle (2013): Self, No Self, and Salvation: DharmakTrti's Critique of the Notions of Self and Person. Vienna: Verlag der Osterreichischen Akademie der Wissenschaften.
Feldman, Joel (2005): “Vasubandhu's Illusion Argument and the Parasitism of Illusion upon Veridical Experience”. Philosophy East and West 55/4: 529-541.
Flew, Antony (symposium contribution, with R. M. Hare, Basil Mitchell, and I. M. Crombie) (1955): “Theology and Falsification”. In: New Essays in Philosophical Theology. Edited by Antony Flew and Alasdair MacIntyre. New York: Macmillan, 96-99.

Franco, Eli / Preisendanz, Karin (2010): “Vorwort” to Erich Frauwallner: Die Philosophie des Buddhismus. Berlin: Akademie Verlag. 5th reprint, XI-XXX.
Frauwallner, Erich (1951): On the Date of the Buddhist Master of the Law Vasubandhu. Rome: Istituto Italiano di Medio e Estremo Oriente.
Frauwallner, Erich (1961): “Landmarks in the History of Indian Logic”. WienerZeitschrift fur die Kunde Sud- und Ostasiens 5: 125-148.
Frauwallner, Erich (2010): The Philosophy of Buddhism. Translated by Gelong Lodro Sangpo and Jigme Sheldron under the supervision of Ernst Steinkellner. Delhi: Motilal Banarsidass.

Funahashi, Naoya (1986): “Textual Notes on the Trimsikavijnaptibhasyam based on the Comparison of Nepalese Manuscripts and a Study of Verses 1 and 2 of the Vimsatika”. Buddhist Seminar 43: 15-30.

Hall, Bruce Cameron (1986): “The Meaning of vijnapti in Vasubandhu's Concept of Mind”. Journal of the International Association of Buddhist Studies 9/1: 7-23.
Hamilton, Clarence Herbert (1938): Wei Shih er Shih Lun. The Treatise in Twenty Stanzas on Representation-only by Vasubandhu. New Haven, Connecticut: American Oriental Society.
Hanneder, Jurgen (2007): “Vasubandhus Vimsatika 1 2 anhand der Sanskrit- und tibetischen Fassungen”. In: Indica et Tibetica. Festschrift fur Michael Hahn, zum 65. Geburtstag von Freunden und Schulern uberreicht. Edited by Konrad Klaus and Jens-Uwe Hartmann. Vienna: Arbeitskreis fur Tibetische und Buddhistische Studien, 207-214.
Harada, Waso (2003): “A Philological and Doctrinal Reconsideration of the Vimsatika Vijnaptimatratasiddhi, Part 3”. Bulletin of Kyushu Ryukoku Junior College 49: 131-188 (in Japanese).

Hayes, Richard P. (1988): Dignaga on the Interpretation of Signs. Dordrecht: Kluwer Academic Publishers.
Hirst, R. J. (1967): “Phenomenalism”. In: The Encyclopedia of Philosophy. Edited by Paul Edwards. New York: Macmillan. Vol. 6, 130-135.
Hugon, Pascale (2008): “Arguments by Parallels in the Epistemological Works of Phya pa Chos kyi seng ge”. Argumentation 22: 93-114.
Kano Kazuo (2008): “Two Short Glosses on Yogacara Texts by Vairocanaraksita: VimsikatTkavivrti and *Dharmadharmatavibhagavivrti”. In: Manuscripta Buddhica, Vol. I: Sanskrit Texts from Giuseppe Tucci's Collection, Part I. Edited by Francesco Sferra. Roma: Istituto Italiano per l'Africa e l'Oriente, 343-380.
Kapstein, Matthew (1988): “Mereological Considerations in Vasubandhu's ‘Proof of Idealism' (Vijnaptimatratasiddhih)”. Idealistic Studies 18/1: 32-54.
Katsura, Shoryu (1978): “Inmyo shori mon ron kenkyu 2 {A Study of the Nyayamukha, part 2}”. Hiroshima Daigaku Bungakubu Kiyo 38: 110-130.
Katsura, Shoryu (1992): “Dignaga and DharmakTrti on adarsanamatra and anupalabdhi”. Asiatische Studien/Etudes Asiatiques 46: 222-231.
Kellner, Birgit (1996): “There Are no Pots in the Slokavarttika. SV abhavapariccheda 11 and Patterns of Negative Cognition in Indian Philosophy”. Journal of the Oriental Institute (Baroda) 46/3-4: 143-167.

Kellner, Birgit (1997): Nichts bleibt Nichts. Vienna: Arbeitskreis fur Tibetische und Buddhistische Studien, Universitat Wien.
Kellner, Birgit (1999): “Levels of (Im)perceptibility. Dharmottara on the drsya in drsyanupalabdhi”. In: Dharmakirti’s Thought and Its Impact on Indian and Tibetan Philosophy. Proceedings of the DharmakTrti Conference Hiroshima, November 4-6,1997”. Edited by Shoryu Katsura. Vienna: Verlag der Osterreichischen Akademie der Wissenschaften, 193-208.
Kellner, Birgit (2003): “Integrating Negative Knowledge into Pramana Theory: The Development of the drsyanupalabdhi in DharmakTrti's Earlier Works”. Journal of Indian Philosophy 31/1-3: 121-159.
Kieffer-Pulz, Petra (2000): “Die buddhistische Gemeinde”. In: Der Buddhismus I. Der indische Buddhismus und seine Verzweigungen. Edited by Heinz Bechert et al. Stuttgart: Kohlhammer, 281-402.
King, Richard (1998): “Vijnaptimatrata and the Abhidharma Context of Early Yogacara”. Asian Philosophy 8: 5-13.
Kobayashi, Hisayasu (2011): “On the Develoment of the Argument to Prove vijnaptimatrata”. In: Religion and Logic in Buddhist Philosophical Analysis: Proceedings of the Fourth International DharmakTrti Conference. Edited by Helmut Krasser, Horst Lasic, Eli Franco, and Birgit Kellner. Vienna: Verlag der Osterreichischen Akademie der Wissenschaften, 299-308.
Kochumuttom, Thomas (1982): A Buddhist Doctrine of Experience: A New Translation and Interpretation of the Works of Vasubandhu, the Yogacarin. Delhi: Motilal Banarsidass.
Krasser, Helmut (2004): “Are Buddhist Pramanavadins non-Buddhistic? Dignaga and DharmakTrti on the Impact of Logic and Epistemology on Emancipation”. Horin 11: 129-146.
Kritzer, Robert (2005): Vasubandhu and the Yogacarabhumi. Tokyo: International Institute of Buddhist Studies of International College for Advanced Buddhist Studies.
La Vallee Poussin, Louis de (1912): “Vasubandhu, Vimsakakarikaprakarana. Traite des vingt slokas avec le commentaire de l'auteur. Traduit par Louis de la Vallee Poussin”. Le Museon n.s. 13: 53-90.
La Vallee Poussin, Louis de (1926): L'Abhidharmakosa de Vasubandhu. Traduit et annote par Louis de la Vallee Poussin. Troisieme chapitre. Paris: P. Geuthner.
Levi, Sylvain (1932): Materiauxpour l’etude du systeme Vijnaptimatra. Paris: Bibliotheque de l'Ecole des Hautes Etudes.
Lusthaus, Dan (2002): Buddhist Phenomenology: A Philosophical Investigation of Yogacara Buddhism and the Ch'eng Wei-shih lun. London/New York: RoutledgeCurzon.
Mills, Ethan (2013): The Dependent Origination of Skepticism in Classical India: an Experiment in Cross-Cultural Philosophy. Dissertation. University of New Mexico, Philosophy Department.
Mimaki, Katsumi, Musashi Tachikawa, and Akira Yuyama (eds.) (1989): Three works of Vasubandhu in Sanskrit manuscript. The Trisvabhavanirdesa, the Vimsatika with its Vrtti, and the Trimsika with Sthiramati's commentary. Tokyo: Centre for East Asian Cultural Studies.
Oetke, Claus (1988): “Ich” und das Ich. Stuttgart: Franz Steiner Verlag.
Oetke, Claus (1992): “Doctrine and Argument in Vijnanavada-Buddhism”. Wiener Zeitschrift fur die Kunde Sudasiens 36: 217-225.
Preisendanz, Karin (1994): Studien zu Nyayasutra III.1 mit dem Nyayatattvakoka Vacaspatimisras II, vol. 2. Stuttgart: Franz Steiner Verlag.
Priestly, Leonard C. D. C. (1996): Pudgalavada Buddhism: The Reality of the Indeterminate Self. Toronto: University of Toronto, Centre for South Asian Studies.
Rahula, Walpola (1978): Zen and the Taming of the Bull. London: Gordon Fraser.
Ratie, Isabelle (2011): Le Soi et l'Autre: Identite, difference et alterite dans la philosophie de la Pratyabhijna. Leiden: Brill.
Schmithausen, Lambert (1967): “Sautrantika-Voraussetzungen in Vimsatika und Trimsika”. Wiener Zeitschrift fur die Kunde Sudasiens 11: 109-136.
Schmithausen, Lambert (1973): “Spirituelle Praxis und philosophische Theorie im Buddhismus”. Zeitschrift fur Missionswissenschaft und Religionswissenschaft 3: 161-186.
Schmithausen, Lambert (1987): Alayavijnana. 2 vols. Tokyo: The International Institute for Buddhist Studies.
Schmithausen, Lambert (2005): On the Problem of the External World in the Ch’eng wei shih lun. Tokyo: The International Institute for Buddhist Studies.
Schmithausen, Lambert (2007): “Aspects of Spiritual Practice in Early Yogacara”. Journal of the International College for Postgraduate Buddhist Studies 11: 213-244.
Shastri, Dharmendranath (1964): Critique of Indian Realism: A Study of the Conflict between the Nyaya-Vaisesika and the Buddhist Dignaga School. Agra: Agra University.
Siderits, Mark (2003): Personal Identity and Buddhist Philosophy: Empty Persons. Aldershot, England: Ashgate.
Skilling, Peter (2000): “Vasubandhu and the Vyakhyayukti”. Journal of the International Association of Buddhist Studies 23/2: 297-350.
Stace, W.T. (1934): “The Refutation of Realism”. Mind 43/170: 145-155.
Steinkellner, Ernst (1979): DharmakTrti’s Pramanaviniscaya, Zweites Kapitel: Svarthanumanam. Teil II. Ubersetzung undAnmerkungen. Vienna: Verlag der Osterreichischen Akademie der Wissenschaften.
Steinkellner, Ernst (1992): “Lamotte and the Concept of Anupalabdhi”. Asiatische Studien/ Etudes Asatiques 46: 398-410.
Taber, John (1994): “Kumarila's Refutation of the Dreaming Argument: the Niralambanavada- Adhikarana”. In: Studies in MTmamsa: Dr. Mandan Mishra Felicitation Volume. Edited by R. C. Dwivedi. Delhi: Motilal Banarsidass.
Taber, John (2012): “Uddyotakara's Defense of a Self”. In: Hindu and Buddhist Ideas in Dialogue: Self and No-Self. Edited by Irina Kuznetsova, Jonardon Ganeri, and Chakravarthi Ram-Prasad. Farnham, England: Ashgate.

Tillemans, Tom J. F. (1995): “DharmakTrti and Tibetans on Adrsyanupalabdhihetu”. Journal of Indian Philosophy 23/2: 129-149.
Tillemans, Tom J. F. and Donald S. Lopez (1998): “What Can One Reasonably Say about Non-existence? A Tibetan Work on the Problem of Asrayasiddha”. Journal of Indian Philosophy 26: 99-129.

Vetter, Tilmann (1968): “Das Problem des metaphysischen Beweises in der logisch-erkenntnis- theoretischen Periode der indischen Philosophie”. Zeitschrift der Deutschen Morgenlandi- schen Gesellschaft 118/2: 351-356.
Walton, Douglas (1999): “The Appeal to Ignorance, or Argumentum ad Ignorantiam”. Argumentation 13: 367-377.
Watson, Alex (2006): The Self’s Awareness of Itself: Bhatta Ramakantha’s Arguments against the Buddhist Doctrine of No-Self. Vienna: Publications of the de Nobili Research Library.


Wayman, Alex (1979): “Yogacara and the Buddhist Logicians”. Journal of the International Association of Buddhist Studies 2: 65-78.
Wood, Thomas E. (1991): Mind Only: A Philosophical and Doctrinal Analysis of the Vijnanavada. Honolulu: University of Hawaii Press.
Yaita, Hideomi (1985): “On anupalabdhi, Annotated Translation of DharmakTrti's Pramanavarttikasvavrtti (2)”. Chisan Gakuho 34: 1-14.
Yamabe, Nobuyoshi (1998): “Self and Other in the Yogacara Tradition”. In: Nihon bukkyo bunka ronso: Kitabatake Tensei hakushi koki kinen ronshu. Kyoto, 15-41.