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

Buddhist Idealism, Epistemic and Otherwise: Thoughts on the Alternating Perspectives of Dharmakīrti

From Chinese Buddhist Encyclopedia
Jump to navigation Jump to search

by Dan Arnold


Some influential interpreters of Dharmakīrti have suggested understanding his thought in terms of a ‘sliding scale of analysis.’ Here it is argued that this emphasis on Dharmakīrti's alternating philosophical perspectives, though helpful in important respects, obscures the close connection between the two views in play (identified by later commentators as ‘Sautrāntika’ and ‘Yogācāra’). Indeed, with respect to these perspectives as Dharmakīrti develops them, the epistemology is the same either way. Insofar as that is right, John Dunne's characterization of Dharmakīrti's Yogācāra as ‘epistemic idealism’ may not, after all, distinguish this perspective from Sautrāntika; indeed, epistemic idealism can be understood as just the view these positions share. Thus, what distinguishes the ‘Yogācāra’ section of Dharmakīrti's texts is simply his making explicit that epistemological commitments the Sautrāntika does (or at least can coherently) hold are already compatible with idealism. Sautrāntika and Yogācāra thus differ only when one turns to the metaphysical arguments that (on the idealist's view) additionally show that only such mental things as sense data could be real.

Keywords Yogācāra.Dharmakīrti.Idealism.Epistemology

Introduction: Sautrāntika and Yogācāra in Dharmakīrti’s Work

Dharmakīrti’s Pramān :avārttika–the work that most influentially advanced a trajectory of thought that subsequent Indian philosophers took as practically

Of the many people who have helped me advance my understanding of the issues discussed here, I would especially like to thank the Sanskrit students who suffered through my attempts to teach the chiefly considered section of Dharmakīrti’s Pramān : avārttika: Colleen Christensen, Michelle Guittar, Sonam Kachru, Katarzyna Pazucha, and Charles Preston. D. Arnold (*) University of Chicago Divinity School, 1025 E. 58th Street, Chicago, IL 60637, USA e-mail:

co-extensive with the ‘Buddhist’ position in matters philosophical–is traditionally read as alternating between arguments for two kinds of views: a representationalist epistemology (of the sort familiar from modern empiricist sense-datum theories), which many later Buddhist commentators characterized as the ‘Sautrāntikaperspective; and the ‘Yogācāraperspective that is typically understood in terms of idealism. It is important to acknowledge the possibly misleading character of the traditional doxographic terms here in play; ‘Sautrāntika,’ in particular, seems to have been applied by later Indian commentators (and, following them, by many exemplars of the Tibetan scholastic tradition)

in ways that better served their own systematic concernsthantheyserveourunderstandingofanyactual,self-identifiedSautrāntikas.1 Insofar, however, as we take the positions so designated as philosophically ‘idealtypical,’ we can nevertheless make sense of the trajectory of doxographical thought that takes Sautrāntika, in particular, to give way to Yogācāra. I will retain this pair of terms, then, not only because of my having read Dharmakīrti through the lens of a later commentator who talks this way (Manorathanandin),2 but also because the familiarity of this doxographic scheme can usefully be exploited to develop the philosophical points I have in mind.3 The significance of Dharmakīrti’s

arguments often depends, in any case, on which of these perspectives is in play, and it can sometimes be difficult to determine, with respect to any particular topic, what his definitive position finally looks like. In view of this, some interpreters have commended the view that Dharmakīrti’s arguments can be seen in terms of something like a ‘sliding scale of analysis’–in terms, that is, of a hermeneutical perspective informed by sensitivity to this alternation.4 On John Dunne’s elaboration of this idea, the perspective from which Dharmakīrti most often argues (which I will call ‘Sautrāntika’) is characterized as ‘External Realism.’ This is not, to be sure, the realism of common sense; the things that ‘really exist’ (i.e., exist independently of awareness), on this view, turn out to be radically unique, fleeting particulars, whose ultimacy–which here means irreducibility–may consist in their not even having spatial extension. The salient point, however, is that this is nevertheless (as the commentators say) the ‘affirmation of external objects’ (bāhyārthavāda).5 In contrast, the Yogācāra perspective–which (Dunne notes)

1 See Cabezón (1990) for insightful reflections on some of the general issues here. 2 Dharmakīrti’s verses are often unintelligible without a commentary; I have followed that of Manorathanandin (on whom, see Sām : kr :tyāyana 1938–1940: i-ii) largely because it is the only one fully extant in Sanskrit. John Dunne’s study of Dharmakīrti (2004) is based chiefly on the earlier commentaries (now extant only in Tibetan translation) of Śākyabuddhi and Devendrabuddhi, on whom commentators like Manorathanandin clearly relied. 3 Thanks to John Dunne for some helpful comments on these issues. 4 For a careful development of this idea, see Dunne (2004, 53 –79). Dunne’s book and that of Georges Dreyfus (1997) provide authoritative and philosophically sensitive overviews of the thought of Dharmakīrti, who is usually taken to have lived c. 600–660 C.E. 5 Dunne (2004, 58 –59). The fact that the ontological primitives, even on this realist account, may lack spatial extension obviously plays into the idealist hand that I will show to come out on top. On questions relating to spatial extension, see the brief discussion of Vasubandhu’s Vim : śatikā below, ‘Epistemic vs. Metaphysical Arguments for Idealism’.

Dharmakīrti maintains ‘consistently only in one significant section of his Pramān :avārttika’6–is to be understood as ‘Epistemic Idealism.’ This is, on Dunne’s understanding, the view that ‘All Entities are Mental.’7 My aim here (as if these issues weren’t already tricky enough!) is to complicate this picture in what I think are some significant ways. What I want to suggest is that this emphasis on Dharmakīrti’s alternating philosophical perspectives, even though helpful in important respects, obscures something of the close connection between the two views in play here. I want to suggest, indeed, that with respect to the Sautrāntika and Yogācāra perspectives as Dharmakīrti develops them, the epistemology is the same either way. To the extent that is right, Dunne’s characterization of Dharmakīrti’s Yog ācāra as ‘epistemic idealism’ may not, after all, distinguish this perspective from Sautrāntika; indeed, epistemic idealism can be understood as just the view these positions share. This is, we will see, certainly what Dharmakīrti argues

when he adopts a Yogācāra perspective. There are, moreover, good philosophical grounds for thinking Dharmakīrti is right: proponents of Sautrāntika and Yogācāra are commonly committed to the view that what we are immediately aware of–which is different from the ontological issue of what there is–is only things somehow intrinsic to cognition. On my understanding, the salient point of this epistemological claim is that mental content is taken to be autonomously intelligible. This is the idea, in other words, that we can know how things seem to us quite apart from any considerations about how things really are–which is to say, the idea that we might find it intelligible that our own thoughts are not about a world. To the extent this idea holds, it can reasonably be thought that what we are immediately aware of–the direct objects, as it were, of our acts of awareness–is only things (sense data, say) somehow intrinsic to awareness. Dharmakīrti’sYogācārainsight is, to a great extent, simply that this is all that needs to be said, epistemologically, to recommend idealism. The difference between Sautrāntika and Yogācāra does not lie, that is, in their epistemology, since the Sautrāntika too can hold that we are immediately acquainted only with the contents of our own awareness; rather, the difference lies in the metaphysical arguments that (the idealist takes it) additionally show that only such mental things as sense data could be real. What distinguishes the ‘Yogācāra’ section of Dharmakīrti’s text, then, is simply his making explicit the extent to which epistemological commitments the

7 Dunne (2004, 59). The capital letters here are from the header to the paragraph in which Dunne sketches this view. Dunne notes that the Indian commentators most often refer to this perspective as antarjñeyavāda, ‘the view that objects of awareness are internal.’ 6 Ibid., 59. Dharmakīrti explicitly adopts a Yogācāra perspective, according to Dunne, ‘at the end of the third chapter, starting with the prologue at vv. 194ff.’ (Ibid., p.60n) The correct order of chapters for the Pramān :avārttika has historically been a matter of controversy (see, e.g., Gnoli 1960, xv-xvii); on the reckoning of most contemporary scholars, the third chapter is the one concerning perception (pratyaks :a). That is how I will take it, even though the edition I will cite (Shastri 1968) represents things otherwise. The section Dunne here identifies comprises the passages chiefly to be considered in the present essay, which span roughly verses 321–353. For a useful pass through something like the same material, see also Dreyfus and Lindtner (1989).

Buddhist idealism, epistemic and otherwise: thoughts on the alternating perspectives of Dharmakīrti

Sautrāntika does (or at least can coherently) hold are already compatible with idealism.8 I want to show this with reference to some of Dharmakīrti’s arguments, as elaborated by the commentator Manorathanandin, for the related doctrines of svasam : vitti (‘self-awareness’) and pramān :aphala (‘result of a pramān :a’). Among the most striking things about these arguments, on the reading I want to develop, is the Yogācāra proponent’s emphasis on epistemological common ground with Sautrāntika–or, as Dharmakīrti humorously puts it in what I will take to be the culminating passage, on the extent to which the Sautrāntika proponent has already (even if perhaps unwittingly) made the epistemological case for idealism. Understanding these arguments, I want to suggest in concluding, can help us understand why, in philosophically more general terms, it may be just as we should expect that a broadly empiricist approach would thus give way to idealism–why, that is, it is right to say with Donald Davidson (2001, 46) that empiricism just is ‘the view that the subjective (‘experience’) is the foundation of objective empirical knowledge.’ It is, I will finally suggest in concluding, the characteristically Buddhist claim that perception is non-conceptual that makes the essential subjectivity of the empiricist’s criteria of knowledge show itself.

Background: Dignāga’s Concise Elaboration of the Doctrine of Self-awareness

Dharmakīrti’s Pramān :avārttika is ostensibly framed as a commentary on the Pramān :asamuccaya of Dignāga.9 While much recent scholarship has emphasized that it is misleading to take these thinkers as exemplifying a monolithic school of thought, Dharmakīrti’s elaboration of the doctrines of svasam : vitti and pramān :aphala particularly closely tracks Dignāga; indeed, as Georges Dreyfus and Christian Lindtner have argued (1989, 27), it is most clear in the case of these doctrines that the works of Dignāga and Dharmakīrti represent the ‘products of a unified intention.’ We can, to that extent, usefully start with Dignāga’s concise presentation of these interrelated ideas. Central to Dignāga’s text (as to Indian discussions of epistemology more generally) is the category of pramān :a, which we might render as ‘reliable warrant’ or ‘valid doxastic practice’–something, at any rate, that picks out its referent’s being

8 An example of metaphysical arguments for this conclusion would be Vasubandhu’s Vim : śatikā, verses 11–15–an argument, we will see (‘Epistemic vs. Metaphysical Arguments for Idealism’), to which Dharmakīrti’s commentators allude in this context. Dunne is not unaware of the sense that his qualifier ‘epistemic’ has, and his explanation effectively identifies the same point I am making: ‘Here, the term “epistemic” is meant to reinforce the notion that Dharmakīrti’s critique of extra-mental entities arises in the context of determining what it is that we know in perception (I admit that “Epistemic Idealism” is a somewhat awkward neologism, but it is the most suggestive and least misleading terminology available).’ (Ibid., pp. 59–60n; emphasis mine). The characterization of this ‘epistemic idealism’ as the view that ‘all entities are mental’ suggests, however, the stronger, metaphysical point–which, I think, is also what is suggested by taking this to distinguish the ‘Yogācāraperspective from ‘Sautrāntika.’ It is with respect to these points, then, that I am trying to add some nuance–though what I am adding, in this regard, is in a sense simply a lengthy elaboration of Dunne’s own explanation, here, of his term ‘epistemic idealism.’ 9 On Dignāga (c. 480–540 C.E.), see especially Hayes (1988). For a very usefully annotated translation of the first chapter of Dignāga’s Pramān :asamuccaya, see Hattori (1968).

a criterion of knowledge. Most Indian philosophers were preoccupied with which doxastic practices (e.g., those involving perception, inference, testimony, comparison) should thus be reckoned as criteria (as pramān :as), and with characterizing the criteria so identified. Dignāga’s text famously argued (as would commonly be held by Buddhists writing after him) that only perception and inference had this status; all other ways of arriving at knowledge, Dignāga argued, are reducible to one of these. The first chapter of his Pramān :asamuccaya is chiefly concerned with characterizing perception as constitutively non-conceptual–a characterization that is central to the case for thinking this to be a basically empiricist approach. In the context of that discussion, Dignāga first elaborates the doctrine of svasam : vitti by way of explaining another claim considered

characteristic of the school of thought that begins with him: that by the word pramān :a we should really understand ourselves as referring not to any doxastic practice such as we might bring to bear on the independent objects thereof, but rather, to the cognition that is generally taken to result from the exercise thereof–to the pramān :aphala (‘result of the pramān :a’), as he says. Putting this point rather strongly, Dignāga says that ‘a pramān :a is real only as result’;10 it is, he explains, only figuratively that the word pramān :a is used as though it picked out something apart from a ‘resultant’ cognition. And, as Dignāga says in the next verse, this ‘result’ is svasam : vitti. Insofar, then, as Dignāga has said that ‘a pramān :a is real only as result’ and that this result is svasam : vitti, it seems we are to understand that pramān :as themselves are finally to be understood, somehow, as finally consisting in svasam : vitti–as consisting simply in ‘self-awareness.’11 By way of a first approximation of what might motivate the claims made here, let us consider some comments from Franz Brentano regarding what he called ‘inner perception’: ...besides the fact that it has a special object, inner perception possesses another distinguishing characteristic: its immediate, infallible self-evidence. Of all the types of knowledge of the objects of experience, inner perception alone possesses this characteristic. Consequently, when we say that mental phenomena are those which are apprehended by means of inner perception, we say that their perception is immediately evident. Moreover, inner perception is not merely the only kind of perception which is immediately evident; it is really the only perception in the strict sense of the word.... [this is because] the phenomena of the so-called external perception cannot be proved true and real even by means of indirect demonstration. For this reason, anyone who in good faith has taken them for what they seem to be is being misled by the manner in which the phenomena are connected. Therefore, strictly speaking, so-called

10 Pramān :asamuccaya 1.8d (pramān :am : phalam eva sat). All translations from Dignāga are my own, and are made from the Sanskrit text reconstructed by Steinkellner (2005); cf. Hattori (1968, 28). 11 The word svasam : vitti (like the semantically equivalent svasam : vedana) is formed from the reflexive pronominal prefix ‘sva-’, and a nominal form of the verbal root sam-√vid (‘to be aware’). I will generally leave it untranslated, allowing my engagement with the considered passages to do the work of showing its significance.

Buddhist idealism, epistemic and otherwise: thoughts on the alternating perspectives of Dharmakīrti 7

external perception is not perception. Mental phenomena, therefore, may be described as the only phenomena of which perception in the strict sense of the word is possible. (1973:91) Brentano’s counter-intuitive claim here–that this so-called ‘inner perception’ is, in the final analysis, ‘really the only perception in the strict sense of the word’–is, I suggest, comparable to Dignāga’s idea that a pramān :a is real only as svasam : vitti. What Brentano thus elaborates is an idea that can be understood to figure in the epistemologies of thinkers as diverse as Locke and Descartes: The idea that it is uniquely the case for ‘inner perception’–for our awareness, let us say, of the occurrence and contents of our own mental events–that the object of this cognitive act is precisely as it seems to us to be. This is because the ‘object’ of the awareness, in this case–what it is an awareness of–just is how the cognition seems to us. There can thus be said to obtain, in the case of Brentano’s ‘inner perception’ alone, an identity between the intentional content and the phenomenological character of such cognitions–between what they are about and how they seem. This is in contrast to all other cognitions, regarding which it is always possible to doubt whether what they are of is really as it seems. We might, then, be wrong in thinking that affairs in the world are just as represented in any particular cognition; but we cannot be wrong about the fact that that is it how it seems to us, insofar as its seeming to us that something is thus and so–which just is to say, our being aware–is finally what we are immediately aware of. We might plausibly suppose that Dignāga similarly means to say that the only thing worth identifying as a criterion of knowledge (as a pramān :a) is, for the same kinds of reasons, simply the way any cognition seems to us. More precisely, Dignāga takes as the criterion the conceptually basic fact that cognition seems some way or another. As he puts it, we may take ourselves to mean by ‘pramān :a’ the ‘instrument’ of an act of cognizing, but what we are really referring to is the bare fact of a cognition’s ‘being that whose phenomenal content (ābhāsa) is an object’12–is, in other words, the bare fact of cognition’s seeming to be of something.

Dharmakīrti on Two Ways of Taking Dignāga’s Claims

The foregoing can serve to give us the relevant background for the passages from Dharmakīrti to which we now turn.13 Having, then, elaborated (in verses 301–319) the pramān :aphala doctrine we have here sketched following Dignāga, Dharmakīrti proceeds to frame a debate between proponents of what Manorathanandin calls ‘Sautrāntika’ and ‘Yogācāra’ perspectives. At issue between the parties to this debate is one of the questions naturally raised by the foregoing pramān :aphala discussion: What, finally, is it reasonable to think we are aware of when we thus experience any seemingly contentful awareness? Is it, indeed, right to think that experience thus consists in awareness of anything at all? Or do the arguments for the foregoing

12 Pramān :asamuccaya 1.9cd vis :ayābhāsatâivâsya pramān :am : .... 13 All references here are to verses in the section demarcated in n.6, above; translations are my own, unless otherwise noted.

doctrine commit us, rather, to the view that awareness just is the occurrence of mental events that are intrinsically characterized by the fact of their seeming to have content? If the latter is the recommended conclusion, does it any longer make sense to distinguish, as it were, the epistemicvehicle’ of mental content from the content borne thereby? These are the kinds of questions in play when Dharmakīrti’s commentator Manorathanandin, early in this section, thus frames the topic: ‘What is said by ordinary people–i.e., “this is experienced by that”–precisely that is investigated by us.’14 Given what I take to be the Yogācāra view, the foregoing questions will be seen as rhetorical ones; this is a view on which there is nothing for awareness to be ‘of,’ since causally describable occurrences of awareness are all that finally exist.15 What the

Yogācāra proponent here wants to argue, though, is only that characteristically Sautrāntika epistemological commitments can be taken already to have prepared the way for that conclusion. It might seem that it would be difficult to motivate such a claim, particularly given the basis for the Sautrāntika position’s intuitive plausibility. The Sautrāntika epistemologist is committed to the credible view that perceptual cognition, in particular, is defined not only by its representing some object, but by its causal connection with that object–a view, one could think, that points precisely away from idealism (for what could better show something’s being ‘objective’ than its actually causing the operation of those transducers that are one’s senses?). Perceptions result, on such a view, from causally efficacious ‘impingements by the world on a possessor of sensory capacities,’ in John McDowell’s apt phrase.16 This view reflects, I take it, a basically empiricist approach to the objectivity of our knowledge–one according to which perception represents the unique point in our cognitive relation to the world at which our cognition is constrained by the world. This can surely seem a promising basis for securing the objectivity of our knowledge; for it is only in our causally describable perceptual encounters that we really ‘come up against’ a world of objects that are as they are quite independently of us.17 Dharmakīrti’s Yog ācāra proponent suggests in verse 323, however, that something like the same intuition can be retained even on the view that only mental

14 Manorathanandin, on verse 321b (Shastri 1968, 196): yad ucyate vyavahartr :bhir idam anenânbhūyate iti tad evêdam asmābhir vicāryate. When (as here) I cite passages from Manorathanandin’s commentary that comprise (because they are restating in prose) parts of Dharmakīrti’s verses, the latter will be given in bold type. 15 See below, ‘Epistemic vs. Metaphysical Arguments for Idealism’, for what I take to be the tradition’s most compelling (and clear) argument for this conclusion. 16 McDowell (1996, xv). Among the many passages from Dharmakīrti that might be cited in support of such a view, consider 3.224ab (Shastri 1968, 168): hetubhāvād r :te nânyā grāhyatā nāma kācana (‘there is nothing at all called being apprehendable apart from being a cause’; cf. Dunne 2004, 85n51); and (giving a solution to the ‘time-lag’ problem) 3.247 (Shastri 1968, 175): bhinnakālam : katham : grāhyam : iti ced grāhyatām : viduh : , hetutvam eva yuktijñā jñānākārārpan :aks :amam (‘If it is asked how something is apprehendable given its having occurred at a different time, those who understand reasoning know [that being apprehendable] is just being a cause which is capable of projecting an aspect into awareness’). 17 The characterization of Dharmakīrti (in his Sautrāntika guise, anyway) as an empiricist is not uncontroversial; see, for example, Dreyfus (1996)–and, in response, Tillemans (2003), 104, 121n31. See, as well, Arnold (2008a).

Buddhist idealism, epistemic and otherwise: thoughts on the alternating perspectives of Dharmakīrti events finally exist: ‘If it is agreed that the mark of being cognizable is resemblance of and causal relation to something, even so, an immediately antecedent cognition (samanantaram vijñānam), having the same content, could be what one is aware of (sam : vedyam).’18 The Yogācāra thus appeals to a standard-issue Buddhist analysis of the causal conditions of any moment of awareness. The idea, commonly elaborated in the Abhidharma literature, is that among the causes of any cognition is an immediately preceding moment of awareness (a samanantarapratyaya); visual awareness, for example, cannot suddenly emerge from factors that do not themselves possess the property of seeing–rather, any such moment must always have among its causes something that itself exemplifies the same properties now displayed. We do not here need to be detained by consideration of the

characteristically Buddhist commitments that motivate this idea, or of what there might be to recommend this view19; let us just suppose that we have here a critique of the idea of emergent properties–a critique, that is, of the idea that moments of awareness could be exhaustively reduced to things that are not themselves aware.20 Be that as it may, the Yogācāra’s point is that this idea can satisfy the Sautrāntika’s criterion of a causal constraint on cognition; if all that exists is mental events, it nevertheless remains the case that such events are causally constrained by what there is, insofar as mental events themselves occur in causally regular ways.21 The Sautrāntika rejoins that this misses the mark, since the salient point of his view is that perceptual cognitions are causally constrained by what they are cognitions of–and surely no one would want to say that an ‘immediately preceding moment of awareness’ is itself the content of any present such moment. Unpacking Dharmakīrti’s verse to this effect (324a-c), Manorathanandin says, ‘an experience is of that regarding which there arises a thought of ascertainment–“this was seen,” or “this was heard”–not of anything else. And there is no ascertainment

of something seen or heard, etc., with regard to an immediately preceding moment of awareness (samanantarapratyaya)–so this is not what is apprehended.’22 The Yogācāra’s response: ‘Precisely this successiveness (pratyāsatti) is what we’re investigating.’23 That is, whether it is right to say that awareness is of something is just what is in question here. If, that is, one wants (as the Yogācāra does) to argue that cognition’s seeming to have content is simply an intrinsic property of awareness, then of course the fact that our judgments seem successively to follow our perceptions–seem, that is, to be ‘about’ what was first just given to our 18 Shastri (1968, 197): tatsārūpyatadutpattī yadi sam : vedyalaks :an :am / sam : vedyam : syāt samānārtham : vijñānam : samanantaram //. 19 Suffice it to say that this is

surely among the categories introduced by Buddhists to try to account for continuity in the context of what is basically a causal-reductionist project. On this idea (and, more generally, on some interesting philosophical problems having to do with Buddhist attempts thus to explain continuity), see Griffiths (1986), passim. 20 This is an idea that figures prominently in Dharmakīrti’s proof of rebirth; see Taber (2003) for a useful overview. 21 It may, however, be that on this view, the contrast between perceptual and other cognitions (which surely can also be described as exhibiting this kind of causal regularity) is lost. 22 Shastri (1968, 197):... idam dr :s :t :am : śrutam : vêdam iti yatrâvasāyadhīr utpadyate tasya so ’nubhavo nânyasya. na ca samanantarapratyaye dr :s :t :aśrutādyavasāyo bhavati tan na grāhyo’sau. 23 Dharmakīrti’s verse 324d (ibid.): sâiva pratyāsattir vicāryate.

receptive capacities in perception–is precisely the kind of thing one would need to explain. This is, then, just what Dharmakīrti claims he is explaining, so it cannot be invoked against him until we have seen his explanation. What, though, could recommend the conclusion that cognition’s seeming to have content is just an intrinsic property of awareness? Why would someone think we have to be able to understand what thoughts are about independently of the things (trees, pots, lunch, sentences) we typically think they are about? What could suggest the idea that we might find the contents of awareness intelligible without also thereby finding the world intelligible? The basic thought here, I

think, is that anything taken as a constraint on the content of cognition must, as some of the commentators put it, be ‘immediately’ (avyavadhānena)24 occurrent with the cognition in question. Nothing that does not thus ‘immediately’ attend cognition could, on this view, be thought invariably to constrain it. The idea of immediacy here can be understood to involve the kind of identity we noted with reference to Brentano: to have a cognition just is for it to seem to a subject that something is the case; insofar, though, as the one thing that is invariably and indisputably true of awareness is only that it thus seems to us to be of something, we have, uniquely, an irrefragable awareness of that ‘object’ which is our cognition’s seeming as it does. Nothing else, this line of thought goes, is immediately (avyavadhānena) related to the occurrence of an act of cognizing; for while it can be doubted whether anything else that is proposed as a constraint on the content of cognition is really as it seems, there is in this case alone an

identity between the intentional content and the phenomenological character of cognition. Indeed, a cognition’s seeming to be of something just is its character as a cognition; the ‘immediacy’ that obtains, then, is of the peculiarly strong sort that goes with identity.25 We can appreciate this point by seeing what is said about other alternatives that are considered as possibly constraining the content of awareness. Dignāga’s commentator Jinendrabuddhi, for example, anticipates the objection that the relative acuity of our sense faculties could explain the determinacy of mental content. In response, he gives a deceptively simple reason why such factors cannot explain what we here want to understand: ‘because of their not being of the nature of cognition, and because, rather, of their being the cause of all cognitions.’26 What is needed, Jinendrabuddhi thus emphasizes, is something itself ‘of the nature of cognition’ (jñānasvabhāva). Things like properly functioning sense faculties, then, may well be 24 This term is used in the same context, for example, by Dignāga’s commentator Jinendrabuddhi (c. 710– 770 C.E.), who postdates (and is much influenced by) Dharmakīrti; see

Steinkellner, et al. (2005), p.66 line 6–and, for discussion of this, Arnold (2008b). Jinendrabuddhi is here exploiting an idea introduced by one of Dharmakīrti’s commentators, who similarly thought that the only thing indisputably related to the occurrence of cognition could be something ‘not separated’ (avyavahita; the word is from the same root as Jinendrabuddhi’s avyavadhānena) from it; see Dunne (2004:261–62, 269). 25 Indeed, this is arguably just the point of the pramān :aphala doctrine–a point suggested by Dunne (2004, 261–2), discussing the commentator Śākyabuddhi’s idea that the pramān :aphala doctrine especially concerns the ‘unmediated’ (avyavahita) effect of a pramān :a. The distinctiveness of such an ‘unmediated’ effect is ‘that, as Śākyabuddhi notes, it “necessarily occurs.” This is so because the effect is not separated (avyavahita) from the instrument; it is, in fact, identical to the instrument itself.’ 26 Translated from Steinkellner et al. (2005), p.66.11–67.1: tasyâjñānasvabhāvatvāt sarvajñānahetutvāc ca. Buddhist idealism, epistemic and otherwise: thoughts on the alternating perspectives of Dharmakīrti 11

among the enabling conditions of awareness–but their functioning is not itself being aware. What is needed, however, just is something that, as identical with being aware, can be thought ‘immediately’ to attend cognition; for while we can doubt (say) whether a seemingly sensory awareness really is produced by the senses, we cannot doubt its character as an awareness. What invariably or ‘immediately’ constrains mental content, we are thus encouraged to suppose, must be just intrinsic properties of awareness. In the present engagement with a Sautrāntika epistemologist, Dharmakīrti’s Yogācāra proponent develops the same point in the following exchange. The Sautrāntika, persisting in the view that what awareness is of is just the external objects that it seems to be of, asks why the Yogācāra proponent can think an account such as his is required: ‘But if an external object is experienced by cognition, then what is the problem owing to which self-awareness is accepted as the result?’ The enigmatic response (given in verse 333): ‘If it is experienced, then there is no problem at all; but it is said that experience alone (anubhava eva) is not of anything external. So what would be added by saying, “An external object is experienced by cognition”?’27While this is not, perhaps, altogether clear, the Yogācāra here seems chiefly intent on denying the first conditional (‘If it is experienced...’) – that, I take it, is the point of his thus asserting that experience, considered simply in terms of its intrinsic properties (‘experience alone,’ anubhava eva), is ‘not of anything external.’ The point of the concluding rhetorical question then seems to be that insofar as one can only know things as they are experienced, nothing

is added to our understanding of experience by reference to anything beyond the experience itself; to be acquainted with things only as they are experienced just is, in effect, to be acquainted with ‘experience alone.’ This point is more clearly developed in the comments on the next verse (334), where the Yogācāra proponent acknowledges that of course, phenomenologically speaking, it seems to us that what we experience is external to awareness – but continues that how we are to explain this is, again, just what is in question. It is true, he thus concedes, that ‘thought possesses a specific aspect (buddhir ākāravises :in :ī), i.e., it is connected with a particular aspect of blue or non-blue, etc.; but it is worth considering whether that thought could arise from an external object, or from something else, i.e., from the constraint of a latent disposition (vāsanā).’28

27 Shastri (1968, 199): nanu yadi bāhyo’rtho jñānenânubhūyate tadā ko dos :ah :, yena svasam : vit phalam is :yate. āha: yady anubhūyate, tadā nâiva kaścana dos :ah :. anubhava eva tu bāhyasya nâstîty ucyate. tathā idam eva kim uktam : syātb āhyo ’rtho jñānenânubhūyata iti. Cf. Dunne’s translation of Dharmakīrti’s verse (2004, 277n): “‘If we maintain that an external object is experienced, what would be wrong with that claim?” There is nothing wrong with it, but what is the point of saying this: “An external object is experienced”?’ 28 Shastri (1968, 199–200): yadi buddhis tadākārā vā bāhyasarūpêty ucyate, satyam[;] asti sā buddhir ākāraviśes :in :ī nīlānīlādyākāraviśes :ayuktā; kim : tu sā buddhir bāhyād arthājj āyeta, anyato vāsanāpratiniyamād vā iti vicāram idam arhati. Cf. Dunne’s translation of Dharmakīrti’s verse (2004, 277n): ‘If awareness has the image of the object, then it must have something that distinguishes [each] image [for each awareness]. It would be wise to look into whether that differentiation must come from something external, or whether it might just as easily come from something else.’

12 D. ArnoldVāsanās are typically understood by Buddhists as the lingering ‘traces’ of mental activity, posited to explain the many occasions on which the effects of karma are realized only well after the relevant action (karma is thus said to have lain ‘dormant’ in the form of vāsanās); the salient point here, though, is simply that they are mental items.29 The Yogācāra thinks, as he goes on to explain, that we cannot finally know anything beyond such items to constrain our awareness since ‘there is no object at all, possessing a distinction from thought (buddhivyatirekin), which is apprehended as being the cause (hetutayā-upalabhyate).’ The point is that anything cognitively ‘apprehended’

is, ipso facto, internally related to an act of awareness; to that extent, what is immediately known, when one knows anything to be a constraint on awareness–the direct object, as it were, of this very act of knowing–could never have the property of ‘possessing a distinction from thought.’ Hence, as Dharmakīrti’s Yogācāra puts this point, no object apart from thought itself can be known as the cause ‘because of the awareness of nothing but the form of cognition.’30 We only experience things, again, as they are experienced. Dharmakīrti elaborates on this line of thought in the next verse (335a-c), which Manorathanandin elaborates thus: ‘In regard to this, because of the nonapprehension of things like blue apart from the qualification which is awareness– and because of the apprehension of blue only when there is apprehension of that [i.e., of

the awareness]–perceiving (darśana), whose content is things like blue (i.e., whose aspect is blue), is based on awareness of the blue and of the thought only together.’31 Theonethingwecannotdoubt,withrespect toanyoccurrent awareness,is the fact (itself constitutive of awareness) of its seeming to be of something; there is no awareness that lacks this property or ‘qualification’ (upādhi). This means, however, that the property of thus seeming, phenomenologically, to have some content must itself be intrinsic to cognition; that is, indeed, just what it means to say that one cannot have an awareness lacking this property. Whenever, then, one is aware of anything (suchasapatchofblue),onewillalso,ipsofacto,beawareofthefactofbeingawareof it; this is Dharmakīrti’s point in thus urging that one can be aware of anything ‘only together’

(sahâiva) with the fact of the awareness. Precisely to that extent, however, the awareness itself must be reckoned as conceptually basic. If, in other words, we believe based on our experience that things out there in the world exhibit such properties as being blue, it is finally only 30 Ibid., 200: na tāvad buddhivyatirekin :ârthah : kaścid dhetutayôpalabhyate, buddhisvarūpamātravedanāt. 31 Shastri (200): tatra darśanena jñānenôpādhinā viśes :an :ena rahitasya nīlāder agrahāt tasya grahe ca nīlasya grahāt sahâiva nīladhiyor vedanāt darśanam : nīlādinirbhāsam : nīlākāram : vyavasthitam : . Cf. Dunne’s translation of Dharmakīrti’s verse (2004, 277n): ‘[1] There is no apprehension of an object devoid of qualification by the experience of it; and [2] when that experience itself is apprehended, the object is apprehended. Therefore for these two reasons, the cognitive appearance of blue is the experience (darśana) of blue. There is no independent ... external object.’

29 These are among the categories meant to address problems of causal continuity; see n.19, above. They are mental just insofar as karma itself is finally mental. Thus, karma is glossed as cetanā (commonly translated ‘intention,’ and certainly something mental) by Vasubandhu in the Abhidharmakośa (4.1); in the Vim : śatikā (verse 7), Vasubandhu then draws the entailed Yogācāra conclusion that, per the principle of ontological parsimony, the effects of karma–what is produced by it (which for Buddhists is everything)– must also be mental.

Buddhist idealism, epistemic and otherwise: thoughts on the alternating perspectives of Dharmakīrti 13 insofar as we have the logically prior experience of its seeming to us that something is blue that we believe it to be so. If this appropriately represents Dharmakīrti’s idea, it should be clear that his argument, as here developed, advances a key version of what Wilfrid Sellars influentially criticised as the ‘myth of the given’–in particular, a version based on the idea that knowledge is built on the foundations of what incorrigibly seems to a subject to be the case. Sellars argued, against such views, that ‘being red is logically prior [to], is a logically simpler notion, than looking red’–and concludes, from his arguments for this claim, that ‘it just won’t do to say that x is red is

analyzable in terms of x looks red to y.’32 To the extent that it carries conviction, this characteristically Sellarsian point in John McDowell’s formulation—that ‘reality is prior, in the order of understanding, to appearance’ (1998b, 410)–would also cut against the argument Dharmakīrti seems to have in mind. On the view here developed following Dharmakīrti, then, only the fact of cognition’s seeming some way is indubitable–and that fact is intrinsic to (indeed, it is constitutive of) awareness. The point Dignāga and Dharmakīrti are after in arguing that the seeming itself is the only line pramān :a–that, as Dignāga had put it, the only thing worth calling a pramān :a (the only criterion of what we know) is the bare fact of cognition’s ‘being that whose phenomenal content (ābhāsa) is an object.’33 This, finally, is why Dharmakīrti can say (as he does in concluding

verse 335) that experience is therefore intelligible without reference to external objects. As Manorathanandin puts this point in elaborating Dharmakīrti’s conclusion, ‘There is no external object (things like colors) independently [of awareness], since what is cognitively accessible (adhyaks :a) — which is what is accepted [by the Sautrāntika] as showing external objects – does not have the capacity’34 for warranting conclusions about what might be the case apart from awareness; for what is ‘cognitively accessible’ is only the intrinsic properties of awareness itself. This, then, is the argument for thinking that something mental—immediately antecedent moments of awareness, ‘latent dispositions’ (vāsanā), or whatever–is all that we are finally entitled to think of as constraining the content of awareness. Dharmakīrti now elaborates that point, in a verse (336) which Manorathanandin begins to unpack thus: On the part of any cognition, whose aspect is things like elephants,35 something (kiñcid eva)–i.e., a cognition–is the awakener of an internally latent disposition (antarvāsanā), whose characteristic is being suitable for producing a determinate cognition (niyatajñāna), internally located as the immediately preceding moment of awareness (samanantarapratyaya)–i.e., it is a causal factor directed towards producing an effect. Based on that–i.e., based on the awakening factor–

32 Sellars (1997, 36). Sellars’s entire discussion at pp. 32–46 is relevant here; see also Robert Brandom’s comments at pp. 136ff. of this edition. 33 Pramān :asamuccaya 1.9cd; see n.12, above. 34 Shastri (1968, 200): bāhyo nīlādir arthah : kevala nâsti, tatsādhakatvenâbhimatasyādhyaks :asyâsāmarthyāt. See n.32, above, for Dunne’s translation of Dharmakīrti’s concluding quarter verse. 35 That is, whose content is ordinary macro-objects. It is perhaps a bit odd that Manorathanandin thus takes ‘elephants’ as an example of what might be the object of awareness, rather than a more stock example such as a pot.

there is restriction of thoughts as being those whose aspects are constrained; [this restriction is] not with reference to an external object.36 Insofar, that is, as it is cognition’s seeming some way–its intrinsically having content–that is conceptually basic, we are, in fact, better off supposing that the content of awareness is constrained by preceding moments of awareness (by those ‘latent dispositions’ that have the capacity to give rise to further moments of awareness), than that it is immediately constrained by external objects themselves. This, finally, is what it means for Dharmakīrti effectively to argue (contra Sellars) that the phenomenological fact of something’s seeming blue is logically prior to our having the idea of something external’s being blue.

Epistemic vs. Metaphysical Arguments for Idealism

Dharmakīrti has, in this way, now developed the point that appeal to latent dispositions (or other such mental items) represents not simply an alternative explanation–one that can as easily be imagined as can the account proposed by the realist about external objects–but the preferred one. Note, however, that what has thus been advanced as the preferred account is just epistemic idealism–the view, that is, that what we are immediately aware of must be understood in terms of the intrinsic properties of cognition. What makes this an instance of epistemic idealism (idealism, that is, only with regard to what we know) is that this remains compatible with an ontological commitment to really existent external objects; all that has been given up is the claim that such existents could be the direct objects of our awareness. One could, for example, hold that this

account’s ‘latent dispositions’ are themselves produced by the causally describable contact between our sense faculties and the world–that what McDowell’s ‘impingements by the world on a possessor of sensory capacities’ produce is (say) mental representations, which themselves are all that we are immediately aware of.37 The case for thinking we are not immediately aware of what thus ‘impinges’ upon us is not, however, the same as the case for thinking that nothing that thus impinges exists. This is, in fact, just as the Sautrāntika (as represented in Manorathanandin’s commentary on verse 336) now rejoins: ‘Even so–i.e., even given the absence of a probative external object, which [you have shown to be] cognitively inaccessible (paroks :a)–there is no proof of absence (nâbhāvasthiti).’38 The proponent of Yogācāra 37 See n.50, below, for Dharmakīrti’s development of this point.

36 Shastri (1968, 200):kasyacij jñānasya gajādyākārasya kiñcid eva jñānam antarvāsanāyāh : samanantarapratyayāntaravarttinyā niyatajñānajananayogyatālaks : an :āyāh : prabodhakam : kāryotpādanābhimukhyakārakam : . tatah : prabodhakavaśāt dhiyām : niyatākāratayā viniyamah :, na bāhyārthavyapeks :ayā. Cf. Dunne’s translation (2004, 277n), which reads Dharmakīrti’s verse 336 continuously with 335 (n.32, above): ‘Instead, something activates the internal imprint for some experience. It is due to that awakening of an imprint that there is the restriction [of a particular image] to a [particular] cognition; that restriction is not dependent on an external object.’ 38 Shastri (1968, 200): na[;] tathâpi paroks :asya bāhyasya sādhakasyâbhāve ’pi nâbhāvasthitir iti cet.I take the initial ‘na’ as syntactically independent of the sentence that follows–as expressing, that is, the Sautrāntika’s initial denial of the account just proposed by the Yogācāra, with the ensuing sentence giving the reason for so denying. Buddhist idealism, epistemic and otherwise: thoughts on the alternating perspectives of Dharmakīrti 15

may have shown, that is, that external objects are not cognitively ‘accessible’; but that is not to have shown that they do not (because it is not to have shown that they cannot) exist. It is, finally, in replying to this objection that Dharmakīrti’s Yog ācāra proponent (again represented in Manorathanandin’s commentary on 336) effectively argues that epistemic idealism is precisely what he and the Sautrāntika have in common: A cognition is appearing; but it does not appear as external (bāhyam : tu na pratibhāsata eva)–our effort (which is dedicated to negating a fiendish external object which is without a pramān :a that is probative of the desired conclusion)39 is only to that extent. But if the desire to refute this [i.e., external objects] is weightier, [the effort] of the master (ācāryīyah :) with respect to the refutation of atoms (by considering whether or not they have parts) should be considered.40 This revealing exchange is remarkable for its expression of a point that is too often missed in discussions

of Yogācāra: the difference between epistemic arguments for idealism (arguments based on the claim that all we immediately know is things that are themselves mental), and metaphysical arguments for the claim that only mental things exist. Here, the Sautrāntika opponent has quite rightly objected that the Yogācāra’s epistemological case for idealism does not by itself establish that only mental events exist. Dharmakīrti’s Yog ācāra has answered, though, by concurring; he emphasizes that all he claims to have shown–‘our effort is only to this extent’–is that whatever the content of a cognition seems to be, its seeming so is not itself external (‘it does not appear as external’). This is true by definition, since its seeming so is just what any putative object thereof would be external to. Cognition’s seeming some way, however, is all that is ever really known for sure. If one wanted, additionally, to argue that only such ‘seemings’ themselves exist– if, as Manorathanandin here puts it, one’s intention to refute external objects was ‘weightier’–it would be necessary to advert to a fundamentally different kind of argument. It would, in particular, be necessary to offer a metaphysical argument of the kind paradigmatically exemplified by the Vim : śatikā of Vasubandhu–the ‘master’ (ācārya), surely, to whose effort this passage clearly alludes. Vasubandhu’s arguments represent a metaphysical case for idealism in the sense that they do not depend on a posteriori analysis of what we experience; rather, they involve a priori analysis of the adequacy of our concepts.41 In particular, the philosophical heart of Vasubandhu’s text42 concerns (as Manorathanandin here says) ‘the refutation of atoms (by considering whether or not they have parts).’ The famous argument thus characterized starts from the premise that anybody who would be a realist about external objects is necessarily committed to atomism; it must, that

39 An unusual rhetorical flourish from Manorathanandin! 40 Ibid.: pratibhāsamānam : jñānam : bāhyam : tu na pratibhāsata evêti tāvatâvābhimatasiddheh : sādhakapramān :arahitapiśācāyamānabahirarthaniśedhenâsmākam ādarah : . yadi tu tanniśedhanirbandho garīyān sām : śatvānam : śatvakalpanayā paramān :upratis :edhe ācāryīyah : paryes :itavyah : . 41 It is, to that extent, presupposed by Vasubandhu’s arguments (as perhaps by metaphysical arguments more generally) that what we can or cannot adequately conceive tells us something about what is actually possible; whether that is a warranted presupposition is a topic for another day. 42 Verses 11–15–on which see, especially, Kapstein (2001).

is, be the case that the variously reducible macro-objects of our experience are ultimately reducible to basic parts that are not themselves further reducible, since otherwise it would not be possible to specify what, precisely, finally exists. The argument is then to the effect that no coherent account of atomism can be given. The reason, most basically, is that properly irreducible atoms would have to lack spatial extension; for spatially extended atoms–atoms that, insofar as they can join with one atom on one side and another on the other side, could coherently be thought the building blocks of macro-objects–would, ipso facto, be reducible (e.g., to their ‘left’ and ‘right’ halves). But nothing without spatial extension could constitute things having spatial extension. Vasubandhu’s arguments, then, are meant to show that realism about external objects is

fundamentally incoherent: the proponent of such a view both must and cannot propose an account of atoms.43 This, it should be clear, is a very different kind of argument from the epistemological arguments that we have seen Dharmakīrti developing. Dharmakīrti does not claim to have shown that only mental events can exist–only that, whatever the case in that regard, we are only immediately aware of mental events. But Dharmakīrti can, as the proponent of Yogācāra that he is taken to be in this section of his text, here readily allow that this is all he has shown, since that already gets him all he needs epistemologically. Insofar, that is, as Dharmakīrti’s concerns are only epistemological–only, as the title of his text tells us, with the nature and status of pramān :as–there is nothing to be added to the ‘Sautrāntika’ case in order for Dharmakīrti’s to be the epistemology of a finally idealist position. Whether, then, the claim is (as Dharmakīrti has argued that the Sautrāntika can allow) that only mental events are immediately known or (as for Vasubandhu) that only these exist, the epistemology remains the same; either way, we have at least epistemic idealism. Dharmakīrti’s elaboration of a ‘Yogācāraperspective in epistemology here consists, then, simply in making explicit the extent to which the empiricist representationalism of Sautrāntika already constitutes the epistemological case for idealism. Anybody who wants, additionally, a metaphysical case for

43 Recent developments in physics suggest that Vasubandhu’s arguments are not a mere historical curiosity. See, for example, Elizabeth Kolbert’s(2007) New Yorker piece on the imminent opening of the Large Hadron Collider. Kolbert quotes physicist Jos Engelen on what is promised by research projects such as this: ‘what we want is to reduce the world to objects that have no structure, that are points, that are as simple as we can imagine. And then build it up from there again’ (emphasis added). The thought, in other words, is that physicists cannot be confident that they have found the most basic sub-atomic particles until they have found something further irreducible–which means they are effectively looking for (what alone has ‘no structure’) mathematical points. Here, then, is a massively funded research project dedicated, in effect, to the empirical discovery

of what is a mathematical abstraction–to the empirical discovery, that is, of something which, though lacking spatial extension itself, somehow explains the spatially extended character of everything else. Vasubandhu’s arguments are meant to show precisely such an endeavor to be incoherent. It should be noted that it is not universally agreed that Vasubandhu should here be understood as arguing for idealism; see, for example, Lusthaus (2002) for an influential case for a non-idealist reading of Vasubandhu. While ‘Yogācāra’ names, of course, a position developed within a vast corpus of works, I confess that I do not see how the Vim : śatikā, at least, can be understood as offering anything other than a metaphysical argument for idealism. Manorathanandin’s explicit differentiation between, in effect, epistemic and metaphysical arguments–between less and ‘more weighty’ (garīyān) intentions to ‘refute external objects’ (bahirarthaniśedha)–seems to me to represent a point that could be enlisted as part of a case for a strong ‘idealist’ reading of Yogācāra. For a cogent philological argument to the effect that nonidealist readings of texts like the Vim : śatikā are unconvincing, see Schmithausen 2005.

Buddhist idealism, epistemic and otherwise: thoughts on the alternating perspectives of Dharmakīrti 17 idealism–anyone, as Manorathanandin puts it, whose desire to refute external objects is ‘weightier’–will have to look elsewhere.

Dharmakīrti’s Clincher: Sautrāntika is Epistemic Idealism

In that case, though, Dunne’s characterization of Dharmakīrti’sYogācāraperspective as ‘epistemic idealism’ may not, after all, distinguish this from the ‘Sautrāntikaperspective from which he more often argues–which should perhaps lead us to suspect that the strictly epistemological intuitions that lend credence to idealism are not as innocent as they may seem. This is what I would like to consider in concluding; first, however, let us see how Dharmakīrti drives home his point. Emphasizing the limited scope of his epistemological case for Yogācāra, Dharmakīrti stresses (again as in Dignāga’s discussion of these doctrines) that all the arguments for the foundational status of svasam : vitti have really shown is that awareness constitutively has various ‘aspects’ (ākāra), which are all that is immediately known–a point that is again developed as one that the Sautrāntika and Yogācāra perspectives can share. Dharmakīrti thus elaborates the view he has developed without the kind of ‘weightier’ concern for refuting external objects that Vasubandhu showed. In verse 337 (again, here unpacked by Manorathanandin), he states the conclusion that should follow from verse 336: Since an external object is not experienced, therefore a single cognition, because of its being overcome by ignorance,44 has two forms–the form of awareness itself, and the form of such objects thereof as color. This is because cognition is experienced in this way (as being of two aspects) by selfawareness, and is remembered at a later time according to that experience. And in this way, because there is no awareness of anything else, awareness of both aspects–i.e. of the form of experience itself and of things like color–is the result. Thus, in this way, what is known (prameya) is the apprehended aspect, the pramān :a is the apprehending aspect, the result is self-awareness–this is what has been shown.45 Here, Dharmakīrti’s discussion closely tracks the elaboration of these doctrines to be found in the earlier work of Dignāga. It is said, for example, that cognition is experienced as

having two ‘aspects’–that, in other words, it is intrinsic to cognition for it to involve that ‘aspect’ (a representation that is itself internal to cognition) 44 The characteristically Yogācāra point here is that when rightly understood, awareness is finally to be seen as altogether lacking an intentional structure; it is only insofar as we are benighted by ignorance (of the sort which it is the constitutive concern of Buddhism to eliminate) that we mistakenly suppose that we stand, as knowing subjects, over and against the world that is known by us. 45 Shastri (1968, 201): yasmādb āhyo’rtho nânubhūyate, tasmād ekam : vijñānam : avidyopaplutatvāt dvirūpam : bodharūpam : nīlādirūpañ câsti. yat yasmāj jñ ānam evam : dvyākāratayânubhūyate svavedanena yathânubhavam : kālāntare smaryate ca. tathā cânyasya sam : vedanābhāvāt, ubhayādyākārasya nīlādyanubhavarūpasya sam : vedanam : phalam : . tad evam : , prameyo grāhyākārah : , pramān :am : grāhakākārah : , phalam : svasam : vid iti darśitam : bhavati. Cf. Dunne’s translation (2004, 277n) of Dharmakīrti’s verse: ‘Therefore that one awareness which is experienced and remembered in that fashion has two aspects (dvirūpa); the instrumental result is the reflexive awareness of both aspects.’

which is taken to be what the awareness is of, as well as the ‘aspect’ which is its reflexive character as an awareness thereof. The latter idea–that cognition invariably involves not only some content, but also an intrinsically reflexive character–is further said, here, to relate to memory (cognitions are ‘remembered at a later time according to that experience’). It was, in fact, largely by appeal to memory that Dignāga had tried to show that all cognitions necessarily exhibit the reflexive character that is part of what he has in mind in discussing svasam : vitti. Thus, Dignāga argued that the memory of a first-order cognition is phenomenologically distinct from the first-order cognition thus remembered just insofar as the memory involves not only the content of the original cognition, but also, explicitly, the awareness of oneself as having

experienced it; this is just what makes the memory phenomenologically distinct. Insofar, however, as one can have available to memory only what has already been experienced, this self-awareness, Dignāga argued, must already have been part of the first-order cognition.46 And it is insofar as every awareness must thus be thought to involve self-awareness that Dignāga concluded, at Pramān :asamuccaya 1.10, just what Dharmakīrti has expressed with his verse 337: that (as Dignāga puts the point) ‘what appears in cognition is [called the] object of awareness (prameya); the properties of being the pramān :a and being the result thereof then belong, respectively, to the apprehending aspect and to self-awareness. These three, therefore, are not separate.’47 As Dharmakīrti and Manorathanandin have now stressed, Dignāga’s last point–that these three factors,

separable for heuristic purposes, are not finally distinct–follows from the fact that all of them are comprised in svasam : vitti; ‘awareness of both aspects,’ they say, ‘is the result.’ It is, then, finally only because ‘cognition is experienced in this way (as being of two aspects) by selfawareness’ that we know anything at all. Dharmakīrti then makes explicit (in verse 338) the point that this is something the Sautrāntika can assent to: When, even given the affirmation of external objects, something other [than thought] (i.e., an external object)–having a nature that is constituted as desired or not desired (i.e., is being desired or not desired, as differentially established owing to cultivation)48–is the object [just] insofar as it is the cause of a

46 On this line of argument, see, inter alia, Ganeri (1999). 47 Pramān :asamuccaya 1.10: yadābhāsam : prameyam : tat pramān :aphalate punah : , gr āhakākārasam : vittyos trayam : nâtah : pr :thakkr :tam. The status of Dharmakīrti’s verse 337 (and of the verse from Dignāga given here) as a locus classicus seems to be suggested by the fact that several editions of Manorathanandin’s commentary append annotations (in the manuscript...?) citing Dignāga’s main verses on the subject. The interlinear reference to Dignāga is then followed by a comment that reads (per Sām : k:rtyāyana 1938–1940, 222): ‘Based on following this group of four verses from Dignāga, four alternatives regarding the result have been expressed; these alternatives are just possibilities–but the conclusion rendered is just Vijñānavāda’ (sūtracatus :t :ayānurodhāc catur āvarttitah : phalavikalpah : ; sambhavamātren :âmī vikalpāh : [;] samāptis tu vijñānavāda eva kr :tā). 48 That is, it is largely owing to one’s own cultivated dispositions that something is found attractive

or not. Reference to the affective character of experience here is another way to make the point that we experience things only as they are experienced; for the experienced fact of something’s being, for example, ‘desired or not desired’ is surely a fact about our awareness, not about the putative objects thereof. Insofar, then, as our experience of things typically involves some such affective response, it is clear that what we are immediately aware of is only a representation that is itself internal to awareness. Buddhist idealism, epistemic and otherwise: thoughts on the alternating perspectives of Dharmakīrti 19 representation (vijñapti) with that form, then the experience of that representation is in this way–i.e., with the aspect of being desired or not desired–and is said to be experience of an object. Accordingly, resemblance of the object is said to be the pramān :a, awareness of an object the result.49 Dharmakīrti thus makes explicit the kind of account (noted above)50 according to which the mental events–vāsanās, samanantarapratyaya, representations, or whatever–that immediately constrain the content of awareness might, for all we know, be caused by external objects. The salient point remains, however, that even for the Sautrāntika, external objects would thus have to be reckoned simply

as ‘the cause of a representation.’ The experience of this representation, though, is all that can be ‘said to be experience of an object.’ This chastened way of understanding the Sautrāntika position is not, Dharmakīrti then makes explicit, in contradiction with the Yogācāra reading that is then summarized in verse 339–a verse that Manorathanandin introduces by saying, ‘Alternatively, even given the Yogācāra position, there is no contradiction’: When it’s accepted that cognition comprises the object (jñānam : savis :ayam)– i.e., because of the establishment of the object, misleadingly, in a phenomenal aspect (ākāra) which is part of cognition–in that case, the ascertainment (niścaya), i.e. awareness, of an object just is the experience of itself, i.e., of an aspect of awareness. Hence, even on the Yogācāra position, there is no contradiction [in saying that] the phenomenal aspect of an object is the pramān :a, awareness of the object is the result.51 That is, the proponent of Yogācāra, too, can say that ‘awareness of the object is the result’ of a cognition; for this is not to say (what the Yogācāra would deny) that the object itself is thus given to awareness. It is, rather, only to say that an awareness is occurrent, having as its content a certain representation–this awareness itself (its seeming as it does), however, is what is experienced. To say there is ‘awareness of an object,’ then, is–according to an analysis that the Sautrāntika, too, has been shown to accept–only to say that there is occurring a cognition whose status as a pramān :a consists in its seeming to be of some object. For the kinds of reasons we have canvassed in the foregoing survey of this discussion, though, that much can be said without committing us to any views at all

49 Shastri (1968, 201): yadā bahirarthavāde’pi paro bāhyo’rtha is :t :o ’nis :t :o ’pi vā nis :pannatadbhāvo bhāvanāvaśād vyavasthites :t :ānis :t :abhāvah : sarūpāyā vijñapter hetuh : san vis :ayo bhavati, tadā tasyā vijñaptes tathā is :t :ānis :t :ākāren :ânubhavo vis :ayasya cânubhava ucyate. tena vis :ayasārūpyam : pramān :am arthasam : vit phalam uktam : . Cf. Dunne’s translation of Dharmakīrti’s verse (2004, 277n): ‘When the object is considered to be other than the mind and established with a nature that is desired or not desired, then the object is the cause of the representation and the effect is the experience of that representation in that way, i.e., as desired or not desired.’ 50 See n.38, above. 51 Shastri (1968, 201): yadā jñānasyâm : śe ākāre viplavavaśāt arthasya vyavasthiter jñānam : savis :ayam ist :am, tadā ya ātmano jñānākārasyânubhavah : sa evârthasya niścayah : sam : vedanam is :yate. tataś ca vijñānavāde ’py arthākārah : pramān :am arthasam : vit phalam aviruddham : . Cf. Dunne’s translation (2004, 277n) of Dharmakīrti’s verse: ‘If awareness includes the object (yadā savis :ayam : jñānam : ) due to the positing (vyavasthiti) of the object as an aspect (am : śa) of awareness [and not as actually external], then the determination (viniścaya) of the object is just the awarenessexperience of itself.’

about the ontological status of external objects. The point in thus urging that there is no contradiction involved in the Yogācāra’s saying pretty much the same thing, then, is that the Sautrāntika’s epistemology already amounts to epistemic idealism; and insofar as Dharmakīrti’s concerns are exclusively epistemological, there is nothing more than that for Dharmakīrti to defend as a Yogācāra–only metaphysical arguments (which Dharmakīrti is typically not in the business of offering) would yield something more. What Dharmakīrti the Yogācāra finally argues in this section (as stated, e.g., in verses 349–350), then, is represented as something the Sautrāntika already holds: Therefore, it’s accepted that just this self-awareness is awareness of an object, since the nature of an object which intrinsically exists as external isn’t perceived; rather, only an aspect

of thought is known. Hence, only the perceiving which is awareness of that [i.e., of thought] is awareness of an object.... even though ultimately being self-awareness, [[[cognition]]] is thought to be awareness of an object.52 It is finally at this point that Dharmakīrti’s Yog ācāra proponent makes clear the game he has been playing. Confronted with the foregoing arguments, the exasperated Sautrāntika now wonders whether there is any longer a point in referring to external objects. His is, after all, a version of the intuitively plausible sort of empiricism that takes our thought to be constrained by a really existent external world; but if the Sautrāntika can, as Dharmakīrti’s Yog ācāra proponent thinks he has shown, allow that his epistemological representationalism essentially amounts to epistemic idealism, then reference to external objects seems to have been evacuated of the explanatory significance he took it to have. The Sautrāntika thus expresses this worry in verse 353: ‘Having given up [the idea that] contentful

cognition53 has the form of an object which is appearing some way, how could cognition be the apprehension of an object?’54 How, that is, could one admit the cogency of epistemic idealism, and yet still think that external objects have a significant explanatory role to play in understanding the determinacy of mental content? ‘So thinks the Sautrāntika,’ Manorathanandin comments, introducing the punch line in Dharmakīrti’s verse; ‘but the Yogācāra, thanking him for his assistance,55 says: “Exactly; I don’t know how, either; how can one say there is apprehension of an object?’”56 Dharmakīrti’sYogācāra’ point, then, is of course that having allowed the cogency of his arguments for epistemic idealism, one cannot any longer think of mental content as requiring reference to an objective world; rather, the whole point of these arguments just is, as the pramān :aphala doctrine makes explicit, that

52 Shastri (1968, 204): iti tasmāt sâivâtmasam : vid arthasam : vid is :t :ā, yatah : svarūpād bahirbhūto’rthātmā na dr :śyate, buddhyākāra eva tu vedyate. atas tadvedanadarśanam evârthavedanam : .... paramārthatah : svavid api satī arthavid matā. 53 ‘Contentful’ renders avabhāsinah : , ‘having an appearance’; ‘cognition’ is supplied by Manorathanandin as the antecedent of the genitive pronoun tasya, which is all that appears in Dharmakīrti’s verse. 54 Shastri (1968, 205): tadā tasya jñānasyâvabhāsino yathā kathañcid is :t :ānis :t :ādinā bhāsamānam artharūpam arthākāram : muktvā katham : kena prakāren :ârthasya grahah : syāt. 55 Literally, ‘considering his assistance’ (sāhāyyakam : manyamāna). 56 Ibid.: ...iti manyate Sautrāntikah : . Yogācāras tu tasya sāhāyyakam : manyamāna āha: [353c1-d:] satyam : na jāne ’ham apîdr :śam // satyam : na jāne ’ham apîdr :śam arthagrahah : katham iti.

Buddhist idealism, epistemic and otherwise: thoughts on the alternating perspectives of Dharmakīrti 21 cognition’s seeming to be ‘of’ something is the only pramān :a–is all, that is, that we can be sure of. The point all along, in other words, has been that awareness itself is conceptually prior to an objective world. This is a view according to which (as John McDowell has said) ‘the “inner” role of [for example] colour concepts is autonomously intelligible, and... [we can] explain their “outer” role in terms of the idea that for an “outer” object to fall under a colour concept is for it to be such as to cause the appropriate visual “inner experience”.’ (1996, 30) And this point, Dharmakīrti here amusingly says, is one that the Sautrāntika himself has, in effect, already made. Dharmakīrti’s Yog ācāra can, then, thus thank the Sautrāntika for his assistance in advancing the case for idealism insofar as a representationalist epistemology already amounts to epistemic idealism–and, to the extent that one’s concerns are (like Dharmakīrti’s) only epistemological, that is all the more ‘idealist’ any view can be.

Conclusion: a Philosophical Critique of Perception as Non-conceptual

At the beginning of this essay, I suggested that the close conceptual relation between Dharmakīrti’sSautrāntika’ and ‘Yogācāra’ moments was less surprising than particularly the committed empiricist might suppose. I suggested, indeed, that the intuitively plausible empiricism that makes it strategically advantageous for Dharmakīrti typically to argue from the Sautrāntika perspective can (as Donald Davidson has said of empiricism generally) be understood as ‘the view that the subjective (“experience”) is the foundation of objective empirical knowledge.’ I would like to conclude by briefly developing what I take to be the significance of Davidson’s observation. I want to suggest, in particular, that characteristically empiricist approaches may entail subjectivism just insofar as they suppose our knowledge can–indeed, that it must–be based on a finally non-conceptual acquaintance with the world. It is, I thus want to suggest, partly because of the characteristically Buddhist commitment to non-conceptual awareness–to the view (as

this commitment is elaborated by epistemologists like Dignāga and Dharmakīrti) that perception, in particular, is constitutively nonconceptual (kalpanāpod : ha)–that Buddhist philosophers are almost ineluctably drawn to idealism. It is the foundational status, for these Buddhists, of putatively nonconceptual perception that makes it a small (and perhaps inevitable) step from the epistemological representationalism of Sautrāntika to the idealism of Yogācāra. The aptness of Davidson’s observation can, I think, be elaborated with reference to the basically Kantian insight that finally informs Sellars’s critique of the ‘myth of the given.’ Telling the story this way, it makes sense to start with an observation that is central to Kant’s development of his whole philosophical project: ‘It must be possible for the “I think” to accompany all my representations; for otherwise something would be represented in me which could not be thought at all, and that is equivalent to saying that the representation would be impossible, or at least would be nothing to me.’ (1787, B131 –32) Kant thus introduces what he called the transcendental unity of apperception–the idea, that is, that experience is intelligible only as had from some perspective. I cannot, in other words, wonder, with respect to any experience I am having, whether it is my experience; for it is a condition of the

possibility of anything’s being my experience that it be (somehow) reflexively ‘indexed’ to my perspective.57 But something more is expressed by the foregoing formulation of Kant’s point: To say that an experience must be such as can be taken as the content of a judgment– that, as Kant has put it, ‘It must be possible for the “I think’” to individuate the experience, which must therefore admit of expression as the object of a ‘that’-clause: ‘I think that X’–is also to say that any genuinely contentful experience must be such as can be brought into what Sellars called the ‘logical space of reasons.’58 That is, experience can only be thought to have any epistemic content to the extent

that it could, at least in principle, figure (whether as premise or as reason) in the inferential activity of justifying beliefs; it must, to put the point more generally, at least in principle admit of conceptually structured expression. But this is of course to affirm what Buddhists like Dignāga and Dharmakīrti are perhaps most concerned to deny; to say that only what can thus find expression as the content of a judgment is a possible experience just is to say that our conceptual capacities must always already be in play in any ‘experience’ worth the name. This way of putting the point reflects John McDowell’s characteristic way of developing the Kantian line of argument behind Sellars’s influential critique. McDowell, I suggested, gives expression to something essentially similar to the trend of Dharmakīrti’s Yog ācāra argument vis-à-vis Sautrāntika when he talks of views in which ‘the “inner” role of [for example] colour concepts is autonomously intelligible.’ The problem with such an account, McDowell continues, is that ‘we might manage to externalize at best a propensity to induce the relevant feature of “inner experience” in us.’ (1996, 31) One of McDowell’s most prominently recurrent lines of argument, though, would have us ask whether we really are ‘entitled to characterize [...intrinsically] inner facts in content-involving terms–in terms of its seeming to one that things are thus and so–at all.’ (1998c, 243) McDowell’s point is thus to ask whether we can really make sense of cognition as having any content at all on views according to which cognition is autonomously intelligible–on views, that is, according to which questions of what cognition is of are taken to be extrinsic to its character, which is thought instead to be intelligible entirely in terms of its intrinsic properties. In this connection, our culminating passage from Dharmakīrti expressed the Sautrāntika’s dawning realization that his own epistemological representationalism amounts to his ‘having given up [the idea that] contentful (avabhāsin) cognition has the form of an object which is appearing

57 On the possibility that at least some Buddhists (notably, Śāntaraks :ita) understood the point of the svasam : vitti doctrine to be similar, see Arnold (2005). 58 Cf. Sellars (1997, 76): ‘ characterizing an episode or a state as that of knowing, we are not giving an empirical description of that episode or state; we are placing it in the logical space of reasons, of justifying and being able to justify what one says.’ John McDowell has concisely stated the sense in which Kant’s passages concerning the transcendental unity of apperception lead to the Sellarsian point: ‘Experiences in which the world is disclosed are apperceptive. Perception discloses the world only to a subject capable of the “I think” of apperception.’ And ‘if an experience is world-disclosing, which implies that it is categorically unified, all its content is present in a form in is suitable to constitute contents of conceptual capacities.’ (1997, 346–348) Buddhist idealism, epistemic and otherwise: thoughts on the alternating perspectives of Dharmakīrti 23

some way.’59 (Dharmakīrti’s Yog ācāra, we saw, chimes in at that point with his clincher: To give that up just is to have reached, epistemologically, the idealist’s position.) What McDowell would have us ask, though, is whether it really is intelligible, after all, to think that cognition could be ‘contentful’ (avabhāsin) on such a view. To the extent that McDowell gives good reasons to doubt the intelligibility of this, it can be asked whether the intuitively plausible commitments of the Sautrāntika’s broadly empiricist approach are, after all, so epistemologically innocent. What, then, might we reasonably think is wrong with the basic idea that Dharmakīrti is thus able to exploit? Whatreasons,thatis,mightMcDowellgiveusfordoubtingtheintelligibility of experience that is ‘contentful’ without also being, intrinsically, ‘world-disclosing’? ‘Consider,’ McDowell suggests, ‘Kant’s advance over Hume. Hume inherits from his predecessors a conception according to which no experience is in its very nature– intrinsically–an encounter with objects.’ (1998a, 344) Hume, that is, inherits from empiricist predecessors such as Locke the idea–comparable to the one developed here with reference to Dignāga and Dharmakīrti–that insofar as we can always doubt whether experience is really of what it seems to be of, we ought therefore to derive our account of knowledge simply from what seems to be the case; the seeming itself, as we have seen Dignāga and Dharmakīrti put the point, is the only pramān :a. Hume famously drew skeptical conclusions from the

predicament he thus took his predecessors to have shown. According to McDowell, however, ‘Kant does not miss Hume’s point. He builds on it: since there is no rationally satisfactory route from experiences, conceived as, in general, less than encounters with objects–glimpses of objective reality–to the epistemic position we are manifestly in, experiences must be intrinsically encounters with objects.’ (Ibid.) On this reconstruction of Kant’s basic transcendental argument, then, the point is that just insofar as we are in ‘the epistemic situation we are manifestly in’–insofar as we manifestly are aware of a world–the fact that there is no viable inference from its merely seeming so to its really being so just means that this cannot be the right direction of explanation. It is, in other words, a condition of the possibility of our experience’s being of an external world at all that it be intrinsically so; awareness can only intelligibly have content at all, on this view, to the extent that (as McDowell elsewhere puts it)

cognitive space’ intrinsically incorporates ‘the relevant portions of the “external” world.’60 Insofar, then, as we cannot get (from its merely seeming so) to the idea that awareness is of a world, it must always already be so–experiences, contra Dharmakīrti, ‘must be intrinsically encounters with objects.’ Of course, it is precisely the claim that experience is of an external world that Dharmakīrti the Yogācāra would have us question; any argument, then, which takes this as ‘the epistemic position we are manifestly in’ would surely be seen by Dharmakīrti as begging the question. Nonetheless, Dharmakīrti may himself be begging the question insofar as he would advance his case for the Yogācāra position by helping himself to the idea that awareness seems to be of a world (that it has 59 See note 54, above. 60 McDowell (1998b, 257–58). I take McDowell’s scare quotes here not to query or ‘bracket’ the external world, but to question the idea of its being definable as constitutively external to our supposedly ‘internal’ cognitive space.

content) at all. On McDowell’s analysis, that is, the very idea of its seeming so is only intelligible given that it is so–this is the point in McDowell’s urging that ‘reality is prior, in the order of understanding, to appearance.’ (1998b, 410) Dharmakīrti’s thought, then, that awareness is autonomously intelligible–that we are, as we saw him put it, only immediately acquainted with awareness’s having ‘the qualification which is awareness’61–may itself presuppose the kind of view he aims to demonstrate.62 Dharmakīrti may, to that extent, be advancing his case by exploiting epistemological intuitions that are not, after all, innocent of the view he means to support thereby; for if McDowell’s arguments are cogent, then one could suppose (as Dharmakīrti does) that mental content is autonomously intelligible only if one already holds that awareness is not

constitutively of a world. But if that is right, then these epistemological intuitions cannot be taken independently to recommend that very conclusion. This, finally, is why it becomes relevant particularly to question Dharmakīrti’s commitment to the non-conceptual character of perception; for McDowell has been perhaps most concerned to argue that if we are coherently to think our awareness is of the world, our conceptual capacities must always ‘have already been brought into play, in the content’s being available to one, before one has any choice in the matter.... conceptual capacities, capacities that belong to spontaneity, are already at work in experiences themselves, not just in judgments based on them....’ (1996, 10, 24) That is, the objectivity of our knowledge is really only intelligible to the extent that the content of our awareness is always already suitable currency for Sellars’s ‘logical space of reasons’; otherwise, we cannot coherently think that perception could ever give us reasons for anything.63 The same arguments, then, that McDowell takes to recommend the view that ‘experiences must be intrinsically encounters with objects’ recommend, as well, the conclusion that even in sensory perception, our conceptual capacities must always already be in play. It would, of course, be a philosophical life’s work to make a satisfying case for anything like the foregoing line of thinking, and I do not begin to take myself, in here sketching it, to have established its truth over and against the position I have attributed to Dharmakīrti and his philosophical fellow travelers. I do, however, hope thus to have suggested some ways of understanding the close conceptual connection 61 See note 32, above. 62 For a different (but, I think, related) argument to the same effect, see Ram-Prasad (2002, 52 –64). While it seems to me that Ram-Prasad (who here follows Śan :kara) errs in taking Vasubandhu’s use (in the Vim : śatikā) of the dream analogy as part of his demonstrative argument for idealism–Vasubandhu’s main argument, I have said, is at verses 11–15 of that text, and I take the appeal to the example of dreams to advance a subsidiary point–he has read Śan :kara as making a similar argument about the conditions of the intelligibility of any concept of externality. 63 Dharmakīrti’s commentator Dharmottara arguably agrees with this thought, even if his avowed

faithfulness as a commentator precludes his putting it so baldly; for on Dharmottara’s reading, the point of the Buddhist pramān :aphala doctrine appears very different than as elaborated (vis-à-vis svasam : vitti) by Dignāga and Dharmakīrti. For Dharmottara, the point is that pramān :a really denotes the ‘result of the pramān :a’ (pramān :aphala) in the sense that only when cognition issues in a resulting judgment is there any epistemic content. On this point, see Arnold (2008a). Cf., as well, Dreyfus (2007, 107n), who says, based on ‘using Dharmottara’s ideas rather than Dharmakīrti’s,’ that ‘it looks as if Dharmakīrti is bound to maintain that perceptual aspects are transformed almost retrospectively into representations by conceptions’.

Buddhist idealism, epistemic and otherwise: thoughts on the alternating perspectives of Dharmakīrti between the ‘Sautrāntika’ and ‘Yogācāra’ perspectives as those are developed in Dharmakīrti’s Pramān :avārttika. I have, in this regard, suggested that the characterization of Dharmakīrti’sYogācāra’ arguments in terms of ‘epistemic idealism’ may, in the end, fail to distinguish these from the ‘Sautrāntika’ arguments he more typically deploys; for a significant point of the arguments developed in the avowedly ‘Yogācāra’section of the text just is to show that the Sautrāntika’s position already amounts to epistemic idealism–to the epistemological view, purportedly neutral with respect to ontological commitments, that what we are immediately aware of is only mental existents. To the extent that the trajectory of argument I have now briefly sketched following McDowell is cogent, there is reason to think this may all derive from the characteristically Buddhist commitment to the non-conceptual character of our selfawareness.

The thought, that is, that uninterpreted sensations (rather than judgments) represent the basis of our experience leads, on this reading, to the interiorizing of awareness; what is uniquely indubitable, from the perspective of such a view, is finally only the character of occurrent awareness as awareness. On the contrasting view I have commended the intrinsically objective (the ‘world-disclosing’) character of our experience necessarily requires reference to such constitutively intersubjective things as concepts and discourse–to the conceptual capacities in virtue of which we are ‘minded.’ Of course, the basic commitment to nonconceptual thought is not one that would be lightly given up by these Buddhists; it is, indeed, a part of the deep grammar of the tradition, and it is surely central at least to some workings-out of basic Buddhist insights.64 This is, most

basically, because the self is the originating example of the kind of conceptually projected abstraction these Buddhists mean to refute. Insofar, that is, as their main target is the conviction that our episodic cognitions represent the states of an enduring self, Buddhists like Dharmakīrti have a strong stake in thinking it possible to have a more basic sort of awareness that does not thus implicate this idea–in thinking it possible, that is, to have an immediate acquaintance only with the cognitive episodes, without also entertaining (what these Buddhists take to be) the inferential belief that these must be the states of a self. There is, then, for this and other reasons, surely more to be said for the Buddhist position here developed; the issues in play here are very far from being settled by the considerations I have introduced. My present purpose has only been to elaborate what I take to be philosophically interesting reasons for thinking that it may be just as we should expect that Sautrāntika should lead inevitably to Yogācāra.

64 But not to all recognizably ‘Buddhist’ workings-out of these insights; I have tried to argue that MadhyamakamayrepresentanalternativeelaborationofBuddhistinsightsthatdoesnotdependonthekinds of presuppositions I have here tried to show to be problematic. See, inter alia, Arnold (2008c).


Arnold, D. (2005). Is svasam : vitti transcendental? A tentative reconstruction following Śāntaraks :ita. Asian Philosophy, 15/1, 77 –111. Arnold, D. (2008a). Dharmakīrti and Dharmottara on the intentionality of perception: Selections from ‘an epitome of philosophy’ (Nyāyabindu). In J. Garfield & W. Edelglass (Eds.) Buddhist philosophy. New York:

Oxford University Press. (forthcoming). Arnold, D. (2008b). Svasam : vitti as methodological solipsism: ‘Narrow content’ and the problem of intentionality in Buddhist philosophy of mind. In J. Garfield, M. D’Amato & T. Tillemans (Eds.), Pointing at the moon: Buddhism, logic, analytic philosophy. (forthcoming). Arnold, D. (2008c). Madhyamaka critiques of epistemology. In M. Kapstein (Ed.) The Columbia guide to Indian philosophy. New York: Columbia University Press. (forthcoming). Brentano, F. (1973). Psychology from an empirical standpoint. Ed. O. Kraus; translated by Antos C. Rancurello, D. B. Terrell, and Linda L. McAlister. London: Routledge and Kegan Paul. Cabezón, J. (1990). The

canonization of philosophy and the rhetoric of Siddhānta in Indo-Tibetan Buddhism. In P. Griffiths & J. Keenan (Eds.) Buddha nature: A festschrift in honor of Minoru Kiyota (pp. 7–26). Reno, Nevada: Buddhist Books International. Davidson, D. (2001). Subjective, intersubjective, objective. New York: Oxford University Press. Dreyfus, G. (1996). Can the fool lead the blind? Perception and the given in Dharmakīrti’s thought. Journal of Indian Philosophy, 24, 209–229. Dreyfus, G. (1997). Recognizing reality: Dharmakīrti’s philosophy and its Tibetan interpreters. Albany, NY: SUNY Press. Dreyfus, G. (2007). Is perception intentional? A preliminary exploration of intentionality in Dharmakīrti. In

B. Kellner, et al. (Ed.) Pramān :akīrti : Papers dedicated to Ernst Steinkellner on the occasion of his 70th Birthday. Part 1 (pp. 95–113). Wien: Arbeitskreis für Tibetische und Buddhistische Studien Universität Wien. Dreyfus, G., & Lindtner, C. (1989). The Yogācāra philosophy of Dignāga and Dharmakīrti. Studies in Central and East Asian Religions (Journal of the Seminar for Buddhist Studies, Copenhagen and Aarhus), 2, 27 –52. Dunne, J. (2004). Foundations of Dharmakīrti’s philosophy. Boston: Wisdom Publications. Ganeri, J. (1999). Self-intimation, memory and personal identity. Journal of Indian Philosophy, 27, 469–483. Gnoli, R. (Ed.) (1960). The Pramān :avārttikam of Dharmakīrti: The first

chapter with the autocommentary: Text and critical notes. Serie Orientale Roma, XXIII. Rome: Istituto Italiano per il Medio ed Estremo Oriente. Griffiths, P. (1986). On being mindless: Buddhist meditation and the mind-body problem. LaSalle, IL: Open Court. Hattori, M. (1968). Dignāga, on perception, being the Pratyaks :apariccheda of Dignāga’s Pram

ān :asamuccaya from the Sanskrit fragments and the Tibetan versions. Cambridge: Harvard University Press. Hayes, R.P. (1988).Dignāga on the interpretation of signs. Dordrecht: Kluwer Studies in Classical India, 9. Kant, I. (1781, 1787). Critique of pure reason. Translated by Norman Kemp Smith. Unabridged edition. New York: St. Martin’s Press, 1965.

Kapstein, M. (2001). Mereological considerations in Vasubandhu’sProof of idealism’. Reason’s traces: Identity and interpretation in Indian & Tibetan Buddhist thought pp. 181–204. Boston: Wisdom Publications. Kolbert, E. (2007). Crash course: Can a seventeen-mile-long collider unlock the universe? The New Yorker, May 14, 2007; available online at < 070514fa_fact_kolbert>. Lusthaus, D. (2002). Buddhist phenomenology: A philosophical investigation of Yogācāra and the Ch’eng

Wei-shih lun. New York: RoutledgeCurzon. McDowell, J. (1996). Mind and world: With a new introduction. Cambridge: Harvard University Press. McDowell, J. (1997). What myth? Inquiry, 50/4, 338–351. McDowell, J. (1998a). The content of perceptual experience. Mind, value, and reality pp. 341–358. Cambridge, MA: Harvard University Press. McDowell, J. (1998b). Knowledge and the internal. In Meaning, Knowledge, & Reality (pp. 395–413). Cambridge, MA: Harvard University Press. McDowell, J. (1998c). Singular thought and the extent of inner space. Meaning, Knowledge, & Reality, 228–259.

Buddhist idealism, epistemic and otherwise: thoughts on the alternating perspectives of Dharmakīrti 27

Ram-Prasad, C. (2002). Advaita epistemology and metaphysics: An outline of Indian non-realism. London: RoutledgeCurzon. Sām :kr :tyāyana, R. (Ed.) (1938–1940). Dharmakīrti’s Pramān :avārttika with a commentary by Manorathanandin. Journal of the Bihar and Orissa Research Society, 24/3–26/3. Schmithausen, L. (2005). On the problem of the external world in the Ch’eng wei shih lun. Studia Philological Buddhica Occasional Paper Series, XIII. Tokyo: The International Institute for Buddhist Studies. Sellars, W. (1997).

Empiricism and the philosophy of mind: With an introduction by Richard Rorty and a study guide by Robert Brandom. Cambridge: Harvard University Press. Shastri, S. D. (Ed.) (1968). Pramān :avārttika, Ācāryamanorathanandivr :ttiyutam. (Dharmakīrtinibandhāvali , vol. 1). Varanasi: Bauddha Bharati. Steinkellner, E. (2005). Dignāga’s Pramān :asamuccaya, Chapter 1: A hypothetical reconstruction of the Sanskrit text with the help of the two Tibetan translations on the basis of the hitherto known Sanskrit fragments

and the linguistic materials gained from Jinendrabuddhi’s T :īkā, downloaded from <http://>; (accessed June 2, 2005). Steinkellner, E., Krasser, H., & Lasic, H. (Eds.) (2005). Jinendrabuddhi’s Vi śālāmalavatī Pramān :asamuccayat :īkā : Chapter 1, Part I: Critical Edition. Beijing and Vienna: China Tibetology Publishing House, Austrian Academy of Sciences Press. Taber, J. (2003). Dharmakīrti against physicalism. Journal of Indian Philosophy, 31, 479–502. Tillemans, T. J. F. (2003). Metaphysics for Mādhyamikas. In G. Dreyfus, & S. McClintock (Eds.) The Svātantrika-Prāsan :gika distinction: What difference does a difference make? (pp. 93–123). Boston: Wisdom Publications.