Buddhist Idealism

Bronwyn Finnigan School of Philosophy Australian National University

Idealism has been a prominent philosophical view in Indian Buddhist thought since the 4th century CE. It was a topic of considerable debate for centuries amongst Buddhist and non-Buddhist philosophers in India and Tibet. It also had a significant influence on the intellectual culture of China and Japan. Much can be gained from contemporary engagement with these arguments. It has the potential to reveal cross-cultural antecedents to contemporary views, new arguments that could be re-mobilised in current debates, and may challenge the presuppositions that frame Western discussions of idealism by providing alternatives.

There are several ways to approach such a study. One approach is comparative; taking the Western philosophical tradition as one’s starting point and considering the extent to which Buddhist arguments might advance or depart from established views. While there are many potential benefits of this approach, it runs the risk of reading Western philosophical commitments into Buddhist thought rather than drawing a genuine comparison. There is also considerable debate on either side of the comparative divide about the nature and presuppositions of idealistic arguments, issues which would need to be carefully addressed and potentially resolved for a genuine comparison.

1 I would like to thank Dan Arnold, Jay Garfield, Jonathan Gold, Shinya Moriyama, Tom Tillemans and the editors of this book for helpful comments on an earlier draft of this chapter. Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

This chapter will take a more modest approach. It will survey some of the most influential Buddhist arguments in defense of idealism. It will begin by clarifying the central theses under dispute and will rationally reconstruct arguments from four major Buddhist figures in defense of some or all of these theses. It will engage arguments from Vasubandhu’s Viṃśikā (or Viṃśatikā) 2 and Triṃśikā (or Triṃśatikā); Dignāga’s Ālambanaparīkṣā; the sahopalambhaniyama inference developed by Dharmakīrti; and Xuanzang’s logical argument that intrigued philosophers for centuries in China and Japan. As will be seen, the arguments themselves are often presented in the form of brief and sometimes obscure inferences or syllogisms. This chapter will attempt to clarify what is being argued and motivate these arguments in terms of their presuppositions. These presuppositions range from views about the nature of mind and metaphysics to epistemology and logic. By making this context explicit, this chapter will introduce central ideas in Buddhist philosophy and suggest ways in which they were mobilised in support of an idealist conclusion.

1. Preliminary Background The historical Buddha arguably lived and taught in India somewhere between 6-4th centuries BCE. Philosophical debates about how best to articulate and defend the truth of his teachings reached their scholarly peak in India between 4-9th centuries CE. These debates spanned issues in metaphysics, epistemology, the nature of mind and logic and resulted in a complicated and diverse array of philosophical positions. Later doxographers classified the major philosophical trends of this period into four Buddhist schools: Vaibhāṣika, Sautrāntika, Yogācāra, and Madhyamaka.3

See Kano (2008) for reasons to prefer the title Viṃśikā 3 Although I shall employ this doxographical terminology in this chapter, current scholarship calls into question the distinctiveness of these schools, drawing out continuities in their Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

Idealism is traditionally associated with Yogācāra, otherwise known as Vijñānavāda (‘the doctrine of consciousness’) or Cittamātra (the ‘mind-only’ school). ‘Idealism’ can stand for a range of philosophical positions, however, and there is much contemporary dispute about whether and in what sense Yogācāra arguments should be conceived as idealist. In the Buddhist context, idealism is associated with (at least) three philosophical theses:

(T1) What we take to be external objects are merely appearances in conscious awareness. (T2) Appearances in conscious awareness are mentally caused. (T3) The mental cause of appearances is more fundamentally real than appearances.

An ‘external object’ (bāhyārtha) is to be understood as something that exists in mind independent, physical reality. The Buddhist doctrine of ‘mind only’ was traditionally conceived to accept all the above theses. It also took T1 to entail that external objects do not exist and T2 to entail that everything is mental. It was thus understood as a strong form of metaphysical idealism. There is both contemporary and historical dispute, however, about whether central Yogācāra arguments entail all of these theses (or, if only some, which ones) and whether they (were intended to) have these implications. One source of contemporary dispute concerns T1. As we will see, all the Yogācārins surveyed accept T1. Without further qualification, however, T1 might seem to be consistent with representational realism. Some contemporary scholars insist that this possibility is left open by the fact that some Yogācāra arguments for T1 are epistemological rather than metaphysical. As such, they demonstrate that external objects are unknowable but do not entail that they neither exist nor are causally related to philosophical commitments as well as a complex interplay of ideas amongst those so categorised. See Kritzer (2003), Arnold (2008), Gold (2015a). Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

appearances in the mind.4 Others insist that epistemological arguments do, and were intended to, entail this metaphysical conclusion.5 Another source of dispute concerns the relation between T2 and T3. As we will see, while many Yogācārins articulate T2, they do not explicitly state T3. It was traditionally understood that, for Yogācāra, T3 was entailed by T2.6 Some contemporary scholars contest this entailment.7 There is also much dispute about which of these theses warrant the English term ‘idealist’. It is impossible to navigate these disputes without first having some idea of the Buddhist arguments under discussion. However, it is difficult to articulate these arguments without adopting some interpretative stance. Moreover, a thorough examination of the issues relevant to each argument could easily constitute the subject matter of a book (or two). This ‘modest’ survey thus turns out to be not so modest. Nevertheless, it would be helpful for those interested in Buddhist forms of idealist arguments to have some overview of the central arguments as a starting point. To navigate these issues, this chapter will adopt the following methodology. It will rationally reconstruct Buddhist arguments that were traditionally understood as defending idealism and will consider the extent to which they can be read as accepting and arguing for T1-T3. It will set aside the question of which of these theses best deserves to be called ‘idealist’ but will treat Buddhist mind-only metaphysical idealism as the conjunction of all three.

See Hayes (1988), Oetke (1992), and Lusthaus (2002) 5 This will need to be nuanced. See the section on Dharmakīrti. 6 Tibetan Madhyamaka philosophers were particularly insistent on this point. Madhyamaka was the main philosophical rival to Yogācāra. According to the Mādhyamika Candrakīrti, all existence is mind-dependent in the sense that entities only ever exist conventionally (saṃvṛtisat) and it is the mind that produces conceptual or conventional designations. From this perspective, both Yogācāra and Madhyamaka agree that objects only ever have mind-dependent ontological status and thus both accept T2. However, Tibetan Madhyamaka thinkers frequently insisted on the Yogācāra acceptance and Madhyamaka denial of T3 as a point of differentiation. See Tillemans (2007: 65). There is reason to think, however, that the distinction between Madhyamaka and Yogācāra may not be so neatly drawn. See Garfield & Westerhoff (2015) for a recent philosophical discussion. 7 See Gold (2015a,b) Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

2. Vasubandhu’s Idealism: An Articulation and Defense Vasubandhu (fl. c.360) was one of the most influential philosophical figures in the development of Buddhist scholarship, contributing seminal texts to three Buddhist scholarly traditions.8 His Twenty Verses and Autocommentary (Viṃśikā & Vṛtti), Thirty Verses (Triṃśikā) and the Three-Natures Treatise (Trisvabhāvanirdeśa) are some of the most important Yogācāra texts.9 The Twenty Verses, in particular, was traditionally read as providing seminal arguments for idealism. It opens with the following verse:

All this is mere appearance of consciousness Because of the appearance of non-existent objects Just as a man with an eye disease Sees non-existent hairs (Viṃś.1)10

The first line straightforwardly states T1. Everything we take to exist in mindindependent reality11 is to be understood as a mere appearance in conscious awareness, just as we understand the distorted images seen by someone with an eye disease to be merely appearances in their conscious awareness. The verse also suggests that we should treat external objects as unreal or non-existent in the same sense in which these distorted

8 See note 3. 9 English translations of all three texts can be found Anacker (1998) and Kochumuttom (1982); English translations of both the Twenty Verses and the Three-Natures Treatise can be found in Tola & Dragonetti (2004), and an English translation of the Three-Natures Treatise alone can be found in Tola & Dragonetti (1983), Wood (1991) and Garfield (2009). 10 There is some dispute about how best to translate this first verse. Some variants can be found in Gold (2015a,b), Garfield (2002: 157), Anacker (1998: 161), Kochumuttom (1982: 260), and Kellner & Taber (2014: 735). The latter query whether this verse was actually written by Vasubandhu or whether it is a summary of the first section of Vasubandhu’s Autocommentary. 11 See Vasubandhu Viṃś.1 vṛ.1 “In the Mahāyāna system it is understood that those belonging to the three worlds are mere appearances of consciousness.” The three realms are the three states into which Buddhists believe living beings may be reborn, understood as constituting the entire universe. Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

images are treated as non-existent.12 What are Vasubandhu’s arguments for this view? His most well known argument is found in Viṃś.11-15 where he attempts to disprove the possibility of external objects. His realist opponents took a mereological approach to ontology. The orthodox Hindu school of Nyāya-Vaiśeṣika maintained that wholes exist in addition to the unified collection of their parts. Buddhist Vaibhāṣika and Sautrāntika were atomistic nominalists and insisted that while the basic atomic constituents of aggregations are real, their aggregation is not.13 Vasubandhu argues against each mereological possibility as a way of accounting for the objects in our sensory experience. This is meant to substantiate T1 as the best (and only remaining) alternative.

A version of this metaphysical argument is revisited by Dignāga and so I shall delay its discussion until the following section. My discussion of Vasubandhu will instead focus on his responses to three initial objections as well as his innovative three nature analysis of conscious experience. Vasubandhu’s responses to these objections are not yet attempts to prove T1 but merely show its viability because not disproved. They are significant, however, in drawing out his commitment to T2. His three-nature analysis, moreover, draws out important Buddhist ideas about the nature of mind which raise interesting issues concerning T3. The first objection concerns how T1 can account for the spatio-temporal dimensions of experience. If (e.g.) color is not caused by contact with an external object, how does one account for the regularity with which it arises in certain times and places and not arbitrarily everywhere? The second objection concerns how it can account for the intersubjectivity of experience. What explains the fact that most people do not experience

12 What exactly that sense is will be discussed later. 13 For a discussion of the differences between Vaibhāṣika and Sautrāntika see Williams (1981), Gethin (1998), Ronkin (2005, 2015) Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

the distorted images seen by someone with an eye disorder. According to the realist, these images count as illusory relative to the normal sense perception of those with undistorted vision, who have similar experiences when placed in similar circumstances. How can an idealist account for this similarity and difference? The third objection concerns the efficacy of our experiences in relation to actions and their subsequent effects. In waking life, seeing water and feeling thirst can prompt an agent to drink and subsequently feel satiated. How can this be explained within an idealist framework? (Viṃś.2)

Vasubandhu responds to the first objection by pointing out that these same regularities can hold in dreams. It is not the case that (e.g.) colors experienced in dreams appear at all times and everywhere (Viṃś.2 vṛ.2). If a blue wren appears in a dream, the blue of its feathers typically remains on its feathers as it flits here and there and there. Later in the argument the realist opponent proposes that such regularity is conditioned on the recollection of some prior sense experience of an external object (Viṃś.17). Vasubandhu responds by arguing that appeal to memory does not prove the existence of an external object as the recollected experience could itself have been merely an appearance in conscious awareness (Viṃś.17 vṛ.17). This is not meant to prove that waking experience is a dream. Rather, since dreams (which the opponent is assumed to agree are mentally caused mental phenomena) have the properties cited in the objection, the presence of these properties in waking experience does not prove that their bearers must be caused by external objects. The objection thus fails to disprove T1. Vasubandhu’s response to the second objection is more obscure and involves reference to pretas or hungry ghosts. Pretas are unfortunate beings which experience and are insatiably compelled to consume what they take to be rivers filled with pus, excrement and urine. Their similarity of experience is explained in terms of the similarity Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

of their karma; it is the karmic fruit of similar past negative actions. If one denies the laws of karma, this reason will not be compelling. But, even if accepted, it would not be decisive as Vasubandhu’s realist opponent also accepts the laws of karma but insists that its fruit is the result of events and objects in a mind-independent reality with which the agent previously interacted. Vasubandhu provides an argument from simplicity against this view. 14 Since his opponent agrees that karma causes events that alter the consciousness of pretas, why not just allow that karma alters their consciousness? What need is there to posit additional elements? (Viṃś.6-7 vṛ 6) To appreciate this argument we need to understand something about karma. In his Commentary on the Treasury of the Abhidharma (Abhidharmakośabhāṣya), Vasubandhu tells us that karma is tied to intentions or volitions for action, which are psychological states (ABKh 4.1). Intentions are said to produce karmic seeds or latent dispositions that are both stored in and made manifest by an underlying mode of consciousness known as the store-consciousness (ālayavijñāna). When conditions ‘ripen’, these seeds bear their karmic fruit in the form of conscious experiences of various kinds. This much was accepted in the context of this debate. Vasubandhu’s point is that all of the relevant components of karma (its seeds, fruit, and operative cause) are either mental phenomena or explicable in terms of mind. Since there is no need to posit an external reality to explain the function of karma, it should not be introduced into its explanation. Since the function of karma explains the similarity of experience amongst pretas and, by analogy, experiential similarities and differences amongst ordinary experiencers, there is no need to posit an external reality to explain these facts and thus one should not introduce it into one’s explanation.15 This implies a commitment to T2. Appearances in conscious awareness are mentally caused.

Siderits calls this the ‘Principle of Lightness’ (2007: 157). Gold (2014) identifies it as an application of Occam’s razor. 15 See Siderits (2007: 157), Arnold (2012:.31, 178). Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

According to the third objection, appearances in waking experience relate to action in a way that those in dreams do not. There is a question-begging and non question-begging way to pose this objection. The former is as the claim that appearances of (e.g.) water in waking experience, unlike those in dreams, causally contribute to an agent performing actions (e.g., drinking water) that alter external reality (e.g., reduces the quantity of water) and cause new experiences (e.g., satiation). Vasubandhu provides a similarly question-begging reply, citing the case of ejaculated semen as the result of an erotic dream (vṛ 4). A non-question-begging way to pose this objection might be to insist that there is a regularity in the appearances of (e.g.) water, of oneself as drinking, of a reduced visible quantity of water and of a feeling of satiation which can only be explained by reference to a corresponding external reality. Vasubandhu’s response to this form of the objection is the same as to the first; such regularity can also be found in dreams and, thus, appeal to regularity does not disprove T1 (Viṃś.4). Vasubandhu does not aim to prove that sensory experience is a form of dreaming but uses the example of dreams to disprove objections to T1. However, as we have seen, the opening verse claims that we should treat sensory experience like the visual illusions of someone with an eye disorder. Visual illusions are similar to dreams insofar as both are mental phenomena. They are importantly distinct, however. Appearances in dreams are taken to be real from the perspective of the dreamer (Viṃś.17). The visual illusions seen by someone with an eye defect, however, are considered by them to be unreal or ‘non-existent’. Vasubandhu suggests that we should treat sensory experiences as similarly ‘non-existent’. What is the relevant sense of ‘non-existent’ in this case and against what is it measured? To understand this, it is helpful to consider Vasubandhu’s innovative ‘three-nature’ analysis of conscious experience advanced in his Three-Nature Treatise. Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

According to this analysis, all conscious experience has three natures: an imagined nature (parikalpita-svabhāva), a dependent nature (paratantra-svabhāva) and a perfected nature (pariniṣpanna-svabhāva). The imagined nature is the common sense way in which we understand our sensory experience; i.e. as the apprehension of external objects in mind-independent reality.16 Vasubandhu emphasises that there are two aspects to this experience; an objective aspect (an object or situation which is experienced) and a subjective aspect (the experiencing of that object). We ordinarily and pre-reflectively take both aspects to correspond to real entities in mind-independent reality. We consider there to be a substantively real self that is interacting with a mind-independent external reality. According to Vasubandhu, however, we are mistaken in both respects. That we are mistaken about the reality of self in addition to external objects might come as a bit of a surprise. This idea draws on the Buddhist doctrine of no-self (anātman). To deny the existence of self (ātman) is to deny that there is an enduring substance, wholly present from moment to moment, that exists separate from and as the owner of events in conscious awareness. Ātman is also thought to be the essential nature of individual persons and the referent of proper names. A central motif of Buddhist thought is to deny that such an entity exists. To deny that there is a self, in this sense, evidently leaves open the possibility of alternative conceptions of self. It also creates pressure to provide alternative explanations of that for which it is posited as explananda. These issues drive much Buddhist philosophical debate. For Vasubandhu, at least some elements of this notion of ātman are present in our common-sense understanding of sensory experience. We ordinarily and pre-reflectively assume that there is a perceiver who perceives objects and to whom perceptions appear. We also consider this perceiver

16 There are several forms of imagined nature (parikalpitasvabhāva) discussed in the Yogācāra school. I present one of two predominant characterisations. The other concerns the commonsense view that words refer to their referents in virtue of the referents’ own nature (svabhāva). See Tillemans (2007: 55). Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

to be ourselves. These assumptions are mistaken. The fact of this mistake is revealed by our understanding the dependent nature of experience; namely, that both the objective and subjective aspects of ordinary experiences are merely appearances in, and constructions of, the mind. What we take to be ourselves, the experiencer of experiences, is merely a subjective aspect of the experiences themselves. While these aspects of experience are real qua qualities of experiential states (our ordinary experiences really do have objective and subjective aspects) they are unreal qua elements of a mind-independent reality. The perfected nature consists in a thorough-understanding that the dual (subjective-objective) aspects of our ordinary experience are merely constructions out of an underlying mode of consciousness which, itself, is empty of subject-object duality.17 Now, there is a lot going on here and much to be said. But certain features of this three-fold analysis of experience are relevant to our discussion. Vasubandhu appeals to this analysis in his Thirty Verses to elaborate T1 and T2.18 For instance:

Depending on the conditions available, the five sense-consciousnesses, Together or separately, originate in the root consciousness, Just as waves originate on water (Triṃś.15)

This alteration of consciousness is just the distinction [between subject and object]; What is thus distinguished [as subject and object], does not exist Therefore this is all mere appearance of consciousness (Triṃś.17)

For discussions of Vasubandhu’s three-nature theory, see Garfield (2002: Chapters 6-8), Gold (2015a,b), Kochumuttom (1982). 18 This need not imply that the three-nature theory, as presented in TSN, either presupposes or was intended to substantiate these theses. See Kochumuttom (1982), Lusthaus (2002) and Garfield (2015a,b) who argue that TSN is best read as a form of phenomenology. Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

The last sentence in this verse mirrors the opening line of Vasubandhu’s Twenty Verses. The three-nature analysis also helps clarify the relevant sense of ‘all’ in this initial claim. It refers not merely to what we take to be external objects in our conscious experience but also to what we take to be a real, perceiving self. Both are to be considered mere appearances in conscious experience. While we have good reason to think that Vasubandhu accepts T1 and T2, he does not assert T3. To count as a ‘mind only’ metaphysical idealist he needs to accept all three theses. Whether he accepts or presupposes T3 turns on how one understands the underlying mode of consciousness. Vasubandhu’s three-nature analysis was traditionally read as the view that external objects and real selves are conventionally unreal (they are false constructions to be rejected), the objective and subjective aspects of experience are conventionally real (they are genuine appearances constructed by mind), but the underlying mode of consciousness which constructs these appearances has a more fundamental or ultimate reality. Thus T3. However, Vasubandhu also claims that the perfected nature is empty of all subject-object duality and, as such, is “inconceivable” (Triṃś.29). Some argue from this that it is not possible to conceive and thereby accept T3 from the ultimate perspective of perfected nature. It is thus inferred that in the final analysis Vasubandhu is not a ‘mind-only’ metaphysical idealist.19

3. Dignāga’s Ālambanaparīkṣā

Dignāga (ca. 480-540 CE) was one of the initial formulators of Buddhist logic and made important contributions to the study of epistemology, perception, language, and mind.

See Gold (2015a,b). Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

Alongside the writings of Vasubandhu, Dignāga’s Ālambanaparīkṣā is a central Yogācāra text.20 Here, Dignāga argues against the possibility of external objects as understood by his Buddhist realist opponents, the Sautrāntika, and advances T1 as the best (and only remaining) alternative. The Sautrāntika defend a nominalist atomistic ontology according to which ordinary material objects (e.g., pots, cups, persons) are to be understood as aggregations of more basic atomic constituents. While the basic atomic constituents are considered to be real, their aggregation is not. Of the various criteria that are offered to establish this position, the most important is causal efficacy. Whatever causal efficacy we (conventionally) attribute to wholes is thought to be entirely reducible to that of its (ultimately real) parts.

Dignāga’s argument turns on the requirements for something to count as an ‘object’ (ālambana) of a conscious mental state. Dignāga, along with his Sautrāntika opponents, assumes that two criteria must be satisfied. First, the object must be causally related to our conscious mental state. Second, it must resemble the appearance in or content of that state. Dignāga’s first move is to demonstrate that these criteria cannot be satisfied by Sautrāntika realism. There are two initial possibilities available to a Sautrāntika to satisfy these criteria; namely, that the object of consciousness is either some basic atomic constituent or an aggregation thereof. Dignāga argues against both possibilities. Against the view that the objects of consciousness are basic atoms, Dignāga argues

20 An English translation and commentary of the Ālambanaparīkṣā can be found in Tola and Dragonetti (1982) and Duckworth et.al (forthcoming). Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

Even if the atoms were the cause of the cognition through the senses, Since the cognition does not appear under the form of those [[[atoms]]]

he atoms are not the object of that [[[cognition]]] In the same way that the sense-organs [are not] (Ālam.1a-d)21

Although the Sautrāntika allow atoms to be causally related to conscious experience, they do not appear in conscious awareness. What we experience are (e.g.) pots, cups and persons, not their atomic constituents.22 Thus, atoms may satisfy the criterion of causal efficacy but not the criterion of resemblance. In this sense they are analogous to our sense-organs; both condition our experience but do not appear in its content. Against the view that the object of consciousness is an aggregation of atoms, Dignāga argues:

That [[[appearance]] of an aggregate] does not arise from that [i.e. an aggregate] because it [the aggregate] does not exist as something real, in the same way as a second moon [does not exist as something real]. Thus, in both cases, [something] external cannot be the object of consciousness (Ālam.2a-d)

While an aggregation of atoms (in the form of a pot or cup etc.) can appear as the content of a conscious mental state, it is not causally related to this state because, according to the Sautrāntika, aggregates are unreal and unreal things cannot cause anything. Aggregates are like a second moon perceived by someone with an eye defect; both

Translation from Tola and Dragonetti (1982). 22 Vasubandhu similarly denies that the “object is experienced…as many single atoms” (Viṃś 11) for the reason that “single atoms are not apprehended.” (vṛ 11). Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

appear in conscious awareness but do not really exist and so cannot have produced this state.23 Aggregates thus satisfy the criterion of resemblance but not the criterion of causal efficacy. Since atoms and their aggregations are the only possibilities admitted by his realist opponent, Dignāga (sub) concludes that external objects are not the objects of consciousness.

One might object that this argument unduly restricts the explanatory options available to the Sautrāntika (or atomic realists more generally). Dignāga considers a further possibility; namely, that atoms can jointly produce a representation or image in conscious awareness. That is, while the image resembles the aggregation (thus satisfying the second criterion), it is explicable by the individual causal contributions of its many atomic constituents (thus satisfying the first criterion). The fact that ‘atomicity’ does not appear as the content of the experience, moreover, can be explained in the same way that solidity does not; i.e. it exists but is not perceptible (Ālam.3a-d). Against this view, Dignāga argues that it wouldn’t allow for the production of different images in conscious awareness (e.g. as of a pot or as of a cup) given that these distinct objects would be caused by the same kind of atoms (Ālam.4a-b). Moreover, he insists it would be inadequate for the realist to respond by grounding this distinction in differences in the size and shape of these objects because those aspects of the object are themselves reducible constructs of more basic atoms, which are uniformly identical (Ālam.4c-d, 5a b).

This argument seems to allow the possibility of atomic aggregation but denies its ontological status. Vasubandhu argues against the possibility of aggregating atomic constituents. This is a complex argument. One of its opening moves involves denying the conceptual possibility of impartite atoms conjoining into spatially extended objects for the reason that it would problematically presuppose that atoms have distinct ‘sides’ (i.e. one atom joining to the ‘left side’, another joining to the ‘right side’), where a side is a part (Viṃś 12 vṛ 12). Detailed studies of this argument can be found in Kapstein (2001) and Siderits (2007). See Gold (2015) for a helpful summary. Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

These arguments attempt to undermine the idea that sensory experience is explicable by a correspondence to external objects in mind-independent reality. Dignāga offers T1 as an alternative. The objects of mental states are simply the appearances in consciousness, the content of mental states, which appear as external but do not actually lie outside of the mind (Ālam.6a-c).24 How does this satisfy the two criteria for objects of

conscious experience endorsed by Dignāga? And is there any reason to think that Dignāga also affirms T2 or T3? These questions are difficult to answer. It seems that, for Dignāga, the criterion of resemblance is satisfied by the fact that an appearance in conscious experience resembles itself. The criterion of causal efficacy, however, is much harder to satisfy. As his opponent complains, how can the content of a conscious mental state cause that state? (Ālam.vṛ 21). Dignāga offers two responses. First, it is a cause in the sense of being a defining characteristic and thus a necessarily concomitant part of the conscious mental state (Ālam.7a). Consciousness depends on appearances or content to the extent that, without them, it would not count as consciousness (considered to be true by definition).25 Second, the appearances in consciousness may (also) be considered the actualised fruit of latent karmic dispositions that were produced by prior mental states with the same content (Ālam.7b). This second explanation satisfies both of Dignāga’s criteria for an object of consciousness insofar as the appearance in consciousness resembles the content of the state on which it (indirectly) causally depends; indirectly, because the appearance actualises a latent disposition that was caused by this original mental state. This demonstrates a commitment to T2 but Dignāga does not go so far as to assert T3.26

See Matilal (1974: 145) and Tillemans (2007: 61). 25 See Arnold (2012: .27). 26 For an alternative characterisation of the criterion and upshot of this argument, see Duckworth, et al (forthcoming). Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

4. Dharmakīrti’s Sahopalambhaniyama Inference Dharmakīrti (ca. 600-660 CE) was Dignāga’s most famous successor and, together, they are responsible for the Buddhist tradition of epistemology and logic known as Pramāṇavāda. This school provoked some of the most systematic and important philosophical debates in ancient India on the nature of metaphysics, mind, language and logic, and it decisively shaped the course of Mahāyāna Buddhist thought.27 The third chapter of Dharmakīrti’s Commentary on Epistemology (Pramāṇavārttika) is traditionally read as both articulating and defending a form of idealism. His conception of the thesis to be proved is most concisely stated in Ascertainment of Epistemology (Pramāṇaviniścaya):

[Apart from consciousness itself] there is nothing else which is experienced by the mind. Nor does the [[[mind]]] have any other experience. As there is nothing apprehended and no apprehender, it is the [[[mind]]] itself which illuminates itself (Pvin 1, k.38)28

Dharmakīrti here asserts commitment to T1. The idea that there is ‘nothing apprehended and no apprehender’ echoes ideas earlier discussed in the writings of Vasubandhu; namely, that there is no substantively real self that perceives a mind-independent external reality. Both the perceiving (subject) and the perceived (object) are simply appearances in conscious experience. The idea that ‘the mind illuminates itself’, however, reflects an original thesis introduced by Dignāga; namely, that conscious mental states ‘self illuminate’ or are ‘self-aware’ (svasaṃvedana or svasaṃvitti). There is much contemporary interest in the idea of self-awareness but no

See Tillemans (2011). 28 Translation from Tillemans (2007: fn.131). Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

uniformity in how it is best understood. 29 Some read it as Dignāga’s way of characterising and defending the idea that a subjective-aspect is a necessary constituent of ordinary consciousness.30 On this view, to say that a mental state is self-aware is just to say that it necessarily involves both an experienced object and the experiencing of that object. This can be understood in one of two ways; either as qualifying the object-aspect (i.e., to be an object of consciousness, the object must be experienced or illuminated)31 or as an awareness of both the experienced object and the experiencing of it. Dignāga’s view is arguably closer to the latter. Like a lamp that illuminates itself at the same time as it illuminates other things, Dignāga maintains that there are ‘three factors’ to conscious mental states: the subject-aspect (illuminating), the object-aspect (illuminated) and awareness of both.32 This might mean either that the experienced object and the experiencing of it are both apparent in the experience (a phenomenological claim) and/or it involves an additional understanding that the experienced and experiencing are both mere aspects of the mind (a metaphysical claim). 33 The latter resonates with

See Williams (2000), Ganeri (1999, 2012), Yao (2005), Arnold (2005), Garfield (2006), MacKenzie (2007, 2008), Kellner (2010, 2011b), Thompson (2011), Coseru (2012). 30 See Kellner (2010, 2011a). 31 This view is often attributed to Śāntarakṣita, who takes svasaṃvedana to be the criterion for sentience. 32 See Dignāga PS 1.10:

Whatever the form in which it [viz. the cognition] appears, that [[[form]]] is [[[recognized]] as] the object of cognition (prameya). The means of cognition (pramāṇa) and [the cognition which is] its result (phala) are respectively the form of subject [in the cognition] and the cognition cognizing itself. Therefore, these three [factors of cognition] are not separate from one another. (Hattori trans. 1968: 29)

33 See Hattori (1968: 46):

The idea that ālayavijñāna or store-consciousness appears or manifests itself (ābhāti, pratibhāti, avabhāti, khyāti) as both the apparent subject (sva-ābhāsa= grāhaka) and the apparent object (artha-ābhāsa = grāhya) is met elsewhere in Vijñanavāda treatises and we need not explain it here. It is clear that Dignāga established his theory of knowledge on the ground of the Vijñanavāda philosophy, although he does not use the term ālayavijñāna. The theory of self Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

Vasubandhu’s dependent nature of conscious experience. It also points towards the position Dharmakīrti will defend. Dharmakīrti’s most well known idealist argument turns on this notion of self awareness. It has come to be known as the sahopalambhaniyama inference because it argues from the ‘necessity’ (niyama) of things only ever being experienced ‘together’ (saha) with ‘experience’ (upalamba). The inference is neatly summarised as follows:

Blue and the consciousness of blue are not different, because they must always be apprehended together (PVin 1.54ab)34

Or, in a longer variant:

Because [something blue] is not apprehended without the additional qualification of consciousness, [and] because [blue] is apprehended when this [qualification of consciousness] is apprehended, consciousness [itself] has the appearance of blue. There is no external object by itself. (PV 3.335)35

There are at least two ways to read this inference, which correspond to two ways of understanding self-awareness. According to the first reading, one is never aware of blue other than being aware or conscious of it; one never apprehends an apprehended object other than by apprehending it. To establish that the experienced object and the

cognition (svasamvit) which marks the specific feature of his theory of direct knowledge is understandable only on the basis of the Vijñānavāda doctrine.

sahopalambhaniyamād abhedo nīlataddhiyuḥ / translation from Tillemans (2011: n.33). 35 Translation from Taber (2010: 291). Yet another version of this argument can be found in PV3.388-391, English translation of which can be found in Matilal (1974: 159-60). Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

experiencing of it are distinct, one would need a case where they are perceived apart. But there is no such case. Thus, the appearance of blue is just the content of a conscious state and is not distinct from it.36 According to the second reading, one is never aware of blue

without also being aware of being aware of it; one never apprehends an apprehended object without also apprehending its apprehension. From this we can conclude that the object and its apprehension are the same insofar as they are invariantly concomitant aspects of mind or awareness. There is much controversy about this inference. Buddhist and non-Buddhist opponents considered it to be a fallacious attempt to derive a metaphysical conclusion from phenomenological facts. While it might be true that external objects cannot be apprehended other than in an act of apprehension, it does not follow that they only exist when apprehended or cannot exist unapprehended. The inference was thus often read as a failed attempt to establish T1. A more charitable reading of this inference is as an epistemological argument with metaphysical implications. Dharmakīrti was an epistemological empiricist. Accordingly, for a metaphysical claim to be warranted it must be epistemically established. If there is no legitimate epistemic way to establish the existence of some object, claims of its existence can be denied. Following Dignāga, Dharmakīrti considered there to be only two legitimate epistemic means (pramāṇa) of establishing the truth of a claim: perception (pratyakṣa) and inference (anumāna). For the reasons given above, the mind-independent existence of external objects cannot be established perceptually. However, it also cannot be established inferentially because we would need some perceptual evidence of their mind-independent existence to serve as the supporting reason for the inference. But, again, for the reasons offered above, this perceptual

ee Chatterjee (1987: 45-8), Arnold (2012: 175). Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

evidence cannot be provided. Since there is no other legitimate epistemic way to establish the truth of this claim, it can be denied.37 Some read Dharmakīrti as not, himself, making any metaphysical claims but simply demonstrating that there are no good epistemic reasons to deny T1 in favor of realism.38 This strategy is evident in his response to the problem of other minds.39 To accept the existence of other minds would seem to deny T1. But if it is denied, some account needs to be provided for how it is that we conceive of other minds. If we cannot form such a conception, how can we cultivate compassion for others?40 How can we make sense of the idea of a Buddha teaching his disciples?41 According to Ratnakīrti, other minds cannot be established from an idealist perspective and thus the idea of their existence must be rejected.42 Dharmakīrti argues against this view. In Proof of Other Minds (Santānāntarasiddhi), he contends that the existence of other minds is only ever epistemically established by means of inference. He argues, however, that the evidence for this inference is the same for the realist and idealist alike; namely, appearances of bodily, verbal behavior in conscious awareness.43 The idea that these appearances are caused by (unobserved) mental events of ‘another mind’ is inferred by analogy to the regularity we observe between the mental states and actions we attribute to ‘ourselves’.

Kellner & Taber (2014) give reasons to think that Dharmakīrti may not, himself, have taken this last step as it involves inferring a negative fact (the non-existence of external objects) from unknowability and Dharmakīrti did not consider unknowability to be a legitimate epistemic means of establishing truth or existence - while it might be reason to negate treating something as existent, by introducing doubt, unknowability does not, thereby, establish non-existence. If this is right, then this epistemic argument, at best, undermines the realist claim that external objects do exist mind-independently but it cannot establish that they do not exist mind-independently (and, for that reason, must be merely appearances in consciousness). 38 See Arnold (2008). 39 See Inami (2001). 40 See Ganeri (2012: 202). 41 See Finnigan (2011a,b). 42 See Ratnakīrti’s Santānāntaradūṣaṇa (English translation in Kajiyama (1965)). Inami (2001). 43 See Dharmakīrti SS 5. 2-6: “Observing that the appearance of our bodily and verbal actions is preceded by our minds, we infer the existence of other minds from the appearance of other persons’ bodily and verbal actions.” (citation from Inami 2001: 466). Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

But on what basis do we distinguish those actions which are our own from that of another? According to Dharmakīrti, this distinction is inferred from the relation between verbal, behavioral and bodily appearances. Those verbal and behavioral appearances we take to be our own actions are typically conjoined with those appearances we take to be of our body, whereas those we attribute to others are typically not.44 The evidence for inferring other minds is thus the same for the realist and idealist alike (i.e. objects in conscious awareness) and so the fact that we conceive of other minds need not undermine commitment to T1.45

5. Xuanzang’s LogicalArgument for Idealism

Xuanzang (玄奘, c.a. 602-664) was a Chinese Buddhist monk who famously pilgrimaged to India during the period between Dignāga and Dharmakīrti to study Yogācāra Buddhist philosophy.46 In a debate in Kanyakubja (modern-day Kanauj in Uttar Pradesh, India) Xuanzang delivered his own proof of idealism. This proof was unrefuted in the context of the debate and, although extremely obscure, was assumed to be correct.47 Interest in this proof waned in China but it continued to puzzle Japanese scholars for centuries. Its obscurity has led some recent scholars to suppose, “Xuanzang himself did not understand However, as Inami (2001: 467) points out:

Dharmakīrti notes that such a difference is not fundamental. There can be some exceptions. For example, although the movement of the stone cast by us and the movement of another person shaken by us appear as separated from our own body, they are caused by our own intentions. Therefore, Dharmakīrti concludes that whether the actions are separated from our own body or not is not a crucial factor in inferring other minds.

See Dharmakīrti SS 15.1-17. 45 Ratnakīrti is surely correct, however, to deny the plausibility of this inference from the ultimate perspective, which is held to be empty of all subject-object duality. See Inami (1965:. 472). 46 See Makeham (2014: 8). 47 See Waley (1952: 64-65), Franco (2005: 201, n.34): “The point, however, is not that nobody dared to contradict Xuanzang but that his inference was perfectly correct.” Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

Indian logic”48 and to judge his inference “perilously near to being nonsense.”49 Shinya Moriyama translates it as follows:

Thesis: From the ultimate viewpoint, commonly accepted colors and forms (i.e. visual objects) are not separate from visual consciousness Reason: Because while being included in the first three [of the eighteen elements] that we too accept they are not included in the sense of vision Example: Like the visual consciousness50

Xuanzang’s inference can be read as an attempt to prove T1. At its core is an argument that visual objects are not distinct from (and thus are nothing but) visual consciousness because they are not included in the sense of vision, like visual consciousness. This has the following logical structure:

A is not distinct from B Because it is not C Like B

It is not difficult to see why some scholars judged this inference to be nonsense. On a simplified reading, it is equivalent to inferring that a philosopher is a banana because she is not a pyjama, like a banana. Why should we accept the reason for the thesis? And how

48 See Nakamura (1958) as translated by Shinya Moriyama in Franco (2005: 202) 49 See Waley (1952: 64) 50 大師立唯識比量云：眞故極成色不離於眼識宗,自許初三攝 眼所不攝故因, 猶如眼識喩 (YRZLS T44:1840.115b25-26). Translation cited in Franco (2005: 205) with modifications from Moriyama (2014: 143). The original Chinese is documented in Kuji’s Yinming ru zhengli lun shu and cited in Moriyama (2014). Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

do the qualifications and example contribute to the argument (what do they mean and how do they support the thesis)? Eli Franco (2005) offers an ingenious explanation of this argument, which draws on, and attributes to Xuanzang, a highly sophisticated understanding of Dignāga’s conception of logic. While Franco concludes by demonstrating that Xuanzang’s inference was in fact refuted by the Korean Patriarch Wōnhyo (618-686), his reconstruction helpfully draws out central features of Buddhist logic and its relationship to idealism. To understand it, we need some grasp of Dignāga’s triple-condition (trairūpya) theory of inferential reasoning and the Buddha’s 18-element compositional analysis of persons.

As mentioned earlier, Dignāga and Dharmakīrti maintained that there were only two legitimate epistemic means of establishing the truth of a claim; perception and inference. For Dignāga, an inference is a cognitive process whereby a new cognition arises from an initial cognition taken as a reason.51 A classic example is that of inferring the presence of fire on a mountain from the observation of smoke. The presence of smoke is here taken as a reason (hetu) to suppose that the subject (pakṣa) ‘mountain’ has the property to be proved (sādhya): ‘fire’. For the presence of smoke to count as a good reason for this thesis it must satisfy three conditions.52 First, it must be ascertained as present with or possessed by the subject of the inference. In this case, smoke must be observed on the mountain. Second, the reason must be ascertained as present with or possessed by those things (other than the subject) that possess the property to be proved.

According to Katsura (1983: 540), Dignāga’s logic stands precisely at a point in time between thinking of inference as an inductive process, aimed at discovering the relationship between a cause and its result, and Dharmakīrti’s consideration of inference as a logical analysis of the result of such a process. It should be thus understood as a form of abduction rather than deduction. For a discussion of Dharmakīrti’s views on the trairūpya, see Tillemans (1999, 2005, 2011), Tanaka (2013) 52 My characterisation of Dignāga’s trairūpya follows Katsura (2005, 1983) and Hayes (1988: 146-9) Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

That is, if we group together all of the things that we know have or involve fire, at least one of these things must have or involve smoke. A classic example from this group is of a (wood-burning) kitchen. Third, the reason must not be ascertained as present with or possessed by those things that lack the property to be proved. If we group together all of the things that we know do not have or involve fire, none of those things have or involve smoke. A classic example is of a misty lake. For Dignāga, an inference can thus be understood to exhaustively divide the universe into three parts: a group of things known to possess the property to be proved, a group of things known not to possess this property, and the subject of the inference, “about which there is some doubt because we do not yet know whether it possesses this property or not.” (Franco 2005: 207). An inference is ‘valid’ if the reason offered for the thesis satisfies the above three conditions. Xuanzang’s thesis is that visual objects are not distinct from visual consciousness. The subject of the thesis is ‘visual objects’ and the property to be proved is ‘being not distinct from visual consciousness’. The reason Xuanzang offers, however, is obscure; i.e., ‘not being (included in) the sense of vision’. To evaluate whether this is a good reason, we must divide the universe into those elements known to possess the property to be proved and ascertain whether at least one of them also possesses the reason. Moreover, of those things that lack the property to be proved, we must also determine that they lack the reason.

What are the relevant elements for logical division in this case? They are the 18 elements (dhātus) which, according to the Buddha, exhaustively compose a person. These elements are organised in terms of the 5 sense faculties (plus a mental faculty), with their corresponding objects and forms of consciousness. Following Franco, it is helpful to numerically list them as follows: Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

[dhātu-D1] sense of vision, [D2] visible objects, [D3] visual consciousness [D4] sense of hearing [D5] sounds [D6] auditory consciousness [D7] sense of smell [D8] odours [D9] olfactory consciousness [D10] sense of taste [D11] flavours [D12] gustatory consciousness [D13] sense of touch [D14] tangibles [D15] tactile consciousness [D16] sense of mind/thought [D17] mind/thought objects [D18] mind-consciousness

The subject of Xuanzang’s thesis (‘visual objects’) is element D2 in this analysis. The property to be proved (‘being not distinct from visual consciousness’) is equivalent to ‘being not distinct from D3’. And the reason (‘not being included in the sense of vision’) can be understood as ‘not being D1’ (which is equivalent to ‘being D2-18’). Understood in this way, Franco reconstructs Xuanzang’s argument as:

D2 is not distinct from D3 because it is not D1.53

As Franco points out, however, this inference is invalid because, without some qualification, the reason fails to satisfy Dignāga’s third condition. It satisfies the first because D2 has the property of ‘not being D1’. It satisfies the second because of all the things known to be ‘not distinct from D3’, the only member of this set aside from the subject, (i.e., D3), has the property of ‘not being D1’.54 The reason fails the third condition, however. It is not the case that, of all those things known not to be ‘not distinct

Or, more precisely, D2 is not distinct from D3 because it is not included in D1. 54 That Franco’s analysis is correct is implied by the fact that D3 (‘visual consciousness’) is the example that Xuanzang himself offers to support his inference. Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

from D3’ (i.e., D1, 4-18) they all lack the property of ‘not being D1’. This property is possessed by D4-18.

Xuanzang’s reason does come with a qualification, however; namely, “while being included in the first three [of the eighteen elements] that we too accept…” In a sense, Xuanzang is simply making the point that, like his opponent, he accepts that visible objects and visual consciousness are ‘visual’ (the sphere composed by the first three elements in the Buddha’s compositional analysis). However, as Franco points out, this qualification makes a logical difference.55 It introduces a second property into the reason that qualifies the first; namely, that of ‘not being D1 but included in D1-3’. The inference can thus be reformulated as:

D2 is not distinct from D3 because it is not D1 but is included in D1-356

As Franco demonstrates, this reformulated inference is valid because the qualified reason satisfies Dignāga’s triple condition. The first is satisfied because D2 has the property of ‘not being D1 but included in D1-3’. The second is satisfied because, of all the things known to be ‘not distinct from D3’, the only member of this set aside from the subject, (i.e., D3), has the property of ‘not being D1 but included in D1-3’. And the third is satisfied because, of all those things known not to be ‘not distinct from D3’ (i.e., D1, 4 18) they all lack the property of ‘not being D1 but included in D1-3’. Xuanzang’s inference for idealism is thus valid by the lights of Dignāgean logic. Should we accept it and thus agree that visual objects are not separate from visual consciousness (in the strong sense that implies T1)? There are reasons to be suspicious.

See Franco (2005: 208): “It is a conspicuous feature of the Indian syllogism that the reason is usually expressed in the form of a single property. If two properties are needed, the second property is expressed in the form of a qualification of the first property.” 56 Or, more precisely, D2 is not distinct from D3 because it is not included in D1 but is included in D1-3. Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

As indicated earlier, the logical structure of Xuanzang’s inference allows for absurd conclusions. Moreover, its validity seems to turn entirely on exploiting the tripartite structure presupposed by Dignāgean logic. To see this, consider a world with only three classes of things: philosophers, bananas and pyjamas. In such a world, the first of Dignāga’s conditions is satisfied because the subject (i.e., a philosopher) has the property of ‘not being a pyjama’. The second is satisfied because, of all the things known to be ‘not distinct from a banana’, the only member of this set other than the subject, (i.e., a banana), has the property of ‘not being a pyjama’. And the third is satisfied because the things known not to be ‘not distinct from a banana’ other than the subject, (i.e., pyjamas), also lack the property of ‘not being a pyjama’. A philosopher is thus not distinct from a banana. However, by this very same reasoning I can also prove that a banana is not distinct from a pyjama because it is not a philosopher and, from these two proofs, I can derive the claim that a philosopher both is and is not distinct from a banana.57 As Franco interestingly points out, Xuanzang’s Buddhist realist opponent cannot appeal to the fact of such transposition as a reason to refute Xuanzang. This is because any rearrangement of the three elements contained in his inference only serves to reinforce his idealism. It tautologously follows from the fact that visual objects are not distinct from visual consciousness that visual consciousness is not distinct from visual objects. And to say that the visual sense is not distinct from visual consciousness (or visual objects) because it is not a visual object (or visual consciousness) would just show that they are nothing but the (karmic) seeds of (the visual objects in) visual

57 Of course, we do not live in such a world and so, in actuality, the final criterion of these inferences fails (it is not the case that everything that is not a banana is not not a pyjama) The absurd conclusion thus does not necessarily follow from Dignāgean logic but rather from enforcing a restriction in the classification of existent things in a way that exploits its tripartite structure. Later Buddhist thinkers (Dharmakīrti most famously) revised this structure, insisting instead on a bipartite universe in which the subject of the inference is included in the class of things that possess the property to be proved. Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

consciousness, which implies T2. The Buddhist opponent also cannot deny the restriction that allows for this transposition (and thus for the validity of the inference) because they also accept that visual objects and visual consciousness are within the visual sphere as constituted by the first three elements in the Buddha’s compositional analysis. Franco closes his discussion, however, by reconstructing a challenge from Wōnhyo which uses the same logic as Xuanzang but transposes the relevant elements to prove the opposite thesis. Rather than directly refuting Xuanzang’s inference (by proving its invalidity), Franco takes Wōnhyo to annul it by establishing the converse and thereby generating a contradiction.58 Moriyama translates Wōnhyo’s counterargument as follows:

From the ultimate view, colour and form (i.e. visual objects), which are well known among the people, are separate from visual consciousness. Because while being included in the first three [[[dhātus]]] that [we] accept, they are not included in visual consciousness, like the sense of vision.59

At its core is an argument that visual objects are distinct from visual consciousness because they are not visual consciousness, like the sense of vision. This argument can be read tautologously (A is not B because it is not B, like C) or more substantively (A is separate from B because it is not included ‘by the Buddha in his enumeration of the 18fold elements’60 in B, like C). Either way, the reasons respectively satisfy Dignāga’s triple-condition for inferential validity and thus Wōnhyo’s counterargument against

According to Franco, Wōnhyo’s argument strategy is that of viruddhāvyabhicārin (antinomial reason), “i.e. construing a contradictory inference whose reason also fulfils the three conditions.” (2005: 212). Xuanzang’s student, Kuji, sought to deny that Wōnhyo successfully provided a viruddhāvyabhicārin refutation. His reasons, however, are not particularly convincing. See Moriyama (2014). 59 Translation cited in Franco (2005: 212)

60 See Franco (2005: 212) Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

Xuanzang’s inference for idealism is valid.61

6. Conclusion

This chapter has intellectually surveyed some of the most influential Buddhist arguments in defense of idealism. It has refrained from speculating on comparative similarities with Western philosophical views or from offering critical commentary. These are important and substantive tasks, but ones which presuppose an initial introduction to the relevant ideas and arguments. The aim of this chapter has been to provide a broad overview which can enable further philosophical inquiry. It identified a range of theses as potential candidates for the title ‘idealism’ and a range of argument strategies and methodologies for establishing these theses. These strategies included:

- arguments from simplicity (against the need to posit external objects) - inferences to the best explanation (having eliminated alternative explanations of sensory experience which appeal to external objects) - epistemological arguments (either inferring the falsity of the realist conception of external objects from unknowability or demonstrating isomorphism in realist and idealist evidential assumptions) - logical arguments (demonstrating that an idealist conclusion can be validly inferred within a particular logical system).

If Franco is right, then my derivation that a philosopher both is and is not a banana can be similarly understood as a viruddhāvyabhicārin (antinomial reason) for rejecting the analogous proofs. Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

Whether or not one finds these arguments convincing will crucially depend on what one thinks of their presuppositions about the nature of mind, metaphysics, epistemology and logic. While these ideas are assumed by the arguments presented in this chapter, they are rigorously defended elsewhere. It must also not be forgotten that these ideas and their arguments were also vigorously opposed by both Buddhist and non-Buddhist philosophers. Idealism is but one philosophical trend in the Buddhist intellectual tradition. This chapter has provided an overview of some of the central Buddhist arguments in defence of idealism, as traditionally understood. While it is beyond the scope of this chapter to survey the lively philosophical dialectic that emerged in response to these arguments, it is hoped that this chapter may stimulate sufficient interest to inspire further inquiry. 62

Abbreviations

Ālam Ālambanaparīkṣā of Dignāga AKBh Abhidharmakośabhāṣya of Vasubandhu PS Pramāṇasamuccaya of Dignāga PV Pramāṇavārttika of Dharmakīrti PVin Pramāṇaviniścaya of Dharmakīrti Triṃś Triṃśikā of Vasubandhu SS Santānāntarasiddhi of Dharmakīrti Viṃś Viṃśikā of Vasubandhu Viṃś vṛ Viṃśikā vṛtti of Vasubandhu YRZLS Yinming ru zhengli lun shu (因明入正理論疏) of Kuiji (窺基)

I would like to thank Dan Arnold, Jay Garfield, Jonathan Gold, Shinya Moriyama, Tom Tillemans and the editors of this book for helpful comments on an earlier draft of this chapter. Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

References

Anaker, Stephen. 1998. Seven Works of Vasubandhu. (Delhi: Motilal Banarsidass). Arnold. Dan. 2005. ‘Is Svasaṃvitti Transcendental?’ Asian Philosophy, 15, 1: 77-111. ______. (2008). ‘Buddhist Idealism, Epistemic and Otherwise: Thoughts on the Alternating Perspectives of Dharmakīrti’ Sophia, 47: 3–28. _______. (2012) Brains, Buddhas, and Believing. (New York: Columbia University Press). Chatterjee, Ashok Kumar (1987) The Yogācāra Idealism. (Delhi: Motilal Banarsidass) Cowherds, The. 2014 Moonshadows. (Oxford: Oxford University Press). Coseru, Christian. 2012. Perceiving Reality: Consciousness, Intentionality and Cognition in Buddhist Philosophy. (Oxford: Oxford University Press) Dreyfus, Georges. 1996. ‘Can the Fool Lead the Blind? Perception and the Given in Dharmakīrti’s Thought’, Journal of Indian Philosophy, 24, 3: 209–229. Duckworth, Douglas., Eckel, M.D., Garfield, J., Powers, J., Thabkas, Y., Thackchoe, S., (forthcoming) Dignāga’s Investigation of the Percept: A Philosophical Legacy in India and Tibet (Oxford: Oxford University Press) Dunne, John. 2004. Foundations of Dharmakīrti’s philosophy. (Boston: Wisdom Publications).

Finnigan, Bronwyn. 2011a. ‘How Can a Buddha Come to Act? The Possibility of a Buddhist Account of Ethical Agency’, Philosophy East and West, 61, 1: 134-60. _______. 2011b. ‘A Buddhist Account of Ethical Agency Revisited: Reply to Garfield and Hansen’, Philosophy East and West, 61, 1: 183-94. Franco, Eli. 2005. ‘Xuanzang’s Proof of Idealism (Vijñanaptimātratā)’ in Aoyama, Takao, Ed. Horin: Vergleichende Studien zur japanischen Kultur, 11:199-212 Frauwallner, Erich. 1951. On the Date of the Buddhist Master of the Law Vasubandhu. Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

(Rome: Istituto Italiano per il Medio ed Estremo Oriente). Ganeri, Jonardon. 1999. ‘Self-intimation, Memory, and Personal Identity’, Journal of Indian Philosophy, 27: 469-83. _______. 2012. The Self. (Oxford: Oxford University Press). Garfield, Jay. L. (2002) Empty Words: Buddhist Philosophy and Cross-Cultural Interpretations, (Oxford: Oxford University Press) _______. 2006. ‘The Conventional Status of Reflexive Awareness: What’s at Stake in the Tibetan Debate’, Philosophy East and West, 56: 201-28. _______. 2015a. ‘I am a Brain in a Vat (Or Perhaps a Pile of Sticks By the Side of the Road)’ in Garfield & Westerhoff, Eds. Madhyamaka and Yogācāra: Allies or Rivals? (Oxford: Oxford University Press), 252-275. _______. 2015b. Engaging Buddhism. (Oxford: Oxford University Press). Garfield, Jay.L. & Westerhoff, Jan. (Eds.) 2015. Madhyamaka and Yogācāra: Allies or Rivals? (Oxford: Oxford University Press). Gethin, Rupert. 1988. The Foundations of Buddhism. (Oxford: Oxford University Press). Gold, Jonathon. 2006. ‘No Outside, No Inside: Duality, Reality and Vasubandhu's Illusory Elephant’, Asian Philosophy, 16: 1–38. _______. 2015a. ‘Vasubandhu’ in Stanford Encyclopedia of Philosophy, Edward N. Zalta, Ed. URL = <http://plato.stanford.edu/archives/sum2015/entries/vasubandhu/>. _______. 2015b. Paving the Great Way: Vasubandhu's Unifying Buddhist Philosophy, (New York: Columbia University Press) Hattori, Masaaki. 1968. Dignaga, On Perception. (Cambridge, Mass: Harvard University Press). Hayes, Richard. 1988. Dignāga on the Interpretation of Signs. (PLACE: Kluwer Academic Publishers). Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

Inami, Masahiro. 2001. ‘The Problem of Other Minds in the Buddhist Epistemological Tradition’, Journal of Indian Philosophy, 29: 465-83. Jaini, PS. 1958. ‘On the Theory of the Two Vasubandhus’, Bulletin of the School of Oriental and African Studies, 21: 48–53. Kajiyama, Yuichi. 1965. ‘Buddhist Solipsism – A Free Translation of Ratnakīrti’s Santānāntaradūṣaṇa’, Journal of Indian and Buddhist Studies, XIII–1: 9–24. Kanō Kazuo. 2008. ‘Two Short Glosses on Yogācāra Texts by Vairocanarakṣita: Viṃśikāṭīkāvivṛti and Dharmadharmatāvibhāgavivṛti” in Sferra, Francesco., Ed. Manuscripta Buddhica, Vol. I: Sanskrit Texts from Giuseppe Tucci’s Collection, Part I. (Roma: Istituto Italiano per l’Africa e l’Oriente), 343–80. Kapstein, Matthew. 2001.. ‘Mereological Considerations in Vasubandhu's "Proof of Idealism”’, in Reason's Traces. (Boston: Wisdom Publications), 181–204. Katsura, Shoryu. 1983. ‘Dignāga on Trairūpya’, Journal of Indian and Buddhist Studies 32, 1: 15-21.

_______. 2005. ‘Pakṣa, Sapakṣa and Asapakṣa in Dignāga’s Logic’ in Aoyama, Takao, Ed. Horin: Vergleichende Studien zur japanischen Kultur (11), 119-128. Kellner, Birgit. 2010. ‘Self-Awareness (svasaṃvedana) in Dignāga’s Pramāṇasamuccaya and –vṛtti: A Close Reading’, Journal of Indian Philosophy, 38: 203-31 _______.2011. ‘Self-awareness (svasaṃvedana) and Infinite Regresses: A Comparison of Arguments by Dignāga and Dharmakīrti’ Journal of Indian Philosophy, 39: 411-26. _______. 2011. ‘Dharmakīrti’s Criticism of External Realism and the Sliding Scale of Analysis’ in Krasser, Helmet, et al., Eds. Religion and Logic in Buddhist Philosophical Analysis: Proceedings of the Fourth International Dharmakīrti Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

Conference. (Wien: Verlag der O¨ sterreichischen Akademie der Wissenschaften), 291-298. Kellner, Birgit. & Taber, John. 2014. ‘Studies in Yogācāra-Vijñānavāda idealism I: The interpretation of Vasubandhu’s Viṃśikā’, ASIA, 68, 3: 709-56. Kochumuttom, Thomas.A. 1982. A Buddhist Doctrine of Experience: A New Translation and Interpretation of the Works of Vasubandhu the Yogācārin. (Delhi: Motilal Banarsidas Publishers Private Limited). Kritzer, Robert. 2003. ‘Sautrāntika in the Abhidharmakośabhāṣya’, Journal of the International Association of Buddhist Studies, 26: 331–84. _______. 2005. Vasubandhu and the Yogācārabhūmi: Yogācāra Elements in the Abhidharmakośabhāṣya. Studia Philological Buddhiaca XVIII. (Tokyo: International Institute for Buddhist Studies). Lusthaus, Dan. 2002). Buddhist Phenomenology: A Philosophical Investigation of Yogācāra Philosophy and the Ch’eng Wei-shih lun. (New York: Routledge Curzon) MacKenzie, Matthew. 2007. ‘The Illumination of Consciousness: Approaches to Self- Awareness in the Indian and Western Traditions’, Philosophy East and West 57, 1: 40-62. _______. 2008. ‘Self-Awareness Without a Self: Buddhism and the Reflexivity of Awareness’, Asian Philosophy, 18, 3: 245-66. Makeham. John. 2014. ‘Introduction’, in Makeham, John, Ed. Transforming Consciousness: Yogācāra Thought in Modern China (Oxford: Oxford University Press) Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

Matilal, Bimal. K 1974. ‘A Critique of Buddhist Idealism’ in Cousins, Kunst & Norman, Eds. Buddhist Studies in Honour of I B. Horner. (Dordrecht, Holland: D. Reidel Publishing Company), 139-169. McClintock, Sarah. 2003. ‘The Role of the “Given” in the classification of Śāntarakṣita and Kamalaśīla as Svātantrika-Mādhyamikas’ in Dreyfus & McClintock, Eds. The Svātantrika-Prāsaṇgika distinction: What difference does a difference make? (Boston: Wisdom Publications), 125–71). Moriyama, Shinzo. 2014. ‘A Comparison between the Indian and Chinese Interpretations of the Antinomic Reason (Viruddhāvyabhicārin)’ in Lin, C.K & Radich, M. (eds.) A Distant Mirror: Articulating Indic Ideas in Sixth and Seventh Century Chinese Buddhism, (Hamburg: Hamburg University Press), 121-150. Nakamura, Hajima. 1958. Introduction to Nyāyapraveśa, in Kokuyaku issaikyō, (Tokyo: Iwanami shinsho) Oetke, Claus. 1992. ‘Doctrine and Argument in Vijñānavāda-Buddhism’, Wiener Zeitschrift für die Kunde Südasiens, 36: 217–25. Ronkin, Noa, 2005. Early Buddhist Metaphysics: The Making of a Philosophical Tradition. (New York: Routledge-Curzon). _______. 2014. ‘Abhidharma’ in Stanford Encyclopedia of Philosophy, Edward N. Zalta, Ed. URL = <http://plato.stanford.edu/archives/fall2014/entries/abhidharma/>. Schterbatsky, Th. 2003. Central Conception of Buddhism and the Meaning of the Word

Dharma. (New Delhi: Asian Educational Services). Siderits, Mark. 2007. Buddhism as Philosophy, (Aldershot: Ashgate Publishers) Taber, John. 2010. ‘Kumarīla’s Buddhist’, Journal of Indian Philosophy, 38: 279-96. Tanaka, Koji. 2013. ‘Buddhist Philosophy of Logic’ in Emmanuel, Steven M., Ed. A Companion to Buddhist Philosophy (Oxford: Wiley-Blackwell), 320-30. Final draft of article in Goldschmidt & Pearce (eds.) Idealism: New Essays in Metaphysics, Oxford (2017)

Tillemans, Tom. J. F. 1999. Scripture, Logic, Language. (Boston: Wisdom Publications) _______. 2005 ‘The Slow Death of the trairūpya in Buddhist Logic: A Propos of Sa Skya Paṇḍita’ in Aoyama, Takao, Ed. Horin: Vergleichende Studien zur japanischen Kultur (11), 83-94. _______. 2007. Materials for the Study of Āryadeva, Dharmapāla and Candrakīrti. (Delhi: Motilal Banarsidass Publishers Private Limited). _______. 2011. ‘Dharmakīrti’ in Stanford Encyclopedia of Philosophy, Edward N. Zalta, Ed. URL = <http://plato.stanford.edu/archives/spr2014/entries/dharmakiirti/>. Thompson, Evan. (2011) ‘Self, No Self? Memory and Reflexive Awareness’ in Siderits, Thompson, Zahavi (eds.) Self, No Self? Oxford University Press Tola, Fernando. & Dragonetti, Carmen. 2004. Being as Consciousness: Yogācāra Philosophy of Buddhism. (Delhi: Motilal Banarsidass). Yao, Zhihua. 2005. The Buddhist Theory of Self-Cognition. (New York: Routledge). Vasubandhu 1990. Abhidharmakośabāṣya, Leo M. Pruden, Trans.( Berkeley, Calif.: Asian Humanities Press). Waley, Alex. 1952. The Real Tripitaka and Other Pieces (New York: Routledge) Williams, Paul. 1981. ‘On the Abhidharma Ontology’, Journal of Indian Philosophy, 9: 227–57. _______. 2000. The Reflexive Nature of Awareness: A Tibetan Madhyamaka Defense, (Delhi: Motilal Banarsidass) Wood, Thomas. E. 1991. Mind Only: A Philosophical and Doctrinal Analysis of Vijñānavāda, (Honolulu: University of Hawaii Press).