Light as an Analogy for Cognition in Buddhist Idealism (vijñānavāda)

by Alex Watson

Introduction

The phrase ‘Buddhist Idealism’ in the title of the paper refers to the vijñānavāda of the Buddhist Epistemological School, specifically as it is expounded in Jayanta’s (850–910) Nyāyamañjarī. Jayanta, drawing directly or indirectly on Dignāga, Dharmakīrti and Dharmottara, gives five arguments1 for vijñānavāda (understood as the view that everything in the universe is of the nature of cognition) before arguing against them in the subsequent section of the text.2 The point of this paper is to expound as clearly as possible one of these arguments (section 1) and to examine the role played in it by the analogy of light (section 2). At the end (section 3), the suitability of the light analogy is evaluated.

The ‘light analogy’ is a shorthand expression for two different ways in which cognition is compared to light. On the one hand cognition is said to be like light, on the other hand it is said to be (a kind of) light.3 We see the first in such cases as when lights or flames are given as the corroborating example in a formal inference involving cognition, or, less formally, when something is said to be true of cognition, and then this is made more plausible by adding that the same is true of light(s). We see the second when cognition is referred to with the word prakāśa (‘light’ or illumination’) or when it is said to shine forth (prakāśate) or illuminate (prakāśayati).4

1 Apart from the argument examined in this article, whose history is commented on below, two of the arguments go back at least to Kumārila (that based on the economy of postulation, kalpanālāghava, and that based on the impossibility of contradictory properties belonging to the same object, viruddhadharmasamāveśa), and the other two originate with Dharmakīrti: the argument from the necessary co-perception of object and cognition

(sahopalambhaniyama) and the argument from the object-specificity of cognition (pratikarmavyavasthā). 2 This portion of the Nyāyamañjarī (Vol. 2, pp. 487,10–504,15), covering both the exposition of the five arguments and their refutation, was re-edited in Kataoka (2003), translated for the first time into a non-Indian language (Japanese) in Kataoka (2006), and translated for the first time into a European language (English) in Watson and Kataoka (2010).

The last of these publications also includes analysis in an introduction, two sets of notes, and a running exegetical commentary. I would like to express sincere gratitude to Kei Kataoka; were it not for having worked with him on that, I doubt I would have been able to begin the present article. 3 For another recent study of cognition as ‘luminosity’, see Ram-Prasad (2007: 51– 99). 4 The following two examples from the Nyāyamañjarī combine both kinds of analogy: na prakāśāntarāpekṣaṇam, svata eva dīpavat prakāśasvabhāvatvāt, ‘[[[cognition]]] does not de

1.1 The Vijñānavādin’s Argument in Jayantabhaṭṭa's Nyāyamañjarī

The difference between Vijñānavāda idealism and Nyāya/Mīmāṃsā realism concerns the nature of the objects of our experience. For the Vijñānavādin they are cognition appearing in a certain form; for the realist they are insentient (jaḍa), external objects. The first response of the realist in the Nyāyamañjarī is that the Vijñānavādin view can be refuted by appeal to direct perception alone:5 if what we experience in direct perception were merely cognition, we

would experience always the same thing, something that is always of the same nature, but in fact what we perceive changes constantly, as one object of perception succeeds the previous. The Vijñānavādin gets around this problem by asserting that cognition is not a formless blank or monotone, but something that has form (ākāra).6 Thus both sides agree about what it feels like to experience the world around us: we see a certain form followed by another, and so on – this in itself weighs neither in favour of one side nor the other. What they disagree about is whether the forms we perceive belong to cognition or to external objects.

One avenue that is not open to the Naiyāyika/Mīmāṃsaka is to argue that we experience both an external object and the cognition that reveals that object.

If these two things were given to us separately in experience, then the realist would win. There would be no scope for any debate; we would be experiencing an object that is separate from cognition, the existence of which is precisely what the Vijñānavādin denies. But the realist agrees with the Vijñānavādin that we only perceive one form, for example blue (not two forms, both the blue and a separate form of its cognition). There is a question mark hanging over this regarding whether it belongs to cognition or to an object.

pend on another illumination, because like a lamp its nature is to shine forth by itself’ (§ 3.2.2 in Kataoka 2003) and prakāśatvāj jñānasya pradīpavat pūrvaṃ grahaṇam, ‘because cognition is (a kind of) light, like a lamp, it must be grasped first’ (§ 4.4.1 in Kataoka 2003). For examples of the first kind in Dharmakīrti, see Pramāṇavārttika 3:329, 3:482ab; for examples of the second kind see Pramāṇavārttika 3:327, 3:446, 3:477, 3:478, 3:480, 3:481, and

Pramāṇaviniścaya 1:38. 5 § 2.1 in Kataoka (2003). This strategy is also found in the Nirālambanavāda chapter of Kumārila’s Ślokavārttika, before he adopts a different strategy in the Śūnyavāda chapter; see Taber (2010: 279–282). 6 Each time the word ‘form’ is used from now on, it is an English rendering of ākāra.

The Vijñānavādin position is that it belongs to cognition; the Naiyāyika/Mīmāṃsaka position is that it belongs to an (external) object.

In the argument that we will look at, the Vijñānavādin exploits this admission on the part of the realist that we perceive only one form, not two. He argues that we perceive cognition. If that is the case, and it is combined with the assumption that we perceive only one thing, not two, then it follows that this one thing must be cognition, not an external object. The argument can be divided into four main stages.

1. Cognition must be grasped first (= before the object is grasped).

2. Since cognition is grasped, it must have form.

3. Thus, since we are both agreed that only one form is grasped, that one form must belong to cognition and cannot belong to an external object.

4. Thus there is no justification for positing the existence of an external object.

Once it has been established in 1 and 2 that cognition must be grasped first and that it has form, it then becomes untenable to suppose that after grasping cognition with form, we grasp its object with form, for we do not experience two forms. It thus becomes not merely redundant to suppose experience of an external object; it becomes incoherent.

The argument moves from svasaṃvedana in 1 (the view that cognition grasps itself), to sākāravāda in 2 and 3 (the view that cognition has form), to vijñānavāda in 4 (the view that no objects outside of cognition exist). To get from sākāravāda to vijñānavāda requires a refutation of the Sautrāntika inference of the existence of external objects, which is carried out by the Vijñānavādin speaker in this same portion of the text.7 Stages 2, 3, and 4 consist of no more than what has already been said about them, but stage 1 is argued for at length. We will now investigate its supporting arguments:8

1 Cognition must be grasped first

1.1 Because it is (a kind of) light (prakāśa). Hence it would not be able to illuminate its object unless it was grasped first, because lights such as lamps are only able to illuminate objects if they are themselves grasped. 1.2 Because it must be grasped at the very moment that it arises (which is before it reaches the object). 1.2.1 Because no obstruction can come between it and itself, and because it does not depend on another illumination/cognition. 1.2.2 If it were not grasped at the moment it arose, it could never be grasped. 1.2.2.1 [Naiyāyika’s objection:] It could be grasped by a subsequent cognition. 1.2.2.2 [Vijñānavādin’s response:] That would result in an infinite regress. 1.3 Because of reflection (pratyavamarśa) on an object as cognized.9 The claim that cognition must be perceived before the object is an unwanted consequence (prasaṅga); the Vijñānavādin’s final position (siddhānta) is of course that cognition and object are perceived simultaneously, for they are (two aspects of) the same thing. But he reasons that if there were an external object, separate from cognition, it could only be perceived after cognition has been perceived.

1.2 Historical Precursors of the argument

Jayanta’s immediate source for the argument is Kumārila. And the relevant section in the Ślokavārttika (śūnyavāda 21–34) is in turn an expansion of the following argument in the Śābarabhāṣya: ‘We do not perceive different forms, one of the object and one of cognition. And for us cognition is perceived. Therefore there is no supposed [[[entity]]] of the nature of an object which is separate from cognition.’10 We have shown11 that 1.3 falls in a line of development stretching back, from Jayanta’s own argument, to the argument given in Ślokavārttika, śūnyavāda 28–29, and further to Dignāga’s famous argument from memory for cognition cognizing

7 § 3.4 in Kataoka (2003). 8 For the full text and translation of what is just given in outline in the following, see Kataoka (2003, §§ 3.2.1–3.2.3) and Watson and Kataoka (2010: 304–312); and for more elaboration see the annotation and commentary to the translation.

9 The fact that we can say ‘this object was cognized by me’ shows that we must have perceived the cognition itself earlier. 10 arthajñānayor ākārabhedaṃ nopalabhāmahe. pratyakṣā ca no buddhiḥ. atas tadbhinnam artharūpaṃ nāma na kiṃcid asti (Śābarabhāṣya 28,14–16). 11 Watson and Kataoka (2010: 308–310).

itself.12 At each step the argument is slightly modified, partly in order to make clear what was not so in the version given by the previous philosopher. So in 1.3 we have a development of Dignāga’s argument from memory; and in 1.2 we have the infinite regress argument that features in Dignāga as a subsidiary part of his argument from memory.13 As in both Kumārila and Dharmakīrti, so in Jayanta, the argument from memory has been separated off from the infinite regress and presented as a self-standing argument. Jayanta concludes 1.2 by citing Dharmakīrti’s famous line: ‘The seeing of an object cannot be established for someone who does not perceive the cognition [of that object]’.14 In short, Jayanta’s argument draws on Kumārila and Dharmakīrti; Kumārila’s argument draws on the Śābarabhāṣya and Dignāga.

2.1 The role played by light in the argument

The light analogy serves to render plausible the claim that cognition is perceived. How does it do that? Light illuminates not only its objects but also itself; so if cognition resembles light, it too should illuminate not only its objects but also itself. That much is well understood, but what I have not seen yet in secondary literature is an attempt to separate out more specific and distinguishable aspects of the light analogy. Here I will separate out two and later I will add two more.

(1) In order to see objects illuminated by a light, you have to see the light itself.

(2) A light requires no illumination other than itself. In order to see an object a lamp is required; but in order to see that lamp, no further lamp is required.

It is the second aspect that is relevant in 1.2.1.15 A lot depends on the first aspect: it is the sole force behind the argument in 1.1,16 and it is what drives the infinite regress in 1.2.2.2.17

2.2 Infinite Regress The structure of the Vijñānavādin’s infinite regress argument as presented by Jayanta is as follows.

12 Pramāṇasamuccaya 1:11cd and vṛtti thereon. 13 On which, see Ganeri (1999), Kellner (2010: 213-215) and (2011). 14 Pramāṇaviniścaya 1:54cd: apratyakṣopalambhasya nārthadṛṣṭih prasiddhyati. 15 See na prakāśāntarāpekṣaṇam: ‘[it] has no dependence on another illumination’. 16 See agṛhītasya dīpādeḥ prakāśasya prakāśakatvādarśanāt: ‘because we do not find that lights such as lamps are able to illuminate if they are not grasped’ and (from §

4.2.3 of Kataoka 2003, where the argument of this section is restated): na cāgṛhītaḥ prakāśaḥ prakāśyaṃ prakāśayati: ‘and a light that is not grasped does not illuminate an object of illumination’. 17 prakāśāgrahaṇe tatprakāśyaparicchedāyogāt: ‘because as long as the illumination remains ungrasped, it is impossible to discern things illuminated by that’.

For cognition to perceive its object, it must itself be perceived. • If it were perceived by a subsequent cognition there would be an infinite regress. • Therefore it must be perceived by itself at the very moment that it arises. To that extent the argument is the same as Dharmakīrti’s.18 But they differ on the matter of what justifies the first point. For Dharmakīrti it is an argument – centering on the claim that unless a cognition is perceived, it cannot condition conceptual determination, language or physical behaviour – that according to Birgit Kellner's recent analysis takes for granted what it needs to prove (Kellner 2011: 419–422). For Jayanta it is the second aspect of the light analogy: in order for an illuminated object to be perceived, its illuminator must be perceived.

2.3 Can the light analogy achieve what it is meant to achieve?

Much hangs, then, on this aspect of the light example. Is it able to carry the weight? Is it able to provide what is lacking in Dharmakīrti’s argument according to Birgit Kellner, namely a valid justification of the claim that cognition must be perceived in order to perceive its object? According to John Taber, no. Commenting on the claim (made by Kumārila’s Vijñānavādin in one of the verses19 that is the source of the argument in Jayanta) that we have to see a light in order to see objects illuminated by it, he writes:

‘[I]t just doesn’t seem true that one has to see the source of illumination in order to see objects illumined by it. All the time we look at objects illumined by the sun and other luminous bodies without also apprehending those sources of illumination. In the case of the moon, most of the time it isn’t even possible to apprehend the sun which is illumining it, for it is blocked by the earth.’ (Taber 2010: 284) Even if an obstruction intervenes between an observer and a source of light, the observer is not prevented from seeing all of the things illuminated by that light:

18 Pramāṇaviniścaya ad 1.54cd, pp. 40,11–41,13. For a detailed study of this argument of Dharmakīrti and an analysis of how it differs from Dignāga’s infinite regress argument, see Kellner (2011). For a different account of how the two infinite regress arguments differ, see Siderits (forthcoming: §5, especially note 4). 19 Ślokavārttika, śūnyavāda 22.

What are we to make of the Buddhist argument – that cognition must be perceived in order for its object to be perceived, because light must be perceived in order for an object illuminated by it to be perceived – in the light of Taber’s objection? I give three ways in which we can respond. The first takes Taber’s objection to undermine the argument. The second and third are defenses of the argument against Taber’s objection.

(1) The Buddhist is using an example that is not capable of proving what he wants it to prove. He puts light forward as an example of something that only when grasped is capable of illuminating objects. In fact it is something that may be grasped when it illuminates an object, but it may not. Yet only if it must be grasped does it follow that cognition must be grasped; and only if cognition must be grasped is the argument in 1.1 successful and is the infinite regress in 1.2 generated. We have a case of unwarranted modal strengthening from ‘can’ to ‘must’; the example is only capable of proving possibility but it is put forward as proving necessity.

(2) The light example does illustrate the mode of necessity: Light must appear, in the sense that it must shine, in order to illuminate. The necessity attaches to its shining forth. If an observer is blind or separated from the lamp by an obstruction, that is a problem for the observer but it says nothing about the lamp itself, which is the thing being used to exemplify the nature of cognition.

The sense in which light ‘must appear’ can be illustrated by observing how it differs from other visible objects in this respect: any other visible object will not be illuminated if there is no light source close to it, or if a light source is close but is blocked from it by a non-translucent barrier. But neither of these two conditions can lead to a light not being illuminated. It cannot be deprived of light by its light source being too far away, because it is its own light source. It cannot be separated from its light source by an obstructing surface, because there is no gap between it

and its light source. In short, light must appear, because whereas other objects can remain in the dark, a light never can. If it is objected that these considerations do not prove that it must be perceived, just that it must shine forth, this could be held to be irrelevant. It is true that in the case of light, there may be no observer in its vicinity, or an observer may be blocked from its view; but neither of those are possible

in the case of cognition, for there that which is illuminating and that which is observing are the same thing.20 There is nothing in the universe other than cognition itself that could serve as an example, for the Vijñānavādin, of something that is both an illuminator and an observer of that illuminating, because it is the only thing in the universe capable of observation. But cognition is obviously inadmissible as an example: it is the proofsubject

(pakṣa). Thus the impossibility of providing an example of something that not only necessarily shines forth, but also is necessarily perceived, cannot be taken as a fault. The way the example works is by showing that light, when illuminating, necessarily shines forth; it thus provides evidence that cognition, when illuminating necessarily shines forth. We then simply have to add the considerations that cognition will always be present at its own shining forth, and can never be separated from it by an obstruction, in order to arrive at the conclusion that cognition, when illuminating, will necessarily appear to itself.

Why will it necessarily appear to itself before the object appears to it? If cognition is like light, then just as some time elapses between a flame arising and light from that flame reaching the object, so some time will elapse between a cognition arising, at which point it becomes aware of itself, and reaching an object that exists outside of it (which is the realist assumption that is in play in this argument).

This interpretation differs from the first one by asserting that light must appear, rather than claiming that it may or may not. This necessity attaches only to light’s shining forth, however. It is accepted that an observer may see an object, but not see the source of light that is illuminating that object. There is one way of interpreting the light example, however, that maintains that an observer must indeed always see the light that is illuminating the object they are looking at.

(3) This involves identifying the light that must be seen as not the source of light, but the light that emanates from that source and illuminates the object when it reaches it. Two considerations, taken together, make this interpretation plausible.

(A) It is common in Sanskrit philosophical sources to regard light as existing not only at the location of the source, e.g. the flame, but also as spreading out from there.21 (This also fits with the understanding implied in such English expressions

20 See Taber (2010: 285). 21 See for example the Nyāyavārttikatātparyaṭīkā sentences given in note 8; and Ślokavārttika, śūnyavāda 157, where, in the compound pradīpaprabhā-, pradīpa refers to the

as ‘cast some light on this’, which is a request not for someone to place a flame or an electric torch/flashlight on the object we are trying to see, but rather for them to bring it close enough so that light can spread out from there and fall on the object.) Whereas the words dīpa and pradīpa refer more commonly to the flame or lamp, words such as ābhāsa and prabhā refer more commonly to the light that emanates from there. The word prakāśa is used in both senses.

(B) This light that spreads out from a source such as a flame or the sun was regarded as visible, perceptible to the eye.22 This is not as strange as it might sound. After all when, on a sunny day, we say ‘there’s a lot of light in this room’, the light we are detecting there through our faculty of sight is not the sun itself. The advantage of this interpretation is that this light must indeed be seen in order to illuminate an object, so it provides a better explanation than the other two do of

lamp and prabhā to the light that spreads out from there, it being maintained in the verse that there is a time difference between the appearance of the pradīpa and the appearance of the prabhā. The distinction between pradīpa, a flame, and prakāśa, the light that spreads out from there to illuminate a pot, is evident in the passage from Śrīdhara’s Nyāyakandalī cited in note 25. See also Paramokṣanirāsakārikāvṛtti ad 34: pradīpaprakāśasya hi

pradīpotpattau krāmataḥ svaviruddhatamonivartanena tatra tatra deśe prasaraṇaṃ matam. ‘For when a flame arises, the light of the flame, moving to various places by removing the darkness that opposes it, is held to spread’. 22 See Nyāyavārttika ad 3.1.38, discussed in Watson and Kataoka (2010: 305–306); and

Nyāyamañjarī Vol. 1 p. 211,7–9: so 'yaṃ sūryaprakāśa ḥ prakāśāntaranirapek ṣacakṣurindriyaprathamagṛhītaḥ, ciram avatiṣṭhamānas tadindriyagrāhya eva viṣaye gṛhyamāṇe karaṇatām upayāti. ‘This light from the sun is first grasped by the faculty of sight without the need for any other light; remaining for a while, it becomes an instrument when the object that is grasped by the same faculty is grasped.’ When we look at a pot that is illuminated by sunlight, our visual faculty first grasps the light in front of the pot, and then the pot. The light referred to here is not the sun itself, but the light that has spread out from there and reached the pot.

why the Buddhist may have thought the second aspect of the light analogy to be valid. Why must it be seen? Whenever we see an object, there will always be some of this light between us and that object, even if the source of the light is hidden from us.

Our faculty of seeing must encounter this light on its journey towards the object. Hence if this light really is perceptible to the eye, we will not be able to see the object without first seeing this light. Thus this interpretation also explains why the light must be seen first. Vācaspati claims that this light that spreads out from a flame, and not the flame itself, is the true illuminator:23 this lends further credence to the idea that it is this light we should look to in order to understand the maxim that illuminators need to be grasped in order for the objects they illuminate to be grasped.

Is there any evidence in the Nyāyamañjarī bearing on which of these three interpretations was the way that Jayanta understood the light example? It seems that he does not regard it as invalidated by Taber’s objection, because although he refutes this Buddhist argument – that cognition must be perceived in order for its object to be perceived, as light must be perceived in order for an object illuminated by it to be perceived – at length and in several different ways (§§ 4.2–4.6), he never challenges its example. That is to say, he never disputes that light must be perceived in order for its object to be perceived; he just disputes that cognition must be per

23 Nyāyavārttikatātparyaṭīkā ad 3.1.32: prabhā hi visāriṇī tam arthaṃ prāpya prakāśayati, na tu pradīpaḥ. ‘For the light that spreads out [from the lamp], having reached the object, illuminates it; it is not the lamp [itself that illuminates it].’ In the previous sentence he has described the lamp as, ‘the light that is densely gathered at the location of the wick’: vartideśasthena piṇḍitena tejasā pradīpena.

ceived in order for its object to be perceived, and he does this by arguing that cognition is not like light. Is there any evidence to suggest whether he was envisaging the light that must be perceived as the source of light or the light that spreads out from there? In favour of the latter is the passage from another part of the Nyāyamañjarī given in note 8. He maintains there, while writing as the siddhāntin, that when light from the sun illuminates a pot, we see that sunlight before we see the pot. And it is clear there that he means by sunlight, not the sun, but light that touches the pot.24

2.4 Two More Features of the Light Analogy

We have separated out two different features of the light analogy.

(1) In order for light to illuminate it must be grasped. (2) Light is not illuminated by another light.

A third can be added, which is an extension of 2:

(3) Light is illuminated by itself (svata eva prakāśyate)

Unlike the previous two this involves a claim of reflexivity. That the three are distinct from each other can be seen from the way that they are susceptible to different refutations. 1 can be refuted by John Taber’s point, but that does not refute either of the other two. 2 is uncontroversial, and is used in contexts other than svasaṃvedana by traditions that do not accept the latter.25 3 can be refuted by arguing that light is never at the same time both illuminating and being illuminated.26 That does not refute either of the other two.

24 The evidence is not completely one-sided, however. For Jayanta can only have intended the example to be the light that spreads out from a lamp if he is using pradīpa (in prakāśatvāj jñānasya pradīpavat pūrvaṃ grahaṇam, § 4.4.1 in Kataoka 2003) to refer not only to the flame, which would be the most natural interpretation, but also to the light that spreads out from there. 25 See for example Nyāyabhāṣya ad 5.1.10: antareṇāpi pradīpāntaraṃ dṛśyate

pradīpaḥ, tatra pradīpadarśanārthaṃ pradīpopādānaṃ nirarthakam, ‘A lamp can be seen even without another lamp; that being the case, to take up a lamp in order to see another lamp is pointless’; and Nyāyakandalī p. 324,7–8: yathā ghaṭādiṣv aprakāśasvabhāveṣu pradīpādeḥ prakāśasvabhāvāt prakāśo bhavati. na tu

pradīpe pradīpāntarāt prakāśaḥ, kiṃ tu svata eva. ‘Just as light occurs in/on pots and the like, which are not of the nature of light, because of lamps and the like, which are of the nature of light. But the light in a lamp is not caused by / does not come from another lamp, rather it is there innately.’ For an early (pre-Dignāgan) example of this feature of light being applied to cognition, see Vākyapadīya 3.1.106: yathā jyotiḥ prakāśena nānyenābhiprakāśyate

jñānākāras tathānyena na jñānenopagṛhyate || ‘Just as a light is not illuminated by another light, so is the form of a cognition not grasped by another cognition.’ 26 That it is never both illuminating and being illuminated can be argued for in different ways. Kumārila does so by maintaining that it is never actually being illuminated: fire and

3 was objected to by those such as Kumārila and Jayanta who do not accept svasaṃvedana and do not accept that there is anything self -illuminating (svaprakāśya), whether light, cognition or anything else.27 But it was also objected to by some traditions that accepted svasaṃvedana, namely Advaita Vedānta,28

other lights do not have a form that is illuminated, because they do not depend on light (na te prakāśyarūpā hi prakāśasya anapekṣaṇāt, Ślokavārttika, śūnyavāda 65cd). This is the same kind of argument as we find in Bodhicaryāvatāra 9:18, and

Prajñākaramati’s commentary thereon: illumination entails the removal of darkness. Since a pot can be covered by darkness, it can be illuminated. Since a light can never be covered by darkness, it cannot be illuminated.

See Garfield (2006: 224, note 3), who regards this argument as consistent with that put by Nāgārjuna in Mūlamadhyamakakārikā 7:9, and by

Buddhapālita, Bhāvaviveka, and Candrakīrti in their commentaries on that. In these arguments we can detect the intuition that it only makes sense to say of something that it is illuminated if, prior to the light falling on it, it required illumination, i.e. existed in an unilluminated state. Light shines forth, but it does not illuminate itself because it was not there prior to the supposed act of illumination. It is impossible for it, at any point, to

exist while requiring illumination. And if it has never required illumination, what sense does it make to say that it has illuminated itself? Even if it is true that light is not illuminated, it is certainly grasped, and it enables things – including itself – to be grasped, so is it not reflexive in this sense at least, that it enables itself to be grasped? In other words is it not simulaneously instrument and object of a single action of grasping? (In

which case it would render plausible that cognition is instrument and object of a single action of grasping.) Kumārila and Jayanta answer no; they maintain that the instrument of its grasping is not itself, but the eye (or the faculty of sight). See Ślokavārttika, śūnyavāda 66ab, and the following passage in the Nyāyamañjarī (Vol. 1, pp. 210,16–211,7): maivam, ekasya kārakasyaikasyām eva kriyāyāṃ karmakaraṇabhāvānupapatteḥ. savitṛprakāśavad iti cet,

na kriyābhedāt. yatrāsau karaṇaṃ na tatra karma, yatra vā karma na tatra karaṇam iti. ghaṭādiviṣayapramitijanmani karaṇam eva taraṇiprakāśaḥ, na karma; tadgrahaṇakāle tu karmaivāsau na karaṇam. kiṃ tarhi tatra karaṇam iti cet, kevalam eva cakṣur iti brūmaḥ, ālokagrahaṇe cakṣuṣaḥ prakāśāntaranirapekṣatvāt. katham evam iti cet, aparyanuyojyā hi vastuśaktiḥ. ghaṭādigrahaṇe cakṣur udyotam apekṣate nodyotagrahaṇa iti kam anuyuñjmahe. An opponent asserts that sunlight is both instrument (i.e. means) and object of the same action, i.e. that when it is grasped, it is also the instrument that enables that grasping,

and that when it is the instrument that enables the grasping of a pot, it is itself grasped. Jayanta states that we have to separate out two distinct types of action. When sunlight is the instrument, it itself is not grasped; when it is itself grasped, it is not the instrument. What, then, is the instrument when it is grasped? Just the eye, for when the eye perceives light it does not depend on the presence of any other light. The advocate of the reflexivity of light wants both the eye and light to be instruments then (in order that light is both instrument and object), so he asks how it can be that

the eye, despite depending on light when it is perceiving all other things, does not when it is perceiving light. Jayanta replies that it is simply the way things are that when the eye perceives a pot, say, it requires light to make that object visible, but when it perceives light it does not. ‘Who can be questioned about that?’, he rhetorically asks in answer to the opponent’s request for an explanation. The implication is that here we have just reached a

fact about the way the world is that cannot be explained further (see also § 4.4.2.2 of Kataoka 2003). Jayanta's position, then, is that light is both instrument and object of perception, but never of the same perception – rather instrument of some perceptions and object of others. This accords with Vātsyāyana's treatment of light ad Nyāyasūtra 2.1.19. 27 See § 4.4.2.2 of Kataoka (2003). 28 See, e.g., Śaṅkara’s Bṛhadāraṇyakopaniṣadbhāṣya ad 1.5.3, p. 220,9–11.

Prābhākara Mīmāṃsā29 and Śaiva Siddhānta.30 They concluded from 2, i.e. from the fact that light is not illuminated by another light, not, as did vijñānavāda, that it is illuminated by itself, but rather that:

(4) Light is not illuminated at all; it shines forth by itself

The contrast between 3 and 4 may be clearer in the Sanskrit: 3 maintains that light svata eva prakāśyate; 4 denies this and asserts that light rather svata eva prakāśate. It is not correct to say that light is illuminated, for that would only make sense if were the recipient of illumination from something other than itself, exactly what is denied by 2, to which all parties agree.

4 is asserted of light by those who accept svasaṃvedana but resist vijñānavāda. For them light appears (prakāśate) by itself,31 but never as the object of illumination; so similarly cognition appears through svasaṃvedana, but never as the object. They are thus able to maintain a firm dualism between cognition and objects, despite acceptance of svasaṃvedana. On one side of the divide is something (the self, with cognition as its nature32) that is never illuminated (prakāśyate), but rather shines forth (prakāśate). On the other side are all other things in the universe; they differ from the first in two ways: they are illuminated, and they are incapable of shining forth.33

29 See Nyāyamañjarī Vol. 2, p. 273,1–11; Bṛhatī p. 64,2–5; Ṛjuvimalā p. 64,11b–8b. These passages deny that the self (or cognition) is the object of illumination (prakāśyate); rather it appears by itself (svata eva prakāśate). Or they deny that self or cognition appear to us as the object of perception (grāhyatayā), or as the object of cognition (saṃvedyatayā); rather they appear as, respectively, the perceiver (grāhakatayā) and cognition itself (saṃvittayā). They do not explicitly deny that light is the object of illumination. So to regard the Prābhākaras as holding the latter view depends on an

inference from the exemplified to the example. 30 See, e.g., Nareśvaraparīkṣāprakāśa p. 18,9–11. 31 I.e. without requiring any means other than itself. All other objects, by contrast, can only appear if they receive illumination from something other than themselves. 32 Although ‘with cognition as its nature’ describes the view of Advaita Vedānta and Śaiva Siddhānta, for Prābhākara Mīmāṃsā we have to substitute for that phrase ‘with cognition as its property/quality’. 33 4 is not refuted by either of the refutations mentioned above. It can be refuted by asserting that light does not appear by itself, since it requires the presence of a perceiver for it

3 Why is light a useful analogy for the Vijñānavādin?

In this final section I turn to the question of the value of the light analogy for vijñānavāda; I argue that although light is in some ways a helpful analogy for the Vijñānavādin, in other ways it is thoroughly inappropriate.

3.1 Helpful aspects of the light analogy

The ways in which light is of use to vijñānavāda can be divided into those that the Vijñānavādin would admit to, and those that he would not admit to.

3.1.1 Things the Vijñānavādin would admit to Light is something that appears, we see it; therefore comparing cognition to light is of help for establishing, against the Naiyāyikas and Bhāṭṭa Mīmāṃsakas, that cognition is perceived. And all four of the specific features of light, or claims made of light, in the previous section are of help to the Vijñānavādin.

(1) That light must be grasped in order to illuminate renders plausible the claim that cognition must be grasped in order to illuminate. Against the Buddhists on this point, the Naiyāyikas and Bhāṭṭa Mīmāṃsakas held that it is quite possible to be an illuminator (prakāśaka) and not be perceived, and

they pointed to the eye, or the faculty of sight, as a clear example. To support their position that cognition, which is an illuminator, is not perceived when illuminating (= enabling objects to be perceived), they used the faculty of sight as a corroborating example of an illuminator that is not perceived when illuminating. Thus we can see that it was incumbent upon the Vijñānavādins to come up with an example of an illuminator that contrasts with the Naiyāyikas’ and Bhāṭṭa Mīmāṃsakas’ example of the faculty of sight by being perceived while illuminating. Light fits this purpose exactly.

(2) That light is not illuminated by another light is of use to the Vijñānavādin when he comes to challenge the Naiyāyika view that cognition is perceived by another cognition (through anuvyavasāya, “subsequent determination”).

(3) We can separate out two things that become plausible of cognition if it is accepted that light illuminates itself: the reflexivity of cognition and the nonduality of perceiver (i.e. cognition) and perceived objects. (A) As is wellknown, those who argued that cognition is aware of itself had to deal with the objection that nothing can act on itself (svātmani kriyāvirodhaḥ): fire cannot burn itself, an axe cannot cut itself, the tip of the finger cannot touch itself

to appear to. There is a less sophisticated and a more sophisticated version of this argument: see Watson and Kataoka (2010: 333–335) for this contrast and for an elaboration of the more sophisticated version as put by Jayanta in § 4.4.2.

etc.34 Light, if it does illuminate itself, can neutralize the force of these examples by providing an example of something that does act on itself. And if light can do it, why not cognition? (B) If light illuminates itself, it shows that subject and object of illumination need not be separate. Hence the non-duality of the subject and object of perception, a kind of illumination, becomes more plausible. If cognition is like light in being able to fill the role of both illuminator and illuminated, why postulate a separate, external object?35

(4) Occasionally Vijñānavādins preferred to use the fourth feature of light rather than the third. One of the reasons for this may have been that to claim that cognition cognizes itself (as light illuminates itself) is liable to be misunderstood as imputing activity (vyāpāra) to cognition, something that was denied by Dignāga at the outset of this tradition of Buddhism.36 Thus we find

Vijñānavādins in this tradition preferring to characterize light as (rather than illuminating itself) simply arising with luminosity (prakāśa) as its nature, and cognition as simply arising with luminosity, or sentience, as its nature. Since it has two forms within it, that of itself and that of its object, its sentience means that it can be characterized as cognizing both of those; but it is more accurate to say just that it arises with those two forms shining forth within it.37

3.1.2 Things the Vijñānavādin would not admit to It is not only empirical features of light that are of help to the Vijñānavādin, but also linguistic features of the word prakāśa. There is an ambiguity in the term which the Vijñānavādin exploits.

Prakāśa can either mean something that illuminates other things, makes them manifest, makes them appear (i.e. it can have the sense of prakāśaka, that which prakāśayati); or it can mean something that shines forth, is manifest, appears (i.e.

34 See for example Tattvasaṅgrahapañjikā ad 1683, p. 585,11–12; Brahmasūtrabhāṣya ad 2.2.28, pp. 398,15–399,2; Tarkabhāṣā p. 16,4–10; Bodhicaryāvatāra 9:17 and Yao (2005: 29, 52, 53, 63, 102, 124 and 148). 35 But see below, sections 3.2.1–3.2.3, for doubts about whether this feature of light really makes the non-duality of subject and object more plausible. 36 Dignāga was here following earlier tradition: see Kellner (2010: 219 and footnotes 49 and 51).

37 See for example the many places where Dharmakīrti writes svayam eva prakāśate or some synonymous expression (e.g. Pramāṇavārttika 3:327, identical to Pramāṇaviniścaya 1:38, and Pramāṇavārttika 3:446, 3:478, 3:480, 3:481). See Manorathanandin’s claim that when Dharmakīrti writes that cognition perceives itself (dhīr ātmavedinī, Pramāṇavārttika 3:329) he is being metaphorical, and both Devendrabuddhi and Manorathanandin’s remarks (given and translated by Kobayashi 2006: 2–3) on the same verse to the effect that ‘light illuminates itself’ is to be understood simply to mean that light arises with luminosity as its nature. And see Śāntarakṣita’s interpretation of svasaṃvedana (discussed by Williams 1998 and Arnold 2005) as meaning not that cognition cognizes itself, but merely that cognition is sentient.

it can have the sense of that which prakāśate).38 For the Vijñānavādin cognition is prakāśa in both senses; for Nyāya and Bhāṭṭa Mīmāṃsā it is prakāśa only in the first sense.

Vijñānavāda uses this ambiguity when, in arguing against the Naiyāyikas and Bhāṭṭas, it explicitly or implicitly uses the assumption that since cognition is prakāśa in the first sense, it must be prakāśa in the second sense.39

3.2 Ways in which the light analogy is not helpful

We will next consider ways in which the light example is not helpful. Having done that we will ask whether the lack of fit between light and cognition is inevitable.

3.2.1 Discord between the beginning and the end of the argument

The first deficiency of the light analogy can be observed by noting that there is discord between the first and the last two stages of our argument. 1 is dealing with a model of cognition needing to be grasped first in order to illuminate an object, in the way that light needs to be grasped first in order to illuminate an object. An object, separate from the illumination, is very much in the picture. Yet in 4 we reach the position that we can do away with a separate object altogether. The justification in 1(.1) for cognition’s being perceived is that otherwise it would not be able

38 For these two meanings of the word, see § 4.4.1 of Kataoka (2003). 39 For example see the following, put by Jayanta’s Vijñānavādin. It closely resembles the argument we examined above (for minor differences see Watson and Kataoka 2010: 327– 328). The Vijñānavādin is arguing that that which the Naiyāyika takes to be the appearance of a perceived object (grāhya) is nothing but the perceiver (grāhaka) appearing in a certain form. grāhakād anyo hi grāhyo jaḍātmā bhavet. grāhakas tu prakāśasvabhāvaḥ, grāhakatvād eva. dvayapratibhāsaś ca nāstīty uktam. tatrānyatarasya prakāśane jaḍaprakāśayoḥ

katarasyāvabhāsituṃ yuktam iti cintāyāṃ balāt prakāśa eva prakāśate, na jaḍaḥ. nirākāraś ca na prakāśaḥ prakāśata iti tasmin sākāre prakāśamāne kuto jaḍātmā tadatirikto 'rthaḥ syāt (§ 4.2.1 in Kataoka 2003). ‘For a perceived object, were it different from a perceiver, would be insentient in nature. A perceiver, on the other hand, is of the nature of an illumination, just because it is a perceiver. And it has already been stated that there are not two appearances. In that case (tatra), given that one of the two appears [the question arises as to] which of the two, the insentient entity or the

illumination, is the appropriate [candidate] to appear. Considering this, obviously it is the illumination that appears and not the insentient entity. And an illumination does not appear without a form. Therefore, given that it appears together with a form, what is the need of an insentient object which is different from that [form-containing illumination]?’ The question is whether the one form that appears to us is the perceiver, or an insentient object. The Vijñānavādin asserts that the perceiver (grāhaka) must be of the nature of prakāśa by virtue of the fact that it is the perceiver. If that is true,

then the Vijñānavādin has won the argument; for if the perceiver is of the nature of prakāśa, then it appears, because it is axiomatic that prakāśaḥ prakāśate (prakāśa appears). But in fact it does not follow that whatever is a perceiver must be of the nature of prakāśa – the Naiyāyika would dispute that. What could be said to follow is that it is of the nature of the prakāśaka (in that it is that which brings about the illumination of the object), but that is not the same thing.

to go on to perceive an object, yet in 3 and 4 we learn that it does not need to go on to perceive an object. The cognition spoken of in 1 is something separate from its object, as light is separate from the object it illuminates; how then can it, in 3 and 4, encompass within itself the form of the object? We have here the same switching of levels between dualism and non-dualism that we have in the sahopalambhaniyama argument, which concludes the non-duality of object and cognition from the fact that they are necessarily perceived together. That argument starts by talking of two things, as implied both by the use of the word ‘together’ (saha), and by the fact that the example of the argument is two moons. Yet we end with one thing. We were dealing with two things, but suddenly one of them disappears by turning out to be part of the other.

For an argument to begin with dualism and end with non-dualism is not in itself a problem. The lack of fit can be justified on the grounds that the beginning of the argument is stated from a provisional, common-sense point of view, whereas the end is stated from a final, post-correction point of view. The problem is the light example. It would only be capable of facilitating this switch from two things to one if it were something that, viewed from one

perspective is two things, but from another is just one. This is indeed the case with the two moons (that are mistakenly seen by someone with an eye defect when they are looking at the one moon). The two moons, at both the beginning and the end of the argument, can stand for cognition and object (since at the beginning of the argument the two moons can be considered to be two things, and at the end they can be considered to be one). Light, by contrast,

stands just for cognition at the beginning, not the object as well, yet has to stand for both at the end, despite not being suited to do so. Why is it not suited? Because light is not something that contains within itself the forms of the objects it illuminates. We have an example that implies dualism between objects and cognition being used to establish their non-dualism. The light example, at worst, undermines the transition from non-dualism to dualism, and at best is completely unable to facilitate it.

3.2.2 Two additional premises

The Buddhist could reply that the disappearance of a separate object between the beginning and end of the argument is brought about validly by two further premises. The jump from ‘we perceive cognition’ to ‘we do not perceive an object that is external to cognition’ is brought about by the combined force of: (1) ‘cognition would not be perceived unless it had form’ and (2) ‘we do not perceive two forms’. But considerations concerning light could be used to block either of these premises. Against the first it could be argued that we do perceive cognition without form, in the way that we perceive light without a specific form. This was the view of Prābhākara Mīmāṃsā, Advaita Vedānta and Śaiva Siddhānta. If the Buddhist insists that when we actually perceive light, it has a certain form, an opponent could point out that it may have form (a certain brightness and a certain colour

somewhere between white and yellow), but it does not have the form of the object that it illuminates. So similarly even if we perceive cognition with form, we do not perceive it to have the form of the object it illuminates. This rejects the second of the two premises (that we do not perceive two forms) by asserting that we perceive cognition with a certain kind of form and an object with a different form. Whichever of these two considerations concerning light is preferred, i.e. whether it is regarded as perceived without form or with form, we are forced to conclude that if cognition were really parallel to light, then we would perceive both cognition and an object separate from that cognition.40 In other words the light example serves better those who combine svasaṃvedana with realism about external objects, such as Prābhākara Mīmāṃsā and Śaiva Siddhānta, than it does those who combine svasaṃvedana with idealism, such as the Vijñānavādins.

3.2.3 Light is a manifestor and cognition is not

Light is the standard example in Indian philosophy of a manifestor (vyañjaka), that is to say something that manifests or reveals other things, enabling them to be known. The concept of a vyañjaka is frequently contrasted with that of an

utpādaka, something that causes other things to come into existence. The test of whether something is a vyañjaka or an utpādaka is whether the things that it reveals existed earlier or not. An utpādaka brings its object into existence at the moment it acts, whereas a vyañjaka can only act on an object that already exists. Insofar as it is agreed that light is a vyañjaka and not an utpādaka, it should be agreed that the objects illuminated by light exist before the moment that the light comes to fall on them. We would surely be surprised if someone were to claim that light is an utpādaka, i.e. that it

brings the objects it illuminates into existence. Yet this is precisely the move that a Vijñānavādin such as Prajñākaragupta is forced into. He is faced with a lack of fit between cognition and that to which vijñānavāda frequently compares it, light. The lack of fit is that light is a manifestor, but cognition is not. It does not meet the requirement to qualify as a manifestor, because the objects it reveals, according to vijñānavāda, are not separate from it and do not pre-exist it, but are rather brought into existence by it and with it. In order to remove this inconsistency he makes the surprising move mentioned above: he denies that light manifests objects, and claims, counter-intuitively, that it rather brings them into existence

(Pramāṇavārttikālaṅkāra p. 353; see Kobayashi 2006: 4–7). It seems, then, that one either has to deny that light is a manifestor, or one has to admit that it is more suited to exemplify the cognition of the realists, which reveals pre-existing objects, than that of the Vijñānavādins, which contains objects within itself and is non-different from them.

40 The third option, that light is not perceptible at the time that it illuminates an object, only the illuminated object being perceptible, is clearly disastrous for the Vijñānavādin. The latter’s claim that cognition is (like) light would then lead to the view that cognition is not perceptible at the time that it illuminates an object: we arrive at the Bhāṭṭa Mīmāṃsaka and Naiyāyika view.

To take stock of the last 3 points: (1) cognition, as represented in stage 1 of the argument, where it is compared to light, does not look like something capable of making a separate object redundant in 3 and 4; (2) the premises used by the Buddhist to enable the jump from ‘we perceive cognition’ to ‘we do not perceive an object that is external to cognition’ can be blocked by considerations concerning light; (3) light is a manifestor of objects separate from it. Involved in all three points is the fact that there is a lack of parallelism between light, and cognition as conceived of by the Vijñānavādins, i.e. as something that makes the existence of separate objects redundant by containing objects within itself and bringing them into existence.41 One could respond in two ways: either by asserting that a certain lack of fit between cognition and light is inevitable; or by searching for a better example.

3.3 Revisionary metaphysics

One could diagnose the lack of fit to be an inevitable result of vijñānavāda’s nature as a piece of ‘revisionary metaphysics’.42 We have seen this revisionism in the way that after reaching the conclusion of the argument, one has to correct the first stage, which assumed an object separate from its illuminator. And we saw it in Prajñākaragupta’s claim that ‘a pot is illuminated by light’ should correctly be understood to mean that a pot is produced by light. On this view, then, the lack of fit is a consequence of the contrast between the world as viewed prior to Vijñānavādin revision, and the world as viewed subsequently to it. Vijñānavāda regards the world of everyday interaction and language as based on ignorance and as unreflective of reality. If it then uses this world of vyavahāra as the source of its examples and analogies, there will inevitably be some discord. Yet it cannot use anything else, for two reasons. Firstly because the only way we can talk about the world is through language. Secondly because if Vijñānavādin arguments against realism are to convince realists, they have to use examples that realists will accept.

3.4 Better example?

Or one could ask whether a better example may be forthcoming. When characterizing the meeting of cognition and objects, we have a choice regarding direction of movement. We can either depict cognition as moving towards the object and illuminating it; or the object (or light from the object) moving to cognition and affecting it. The second seems more suitable for sākāravāda, and indeed the Sāṅkhya Sākāravādins opted for the second, comparing cognition to a mirror. This is a bet

41 Even for the Sautrāntika acceptance of separate external objects and at the same time of form-containing cognition, light will struggle to perform its role, for light is not coloured by the objects it illuminates; it does not take on their form. 42 I use this expression to denote a system of metaphysical thought that aims – unlike ‘descriptive metaphysics’ (both expressions having been coined by Peter Strawson) – at fundamentally revising our ‘common-sense’ concepts and beliefs.

ter example of something that bears the form of its object, than light. It is more suggestive of the sākāravāda of the Sautrāntikas, however, than that of the Vijñānavādins, since

the forms on the surface of a mirror do not make redundant the supposition of external objects, but rather depend on external objects as their cause. Ideally what is needed is an example of something that on the one hand contains form, and on the other hand makes redundant those things outside of itself that its forms are taken to represent. There is such an example and it is actually one that is of the nature of light. When we look at a film projected onto a screen, we are looking just at coloured lights, but we take them to be a real person, say, in a real car in a real street with real shops on it. So I end with the contention that the existence of projectors and in particular the coloured lights that they project, which depict certain forms (ākāras) that are mistakenly taken to be of external objects, makes light a more compelling example for vijñānavāda than it was before the invention of these machines.43

Primary Sources

Aṣṭaprakaraṇa: Dvivedī, B. (ed.), Tattvaprakāśa, Tattvasaṅgraha, Tattvatrayanirṇaya, Ratnatraya, Bhogakārikā, Nādakārikā, Mokṣakārikā, Paramokṣanirāsakārikā. Vārāṇasī 1988.

Ṛjuvimalā: Ṛjuvimalā of Śālikanātha Miśra. See Bṛhatī. Tarkabhāṣā: Iyengar, H. R. R. (ed.). Tarkabhasha and Vadasthana of Mokshakaragupta and Jitaripada. Mysore 1944. Tattvasaṅgraha: Shastri, D. (ed.). Tattvasangraha of Ācārya Shāntarakṣita with the Commentary ‘Pañjikā’ of Shri Kamalashīla. 2 Volumes. Vārāṇasī 1981–1982. Nareśvaraparīkṣā: Shastri, M. K. (ed.). Nareśvaraparīkṣā of Sadyojyotis with the Prakāśa Commentary of Rāmakaṇṭha. Delhi 1989 (Original Edition Śrīnagar 1926).

Nyāyakandalī: The Praśastapāda Bhāshya with Commentary Nyāyakandali of Sridhara ed. V. P. Dvivedin. Sri Garib Dass Oriental Series 13. Delhi: Sri Satguru Publications, 1984. [First Edition Banaras 1895.]

43 Alternatively, instead of these considerations leading us to disparage the worth of the light analogy prior to the invention of projectors, they could encourage us rather to adjust our understanding of it: by seeing it as appealing to lights of many colours. This makes it much

Nyāyabhāṣya: Thakur, A. (ed.). Gautamīyanyāyadarśana with Bhāṣya of Vātsyāyana. New Delhi: Indian Council of Philosophical Research 1997. Nyāyamañjarī: Varadacharya, K. S. (ed.). Nyāyamañjarī of Jayantabhaṭṭa, with Ṭippaṇi—Nyāyasaurabha by the Editor. 2 Volumes. Mysore 1969 and 1983. Nyāyavārttika: Thakur, A. (ed.). Nyāyabhāṣyavārttika of Bhāradvāja Uddyotakara. New Delhi: Indian Council of Philosophical Research 1997. Nyāyavārttikatātparyaṭīkā: Thakur, A. (ed.). Nyāyavārttikatātparyaṭīkā of Vācaspatimiśra. New Delhi: Indian Council of Philosophical Research 1996. Nyāyasūtra: see Nyāyabhāṣya.

Paramokṣanirāsakārikāvṛtti: see forthcoming edition and translation by Alex Watson, Dominic Goodall and Anjaneya Sarma. See also Aṣṭaprakaraṇa. Pramāṇasamuccaya: Steinkellner, E. (ed.). Dignāga’s Pramāṇasamuccaya, 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 Jinendrabuddhiʼs Ṭīkā. URL: http://ikga.oeaw.ac.at/Mat/dignaga_PS_1.pdf.

Pramāṇavārttika 3 (pratyakṣa chapter): Tosaki, H. (ed.). Pramāṇavārttika III. Bukkyo-ninshikiron no kenkyu [*Studies in Buddhist Epistemology]. 2 Volumes. Tokyo: Daitoshuppansha 1979 and 1985.

Pramāṇavārttikālaṅkāra: Sāṅkṛtyāyana, R. (ed.). Pramāṇavārttikabhāṣyam or Vārtikālaṅkāraḥ of Prajñākaragupta. Patna 1953. Pramāṇaviniścaya: Steinkellner, E. (ed.). Dharmakīrti's Pramāṇaviniścaya; Chapters 1 and 2. Vienna, Beijing: Austrian Academy of Sciences, China Tibetology Research Centre 2007.

Bṛhatī: Sastri, P. A. C. (ed.). Bṛhatī, a Commentary on Śābarabhāṣya by Prabhākara Miśra with the Commentary Ṛjuvimalā of Śālikanātha Miśra. Benares 1929–1933.Bṛhadāraṇyakopaniṣadbhāṣya: Kāśināth Śāstrī Āgāśe, Ve. Śā. Rā. Rā. (ed.). Bṛhadāraṇyakopaniṣat; Ānandagirikṛtaṭīkāsaṃvalitaśāṃkarabhāṣyasametā . Ānandāśramasaṃskṛtagranthāvali 15. Ānandāśrama 1982.

Bodhicaryāvatāra: Vaidya, P. L. (ed.). Bodhicaryāvatāra of Sāntideva with the Commentary Pañjikā of Prajñākaramati. Buddhist Sanskrit Texts 12. Darbhangha: Mithila Institute 1960.

easier to understand the Vijñānavādin claim that objects are of the nature of prakāśa, or that

Brahmasūtrabhāṣya: Iyer, T. K. B. (ed.). Śrīśāṅkaragraṃthāvaliḥ, Complete Works of Sri Sankaracharya in the Original Sanskrit: Volume 7, Brahmasutra Bhashya. Madras: Samata Books 1983 (First Published 1910).

Vākyapadīya: Rau, W. Bhartṛharis Vākyapadīya; Versuch einer vollständigen deutschen Übersetzung nach der kritischen Edition der Mūla-Kārikās, hrsg. von Oskar von Hinüber. Stuttgart: Franz Steiner Verlag 2002.

Śābarabhāṣya: Frauwallner. E. (ed.). Śabarasvāmi’s Bhāṣyam zu den Mīmāṃsāsūtren 1.1.1–5: Materialien zur ältesten Erkenntnislehre der Karmamīmāṃsā. Veröffentlichungen der Kommission für Sprachen und Kulturen Süd- und Ostasiens 6. Wien 1968.

Ślokavārttika: Śāstrī, D. (ed.). Ślokavārttika of Śrī Kumārila Bhaṭṭa with the Commentary Nyāyaratnākara of Śrī Pārthasārathimiśra. Varanasi: Tara Publiations 1978.

Secondary Sources

Arnold, D. (2005). Is Svasaṃvitti Transcendental? A Tentative Reconstruction Following Śāntarakṣita. Asian Philosophy, 15(1), 77–111. Ganeri, J. (1999). Self-Intimation, Memory and Personal Identity. Journal of Indian Philosophy, 27(5), 469–483.

Garfield, J. (2006). The Conventional Status of Reflexive Awareness. What's at Stake in a Tibetan Debate? Philosophy East and West, 56(2), 201–228. Kataoka, K. (2003). <http://www2.lit.kyushuu.ac.jp/~kkataoka/Kataoka/NMvijR.pdf> This is an updated version of "Critical Edition of the Vijñānādvaitavāda Section of Bhaṭṭa Jayanta's Nyāyamañjarī." The Memoirs of the Institute of Oriental Culture 144, 115–155. The updated version, rather than the printed version, should be used, because the section numbers are different, and the section-numbering I refer to in this article is that of the updated version.

prakāśa includes within itself the form of objects.

Kataoka, K. (2006). "Jayanta no Yuishiki Hihan: Nyāyamañjarī Ninshiki Ichigenron Hihan Wayaku." [An annotated Japanese translation of the Vijñānādvaitavāda Section of Bhaṭṭa Jayanta's Nyāyamañjarī.] Tetsugaku Nenpo (Faculty of Letters, Kyushu University) 65, 39–85. Kellner, B. (2010). “Self-Awareness (svasaṃvedana) in Dignāga’s Pramāṇasamuccaya and -vṛtti: A Close Reading.” Journal of Indian Philosophy 38: 203–231. Kellner, B. (2011). “Self-awareness (svasaṃvedana) and infinite regresses: a comparison of arguments by Dignāga and Dharmakīrti.” Journal of Indian Philosophy 39/4–5, 411– 426.

Kobayashi, H. (2006). “The Proof of the Self-Luminous Nature of a Cognition.” Paper given at the 2006 World Sanskrit Conference in Edinburgh. Ram-Prasad, C. (2007). Indian Philosophy and the Consequences of Knowledge: Themes in Ethics, Metaphysics and Soteriology. Aldershot: Ashgate. Siderits, M. (forthcoming) “Dignāga’s Philosophy of Mind.” English version of an article due to appear in Korean in a Seoul National University volume dedicated to the Philosophy of Mind.

Taber, J. (2010). "Kumārila's Buddhist." Journal of Indian Philosophy 38, 279–296. Watson, A. and Kataoka, K. (2010). “Bhaṭṭa Jayanta's Refutation of the Yogācāra Buddhist Doctrine of Vijñānavāda: Annotated Translation and Analysis.” South Asian Classical Studies 5, pp. 285–352.

Williams, P. (1998). The Reflexive Nature of Awareness. A Tibetan Madhyamaka Defence. Richmond: Curzon Press. Yao, Z. (2005). The Buddhist Theory of Self-Cognition. London and New York: Routledge.