Skip to main content

There is Something Wrong with Raw Perception, After All: Vyāsatīrtha’s Refutation of Nirvikalpaka-Pratyakṣa


This paper analyzes the incisive counter-arguments against Gaṅgeśa’s defense of non-conceptual perception (nirvikalpakapratyakṣa) offered by the Dvaita Vedānta scholar Vyāsatīrtha (sixteenth century) in his Destructive Dance of Dialectic (Tarkatāṇḍava). The details of Vyāsatīrtha’s arguments have gone largely unnoticed by subsequent Navya Nyāya thinkers, as well as by contemporary scholars engaged in a debate over the role of non-conceptual perception in Nyāya epistemology. Vyāsatīrtha thoroughly undercuts the inductive evidence supporting Gaṅgeśa’s main inferential proof of non-conceptual perception, and shows that Gaṅgeśa has no basis for thinking that non-conceptual perception has any necessary causal role in generating concept-laden perceptual awareness. He further raises a number of internal inconsistencies and undesirable consequences for Gaṅgeśa’s claim that non-conceptual states are introspectively invisible. His own causal theory of perception is more parsimonious than the Nyāya account, and is equally compatible with direct realism. I conclude by noting several striking parallels between Vyāsatīrtha’s views and the conceptualism of John McDowell, while also suggesting that Vyāsatīrtha own conceptualism is not unduly constrained by some of McDowell’s limiting assumptions about concepts and perceptual contents.

This is a preview of subscription content, access via your institution.


  1. 1.

    See Chakrabarti (2000, 2001, 2003, 2004), Phillips (2001, 2011, pp. 37–38); and Chadha (2004, 2006, 2009, 2014, 2016). Bronkhorst (2011) and Maitra (2017) also weigh in on the debate.

  2. 2.

    NSMD 58, 252.12-13. See Bhattacharyya (1990a, pp. 52–59) for Gadādhara's discussion of why the viṣayatā of nirvikalpaka and savikalpaka states must be distinct.

  3. 3.

    Sharma (2008, pp. 306–311) gives a brief summary of Tarkatāṇḍava. Duquette (2018) examines portions of the Tarkatāṇḍava in which Vyāsatīrtha refutes of Gaṅgeśa’s inferential proof of God’s existence and criticizes certain aspects of Mīmāṃsā hermeneutics. For a discussion of Vyāsatīrtha’s relation to previous Nyāya thought, see Williams (2014). McCrea (2015) and Stoker (2016) situate Vyāsatīrtha within the sectarian context of Vedāntic polemics and politics in 16th-century South India.

  4. 4.

    Two of the three objections taken up by Tatacharya don’t actually originate with Vyāsatīrtha. The first objection was already made by Gaṅgeśa’s commentators, and is discussed in footnote 16 below; the second objection was rejected by Gaṅgeśa at TCM, 859.01-860.06. While the third objection (TT, 489.02-06) has deeper implications regarding the introspective invisibility of non-conceptual perception, it’s ultimately just a flippant tu quoque that Vyāsatīrtha wouldn’t defend in its own right.

  5. 5.

    Phillips and Tatacharya (2009, p. 611) tentatively identify Gaṅgeśa's pūrvapakṣin as a Prābhākara Mīmāmṣaka.

  6. 6.

    See TCM, 857.03-04; BP 58ab, 254.01.

  7. 7.

    TB, 36.29-30.

  8. 8.

    TCM, 857.05-06; TT, 484.06-08.

  9. 9.

    Tatacharya (1992, p. 320.05-09) confirms that a nirvikalpaka awareness could not properly be said to have the form, “This is something,” as “somethingness” (kiṃcittva) would become the awareness's predicative content/prakāratā, thereby rendering the awareness to be savikalpaka in nature.

  10. 10.

    TCM, 857.04-05.

  11. 11.

    The two weakest inferences are raised at TT, 485.08-11. The first inference is mentioned by Śaśadhara (NSD, 45.21-22), and claims that the property of being an awareness (jñānatva) shares the same locus as the property of not having predicative content (niṣprakārasamānādhikaraṇa), because awarenesshood resides in all awareness-states (sakalajñānavṛttitva), just like the property of being existent (sattā). The example of sattā is a corroborating example (sapakṣa) insofar as it, being a property of all existing things, is a property of all awareness-states, and is also located in other things which technically lack predicative content as well, like a pot. In this way, every awareness is shown to lack predicative content, and hence be non-conceptual in nature. Yet, both Śaśadhara and Vyāsatīrtha agree that the proving property here—“being a property of all awareness-states”—is lacking in genuine probative power (aprayojaka). There is no supporting argument—no anukūla-tarka—which would conclusively prove that the proving property (hetu) cannot be found anywhere the target property (sādhya) is absent, in which case an opponent could suspect that the property of residing in all awareness-states could belong to awarenesshood even though awarenesshood does not share the same locus as the property of not having predicative content. In other words, the opponent could suspect that the hetu is present in the pakṣa without the sādhya also being present there. This inference would also absurdly result in an over-extensive application (atiprasaṅga): As Rāghavendratīrtha (ND, 485.12) explains, the same hetu could be used to prove that awarenesshood shares the same locus as pothood (ghaṭatvasamānādhikaraṇa). It’s also possible to corroborate such a claim using the example of sattā, since sattā resides in all awareness-states and all pots. In a similar vein, Śaśadhara (NSD, 45.22) notes that the inference suffers from an undercutting defeater (upādhi), namely the property of residing in something which is not an awareness (jñānātiriktavṛttitva). This property would pervade the target property, given that whatever shares the same locus as the property of not having predicative content—such as sattā—can also be found to reside in something which is a not an awareness, like a pot. The proving property obviously cannot be found to reside in something which is not an awareness. Hence, the proving property is undercut—knowing that something resides in an awareness is not sufficient for knowing that it also shares the same locus with that which lacks predicative content, i.e., is non-conceptual.

    The second inference tries to prove that the visual sense faculty is the chief instrumental cause of an awareness which is distinct from a concept-laden visual awareness (cākṣuṣasavikalpakātiriktajñānakaraṇa)—i.e., a non-conceptual visual awareness—because the visual sense faculty is a chief instrumental cause of awareness-states (jñānakaraṇatva), just like the olfactory sense faculty. Śaśadhara mentions an undercutting defeater for this inference, as well: Anything that is the chief instrumental cause of an awareness which is distinct from a concept-laden visual awareness also has the property of being different from imperceptible light, whereas not every instrumental cause of awareness-states has such a property. According to Nyāya-Vaiśeṣika, the visual sense faculty is constituted by the material element of fire or light (tejas); this light illuminates the objects of vision although it itself is imperceptible (see NS 3.1.34, 3.1.42). The other non-visual sense faculties are constituted by material elements which are different from such imperceptible light, and they are the chief instrumental causes of non-visual concept-laden awareness-states. However, the visual sense faculty is not different from imperceptible light, so there is an instance of something which is a chief instrumental cause of awareness-states and which does not produce non-visual concept-laden awareness-states. More to the point, Gaṅgeśa’s prima facie opponent (TCM, 863.03-04) and Vyāsatīrtha (see ND, 485.13) both agree that this sort of inference results in the absurd consequence that, rather than proving the existence of non-conceptual perceptions, it could prove the existence of some unrecognized third type of awareness that is neither non-conceptual nor concept-laden.

  12. 12.

    NSD, 45.03-04: “jāgarādyasavikalpakaṃ janyaviśeṣaṇajñānajanyaṃ janyaviśiṣṭajñānatvāt. daṇḍī puruṣa iti jñānavaditi.”.

  13. 13.

    A separate objection to Śaśadhara’s inference is raised by Gaṅgeśa’s prima facie opponent in Tattvacintāmaṇi. Non-conceptual awareness is not the only possible trigger for activating the relevant memory trace of a past experience with cows. The trigger does not have to be a state of awareness at all; it could be the case that whatever caused one to wake is also the cause of the memory trace’s activation, which then produces the initial concept-laden awareness of the cow. There is no uniformity as to what triggers the activation of memory traces; it could be a certain cognitive state, it could be some unobservable karmic force (adṛṣṭa), or it could be any of the other disparate factors listed in NS 3.2.41, which is likely the “praṇidhānādi-sūtra” directly referenced by Gaṅgeśa (TCM, 862.01) (contra Bhattacharyya 1993, p. 43 and Phillips and Tatacharya 2009, p. 622, who read the phrase as a reference to the Yogasūtra. Also, see TCMP, 862.18-19: “nirantarābhyāsa ityādiśūtraṃ”, and NS 3.2.41, 877.06: “…nibandhābhyāsa….”). Any of these factors may be present, and still the memory trace may fail to be activated; none of them is an invariable cause of activation (TCM, 861.16). The ultimate point is that non-conceptual awareness need not be posited in order to explain how an activated memory trace may produce the initial concept-laden awareness of the cow upon waking.

    One more objection to Śaśadhara’s invocation of memory traces is offered by Gaṅgeśa’s commentator Jayadeva, namely that if the concept-laden awareness of the cow upon waking is to be produced by an activated memory trace, then that awareness would take the form of a recognition, i.e., “This is that cow that I remember seeing before” (TCMA, 817.14-16; cf. Bhattacharyya 1993, pp. 41–42). According to Navya Nyāya, when a memory trace contributes some predicative content—in this case, the property of cowhood—to a current perceptual state, we get a hybrid concept-laden recognition (pratyabhijñā) that is part perceptual and part recollective. Memory states carry an implicit grasp of their causal origins in past experience, such that their content is marked by an additional property of having been experienced in the past—Gaṅgeśa terms this property “thatness” (tattā) (see TCM, 886.01; Phillips and Tatacharya 2009, pp. 671–673). So, the awakened memory trace of previously experienced cowhood would lead to a perceptual recognition of the cow as being identical with “that” cow which was seen before. But, this perceptual recognition is not what has been postulated as the inferential site (pakṣa); that is, the sort of concept-laden awareness that occurs immediately after waking is not supposed to be recognitional (pratyabhijñā). Thus, Śaśadhara’s strategy of citing activated memory traces to prove non-conceptual awareness cannot be salvaged, which is why Gaṅgeśa will offer another inferential proof of his own.

  14. 14.

    TCM, 863.05-06: “prāthamikaṃ gaur iti pratyakṣaṃ jñānaṃ janyaviśeṣaṇajñānajanyaṃ janyaviśiṣṭajñānatvāt anumitivat.” The adjective “generated” (janya) is added as an adjective to the probandum (“janyaviśeṣaṇajñānajanyam”) and the probans (“janyaviśiṣṭajñānatvāt”) in order to make an exception for the states of perceptual awareness belonging to God (īśvara), which are eternal and so could never be causally generated by some prior awareness-states. The rule that qualificative states of awareness are produced by a prior awareness of a qualifier holds just for mortal, non-omniscient beings; I will be referring just to these sorts of mortal awareness-states for the rest of the essay, and will ignore translating “janya” as an adjective in compounds such as “janyaviśeṣaṇajñāna,” “janyaviśiṣṭajñāna,” and so on.

  15. 15.

    Gaṅgeśa's own definition of perception as “jñāna-akaraṇaḳam” (TCM, 595.01) allows that a viśiṣṭa-jñāna can be perceptual even though it is directly preceded by another awareness. That is, any awareness which does not have another awareness as its chief instrumental cause (karaṇa) counts as an instance of perception. A chief cause is understood as bringing about its effect through the operation of some intermediate cause (vyāpāra); in the case of perception, the connection of an object with the sensory organs gives rise to a nirvikalpaka awareness, which in turn produces a savikalpaka awareness. See Phillips and Tatacharya 2009, pp. 334–335 for more discussion.

  16. 16.

    It’s not clear whether having a prior acquaintance with a probandum entails that one also should know that the probandum pervades a certain probans; according to Matilal (1986, p. 347), the idea of sādhya-prasiddhi is just that if one is going to draw proper inferences about some probandum like fire, one ought to have knowledge of what fire is through some acceptable pramāṇa. Hence, the fallacy of sādhya-aprasiddhi is one where the probandum is not “well-known,” as for example when someone tries to infer from the presence of smoke that there is fire which made of gold (kāñcanamayavahni); here, the delimiting property of the probandum, i.e., the property of being made of gold, does not actually exist in the probandum itself, i.e., fire. Once one is aware of this fallacy, one will be blocked from thinking that the smoke on a particular mountain is pervaded by such golden fire, thereby preventing one from drawing the conclusion that golden fire exists on the mountain. But that being the case, sādhya-prasiddhi may consist in the knowledge of what fire is such that it would pervade smoke; the fallacy of sādhya-aprasiddhi would thus be an instance of the general fallacy of not establishing that a probans is pervaded by or concomitant with a probandum (vyāpyatvāsiddhi—see Vattanky 2003, pp. 342–344).

    Mentioning that sādhya-prasiddhi entails the knowledge that a probandum pervades a probans allows us to bypass Vyāsatīrtha’s first objection to Gaṅgeśa (TT, 487.07-488.03), one which was already raised by Gaṅgeśa’s commentators Rucidatta and Jayadeva (see Bhattacharyya 1993, pp. 18–21). Rucidatta points out that, strictly speaking, one’s antecedent understanding of a probandum (sādhyaprasiddhi) is not the actual cause (kāraṇa/hetu) of inferential knowledge (anumiti)—the mere awareness of fire as such (vahnijñāna) does not directly cause one to infer that fire is on the mountain (“parvato vahnimān”). Rather, the actual cause of inferential knowledge is one’s knowledge of a relevant concomitance rule (vyāptijñāna), e.g., that smoke is pervaded by fire. As he puts it, when an awareness of a vyāpti is present, the inferential conclusion is not delayed in arising by a delay in the awareness of the probandum as such (TCMP, 858.22-24). Jayadeva further explains that if sādhya-prasiddhi were an awareness which preceded the awareness of concomitance, then it would be twice removed from the inferential knowledge, since it is a cause of the vyāpti-jñāna which is itself the direct cause of the anumiti. Just as a potter’s father is considered to be causally irrelevant (anyathāsiddha) for the production of a pot, the prior awareness of the probandum, being a cause of a cause (hetuhetu) would be also be causally irrelevant, and hence would not serve as evidence for the causal rule that qualificative awareness-states must be produced by a prior awareness of a qualifier. (Vyāsatīrtha raises this same objection at TT, 488.01-03.) Jayadeva’s solution is to identify sādhya-prasiddhi with the vyāpti-jñāna, making it the cause of inferential knowledge and presumably restoring its status as the awareness of a qualifier—e.g., the awareness of fire as a pervader of smoke—that brings about a qualificative state of inferential knowledge (TCMA, 811.15-812.13). Vyāsatīrtha would seem to reject Jayadeva’s solution; see discussion of TT, 494.07-10 in Sect. 4 below. In short, the problem is that a vyāpti-jnāna doesn’t cause inferential knowledge in the way that an awareness of a qualifier generally causes qualificative awareness, since the content of the vyāpti-jñāna doesn’t become the predicative content of the inferential knowledge-state. So, the case of inference would not support Gaṅgeśa’s general causal rule.

  17. 17.

    TCM, 860.04-06; TT, 487.02-05.

  18. 18.

    Perception is thus spoken of by Naiyāyikas as being the “pre-eminent” source of knowledge (jyeṣṭhapramāṇa), and as being that upon which inference depends for its epistemic livelihood (pratyakṣopajīvakatva; see Phillips 1997, pp. 129, 356 n. 34).

  19. 19.

    e.g., NS 3.1.21, 745.

  20. 20.

    See Sinha 2014, pp. 156–161 for a discussion of Nyāya-Vaiśeṣika accounts of autonomic willing (jīvanayoniyatna) which are produced by jīvana-adṛṣṭa.

  21. 21.

    TT, 485.11-486.02; ND, 486.16-17.

  22. 22.

    TT, 486.03. Gaṅgeśa and Vyāsatīrtha mention one more negative consequence of accepting that sense-object contact (indriyārthasannikarṣa) can cause a memory instead of a perception, and thus of admitting that the set of causal conditions (sāmagrī) of memory is stronger than that of perception. One’s knowledge of the Vedas is derived from reliable testimony, and reliable testimony must ultimately originate from someone who has directly experienced the object of testimony. Thus, the ancient sages to whom the Vedas were originally revealed must have a direct experience (anubhava) of the Vedic sounds and their meanings. However, if the causal conditions for memory (smṛtisāmagrī) are stronger than that of direct experience (anubhavasāmagrī), then even the sages would only have a memory of the Vedas, rather than directly experiencing them (TCM, 863.18: “… nityānāṃ vaidikārthānāṃ ca smaraṇaṃ syāt na tv anubhavaḥ.”). But, the sages are supposed to be “seers” of the Vedas, to whom the Vedic sounds and meanings are directly revealed. Hence, this inability to accommodate the direct experience of Vedic sages serves as reductio of the claim that, in having a stronger set of causal conditions, memory can arise instead of a direct perceptual awareness, and specifically that memory could arise from sense-object contact instead of a non-conceptual perception in order to generate one’s first concept-laden awareness of a cow.

    There are two curious points to note with this argument. First, Naiyāyikas actually reject the view that Vedic sounds/syllables are eternal, or that they are eternally linked with their meanings. Instead, the Vedic mantras were composed by God (īśvara), and got their meanings/referents through the linguistic conventions he created. We might then suspect that Gaṅgeśa is drawing out an undesirable consequence for the Prābhākara Mīmāṃsaka, who would take the Vedas to be eternal, and who is the ostensible target of Gaṅgeśa’s arguments in his Nirvikalpakavāda (Phillips and Tatacharya 2009, p. 611). Second, it is peculiar that Gaṅgeśa only mentions this consequence concerning the Vedas, as it overlooks the more obvious consequence that we would continually recollect objects experienced in past lives whenever our sense faculites made contact with them in this life. This undesirable consequence is supplied by Vyāsatīrtha (TT, 486.07-08: “anyathā janmāntarānubhūtānāṃ etajjanmani indriyasannikṛṣṭānām….”); it is unclear whether his doing so suggests a lacuna in received TCM manuscripts, or is instead a charitable addition to Gaṅgeśa’s argument.

  23. 23.

    TT, 486.09-12. Gaṅgeśa raises different forms of the regress objection in both the pūrvapakṣa and siddhānta portions of his Nirvikalpakavāda. In the pūrvapakṣa section, a prima facie defender and a denier of non-conceptual perception both accuse each other’s positions of leading to an infinite regress. The defender invokes the version of the regress stated above, which would result if a qualificative awareness were caused by a prior viśeṣaṇa-jñāna that was itself qualificative (TCM, 861.11-12). The denier of non-conceptual perception responds by arguing that the defender can’t avoid inferring that all awareness is produced by a prior viśeṣaṇa-jñāna, in which case a regress follows because even non-conceptual awareness would have to be produced by a prior viśeṣaṇa-jñāna (see Phillips and Tatacharya 2009, pp. 621–622). The first mention of the regress objection in Gaṅgeśa’ own voice comes at TCM, 864.01-03: “na ca gotvajñānaṃ gotvajñānajanyaṃ janyajñānajanyaṃ janyaviśeṣaṇajñānajanyaṃ cety anyatra darśanāt prathamaṃ na gotvānubhava iti vācyam. ekasya eva janyajanakatvānavacchedakatvāt anavasthāpātāc ca.” “Nor can the following objection be stated: ‘Because it is observed elsewhere that the awareness of cowhood is generated by another awareness of a cowhood, which is generated by a prior awareness that is itself generated by a prior awareness of a qualifier, the first awareness [in the sequence that leads to the first-time qualificative perception of a cow] is not a non-mnemonic experience of cowhood (gotvānubhava).’ That is because the same thing cannot be the delimiting property of something’s being a cause and something’s being an effect, and because there is an infinite regress.” The responses of Vyāsatīrtha and Gaṅgeśa’s commentator Rucidatta to the regress objection are discussed in Sect. 6.

  24. 24.

    Indeed, even in the qualificative perception of a cow as a cow, the direct non-conceptual acquaintance with cowhood remains; the property of cowhood is not itself qualified or delimited by any other property. See Phillips and Tatacharya 2009, pp. 631–632.

  25. 25.

    TT, 489.07-490.05.

  26. 26.

    Matilal (1986, p. 351) misleadingly asserts that only the qualifying property is presented in a non-conceptual awareness, and that the qualificand is presented only in a subsequent concept-laden perception. Instead, the Navya Nyāya view is that the qualifier and qualificand must both be presented in a non-conceptual awareness, if not also the relation that links them (for discussion, see Bhattacharyya 1990b, pp. 172–176; see also the arguments in Mahādeva’s NyāyakaustubhaNKau, 195.08-15). Though, Matilal is of course not without a point here. Even though the qualificand should be presented in a non-conceptual perception, later Naiyāyikas would insist that the non-conceptual awareness of the qualificand does not play the same sort of causal role that the awareness of the qualifier plays in generating a subsequent qualificative awareness. The reasons for attributing the viśeṣaṇa-jñāna with an exclusive causal role are explained below. Perhaps it is to more sharply distinguish this exclusive causal role that later Naiyāyikas deny the causal necessity of a viśeṣya-jñāna, and instead attribute a causal role merely to the connection between one’s senses and a qualificand (viśeṣyendriyasannikarṣa). This sensory connection is sufficient for presenting in order for the qualificand to appear in a qualificative perception; see NKau, 199.08-12.

  27. 27.

    TCMA, 813.17-814.15.

  28. 28.

    It is somewhat curious why Gaṅgeśa does not straightforwardly explain the necessary causal role of a viśeṣaṇa-jñāna in producing a viśiṣṭa-jñāna as being due to a qualificative relation’s (vaiśiṣṭya) structure being determined by its qualifying relatum (pratiyoginirūpyatva), instead opting to inductively establish the general causal relation between viśeṣaṇa-jñāna and viśiṣṭa-jñāna on the basis of specific instances of this relation (e.g., inference, verbal understanding, perceptual illusion, etc.). Jayadeva suggests that Gaṅgeśa settles for generalizing his rule from particular instances because they are easier to focus on (samādhisaukaryāt) (TCMA, 815.18-816.12).

  29. 29.

    TCM, 441.01-02.

  30. 30.

    TCM, 434.01-437.02.

  31. 31.

    Leaving aside their total rejection of the existence of universals, Dvaita Vedāntins offer a “fresh” (abhinava) version of Nyāya’s “misattribution” theory of perceptual error (anyathākhyāti), one which denies the Nyāya belief that real snakehood is included as the content of the rope-snake illusion. In seeing a rope as a snake, Dvaitins deny that we are directly aware of real snakehood, present elsewhere in actual snakes but absent in the rope before me, through an extra-ordinary sensory relation (alaukikasannikarṣa). Rather, the snakehood which is erroneously superimposed (āropita) onto the rope is simply non-existent; the illusion just consists in the presentation of something non-existent as existing in the object which is actually being seen. The veridical perception of a real snake may indeed be responsible for my memory trace of snakehood, but whether or not the snakehood superimposed by memory onto the rope actually exists elsewhere is irrelevant to the content of the illusion. In short, it doesn’t make sense to the Dvaitin that a non-existent entity could enter into a sensory relation with the mind. See Williams (2017) and Sharma (1962, pp. 131–136) for more discussion.

  32. 32.

    Vyāsatīrtha points out how the role of an operative sensory connection (sannikarṣa) in presenting the mind with real objects also vitiates the support that perceptual illusion is supposed to lend to Gaṅgeśa’s general causal rule that all qualificative awareness is produced by an awareness of a qualifier. On the Nyāya account, the qualificative, albeit illusory, awareness of a rope as a snake is caused by the mistriggered memory of snakehood; the memory serves as an awareness of the qualifying property (viśeṣaṇajñāna) that leads to the erroneous attribution of snakehood to a rope in the subsequent illusory awareness. As Vyāsatīrtha notes, however, Naiyāyikas believe that the memory-awareness itself serves as an extra-ordinary sensory connection (alaukikasannikarṣa) with the qualifying property, such that through the memory of snakehood, one is directly aware of that real property in the world even as it is being falsely ascribed to the seen rope. If that is the case, then the memory-awareness is causing the subsequent perceptual illusion not by virtue of being an awareness of a qualifier (viśeṣaṇajñānatva), but by virtue of its serving as a sensory connection with that qualifier (viśeṣaṇasannikarṣatva). Further, it is more parsimonious for the causal role of memory to be delimited by its being a sensory connection with a qualifier, and not by its being an awareness (jñāna) of a qualifier, given that Naiyāyikas also admit that non-cognitive states like desire (icchā) and volition (kṛti) can also produce a qualificative perceptual awareness through serving as the extraordinary sensory connection with its contents (see JLVR, 60ff and Mohanty 2011, pp. 401–404). In sum, if viśeṣaṇa-sannikarṣatva ought to be the property which delimits memory’s causal role in producing perceptual illusions, then perceptual illusion would no longer provide evidence for Gaṅgeśa’s rule that a prior state of awareness produces a qualificative awareness (viśiṣṭajñāna) just by virtue of being an awareness of the relevant qualifier (i.e., by virtue of its viśeṣaṇajñānatva); see TT, 493.04-07.

  33. 33.

    NB 1.1.4, 94.02; see also VS 3.1.16, 111.

  34. 34.

    See PDS, 182-3; NK, 237-238.

  35. 35.

    TT, 491.13-492.04.

  36. 36.

    There are two other counter-examples mentioned by Vyāsatīrtha (TT, 491.02-05). The first is the perceptual awareness of two things as related by conjunction (saṃyoga) or inherence (samavāya), as expressible in the form, “These two are conjoined,” or, “These two are inherent” (“saṃyuktāv imau”; “samavetāv imau”). In these cases, the conjunction and inherence relations are presented as qualifying the two objects which are related. If a prior awareness of the qualifier is a necessary cause of any qualificative awareness, then there must be a prior awareness of the qualifying saṃyoga or samavāya relations. However, this prior awareness of the relation itself could not be non-conceptual, because Navya Naiyāyikas generally believed that only the qualifier and qualificand appear in non-conceptual perception (see TSBP, 25.07-14; though, cf. TSBP, 137.01). According to them, an awareness of a relation (sambandha) requires that one first is aware of its relata (sambandhin). So, the awareness which first presents these relata—i.e., the qualifying relatum and qualified relatum—must be a non-conceptual, pre-predicative perception in which the relational tie itself is not yet presented (vaiśiṣṭyānavagāhin). But, if it is impossible for a relation to be presented in non-conceptual awareness, then the qualificative awareness of two objects as being qualified by a relation couldn’t be produced by a prior non-conceptual awareness of that qualifying relation. Instead, this qualificative awareness arises without its qualifier being previously known in a separate awareness. That would contradict Gaṅgeśa’s rule that all qualificative awareness is produced by the prior of awareness of the relevant qualifier.

    The second counter-example is an introspective awareness (anuvyavasāya) of oneself as possessing a predicative awareness (vyavasāya) of some object. Say that you have a qualificative perception expressible in the form, “That is silver.” The perceived object is identified as silver because it is seen as being qualified by the property of silverhood (rajatatva). The subsequent introspective awareness of this first-order perception identifies one’s self as possessing a qualificative awareness in which silverhood is presented as a qualifying predicate (prakāra). More specifically, the self is qualified by the first-order perception of silver, and that first-order perception is further qualified or identified by its having silverhood as its predicative content (rajatatvaprakārakatva). This introspective awareness would be expressible in the form, “I have an awareness of silver as qualified by silverhood” (“rajatatvaprakārakarajataviśeṣyakajñānavān aham”). (If that sort of introspective report seems far-fetched, we should bear in mind with Viśvabandhu Tarkatīrtha (Shaw 1996, p. 236) that while the content of introspective awareness is propositionally structured and hence linguistically expressible, it is not to be conflated with its linguistic form, nor does language play a necessary role in generating the awareness. The perceptual nature of an anuvyavasāya implies that it is to be distinguished from the verbal report that may follow from it.)

    The problem is that, prior to the introspection of the first-order perception, there is no awareness of the qualifying property, “having silverhood as its predicative content.” Sure, the first-order perception possesses silverhood as its predicative content; but, while it is aware of silverhood, the perception is not reflexively aware of its own property of having silverhood as its predicative content. That property is only cognized when the introspective awareness identifies the first-order perception; there is no awareness of it prior to the introspection. Consequently, we have another case where a qualificative awareness—i.e., the anuvyavasāya—arises without depending upon a prior awareness of the relevant qualifier.

  37. 37.

    TT, 491.01-07: “api ca tvayā api ghaṭābhāvavad bhūtalam iti jñāne abhāvarūpaviśeṣaṇasya… pūrvam ajñātasya eva bhānam iti svīkṛtatvāttatra vyabhicāraḥ.”

  38. 38.

    To complete Gaṅgeśa’s causal story of absence-perception, the prior perception of the pot-absence as qualified by the floor can be preceded by the non-conceptual perception of the floor. Something else to note: In the qualificative perception of the absence-as-qualificand, there is also the extra-ordinary sensory connection with the absent pot established by a prior memory of a pot.

    Yet, there are still some basic flaws in this story. Specifically, it is not obvious how the pot-absence is cognitively transferred from being a qualificand in the first awareness to a qualifier in the second awareness. In their translation and comments, Phillips and Tatacharya (2009, pp. 637–638) attribute Gaṅgeśa with the view that a quick, “one-step” inference bridges the gap: After perceiving the absence as qualified by the floor, one then inferentially concludes that the floor is qualified by the absence. Phillips and Tatacharya write: “Admittedly, there occur absential cognitions to be analyzed as having the floor not as qualifier but as qualificandum, with an absence of a pot as the qualifier. But these are not perceptual. They are the results of one-step inferences from perceptual cognitions, such as from a perception where the qualifier is the floor and the qualificandum the absence of a pot” (2009, p. 638).

    This interpretation is highly peculiar, though. As far as I can tell, it has no grounding in Gaṅgeśa’s text or its commentaries. The perceptual awareness of a floor as qualified by the absence of a pot (“ghaṭābhāvavad bhūtalam”) is a canonical example of absence-perception (TS, 137.18-19). Gaṅgeśa would take this awareness to be generated through the specific sensory relation of indriya-sambaddha-viśeṣaṇatā (TCM, 613.05; later Naiyāyikas would more precisely call it saṃyukta-viśeṣaṇatā), whereby the sense faculty is connected with the pot-absence through that absence’s being the qualifier of the floor which is in physical contact with the sense faculty. This type of sensory relation was posited by Naiyāyikas as early as Uddyotakara, in part to account for the perceptibility of absences.

    That Gaṅgeśa thinks the awareness of the floor as qualified by a pot-absence is perceptual should be clear from the context of his discussion in the Nirvikalpakavāda. Indeed, the whole motivation for Gaṅgeśa’s story about absence-perception is that he is trying to square his acceptance of the fact that an absence can be perceived as a qualifier of the floor with his insistence that all qualificative perceptions must be generated by a prior awareness of a qualifier. Since an absence can’t be perceived non-conceptually, Gaṅgeśa thus posits that it is first perceived as a qualificand. But, the subsequent awareness of the absence-as-qualifier ought to be no less perceptual.

    So, if it is possible for the awareness of the absence-as-qualifier to be perceptual, then it is mysterious as to why, even when the relevant sensory connections are present, an inference would be needed to generate such an awareness. I am not aware of any Nyāya thinker who believes that inferences can produce states of perceptual awareness. Also, it’s not clear why an inference is needed to transition from the awareness of absence-as-qualificand to the awareness of absence-as-qualifier, when no inferential process is invoked to transition from a non-conceptual awareness of a qualifier-not-as-qualifier to the qualificative awareness of that qualifier-as-qualifier. Additionally, there is no obvious way of understanding how such a “one-step” inference is supposed to be structured—it wouldn’t seem to follow any standard form of anumāna.

    Part of the problem may be that Phillips and Tatacharya slip up in interpreting Gaṅgeśa’s answer to the following objection: If an absence is perceived through the sensory relation of indriya-sambaddha-viśeṣaṇatā, that is, if it is perceived as the qualifier of an object which is connected with a sense faculty, then how is that absence first supposed to be perceived as a qualificand (viśeṣya)? Gaṅgeśa replies by claiming that the absence is grasped as a qualificand through the presentation of the floor as its qualifier. Being-a-qualifier (viśeṣaṇatā) is a self-linking relation (svabhāvapratyāsatti, or svarūpasambandha) between the qualifier and its qualificand; the qualifier’s relation to a qualificand is not ontologically distinct from its being a qualifier. Hence, through apprehending the floor as a qualifier, one will also apprehend the absence-as-qualificand. Since the absence is not being presented as a qualifier, the absence would not need to be perceptually presented through the sensory relation of viśeṣaṇatā. (See TCM, 867.05-07. Later Naiyāyikas will just say that the absence is perceived through a separate sensory relation of saṃyukta-viśeṣyatā, which presents the absence as a qualificand of what is in contact with a sense faculy.)

    However, in their translation, Phillips and Tatacharya evidently jump to the conclusion that the absence does not need to be perceptually presented at all: “[The absence] is grasped through the qualifierhood relation to the floor, that is to say, through a sensory connection that is self-linking—this is what we mean (by our talk of qualifierhood as the sensory connection in perception of absences). We do not mean that an absence appears (perceptually) as a qualifier” (2009, p. 637). It is true that an absence does not perceptually appear as a qualifier in the first awareness, “There is no pot on the floor.” But, Gaṅgeśa certainly thinks that the absence perceptually appears as a qualifier in the subsequent awareness, “The floor lacks a pot.” His statement that the absence does not appear as a qualifier is just referring to the first awareness of the absence-as-qualificand, which he also clearly thinks is perceptual. Gaṅgeśa even admits that one can visually perceive the absence of color as qualifying air, which is itself invisible (Ibid., 637–638). If this awareness of invisible air can count as a visual perception of an absence as a qualifier, then surely the awareness of a pot-absence as qualifying a floor can be perceptual, rather than inferential.

  39. 39.

    ND, 491.21-492.12.

  40. 40.

    TT, 491.05-07.

  41. 41.

    TT, 494.01-03.

  42. 42.

    TT, 494.05-08: “kiṃ ca phalībhūtajñāne yadviśeṣaṇatvena bhāsate tadviṣayakajñānena janyaṃ yajjñānaṃ tattatprakārakamiti vā niyamaḥ…? nādyaḥ, ātmāśrayāt”; see also ND, 494.20-22.

  43. 43.

    TT, 494.07-10: “… yajjñānajanyaṃ yajjñānaṃ tattatprakārakamiti vā? … nāntyaḥ, anumityādau vyāptyādeḥ prakāratvāpātāt. tasmāttvayāpi prakāratvaniyamo bhojakādṛṣṭādibhireva vācyaḥ.”

  44. 44.

    The same objection just raised concerning absence-perception was addressed by Gaṅgeśa in the context of inferential knowledge (TCM, 860.03-06; Phillips and Tatacharya 2009, pp. 617–618). His opponent claims that it is not merely in virtue of its being an awareness of a qualifier (viśeṣaṇajñānatvena) that the prior awareness of fire serves as a cause for the inferential conclusion that the mountain is on fire. That is because if one were aware of the qualifying fire as being a substance, rather than as being fire as such, then one would instead infer that the mountain is merely substance-possessing. Thus, the prior awareness of the qualifier must be a viśeṣaṇatā-avacchedaka-prakāraka-jñāna, that is, an awareness whose predicative content is that property which will serve as a delimiting mode of presentation for the qualifier in the inferential conclusion. In other words, when the prior awareness predicates fire with the property of firehood (vahnitva), and not substancehood (dravyatva), then the mountain can be predicated by that fire qua fire, not qua substance; the property of firehood will thereby delimit the fire’s being a qualifier of the mountain. In response, Gaṅgeśa insists that the more general causal rule still holds even in the case of inference—the prior awareness of fire-as-fire is still just a specific kind of awareness of a qualifier (viśeṣaṇajñāna), so we can still generalize from the example of inferential knowledge to claim that all qualificative awareness is similarly produced by a prior awareness of a qualifier.

  45. 45.

    TT, 496.07-497.03.

  46. 46.

    TT, 497.03-05. Vyāsatīrtha is here drawing on a similar move made by Rucidatta, whose own arguments will be discussed below; see TCMP, 865.17-18. Rucidatta goes further than Vyāsatīrtha in giving an explanation for the loss of parsimony that results from restricting the awareness of a qualifier to be the cause only of qualificative states of experiential/non-mnemonic awareness (i.e., viśiṣṭānubhavatva). Absent this restriction, the delimiting property of something’s being an effect of a viśeṣaṇa-jñāna—i.e., the janyatā-avacchedaka or kāryatā-avacchedaka—would just be the property of being a non-mnemonic experience, or anubhavatva. Gaṅgeśa would accept that the property of anubhavatva is a natural kind (jāti—see Phillips 2011, pp. 26–28 for discussion), whereas the property of viśiṣṭānubhavatva is presumably not. As a result, the unqualified causal rule would be more parsimonious due to having more explanatory power (balavattva) through its tracking the natural-kind property of anubhavatva. For his part, Vyāsatīrtha wouldn’t make such an appeal, since Dvaita Vedāntins are nominalists who don’t take natural kinds to be real (see Sharma 1962, pp. 69–71). But, given that Gaṅgeśa does take anubhavatva as a natural kind, it seems that he wouldn’t be able to resist Rucidatta’s parsimonious rejection of the restricted causal rule, in which case he couldn’t resist the undesirable consequence that non-conceptual perception, being a state of experiential awareness (anubhava), would also be produced by an awareness of a qualifier.

  47. 47.

    TT, 497.06-07.

  48. 48.

    TT, 497.07-11: “na caivam anavasthā. kvacid viśiṣṭajñānasya smaraṇarūpatayā tatra viśeṣaṇajñānānapekṣaṇāt. janyaviśiṣṭānubhavaṃ praty eva hi janyaviśeṣaṇajñānaṃ hetuḥ. na tu smṛtiṃ praty api ity uktam.” “Nor does a vicious regress follow [if non-conceptual awareness does not exist and all awareness is instead qualificative], because at some point in the causal series, there is a qualificative awareness which, by virtue of its being a memory, does not depend on a prior awareness of a qualifier. For an awareness of a qualifier is the cause only of qualificative states of non-mnemonic experience (viśiṣṭānubhava)—it is not also the cause of states of memory, as was said earlier [i.e., in the section starting with TT, 496.05-07, according to ND, 497.19-20].” Vyāsatīrtha is again following the suggestion of Rucidatta (see footnote 46), who also states that the awareness of a qualifier serves as a cause only for non-mnemonic experiential states, not for memory, in which case the cause of a first-time qualificative awareness can be a memory, instead of a non-conceptual perception, without triggering a vicious regress; see TCMP, 865.18-20.

  49. 49.

    ND, 497.16-17.

  50. 50.

    See, e.g., NS 3.1.18-24 and commentaries thereon. TCMśa, 159.02-04: “ata eva stanapānapravṛttāvapy upāyecchākāraṇatvena gṛhītasyeṣṭasādhanatvajñānasyāpi kalpanaṃ dṛṣṭānurodhitvāt kalpanāyāḥ.” “Therefore, in the case of intentional activity such as the suckling of a mother’s breast, there is the postulation of even an awareness of [the breast’s] being a means for achieving what is desired—this awareness being understood as the cause of a desire for attaining the means—since this postulation accords with what is observed.”

  51. 51.

    TCM, 881.02-03.

  52. 52.

    NVTṬ 1.1.4, 116.14-16.

  53. 53.

    TT, 498.01-08; compare with TCMP, 864.11-16. Rucidatta first addresses Gaṅgeśa’s charge (TCM, 864.01-03; see footnote 23 above) that if the (qualificative) awareness of cowhood were caused by just another (mnemonic) awareness of cowhood, then the exact same property of being an awareness (jñānatva) would delimit both the effect and the cause, which would presumably make the cause and effect indistinguishable from each other. In response, Rucidatta specifies that the kāryatā- or janyatā-avacchedaka should be the property of being an awareness, whereas the kāraṇatā- or janakatā-avacchedaka should be the property of being an awareness of a qualifier (viśeṣaṇajñānatva)—now, the two delimiting properties are distinct. Gaṅgeśa also can’t claim that the inclusion of the term “viśeṣaṇa” in the kāraṇatā-avacchedaka would not be parsimonious, since his own proposed rule would meet the same objection, as it too delimits the cause of qualificative awarenesses in terms of its viśeṣaṇa-jñānatva, and would render cause and effect indistinguishable from each other if viśeṣaṇa-jñānatva weren’t the kāraṇatā-avacchedaka. Finally, Rucidatta says that it doesn’t matter whether or not the same property delimited both kāryatā and kāraṇatā—either way, a vicious regress would not follow. Just as seeds can ultimately produce more seeds in an indefinite causal series, in which case the property of seedhood (bījatva) would delimit both cause and effect, let one awareness generate another in a similarly indefinite causal series, with awarenesshood (jñānatva) delimiting both cause and effect. Or, if one thinks that cause and effect can’t be delimited by the same property, then delimit the seed’s or awareness’s causehood in terms of its existing prior to its effect (i.e., its pūrva-bījatva or pūrva-jñānatva). Regardless, if these causes and effects still come into existence in dependence on their own causes, then any regress that follows would be benign rather than vicious.

  54. 54.

    TT, 498.09-15; TCMP, 864.15-18. Note that Phillips and Tatacharya (2009, p. 632) are wrong to imply that Rucidatta himself accepted the objection that a regress of states from past lives would make death impossible.

  55. 55.

    Rucidatta goes on to suggest that his response to the objection from the impossibility of death has the further result of rendering the inclusion of the term “qualificative” in Gaṅgeśa’s inference meaningless (TCMP, 864.18-20). Recall the form of Gaṅgeśa’s inference: The first-time perception of a cow as a cow must be generated by a prior awareness of a qualifier, because that perception is a qualificative awareness. Additionally, that prior awareness of a qualifier has to be non-qualificative/non-conceptual, to avoid an infinite regress. One initial objection which can be raised against this inference is that if non-qualificative awareness is not yet known to exist prior to the statement of this inference, then it is pointless to include the term “qualificative” (“viśiṣṭa”) in the probans “qualificative awareness,” since we don’t yet know what is meaningfully distinguished by that term. (On the other hand, if non-qualificative awareness is already known to exist prior to the statement of the inference, then the inference itself is pointless.) However, if the term is dropped so that the probans just becomes “because [the first-time perception] is an awareness” (“jñānatvāt”), then even the awareness of a qualifier (viśeṣaṇa-jñāna) would also have to be generated by an awareness of a qualifier, leading to an apparently vicious regress. Plus, if non-qualificative awareness-states did exist, as the inference was originally intended to establish, then they present a counter-example to the concomitance rule of the inference as now stated, since they are states of awareness which aren’t supposed to be produced by a prior viśeṣaṇa-jñāna (TCM, 862.01-04). Gaṅgeśa responds to these objections by arguing that the term “qualificative” is not pointless. Without it, an invariable concomitance rule couldn’t be ascertained between the properties of being an awareness and being produced by an awareness of a qualifier, because such a concomitance would be defeated by the infinite regress and the counter-example of non-conceptual perception. The term “qualificative” needs to be included in the probans (“being a qualificative awareness”) precisely to prevent the concomitance rule from being defeated (TCM, 865.09-13). However, Rucidatta denies Gaṅgeśa’s line of reasoning here, claiming that the term “qualificative” is still pointless and should be dropped, since, in having warded off the impossibility of death, there would be no vicious regress and hence no counter-example of non-conceptual perception which needs to be warded off.

  56. 56.

    The above objection and response allude to an inferential proof of non-conceptual perception defended by Śaśadhara and abandoned by Gaṅgeśa. The inference claims that a perception of some object as being qualified by a newly arisen sensory property like color must be generated by a prior non-conceptual perception, because that perception of the object is a qualificative perception. Since this recently arisen color has never been experienced before, the qualificative perception could not have gotten its predicative content from a prior memory of the color, so it must be getting its content from a non-conceptual perception instead (NSD, 46.13-14; TCM, 860.13-14). Gaṅgeśa’s prima facie opponent rejects this line of argument (TCM, 861.01-06). For example, the qualificative perception of a pot would classify it as possessing a certain type of color, e.g., a shade of white. So, the perception would be doubly qualificative (viṣiṣṭavaiśiṣṭyabodha), in that it presents the pot which is qualified by a white color, and that white color as qualified by the property of being white (śuklatva). Now, a doubly qualificative awareness must also be preceded by an awareness of a qualifier, which in this case would be a qualificative perception of the newly arisen color which is being classified as a shade of white. This qualificative perception of the new color as being white can in turn get its predicative content from a prior memory of whiteness—not a non-conceptual perception of whiteness—since one will have experienced whiteness in one’s prior experience with other shades of white. Alternatively, the memory of whiteness can also have as its content the newly arisen color itself, given that Naiyāyikas allow the whiteness-universal to serve the function of establishing an extraordinary cognitive relation (sāmānyalakṣaṇasannikarṣa) with all past, present, and future instances of whiteness (see Sinha 1934, pp. 79–81).

    Śaśadhara rejects the appeal to such an extraordinary cognitive relation (NSD, 46.14-17). Through knowing the whiteness-universal, one cognizes all instances of whiteness only in a general way. But, the current qualificative perception demonstratively presents the pot as being qualified by “this” newly arisen color—i.e., “The pot possesses this color” (“etadrūpavān ghaṭaḥ”), not “The pot possesses a color” (“rūpavān ghaṭaḥ”). Because the prior memory does not demonstratively cognize the new color as being currently close at hand, it cannot be what supplies the predicative content to the current qualificative perception of the color, since it’s a rule that if a prior awareness of a qualifier is itself qualificative, then it must present that qualifying content in the same way as it is presented in the subsequently produced qualificative awareness (or else there would be the absurd possibility that a totally unrelated awareness could supply the same qualifying content to the subsequent awareness). Non-conceptual perception is exempt from this rule, as it does not present its contents as being a certain way whatsoever.

    All that said, Śaśadhara overlooks the response mentioned by Gaṅgeśa and hinted at by Vyāsatīrtha, namely that the memory of whiteness would first contribute to the qualificative perception of the new white color, before producing the doubly qualificative perception of the pot as possessing that white color. The intermediate stage of perceiving the new color as white (“This color is white”) can be the place where the demonstrative content “this white color” enters, being supplied directly by the contact of the visual faculty with the white color itself, and being presented as a delimiting property of the qualificand (i.e., “thisness”/“idantā” would be the viśeṣyatā-avacchedaka; see Ganeri 2011, p. 140). Then, this white color can be presented in the doubly qualificative perception as belonging to the pot.

  57. 57.

    TT, 499.07-08; TCMP, 864.22-25.

  58. 58.

    Rāghavendratīrtha (ND, 499.14-15) indicates that since Rucidatta’s arguments originate from some undesirable assumptions (arucibīja), Vyāsatīrtha’s own view, i.e., that the regress of viśeṣaṇa-jñānas is broken by memory, is what should be understood in the last analysis.

  59. 59.

    We may note that the later Naiyāyikas Gadādhara and Mahādeva both apparently abandon Gaṅgeśa’s causal rule in light of the counter-example posed by memory, which is again a form of qualificative awareness that is not produced by a prior awareness of a qualifier. The kāryatā-avacchedaka in the rule can’t be either viśiṣṭa-jñānatva or viśeṣaṇa-jñānatva, since memory is both qualificative and can serve as an awareness of a qualifier without itself having an awareness of a qualifier as its cause. One way of defusing memory as a counter-example would be to let the cause of a qualificative awareness just be whatever has a relevant qualifier as its content, whether or not it is an awareness-state itself. (The kāraṇatā-avacchedaka is now viśeṣaṇa-viṣayakatva.) Memory-traces (saṃskāra) are not states of awareness, but they have contents and are responsible for generating qualificative memory-states. However, this move won’t work either, since a memory-trace which remains unactivated will have a relevant qualifier as its content even though it does not generate a memory-state. The counter-example of memory can’t be removed by restricting the effects in question to either qualificative forms of non-mnemonic experience (viśiṣṭānubhava) or non-mnemonic states in general (smṛtyanyajñānatva, i.e., anubhavatva), because the rule would become unparsimonious given that the property of being a non-mnemonic experience is not actually a natural kind (cf. footnote 46). The property of anubhavatva suffers from the defect of cross-cutting (sāṅkarya): There are types of non-mnemonic experience which are indirect (parokṣa), but mnemonic states are also indirect. Finally, the counter-example of memory could be removed without altering the causal rule, if we admitted the existence of non-conceptual states of memory (nirvikalpakasmaraṇa). Just as qualificative perceptions are generated by a prior non-conceptual perception of a relevant qualifier, so too would qualificative memories be generated by a prior non-conceptual memory of a relevant qualifier. But, neither Gadādhara nor Mahādeva find the notion of non-conceptual memory plausible. In the end, the counter-example of memory can only be removed if the causal rule is restricted so that only qualificative states of perception are caused by a prior awareness of a qualifier, thus abandoning Gaṅgeśa’s goal of establishing a causal uniformity across all states of qualificative awareness in general. See VJHV, 76.30-77.03, 77.08-09; see also NKau, 200.22-201.06, 207.04-08.

  60. 60.

    TT, 499.10-12.

  61. 61.

    Tatacharya (TSBP, 33.07-16) notes that there is a shift in how “old” and “new” Naiyāyikas understand the compound “yogya-anupalabdhi.” The “old” reading interprets the compound as a genitive tatpuruṣa (yogyasya anupalabdhiḥ), i.e., “the non-awareness of what is yogya.” What is yogya is some object which is fit to be perceived by the senses; in other words, “yogya” means “perceptible,” and is synonymous with “dṛśya.” The non-observation of what is otherwise perceptible can serve as evidence that the perceptible object is not present, since if it were present, then it would’ve been perceived. The “new” reading takes “yogya” to be an adjective of “anupalabdhi”; now, it is the non-awareness itself which is “fit.” An absence of awareness is fit for proving the absence of an object when the absent awareness would have been present were the object in question also present. Of course, the absent object still has to be perceptible in order for its absence to be known from a “fit” non-observation of it. See NSM 62, 262.06-265.01.

  62. 62.

    As Udayana explains (NKus 3.20, 427.05-428.06), a yogya-anupalabdhi can contribute to the knowledge of Maitra’s absence in one of two ways, depending on whether or not I am explicitly aware of having that anupalabdhi. Perceptual knowledge of Maitra’s absence doesn’t rely on my being explicitly aware that I lack the sort of observation which hypothetically would’ve occurred if Maitra were present. Though being unknown, this non-observation can serve as an auxiliary causal condition which assists my sense faculties in directly perceiving Maitra’s absence. This is in keeping with the general rule that, when the causal conditions which give rise to an awareness can operate without a subject being aware of them—as is the case with the sense faculties—then the awareness which arises will be directly perceptual in nature (sākṣātkārin). When the causal conditions for knowledge-states must themselves be known to a subject in order for them to generate the knowledge—as is the case with inference, where the subject must be aware of each step of the inference in order to arrive at its conclusion—then the resulting knowledge is not directly perceptual in nature. Distinguishing between the different conditions under which yogya-anupalabdhi can contribute to the knowledge of absence is relevant for appreciating why Vyāsatīrtha’s objection specifically targets the inferential knowledge of absence—see ND, 499.18-500.12.

  63. 63.

    TCM, 723.10; see Phillips and Tatacharya (2009, pp. 430–435).

  64. 64.

    As it turns out, Vyāsatīrtha goes on to argue that non-conceptual perceptions would generally prevent any subsequent concept-laden awareness from arising at all (TT, 500-502). First, we need to assume with Naiyāyikas that a subject has only one state of awareness (jñāna) at a time (NS 1.1.16), and that most every type of awareness-state persists for two moments (see NSM 108, 461.03-04; Shaw 1996, p. 258, n. 12, 259, n. 16; Phillips and Tatacharya 2009, p. 604). Assume also Gaṅgeśa’s particular story that an introspectively invisible non-conceptual perception arises at t1 and persists until t2, at which point (t3) it yields a concept-laden perception. Now, the specific causal conditions which produced the non-conceptual state at t1—e.g., sense-object contact—would still be present prior to t3, along with the additional condition that is responsible for generating the concept-laden perception—namely, the prior occurrence of an awareness of a qualifier, i.e., the non-conceptual perception. Vyāsatīrtha first suggests that the prior co-existence of these two sets of causal conditions (sāmagrī) would consequently cancel each other out and prevent either type of perception from arising in the next moment, just as two equally valid but contradictory inferences (i.e., a case of satpratipakṣa) can prevent one from drawing a definitive inferential conclusion.

    One possible reply is to claim that the prior existence of the non-conceptual state at t2 prevents another non-conceptual perception from arising at t3, allowing the concept-laden perception to arise as expected. But, this move doesn’t work, because Naiyāyikas also admit the existence of serial perceptual awareness (dhārāvāhikapratyakṣa), wherein one has a series of perceptual states all of the same object (with the latter perceptions being states of knowledge about the object just as much as the first perception). The existence of the prior perceptual states in the series don’t prevent the later perceptions from arising. (Whereas in the case of inference, one’s knowledge of an inferential conclusion or anumiti—together with a lack of any desire to prove the conclusion once again (see Mohanty 1992, pp. 102–103)—will block that same anumiti from subsequently arising.) So, the existence of a non-conceptual perception shouldn’t be sufficient for blocking a non-conceptual perception from arising immediately after, which means that the first objection—i.e., the causal conditions for non-conceptual and concept-laden perceptions would cancel each other out—still stands.

    The Naiyāyika couldn’t then attempt to argue that the two sets of causal conditions, rather than canceling each other out, instead cooperate to produce a single perception that is jointly non-conceptual and concept-laden. Indeed, Gaṅgeśa takes the view that concept-laden, predicative perceptions can also partly non-conceptual/non-predicative, in that there can be some qualifier which isn’t itself presented as being qualified by anything else. For example, the qualificative perception of a pot would concept-laden with respect to the pot which is qualified by pothood, but non-conceptual with respect to the pothood that is not presented under any other qualifiers (see TCM, 868). He doesn’t consider this partly concept-laden, partly non-conceptual structure to be contradictory, just as it’s not a contradiction for Viṣṇu to have a form that is part-man and part-lion (narasiṃhākāra; TCM, 856.04).

    Vyāsatīrtha evidently disagrees, though. The perceptual awareness “This is a pot” is conceptual/qualificative because it takes pothood as its predicative content (i.e., the perception is ghaṭatva-prakāraka); a perception of the exact same objects would be non-conceptual only if it didn’t take pothood as a predicative content (ghaṭatva-aprakāraka). There is one type of conjunctive perceptual awareness (samūhālambanajñāna) that takes multiple objects as its contents—e.g., “There is a pot and a cloth”— such that one could say part of the awareness takes pothood as a predicative content, and another part of the awareness doesn’t. But, this is true only in the sense that the awareness also takes clothhood as an additional predicative content. It isn’t contradictory for a single awareness to present both pothood and clothhood as qualifiers (of different qualificands), so no contradiction results from the cooperation of the distinct causal conditions for perceiving a pot and perceiving a cloth. However, it would be a contradiction for a single awareness of a pot to both take and not take pothood as its predicative content; hence, the distinct causal conditions for conceptually and non-conceptually perceiving the same pot can’t cooperate together to produce a single awareness of that pot. In that case, the first objection still stands—the opposing sets of causal conditions would prevent any perceptual awareness from arising after a non-conceptual perception.

    The last and most promising response is to claim that the causal conditions for concept-laden perception are more powerful than the conditions for producing non-conceptual perception, so that even if both sets of conditions are present at t3, only a concept-laden perception would arise, thus resolving the problem of mutual causal cancellation. Vyāsatīrtha parries this response by deriving an unwanted consequence for the Naiyāyika’s story of how a doubly qualificative awareness like “The man has a stick” (“daṇḍī puruṣaḥ”) is produced. In this awareness (at t5), the man is primarily qualified by the stick, and the stick is secondarily qualified by the property of stickhood. So, there must be a prior qualificative perception of the stick as qualified by stickhood (at t3-4), and that qualificative perception must be preceded by a non-conceptual perception of the stick and stickhood (at t1-2). Throughout this sequence, one maintains sensory contact with both the man and the stick. The problem, however, is that if the causal sāmagrī for generating concept-laden perceptions is stronger than the sāmagrī for non-conceptual perception, then after there is a non-conceptual awareness of the stick and stickhood, the causal conditions are in place for also generating an awareness of the man as qualified by the stick which is itself presented as shorn of any qualifying property (muṇḍitadaṇḍa). This kind of qualificative perception would instead arise prior to, and in place of, the perception of the stick as qualified by stickhood, with the stickhood also being presented as shorn of any qualifying properties. Of course, neither Vyāsatīrtha nor the Naiyāyika would say that we ever experience such an awareness of the man as qualified by the unqualified stick. The trouble is that, once the Naiyāyika accepts the general dominance of the savikalpaka-sāmagrī over the nirvikalpaka-sāmagrī, there is nothing in principle within the Naiyāyika’s causal story to prevent such an awareness from arising after the non-conceptual perception of stick and stickhood: At t3, there exists both sensory contact with the man (i.e., viśeṣyendriyasannikarṣa), and a prior awareness of the qualifying stick (i.e., a viśeṣaṇajñāna), so all the necessary conditions are in place for perceiving the man as qualified by the unqualified stick, rather than having to perceive just the stick as qualified by stickhood. Especially given the difficulties outlined in Sect. 4 concerning how to determine which non-conceptual content goes on to serve as the predicative content in a subsequent qualificative awareness, there doesn’t seem to be anything preventing the unqualified stick from being the qualifier in the subsequent qualificative perception of the man. If anything, this undesirable consequence drawn by Vyāsatīrtha highlights the convoluted and ad hoc character of the Nyāya account of doubly qualificative perception, making Vyāsatīrtha’s own alternative account more attractive by comparison. On his view, the perception of a man as possessing a stick can arise at t1 simply from the direct sensory connection with the man, the stick, and the contact-relation obtaining between them.

  65. 65.

    TT, 500.01-04.

  66. 66.

    Though, staying with its relevance to the inferential knowledge of absences, Vyāsatīrtha’s argument suggests yet another problem with Gaṅgeśa’s account. The atīndriyatva of non-conceptual awareness turns one of Gaṅgeśa’s own objections to the Bhāṭṭa Mīmāṃsā account of anupalabdhi back on himself. To vindicate their belief in the absence of awareness (anupalabdhi) as being the sole source of knowledge about absences in the world, Bhāṭṭa Mīmāṃsakas propose that the absence of an awareness can by its own nature directly grasp the non-awareness of an object, without itself being known. This claim is made in order to avoid the infinite regress that would result if the anupalabdhi which makes the absence of an object known must itself be known. (The anupalabdhi is itself an absence, so there would need to be a 2nd anupalabdhi to make the 1st anupalabdhi known, and so on.) Yet, in cases where one has to infer the absence of an object on the basis of one’s lack of an awareness of that object—as when I assert that Maitra was absent in the morning because I don’t remember seeing him—then the lack of awareness must itself be known. To avoid triggering a regress in such cases, the Bhāṭṭa proposes that a second-order anupalabdhi remain unknown while making known the first-order non-awareness of Maitra. Presumably, an analogy with the sense faculties can be drawn: Just as a sense faculty can by its own nature generate knowledge of a certain object without itself needing to be known, a second-order non-awareness can by its own nature make it known that there is no first-order awareness of a certain object, without the second-order non-awareness itself needing to be known.

    In response, Gaṅgeśa points out that because the Bhāṭṭas think that all states of awareness are atīndriya or introspectively invisible, it isn’t possible for an absence of awareness to be known by means of anupalabdhi, that is, through another absence of awareness. If an awareness-state is not yogya or fit to be perceived, then the fact that I’m not aware of the state obviously cannot provide me with evidence that the state is actually absent (TCM, 717.01-03, 722.08-10). And so, if I am unable to know that I didn’t have an awareness of Maitra, then I can’t infer on the basis of that awareness-absence that Maitra himself was absent. But, this exact same objection concerning anupalabdhi-based inferences will rebound back to Gaṅgeśa, since he too admits the existence of a certain type of introspectively invisible awareness, namely non-conceptual awareness. Gaṅgeśa’s own objection against the Bhāṭṭas thus doesn’t have a leg to stand on.

  67. 67.

    TT, 493.01-03. According to Nyāya, desires and aversions have intentional contents which are predicatively structured. These states are generated by a prior awareness that classifies an object as being pleasurable or painful, or as being a means of attaining some other desired or undesired object (see NSMK 145-146, 472.16-17 & 26-29). As a result, states of desire and aversion will themselves identify those objects as being the sorts of thing which are desirable or undesirable, or which conduce to the satisfaction of another desire. While Naiyāyikas would claim that there is no reason to admit the existence of pre-predicative desires and aversions, the fact that pre-predicative states are necessarily imperceptible still opens the door to their being present without our knowledge.

  68. 68.

    TT, 503.08-504.01.

  69. 69.

    ND, 503.20-21.

  70. 70.

    TT, 503.01-03.

  71. 71.

    Because nirvikalpaka states of awareness pose an obvious counter-example to Annambhaṭṭa’s definition of awareness as the cause of all linguistic expression, his commentator Nīlakaṇṭha instead interprets him as stating some of the properties of awareness (svarūpakathana), rather than pinpointing a defining characteristic (lakṣaṇa) which encompasses all states of awareness and nothing else. See TSBP, 462.19-25 for further discussion of the reasons why awareness can’t literally be defined as the cause of all linguistic expression. The defining characteristic of awareness ultimately is nothing more than the natural-kind property (jāti) of being a state of awareness (jñānatva; TSP, 114.20), which still belongs to non-conceptual states of awareness. This is despite the fact that our normal way of knowing that some mental state is an awareness is through introspecting its awarenesshood—that is, through introspectively noticing that I am aware of something (“jānāmiiti)—which is unavailable in the case of unintrospectible non-conceptual states.

  72. 72.

    TCMśa, 154.09-155.06. See also NSM 146-147, 503.01-504.03. Explicit elaborations on the predicative structure of the awareness-states which generate desires and volitions can be found in NSMK 146-147, 472.26-29 & 473.31-474.10.

  73. 73.

    TCM, 441.01-02.

  74. 74.

    TCM, 434.01-437.01-02.

  75. 75.

    TT, 503.03-07.

  76. 76.

    TCM, 567.01.

  77. 77.

    TCM, 866.09-10: “daṇḍasambandha eva puruṣe daṇḍajñānaṃ vinā na jñāyata iti brūmaḥ.” “We state that without a [prior] awareness of a stick in regards to a man [possessing that stick], a relation with a stick cannot be known.” TT, 495.01-03: “uktena vyabhicārena eva, ‘viśiṣṭajñānaṃ sambandhaviṣayakaṃ, sambandhajñāne ca sambandhijñānaṃ hetuḥ, sambandhi ca viśeṣaṇamiti nirastam.” “The previously stated counter-examples refute the claim that, ‘A qualificative awareness has a relation as its content; the awareness of a relatum is the cause with respect to an awareness of a relation; and, the qualifier [in the qualificative awareness of the relation] is a relatum.’”

  78. 78.

    Vyāsatīrtha at TT, 504.11-505.11 provides a helpful synopsis of the general difficulties or defeaters (bādhaka) which face Gaṅgeśa’s causal account of perception while not threatening the Dvaitin account, as well as the defeaters that specifically undermine Gaṅgśa’s posit of non-conceptual perception. The specific defeaters are discussed in Sect. 7 and particularly footnote 64.

  79. 79.

    PP, 188.07-09. Jayatīrtha’s commentators point out that a non-conceptual awareness cannot be responsible for generating the memory of either the name or absentee that will then serve as qualifiers in the subsequent qualificative perception of the named object or absence. Some commentators (e.g., Rāghavendratīrtha (PPBD, 189.05-06), Viṭṭalabhaṭṭa (PPṬ, 191.12-14)) suggest that a sensory connection is directly responsible for generating the relevant memory, while other commentators (Janārdanabhaṭta (PPV, 188.15-18), Vedeśatīrtha (PPBV, 190.16-17)) claim that a sensory connection first gives rise to a predicative awareness that then generates the memory—for instance, one first perceptually classifies something as being an animal before one then recalls and applies the linguistic term “animal.” In either case, non-conceptual perception is cut out of the picture.

  80. 80.

    The first-sight argument may also be the Naiyāyika’s best way of defending Gaṅgeśa’s proof of non-conceptual perception. For instance, having raised the objection that memory traces from past lives could be involved in generating a first-time qualificative perception—an objection which Rucidatta and Vyāsatīrtha believe does not lead to a vicious regress—all Gadādhara has to say in response is that non-conceptual perception is necessary just in order to account for the perception of properties which are being experienced for the first time in any life; without a prior memory of that property, the qualificative perception of it must get its predicative content from a non-conceptual perception. See VJHV, 76.28-29, NSMañ, 15.23-16.02.

  81. 81.

    I will be discussing the brand of conceptualism developed by McDowell in Mind and World (1994a), not the revised position that he adopts in his later essay, “Avoiding the Myth of the Given” (McDowell 2009, pp. 256–274).

  82. 82.

    In a later shift, McDowell (2009) abandoned the idea that perceptual experience has to “take the world as thus and so” and share the same type of propositional content as judgments. Perceptual content still has some unified structure that allows it to be discursively articulated by a perceiver, and that entitles the perceiver to make propositional judgments about what is perceived, but the perceptual content itself does not take a propositional form.

  83. 83.

    Mohan Matthen (2005a, b) is one contemporary philosopher who offers an account of perception which is also unconstrained by these unnecessary dichotomies. He defends the “Sensory Classification Thesis,” which claims that sensory systems operate by sorting and assigning perceived external objects according to classes. This thesis comes close, I think, to articulating how Gaṅgeśa and Vyāsatīrtha would understand the conceptual structure of perceptual content. Matthen writes, “The Sensory Classification Thesis claims, in effect, that sensory awareness can be expressed in terms of a set of singular propositions, messages to the effect that a particular individual is assigned to a certain class, and is identified as exemplifying a certain property. This goes against the traditional distinction between sensation and perception, which implies that the former is more like an image from which such propositions may be extracted, but not in itself articulated in such terms, such articulation being attributed to the post-sensory process called perception” (2005a, p. 14).


Primary Sources

  1. BPBhāṣāpariccheda of Viśvanātha Pañcānana. [see NSM]

  2. JLVRJñānalakṣaṇāvicārarahasya of Harirāma Tarkavāgīśa. 1958. Bhaṭṭācārya, G., ed. Jñānalakṣaṇāvicārarahasyam of Śrī Harirāma Tarkavāgīśa: with the commentary of Vimarśiṇī. Calcutta: Sanskrit College.

  3. NBNyāyabhāṣya of Vātsyāyana. [see NS]

  4. NDNyāyadīpa of Rāghavendratīrtha. [see TT]

  5. NKNyāyakandalī of Śrīdhara. [see PDS]

  6. NKauNyāyakaustubha of Mahādeva Punataṃkara. 1930. Kaviraj, G., ed. The Nyāyakaustubha (Pratyakṣa Khaṇḍa). Benares: Vidya Vilas Press.

  7. NKusNyāyakusumāñjali of Udayana. 1957. Upādhyāya, P., & Dhundhirāja, S, eds. The Nyāyakusumāñjali of Śrī Udayanācārya with Four Commentaries: The Bodhinī, Prakāśa, Prakāśikā (Jalada) and Makaranda by Varadarāja, Varddhamānopādhyāya, Mecha Thakkura and Rucidattopādhyāya and with Notes by Śrī Dharmadatta (Bachchā Jhā), Benares: Chowkhamba Sanskrit Series.

  8. NMNyāyamañjarī of Jayanta Bhaṭṭa. 1969. K.S. Varadacharya, K.S., ed. Nyāyamañjarī: with Ṭippaṇī Nyāyasaurabha. Vol. 1. Mysore: Oriental Research Institute.

  9. NSNyāyasūtra of Akṣapāda Gautama. 1985. Nyayatirkatirtha, T. & Tarkatirtha. A., eds. Nyāyadarśanam: with Vātsyāyana’s Bhāṣya, Uddyotkara’s Vārttika, Vācaspati Miśra’s Tātparyaṭīkā & Viśvanātha’s Vṛtti. New Delhi: Munshiram Manoharlal.

  10. NSDNyāyasiddhāntadīpa of Śaśadhara. 1976. Matilal, B.K., ed. Śaśadhara’s Nyāyasiddhāntadīpa with Ṭippaṇa by Guṇaratnasūri. Ahmedabad: L. D. Institute of Indology.

  11. NSMNyāyasiddhāntamuktāvalī of Viśvanātha Pañcānana. 2002. Jere, A.N., ed. Kārikāvalī of Viswanātha Nyāyapañcānana Bhattācārya with the commentaries Muktāvalī, Dinakarī, Rāmarudri. New Delhi: Rashtriya Sanskrit Sansthan.

  12. NSMañNyāyasiddhāntamañjarī of Jānakīnātha Bhaṭṭācārya Cūḍāmaṇi. 1916. Misra, J., ed. Nyāyasiddhāntamañjarī: Śrīyādavācāryaviracitayā Nyāyamañjarīsārākhyaṭīkayā sahitā. Benares: E.J. Lazarus and Co.

  13. NSMDDinakarī of Dinakara. [see NSM]

  14. NSMKKiraṇāvalī of Kṛṣṇavallabhācārya. 1990. Sastri, N. & Sastri. S., eds. Nyāyasiddhāntamuktāvalī of Śrī Viśwanātha Pañcānan: with the commentary Kiraṇāvalī. Varanasi: Chaukhambha Sanskrit Sansthan.

  15. NVTṬNyāyavārttikatātparyaṭīkā of Vācaspati Miśra. [see NS]

  16. PDSPadārthadharmasaṅgraha of Praśastapāda. Jetly, J.S., ed. 1991. Nyāyakandalī, being a commentary on Praśastapaādabhāṣya, with three subcommentaries. Vadodara: Oriental Institute.

  17. PPPramāṇapaddhati of Jayatīrtha. 2004. Pandurangi, K.T., ed. Pramāṇapaddhatiḥ: Śrījanārdanaviracitā Jayatīrthavijayanāmnī ṭippaṇī, Śrīrāghavendratīrthaviracitaḥ Pramāṇapaddhatibhāvadīpaḥ, Śrīvijayīndratīrtha viracitā Pramāṇapaddhatiṭippaṇī, Śrīvedeśabhikṣuviracitā Pramāṇapaddhatiṭippaṇī, Śrīsatyanāthatīrthaviracitam Abhinavāmr̥tam, Śrīvyāsapurī Narasiṃhaviracitaḥ Ādarśaḥ, Śrīnivāsatīrthaviracitā Vākyārthamañjarī, Śrīviṭṭalabhaṭṭaviracitā Pramāṇapaddhatiṭippaṇī iti aṣṭa ṭippaṇī sahitā. Bangalore: Dvaita Vedanta Studies and Research Foundation.

  18. PPBDBhāvadīpa of Rāghavendratīrtha. [see PP]

  19. PPBVBhāvavivaraṇā of Vedeśatīrtha. [see PP]

  20. PPṬṬippaṇī of Viṭṭalabhaṭṭa. [see PP]

  21. PPVVivaraṇa of Janārdanabhaṭṭa. [see PP]

  22. PSPramāṇasamuccaya of Dignāga. 2005. Steinkellner, E., ed. Dignāga’s Pramāṇasamuccaya, Chapter 1. A hypothetical reconstruction with the help of the two Tibetan translations on the basis of the hitherto known Sanskrit fragments and the linguistic materials gained from Jinendrabuddhi’s Tīkā. <>

  23. TBTarkabhāṣā of Keśava Miśra. 1979. Iyer, S.R., ed. and trans. Tarkabhāṣā of Keśava Miśra. Varanasi: Chaukhamba Orientalia.

  24. TCMTattvacintāmaṇi of Gaṅgeśa Upādhyāya. 1973. Tatacharya, N.S.R., ed. Tattvacintāmaṇi: With Prakāśa of Rucidattamiśra and Tarkacūḍāmaṇi on Prakāśa of Dharmarājādhvarin. Vol. 1: Pratyakṣakhaṇḍa. Tirupati: Kendriya Sanskrit Vidyapeetha.

  25. TCManTattvacintāmaṇi (anumānakhaṇḍa) of Gaṅgeśa Upādhyāya. 1990. Kamakhyanath, T., ed. The Tattvacintāmaṇi of GaṅgeśaAnumāna Khaṇḍa: With the Commentary’Rahasya’ by Shri Mathuranatha Tarkavagisha. Vol. 2: Part 1. Delhi: Chaukhamba Sanskrit Pratishthan.

  26. TCMśaTattvacintāmaṇi (śabdakhaṇḍa) of Gaṅgeśa Upādhyāya. 1990. Kamakhyanath, T., ed. The Tattvacintāmaṇi of GaṅgeśaŚabda Khaṇḍa: With the Commentary’Rahasya’ by Shri Mathuranatha Tarkavagisha. Vol. 4: Part 2. Delhi: Chaukhamba Sanskrit Pratishthan.

  27. TCMAĀloka of Jayadeva Miśra. 1990. Kamakhyanath, T., ed. The Tattvacintāmaṇi of GaṅgeśaPratyakṣa Khaṇḍa: With Extracts From the Commentaries of Shri Mathuranatha Tarkavagisha & Shri Jayadeva Mishra. Vol. 1. Delhi: Chaukhamba Sanskrit Pratishthan.

  28. TCMPPrakāśa of Rucidatta Miśra. 1973. Tatacharya, N.S.R., ed. Tattvacintāmaṇi: With Prakāśa of Rucidattamiśra and Tarkacūḍāmaṇi on Prakāśa of Dharmarājādhvarin. Vol. 1: Pratyakṣakhaṇḍa. Tirupati: Kendriya Sanskrit Vidyapeetha.

  29. TSTarksaṃgraha of Annambhaṭṭa. 2008. Veezhinathan, N. 2008. The Tarka- saṅgraha-dīpikā-prakāśikā: With the Commentaries Bālapriyā and Prasāraṇa. Chennai: Sri Sri Sri Manalakshmi-Mathrubhutheswarar Trust.

  30. TSBPBālapriyā of N.S. Rāmānuja Tatacharya. [see TS]

  31. TSPPrakāśikā of Nīlakaṇṭha. [see TS]

  32. TTTarkatāṇḍava of Vyāsatīrtha. 1932. D. Srinivasachar, D. & and Madhvachar, V., eds. Tarkatāṇḍavam of Śrī Vyāsatīrtha with the Nyāyadīpa of Śrī Rāghavendratīrtha. Vol. 1. Mysore: Oriental Research Institute.

  33. VJHVViśeṣaṇajñānahetutāvādavīci of Gadādhara Bhaṭṭācārya. 1936. Misra, B., & Shastri, D., eds. Vādavāridhi by Śrī Śrī Gadādhara Bhaṭṭācārya and others. Benares: Chowkhamba Sanskrit Series Office.

  34. VSVaiśeṣikasūtra of Kaṇāda. 1969. Misra, N., ed. Vaiśeṣikasūtropaksāra of Śriśaṅkaramiśra. Varanasi: Chowkhamba Sanskrit Series Office.

Secondary Sources

  1. Bagchi, S. (1953). Inductive reasoning: A study of tarka and its role in Indian logic. Calcutta: Munishchandra Sinha.

    Google Scholar 

  2. Baillargeon, R. (2008). Innate ideas revisited: For a principle of persistence in infants’ physical reasoning. Perspectives on Psychological Science,3(1), 2–13.

    Article  Google Scholar 

  3. Bhatt, G. P. (1989). The basic ways of knowing: An in-depth study of Kumarila’s contribution to Indian epistemology. Delhi: Motilal Banarsidas.

    Google Scholar 

  4. Bermúdez, J. L. (2007). What is at stake in the debate about nonconceptual content? Philosophical Perspectives,21, 55–72.

    Article  Google Scholar 

  5. Bermúdez, J., & Cahen, A. (2015). Nonconceptual mental content. In E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy (Fall 2015). Retrieved October 15, 2019, from

  6. Bhattacharyya, S. (1990a). Gadādhara’s theory of objectivity, part 1. New Delhi: Indian Council of Philosophical Research.

    Google Scholar 

  7. Bhattacharyya, S. (1990b). Gadādhara’s theory of objectivity, part 2. New Delhi: Indian Council of Philosophical Research.

    Google Scholar 

  8. Bhattacharyya, S. (1993). Gangeśa’s theory of indeterminate perception, part 2. New Delhi: Indian Council of Philosophical Research.

    Google Scholar 

  9. Bronkhorst, J. (2011). A note on nirvikalpaka and savikalpaka perception. Philosophy East and West,61(2), 373–379.

    Article  Google Scholar 

  10. Burge, T. (2003). Concepts, conceptions, reflective understanding: Reply to Peacocke. In M. Hahn & B. Ramberg (Eds.), Reflections and Replies: Essays on the Philosophy of Tyler Burge, edited by Martin Hahn (pp. 383-396). Cambridge, MA: MIT Press.

  11. Chadha, M. (2004). Seeing without recognizing? More on denuding perceptual content. Philosophy East and West,54(3), 365–367.

    Article  Google Scholar 

  12. Chadha, M. (2006). Yet another attempt to salvage pristine perceptions! Philosophy East and West,56(2), 333–342.

    Article  Google Scholar 

  13. Chadha, M. (2009). Contents of experience. Sophia,48(3), 237–251.

    Article  Google Scholar 

  14. Chadha, M. (2014). On knowing universals: The Nyāya way. Philosophy East and West,64(2), 287–302.

    Article  Google Scholar 

  15. Chadha, M. (2016). Contents of experience: Revisited. In P. Bilimoria & M. Hemmingsen (Eds.), Comparative Philosophy and J.L. Shaw (pp. 141–157). Cham: Springer International Publishing.

  16. Chakrabarti, A. (2000). Against immaculate perception: Seven reasons for eliminating nirvikalpaka perceptions from Nyāya. Philosophy East and West,50(1), 1–8.

    Article  Google Scholar 

  17. Chakrabarti, A. (2001). Reply to Stephen Phillips. Philosophy East and West,51(1), 114–115.

    Article  Google Scholar 

  18. Chakrabarti, A. (2003). Perception, apperception and non-conceptual content. In A. Chatterjee (Ed.), Perspectives on consciousness (pp. 89–107). New Delhi: Munshiram Manoharlal.

    Google Scholar 

  19. Chakrabarti, A. (2004). Seeing without recognizing? More on denuding content. Philosophy East and West,54(3), 365–367.

    Article  Google Scholar 

  20. Chakrabarti, A. (2016). Remembering Matilal on remembering. Sophia,55(4), 459–476.

    Article  Google Scholar 

  21. Chakrabarti, A. (2019). Realisms interlinked: Objects, subjects, and other subjects. London: Bloomsbury Academic.

    Book  Google Scholar 

  22. Chatterjee, A. (2016). Computational traits in Navya-Nyāya? Sophia,55(4), 543–551.

    Article  Google Scholar 

  23. Dasti, M. R. (2012). Parasitism and disjunctivism in Nyāya epistemology. Philosophy East and West,62(1), 1–15.

    Article  Google Scholar 

  24. Duquette, J. (2018). On Appaya Dīkṣita’s engagement with Vyāsatīrtha’s Tarkatāṇḍava. Journal of Indological Studies,27, 1–24.

    Google Scholar 

  25. Ganeri, J. (2011). Artha: Meaning—testimony and the theory of meaning in Indian philosophical analysis. Oxford: Oxford University Press.

    Book  Google Scholar 

  26. Ganeri, J. (2012). The self: Naturalism, consciousness, and the first-person stance. Oxford: Oxford University Press.

    Book  Google Scholar 

  27. Gennaro, R. J. (2012). The consciousness paradox: Consciousness, concepts, and higher-order thoughts. Cambridge, MA: MIT Press.

    Google Scholar 

  28. Hanna, R., & Chadha, M. (2011). Non-conceptualism and the problem of perceptual self- knowledge. European Journal of Philosophy,19(2), 184–223.

    Article  Google Scholar 

  29. Heck, R. K. (2007). Are there different kinds of content? In J. Cohen & B, McLaughlin (Eds.), Contemporary debates in philosophy of mind (pp. 117–138). Oxford: Blackwell (Originally published under the name “Richard G. Heck, Jr.”)

  30. Kriegel, U. (2012). Personal-level representation. In S. Miguens & G. Preyer (Eds.), Consciousness and subjectivity (pp. 109–146). Frankfurt: Ontos Verlag.

    Google Scholar 

  31. Maitra, K. (2017). The Nyāya theory of concepts. In J. Tuske (Ed.), Indian epistemology and metaphysics (pp. 381–396). London: Bloomsbury Academic.

    Google Scholar 

  32. Matilal, B. K. (1968). The Navya-nyāya doctrine of negation: The semantics and ontology of negative statements in Navya-Nyāya philosophy. Cambridge, MA: Harvard University Press.

    Google Scholar 

  33. Matilal, B. K. (1985). Awareness and meaning in Navya-Nyāya. In B. K. Matilal & J. L. Shaw (Eds.), Analytical philosophy in comparative perspective: Exploratory essays in current theories and classical Indian theories of meaning and reference (pp. 373–391). Dordrecht: Springer.

    Chapter  Google Scholar 

  34. Matilal, B. K. (1986). Perception: An essay on classical Indian theories of knowledge. Oxford: Clarendon.

    Google Scholar 

  35. Matilal, B. K. (2002). A note on Śaṅkara’s theodicy. In J. Ganeri (Ed.), Ethic and epics: The collected essays of Bimal Krishna Matilal (pp. 421–432). New Delhi: Oxford University Press.

    Google Scholar 

  36. Matthen, M. (2005a). Seeing, doing, and knowing: A philosophical theory of sense perception. New York: Oxford University Press.

    Book  Google Scholar 

  37. Matthen, M. (2005b). Visual concepts. Philosophical Topics,33(1), 207–233.

    Article  Google Scholar 

  38. McCrea, L. (2015). Freed by the weight of history: Polemic and doxography in sixteenth century Vedānta. South Asian History and Culture,6(1), 87–101.

    Article  Google Scholar 

  39. McDowell, J. (1994a). Mind and world: With a new introduction. Cambridge: Harvard University Press.

    Google Scholar 

  40. McDowell, J. (1994b). The content of perceptual experience. Philosophical Quarterly,44(175), 190–205.

    Article  Google Scholar 

  41. McDowell, J. (1995). Knowledge and the internal. Philosophy and Phenomenological Research,55(4), 877–893.

    Article  Google Scholar 

  42. McDowell, J. (2009). Having the world in view: Essays on Kant, Hegel, and Sellars. Cambridge: Harvard University Press.

    Google Scholar 

  43. Mohanty, J. N. (1992). Reason and tradition in Indian thought: An essay on the nature of Indian philosophical thinking. Oxford: Clarendon Press.

    Google Scholar 

  44. Mohanty, J. (2000). Classical Indian philosophy. Lanham: Rowman & Littlefield.

    Google Scholar 

  45. Mohanty, J. (2011). Jñānalakṣaṇāvicārarahasyā. In K. Potter & S. Bhattacharyya (Eds.), Encyclopedia of Indian Philosophies: Nyāya-Vaiśeṣika philosophy from 1515 to 1660 (pp. 401–404). Delhi: Motilal Banarsidass.

    Google Scholar 

  46. Peacocke, C. (2001). Does perception have a nonconceptual content? Journal of Philosophy,98(5), 239–264.

    Article  Google Scholar 

  47. Phillips, S. H. (1997). Classical Indian metaphysics: Refutations of realism and the emergence of “new logic”. Delhi: Motilal Banarsidass.

    Google Scholar 

  48. Phillips, S. H. (2001). There’s nothing wrong with raw perception: A response to Chakrabarti’s attack on Nyāya’s “nirvikalpaka pratyakṣa”. Philosophy East and West,51(1), 104–113.

    Article  Google Scholar 

  49. Phillips, S. H. (2011). Epistemology in classical India: The knowledge sources of the Nyāya school. London: Routledge.

    Google Scholar 

  50. Phillips, S. H., & Tatacharya, N. S. R. (2009). Epistemology of perception: Transliterated text, translation and philosophical commentary of Gaṅgeśa’s Tattvacintāmaṇi (jewel of reflection on the truth), pratyakṣa-khaṇḍa, the perception chapter. Delhi: Motilal Banarsidass Publishers.

    Google Scholar 

  51. Pylyshyn, Z. W. (2007). Things and places: How the mind connects with the world. Cambridge, MA: MIT Press.

    Book  Google Scholar 

  52. Raghavendrachar, V. H. N. (1941). The Dvaita philosophy and its place in the Vedānta. Mysore: University of Mysore.

    Google Scholar 

  53. Raftopoulos, A. (2009). Cognition and perception: How do psychology and neural science inform philosophy?. Cambridge, MA: MIT Press.

    Book  Google Scholar 

  54. Roskies, A. L. (2008). A new argument for nonconceptual content. Philosophy and Phenomenological Research,76(3), 633–659.

    Article  Google Scholar 

  55. Sarma, D. (2016). An introduction to Mādhva Vedānta. London: Routledge.

    Google Scholar 

  56. Searle, J. R. (1991). Consciousness, unconsciousness and intentionality. Philosophical Issues,1, 45–66.

    Article  Google Scholar 

  57. Sellars, W. (1991). Science, perception and reality. Atascadero, CA: Ridgeview Publishing Company.

    Google Scholar 

  58. Sinha, J. (1934). Indian psychology: Perception. London: Kegan Paul, Trench, Trubner & Co.

    Google Scholar 

  59. Sinha, S. (2014). Self and agency: A defence of Praśastapāda’s differential naturalism. Doctoral thesis (PhD), University of Sussex.

  60. Speaks, J. (2005). Is there a problem about nonconceptual content? Philosophical Review,114, 359–398.

    Article  Google Scholar 

  61. Sharma, B. N. K. (1962). Philosophy of Śrī Madhvācārya. Bombay: Bharatiya Vidya Bhavan.

    Google Scholar 

  62. Sharma, B. N. K. (2008). History of the Dvaita school of Vedānta and its literature: From the earliest beginnings to our own times. Delhi: Motilal Banarsidass.

    Google Scholar 

  63. Shaw, J. L. (1996). ’Cognition of cognition’, part II, anuvyavasāya of Pandit Visvabandhu, translated From Bengali with explanatory notes. Journal of Indian Philosophy,24, 231–264.

    Article  Google Scholar 

  64. Spelke, E. S. (1990). Principles of object perception. Cognitive Science,14(1), 29–56.

    Article  Google Scholar 

  65. Stoker, V. (2016). Polemics and patronage in the city of victory: Vyāsatīrtha, Hindu sectarianism, and the sixteenth-century Vijayanagara court. Oakland: University of California Press.

    Book  Google Scholar 

  66. Taber, J. (2005). A Hindu critique of Buddhist epistemology: Kumārila on perception: The “Determination of Perception” chapter of Kumārila Bhaṭṭa’s Ślokavārttika. London: RoutledgeCurzon.

    Google Scholar 

  67. Tatacharya, N. S. R. (1992). Pratyakṣatattvacintāmaṇivimarśaḥ. Tirupati: Rāṣtṛiya Saṃskṛta Vidyāpiṭha.

  68. Toribio, J. (2008). State versus content: The unfair trial of perceptual nonconceptualism. Erkenntnis,69(3), 351–361.

    Article  Google Scholar 

  69. Vattanky, J. (2003). A system of Indian logic: The Nyāya theory of inference. London: Routledge.

    Google Scholar 

  70. Williams, M. (2014). Mādhva Vedānta at the turn of the early modern period: Vyāsatīrtha and the Navya-Naiyāyikas. International Journal of Hindu Studies,18(2), 119–152.

    Article  Google Scholar 

  71. Williams, M. (2017). Jayatīrtha and the problem of perceptual illusion. In J. Ganeri (Ed.), The Oxford handbook of Indian philosophy (pp. 559–576). New York: Oxford University Press.

    Google Scholar 

  72. Wright, W. (2015). Nonconceptual Content. In M. Matthen (Ed.), The Oxford handbook of philosophy of perception (pp. 181–197). New York: Oxford University Press.

    Google Scholar 

  73. Xu, F. (1999). Object individuation and object identity in infancy: The role of spatiotemporal information, object property information, and language. Acta Psychologica,102(2–3), 113–136.

    Article  Google Scholar 

Download references


I owe the inspiration for this article to Arindam Chakrabarti. I am especially indebted to Prof. Veeranarayana Pandurangi of Karnataka Sanskrit University and Prof. D. Prahladachar of Poornaprajna Samshodhana Mandiram for introducing me to the discussions of non-conceptual perception in several Navya Nyāya and Dvaita Vedānta texts. Finally, I thank the two anonymous referees for suggesting improvements.

Author information



Corresponding author

Correspondence to Amit Chaturvedi.

Ethics declarations

Conflict of interest

The corresponding author states that there is no conflict of interest.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Chaturvedi, A. There is Something Wrong with Raw Perception, After All: Vyāsatīrtha’s Refutation of Nirvikalpaka-Pratyakṣa. J Indian Philos 48, 255–314 (2020).

Download citation


  • Non-conceptual perception
  • Dvaita Vedānta
  • Navya Nyāya
  • Direct realism