Against Immaculate Perception: Seven Reasons for Eliminating Nirvikalpaka Perception from Nyāya


By Chakrabarti, Arindam

Philosophy East and West

v.50 n.1(January 2000)


Copyright 2000 by University Press of Hawaii


Whether nonconceptual, nonpropositional sense experience is possible or not is still a live debate in contemporary Western epistemology. Some philosophers, following Wilfrid Sellars' influential attack against the "myth of the given," argue against the very intelligibility of the notion of a state that claims to be cognitive without involv­ing the use of any recognitional capacity or concept. Others, without resurrecting the well-wrecked sense-data version of the theory of perceptual immediacy, point out that unless we conflate knowledge with knowledge claims, a cognitive state con­sisting of something's phenomenally looking some way to someone can be trans­parently and directly intentional without involving concept predication. Indeed, such direct perceptual acquaintance has been considered by these "given"-friendly philosophers to be "what gives perception its distinctive character" as against memory, Judgment, reasoning, wondering, and other forms of abstract thought.[1]

        In classical Indian epistemology, Buddhists have sometimes gone to the extreme of defining all perception as nonconceptual, nonlinguistic, and untainted by any predicative construction. What is intriguing is that the Mīmāṃsā (Vedic Hermeneutic Ritualist School) realists themselves, while vigorously opposing Buddhist phenomenalism or the Buddhist mistrust of language as fiction-mongering rather than knowledge-generating, have generally made room for this kind of nonconceptual perception in their epistemology, sometimes illustrating it with the sense experience of a newborn. It is surprising that apparently while arguing against the Buddhist definition of perception as "Experience untouched by conceptual qualification," Śālikanātha (the Prābhākara Mīmāṃsaka) would say: "All awareness arises in the form 'Thus' (evam). The word Thus' means a qualifier (prakāra). Universals and qualities function as qualifiers of objects.... [H]ence the expression 'untouched by conceptual qualification' is empty (applies to no awareness)"[2]—and yet while clas­sifying perception, the same author would admit that it is of two types, qualificative and nonqualificative! This distinction also seeped into the hard-core realist Nyāya school although the objects of nonqualificative perception would not be anything other than those very material (and mental), ordinary, more-or-less stable objects that are known by more conceptually structured qualificative perceptions, whereas in Buddhism nonqualificative perception would have special qualityless, essenceless pure particulars (svalakṣaṇas) as their objects. The postulation of such special im-


mediate objects of pure sensation, which, like the notorious sense data or phenomenal qualities in the West, have always had a pull toward privacy or single-subject accessibility, has been naturally resisted by commonsense realists of the Nyāya ilk.
        The common public world revealed to us through property-ascribing, verbally communicable qualificative perceptions is, according to Nyāya, externally real and directly perceived—often correctly and sometimes "otherwise."[3] Why, then, do the Nyāya realists need such a primary state of "raw," nonqualificative perception at all, in comparison to which qualificative verbalized perception would always— misleadingly—look "cooked" and constructed and hence not quite "as it is"?!
        By trying to jettison the notion of nonqualificative perception within and on the basis of the Nyāya epistemological framework, this essay tries to understand the deeper relationships between direct realism and concept-enriched perception. If my broadly Nyāya-minded arguments against the traditional Nyāya doctrine of Indeter­minate Perception (nirvikalpaka pratyak
a) can be met, then there must be some deep link between direct realism about the external world and the admission of the possibility of preconceptual sensory awareness of objects and qualities; otherwise such a doctrine is indeed out of place in Nyāya, and we should get rid of the myth of the unqualified given even in Indian Realist thought.

The idea of a "veil of perception" seems seductive when one considers the possi­bilities of perceptual content being colored by illusion, interpersonal or inter-species divergence of points of view, and linguistic and cultural training. If you look at a stone and find it quite clearly a sacred emblem of Siva while 1 look at it and see a geologically interesting sample of granite, it sounds idiomatic to remark "You and I are not seeing the same rock." But the Nyāya realists have shown exemplary robustness in resisting such idiom slippage. That you and I can see the same rock even if we are seeing two different sides of it, that the rock I had inferred or talked about or heard people talk about could be exactly the same rock that I now see, that I can touch the rock that 1 saw from a distance first, and that the insensible particles "making up" the rock are real causes of but not reductively identical with the per­ceived rock are some of the pivotal tenets that such a direct realist has had to hold onto in the face of immense opposition from idealist, phenomenalist, and scientific-realist quarters. It follows from such a picture of ordinary, objectively featured par­ticulars that there is no need to see the rock minus all its qualities and properties and movements in order to see it as it is. Indeed, since the rock is ready black and really granite, to see it as a colorless, propertyless particular would be to misperceive it. In society as well as in nature, we do not need to strip or dissect bodies in order to see them "properly." Even if we are not seeing all of it we are still seeing it.

The effort to peel all concepts and qualities and predicates and classifications away from perception to make it faithful and "fiction-free" (kalpanāpoham) comes from a kind of metaphysical "nudism" that the Nyāya philosophers simply reject thanks to their belief that real substances are really rich with properties and relations inherent in them and that qualification and differentiation are not imposed but dis­covered by perceivers.



        Given such a realism about propertied particulars, it is odd and disappointing that there should be room in Nyāya epistemology for pre-predicative or non-qualificative "pure" (nirvikalpaka) perception!

        The following is an attempt to show why we can easily do without nirvikalpaka perception inside the Nyāya epistemology and why it creates more problems than it solves. A full-scale attack against Gageśa's theory of pre-predicative perception is launched by the great Māddhva logician Vyāsatīrtha in his book Tarkatāndava (The deadly dance of dialectic).[4] I have only minimally used his arsenal of anti-Gageśa arguments here. Most of the arguments here are my own. For my part, I would be happy to see these objections answered and the concept of nonqualificative per­ception (defended by such great philosophers as Vācaspati and Gageśa inside Nyāya and by Kumārila and Śālikanātha in Mīmāṃsā) reinstated. But until such answers are available I would like to try to reconstruct Nyāya epistemology shedding the idea of a nirvikalpaka perception, letting all that count as cognition or awareness be uniformly qualificative in their content.

        There is a general rule that unless I instantly fall asleep immediately after undergoing a mental state, I can usually "apperceive" or be immediately aware of the special qualities of my own self, that I can know that I am happy or hateful or hearing or seeing when I am happy, hateful, hearing, or seeing. The state of a non­qualificative perception is supposed to be one exception to that general rule. Even if I am fully attentive to my own perceptual situation, I can never apperceive or make myself conscious of the allegedly indeterminate perception that I have just now had. Since such a perception is nameless, featureless, qualityless, and relationless I can­not intuit it under any description and hence cannot introspect or recall it even as a perception of mine! If we can drop this assumption of a nonqualificative perception, then such an odd exception can be avoided.

        Perhaps a step-by-step registering of the qualifier first and then the qualificand and the relation can be admitted at the sensory level, given the one-item-at-a-time nature of the atomic internal sense organ (manas). But there seems to be no ground for calling such sensory registering "awareness."

        However, two very grave-sounding objections could be made against my proj­ect of purging Nyāya of indeterminate perception.[5] First, awareness itself is taken in Nyāya to be formless. So at least in one instance it has to bear its structurelessness on its face by lacking any subject-predicate sort of intentionality and by grasping objects and qualities loosely without any cementing bond. Second, the world is a plurality of particulars, universals, motions, inherence, and so forth. In judgmental perception this discreteness of the elements is not fully reflected. The juxtaposition without relation that nirvikalpaka brings in makes their separateness vivid (while the verbal content-representation of a determinate perception of a rabbit would be "a rabbit qualified by rabbitness," the verbal content-representation of the inherently unverbalizable, indeterminate perception preceding such an identificatory perception would be "rabbit, rabbitness").

        However, neither of these two points seems to be compelling. First, the form­lessness of awareness championed by Nyāya only consists of denying the possibility


of awareness supplying its own object from within itself. It is the thesis that no cog­nition can exist without an external (not necessarily physical) object—the flip side of their denial of the self-illuminative nature of awareness. Just as the knowledge-independence of objects need not be proved by performing the impossible feat of producing an unknown object, the formlessness of cognition need not be proved by making room for an actually formless awareness.

        By comparison and contrast, we can show that my seeing of a white piece of chalk does not contain the white chalk within itself, because the "seeing" aspect of it would have remained the same even if I happened to see a red cherry. The chalk or the cherry could exist without my seeing them, and my seeing is only externally related to them, but all the seeings that we register are already saddled with an ex­ternally structured content.

        Similarly, it is absurd to try to find actual perceptual evidenced a world where pure particulars and floating universals exist unhinged, because there is no such world and there are no such introspectively accessible perceptions. The plurality of the distinguishable elements of any structured state of affairs in the world is a piece of philosophical knowledge gained by reflection on the very fact that, as Śakara Miśra insists: 'There is no single cognition where a difference does not figure as an object." It is because all cognitions are relational that the world is known to be a multiplicity of things that fit into each other like the links of a chain, rather than be­cause, in rare lucid moments, we have nonqualificative awareness of first one thing and then another. Many ness can only be appreciated by savikalpaka perception— for example, of mutual otherness. When Mohanty remarks about indeterminate per­ception that "I believe this admission is vital to Nyāya realism/ for only such an awareness opens an access to things as they really are...," [6] he is doubly mistaken. First, since things in the world are qualified in themselves, a perception that shows them to be unqualified and disjointed would not give us access to them as they re­ally are. Second, if this remark were correct, then all qualificative perception would involve a little bit of distortion or dressing up—precisely the point that Nyāya real­ism has fought tooth and nail not to concede to the Buddhist. Jayanta says that qualificative perception is the very life-breath of Nyāya philosophers.[7] This is so because such predicative perceptions alone give us access to things as they really are!

        Things really are qualified. Nowhere in the world is there an inherence that does not hold between two other entities, because in order for it to be an inherence it has to be a relation of a pair of things that are not separately available. So a non­qualificative cognition—had it been a consultable "access" to reality—would have been misleading. Luckily it is not an access we can consult as evidence! Also Mohanty inadvertently admits that such pre-predicative awareness does not have even the basic epistemic features of having an object-directedness or viayitā. Hence it is not an epistemic state. In being pre-predicative it is also pre-cognitive, and nothing pre-cognitive is a perception. It may be a stimulation of nerves of a certain kind, but it is not an awareness. Bertrand Russell was careful enough not to identify his "knowledge by acquaintance" with the mere acquaintance (the converse of the



relation of the object's being presented to the senses) by which such knowledge is generated.

So, here are the seven reasons for rejecting non-predicative perceptions:

        1. It all started with the adjective "nonverbal" (avyapadeśyam) in the Nyāya Sutra definition of perception. But there is no good reason to interpret Gautama's Nyāya Sūtra 1.1.4 to be partly a definition and partly a bipartite division of perception. The next Sūtra is quite clearly so. With Just one word "tat-pūrvakam" for a definition, 1.1.5 hastens to classify inferences into three kinds. But Vātsyāyana as well as Jayanta and others have taken all five terms of the definition of perception to be applicable to all sorts of perceptions.

        Thus "avyapadeśyam" can very easily apply to savikalpaka perceptions/ insofar as even a piece of sensory knowledge that can be communicated in words does not need those (or other) words as its cause or object. The perception of a leaf as green neither is caused by nor is a perception of the word "green." Arid "nonverbal" in the definition of perception does not mean unmentionable or undenotable by words. Everything that exists or happens is mentionable and denotable by words in the Nyāya scheme of things. If a certain kind of perception were really beyond words then even descriptions like "avyapadeśyam" or "nirvikalpakam" or "ghaṭa-ghaṭatve-ityākārakam" would not have been applicable to it. So Vācaspati was simply mistaken when he arbitrarily made a concession to the Buddhists by making one term of the general definition of perception apply to only one sort of pre-predicative perception. Another unwelcome consequence of his partitioning of the definition into classificatory terms was that the term "doubt-free" or "determinative" (vyavasāyātmakam) would be meant to be inapplicable to pre-predicative percep­tion, giving the false impression that the so-called pure perceptions are uncertain or doubtful! As perceptual knowledge all perceptions should be determinative and initially sure. Vācaspati was nervous about this and that is why, after having done this particular hermeneutic violence to the text, he appeals to authority saying that he is following his Guru Trilocana in doing so.

        2. The rule "Every qualificative awareness is caused by an awareness of the qualifier" is especially suspect as a causal claim, because the only example given of that vyāpti (generalization) is inferential awareness (anumiti), which is caused by parāmarśa and so forth. But a mediate knowledge such as an inference, which is caused by another knowledge, is precisely what immediate sense perception should not be compared to. As immediate awareness, perception, even when determinate, should not be caused by another perception, on pain of losing its immediacy' We can intuitively make sense of the claim that one cannot descry a featured object unless one descries the feature in question—that one cannot recognize something to be a rabbit unless one is acquainted with rabbithood as a feature. But there is no need to take this logical "unless" as a causal "unless." The feature need not be first described as an unbonded property before we can see an animal as qualified by it. One cannot spell "London" unless one first utters or writes "L." But one does not first have to spell "L" in order to spell "London."

        That rule, in any case, has too many exceptions. In seeing a sheer absence, it is



admitted that we can see the floor to be devoid of the carpet. In the content of this seeing, the absence of the carpet qualifies the floor. But insofar as a nonqualificative awareness of absence is, for good reasons, recognized to be impossible, we cannot insist on a prerequired pure acquaintance with the qualifier (in this case/ an ab­sence). Gageśa, who is committed to pure acquaintance, takes care of this by an ad hoc adjustment. The form of the perception is reversed: the absence is made the quaiificand and the floor the qualifier.

But this is totally counterintuitive. We most certainly see the floor to be charac­terized by a carpetlessness rather than spot a lack of carpet first and then embellish it with its location!

        3. Gaṇgeśa himself, after rejecting several traditional definitions of perception including Gautama's, defines perception as an awareness of which no other aware­ness is the instrumental cause (jñāna-akaraaka jñānam). Now, in most other places he rejects the "cause with an operation" (vyāpāravat kāraṇam) definition and accepts the "immediate antecedent, that upon which the effect immediately fol­lows" (phalāyogavyavacchinna) definition of instrumental cause. If he did that here, then many savikalpaka perceptions, especially those at the first moments of awakening, which are immediately preceded (allegedly) by nirvikalpaka aware­ness, would be disqualified by Gaṇgeśa's definition, insofar as they would be instrumentally caused by another awareness.

        Thus, if we are to preserve Gaṇgeśa's beautifully insightful definition of percep­tion, we must drop nirvikalpaka.

        4. As far as sannikara (the operative sense-object connection) goes, except in the case of hearing sounds and qualities of sounds and so forth, all perceptually operative sense connections happen via contact (sayoga). When the visual sense organ or skin is in touch with a substance and perceives a universal such as cupness or a textural feature such as roughness that inheres in it, it is simply counterintuitive to say that it first sees the universal or feels the texture-quality without any clue that it is a cup or a bark that has that texture and then goes on to perceive, predicatively, that—ah!—it must be the cupness in that cup or the roughness of that bark. If caus­ally one is seeing a quality that is inherent in a substance that one's sense organ is in touch with, then it makes sense to say that epistemically the first encounter with the smile in a face is also as a smile in a face.

        5. The worst penalty that Nyāya pays for accommodating this myth of the pure given is that perception, which is introduced, taxonomically, as a variety of veridical awareness (pramā), comes to have a sub-variety that is not a veridical awareness, insofar as veridicality, in Nyāya, requires predicative structure and nirvikalpaka has none. If all perceptions have to be either true or false, then no per­ception can be pre-predicative.

        6. The heart of Nyāya realism lies in accepting the qualificative structure as lying out there in the world. Greenness will inhere in an emerald even when nobody will look at it or when some Goodmanian crackpots will look puzzlingly at it and fail to figure out whether it is "grue" or "bleen." Since we do not contribute this featured-ness (vaiśiṣṭya) onto the object, why do we need to see it in a causally layered way,



first the qualifier alone and then the qualified complex? That is already succumbing to the antirealist pressure that, after all, seeing the pure particular would have been the most immaculate way of seeing (as a baby or a mute person sees!) without the noise factor of language! Well, the Naiyāyika can never hope to make friends with the Buddhist this way, because language is not a distraction failing to touch per­ceptible reality in Nyāya. Words touch the same reality as perception. And, to make a Kantian point without the Kantian doctrine of self-conscious synthesis, to the extent that the baby 01- the speechless person notices something outside themselves, their experience must involve some application of concepts or some qualification, in any case, Buddhists would not be happy when they learn that the Naiyāyika immaculate perception often takes the form of direct acquaintance with a universal! How on earth, the Buddhist nominalist will wonder, could they start calling the bare aware­ness of a sheer vikalpa "a nirvikalpaka perception"?

        7. Finally, all informative awareness (anubhava), in order to count as awareness, must have intentional directedness toward some object or object complex. By virtue of being known about or believed, those objects, in turn, acquire a titular property of "objecthood" (viṣayatā). There are only three major kinds of roles or objecthood properties recognized by Nyāya, namely qualifierhood, qualificandhood, and relationhood. Now, it is admitted that a nirvikalpaka awareness does not endow its object or objects with any one of these roles. A pre-predicative perception does not have either a qualifier or a qualificand or a relation in its content. With the help of an ad hoc dodge, these unstructured perceptions are still made to have some other kind of objecthood or intentionality associated with them. But the concept comes very dangerously close to that of an awareness without any object directed­ness/ which is a contradiction in terms in Nyāya. Such an objectless awareness may make an Advaita Vedāntin (who admits immediate nondiscursive awareness even out of a contemplative listening to the words "You are That Brahman"!) happy. But it is too steep a price to pay for Nyāya—a school that would rather have its soteriological summum bonum be an unconscious state than have consciousness without objects.



I am grateful to two anonymous referees and to my friend Stephen Phillips for com­ments on an earlier draft. I sincerely hope to be answered, perhaps by a Gaṇgeśa enthusiast like Phillips himself, so that I know why immaculate perception is needed by Nyāya after all.

[1] - See William P. Alston, "Sellars and the Myth of the Given" (paper presented at the Eastern Division Meeting of the American Philosophical Association, Wash­ington, D.C, 30 December 1998).  [back]

[2] - Śātikanāthamisra, Prakāraapancikā, Banaras Hindu University Darsana series, no. 4 (Varanasi: Visvavidyalayena Prakasita, 1961), p. 142. [back]

[3] - For the clearest account of the Nyāya theory of pre-predicative perception, see Sibajiban Bhattacharyya, Gageśa's Theory of Indeterminate Perception: Con­taining the Text of Gaṅgeśa 's Nirvikalpakavada with an English Translation and Explanatory Notes, pt. 2 (New Delhi: Indian Council of Philosophical Research; distributed by Munshiram Manoharial Publishers, 1993).  [back]

[4] - Recently republished (Sri Vyasatirthaviracitam, Tarkatandavam [Mysore: Pracyavidyasamsodhanalayah, 1985]), this work is still untranslated. A very sketchy gist of its attack against indeterminate perception is to be found in Surendra Nath Dasgupta, A History of Indian Philosophy (Delhi: Motilal Banarsidass, 1975)/ vol. 4/ p. 183.  [back]

[5] - See Jitendra Nath Mohanty, Reason and Tradition in Indian Thought: An Essay on the Nature of Indian Philosophical Thinking (Oxford; Clarendon Press; New York: Oxford University Press/ 1992), p. 174.  [back]

[6] - Ibid.  [back]

[7] - See Arindam Chakrabarti, "Concept-possession, Knowledge of a Language and Sense-Experience," in The Philosophy of P. F. Strawson, ed. Lewis Edwin Hahn (Chicago: Open Court, 1998).  [back]