It is well known that Gautama's sutra (NS 1/1/4) on perceptual knowledge states that perception (pratyaksa) is that cognition (jnana) which is due to sensory contact with the object (indriyarthasannikarsotpanna) and which is nonlinguistic (avyapadesya), certain (vyavasayatmaka) and non-deviating from the object (avyabhicarin). According to Vatsyayana and his followers this sutra states only a definition and no division of perceptual knowledge while according to others it represents both a definition and a division. The first interpretation asserts that all the words in the sutra except that for the definiendum (‘perception’ = 'pratyaksa') are constituents of the defmiens. But the second interpretation treats only three out of these five as constituents of the definiens, the other two, namely, 'non-linguistic' ('avyapadesya') and  'certain' ('vyavasayatmaka) as indicative of a division. The definition, in whatever way it is viewed, is basically a causal definition though it does not enumerate the causal conditions exhaustively. The causal factor that is mentioned in the definition is sensory contact with the object. But such a contact can generate not only a cognitive state but also an affective state in appropriate circumstances. Hence the word 'cognition' is very much necessary for the definition. In the preceding sutra Gautama has enumerated the varieties of knowledge (prama) including the perceptual variety. Truth which is necessary for knowledge seems to have been represented by the word 'non-deviating' (from the object) in Gautama's definition of perception. This avowedly forms a part also of the definitions of other varieties of knowledge and is not anything exclusive to the definition of perceptual knowledge.
The word 'certain', that is to be; taken as a part of the definition, has been incorporated to exclude doubt, as doubt is characterised not by certainty but by oscillation between opposing positions. Doubt can be eliminated even if this component is not incorporated by stipulating that strictly speaking it is deviating from the object in respect of the false alternative involved in doubt. Alternatively, if 'sensory contact' as occurring in the definition is understood as contact with any of the five external sense organs, then if doubt is held to be necessarily mental as some philosophers hold and thus not due to such a contact there would be no need to introduce the word 'certain' to exclude doubt.

The word 'non-linguistic' has been interpreted by some as useful for excluding what is usually referred to as ubhayajajnana (a cognition which is produced by the dual factors of sensory contact and testimony). Supposing that a person is perceptually aware of the features of a perceptible object but is not conversant with the word for its kind then on the basis of the testimony of another to the effect that the object is an X he may come to know the object as perceived in that situation as an X. Such knowledge is thus due to two sources of sensory contact and testimony. Since the presupposition of the definition is that perceptual knowledge is not due to testimony, the component under consideration looks useful for excluding such knowledge produced by dual sources as noted here.

Many philosophers do not however accept the thesis of ubhayajajnana¹ They interpret the component 'non-linguistic' in the definition as indicative of refutation of the theory according to which every cognition takes its rise as wrapped up in language – (as sabde jata). Linguistic clothing of cognition, according to the Naiyayikas, is not essential for perception even though such a relationship may appear to be obvious for other varieties of cognition. Thus the usefulness of the component lies in asserting the existence of perceptual cognitions viewed as intrinsically non-linguistic in character. The implication then is that the component has not been inserted to secure an act of exclusion which a component normally performs in a definition. Later interpreters of the sutra thus treat only the three words 'cognition', 'due to sensory contact with the object’ and ‘non-deviating’ (from the object) as the genuine components of the definition. They do not take the remaining two words ‘avyapadesya’ (translated so far as ‘non-linguistic’) and ‘vyavasayatmaka’ (translated so far as 'certain') as components of the definition. The two words under reference are thus interpreted by these philosophers as representing a division into indeterminate (nirvikalpaka) and determinate (savikalpak) varieties of perception. The proposal was explicitly floated in the Nyaya tradition first by Vacaspati in Tatparyatika and later supported by Gangesa in the Pratyaksakhanda of his Tattvacintamani and also by Visvanatha in his Vrtti.². But the interesting point to be noted here is that neither Gautama nor Vatsyayana has spoken of this distinction anywhere in their works. Therefore, the thesis of nirvikalpaka is not a theory of the masters. It is thus natural to suppose that it has been smuggled into Nyava from the work of the sister system of Vaisesika. It seems to us that the theory has actually been borrowed from Prasastapada.³ The theory was ably defended and greatly developed by Gangesa within the backdrop of the system of Navya Nyaya propounded by him. But I am afraid that Gangesa differs from Prasastapada in significant respects. This will be clear as we proceed.

Next..