Nyaya Ratnamala of Parthasarathi Misra 5.2

Others, however, assert that the sentence is strictly a cognition of this kind, and the sentential meaning is simply the subsequent cognition in which multiple universals, qualities, substances, and actions appear; and the relation between the sentence and the sentential meaning is simply the relation of cause and effect. On both views, the comprehension of an imagined relation between words and word meanings is an expedient means. / Still others declare that, like a word, the sentence too has parts—via the path [which holds] that the final phoneme assisted by the impressions of the preceding ones is the expressive unit, and so on—and that it expresses the sentential meaning, having as its means the cognition of word meanings generated by the words that are its parts, whose relation has been apprehended. / The revered author of the Commentary holds that the words themselves denote their own meaning as connected with other word meanings. / The Master holds that the word meanings themselves, denoted by the words, make known the sentential meaning. Among these, the view that the sentence and sentential meaning are of the nature of cognition is refuted already in the refutation of the Buddhist idealist/voidist doctrine (śūnyavāda), because of the contradiction of an operation upon oneself and because [the object] appears externally. The alternative regarding the difference or non-difference of words and word meanings is also resolved by resorting to the doctrine of non-exclusivity (anekānta). For a sentence is indeed nothing but words whose sole purpose is to make known a single qualified meaning; the sentential meaning too is simply the word meanings themselves, possessing the nature of mutually interwoven items. Thus, even if absolute non-difference or difference cannot be asserted, nevertheless, since an intermediate state is established, the two cannot be denied as external entities. The view of a partless sentence is also rejected simply because parts are cognized. If it is objected: 'Just as in the case of a word'—no, because [the example] is unproven. There, this might be said: 'Just as, despite the cognition of parts in the form of phonemes, a single partless word is adopted as expressive of the word meaning because of the appearance of a single word, because the comprehension of meaning cannot arise from the phonemes, and because the cognition of parts is erroneous—so too is the sentence. For there too, there is the reflexive cognition: "This is one sentence"; and that very [cognition] rejects multiplicity, since multiplicity and unity are contradictory. And since the comprehension of sentential meaning has no other means, it postulates a partless sentence.' This is not so, because the example is unproven. For there exists for us no partless entity of the nature of a word over and above the phonemes. As for the reflexive cognition 'this is one word', it has as its cause the comprehension of a single meaning, even though there are many [phonemes]. For a unity grounded in a single effect indeed belongs even to many things, as in the case of 'an army', or as in the case of 'the Rājasūya'. For this very reason, there is no reflexive cognition of a single word for those who do not know the unity of the meaning. By perception, however, an unconditional single nature of a word would be manifest without distinction both to the educated and to others whose sense organs are unimpaired, merely upon the recitation of the continuous text (saṃhitā). Therefore, a word distinct from phonemes is not established from the appearance of a single word; on the contrary, its very non-existence is ascertained through non-apprehension of the perceptible. / Nor is another word established through the comprehension of meaning; / for, arising immediately after the phonemes, that [comprehension of meaning] has them alone as its cause for us.

When it is established that phonemes alone are the objects of sense perception and nothing else, then the comprehension of meaning that arises immediately after phonemes is understood to be caused by them alone. The expressive capacity of any other entity, whose existence is refuted by the non-apprehension of sensory perception, is completely dismissed; nor can it be postulated even on the basis of [otherwise] unaccounted-for fact. Therefore, because we observe that comprehension of meaning occurs immediately after phonemes that have attained succession through articulation by a single speaker, possess specific accents, are combined, and have a specific sequence, it is determined that only such phonemes denote meaning. How, again, can those occurring in sequence combine? We reply: through the force of an intermediate operation. For each preceding phoneme, as it is apprehended, generates an intermediate operation termed an impression. And accompanied by that, the final phoneme produces the comprehension of meaning. 'If that is so, then many impressions would have to be postulated for each phoneme; better, then, that a single linguistic reality be postulated.' He says: no. Indeed, you cannot accept the existence of that whose reality has been dismissed by the non-apprehension of sensory perception. But an impression, postulated to account for what is observed, does not bring excessive embarrassment; even multiple factors may be postulated to account for what is observed, but nothing contrary to what is observed [may be postulated]. Or, there is no postulation here of an unestablished additional impression: an operation distinct from memory is accepted as already established, but generating that very memory, they also become the means for the comprehension of meaning, because only phonemes that are remembered are expressive. Objection: Just like experiences, memories occurring successively in relation to phonemes are not possible in the comprehension of meaning. Now, if there is a simultaneous recollection even of those experienced in sequence, then they would be expressive even when articulated in reverse order, because even in that case the impression causing memory is generated, and recollection arises simultaneously with respect to all of them, disregarding error. If, on the other hand, the order of articulation is also accepted as an auxiliary, that being so, that itself would be the postulation of an unseen factor; for an impression that causes memory also arises in another order, and memory is caused by it. In that case, what else [could account for it] by a specific sequence other than the postulation of an unseen factor to produce the effect? Nor can even an unseen factor be postulated, because even without articulation and its sequence, a specific meaning is comprehended from the words 'jarā' and 'rāja'—consisting of a noun and a verb—when mounted upon memory. If phonemes alone are the word, then