The Meaning of Nouns: Semantic Theory in Classical and Medieval India . – Namartha-nirnaya of Kaundabhatta

The Meaning of Nouns: Semantic Theory in Classical and Medieval India . – Namartha-nirnaya of Kaundabhatta – book reviews

J. Ganeri

The study of language is central to the classical Indian philosophical debate, and at least four of the major schools, the Grammarians, Mimamsakas, Naiyayikas and Buddhists, generated a vast body of literature, most of which has still to be properly researched. There has been, however, a marked growth of interest in this material among philosophers and Indologists alike in recent years, perhaps at least partly due to renewed enthusiasm for realist theories of meaning. Among many recent publications in this field, special mention may be made of B. K. Matilal’s The Word and the World: India’s Contribution to the Study of Language (reviewed in Mind by A. Chakrabarti, 1992) and M. Siderits’ Indian Philosophy of Language. Deshpande’s book, published in the prestigious Studies of Classical India Series, is an important new addition to this list. He has presented a highly readable, and densely annotated, translation of a work in the Grammarian tradition, the Namartha-nirnaya (“Determination of the Meaning of Nouns”) chapter of Kaundabhatta’s Vaiyakarana-bhusana, and has supplemented his translation with an extended introduction providing the historical context for the doctrines discussed. His hope, he says, is “to present a picture of a vibrant period in the history of Indian philosophy of language”, and this he has certainly done.

The Vaiyakarana-bhusana is a relatively modern work, dating from the first half of the seventeenth century, and is a commentary on a short (seventy-four verse) book on grammar by Bhattoji Diksita. It is a comprehensive yet critical and innovative survey of the rival theories about the meaning of the various morphological categories, such as nominal stem, verbal root, nominal and verbal affix, etc., and is justly popular as an introduction to Sanskrit philosophical grammar in India. Three chapters have already been translated into English, in J. Gune’s The Meaning of Tenses and Moods, S. D. Joshi’s Kaundabhatta on the Meaning of Sanskrit Verb-roots and also his Sphota-nirnaya of Kaundabhatta (the latter two being of the shorter version called Sara).

In the Namartha-nirnaya, Kaundabhatta addresses such questions as whether the significance of a noun or description is particular or general, whether gender is a syntactic or a semantic phenomenon, and whether nouns are ever selfreferential. This last question raised some important issues connected with the use-mention distinction and with Frege’s Puzzle, and I shall devote most of the rest of this review to what Kaundabhatta says about this. First, though, a few brief comments about the other questions. The doctrine, well-known at least since Donnellan, that definite descriptions have both a “referential” and an “attributive” use, partly echoes the Indian debate between those who claim that the semantic value of a nominal stem is a particular and those who claim that it is a universal (articles, it should be noted, are rarely used in Sanskrit). The Universalists, in a manner reminiscent of the modern Russellians, argue that in sentences such as “Bring [the] cow”, the reference to an individual is to be explained by appeal to a non-literal, derivative mode of referring called laksana. The Universalist argues further that if the reference of a noun like “[the] cow” is taken to be a particular, then since it must refer to different particulars on different occasions, a noun becomes radically homonymous. The Individualist counters by pointing to the existence of invariant elements in the meaning of context sensitive expressions, in this example the feature cowhood which is called the “basis for application” pravrtti-nimitta or “delimitor of reference” sakyatavacchedaka. Thus, they argue, a noun can very well refer to individuals without thereby becoming homonymous. Kaundabhatta reports the intricacies of this debate at great length, but since they have been discussed extensively elsewhere in the modern literature, I shall not go further here. He has also an extremely interesting discussion of the semantics of gender. In view of obvious counter-examples to the naive view that gendered nouns signify physical sex, the Vaisesikas postulated that there are special generic properties of gender residing in objects, while the early grammarian Patanjali endorsed the Samkhya idea of gender as “reflections of different conditions of the ultimate elements of the primal matter (prakrti)”. Kaundabhatta criticises both these views at length, and defends the modern sounding view that gender is merely a syntactic feature of the word. Yet it is interesting to see just how far a realist theory of gender can be taken.

More than a third of the text is given over to a discussion of the initially strange question, do nouns refer to themselves? The question arises because of such usages as “Pronounce cow” (gam uccaraya) (p. 193). Sanskirt lacks a device for systematic quotation (though the particle iti often serves the purpose), nor does it have the means to distinguish direct from indirect discourse. It is clear, nevertheless, that this sentence is correctly interpreted as a command to utter the word “cow”, and Kaundabhatta notes a series of rival semantic hypotheses to account for this linguistic datum. One is that what he calls the “imitation” word (anukarana), e.g. “cow” as it occurs in “Pronounce cow”, is a different word from the phonetically identical “imitated” word (“cow” as it occurs in e.g. “Bring [the] cow”). The imitation word is a name of the imitated word. A problem for this view arises, however, when we note that it is the nominal stem go, not the accusative form gam, which is to be uttered. Deshpande mentions an analogous construction in English, “I will not tolerate any ifs and buts”. Kaundabhatta also notes constructions like “[The verbal root] bhu occurs in the meaning of existence” (bhu sattayam) (p. 240), where the verbal root bhu is uninflected, but if it is a name it should take a nominal affix.

Two other suggestions are that each word has itself as one of its semantic values, or that there is a non-literal, “pragmatic” relation between each word and itself (a suggestion due to the Naiyayika Gangesa). According to Kaundabhatta, however, all these proposals involve an unnecessary prolixity. His own ingenious suggestion is that we should retain the ordinary meaning clause “word w refers to object x”, but relax the rule of interpretation. “An [entity]”, he says, “can be the object of verbal cognition, if and only if, that [entity] is related to the primary signification function, either by the relation of being its describer (nirupaka), or by the relation of being its locus” (p. 197). In other words, either relatum of the meaning clause “w refers to x” can, in appropriate circumstances, be the interpretandum. When, in a sentence like “Pronounce cow”, the word is, as we would say, mentioned, the correct interpretation is obtained by substituting the adjunct (i.e. the word itself) of the meaning clause.

There is a passage where Kaundabhatta comes strikingly close to giving a sententialist solution to Frege’s Puzzle. Let me quote it in full:

Therefore, what the author of the Kavya-Prakasa says on the line “rep-

etition of a word, already once used, is not a fault in some cases” is jus-

tified. He says: “There cannot be a cognition similar to that [caused by

the word tamra ‘red’] unless the word tamra is [used again], in the

verse: “The sun is red (tamra) when it rises, and it is also red (tamra)

when it sets. Similarly, the noble people remain the same in prosperity

and calamity”. [If] there were no difference in the cognition of a pot

from the two words, ghata “pot” and kalasa “pot”, [then] one could

certainly get an identical meaning cognition from the word rakta [and

the word tamra]. For this reason, [if a word does not figure in the

meaning cognition], it would be inappropriate to say that one cannot

have [an identical] meaning cognition from a different word… In our

view, however, the word-form tamra figures [in the verbal cognition

from the word tamra]. Due to this distinctive feature, it is different

[from the verbal cognition derived from the word rakta]. (pp. 226-8)

In his explanatory notes, Deshpande takes rakta “red” and tamra “red” to be synonyms, but this tends to obscure the significance of the passage. For here Kaundabhatta has given an example the correct understanding of which requires the hearer to know that the two occurrences are co-referential. In other words, even though “tamra” and “rakta” both refer to the colour red, the sentences (1) When it rises the sun is tamra, tamra too when it sets. (2) When it rises the sun is tamra, rakta too when it sets. have different cognitive values: someone who does not know that “tamra” and “rakta” co-refer won’t understand the sentence as involving an identity. Kaundabhatta’s explanation of this is that the word itself is a “qualifier” (prakara) in the interpretation: (1) and (2) will then have different cognitive values because although “tamra” and “rakta” co-refer, they are different words.

It would be interesting to see what the later commentators had to say on this, and how Kaundabhatta’s ideas compare to some modern proposals. Deshpande, though, has not tried to examine Kaundabhatta’s arguments in a comparative way, for he rightly emphasises the need first to understand them in their proper historical setting. As he says, “[t]here is no such thing as the Indian philosophy of language. Similarly, there is no such thing as the view of Grammarians, Mimamsakas, or Naiyayikas. There are multitudes of different opinions on every topic, and each argument needs to be placed in its right historical context” (p. 84). With meticulous cross-referencing and annotation, he has accomplished this task superbly.

COPYRIGHT 1994 Oxford University Press

COPYRIGHT 2004 Gale Group