Page 532 - Encyclopedia of Philosophy of Language

Key Figures
representation' (Chomsky 1972:95); his 'indi- vidualism,' which claims that the explanation of linguistic phenomena rests ultimately on the proper- ties of individual human beings, rather than on their social interactions; his 'mentalism,' which posits that in some real sense speakers 'know' the grammars of their language; and his 'rationalism,' which claims that human language learning is mediated by innate mental schemata. Chomsky defends his views in vari- ous works (1972; 1975b; 1980; 1986); for a com- prehensive defense of (the bulk of) Chomsky's system of ideas, see D'Agostino (1986).
5. The Further Development of Chomsky's Ideas about Meaning
Returning to more strictly linguistic themes, the dec- ade after the publication of Syntactic Structures was also a time of various attempts to integrate a semantic theory into generative grammar. Chomsky, in a 1962 presentation, set the course for this development by raising the question: 'What are the substantive and formal constraints on systems of concepts that are constructed by humans on the basis of presented data?'(1964:51-52).
Katz and Fodor (1963) attempted to answer Chom- sky's question in the following way. First, they dis- tinguished between two faculties involved in the interpretation of a sentence: that provided by a uni- versal theory of meaning, whose primitive terms and principles form part of our strictly grammatical abili- ties; and that derived from extralinguistic beliefs about the world. The goal of semantic theory would be to explicate only the former faculty, a component of linguistic competence. Second, they developed an analogy between phonetics and semantics. Just as phonetic representations are based on a universal sys- tem of phonetic features, semantic representations would be built out of primitive conceptual elements. A reading for a sentence, then, would be determined by the syntactic structure of the sentence and the sem- antic features ('markers') in the lexical items that com- prise it, similar to construction of the phonetic representation of a sentence on the basis of the phono- logical distinctive features characterizing each lexical item and the language's particular phonological rules.
Chomsky endorsed the Katz-Fodor approach in Aspects of the Theory of Syntax. Nevertheless, as time has passed he has become increasingly skeptical that there is a universal semantic system, parallel to a uni- versal phonetic system. As he put it succinctly: 'I doubt that one can separate semantic representation from beliefs and knowledge about the world' (1979:142).
6. The Generative Semantics Challenge to Chomsky's Philosophy
It seems to be the case that Chomsky believes that the danger of admitting a substantive theory of meaning
into generative grammar is a prescription for the ulti- mate abandonment of a rationalist theory of language in favor of a return of an empiricist one. He would surely point to generative semantics (McCawley 1976) as an object lesson illustrating this point. This approach to grammatical description flourished as a current within generative grammar in the late 1960s and early 1970s. In brief, it took the Katz-Fodor ideas about semantics seriously and attempted to push them to their logical conclusion. Given these ideas, and the related one that deep structure is the locus of semantic interpretation (the 'Katz-Postal Hypothesis'), gen- erative semanticists came more and more to deny that any sensible boundary could be drawn between the syntactic and semantic components of grammar. They had many reasons for coming to this conclusion, but one of the central ones (and the most important for the present discussion) was that, given the existence of a universal semantic system, there exists an overlap between semantic constructs and those participating in what would appear to be strictly grammatical rules (e.g., constructs such as animacy, gender, and the
common/proper distinction among nouns). Gen- erative semanticists argued that the redundancy seem- ingly entailed by this overlap could be eliminated only by erasing the line between syntax and semantics (for detailed discussion of the steps that led them to this conclusion, see Newmeyer 1986).
As this model progressed, it came to challenge any dividing line between semantic and pragmatic facts as well. (Given Chomsky's hypothesis that no such dividing line exists, it would follow naturally that they would be led to this conclusion.) Thus by 1975 or so, the idea that a universal theory of semantic constructs exists had led the competence-performance dichot- omy, the linchpin of linguistic rationalism, to be aban- doned by generative semanticists.
7. Chomsky and Approaches to Meaning
Chomsky has been equally adamantly opposed to approaches to semantics with roots in the logical tra- dition, in which, by means of a model, an arbitrary sentence of a language is assigned a truth value with respect to a possible state of affairs. Such approaches began to gain currency among linguists in the mid- 1970s and, in one version or another, continue to dominate linguistic semantics today (see Dowty, et al. 1981).
Chomsky argues that anyone who believes in poss- ible world semantics is forced to make one of two choices about the status of the constructs that popu- late such models, and both of them are (in his opinion) unpalatable. On the one hand, they could be regarded in parallel to the way that constructs of syntax and phonology are regarded, namely, as elements of a theory of mind. But, Chomsky argues, it is not at all clear how possible worlds are mentally represented or how people can have access to calculations using
510