Blackwell Reference Online

Extract

Competent speakers of a natural language know what the sentences of that language mean. A theory of meaning for a natural language, if correct, specifies what each well-formed declarative sentence of that language means. Thus, the following question naturally suggests itself: what sort of relationship, if any, obtains between speakers of, and a correct theory of meaning for, a given natural language? In this paper I shall examine a number of answers that have been given in response to this question. In particular, I shall be considering whether any account of the relationship can provide an adequate justification for what has been an article of faith of those engaged in the construction of systematic theories of meaning for natural languages: the requirement that such theories be compositional. A theory of meaning is compositional if and only if (a) it has only finitely many proper (non-logical) axioms, and (b) each of the meaning-delivering theorems (“meaning-specifications”) served up is generated from the axiomatic base in such a way that the semantic structure of the sentence concerned is thereby exhibited.What motivation is there for seeking compositional semantic theories in preference to their more readily available, non-compositional counterparts? Why should the construction of a semantic theory be constrained by the requirement that it reflect the semantic structure of ... log in or subscribe to read full text