Comments

Abstract

Realism rests upon — or, better, consists in — an adherence to a truth-conditional semantics for our language. Where ‘semantics’ is taken to mean ’theory of meaning,’ and a theory of meaning is equated with a theory of understanding—an account of a speaker’s mastery of a language in terms of what he has to know to know the language — there are acute difficulties in such a semantics, which will involve attributing to a speaker, as that in which his understanding of each sentence consists, a knowledge of the condition that has to hold for it to be true. First, there is the problem of explaining how we derive from our grasp of the meanings of our sentences the actual use we make of them: in particular, if the meaning is not, in general, given in terms of how we recognise a sentence as true, but of what it is for it to be true, independently of our capacity to recognise it as such, an explanation is called for, and is difficult to provide, of how our grasp of that meaning leads us to treat this or that as conclusive evidence, or as evidence falling short of being conclusive, for its truth. (The classical ‘problem ofinduction’ is one aspect of this difficulty.) Secondly, there is a difficulty about what it even means to attribute to a speaker a knowledge ofthe condition for the truth of a sentence. At one end of the scale, where a grasp of reference cannot be thought of as mediated by anything, this appears to involve ascribing to the speaker an immediate association between word and object, the object becoming part of his private world; at the other end, truth, taken as subject to the principle of bivalence, appears to transcend our capacity for recognising it as attaching to the sentence, and a grasp of the condition for the truth of such a sentence therefore becomes something not exhaustively explicable in terms of actual linguistic behaviour.