Analytic philosophy has seen a resurgent interest in the possibility of explaining linguistic meaning in terms of truth, which many philosophers have seen as considerably more tractable than meaning. The core suggestion is that the meaning of a declarative sentence may be given by specifying certain conditions under which it is true. Thus the declarative sentence ‘Venus is red’ is true just in case the condition that Venus is red obtains; and this is exactly what the sentence means.
As it stands, however, this suggestion provides us with no explanation of the meanings of the words and phrases that make up sentences, since in general they are not expressions that have truth-conditions. (There are no conditions under which the word ‘Venus’ is true.) Furthermore, it needs to be supplemented by some method of circumscribing the truth-conditions that embody the meanings of declarative sentences, since there are many conditions under which any given sentence is true: ‘Venus is red’ is true not merely when Venus is red, but also, for example, when Venus is red and 7 + 5 = 12; but it does not mean that Venus is red and 7 + 5 = 12.
Evidently the first problem can be solved only by finding other semantic properties which indicate the meanings of words and phrases. For example, it is sometimes thought that the meaning of a name can be specified by saying what it refers to; and that of a predicate by saying what it is true of. But notice that since the meaning of a declarative sentence can be grasped by first grasping the meanings of its basic components, meaning-indicating ascriptions of semantic properties to those components must entail a meaning-indicating statement of its truth-conditions. Semantic properties such as ‘referring to’ and ‘being true of’ satisfy this requirement, at least in the context of what is sometimes called a ‘truth theory’ for a language.
This still leaves the problem of how to circumscribe the right meaning-indicating statement of truth-conditions for declarative sentences. Indeed we now have a further problem. For if the meanings of the components of sentences are not stated directly, but merely in terms of what they refer to or are true of (say), then we must also find a way of determining which of the many ways of specifying what they refer to, or the conditions under which they are true of something, is meaning-indicating. These problems may arguably be solved by placing an appropriate truth theory for a language in a setting that allows us to appeal to the general psychology of its speakers.
Attempts to elucidate meaning in terms of truth-conditions induce a plethora of further problems. Many are a matter of detail, concerning the kinds of properties we should associate with particular idioms and constructions or, equivalently, how we are to produce truth theories for them. As a result of Tarski’s work, we have a good idea how to do this for a wide range of categories of expressions. But there are many which, superficially at least, seem to resist straightforward incorporation into such a framework. More general difficulties concern whether truth should be central at all in the analysis or elucidation of meaning; two objections are especially prominent, one adverting to antirealist considerations, the other to the redundancy theory of truth.
