Truth in Semantics

Semantic theories for natural languages purport to describe a central aspect of the meaning of natural language sentences. In doing so, they usually employ some notion of truth. Most semanticists, even those who have no objections to invoking propositions, will define a truth-predicate that applies to sentences. Some will also employ a notion of propositional truth. Both types of semanticist face the question whether and how the semantic notion(s) of truth they are employing is (are) related to the ordinary, pre-theoretic notion(s) of truth. It seems immediately problematic to say that the semantic truth notion is a pre-theoretic notion. For pretheoretically, we do not seem to apply “true” to sentences. So, if the semantic truth notion is in any interesting way related to a pre-theoretic notion of truth, then the relationship is more complex.As I shall explain in Section 2, however, there are some straightforward ways of postulating an analytic link between semantic truth notions and ordinary truth. Now, recently there have been a host of proposals to adopt a “relativistic” semantics of certain expressions, such as predicates of personal taste, epistemic modals, sentences about the contingent future, evaluative sentences generally, sentences attributing knowledge etc. All these proposals involve the claim that the truth of utterances or propositions depends on an extra factor. As we shall see, the answers standard semanticists can provide to the question of the status of their semantic truth-predicate do not transfer easily to these new, relativistic theories. In Section 3, I shall make some proposals as to how relativists can deal with these