Syntactic Interpretations of Truth and Semantic Underdetermination

I shall discuss part of William Lycan's vigorous defense of Davidson's semantic program in Logical Form in Natural Language. In particular, I will consider Lycan's response to Stich's (1976) attempted trivialization of Davidson's constraints on semantic theory. Lycan's response to Stich's example seems apt, but I will argue that it does not apply to another sort of example which I shall develop. An example of this kind takes the form of a skeptical hypothesis pitted against an assumed correct description of an object language in Tarskian terms; the skeptical hypothesis ascribes a syntactically characterized truth predicate to the object language. I will go on to consider some responses to such examples which are based on general considerations about the explanatory role of a theory of logical relations in a natural language. It is argued that the desired explanations are forthcoming on the basis of the suggested syntactic interpretations. I shall conclude that either Lycan's reconstruction of Davidson's framework does not afford an adequate conception of evidence in semantics, or semantic structure is empirically underdetermined in a particularly radical way.