Meaning, Truth and Interpretation
We owe to Donald Davidson the suggestion that a truth theory used as an interpretation theory for a speaker can do duty as a meaning theory for his language.† This is a brilliant suggestion, but there are some oddities in the development of this idea in Davidsonís work which need to be brought to light, and the project, in the form it takes in Davidsonís hands, in the end is too ambitious to succeed.† I begin by distinguishing three questions:
1.What is it for a word or expression to be meaningful?†
2. What is it for a word or expression to mean what it does?
3. How do the meanings of complex expressions in a language depend upon those of their parts?
A full answer to 1 would include answers to 2 and 3; but it is clear one could pursue question 3 independently of 1 and 2, and 2 independently of 1.† My discussion of the development of Davidsonís program in semantics is organized around the distinction between these three questions.†† First, I discuss Davidsonís development of the project of employing a truth theory in a central role in a recursive meaning theory for a natural language.† Davidsonís project, as first stated in ďTruth and MeaningĒ, undergoes an important but easily overlooked transformation from having the relatively modest goal of answering question 3 to that of answering simultaneously questions 1-3.† Plausibly, a truth theory that meets certain constraints can be pressed into service to answer question 3.† But the unremarked shift to the more ambitious goal imposes the requirement that the constraints not appeal to semantic facts.† This obscures a relatively straightforward answer to the question what constraints such a theory needs to meet to serve a central role in a meaning theory, if we put aside questions 1 and 2.† When Davidson turns from his first proposal about what constraints are needed to his later proposal in ďRadical Interpretation,Ē that the theory be confirmed by a radical interpreter, the effect of not having laid out the straightforward constraints is to leave unclear what the radical interpreter has to confirm to succeed.† Separating more cleanly the projects of answering question 3, and answering 1-2, helps us to evaluate the prospects of each, and, in particular, the prospects for success in radical interpretation.† Next I consider the prospects for answering questions 1 and 2 by describing the procedures a radical interpreter, who takes a recursive truth theory to give the basic structure of a meaning theory for anotherís language.† I distinguish a modest and ambitious program in Davidsonís work, and argue that Davidson is best understood as engaged in the ambitious one.† The ambitious program assumes an a priori guarantee of success at radical interpretation.† The modest program eschews this.† Failure of the ambitious program would undermine a number of important results Davidson grounds in reflection on radical interpretation.† I argue the prospects for the ambitious program are dim, and that this becomes clear once we see what success in radical interpretation requires.† Next, I consider two a priori arguments for the ambitious program suggested by some of Davidsonís recent work, and argue neither succeeds.