Abstract

In the present thesis I analyze and defend Davidson's proposal to generate a theory of meaning out of a Traski style theory of truth. The criterion of a valid theory of meaning is whether it suffices for the speaker's mastery of a language. I argue that Davidson's theory of meaning is a "full-blooded" theory in the sense that it explains the speaker's linguistic knowledge. Sentences for which there is no effective procedure to determine whether or not their truth conditions are fulfilled can be accommodated in a truth conditional semantics, provided by 'knowledge of the truth condition' we mean the ability to state the truth condition and not the ability to observe whether or not the sentence is true. ;My defense of truth conditional semantics leads me to support a version of what I call 'non-reductive extensionalism'. Taking clues from Quine and Davidson I explain the reasons for a theory of meaning on the basis of extensionalism. I show that non-reductive extensionalism does not deny "creativity" of the speaker's knowledge, nor does it make the speaker's linguistic knowledge extremely limited. Extensional truth conditional semantics does not generate bizarre T-sentences of the form "Snow is white" is true, if and only if 'grass is green' if we take into consideration the holistic structure of a theory of meaning. ;This leads me to a discussion of semantic holism. I make a distinction between extreme and moderate versions of holism and I argue for a moderate holism. I show that if a theory of meaning is supposed to represent the speaker's linguistic knowledge, then it must possess a holistic structure. Disagreement about the truth value of a sentence makes sense only in the background of a set of shared beliefs and only a holistic theory of meaning can represent the speaker's knowledge of those shared beliefs, the sentences which both hold true. I also argue that the denial of the compositional structure of a theory of meaning leads to absurdities. ;Extensionalism along with its holism leads to indeterminacy of meaning. I argue that the notion of "radical translation" or "radical interpretation" is a logically coherent notion. Radical interpretation is possible in the sense that we can have a theory of truth for the language L which will act as a theory of interpretation for L. I explain that indeterminacy thesis does not reject meaning, it only makes room for the possibility of more than one interpretation of the language. I introduce the notion of semantic supervenience and show that both extensionalists and anti-extensionalists accept semantic supervenience, but they disagree about the range of non-semantic facts. I argue that a better understanding of each other's language cannot eliminate indeterminacy. Here I bring in the "just more theory" objection to counter the arguments which attempt to restore determinacy. ;In the conclusion I explain, very briefly, some ways of extending the truth conditional semantics to different parts of natural language. Various philosophers, including Davidson himself, have worked towards that aim