Abstract

This dissertation deals with the problem of truth and semantical paradox. It begins by motivating the premise that the Tarskian biconditionals are collectively exhaustive and individually constitutive of the concept of truth. That is, the concept of truth is completely and correctly defined by the totality of all the biconditionals of the form 'x is true iff p', where 'p' is replaced by any sentence and 'x' by any name of that sentence. ;The argument of the dissertation has three stages. First, I argue in Chapter One that the premise above entails that the concept of truth is circular. For the Tarskian biconditionals of certain sentences specify circular truth conditions for those sentences. For example, the Tarskian biconditional of the truth-generalization 'No sentence is both true and false' shows that the truth conditions of this generalization essentially involve reference to truth itself. The circularity of the concept of truth in bivalent languages gives rise to a revision process in which the semantical status of each sentence is determined by its Tarskian biconditional once a totality of nonsemantical facts and an initial extension of the truth predicate are posited. The theory whose thesis is what the preceding sentence asserts is called the revision theory of truth. ;In the second stage, I argue that the revision theory is consistent. I describe in Chapter Two a type of formal semantics called stability semantics. Since each semantical system of stability semantics is a formal representation of the revision process, the consistency of any such system demonstrates that the revision theory is consistent. ;Finally, I show that the revision theory is materially adequate. In Chapter Four I develop a system of stability semantics that represents the revision process faithfully and that yields the intuitively correct results in a representative sample of cases. I also define and study a notion of logical consequence congruous to this semantical system. In order to understand fully how this system succeeds in representing the revision process faithfully, I first examine how other systems fail. Thus in Chapter Three I give a critical exposition of Herzberger's, Gupta's, and Belnap's systems.