Abstract

In his 1936 paper, "The Concept of Logical Consequence," Alfred Tarski set out the first formulation of the model-theoretic definition of logical consequence. This definition has hence become a central canon of contemporary logical inquiry. However, it has come under recent attack. In this dissertation, I present a defense of the standard model-theoretic analyses of the concepts of validity and consequence against various objections. ;In the first chapter, it is argued that the model-theoretic analysis of natural language argumentation is based on determining relationships between structural properties of mathematical models and those of the language under study. This use of model theoretic structures involves the application of mathematical models to the description of a natural language; and as such, the determination of the analysis can best be decided in the context of discussing the analysis, qua mathematical analysis. ;In the second chapter, I discuss the distinction between representational semantics and interpretational semantics. I argue that Etchemendy presents us with a false dilemma: the semantics underlying the model-theoretic analyses of the logical properties is neither a representational semantics, nor an interpretational semantics. I then formulate a precise and defensible version of Tarski's Thesis. ;In chapter 3, I take up the two main lines of argument directed against Tarski's Thesis. First, I argue that neither Etchemendy's under-generational and over-generational examples, nor McGee's reliability problem argue against the formulation of Tarski's Thesis given in this work. Second, I argue that the claims brought out in Etchemendy's discussion of the Reduction Principle and in McGee's discussion of the Contingency Problem are mistaken, and that the conclusions they draw from these claims are unwarranted. ;Finally, chapter 4 takes up historical issues surrounding both the discrepancies between Tarski's original formulation of the model-theoretic definitions and their present day incarnations, and Etchemendy's attribution to Tarski of "Tarski's Fallacy." I argue that the historical account given Tarski's 1936 paper by Etchemendy is mistaken, and that only the most narrow and unkind reading supports the fallacy attribution