Logical semantics and epistemic warrant: Towards an informational account of logical consequence

The purpose of the dissertation is to show that the model-theoretic account of logical consequence does not explain how we may acquire justification for believing the conclusion of a valid argument (in, for example, English) provided only that we know the premises and recognize the validity of the inference. Chapter 1 explains the development of the model-theoretic account of logical consequence, and argues that instead of delivering an analysis of correct reasoning, it describes relations that formal language sentences bear in virtue of facts about mathematical structures. Chapter 2 shows that the model-theoretic account of logical consequence is inapplicable to natural languages, and cannot be made to apply to natural languages without thereby rendering the model-theoretic account circular. Chapter 3 shows how model-theoretically valid inferences do not preserve epistemic warrant. That is to say, the warrant that one has for believing the premises of a valid argument does not necessarily warrant belief in the conclusion. Since the model-theoretic account cannot explain the epistemic role of correct reasoning, and is inapplicable to the languages in which we inquire and infer, alternative accounts deserve renewed consideration. Chapter 4 considers the approach to semantics on which logical relations hold between sentences (or other means of representation) in virtue of the information they convey. On this view, one sentence is a logical consequence of another if and only if the information conveyed by the first is contained in the information conveyed by the second. Despite important developments by Barwise, Etchemendy, and Hintikka, none of the information-based programs solves both the problems of natural language applicability and warrant transfer. Chapter 5 develops and defends an approach that solves both problems by utilizing epistemic concepts of truth and information.