Versions of naturalized epistemology that overlook or reject apriority ignore innate belief-forming processes that provide much of the grounding for epistemic warrant. A rigorous analysis reveals that non-experiential ways of viewing apriority, such as innateness, establish the domain for a plausible naturalistic theory of a priori warrant. A moderate version of naturalistic epistemology that embraces the non-experiential feature of apriority and motivates future cognitive scientific research is the preferred account.

A glass couldn't contain water unless it contained H2O-molecules. Likewise, a man couldn't be a bachelor unless he was unmarried. Now, the latter is what we would call a conceptual or analytical truth. It's also what we would call a priori. But it's hardly a conceptual or analytical truth that if a glass contains water, then it contains H2O-molecules. Neither is it a priori. The fact that water is composed of H2O-molecules was an empirical discovery made in the eighteenth century. The fact that all bachelors are unmarried was not. But neither is a logical truth, so how do we explain the difference? Two-dimensional semantics is a framework that promises to shed light on these issues. The main purpose of this thesis is to understand and evaluate this framework in relation to various alternatives, to see whether some version of it can be defended. I argue that it fares better than the alternatives. However, much criticism of two-dimensionalism has focused on its alleged inability to provide a proper semantics for certain epistemic operators, in particular the belief operator and the a priori operator. In response to this criticism, a two-dimensional semantics for belief ascriptions is developed using structured propositions. In connection with this, a number of other issues in the semantics of belief ascriptions are addressed, concerning indexicals, beliefs de se, beliefs de re, and the problem of logical omniscience.