Belief in Context
Belief is context sensitive. More precisely, belief is sensitive to exactly those contextual factors that epistemic contextualists claim are relevant to the semantics of words like “know”. In particular, whether an agent believes p depends on the not-p alternatives salient to the agent, and the practical importance of p for the agent. I call the view that belief is thus context sensitive sensitivism. My dissertation argues for sensitivism about both outright belief and partial belief, and outlines a sensitivist formal model of belief.
My discussion of sensitivism bears on two disparate literatures: the debate over contextualism and its rivals on the one hand, and debates over the relationship between outright and partial belief, as well as Bayesian epistemology generally, on the other. First, some standard arguments for contextualism, such as DeRose's Bank Cases, can be interpreted equally well as arguments for sensitivism. Furthermore, controlling for the effects of shifting contexts on belief can help us isolate contextual effects on knowledge; therefore, attending to the case for sensitivism (about belief) helps us better understand the case for contextualism (about knowledge attributions). Second, sensitivism about partial belief helps us give a more satisfying foundation to Bayesian epistemology than is otherwise possible, in part by reconciling partial belief with outright belief.
The dissertation has four chapters. Chapter 1 introduces sensitivism, providing some intuitive support for the view and connecting it with the debate over contextualism. Chapter 2 provides a positive argument for sensitivism: without sensitivism, we cannot make proper sense of the platitude that an assertion is sincere iff the speaker believes what she asserts. A preliminary formal model of sensitivist belief is also sketched and discussed. Chapter 3 applies the model of Chapter 2 to the preface paradox. Preface writers are seen not to exhibit inconsistency in any undesirable form. Chapter 4 concludes with an extension of sensitivism to degrees of belief. This extension allows an easy and paradox-free solution to the problem of reconciling degrees of belief with outright belief: outright belief is identified with degree of belief 1. The usual reasons for rejecting this identification lose their force in the face of sensitivism about degrees of belief.