Robert Steel

Sometimes a person has not only some first-order evidence that bears directly on a proposition, but also some higher-order evidence that bears on how reliable they are at gauging this first-order evidence. This commonly happens in cases of peer disagreement; that a smart, informed interlocutor comes to an opposite conclusion is higher-order evidence that one’s judgment is unreliable. This dissertation examines how people in possession of both types of evidence ought to respond. It argues that they should adjust their level of confidence to that indicated by the higher-order evidence alone, regardless of the disposition of the first-order evidence. In defending this “calibrationist” treatment of higher-order evidence, this dissertation develops arguments on a broad range of fundamental topics in epistemology, including the relation between evidence and rational belief, the intelligibility of epistemic akrasia, and the extent to which rationality has an internal character.