Abstract

I review a familiar skeptical argument from Hume, and conclude that it requires us to accept that there is no necessary relation between beliefs about the world and their evidential grounds; that is, there is no logical or quasi-logical relation between empirical beliefs and their grounds, such that their grounds entail them, or even make them probable. I then argue that generic reliabilism can accommodate this fact about evidential grounds in a non-skeptical way. According to reliabilism, the grounds for our beliefs constitute good evidence so long as they are contingently reliable. Next I argue that agent reliabilism successfully addresses two related problems for other versions of reliabilism: the fact that beliefs can be reliably formed by accident, and the need for subjective justification. Finally, I explain why agent reliabilism is properly conceived as a version of virtue epistemology.