Abstract

Every epistemological theory needs to be able to articulate some version of the following principle: If S's belief "q" is to make S's belief "p" justified (or is to make "p" something S knows), then "q" must possess some positive epistemic merit. This paper argues that naturalizing epistemologies do not have access to this principle. The central problem is that of providing a naturalistic account of the notion of a reason-for-which one believes while avoiding internalist commitments. The discussion, which focuses on the work of Alvin Goldman, is part of a general argument against causal accounts of reasons-for-which.