Abstract

I address the question of giving a clear account of knowledge which is "independent of experience". After clarifying the project and its motivation, I briefly consider some familiar accounts of apriority--in terms of innateness, intuition, certainty, psychological determinism, linguistic convention, and Wittgenstinian "forms of life"--and argue that each is defective. ;Next, I consider Philip Kitcher's analysis of a priori knowledge along reliabilist lines. His analysis suffers from unclarities associated with his "psychologism" and is vulnerable to a reductio, but one of its guiding ideas can be divorced from its reliabilist roots to yield a good account of "minimally" a priori justification: that which does not depend on states external to the knower. ;I explain the dependence in question and argue that such justification has philosophical importance. "Knowledge of necessary truths based on minimally a priori justification" comes very close, I argue, to giving extensionally adequate necessary and sufficient conditions for a priori knowledge as traditionally conceived, though it is not intensionally adequate. I then consider "fully" a priori justification: a justification is fully a priori just in case it is minimally a priori and contains no beliefs which entail the existence of any sensory experience or experiencer. It is plausible that knowledge traditionally considered a priori is knowable on the basis of such justification, if at all. ;I conclude by considering Kripke's examples of allegedly contingent a priori truths, and argue that they fail to meet the requirement of even minimal apriority as originally stated; at best, they can be modified so as to meet this requirement, but not that of full apriority. Thus I succeed in giving a clear definition of 'a priori knowledge' which does not trivially entail, but which renders plausible, the view that only necessary truths can be known a priori