Testimony and a Priori Knowledge

Tyler Burge offers a theory of testimony that allows for the possibility of both testimonial a priori warrant and testimonial a priori knowledge. I uncover a tension in his account of the relationship between the two, and locate its source in the analogy that Burge draws between testimonial warrant and preservative memory. I contend that this analogy should be rejected, and offer a revision of Burge’s theory that eliminates the tension. I conclude by assessing the impact of the revised theory on the scope of a priori knowledge. The epistemic status of testimony bears on the scope of a priori knowledge in at least two ways. If testimonial knowledge is a priori, then the scope of a priori knowledge expands beyond its traditional bounds. On the other hand, the premise that testimonial knowledge is a posteriori plays a pivotal role in an influential argument against the existence of a priori knowledge.1 If this argument is sound, then the scope of a priori knowledge recedes from its traditional bounds. Tyler Burge (1993) offers a theory of testimony that allows for the possibility of both testimonial a priori warrant and testimonial a priori knowledge.2 I uncover a tension in his account of the relationship between the two, and locate its source in the analogy that Burge draws between testimonial warrant and preservative memory. I contend that this analogy should be rejected, and offer a revision of Burge’s theory that eliminates the tension. I conclude by assessing the impact of the revised theory on the scope of a priori knowledge. Section I distinguishes three questions regarding the epistemic status of testimony and identifies the primary focus of my investigation. Section II uncovers a tension in Burge’s account of the relationship between testimonial warrant and testimonial knowledge, and argues that it derives from the fact that his account of testimonial warrant is informed by two different, but incompatible, analogies: one with preservative memory and one with perception. Sections III, IV and V address three approaches to resolving the tension: rejecting the analogy with preservative memory, rejecting the analogy with perception, and reconciling the two analogies. The first approach is defended. Section VI assesses the impact of this approach on the scope of a priori knowledge. 322 E P I S T E M E 07 DOI: 10.3366/E1742360007000111 TESTIMONY AND A PRIORI KNOWLEDGE