Memory, Anaphora, and Content Preservation

Tyler Burge defends the idea that memory preserves beliefs with their justifications, so that memory's role in inference adds no new justificatory demands. Against Burge's view, Christensen and Kornblith argue that memory is reconstructive and so introduces an element of a posteriori justification into every inference. I argue that Burge is right, memory does preserve content, but to defend this view we need to specify a preservative mechanism. Toward that end, I develop the idea that there is something worth calling anaphoric thinking, which preserves content in Burge's sense of "content preservation." I provide a model on which anaphoric thought is a fundamental feature of cognitive architecture, consequently rejecting the idea that there are mental pronouns in a Language of Thought. Since preservative memory is a matter of anaphoric thinking, there are limits on the analogy of memory and testimony.