In this chapter I address an underappreciated problem for reliabilism. The problem concerns how a reliabilist approach in epistemology ought to understand defeating evidence. In more colloquial language: How should reliabilism understand the notion of counterevidence, or evidence against one's beliefs? The problem arises because we have certain ­pre-theoretical intuitions about counterevidence, and a reliabilist approach in ­epistemology seems inconsistent with these intuitions.

THE PROBLEM

Here is a quick statement of the problem. Suppose that S believes that p and that S's believing that p is the result of a highly reliable process. But suppose also that S has a lot of misleading evidence against p, even evidence that entails not-p. In such a case, it would seem, any positive epistemic status that S has for her belief is defeated by this counterevidence. But how can reliabilism account for this? What should matter for knowledge on a reliabilist account is that S's believing that p results from a reliable process. It should not matter that other things S believes count against p. After all, reliabilism is a form of externalism. What makes for knowledge on a reliabilist account are the facts about the genesis of S's belief. And in the case under consideration, S's belief is in fact reliably formed.