Traditionally, philosophers have argued that epistemology is a normative discipline and therefore occupied with an a priori analysis of the necessary and sufficient conditions that a belief must fulfill to be acceptable as knowledge. But such an approach makes sense only if human knowledge has some normative features, which conceptual analysis is able to disclose. As it turns out, philosophers have not been able to find such features unless they are very selective in their choice of examples of knowledge. Much of what we intuitively think functions as knowledge, both in human and non-human animals, does not share these normative features.