There are a number of relevant alternatives accounts of knowledge in the literature, including those by contextualists (like Lewis and Cohen), and invariantists (like Dretske). Despite widespread discussion of such views, an explication of the notion of relevance is conspicuously absent from the literature. Without a careful explication of that notion, relevant alternatives accounts resist evaluation. This paper attempts to aid in the evaluation of those accounts, by providing an account of relevance. The account rejects two common presuppositions about the (...) notion of relevance. The account holds that worlds, rather than alternatives, are relevant, and that distant worlds can be relevant. Relevant worlds turn out to be those worlds at which an alternative to one’s belief obtains, and is such that one’s epistemic position (with respect to what one believes at the actual world) is worse than it is at the actual world. (shrink)