Many epistemologists agree that there are things we epistemically shouldn’t believe, i.e., that we have negative epistemic duties. But it is a matter of controversy whether we have any positive epistemic duties, i.e., whether we epistemically should have certain beliefs. In this paper, I argue that, in certain cases in which one acquires counterevidence against what one believes (p), one epistemically should give up one's belief by reasoning in a way that involves forming a belief in its negation (in not-p). (...) In these cases, one epistemically should believe not-p and thus has a positive epistemic duty. I reinforce this argument by showing how our negative epistemic duties depend on our facts about how we should reason. I go on to argue for positive epistemic duties in other cases, including ones in which one trying to make up one’s mind. (shrink)