One goal of epistemology is to refute the skeptic. Another, with an equally dist- guished if briefer pedigree, is to make sense of science as a knowledge-acquiring enterprise. The goals are incompatible, in that the latter presupposes that the skeptic is wrong. The incompatibility is not strict. One could have both goals, conditi- ing the latter upon success at the former. In fact, however, epistemologies aimed at the skeptic tend not to get anywhere near science. They've got all they can handle guring out how we can know we have hands. I come to epistemology from the philosophy of science, my original interest in which was epistemological. Philosophers of science are concerned with epistemic justi cation, but their question about it is how far it extends. They take justi cation to be unproblematic at the level of ordinary experience; their worries begin with the interpretation of experience as evidence for theory. They are interested in the scope of scienti c knowledge. Having taken a position on this question (1997), - guing that justi cation extends to theoretical hypotheses, I came to wonder about the nature of justi cation generally. This is not a belated discovery of the skeptical problem or a reconsideration of what I took to be unproblematic. It is simply an interest in the possibility of locating epistemic advance in science within a broader understanding of the nature of epistemic justi cation. Now that I know that just- cation extends to theory, I am taking a step back and asking what justi cation is.