A central problem in epistemology concerns the justification of beliefs about epistemic principles, i.e., principles stating which kinds of beliefs are justified and which not. It is generally regarded as circular to justify such beliefs empirically. However, some recent defenders of foundationalism have argued that, within a foundationalist framework, one can justify beliefs about epistemic principles empirically without incurring the charge of vicious circularity. The key to this position is a sharp distinction between first- and second-level justifiedness.In this paper I (...) first argue that such versions of foundationalism end up giving their approval to circular chains and are therefore unmotivated; if circular chains are acceptable, the classic regress argument for foundationalism does not go through. I then consider and reject two other ways in which the foundationalist might motivate his position. At the end of the paper I draw from this discussion a moral concerning the airns of epistemological theorizing. (shrink)