Foundationalism and the Regress Problem

Suppose a belief that p is justified only if the believer has some further justified belief, that q, which supports it. It is then natural to ask what justifies the belief that q, and if the answer is a further belief, r, the same question can be asked of r, and so on. This is the epistemic regress problem, and it appears to allow only three possible outcomes: the chain of justifying beliefs continues without end, an infinite regress; the chain loops back on itself, so that some belief eventually serves, directly or indirectly, as part of its own justification, a circle; or the chain terminates in some belief that is justified without needing support from any further belief, a foundation. This trilemma is already stated in essentially its modern form in Sextus Empiricus's Outlines of Pyrrhonism (Pyrrhoneioi Hypotyposeis, 2nd century CE), as part of the Five Modes of Agrippa, specifically the mode of infinite regress and the mode of hypothesis, which Sextus deploys for skeptical purposes, to argue that no belief can ever be genuinely justified at all, since none of the three horns seems acceptable.