Kevin McCain

An evidentialist can be extreme about epistemic requirements in a couple of different ways. At the reductionist end of the spectrum are those who think our epistemic obligations are fully satisfied by the mere having of evidential fit—where having implies nothing about doing. Your beliefs ought to align with your evidence, in other words, but there’s nothing you’re obligated to do in order to get yourself into the epistemically optimal position. At the expansionist end of the spectrum are those who think we ought to do more—epistemically-speaking—than just bring about a state of evidential fit. We ought to get more evidence, for instance, if it’s somehow relevant to our beliefs. Rather leery of extremists, we prefer a moderate account of epistemic obligation. It’s a conjunction of three theses: (1) We never have an epistemic duty to believe any particular proposition; (2) We’re epistemically obligated to reflect on our evidence in order to produce a state of evidential fit; (3) We’re never epistemically obligated to acquire any additional evidence. Our aim is to make these plausible.

In this paper we distinguish between epistemic dilemmas, epistemic quasi-dilemmas, and quasi epistemic dilemmas. Our starting point is the commonsense position that S faces a genuine dilemma only when S must take one of two paths and both are bad. It’s the “must” that we think is key. Moral dilemmas arise because there are cases where S must perform A and S must perform B—where ‘must’ implies a moral duty—but S cannot do both. In such a situation, S is doomed to violate a moral obligation. Analogously, S face’s a genuine epistemic dilemma only if she has an epistemic duty to believe that p and at the same time an epistemic duty to not believe that p (either because she has a duty to believe ~p or a duty to suspend judgment). We argue that such cases never arise because there is no epistemic duty to adopt a particular doxastic attitude towards any particular proposition (ever). Hence, there are no epistemic dilemmas. Nevertheless, one might suppose there are situations in which one’s evidence pulls in different directions without determining a uniquely justified doxastic attitude. Such epistemic quasi-dilemmas aren’t full-fledged dilemmas, but they are arguably close enough to pose a genuine problem for the believer. But we argue that there are no genuine epistemic quasi-dilemmas either: They are all resolvable in principle. Fans of dissonance will not be totally disappointed, however, since we argue that there may well be quasi-epistemic dilemmas. These are genuinely unwinnable situations in which one’s moral or pragmatic obligations to do epistemically-relevant things make incompatible demands.