Appreciating the need for autonomy, or recognizing the truth of evidentialism?

ABSTRACTIn this article I raise two challenges for the primary argument of Adam Carter's Autonomous Knowledge. The first is that his argument for thinking that internalist theories of epistemic justification face problems from TrueTemp-style examples is unsound. The second is that the autonomy condition that Carter argues for is not new as it seems to be already implicit in Earl Conee and Richard Feldman's account of well-founded belief. Ultimately, I do not see these challenges as undermining the value of Carter's project. Rather, these challenges help illuminate that what Autonomous Knowledge does is provide a powerful argument for Evidentialism.KEYWORDS: autonomyAdam CarterEarl ConeeEvidentialismRichard FeldmanTrue-Temp Disclosure statementNo potential conflict of interest was reported by the author(s).Notes1 More precisely, the principle says: “for any version of a TrueTemp case, T, and any kind of epistemic justification condition, EJ, on knowledge that we might appeal to explain why Mr. TrueTemp fails to know in T, there will be a possible variation on T, T*, in which Mr. TrueTemp continues to fail to know and yet EJ is satisfied.”2 See, in particular, their Citation2004 collection of essays and their Citation2011 replies to critics. See also McCain (Citation2014) for development and defense of this sort of mentalist Evidentialism. Although there are externalist versions of Evidentialism (see Bergmann Citation2018 for helpful discussion), throughout this article I will use the term “Evidentialism” to refer to the internalist variety.3 There are various answers to what sorts of mental states in particular count as evidence. For example, as discussed below, Conee and Feldman (Citation2008) claim that experiences of various kinds are ultimate evidence, and beliefs justified by those experiences are intermediate evidence. McCain and Moretti (Citation2021) agree, though they offer a more precise take on which experiences count as ultimate evidence by insisting that only seemings/appearances are evidence.4 There are different views of what is required for a belief to be held on the basis of a given body of evidence. Some insist that one must have a meta-belief to the effect that the evidence sufficiently supports the belief (see for example, Lehrer Citation1971; Tolliver Citation1982; Foley Citation1987). Others insist that one’s evidence must cause the belief in the appropriate way (see, for example, Moser Citation1989; Pollock and Cruz Citation1999; McCain Citation2012). Still others suggest that one must simply satisfy one or the other of these conditions (see Korcz Citation2000; Mittag Citation2002 for criticism). Fortunately, we don’t have to settle the issue here because TrueTemp* doesn’t satisfy any of these conditions with respect to his belief that q.5 A number of Evidentialists are explicit in claiming that ultimate evidence is experiential, and beliefs are only intermediate evidence. In particular, they insist that the only time a belief is evidence is when that belief is itself justified. See, for example, Conee and Feldman (Citation2008), McCain (Citation2014), and McCain and Moretti (Citation2021).6 I am assuming that the reason that TrueTemp* has for believing that p isn’t undermined by other evidence he has and that he doesn’t have equally good reasons for believing ∼p. In other words, I’m granting in this case that, all things considered, the evidence TrueTemp* possesses actually provides propositional justification for believing that p.7 See McCain (Citation2012) for elaboration and defense of this account of the basing relation. Other causal accounts of the basing relation, such as Moser (Citation1989) and Pollock and Cruz (Citation1999), seem to deliver the same result that TrueTemp*’s belief is reasons-responsive. The sort of dispositional account of basing defended by Evans (Citation2013) appears to have this consequence as well.8 Other Evidentialists deny that the appreciation described by Conee and Feldman is required for evidence to support believing a proposition (see, for example, Smithies Citation2019; McCain and Moretti Citation2021). However, these Evidentialists allow that such appreciation is a necessary component of basing, which is required for doxastic justification. Hence, on these alternative Evidentialist pictures of evidential support bypassing or preempting beliefs would still fail to be well-founded.9 It’s worth noting that if we understand Conee and Feldman’s conditions for well-founded belief as an epistemic justification condition in the sense the term is used in the Iteration Principle, then we have additional reason to doubt that principle. After all, a belief that satisfies Conee and Feldman’s requirements for well-founded belief will also satisfy the conditions of Carter’s HSEEA***. Since satisfying HSEEA*** rules out being in a TrueTemp case, it won’t be possible for a TrueTemp character to satisfy Conee and Feldman’s conditions for well-founded belief and fail to have knowledge.10 Thanks to Jon Matheson and an anonymous reviewer for helpful comments on an earlier draft.