![Sample our Humanities journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5939/TOC-Subject-Sample-2021-Humanities.jpg"})
ABSTRACT
Phenomenal conservatism (PC) is the internalist view that non-inferential justification rests on appearances. PC’s advocates have recently argued that seemings are also required to explain inferential justification. The most developed view to this effect is Michael Huemer ’s theory of inferential seemings (ToIS). Luca Moretti has recently shown that PC is affected by the problem of reflective awareness, which makes PC open to sceptical challenges. In this paper I argue that ToIS is afflicted by a version of the same problem and it is thus hostage to inferential scepticism. I also suggest a possible response on behalf of ToIS’s advocates.
Disclosure statement
No potential conflict of interest was reported by the author.
Notes
1. For more detailed introductions to PC, see Tucker (Citation2013), Moretti (Citation2015) and Huemer (Citation2018).
2. JS has been introduced in Huemer (Citation2001) and it is the principle actually at stake in most conversation about PC. Huemer (Citation2007) defends a variant of JS according to which if it seems to S that P, in the absence of defeaters S has at least some degree of justification for believing P. Huemer fears that a weak and wavering seeming that P couldn’t give S justification sufficient to believe P (cf. [Citation2007, 30n1]). To get round this difficulty, whenever I speak of seemings or appearances, I always mean clear and firm seemings. JS concerns propositional justification but it can easily be re-formulated to apply to doxastic justification (see for instance Huemer Citation2018).
3. For instance, perceptual, a priori, moral, and mnemonic beliefs (cf. Moretti Citation2015).
4. For other asserted virtues of PC see Tucker (Citation2013).
5. PC has been targeted by various arguments but it is dubious it has been lethally struck. For objections and responses see Tucker (Citation2013), Moretti (Citation2015) and Huemer (Citation2018).
6. A referee of this journal suggests that PC’s advocates might try to reject this claim by insisting that:
(*)S’s conceiving of a deceptive scenario SH incompatible with P need not defeat S’s justification for P, unless S has reason to think that SH obtains or is likely to obtain.
If (*) is correct, even if S lacked independent justification for taking SH to be false or improbable, S could still have justification for believing P. The referee helpfully indicates that a general problem of (*) is that it enables bootstrapping reasoning, which is counterintuitive. In short, once S acquires justification for believing P and competently deduces Not-SH from P, S will acquire justification for believing Not-SH, and thus disbelieving SH. This appears to be a too easy way to earn justification for ruling out SH. (I return to bootstrapping in §6.) Another problem of (*) is that it has elementary counterexamples. Suppose S conceived of a deceptive scenario SH incompatible with P and found equal reasons for SH and Not-SH. (Imagine that S found out that SH and Not-SH have equal chance to obtain.) It is intuitive that S’s appearance-based justification for believing P would normally be defeated in these circumstances, though S wouldn’t have reason to think that SH obtains or is likely to obtain (cf. Wright Citation2007; Pryor Citation2018). Irrespective of these considerations, (*) appears false if one considers carefully the predicament of S in the thought experiments illustrating the problem of reflective awareness. When S comes to entertain the belief that she has a seeming that P and conceives of an array of alternative and incompatible explanations of her seeming, S could rationally believe one of these explanations only if she had independent reasons to disbelieve all the others. S’s selecting one specific explanation without possessing such reasons would be arbitrary. This looks straightforward. Consequently, S could justifiedly believe the specific explanation according to which she perceives that P (and so she could justifiedly believe P) only if she had independent justification for taking the other error explanations to be false or improbable.
7. ToIS was already presupposed in some of Huemer’s earlier work – for instance, Huemer (Citation2001, 112–113; Citation2007, 30n1).
8. An inferential seeming doesn’t merely represent that E and P stand in a certain logical or epistemic relation to one other. Huemer fears that a seeming of this type would engender a variant of Lewis Carroll’s infinite regress problem (cf. [Citation2016, 146–147, 152–153]). For Huemer, an inferential seeming is rather one that represents that P is true or probable under the assumption that E is true (cf. [Citation2016, 149–150]).
9. A robust defeater is one capable of making P completely unjustified (cf. [Citation2016, 151]). If we focus on plain justification rather than degrees of it, the qualification ‘robust’ is irrelevant.
10. I follow Huemer (Citation2016) in taking the verb to see not to be necessarily factive in this context.
11. It is notoriously difficult to give a characterization of a deviant causal link. Huemer provides this example: S justifiedly believes E and sees that E supports P, but she refuses to believe P because this makes her unhappy. Nevertheless, an epistemically benevolent brain scientist detects S’s justified belief that E and S’s inferential appearance. And this detection causes the scientist to use a brain-manipulation tool to induce in S a belief that P. Clearly, S’s belief that P isn’t in this case inferential.
12. Where E includes everything that S takes to be relevant to P’s truth.
13. In his First Meditation Descartes introduces the conjecture of a deceiving God, then reformulated as the demon conjecture, which raises the possibility that we might constantly be deceived in our elementary arithmetical reasoning. It is easy to extend this conjecture to threaten all our elementary inferences.
14. A referee of this journal suggests that advocates of ToIS might respond that in order to rationally believe P, S need not have independent justification for ruling out the error conjectures she has conceived of, at least so long as S doesn’t have a seeming that an error conjecture is likely. I find this possible response misguided. In the circumstances I’ve described, S can rationally believe P on the basis of E only if she has an epistemic reason to prefer the hypothesis that her inferential seeming is veridical to any alternative error hypothesis she has conceived of. To do so, S does need independent justification for ruling out all these alternatives.
15. Hereafter, I leave reference to B implicit.
16. A controversial hypothesis, which Huemer (Citation2016) nevertheless appears to flirt with, is that non-deductive inferential appearances depend on a faculty of S that tracks logical probabilities. ‘Probability’ could also be interpreted subjectively. In this interpretation, inferential seemings are not requested to track logical relations or objective chances. However, note that subjective probabilities must be coherent – they must conform to the probability calculus. In this interpretation, inferential seemings must depend on some cognitive faculty that guarantees this (weak) form of objectivity. This opens the door to error conjectures.
17. See also Fumerton (Citation1995, 187–190).
18. Although Huemer (Citation2016) doesn’t explicitly state arguments of this type, the final paragraph of his paper indicates that he has them in mind.
19. In the above quotation Huemer writes that ‘we are justified in trusting our appearances’ unless we acquire grounds to distrust them. Huemer is speaking loosely here. For one of the celebrated strengths of appearances is that their justifying power doesn’t require justification for trusting them. Huemer must mean that our appearances have justifying power by default unless we have reasons to distrust them.
20. I’m grateful to a referee of this Journal for inviting me to consider this issue and suggesting possible responses.
21. For general discussion about bootstrapping and epistemic circularity see for instance Cohen (Citation2005; Cohen Citation2010), Weisberg (Citation2012) and Lammenranta (Citation2018).
22. It is quite natural to presume that the advocates of ToIS would endorse PC.
23. Some of these authors claim that PC is problematic because it enables bootstrapping and epistemically circular reasoning. I doubt this is true. I think that a subject S couldn’t competently use inferences of these types to try to rule out error conjectures without falling prey to the problem of reflective awareness and thus losing her initial non-inferential justification. So it appears true that PC doesn’t enable bootstrapping or epistemically circular reasoning.
24. For example, S might have an inferential appearance that Not-SH is probable in light of SH’s being far-fetched and overcomplicated, or a non-inferential appearance that Not-SH is probable relative to S’s scientific background information.
25. Precisely, according to Wright, we are entitled to accept certain hinge propositions (cornerstones or presuppositions) inclusive of these logical negations. Sceptical conjectures are characterized by Wright as error or deception hypotheses such that the evidence in favour or against them can only be none or very little at best.
Additional information
Notes on contributors
Luca Moretti
Luca Moretti is a Reader at the Department of Philosophy of the University of Aberdeen and a (five-year) Visiting Professor at the Munich Center for Mathematical Philosophy. His current areas of investigation are general epistemology, social epistemology, formal epistemology, philosophical logic and metaphysics.