1 Introduction

Evidentialism is, roughly, the view that a subject’s “doxastic attitude towards any given proposition is determined by [their] evidence” (1). This view might seem so intuitive that it hardly needs defense.¹ Nevertheless, a main recent source of resistance to evidentialism is a concern about its potential skeptical implications (after all, doesn’t one need evidence for one’s evidence, and so on, yielding a regress?). Accordingly, many of the contributions to Non-Evidentialist Epistemology share the goal of assessing whether and how a denial of evidentialism can contribute to a plausible response to skepticism. This is an essential work for those interested in the latest developments in this recent branch of epistemology.

A strength of the volume is its first part, devoted entirely to hinge epistemology. ‘Hinges’ are usually construed as themselves lacking evidence, although they support our practices of evidence-based epistemic assessment. In this part, we see three contributions that engage directly with a prominent version of hinge epistemology championed by Annalisa Coliva, Extended Rationality. This is followed by Coliva’s response to each. Though I do not find Coliva’s responses entirely convincing, the criticisms raised and her responses undeniably push this field forward by developing new stances on the relation of hinge epistemology to epistemic relativism and to theories of truth, and in its anti-skeptical ambitions.

The second part of the volume is devoted to criticisms of evidentialist and non-evidentialist epistemology. Here, the focus is on two epistemological theories commonly thought to comprise non-evidentialist positions: (1) Crispin Wright’s Entitlement Theory, according to which we have non-evidential warrant to accept the presuppositions of certain cognitive projects, and (2) Epistemic Conservatism, according to which merely having a belief confers some positive epistemic status on its content. Two important themes emerge in this section regarding the viability of a non-evidentialist epistemology: first, the use of epistemic consequentialism as a framework for justifying the rationality of accepting certain propositions without evidence; and second, a discussion of the nature of evidence itself.

The third part of the volume considers exciting extensions of non-evidentialist epistemology beyond its standard use in dissolving the problem(s) of radical skepticism. Here, non-evidentialism is applied in analyzing stereotype beliefs, delusions, and mathematical knowledge. Notably missing, however, is discussion of some prominent approaches to basic self-knowledge. One of the characteristic marks of basic self-knowledge (e.g., my knowledge that I want some more coffee) is its apparent epistemic baselessness (see, e.g., Bar-On 2004). The literature on self-knowledge is thus arguably an area of epistemology where non-evidentialism is already well established as a leading view.

In what follows, I discuss the contributions to the volume, some in a brief exegetical spirit, others in more critical detail.

2 Hinge Epistemology

Luca Zanetti, in “Transcendental Hinge Epistemology,” makes a distinction between ‘escapable’ and ‘inescapable’ hinges. The former are intellectual commitments one could rationally doubt without having to take that very commitment itself for granted in doing so. The latter are intellectual commitments that cannot be doubted without assuming their truth. Zanetti argues that only inescapable hinges can support an attractive modest anti-skeptical position, and that because Coliva’s Extended Rationality only countenances escapable hinges, it fails to support this kind of anti-skepticism.

In response, Coliva argues that a transcendental hinge epistemology would face the difficult question of explaining why the self-stultifying character of doubts about inescapable hinges should give us any reason to think that those hinges are true. It remains somewhat unclear, however, in virtue of what Coliva’s hinge epistemology escapes this difficult question. Coliva suggests that we ‘go local’ about hinges and then address this worry by rejecting correspondence intuitions about local hinge truth in favor of a deflationary approach. But the reader (this reader, anyway) is left wondering: why is this general deflationary strategy not also available to the proponent of inescapable hinges?

Natalie Alana Ashton, in “Extended Rationality and Epistemic Relativism,” takes up the issue of the relation of hinge epistemology to epistemic relativism. Ashton criticizes Coliva’s arguments against epistemic relativism, and furthermore suggests that Coliva’s own Extended Rationality can be adapted into a plausible form of epistemic relativism. A major contribution of this paper is the care with which Ashton critically engages Coliva’s assumptions about relativism, with the compelling result that, when relativism is properly understood, hinge epistemologists should gladly embrace the relativist leanings of their view.

Coliva takes issue with Ashton’s construal of what a plausible form of relativism would look like. For instance, Ashton, following Kusch (2016), takes there to be a problem with proposing that different systems could be equally valid from a relativist perspective. Claiming that different systems are ‘equally valid’ appears to presuppose a neutral position from which to assess the validity of various systems, and the availability of such a neutral position is clearly antithetical to the core relativist idea. Coliva responds that equal validity is “at the core” of most relativist positions (103); it is a key component of genuine relativism and so relativists are tasked with making good sense of it—and if they can’t, so much the worse for relativism. However, Ashton is interested in the most plausible versions of relativism. If there is a conceptual problem with equal validity, and there are alternative versions of relativism that do not endorse it, we should devote attention to those versions. And it does seem possible to articulate a genuinely relativist view along these lines; relativists should refuse to subject alternative systems to a neutral ranking system, regardless of whether such a ranking would return the verdict that they are all equally valid.

Sebastiano Moruzzi, in “Hinge Epistemology and Alethic Pluralism,” considers the relation of hinge epistemology to alethic pluralism. Moruzzi’s particular concern is with what Coliva’s Extended Rationality should say about the truth property for hinge propositions. Coliva (2018) maintains a general commitment to alethic pluralism but proposes that hinges have only deflationary truth. Moruzzi argues, contrary to Coliva’s general pluralist stance, that the best option for Extended Rationality is to endorse monist alethic deflationism. If correct, Coliva’s view would be committed to deflationism about truth across the board (since she is committed to deflationism about hinges).

Coliva takes up the question of whether deflationism is appropriate to hinges. One general obstacle to deflationism (entertained by Moruzzi) comes from the Inflationary Argument (Wright 1992), which relies on the point that truth and warranted assertability “do not commute in the same way when embedded in negation” (110)—this is problematic for deflationist views on which truth ascriptions are merely endorsements of some content as warrantedly assertible. Coliva contends that the inflationary argument does not even arise for hinges, for neither hinges nor their negations are appropriate candidates for warranted assertion, so (trivially) predicating truth of them is not equivalent to endorsing them as warrantedly assertible.

It seems to me that this response is at best a pyrrhic victory. Given hinge deflationism, if hinges and their negations can never be warrantedly assertible, it seems they would not even be truth-apt, contrary to the propositional framework Coliva prefers. Coliva suggests that an alternative conception of deflationism may be needed; hinges may possess simply plain truth (Lynch 2013)—their truth is “exhausted by the Equivalence Schema and similarly harmless platitudes” (115), though unlike other instances of plain truth (such as logical truths), hinges are often only contingently true.

However, prima facie, hinge propositions appear to belong to a variety of domains, suggesting, given a general alethic pluralist background, that they can have different truth properties. This is in tension with Coliva’s contention that hinges are just plainly true. Also, on the assumption that the truth-property of a proposition is essential to it, we would have it that all hinges are essentially plainly true. This forecloses the possibility that a proposition might be a hinge commitment relative to one individual, or one area of inquiry, but not relative to another. This would conflict with the idea that hinges are such in virtue of the role they play in the cognitive economy of individuals or in the functioning of areas of inquiry.

3 Criticisms of Evidentialist and Non-Evidentialist Epistemology

Luca Moretti, in “Problems for Wright’s Entitlement Theory,” (unsurprisingly) poses problems for Wright’s Entitlement Theory. Wright’s entitlements are a form of non-evidential and unearned warrants to accept a proposition. Entitlements are particularly useful in responding to radical skeptical arguments: it can be argued that skeptical arguments simply ignore this category of warrant, mistakenly assuming that if we cannot justify our basic epistemic commitments by appeal to evidence, we cannot rationally claim warrant for them at all.

Moretti presses two objections. I discuss only the first, which targets strategic entitlement (one of four varieties Wright countenances). Strategic entitlements are warranted because accepting them is a dominant strategy for agents interested in holding true beliefs. Against this dominant strategy vindication of entitlements, Moretti points out (following Pedersen 2009) that such vindication crucially depends on the assumption that forming true beliefs is our only epistemic goal. If we are generally as interested in avoiding false beliefs as we are in forming true ones, the dominant strategy vindication fails. Moretti considers the possibility (raised in Pedersen 2020) that there are yet more epistemic goals we should consider, such as coherence, that would favor accepting entitlements, but he rejects this on the grounds that such further epistemic goals are valuable only insofar as they contribute to gaining true beliefs and avoiding false ones.

Junyeol Kim, in “Epistemic Entitlement: Intellectual Desires and Epistemic Rationality,” raises the question whether epistemic rationality depends on intellectual desire. Kim argues that Wright’s notion of epistemic entitlement leads to a positive answer. It is significant that in defending the connection between entitlements and intellectual desires, Kim establishes that entitlement theorists must endorse a subjective conception of epistemic telos (assuming a teleological explanation of non-evidential warrant).² Applying this subjective conception of epistemic telos in articulating entitlement theory yields the result that whether S is entitled to trust that p depends on whether S desires the epistemic accomplishment of the project for which p is a presupposition. The result is that entitlement candidates are not necessarily propositions that all epistemic agents are entitled to simply in virtue of being epistemic agents. It is worth considering whether this result leads to a kind of epistemic relativism.

Kevin McCain, in “Epistemic Conservatism: A Non-Evidentialist Epistemology?,” explains the relationship between two positions: Epistemic Conservatism (ec)—roughly, the idea that merely believing that p confers some minimal positive epistemic support on p—and evidentialism. The question of whether ec is compatible with evidentialism comes down to whether S’s belief that p can constitute evidence in favor of p for S. McCain defends the answer that yes, it can. A key step in defending this initially surprising claim is to clarify that the sort of evidence belief that p provides for p is quite minimal—as William Lycan puts it, “vanishingly close to zero” (quoted on p. 155).

Tommaso Piazza’s “Weak Non-Evidentialism” continues with many of the themes animating McCain’s contribution. Both authors take up the question of the nature of evidence, considering in some detail whether the ontology of evidence is best understood in terms of beliefs, or in terms of propositions, and the relation of ec to evidentialism. On these issues, they take conflicting views: McCain prefers psychologism about the ontology of evidence, whereas Piazza argues for propositionalism; and Piazza assumes that evidentialism is incompatible with ec, contrary to McCain’s argument.

Overall, Piazza argues that evidentialism is appropriate for inferentially justified beliefs, but not apt to explain the justification for perceptual beliefs. First, he argues that a psychologist ontology of evidence (on which evidence is constituted by mental states) would, if correct, be able to explain perceptual justification consistently with evidentialism, but not inferential justification. By contrast, propositionalism (= evidence is constituted by propositions) would be able to explain inferential justification (consistently with evidentialism), but not perceptual justification. The evidentialist then, cannot have it both ways; whichever ontology of evidence she chooses (between psychologism and propositionalism), there is a category of justification that remains unexplained. Second, Piazza provides a novel general argument in favor of a propositionalist ontology for evidence, resulting in a limited (weak) form of evidentialism, one that applies to inferential but not perceptual belief.

4 Extensions of Non-Evidentialist Epistemology

The third part of this volume contains extensions of non-evidentialist epistemology beyond its typical use as a strategy for addressing radical skepticism.

Ann Meylan, in “Radical Skepticism, Stereotypes, and the Pragmatist Stance” argues, contrary to philosophical dogma, that practical considerations do ordinarily sometimes count as reasons for/against holding certain beliefs. Specifically, moral considerations provide reasons against holding beliefs whose content encodes a stereotype about a social group, even where stereotype beliefs have some evidential support. Meylan utilizes this point to derive a conclusion concerning radical skepticism. Epistemologists typically reject practical reasons as relevant at all to addressing the skeptic’s arguments. But, Meylan suggests, if practical reasons can defeat epistemic ones for stereotype belief, why not also for commonsense/anti-skeptical belief? While this idea initially appears promising, on reflection it appears too hasty a conclusion. In many cases, practical reasons clearly do not defeat epistemic reasons for belief. Instances of wishful thinking come to mind, where one’s desire for some proposition to be true irrationally leads one to form the belief that it is true. We need a principle for determining when and why practical considerations defeat epistemic ones before we can conclude that they do so in the case of commonsense and anti-skeptical beliefs.

Jakob Olhorst, in “The Certainties of Delusion,” makes a case for thinking of delusions (and hinges) as subcategories of certainty; that is, as beliefs held with such a high degree of conviction that no other belief or evidence would change this degree of conviction. In the case of hinges, this evidential insensitivity is explained by the role they play in our cognitive lives; hinges inform how we interpret evidence, and what gets to count as evidence, and so are not themselves assessable in evidential terms. To the extent that hinges are distinct from delusions, it is because hinges underly much of our cognitive lives whereas delusions are more localized. As Olhorst frames it, delusions are typically bizarre beliefs that are added on to an otherwise normally functioning cognitive system. But Olhorst acknowledges room for some overlap: a delusion can be a hinge when it infects enough of the victim’s total belief system.

Nikolaj Jang Lee Linding Pedersen, in “Cornerstone Epistemology: Scepticism, Mathematics, Non-Evidentialism, Consequentialism, Pluralism,” presents an argument against a novel form of mathematical skepticism. While regress arguments are common in mainstream epistemology, they have not been much explored regarding mathematical knowledge. Leveraging Gödel’s Incompleteness Theorem in the context of first-order logic, Pedersen derives a semantic version of the theorem concerning the satisfiability of Peano arithmetic (pa). The result, he argues, is that the satisfiability of pa cannot be proven within pa, nor within a subsystem of it, nor any other theory of equal strength. The only remaining anti-skeptical option is to prove the satisfiability of pa within a theory of greater strength. But then the very same concerns about proving satisfiability arise for this stronger theory, and a regress looms.

Pedersen addresses this mathematical skeptical argument with a non-evidentialist epistemology. The idea is that, while we lack evidence (in the form of proof) for the satisfiability of Peano Arithmetic, we are still rational to claim warrant for it, because accepting it maximizes epistemic value. In support of this, Pedersen argues for a pluralist epistemic consequentialism, in which rational acceptance of a proposition is determined by appeal to the distinct epistemic goals of (i) forming true beliefs, (ii) avoiding error, and (iii) achieving “meta-cognitive coherence.” Insofar as the satisfiability of Peano Arithmetic is a cornerstone of mathematics, we have warrant to accept it in mathematical theorizing even in the absence of evidence.

5 Conclusion

In concluding, I would like to make a general comment on the epistemic consequentialist stance that figures in several contributions to this volume. The motivation for epistemic consequentialism in the context of non-evidentialist epistemology comes from the thought that there must be some plausible positive answer to the question: “In virtue of what is it rationally permitted to accept a proposition for which one cannot have evidence?” I think non-evidentialists have been too concessive to evidentialist intuitions by accepting that a positive answer to this question is required. I would recommend a more radical form of non-evidentialism that answers this question with: “Nothing.” I see this radical form of non-evidentialism as having a basis in Pritchard’s version of hinge epistemology, an instance of non-evidentialism under-discussed in the volume. According to Pritchard (2016), hinges are removed from direct rational evaluation altogether; indeed, the lack of a positive epistemic status for hinges is necessary to their functioning as hinges. The availability of a compelling radical form of non-evidentialism undercuts the motivation for epistemic consequentialism.³

Overall, despite the prima facie counterintuitiveness of non-evidentialist views, the sophisticated responses to skepticism they make available renders them, and Non-Evidentialist Epistemology, well worth serious consideration for anyone concerned with skepticism, evidence, hinge epistemology, and entitlement. The contributions to this volume are each admirably rigorously argued, and collectively center the relatively young literature on non-evidentialism around a program of related themes and concerns (albeit with some notable omissions).