2. The Framework Reading

The title of “hinge epistemology” is often used somewhat broadly to include a constellation of antiskeptical strategies inspired by Wittgenstein’s remarks in On Certainty, which developed into a wide variety of neighboring epistemological views—from Wright (Reference Wright1985, Reference Wright2004) and Williams’ (Reference Williams1991) entitlement (or epistemic) views, Conant (Reference Conant1998) and Strawson’s (Reference Strawson1985) therapeutic accounts, as well as Coliva (Reference Coliva2015, Reference Coliva2016), Pritchard (Reference Pritchard2015), and Moyal-Sharrock’s (Reference Moyal-Sharrock2004, Reference Moyal-Sharrock and Brenner2005) nonepistemic views.Footnote ⁶ Very roughly, this family of views can be seen as originating from a common desire to capture in full the richness of Wittgenstein’s insightful intuitions. One of these in particular stands out for having inspired the christening of this epistemological tradition:

From this shared point of departure, different hinge epistemologies can be distinguished based on the way they characterize (a) the features they attribute to these special “hinge propositions” (or hinges for short), and (b) the role they take hinges to play with respect to the rational system of which they are part. In what follows, I will be primarily concerned with one of the most popular formulations of hinge epistemology to date, known as the “framework reading.” Proponents of this view, popularized by influential work by Coliva (Reference Coliva2015) and Pritchard (Reference Pritchard2015), can be seen to converge over the following two claims:

The Nonepistemicity Thesis is supposed to capture the framework theorists’ commitment to the idea that hinges (which are taken to be very general, nonempirical commitments such as that “I am not radically deceived,” “I am not a brain in a vat,” “there are physical objects,” and “my perceptual faculties are generally reliable”) are rationally inert—that is, that they are “visceral,” “nonoptional” commitments that are not responsive to positive or negative rational evaluation (like justification or doubt). The Basic Hinge Commitment, on the other hand, captures the framework theorists’ broadly foundationalist commitment about the structure of rational evaluation.

Combining these two theses together, we obtain the following concise summary of the main commitment of the framework theory:

In summary, the framework theory is a particular version of hinge epistemology that takes hinges to be the (a) not epistemically assessable (b) ground of our rational system. The picture of the structure of our rational evaluation that the framework reading invites us to buy, then, is one where a set of propositions, hinges, are, at one time the ground and the limit of our rational system. They are the ground in virtue of being those basic assumptions from which our belief system draws its rational validity; and they are its limit in virtue of lying outside epistemic appraisal.

Now, how does the framework theorists’ commitment to the Locality claim help them offer the unitary and charitable response to the problem of skepticism they promised?Footnote ¹¹ Start with Cartesian skepticism. The core challenge of this form of skepticism is directed at what we can sensibly claim to have justification or knowledge of, and it is about the scope or boundaries of our epistemic practices. To carry out this challenge, the Cartesian skeptic typically begins conjuring radical skeptical scenarios where nothing is as it seems, and where the gap between appearance and reality appears insurmountable (e.g., think of Matrix-style scenarios, evil demon possibilities, dream-like experiences, and so forth). On the basis of the ineliminability of these skeptical hypotheses, a Cartesian-style type of skepticism about the external world can be raised that makes leverage on the closure principle for justification (or CPJ)Footnote ¹² for knowledge or justification in the following way:

Virtually applicable to any empirical fact we take ourselves to know or believe justifiably, this argument provides a powerful, universal challenge to our knowledge of the external world.

In some sense, the advantage that endorsing the framework reading can give with respect to skepticism of the external world is clear. The Cartesian skeptical challenge relies on the closure principle, which is about extending one’s justification via competent deduction. Proponents of the framework reading start from this observation to derive the following condition for the applicability of the CPJ:

This is intuitive: after all, how would it be possible to extend one’s justification via inference if the entailed proposition was not available for epistemic assessment? In such a case, the CPJ would not be applicable. Crucially, a consequence of buying into the framework view is that the inference that goes from our justification of everyday propositions to our justification of the denial of the skeptical hypotheses (e.g., that I am not radically deceived) is precisely one where the consequent (i.e., a hinge proposition) is not in the market for epistemic appraisal. It follows then that the CPJ cannot be legitimately applied to the couple of propositions (i.e., E and ¬SH) that make up the Cartesian skeptical challenge.Footnote ¹⁵ For the framework theorist, then, to the extent that it purports to generate, through the principle of closure, a warrant for something (i.e., a hinge commitment) which is not itself warrantable (because it is one of the very conditions for the possibility of warrantability practices) then, premise 2 of the Cartesian argument should be rejected. By making a distinction between the denial of closure and its failure of application, then, the framework theorist can meet the challenge by exploiting a unique feature of hinge commitments which provides them with a principled way of rejecting premise 2.Footnote ¹⁶ At the cost of buying into a revisionary story about the structure of our rational evaluation, they can claim to have obtained a powerful antiskeptical argument that allows one to retain justification and knowledge of the external world in spite of one’s inability to know the denial of the skeptical hypotheses—without, with this, denying closure.

So far so good. But what about the other skeptical challenge? Recall that Pyrrhonism is not (primarily, at least) interested in undermining our grasp of empirical facts. Its doubt is cast over the standard of validity of our epistemic practices—that is, it concerns whether we can sensibly claim to possess a valid way of rationalizing our beliefs. In particular, the challenge gets started from the observation that epistemic justification is a matter of believing on good grounds, and that there can be such a thing as a justified belief. On this basis, the Pyrrhonian skeptic carries out the following considerations about rational support: if we take our beliefs to be justifiable, then they are justified either independently of reasoning (i.e., they are basic) or by another belief (i.e., they are justified inferentially). The Pyrrhonian skeptic rejects the first horn on the ground that basic justified beliefs are either rationally arbitrary—in which case they are not rationally compelling—or they are not—in which case they are indeed supported by reason, and so they are not basic after all. And they reject the second horn too by noting that, when we attempt to justify a belief by means of another belief, the support ultimately turns out to either be circular (mode of circularity) or terminate in an unjustified belief (mode of hypothesis) or continue endlessly (mode of infinite regress). Given that none of these “modes” to end inquiry is acceptable, the skeptic concludes, there is no valid way we can provide justification for our beliefs.

In summary, we can identify three main claims from which the Pyrrhonian skeptic mount their challenge:

where the latter two (the Pyrrhonian demand and the Agrippan modes) can be seen as specifying when the grounds mentioned in the former (Cliffordian Principle) are indeed good grounds. Hence, the Pyrrhonian challenge can be roughly characterized as setting the following desideratum, which cannot be satisfied:

Can the framework reading help us deal with this challenge? To start with, one could notice that by endorsing Locality, the framework reading seems to be committed to a broadly foundational response to Pyrrhonism, though perhaps only so in spirit. The response is foundationalist to the extent that it aims to resist the challenge by stopping the regress—hinges, according to the framework reading, are “regress-stoppers,” the solid foundation on which our rational system is built. But it is nonetheless a foundationalism sui generis, since standard foundationalist views normally take the grounds to be themselves rational. Footnote ¹⁷

In fact, while foundationalist views normally offer overriding responses to Pyrrhonism by arguing that there are some beliefs that are basically justified (i.e., by denying the Pyrrhonian demand), framework theorists concede that to Pyrrhonian skeptics, and instead claim that some of our core “beliefs” are not the kind of thing that can be subject to rational assessment. At this point, one may wonder whether this really is a response to Pyrrhonism. For notice that, if the solution offered by the framework reading simply is that such propositions ought to be taken for granted, then this does not seems to be significantly different from embracing the Agrippan mode of hypothesis. Footnote ¹⁸ Clearly, however, this would not be a very desirable result. Not only because it would consist in a capitulation, more than a solution, to Pyrrhonian skepticism—which in itself is a rather bad result for any self-proclaimed antiskeptical strategy. Even worse, this would give the skeptic a new way of reinstantiating skepticism of the external world—the one that targets justification and knowledge of everyday empirical beliefs.Footnote ¹⁹

How so? Assume, with the Pyrrhonian skeptic, that a belief is justified only if it rests on a good ground (recall the Pyrrhonian demand). This means that a belief B is rationally grounded only if it rests on some ground, G. As per the Pyrrhonian desideratum, however, B will not ultimately be rationally grounded unless G itself rests on a good ground (G*). Iterating this reasoning (within the foundationalist structure the framework theorists endorses), we have that any everyday empirical belief is ultimately grounded on hinge propositions, which are, by definition, rationally unresponsive. Since the ultimate ground of any empirical belief is not rationally held, and since a belief ought to sit on solid ground in order to be rational, it follows that none of our empirical beliefs would be rationally grounded.

You can put it this way: the framework reading shares with foundationalism a picture of the rational validity of the system of our beliefs as flowing upstream from the rationality of its foundations. At the same time, however, since it is a sui generis form of foundationalism, it does not enjoy the resources available to standard foundationalist views for drawing rational force from basic justified beliefs—hinges are the ground, yes, but they are not rationally responsive. And from the groundlessness of hinges, nothing prevents one from inferring the groundlessness of all the beliefs resting on them, including ordinary empirical propositions about the external world.

This is a problem the framework theorist must avoid. I will call it “The Central Challenge,” and reconstruct it thus:

Since premises 1 and 3 come with acceptance of the framework reading, there are only two possible ways its proponents can address the Central Challenge: deny premise 2, and show that hinges can be rationally held, even if not in the way envisaged by the Pyrrhonian demand; or deny the validity of the argument, and show how our belief system can be rationally valid despite the arationality of hinges. In the next sections, I will address each option in turn.

3. Extending Rationality

Denying premise 2 allows proponents of the framework reading to respond to the problem by arguing that, despite being rationally unresponsive, and thus being in this sense arational, commitment to hinges is still somehow rationally compelling. But how exactly? One influential proposal to this effect has recently been defended in a series of influential works by Annalisa Coliva (Reference Coliva2015, Reference Coliva2016, Reference Coliva2020). The fundamental insight of her proposal is simple: instead of attempting to bring hinges under the received notion of epistemic rationality (by showing how they can be mediately or immediately justified), she argues that we should instead extend the notion of rationality itself to include hinges.

The idea, in a nutshell, is to take hinges to be rational*Footnote ²¹ not because they draw their justification from other beliefs or from other facts, but in virtue of the role they play in the structure of our rational system. This role is often explained by analogy with rule-based games like chess: proponents of this strategy take hinge propositions to play a similar role to the one rules play in the game of chess—they are the instructions that make it possible to determine the practice of forming, assessing, and withdrawing ordinary empirical beliefs. The core insight Coliva draws from this analogy is that, like rules, which are a constitutive part of a game, hinges too are constitutive of our system of rational evaluation.

Following this line of thought, Coliva motivates a way of rethinking the notion of rationality to include—in addition to beliefs that inherit their justification status, mediately or immediately, from other beliefs or other facts—also those assumptions (read: hinges) that are constitutive of this practice—that is, assumptions that “make justifications within inquiry possible in the first place” (2020, 15). In this way (i.e., by extending rationality to the very conditions of validity of our epistemic practices), proponents of the framework reading can arguably have their cake and eat it too: they can preserve the sense in which hinges lack evidential support (because they are nonevidential) while at the same time proposing a way hinges can still be considered rationally* held (i.e., because they are constitutive of the notion of rationality itself).Footnote ²² Because it does the former, the framework theorist is capable of retaining the advantage of their position with respect to the closure-based challenge; and because it does the latter, the Pyrrhonian regress can be blocked in a way that preserves the rationality* of the system as a whole.

If this is correct, and so long as one is disposed to endorse a suitable (i.e., constitutive) interpretation of the rationality* of hinges, then this proposal seems to offer an elegant way out of the rational collapse threatened by the Central Challenge. But does it really? Here’s Crispin Wright’s reason to think it does not: if hinges lack evidential support, there must be no (rational) criterion to assess their “objective correctness”—that is, our endorsement of hinges must be completely arbitraryFootnote ²³ (Wright, Reference Wright2004, 205). And if hinge commitments are arbitrary, Wright notes:

Let me explain. For Coliva, hinges are rational* because the very fact that we engage in epistemic practices of a certain kind tells us that there must be some set of rules in place that makes this epistemic activity possible in the first place. The question raised by Wright can thus be expressed like this: if that is so, how can we tell which rules one should be committed to? Would not some epistemic practice work just as fine if one were committed to radically different hinges and incompatible with our own? Hinges rational unresponsiveness seems to rule out the possibility of finding a rationale for evaluating different sets of hinge commitments. Without a rationale, a criterion, however, any set of commitments would be rationally on a par—i.e., even systems based on bizarre, harmful, or aberrant grounds (e.g., one committed to “wood spirits, ectoplasm, and gods”).

The issue raised by Wright (what he refers to as “the problem of demarcation”) is a serious one, and one that, if unaddressed, leads proponents of this view to committing to an undesirable form of extreme relativism—that is, a form of relativism licensing an unlimited number of self-warranting systems of thought that are equally rationally valid.Footnote ²⁴ My interest is in two aspects that make such extreme relativism epistemically problematic: first, the fact that it allows justification for beliefs we would not otherwise take to be justified; and second, the fact that it allows basic Footnote ²⁵ methods for justifying beliefs that are intuitively problematic.Footnote ²⁶ Accordingly, a proponent of the framework reading who wants to resist the charge of relativism will have to explain two things: first, why it is not possible, according to their view, for the methods we use to justify our beliefs to justify different and incompatible beliefs; second, and related, why it is not possible for beliefs to be justified by different and unacceptable methods.

One way of doing the latter would be to argue that there is only one set of methods that can legitimately be used to rationalize beliefs, and it is our own. Why think so though? An obvious reason could be to say that any rational enquirer, like us, ultimately forms their beliefs on the basis of the deliverance of their senses or their rational faculties (Boghossian, Reference Boghossian2006; Coliva, Reference Coliva2015, 144). But this seems to be misguided. Imagine a community of people whose everyday beliefs are grounded, for the most part, on information obtained by consulting celestial bodies.Footnote ²⁷ Their methods may well be not very reliable, and their beliefs may ultimately turn out to be unjustified. Still, their rational system, we can imagine, is roughly structured like ours: they, like us, search for reasons to support their beliefs, and are able to give them upon being asked. The only difference is that while we base our beliefs about the weather by looking at the sky, say, they do so by looking at the position of the planets. Why would not it be possible to even conceive of some such community?

The framework theorists could respond by saying that consulting celestial bodies is not a basic method for justifying a belief (Coliva, Reference Coliva2015, §4.4). Someone who relied on this method must have used their eyes to consult the planets, and possibly infer from what they saw following basic rules of logic (and the same would go for other funky practices as well). This, however, will not be entirely satisfactory either. For it does not follow, from the fact that some justificatory practices often function only if coupled with others, that they are not epistemically basic. Testimony is made possible by the fact that we have eyes and ears to read and listen, but it would be quite of a stretch to draw from this the conclusion that testimony can be reduced to perception.Footnote ²⁸ When a person from a remote community comes to believe that they are going to have a bad Monday because their Jupiter is in Pluto, a truly nefarious occurrence, their belief is not justified by their having seen (or read) that it is so—or at least, not only: for they could have seen or read about Jupiter and still not formed the belief. What makes the belief justified for them is having learned about the position of Jupiter coupled with their taking astrology to be an adequate means for rationally forming their beliefs.Footnote ²⁹

Still, a point could be made by the proponents of this strategy that, since perception is often a key ingredient for belief justification, members of this faraway community would not end up forming beliefs that are substantially different from our own, and so the relativist threat would not be a serious one. However, this may lead to an unjustifiably chauvinistic image of rational inquiry. For why think that all rational inquirers ought to be like us and justify beliefs the way we do? Would not it be possible to at least conceive of alternative practices of epistemic inquiry? Would not a God endowed with divine intuition, allowing her to grasp truths about her environment without the use of (humanoid) senses, count as a rational inquirer? And what of alien creatures? We could imagine aliens (who may as well be, all in all, ultimately not that different from us) who enjoy a felicitous attunement with their environment, so much so that facts that are relevant for them in a particular circumstance would cause them to form (defeasibly) true beliefs about them. What could a proponent of the framework reading do to block this line of thought?

So much then for epistemic practices intended as the methods by which we justify our beliefs. Now, what about the possibility that the same methods justify different and incompatible beliefs? Could (say) my perceptual experience as of a cup justify not only ordinary beliefs about the presence of that cup but also, say, the skeptical belief about the mere appearance of a cup? Presumably, the answer to this question will depend on one’s preferred view about perceptual justification. One option then could be to pick and defend a strongly externalist view of perceptual justification that understands perception as a reliable process that puts us in direct contact with the world. This view would have the advantage of ruling out the possibility that some odd beliefs (different and incompatible to beliefs we ourselves form via perception) could be perceptually justified. This might seem quite ad hoc, but theoretically feasible.

A problem with this, however, is that framework theorists take hinge commitments to play a central role with respect to the justificatory status of individual beliefs. Recall that, independently of one’s preferred internalist or externalist constraints on justification, framework theorists take hinges as the condition of possibility for justification. Referring to perceptual justification in particular, for example, Coliva writes that:

The idea then is that, irrespectively of the particular view (in this case, of perception) adopted by the framework theorist, what can be justified will ultimately depend on the hinges one is committed to. Imagine you have an experience as of a cup in front of you. For this experience to provide justification (a perceptual warrant) for the belief that there is a cup, it must be the case that one is also assuming that there is an external world out there made out of, among other things, that very cup one sees in front of them—and possibly that they are not radically deceived too—or so the thought goes. If one did not assume these very general propositions (if one were committed to the thought that, say, perception is a grand illusion), then one would not have justification for that belief but, perhaps, another radically different and incompatible one—say, that there is a very elaborated virtual reconstruction of a cup in your head.

The antirelativist gambit of proponents of this view consists in the naturalist strategy of showing how our system of rational evaluation is de facto unique.Footnote ³⁰ They do this either by arguing that alternative basic methods are not available, or that communities holding radically different beliefs would ultimately have a system of beliefs not radically different from our own. If I am right, however, it seems implausible to think that there is anything unique about our system of thought. Most importantly, however, even if this were actually the case, this line of argument hardly suffices to quiet the worry that other radically different and incompatible rational systems could not be possible at least in principle. This, recall, is precisely the crux of the extreme relativistic worry that, given a conception of hinge commitments that are rational* but unresponsive to reasons, “there will be no obstacle in principle to the idea of alternative, equally valid ways of conceiving the substance of the world” (Wright, Reference Wright2004, 206, my emphasis).

The key intuition of this strategy is that the structure of our rational practices mandates that some propositions must be taken for granted. From this, it would indeed be tempting to infer that the propositions that are to be taken for granted must necessarily coincide with our fundamental commitments (i.e., that there is an external world, that we are not radically deceived, and so on). The transcendental consideration that we engage in particular epistemic practices, like that of giving and asking for reasons, however, concerns the structure of epistemic evaluations—what Wittgenstein calls “the logic of our investigations.”Footnote ³¹ Precisely because it concerns its logic, however, this view is by its very nature neutral with respect to the content of the system thus constituted. We can, in fact, come up with radically different thought systems with a structure similar to our own—and we can do that quite easily. Think, for instance, of a community—call them Bostromians—whose members, perhaps seduced by Elon Musk’s own endorsement of Bostrom’s simulation theory,Footnote ³² have come to accept that the reality we live in is a computer simulation. What is interesting about Bostromians is that, when presented with chair-like objects, they do not tend to form beliefs about chair-looking physical objects but about, say, mere appearances of a chair (or about “virtual” chairs). Like us, Bostromians would go about in their everyday epistemic business forming beliefs about things around them, and upon being questioned, they would give the same perceptual or testimonial reasons we would also give. Only, they would take themselves to be surrounded by “virtual” objects, products of an elaborate simulation, created by a super-intelligent computer, and fed to us (we can suppose) through something like a Matrix-style brain implant. Is it possible to imagine this within the constraints of this view? I do not see why not. On the contrary, the framework theorist does not seem to be in a position to explain how such beliefs could not be justified.

If this is right, then, and despite attempts by proponents of this strategy to argue for the de facto uniqueness of our epistemic practices, the point has not gone away that their view exposes itself to the in principle arbitrariness of the rational* basis of our system of rational evaluation. And to the extent that this arbitrariness, as I have shown, engenders an undesirable form of extreme relativism, the framework theorist’s putative merits in dealing with the Central Challenge are thereby also put in jeopardy.

4. Surrendering to the Angst

Recall the Central Challenge: on the one hand, we have the Pyrrhonian demand that, in order to be rational, a belief must be supported by another fact or a belief. On the other hand, we have the framework theorist’s commitment to a sui generis foundationalism according to which a belief’s chain of support ultimately bottoms down on a set of fixed propositions that lack evidential support. The Central Challenge then asks how we can preserve the rationality of our belief system if its validity rests on the rationality of its grounds.

In the previous section, I have looked at an attempt to steer clear of the problem by questioning the Pyrrhonian demand and argue that there are other ways in which beliefs (and, in particular, hinges) can be said to enjoy some kind of rational standing (i.e., by being constitutive of rationality). But that move appears to lend itself to a characterization of rationality that is relativistic in a bad way. So what else can the framework theorist do to avoid the Central Challenge?

Another option could be to endorse the Pyrrhonian demand, admit that, ultimately, our rational system is groundless, but at the same time resist the temptation to conclude from this that our everyday empirical beliefs are not justified.Footnote ³³ After all, the idea goes, if we could preserve the rationality of (nearly) all empirical beliefs individually, the arationality of our belief system as a whole would not pose that much of a Challenge, would it?

Maybe. But how so? One option could be to distinguish between two ways of rationally evaluating a system of beliefs: a summative and a nonsummative one. Summative rational evaluation would take the rationality of the belief system taken as a whole to be tightly connected to the rationality of the individual beliefs that constitute the system—such that, quite roughly, the system would be rational if all (or most) beliefs that constitute the system are themselves rational. Nonsummativism about rational evaluation, on the other hand, would simply deny the summative claim and decouple the rationality of the system from the rationality of the individual beliefs.

The interesting thing about this distinction is that it makes it possible (if one is disposed to endorse nonsummativism about rational evaluation) to conceive of a belief system that, while being nonrational when considered as a whole (say, due to the fact that it sits on arational grounds), is nonetheless constituted by a series of perfectly rational beliefs. According to Pritchard, this result is, in fact, a natural consequence of the framework theorist’s commitment to Locality.Footnote ³⁴ Because hinges are exempt from epistemic appraisal, any attempt to ground ordinary empirical beliefs on them is doomed to failure in a similar way in which the skeptic’s attempt to infer the lack of groundedness of our everyday empirical beliefs from the lack of groundedness of the denial of skeptical hypotheses is also, from a hinge theoretical point of view, fundamentally misguided. For Pritchard, this is what commitment to an image of rational evaluation as local entails—that is, it entails that the rational validity of our empirical beliefs does not flow upward all the way from the very bottom of the system (i.e., from hinges) but, say, from other, more “local” sources of supportFootnote ³⁵—or so at least the thought would go.

If this is correct, then, it could be possible to solve the Central Challenge by endorsing nonsummativism about rational evaluation and reject the connection between the lack of groundedness of our belief system and that of our individual beliefs and argue that, though hinges’ arationality does ultimately undermine the rational validity of the system taken as a whole, provided one had good reasons for them individually, each and every individual belief could still be rationally grounded.

Will this strategy be successful? I do not think so, and here is briefly why: according to nonsummativism, at a local level—that is, when considering the rational standing of some everyday belief—the rational standing of a belief (about, say, the fact that the moon is full) must be independent of the ultimate ground on which it stands (according to the framework theorist, the hinge that, say, we are not radically deceived in our perceptual judgments). In other words, it must be only the local perceptual reason (say, that I see that the moon is full) that rationalizes the belief. However, as per classical skeptical scenarios, we can conjure a situation in which an epistemic subject is in the exactly same internal state as ours but, unluckily, finds themselves in an epistemically “bad” environment—they could be a brain in a vat, for instance.Footnote ³⁶ Arguably, then, all their perception can justify is, at best, beliefs about appearances—such as that it looks to them as though the moon were full. However, if this is plausible, whichever rational support we (in the good case) have for our belief is also available to the subject in the bad case (since, by stipulation, the two subjects are internally identical). But if the evidence in the two cases is the same, and it can at most justify beliefs about appearances, how can we claim to have evidence for the ordinary belief that there is a full moon today (and not the mere appearance of one)? More exactly, this point can be made by relying on the following:

In this case, the underdetermination principle can be employed to show that, since the evidence locally available to a subject in the good scenario (us) is the same as that available to a subject in the bad scenario, and the latter can only justify appearance beliefs, then there is no evidence available to us that can favor our empirical belief about the moon over, say, the belief that there is only the appearance of a full moon. For each and every individual empirical belief, then, a new form of radical skepticism about the external world ensue:

In response, it could be pointed out that the soundness of the argument threads on an internalist conception of local support. Consider the following:

The plausibility of this lemma rests on a broadly internalist notion of evidence, whereby evidence supervenes exclusively on facts internal to the subject (namely, on the subject’s mental states). If this is true, since externalist evidence is not available to the subject in the bad scenario, it would suffice to replace the notion of evidence at play, in this case with an externalist one to get rid of SEL, and with it the problem that triggers the underdetermination principle.

Notice however that not any kind of externalist notion of evidence will do (if any), as the framework theorist is under some pressure to accept at least some internalist constraint on the notion of justification. This is because not doing so would undermine a strong motivation for the view itself. If you recall, the framework theorists promised to offer a charitable response to skepticism capable of bringing internalism home safe and, possibly, of carving out a skeptic-free conception of internalist rational support (no less). Following Bergmann (Reference Bergmann2006), then, a minimal internalistic requirement for this strategy could be phrased as follows:

where the fact that X “depends in some important way” on the S’s subjective perspective is what makes X an internalistically conceived reason. With this requirement in place, we can then formulate a new underdetermination principle that embeds this demand in the following way:

Naturally, UP_SCD licenses the construction of a skeptical challenge similar to the one before, and one that the framework theorists cannot ignore, given their commitment to SCD. Notice, however, how SCD (and consequently UP_SCD as well) is compatible with the idea that some other, noninternalist reason may play a role in supporting our ordinary beliefs. That is precisely what makes it a minimal requirement: so long as it is possible to reserve some space for internalistically conceived reasons, the SCD is happy to concede, that will do.

This may be good news, for it seems to open the door to the possibility of combining together externalist and internalist strategies for epistemic justification. Imagine, for instance, we could have externalist support that was also rationally grounded: being externalist, it would elude the sameness of evidence lemma by granting favoring support to epistemic agents in the good case that would not be available to those in the bad case. Being rationally grounded, on the other hand, it would guarantee that this favoring support is also cashed out in internalist-friendly terms, thus offering the kind of response to the skeptic that the framework theorist is committed to giving. In his book, Epistemic Angst, Duncan Pritchard motivates precisely one such idea by embedding it in his disjunctivist treatment of underdetermination based radical skepticism. According to Pritchard’s epistemological disjunctivism, agents in the good and bad cases do not possess the same evidence because perception in the good case provides them with reflectively accessible factive support—namely, a kind of externalistically conceived support that the perceiver can also access reflectively.

But how does this work more precisely? Recall that the SCD asks that our individual empirical beliefs are also supported by internalistically conceived reasons. For this strategy to work, then, it must be the case that the support enjoyed by my empirical belief (say, the belief that p: “there is a tree in front of me”) is not provided only by local external facts (like, say, that “I see p”), but also by some other internalist condition. One straightforward way of satisfying this requirement (and the one Pritchard proposes) is to cash out this internalistically conceived condition in simple access internalist terms. Roughly, according to access internalism, one is justified in believing p if one can have reflective access to one’s basis for believing that p. Overall, then, according to this strategy, in paradigmatic cases of perception, we can have local support of our empirical beliefs because: (1) we have a factive reason for them—namely, the fact that we see that p, which entails the truth of p—, and (2) we have reflective access to this factive reason that satisfies the internalist condition.

Combining these requirements in a coherent disjunctivist view of epistemic justification has proven to be exceptionally difficult.Footnote ³⁹ In what follows, however, I want argue that even if epistemological disjunctivism were capable of offering the kind of hybrid support that is needed to satisfy the minimal internist requirement and reject SEL, it would not be able, as part of the broader antiskeptical program proposed by the framework theorist, to avoid the view’s relapse into skepticism.

To see this, take a closer look at the access requirement in (2). A natural way to conceive of reflective access is in terms of a subject’s awareness of an object (read: a fact, a reason) as something that is relevant to or contributes to the justification of the relevant belief. The reason for this is clear, and I take it is part of what motivates this strand of internalism about epistemic justification in the first place—roughly, the thought that reasons are precisely those things that we can conceive of as that which justifies our beliefs. Footnote ⁴⁰ Following Bergmann (Reference Bergmann2006), let us call this version of access internalism strong access internalism. According to a strong access internalist version of epistemological disjunctivism, then, claim (2) ought to be understood as maintaining that when we have reflective access to a factive reason for p, we have reflective access to this reason as a basis for our belief that p.

The problem with strong access internalism is that it is vulnerable to generating familiar Pyrrhonian worries. How so? By the light of strong access internalism, in order to count as a reason for my believing that p, the factive support that “I see that p”, must be coupled with my awareness of it as a basis for my belief that p. But naturally, my awareness of this factive support is itself susceptible to being evaluated with respect to its justification status.Footnote ⁴¹ For how can I be sure (say) that this awareness is genuine, and I am not confused? Or: how can I rule out the possibility that I wasn’t utterly irrational or insane in conceiving of that support as a basis for my belief? Crucially, answers to this question will have to meet the standard imposed by the Pyrrhonian demand—i.e., not be arbitrary, circular, or leading to an infinite regress. Can proponents of this strategy satisfy the demand?

I think we should give a negative answer to this question. For what would a good ground look like in this case? On the one hand, proponents of this strategy cannot appeal to hinges, for this would establish a direct grounding relationship between individual empirical beliefs and hinges, which would undermine their commitment to a nonsummativist view of hinges’ foundational role. On the other hand, however, nothing short of hinges would do. For suppose some relevant nonhinge belief could somehow stand up to the Pyrrhonian demand. Even if this was possible (and it would be indeed rather puzzling if it were possible), this would not be a good result, since it would undercut the framework theory’s core antiskeptical insight about hinges’ role as “regress stoppers.”

So maybe the proponent of this strategy may want to opt for a weaker form of access internalism, one that need not commit to conceiving (whether conceptually or not) of the factive support as relevant to the justification of the belief. That is, they could endorse weak access internalism Footnote ⁴² and reject the claim that one’s awareness of that which justifies a belief ought necessarily to involve that one conceives of it as that which does that job. In this way, they could retain the intuition that subjects have some conception of the factive support for their belief that p, while at the same time blocking the emergence of the Pyrrhonian worry by leaving the entirety of the “justificatory work” in the hands of the (externalistically conceived) factive reason.

The obvious problem for framework theorists that choose this strategy, however, is that the thought motivating strong access internalism constitutes a main advantage of internalist views of justification compared to externalist ones, and the crucial motivation for imposing an awareness requirement in the first place. For weak access internalists, someone who bases their belief that p on good grounds, since they are not aware of such ground as a basis for that belief, would be in the same subjective state concerning the status of their belief as a subject for whom the justifiedness of their belief may as well be a mere accident. Naturally, however, this clashes with the hinge epistemologists’ commitment to the subjective condition demand, according to which something counts as an internalistically conceived reason only if it depends in some important way on S’s internal subjective perspective. Clearly, if weak access internalism is compatible with someone being completely in the dark about the role of the accessed belief as a justifier, there is no way (let alone “important”) in which S’s justification can be said to depend on S’s own subjective perspective.

The minimal internalist requirement is supposed to pick out a condition no internalism can give up without undermining some key motivating reasons for their view.Footnote ⁴³ Even if this was not a conditio sine qua non for epistemic internalism, it does constitute an important insight one should not want to give up too easily. A main selling point of the framework reading is that it promises to offer a solution to skepticism that is charitable precisely because it does not sacrifice key internalist insights that put internalist theories of epistemic justification in a position of advantage over externalist ones. By adopting a weak version of access internalism, proponents of this strategy give up precisely one of these benefits, thus undermining what motivated the main appeal of the hinge project in the first place.

In conclusion then, framework theorists who choose to rely on the virtues of epistemological disjunctivism in order to steer clear of the underdetermination based skeptical worry without giving up the minimal internalist requirement are confronted with two infelicitous alternatives: endorse strong access internalism and face the same Pyrrhonian puzzle they attempted to escape, or opt for weak access internalism, and undermine what motivated the choice of their strategy in the first place.

5. CODA

The main appeal of hinge epistemology is the fact that it purports to offer a unitary solution to skepticism without sacrificing fundamental internalistic insights about the nature of justification. If I am right, however, the two most prominent formulations of the framework reading fail to deliver their promise. The strategy of “extending rationality,” championed by Annalisa Coliva, may be able to successfully salvage an internalist conception of justification, but it does so at the cost of committing to a radically relativistic structure of rational evaluation. The strategy of “surrendering to the angst,” on the other hand, championed by Duncan Pritchard, may be able to provide some solace from one formulation of the skeptical threat, but only at the cost of renouncing key internalist insights about epistemic justification. Since this was a major advantage of the framework theory compared to competing antiskeptical strategies, however, this strategy ends up undermining a crucial motivation for the view itself.