Beliefs, delusions, hinge commitments

Abstract

It is argued that in order to properly engage with the debate regarding the ethics of belief one first needs to determine the nature of the propositional attitude in question. This point is illustrated by discussing a related topic from social philosophy, broadly conceived, concerning the nature of, and inter-relationship between, delusions and the Wittgensteinian notion of a hinge commitment. Are we to understand either or both of these notions as beliefs? Are delusions a kind of hinge commitment? In answering these questions we will appeal to a distinction between folk belief and knowledge-apt belief. It is argued that while both delusions and hinge commitments count as beliefs in the former sense, neither is a belief in the latter sense. Moreover, once we understand what is involved in the notion of a hinge commitment, it will also become clear why delusions are not hinge commitments. It is claimed that by gaining an understanding of delusions and hinge commitments, and thereby of two fundamental ways of thinking about belief, we will be in a better position to determine what is at issue in the ethics of belief debate.

1 The ethics of belief

Central to the ethics of belief debate is the question of whether it can be appropriate to form one’s beliefs in ways that are unresponsive to genuinely epistemic considerations. In simple terms, can it be appropriate to form a belief that p even when one lacks any rational basis for treating p as true? Call a negative response to this question evidentialism and a positive response to this question non-evidentialism.¹ My primary interest in this paper is to examine the way in which the Wittgensteinian notion of a hinge commitment relates to this debate. This is because, on the face of it at least, taking this notion seriously entails embracing non-evidentialism. The reason for this is that Wittgenstein is quite clear both that hinge commitments are an essential feature of our practices and that they are completely lacking in rational support. Moreover, while we might be ordinarily unaware of the groundlessness of our hinge commitments, Wittgenstein is explicit that becoming aware of their groundlessness does not lessen our conviction in them (because nothing can lessen our conviction in them, at least in normal conditions). Accordingly, it seems that embracing hinge commitments would entail one taking a specific line on the ethics of belief debate.

I will be arguing that the manner in which hinge commitments relate to the ethics of belief debate is in fact more complicated than this initial overview might suggest. Indeed, I will be suggesting that for this reason reflecting on the notion of a hinge commitment offers us a constructive way of engaging with the ethics of belief debate. This is because it prompts us to draw an important distinction between two ways of thinking about belief: the everyday notion of belief (folk belief, as I term it) and a more specific notion of belief that is of particular interest to epistemologists (knowledge-apt belief, or K-apt belief for short, as I term it). This distinction can feed, in turn, back into the ethics of belief debate, as we will see.

In order to bring out the relevance of hinge commitments to the ethics of belief debate, however, it will also be necessary to consider another kind of propositional attitude, that of a delusion. Now one might think that delusions don’t have anything to teach us with regard to the ethics of belief debate, in that they concern beliefs that are both inappropriate and irrational. Accordingly, one can accommodate them without having to reject an evidentialist line on the ethics of belief. The relevance of delusions for our purposes arises from the fact that there has recently been a growing literature that explores the idea that delusions could be a kind of hinge commitment. Accordingly, insofar as hinge commitments are appropriate beliefs to hold, then it follows that delusions, qua hinge commitments, can be appropriate beliefs to have as well, even despite their lack of an adequate rational standing. I think that understanding delusions as hinge commitments is a mistake and so will be arguing against this account. The advantage of bringing this hinge conception of delusions into our discussion is that it will enable us to sharpen our understanding of the distinctive kind of propositional attitude at issue in our hinge commitments. Moreover, as we will see, the distinction between folk belief and K-apt belief also has an important application to the delusion debate, thereby providing further independent theoretical support for taking this distinction seriously.

Here is my plan of action. I will begin by exploring the notion of delusion (Sect. 2). I will use this notion to introduce my distinction between folk belief and K-apt belief, in order to explain why delusions only count as beliefs in the former sense. I will then articulate the notion of a hinge commitment and explore why one might think that delusions are hinge commitments (Sect. 3). A key part of the discussion here will be the claim that neither hinge commitments nor delusions are K-apt beliefs. I will then argue that delusions are not hinge commitments (Sect. 4), which will enable me to distinguish between different kinds of non-K-apt belief. I will then conclude (Sect. 5) by bringing these threads together to examine their relevance for the ethics of belief debate.

2 Delusions and belief

Delusions come in many forms. For example, one might call a person who believes that he is Napoleon deluded. Or consider severe types of paranoia, which are also often classed as delusions. Some delusions take a particularly striking form, such as the Cotard delusion that one is dead or the Capgras delusion that a loved one has been replaced by a doppelgänger. It is usually accepted that delusions are the mark of mental illness of some kind. Relatedly, they are often thought to be manifestations of a kind of extreme irrationality, as the subject fixates on a particular belief even in the face of massive amounts of contrary evidence.

While delusions are usually false, it doesn’t seem to be essential to them that this be so (although, historically at least, they have usually been defined as such by the relevant medical bodies).² What seems to be more relevant is that the subject has a complete conviction in the truth of what they believe even despite the massive amounts of contrary evidence in play. That is, it is the combination of certainty along with not only a lack of supporting evidence (much less the kind of evidence that would warrant such a level of certainty) but also a complete resistance to any counter evidence. This irrationality could conceivably coexist with the target belief being, as it happens, true. So long as the subject had no rational basis for the truth of this belief, then the particular kind of irrationality that is distinctive of delusions would remain and that seems to be more central to the notion.

Similarly, it is important to delusions both that they run contrary to what other reasonable people believe (or would believe, in the circumstances) and, relatedly, that they are also in at least apparent conflict with the subject’s other beliefs. Normal, rational people do not believe that their loved ones have been replaced by doppelgängers, or that they are Napoleon. Moreover, our deluded subject will have lots of reasons at their disposal to indicate to them that their beliefs are irrational. For example, they will have plenty of evidence available to them that indicates how implausible it is that they are really dead or that they are really Napoleon, not least in the form of those around them telling them that these beliefs are obviously false. And yet a sign of the delusion is that they will persist in being certain of this belief regardless. While they may have some partial insight into the nature of their irrationality, they are unable to fully recognize it.³

That delusions are in conflict with what ordinary people believe, and in conflict with other beliefs that one holds, explains why it would be a mistake to class, for example, ordinary religious beliefs as a delusion.⁴ While such beliefs might have a similar kind of certainty attached to them, and also a similar resistance to evidence, they seem different to delusions on these fronts. In particular, religious believers belong to a community of like-minded people who also hold these beliefs, so there is not the conflict with what ordinary people believe that is typical of delusions. Moreover, in the ordinary religious case there is not usually the kind of direct tension between the putative delusional belief and the subject’s other beliefs. By being part of a religious community one also partakes in a narrative that squares one’s religious beliefs with other claims one might be inclined to hold (such as regarding contemporary science), at least to the extent that there is no direct contradiction anyway. Relatedly, just as ordinary religious beliefs wouldn’t qualify as delusions, so it would also be a mistake to treat radically false beliefs that were widely held with complete conviction in the past as delusions, even if the same beliefs held now in this fashion would so qualify (consider the case of flat-earthers).

Note that I use the word ‘belief’ to describe delusions. In philosophical circles at least, this is controversial, even though in medical definitions of delusions this is common practice.⁵ One can appreciate why one might hesitate to refer to delusions as beliefs given the properties that have just been ascribed to them. After all, in what sense can one really believe that one is dead if one acts in every other way as if one is alive (which is the usual situation with cases of Cotard delusion)? Beliefs are manifest in action, after all—if someone acts as if they believe that p, then we would normally ascribe that belief to them even if they claim to not believe that p. Even if we treat that claim as sincere, we might nonetheless think that the agent is confused or otherwise mistaken about what they in fact believe, given their actions. With this in mind, shouldn’t we instead treat delusions as a different kind of propositional attitude to belief?⁶

I think we should resist this line of argument, or at least I would suggest that what motivates this line of argument is better captured in a slightly different fashion. The joker in the pack in this discussion is the folk notion of belief. Theorists often talk about belief as if it is a clearly understood notion, and yet I think it is also transparent on reflection that they often mean very different things by it. (Philosophers are no exception). The folk notion of belief, for example, seems to be incredibly permissive, in that a whole range of distinct propositional attitudes appear to count as beliefs in this sense.⁷ The folk treat both religious faith and scientific endorsement as belief, for example, even though these are clearly very different propositional attitudes.⁸ Moreover, there doesn’t seem to be any impetus to distinguish between belief and other notions in the vicinity that philosophers usually contrast with belief, such as acceptance.⁹ In particular, so long as one sincerely endorses that p, then that will usually suffice for one to be counted as believing that p on the folk conception, regardless of whether one also satisfies any of the other conditions that one might normally demand of belief. For example, the folk conception doesn’t seem to have any problem with treating obviously irrational commitments, or even contradictory commitments, as beliefs. (E.g., if you sincerely claim that you believe that the earth is flat, then we’ll take you at your word, even if you just got back from a transatlantic flight). So long as you explicitly endorse the relevance claims, then you count as (folk) believing them. If that’s right, then the medical establishment are quite correct to treat delusions as beliefs, at least insofar as they are using this term in its everyday manifestation, as they really are beliefs in this sense. The bar for the folk notion of belief is very low.

That’s not to say that the concerns about treating delusions as beliefs lack merit, however. The point is rather that they are misplaced when directed at the folk conception of belief (‘folk belief’, for short). Consider, for example, the following type of belief that I have elsewhere christened K-apt belief.¹⁰ K-apt belief is concerned with that propositional attitude that is a constituent part of rationally grounded knowledge. Just as philosophers in general, like many other kinds of theorist, are inclined to use the notion of belief in very different ways without marking this difference, I take epistemologists to usually have this specific notion of belief in mind, as opposed to the folk notion. What is distinctive about K-apt belief is that it has certain basic conceptual connections to reasons and truth that the folk conception lacks. One core condition for K-apt belief, which doesn’t apply to folk belief, is that one cannot K-apt believe that p while recognizing that one has no rational basis for the truth of p. (Note that this doesn’t exclude groundlessly K-apt believing that p, but only that one cannot be aware that one is groundlessly K-apt believing that p). Insofar as one recognizes that one has no rational basis for the truth of p, then so long as one retains one’s commitment to p that commitment cannot be a K-apt believing, but must be a different propositional attitude altogether (such as a wishful thinking that p, or a hoping that p, etc.,).

Interestingly, however, it seems that one can believe that p in the folk sense even while recognizing that one has no rational basis for the truth of p. Indeed, the folk are quite happy to attribute belief in these cases, so long as the individual sincerely avows that p. Some instances of religious faith are just like this, for example, in that the ‘believer’ maintains that their commitment is not grounded in reasons at all, but is rather a pure expression of faith. Or take the testimony of a parent who is convinced of the innocence of their child who has been charged with a heinous crime, even despite the massive preponderance of evidence for their guilt. The parent might well grant that all the evidence points against their child’s innocence, and yet maintain their belief in this regardless. It seems clear that the folk would be happy to attribute belief in such a case. This indicates, as we’ve previously noted, that the threshold for folk belief is fairly low (and certainly very low compared with a narrower conception of belief like K-apt belief).¹¹ This confirms the previous point that we made about folk belief, which is that it suffices for belief in this sense that the subject sincerely endorses the target proposition. Accordingly, even if the commitment in question lacks many of the properties that we usually associate with belief, such as by being consciously groundless and completely unresponsive to counterevidence, or by being in direct conflict with the subject’s other beliefs, it can still count as a belief in the folk sense of the term.

If the foregoing is correct, then we can offer an irenic slant on the delusions and belief debate. On the one hand, granting that delusions are beliefs in the folk sense ought to be relatively harmless, given the permissive nature of how the folk employ this notion. That delusions are certainties that are groundlessly held, resistant to counterevidence, and in direct conflict with one’s other beliefs, need not be a barrier to them counting as folk beliefs. On the other hand, one can also consistently argue that delusions need not be beliefs in a more restrictive sense of belief than the folk conception. In particular, K-apt belief seems particularly relevant here, given the way in which delusions tend to be consciously ungrounded, resistant to counterevidence, and in direct conflict with the subject’s other beliefs. It should be clear that delusions fail the test for K-apt belief noted above, given that they involve a commitment to the truth of the target proposition—indeed, involve the subject being absolutely certain of this proposition—that continues even once the subject recognizes the lack of a rational basis for this commitment. Delusions are thus not K-apt beliefs, and hence there is at least one more narrow theoretical sense of belief that they do not satisfy.¹²

3 Hinge commitments

In his final notebooks, published posthumously as On Certainty [OC], Wittgenstein describes a radical new way of thinking about the structure of rational evaluation. Rather than seeking self-supporting rational foundations for our most certain commitments, he maintains that the certainties that lie at the heart of our practices are arational. Instead of revealing that there is something inherently problematic about our system of beliefs, however, the Wittgensteinian thought is that it is in the very nature of what it is to be rational that one has these fulcrum arational certainties at the core of one’s beliefs. To be rational is thus to have so-called ‘hinge’ commitments (OC, pp. 341–43), where the very system of rational evaluation turns on the existence of these arational hinge certainties.¹³

It has been suggested that one way of thinking about delusions—at least of a certain fundamental kind anyway—is that they are a variety of hinge commitment.¹⁴ One can see the attraction of the proposal. Delusions are often held with complete conviction, and moreover they seem to play a fixed role in the subject’s rational practices, in that like hinge commitments they are held fast relative to other beliefs that are open to revision. Relatedly, like hinge commitments they also seem to neither be the product of rational processes nor responsive, at least in the usual ways, to rational considerations. Indeed, this is why some commentators (as we noted above) have suggested that delusions cannot be beliefs, a point that has also been made about hinge commitments (as we will see in a moment). It is thus tempting to suppose that delusions may be arational hinge commitments. Nonetheless, I think once we move beyond these particular features of hinge commitments and consider their broader nature, then it becomes clear that delusions simply cannot be hinge commitments. With this in mind, let us now examine the Wittgensteinian notion of a hinge commitment more closely.

The point of the hinge metaphor for Wittgenstein is that all rational evaluation necessarily occurs against a backdrop of certainty. It is this backdrop of certainty that enables these rational evaluations to occur. (OC, pp. 341–343) Since this certainty is required for rational evaluation to be possible, it cannot be itself rationally grounded, and hence it is arational. This overarching certainty is essentially a conviction in the general truth of one’s worldview; that one’s beliefs as a whole are not fundamentally awry. This is what needs to be in place in order for a subject to even enter into the space of reasons. As Wittgenstein notes, the ‘game of doubting itself presupposes certainty’ (OC, p. 115), where what applies here to doubt would equally well apply to (K-apt) belief too.¹⁵

It is important to emphasize that this overarching hinge certainty is not an assumption or an hypothesis on one’s part. (OC, p. 343) Indeed, for it to be in place doesn’t even entail that one will ever have an occurrent thought with this content (something that is surely required for an assumption or hypothesis). Moreover, hinge commitments are a very different kind of propositional attitude to assumptions or hypotheses. One’s assumptions and hypotheses can be adopted as a strategy, and that strategy can be rationally assessed. Relatedly, one can be agnostic about the truth of one’s assumptions and hypotheses. None of this applies to one’s hinge commitments. This is a kind of absolute certainty that is already manifest in one’s everyday action. There is no room for agnosticism, nor it is a stance that one adopts (and hence there cannot be any strategy, rational or otherwise, involved in its adoption). It is not that we come to have this conviction because we recognize its necessity to our rational practices. Wittgenstein’s point is rather that one necessarily absorbs this certainty as part as acquiring one’s worldview. Having such certainty in the background is just what it is to be a rational person who has a worldview at all. It is thus a certainty that is grounded in action rather than reasons. Wittgenstein approvingly quotes Goethe in this regard: ‘In the beginning was the deed.’ (OC, p. 396) The broader point is that this certainty is not the product of ‘ratiocination’, but rather ‘primitive’, visceral, ‘animal’. (OC, pp. 359, 475) It is, as Wittgenstein puts it at one point, simply ‘there—like our life.’ (OC, p. 559; cf. OC p. 344).¹⁶

Similarly, just as this conviction is not grounded in reasons, so it is not responsive to reasons either. Part of the issue here is that nothing could rationally count against the overarching certainty in one’s worldview, because in trying ‘to doubt everything you would not get as far as doubting anything.’ (OC, p. 115) But the deeper point is that it is in the very nature of this conviction that it is not rooted in reasons at all, but rather a primitive commitment that underlies one’s rational practices. One can, of course, say that one has radical doubts about one’s commitments as a whole, but Wittgenstein’s contention is that such a claim is in a certain sense empty, as one’s actions reveal that one’s underlying certainty in one’s worldview remains unaffected.

This overarching conviction is manifest in a complete certainty in a range of specific everyday claims that are such that to be wrong here would entail radical error in one’s worldview. In normal circumstances, for example, one’s certainty that one has hands (e.g., OC, p. 250) is such that one cannot even make sense of the possibility of a mistake. To be wrong about whether one has hands in normal conditions would be for one’s conception of the world as a whole to be in error. This is why in such conditions one treats having hands as a fixed point relative to which other things are rationally evaluated. (OC, p. 125) Similar points apply to other core commonsense nodes of one’s worldview, such as that one’s name is such-and-such (OC, p. 425), or that one is speaking English (OC, p. 158). Just as the overarching certainty in one’s worldview is necessarily groundless and unresponsive to reasons, so one’s commitment to these specific hinge commitments has the same properties.¹⁷ We thus get the advertised overlap between hinge commitments and delusions, given that the latter are also both groundless and resistant to rational considerations.

It follows from these features of hinge commitments that they are not K-apt beliefs, even though they qualify as beliefs in the folk sense of the term.¹⁸ They are clearly beliefs in the folk sense due to how they involve an outright commitment to the truth of the target propositions (indeed, they involve a complete certainty), even if this is usually tacit. It is crystal clear from one’s actions that one is, in normal conditions, certain that, for example, one has hands. Moreover, one would not hesitate, in these conditions, to assent to this proposition if asked to do so (though it is difficult, as Wittgenstein points out, to imagine a non-philosophical reason why such an issue would ever arise in normal conditions, which is why most people would likely regard the question as puzzling, even though—and to a large extent because—they found the answer to the question to be completely obvious).

Hinge commitments cannot be K-apt beliefs, however, as they fail the test for this kind of belief that we noted above. In particular, it is characteristic of our hinge commitments that one continues to be completely certain of them even when it becomes apparent that they are groundlessly held. This reflects the fact that this is a primitive certainty that is not rooted in reasons at all. Again, then, we find there is an overlap between our hinge commitments and delusions, as the latter (as we noted above) also fail to be K-apt beliefs, and for the same reason (i.e., because they don’t satisfy the test for K-apt belief that we have noted). In addition, both delusions and hinge commitments qualify as beliefs in the folk sense.

Given these commonalities between delusions and hinge commitments, it is thus natural that some commentators might be tempted to regard the former as a species of the latter. As we will see, however, our characterization of hinge commitments has already exposed some important disanalogies between hinge commitments and delusions. The devil is thus in the detail, in that once we gain a deeper understanding of hinge commitments fundamental distinctions between delusions and hinge commitments emerge.

4 Delusions are not hinge commitments

Let’s start with the point that one’s hinge commitments are manifestations of one’s overarching certainty in one’s worldview: that one is not fundamentally mistaken in one’s beliefs. It follows that one’s hinge commitments cannot be in conflict with one’s wider beliefs, as the particular hinge commitments that one holds are effectively expressions of those beliefs. This is the sense in which one’s hinge commitments are essentially commonsense certainties, as they are nodes of one’s worldview that are such that to reject them would call one’s worldview into question. Of course, as we’ve noted, one can say that one rejects them, but the point is that such statements would be in a sense idle, as one’s actions would reveal that one is no less certain of them (just as one’s actions continue to reveal one’s overarching certainty in one’s worldview).

This feature of hinge commitments marks an important contrast with delusions. As we have previously commented, delusions are usually in conflict with other beliefs that the subject holds, as when a person with Cotard delusion sincerely claims to be certain that they are dead and yet in many respects acts as if they are still alive. This aspect of delusions is one key reason why some commentators are hesitant to call them beliefs, as we noted above. In contrast, our hinge commitments cannot be in tension with one’s wider beliefs in this way, as they are effectively expressions of one’s worldview. Indeed, it is notable that while both hinge commitments and delusions are groundless, only delusions are contrary to the prevailing evidence. In the normal conditions that hinge commitments function, there is not, for example, evidence against one’s commitment that one has hands or that one’s name is such-and-such (regardless of whether the ‘evidence’ is evaluated from one’s own point of view only, or from the point of view of the wider community). This is in marked contrast to delusions where the contrary evidence is overwhelming.¹⁹ This highlights an important difference in the way that delusions and hinge commitments fail to be K-apt beliefs, in that for the former this is because they are essentially irrational while for the latter this is because they are essentially arational.

I take this observation about how hinge commitments are not in conflict with one’s wider beliefs or the prevailing evidence to be an especially important point of contrast between delusions and hinge commitments, though it is by far not the only difference. Consider, for example, the point about hinge commitments representing one’s most commonsense convictions. This is salient here because there is nothing commonsense about delusions; on the contrary, they are clearly in tension with commonsense. Of course, ‘commonsense’ can be in a certain sense in the eye of the beholder. It used to be commonsense (in some quarters at least) that the earth was flat, for example. Accordingly, one might hold that this point is largely useless at distinguishing delusions from hinge commitments, for couldn’t they also count as commonsense, or at least not in conflict with commonsense, in the right conditions?

I think this response is not credible. For one thing, it is not clear that delusions ever represent a commonsense commitment, even on the part of the deluded individual. At the very least, however, they are very different to the kind of commonsense claim that is involved in a hinge commitment. Recall that these are propositions that are so obvious to members to that community that they are not normally even commented on. They are completely unremarkable claims that, as a consequence, ‘lie apart from the route travelled by inquiry.’ (OC, p. 88) Delusions are very different in this regard. For even if the deluded individual treats the target claim as obvious, they are constantly confronted with recalcitrant evidence, both in terms of their own experiences and beliefs that conflict with the target claim and in terms of the opposing testimony of others who are exposed to this delusion. There is nothing remotely similar at work in the case of hinge commitments.

Even setting this point to one side, there is another important aspect to the notion of commonsense that is salient here, which is that commonsense is always by its nature shared—one cannot have commonsense convictions in isolation. So while not everyone might share your conception of commonsense, it must be shared with others in one’s community or else it wouldn’t qualify as commonsense at all. Indeed, Wittgenstein emphasizes that one acquires one’s hinge commitments, just as one acquires one’s worldview, by being inculcated into a set of social practices. As he puts it, one ‘swallows down’ the hinge commitments along with the elements of one’s worldview that one is explicitly taught. (OC, p. 143) This is in marked contrast to delusions. Not only are they not the product of being immersed into a form of life—much less are they tacit commitments that are part of the necessary backdrop of a form of life—but they are also, relatedly, typically held in complete opposition to one’s wider community.²⁰

This leads to a further difference between delusions and hinge commitments, which also relates to the latter’s essentially commonsense status. For while what counts as commonsense can be to a certain extent at least a variable matter, the sense in which hinge commitments are commonsense implies a shared core of beliefs. This is because while Wittgenstein grants that there can be divergences in one’s set of hinge commitments, just as there can be divergences in worldviews, he also emphasizes that this is subject to certain limitations. In order to be occupants of the space of reasons at all—such that the moves that we make within that space are intelligible—we must have substantially overlapping beliefs.

Wittgenstein expresses this point in terms of the coherence of treating someone as making a mistake. He argues that for this to be possible one ‘must already judge in conformity with mankind.’ (OC, p. 156) Moreover, he maintains that if one’s apparent conflicting judgement is too much in conflict with what is commonly believed, as would be the case if one were to genuinely doubt a hinge commitment, then we will treat it not as a mistake but rather as a ‘mental disturbance’. (OC, p. 71) In short, rather than seeking reasons, we will seek causes (has one had a bump on the head?). This point is significant, since it shows that Wittgenstein is thinking of our hinge commitments in a way that contrasts with mental disturbances, at least mental disturbances of a dramatic kind anyway. That is, it is not our hinge commitments that are akin to mental disturbances (as the idea that delusions are hinge commitments would imply), but rather the lack of these commitments.

Moreover, while both hinge commitments and extreme mental disturbances may be properly thought of as arational, the explanation why is different in each case. Our hinge commitments are arational because they supply the framework relative to which the space of reasons functions. In contrast, extreme mental disturbances, on this picture, are arational because they lie outside of the space of reasons altogether. That is, in order to judge that someone’s behavior is irrational, such that we can make sense of them as making a mistake, one must be able to understand the person exhibiting that behavior in broadly rational terms. When the behavior is utterly bizarre, however, then the person ceases to be explicable along these lines, and hence one seeks causes for this behavior instead. On this way of thinking, then, delusions are at best irrational and at worst—i.e., when it comes to acute forms of delusion—cannot even be understood in terms of reasons at all. This makes delusions very different to the kind of arationality that applies to hinge commitments.

There are other important contrasts between delusions and hinge commitments, but I want to focus on just one more such difference for our purposes. This concerns how delusions and hinge commitments involve distinct forms of certainty. I noted above that one’s hinge commitments cannot be in tension with one’s wider beliefs since they are core nodes in one’s worldview. This highlights a crucial point about hinge commitments, which is that not just any optimal certainty counts as a hinge commitment. This is a detail that Wittgenstein emphasizes at a number of junctures, though it is sometimes overlooked. In particular, if one pays close attention to the text it becomes clear that not every certainty that Wittgenstein discusses is meant to be a hinge certainty. Instead, Wittgenstein is contrasting the special role that our hinge certainties play in our practices when compared with other certainties. What makes a certainty a hinge certainty is the way that it is central to one’s worldview, such that to call this claim into question is to call one’s worldview as a whole into question. As Wittgenstein notes, however, not all of one’s certainties are like this.

Consider, for example, one’s certainty that water boils when heated. Wittgenstein notes that if it turned out that water didn’t boil despite being heated, then one would be puzzled and one would likely seek an expert explanation of why this happened, but it would not undermine one’s worldview. In contrast, one cannot even make sense of what it would mean, in normal conditions, for one to be mistaken about one’s name. As Wittgenstein puts it, unlike the scenario where the water fails to boil, in this scenario ‘a doubt would seem to drag everything with it and plunge it into chaos.’ (OC, p. 613) Clearly delusions are not certainties in the sense that it is certain that water boils when heated, not least because the latter (unlike both delusions and hinge commitments) is known to be supported by a body of accepted evidence (even if one isn’t always able to describe this evidence in detail).²¹ Still, they are a different kind of certainty, and this is evident by the contrasting ways that one responds to discovering that they are groundlessly held.

We have previously noted that hinge commitments are not contrary to the prevailing evidence in the way that delusions are. This in itself entails that these two commitments are challenged in different ways: hinge commitments are challenged by being shown to be groundless, while delusions are challenged by being shown to be not only groundless but also contrary to the prevailing evidence. It’s true that subjects in both cases respond by retaining their certainty in the target propositions regardless. Nonetheless, there are crucial differences. The subject with the hinge commitments is not now retaining their certainty in the presence of opposing evidence, as occurs with the deluded person. This is one important reason why we treat deluded people as irrational, but we would not normally regard someone as irrational for continuing to be certain of their hinge commitments after they recognize that these commitments are groundless.

The more significant point for our current purposes is that the deluded person is typically not perturbed by this challenge at all, at least in a specific sense that is relevant for our discussion. That someone is challenging them can often be upsetting, of course. But the point is that the challenge itself is regarded as being without intellectual merit. So while they might be emotionally upset by being challenged in this way, they do not see in the challenge any cause for concern from an intellectual point of view. In contrast, as many commentators have noted, including Wittgenstein himself, the recognition that one’s hinge commitments are groundless is intellectually disturbing, even though it cannot generate actual doubt. As Wittgenstein puts it, the ‘difficulty is realizing the groundlessness of our believing.’ (OC, p. 166) Note that the difficulty in question here is not being posed by the groundlessness, but by one’s recognition of the groundlessness.

A key part of the reason why becoming aware of the groundless nature of one’s hinge commitments is disquieting is that this feature of them is ordinarily hidden in everyday life. It is not that we routinely suppose—falsely—that one does have a rational basis for one’s certainty that, for example, one has hands. We do suppose this once the issue arises, but as previously observed there are no normal contexts of inquiry where such a question is posed. That’s the nub of the matter: we are not usually even aware of the rational standing of these everyday certainties, as this is a certainty that functions as a tacit given in our practices. Accordingly, we are not aware of their groundlessness either. It is thus unsurprising that once we do become aware of them, then this can strike us as somewhat unsettling. Notice, however, that this unsettling aspect of recognizing one’s hinge commitments qua hinge commitments does not—indeed, cannot—lead to actual doubt. The anxiety in play is thus in a certain sense exclusively at higher-order and disengaged from one’s actual practices, as it cannot impact the hinge certainty that is manifest in one’s day-to-day actions. This makes the groundlessness of these certainties even more disconcerting. It turns out that that which is most everyday is in fact quite strange once properly understood; uncanny even.²²

Elsewhere I have described this intellectual anxiety that arises in response to becoming aware of one’s hinge commitments qua hinge commitments as epistemic vertigo.²³ This terminology plays on the popular (albeit false) equating of vertigo with a fear of heights to capture a kind of phobic response that one might be naturally led to when engaged in a kind of intellectual detachment (or ‘ascent’, if you will). One can be aware, when one is high up, that one is not in danger and yet naturally feel anxious nonetheless. Similarly, when one is in the kind of unusual situation of being aware of one’s hinge commitments qua hinge commitments—and hence is to this extent no longer fully attuned with one’s ordinary practices—one can naturally feel anxious about one’s rational situation. Discovering that one’s hinge commitments are groundless entails, after all, recognizing that there is no rational basis for one’s worldview, for one’s system of beliefs as a whole. Like ‘vertigo’, however, this reaction can be completely phobic, in that even if one follows Wittgenstein in coming to see that having hinge commitments is an inevitable consequence of being a rational subject, the anxiety may remain.

The crux of the matter is that there is no parallel between hinge commitments and delusions on this score. For one thing, the deluded individual will simply not recognize that their groundless certainties are groundless, even in the face of large amounts of evidence indicating that this is the case. In contrast, what prevents someone from recognizing the groundlessness of their hinge commitments is usually only a lack of the kind of philosophical insight that would bring this unusual feature of our rational practices to light. Moreover, while the certainty in the target proposition would remain in place in both cases regardless of the groundlessness of this commitment being exposed, in the case of one’s hinge commitments this certainty, manifest in one’s actions, is likely to coexist with a higher-order anxiety about the overall nature of one’s rational position. In contrast, the deluded individual will typically be entirely sanguine in the face of overwhelming rational opposition.

5 Hinge commitments and the ethics of belief

We have seen that although there are some significant overlaps between delusions and hinge commitments, there are also a range of crucial differences, such that the former cannot be plausibly understood to be a variety of the latter. I think this conclusion is important for our understanding of both notions.

On the one hand, if we align delusions with hinge commitments then we are led to a conception of the latter that is very different to that set out by Wittgenstein. The radical picture that he was proposing with regard to the structure of rational evaluation was not one on which any kind of groundless certainty that is impervious to counterevidence is thereby a hinge certainty. Instead, it is crucial to a hinge certainty in a specific proposition that it manifests the overarching hinge certainty in one’s worldview. This poses inevitable restrictions on which kinds of certainty can be hinges, with delusions now featuring within the non-hinge varieties of certainty. Although it has not been my concern to defend Wittgenstein’s proposal here, but merely to properly describe it, I suggest that recognizing this point helps highlight the plausibility of Wittgenstein’s account, even despite its radical nature. If it were true that delusions are hinge commitments, then that would obviously be a problem for any epistemology that embraces hinge commitments.

On the other hand, treating delusions as hinge commitments is also unhelpful to our understanding of delusions. This is because it is important to our conception of delusions that they capture atypical mental phenomena. And yet hinge commitments are held to be not only normal features of our rational practices but necessary preconditions of those practices. Conceiving of delusions as hinge commitments thus has the effect of normalizing a phenomenon that ought to be by its nature abnormal.

How does all this relate back to our initial concern with the ethics of belief? Recall that the issue in hand concerns how embracing hinge commitments seems to entail a particular kind of response to the ethics of belief question that allows that there can be an appropriate type of belief that is not formed on a rational basis. This would be a form of non-evidentialism, in that it would allow that a belief can be appropriately formed in the absence of supporting evidence. More specifically, one’s belief in one’s hinge commitments can be appropriate even though one lacks a rational basis for their truth, and even if one comes to recognize this fact. If delusions were hinge commitments, then this would throw a further spanner into the works, as we would now have a form of appropriate belief that is not just arational (like hinge commitments) but essentially irrational. Fortunately, as we have noted, delusions are not hinge commitments, so this complication doesn’t arise. Moreover, we have also seen that the debate about both delusions and hinge commitments highlights the need to distinguish between two notions of belief: folk belief and K-apt belief. Despite their other differences, both delusions and hinge commitments share the feature of being beliefs only in the folk sense, and not also in the K-apt sense.

That our hinge commitments are not K-apt beliefs offers us a way of embracing this notion while nonetheless responding to the ethics of belief debate along broadly conservative lines. The thought would be that the ethics of belief debate, properly understood, should be cast at the level of K-apt belief rather than folk belief. It is only K-apt belief, after all, that is a plausible candidate for the intuition that appropriate belief is by its nature belief that is responsive to rational considerations. Since our hinge commitments are not K-apt beliefs, they pose no challenge to the conservative response to the ethics of belief debate when that debate is understood as exclusively concerned with K-apt belief.

Moreover, insofar as one construes the ethics of belief debate specifically in terms of the folk notion of belief, then the non-evidentialist option just noted looks to be both inevitable and also largely harmless. The folk notion of belief, as we have noted, is so liberal that it encompasses a range of propositional attitudes, at least some of which do not seem to bear any essential conceptual connection to reasons and truth. It is no wonder, then, that non-evidentialism might be the upshot of the ethics of belief debate if belief is understood along these lines. (In contrast, notice just how counterintuitive non-evidentialism would be if applied to the ethics of belief debate construed in terms of K-apt belief).

The merit of engaging with the ethics of belief debate via consideration of the propositional attitudes involved in hinge commitments and delusions is thus that it enables us to make vivid the importance of the distinction between the liberal notion of folk belief and the more restrictive notion of K-apt belief. Once that distinction is independently motivated, however, then it can be applied to the ethics of belief debate in a way that demonstrates, at the very least, that embracing hinge commitments needn’t entail that one is committed to adopting a non-evidentialist response to that debate that applies to K-apt belief. More generally, the moral to be extracted from our discussion is that getting a handle on the ethics of belief debate requires us to examine more closely what is meant by belief in this regard.²⁴