Incoherence, inquiry, and suspension

Abstract

I consider two possible evidentialist responses to Schmidt. According to the first, all of the reason-giving work in the relevant cases is being done by evidence. According to the second, even if the ‘incoherence fact’ sometimes provides a reason, what it provides a reason for is not a doxastic attitude, or at least not one that is an alternative to belief. I argue that the first response is not satisfying, but the second is defensible.

1 Introduction

If your beliefs are incoherent, consider doing something about it. That seems like good advice. This seems to support the following claim: if your beliefs are incoherent, then you have a reason to do something that will resolve the incoherence. But what exactly do you have a reason to do, and what fact provides that reason?

In her illuminating and thought-provoking article, Eva Schmidt argues that in some cases, the fact that your beliefs are incoherent—the incoherence fact—itself provides a reason, and what it provides a reason to do is suspend judgment on all of the propositions that constitute the incoherence. The incoherence fact does not itself seem to be evidence as to which, if any, of the believed propositions is true. Thus, Schmidt claims, in these cases the incoherence fact is a non-evidential epistemic reason for suspension. Therefore, not all epistemic reasons are given by evidence.

Schmidt’s argument is powerful. In this commentary, I will consider two possible evidentialist responses. First, the evidentialist may maintain that, in the relevant cases, all of the reason-giving work is being done by evidence after all. While this response initially looks plausible, I will suggest that it may not fully account for the phenomena. Second, the evidentialist may maintain that, even if the incoherence fact sometimes provides a reason, what it provides a reason for is not a doxastic attitude, or at least not one that is an alternative to belief. I will suggest that this response is more promising, mainly by raising doubts about Schmidt’s claims concerning incoherence, inquiry, and suspension. I will close by briefly raising a question about the role of coherence in relation to intellectual virtue.

2 What gives reasons?

On the face of it, there is a straightforward evidentialist account of why you have reasons to revise your attitudes when you are incoherent. Suppose, for instance, that you believe <p> and you believe <q>, and it comes to your attention that these propositions are inconsistent. In that case, your evidence for <p> is evidence against <q>, and conversely. So, plausibly, the evidence for <p> is a reason to stop believing <q>, and conversely. Moreover, abandoning either belief would resolve the incoherence. Thus, in this case, it seems that we can vindicate the idea that you have a reason to do something that will resolve the incoherence in purely evidentialist terms. Your evidence provides reason to abandon one or both beliefs, thereby restoring coherence.

Schmidt presents three cases in which she takes it that the incoherence fact itself provides a reason to revise your attitudes. While they each raise different issues, it seems to me that in all three of them, the evidentialist might argue that it is really evidence that is doing all of the reason-giving work. History v. Philosophy involves flatly inconsistent beliefs and as such is potentially amenable to the sort of evidentialist treatment described above. In Marple and Poirot, the incoherence-constituting beliefs (that not-v, and that the evidence indicates that v) are not inconsistent, but it is nonetheless plausible that evidence for each is evidence against the other. Thus the evidentialist can again claim that the evidence gives a reason to abandon one or both beliefs.¹

There are different ways in which evidentialists might handle 6/49 Lottery. Some might claim that, for each belief in the inconsistent set, the fact that it could easily be false is evidence that provides some reason for suspension. Others might argue that purely statistical evidence does not provide sufficient reason for belief, and so suspension on all of the propositions is the appropriate response here—but this is because the evidence fails to provide sufficient reason to believe any of them, not because of anyone’s incoherence.²

In support of these treatments, the evidentialist might argue that suspension seems no worse supported in variants of these cases where the subject has not yet formed any doxastic attitudes, and so there is not yet any incoherence fact to provide a reason for anything. For instance, if you are confronted with Lola’s evidence in 6/49 Lottery before forming any attitude to any of the propositions in the inconsistent set, you arguably have reason to suspend on all of them. This can’t be because you have incoherent beliefs, because you don’t. But then it seems that, even if Lola has incoherent beliefs, this fact is not required in order to explain why she has reason to suspend—we can instead appeal to whatever facts explain why you, who are not incoherent, should also suspend.³

All of that said, it does seem plausible that incoherence facts can have a kind of normative significance of their own. Consider: if you believe <p> against the evidence, you make a mistake. But if you believe <p> against the evidence and also believe <q>, which is inconsistent with <p>, you seem to be making two mistakes, not just one. But <q> may be sufficiently supported by the evidence, and indeed true. In that case the second mistake is not another one of believing against the evidence, or believing falsely. If this is right, it seems that the incoherence itself must be a mistake. And you have reason to rectify both mistakes – to stop believing against the evidence, and to stop being incoherent. Of course, you could rectify both mistakes at once by abandoning your belief that p. But if you were instead to abandon your belief that q, thus resolving the incoherence in, so to speak, the wrong direction, you would arguably still be improving matters in one respect, even if making them worse in another. Thus, it’s not clear that the evidentialist strategy described above can capture all of the mistakes made by the incoherent, or all of the reasons they have.⁴

3 Reasons for what?

Suppose, then, that incoherence facts at least sometimes provide reasons. Schmidt’s claim is more specific: at least sometimes they provide reasons to suspend on all of the believed propositions. We could accept that incoherence facts provide reasons for something without accepting this more specific claim about what they provide reasons for. Perhaps incoherence facts as such provide reasons only for, say, reconsidering your attitudes, or revising them somehow, and the evidence then determines which particular revisions you have reason to make.

However, Schmidt has an intriguing argument for her claim. It turns on the connections between incoherence, inquiry, and suspension. In the relevant cases, she claims, incoherence gives a reason ‘to re-open inquiry into all matters about which one had previously formed belief’ (Schmidt, this vol., Sec. 4). And inquiry ‘presupposes’ suspension on all of those matters (ibid.), since genuine inquiry requires an open mind.⁵ So, incoherence provides a reason to suspend on all of them.

I have three doubts about this argument. First, even if inquiry requires suspension, does it follow that a reason for inquiry is a reason to suspend? It would if reasons for responses are always transmitted to necessary means for those responses. But, while they may do so in the case of action, it is less clear that reasons transmit to necessary means when the means are attitudes. At any rate, those who deny that there are pragmatic or otherwise ‘wrong-kind’ reasons for attitudes will deny that reasons generally transmit to necessary attitudinal means, since such reasons, if genuine, would often be pragmatic or ‘wrong-kind’.⁶ Indeed, this would be true of many reasons for inquiry if they transmitted to suspension. The fact that you need to stop at the shop might be a reason to inquire into whether it closes soon, and responding to this reason might entail suspending on whether it closes soon. But the fact that you need to stop at the shop does not seem to be a ‘right-kind’ reason to suspend on whether it closes soon.

Schmidt might reply that reasons for inquiry nonetheless sometimes transmit to suspension, and, in particular, that reasons given by incoherence do so. If so, I would like to understand why they transmit in some cases and not others.

Here is my second doubt about the argument: is it obvious that inquiry requires suspending on all matters inquired into? Certainly, inquiry requires a kind of openness to its own outcome—by the nature of inquiry, its conclusion is not settled in advance. But being open as to whether p or q in the context of your inquiry is not obviously the same as suspending judgment on whether p and on whether q. Suspension of judgment is widely thought to be an attitude in its own right, and thus a kind of commitment—‘committed neutrality’, as Sturgeon (2010) puts it.⁷ As such, to suspend judgment is to have, temporarily at least, a settled, albeit neutral, stance on a question. So understood, suspension seems more like a possible conclusion of inquiry than an attitude you must hold while inquiry is ongoing and the question remains unsettled.

Schmidt does not think of suspension in this way. Following Staffel (2019), she suggests that suspension can be a ‘transitional’ attitude, one held while inquiring only to be abandoned when, all going well, inquiry concludes in belief or disbelief.

Fair enough. It’s not clear to me that inquiry requires you to hold even a transitional attitude of suspension. Couldn’t you inquire while not yet having any attitude at all, or while ‘bracketing’ any attitudes you have? And even if inquiry does require transitional suspension, is transitional suspension an alternative to belief, in the sense that a reason for it is a reason against belief? If not, then it may yet be that all epistemic reasons for and against belief and doxastic attitudes that are alternatives to belief are provided by evidence. This would presumably go a long way to satisfying the evidentialist.⁸

Let me briefly elaborate on this point. Belief is not itself a transitional attitude; you can’t undertake the commitment involved in belief only ‘for the time being’, intending to change your mind later. It is a possible conclusion or outcome of inquiry rather than a part or condition of ongoing inquiry. A natural thought is that the alternatives to belief must be other possible conclusions or outcomes of inquiry. This might include a settled (or ‘terminal’) attitude of suspension, but not a transitional one whose point is precisely to enable inquiry while it is ongoing.⁹

A related natural thought is that alternatives to belief must be such that the considerations that are reasons for them must be able to compete, within reasoning, with reasons for belief. It is not clear that this condition is met by transitional suspension. Reasons for it do not seem to bear on the same question as reasons for belief. Consider: ‘Is it the case that p? Well, on the one hand, the reliable experts say that p. But on the other hand, there is reason to inquire into whether p and doing so requires suspending on whether p’. The two parts of this purported bit of reasoning do not seem to engage with each other—they bear on different questions.

Or consider, most pertinently: ‘On the one hand, the reliable experts say that p. But, on the other hand, I have incoherent attitudes towards <p>’. Here again, the second part of the purported reasoning seems to change the subject. The incoherence fact does not seem to compete with the first, evidential fact.¹⁰ Thus, if the incoherence fact is a reason for some attitude, it is not clear that this attitude is an alternative to belief. Indeed, this line of argument does not seem to depend on exactly how we conceive of the attitude that the incoherence fact supports, nor on its relation to inquiry. It thus directly threatens Schmidt’s main thesis, not only the argument for it I am discussing here.

Schmidt may avoid this worry by rejecting my assumption about how reasons must be able to feature in reasoning. This seems to me an unattractive move, but Schmidt might be happy to make it.¹¹

My third doubt about Schmidt’s argument concerns whether you have a reason to inquire into a matter just because you have incoherent attitudes about it. Inquiry, even if not very effortful, costs scarce cognitive resources. Presumably, reasons to inquire are given by the interest or importance of a subject matter, and/or by the prospect of acquiring knowledge about it. If a matter is dull, unimportant, and unknowable, it’s not clear why your having incoherent attitudes about it is a reason to inquire into it. For example, you have, let’s suppose, no reason to inquire into the exact lengths of the blades of grass in your garden. If you form an incoherent set of beliefs about that matter, do you thereby give yourself a reason to inquire into it?¹²

Of course, Schmidt may say that incoherence gives a reason to inquire only when the subject matter of the incoherent beliefs is interesting, important, and knowable. But in such cases, we might wonder whether the incoherence is really contributing to the reason. The interest, importance, and prospect of acquiring knowledge seem sufficient to explain why there is a case for inquiry.

In sum, and in reverse order: I am not sure that incoherence gives a reason for inquiry, nor that inquiry requires an attitude of suspension that is an alternative to belief, nor that a reason to inquire would necessarily be a reason to suspend even if inquiry did require suspension.

4 Intellectual virtue and coherence

My discussion has ignored an important claim of Schmidt’s. Incoherence provides epistemic reasons for suspension, she claims, because ‘the intellectually virtuous response to incoherent doxastic attitudes is to suspend, so that one can properly reevaluate one’s epistemic situation’ (Schmidt, this vol., Sec. 5). Schmidt’s thought here may be that the intellectually virtuous agent will, on discovering incoherence, inquire into the matter(s) on which they are incoherent, and they must suspend judgment on all relevant propositions in order to do so. This would be a way of supplementing the account that I expressed doubts about above; the doubts would still apply. Alternatively, though, Schmidt could simply claim that coherence itself is a requirement or aim of intellectual virtue, and that coherence mandates suspension in the relevant cases. In this way, she could argue for coherence-based epistemic reasons without relying on claims about inquiry, and thus without facing the issues I raised in the previous section.

Of course, the claim that intellectual virtue aims at coherence would raise its own questions. What is it that philosophers who see nothing to be said for coherence as such (e.g. Kolodny, 2005) are missing? How does the aim of coherence relate to other candidate aims of intellectual virtue, such as truth, knowledge, understanding, humility, or wisdom? But there may be plausible answers to these questions. If so, Schmidt’s reflections on intellectual virtue, coherence, and reasons will doubtless lead us towards them.¹³