Replies to critics

Abstract

In these replies, I react to comments on my paper “Facts about Incoherence as Non-Evidential Epistemic Reasons”, provided by Aleks Knoks, Sebastian Schmidt, Keshav Singh, and Conor McHugh. I discuss potential counterexamples to my claim that the fact that the subject’s doxastic attitudes are incoherent is an epistemic reason for her to suspend; whether such incoherence-based reasons bear on individual attitudes or only on combinations of attitudes; the prospects of restricting evidentialism about epistemic reasons to reasons to believe; whether incoherence-based reasons are truly epistemic; the alleged normative and motivational expendability of incoherence-based reasons; the possibility of incoherence-based reasons to suspend without actual belief in the incoherent propositions; the relationship between suspension, inquiry, and incoherence; and the nature of suspension of judgment.

Let me start by thanking Jie Gao, editor of the Asian Journal of Philosophy, for inviting me to write the lead article of this symposium. This is a great format for philosophical debate, and it has been a thrilling experience to have my work scrutinized by four such sharp-sighted philosophers. I am very grateful to Aleks Knoks, Sebastian Schmidt, Keshav Singh, and Conor McHugh for taking the time to engage with my work and to provide penetrating and thought-provoking comments on my paper “Facts about Incoherence as Non-Evidential Epistemic Reasons.” I will use this response to clarify some of the assumptions I make there and to try to defend my proposal. Due to space limitations, I am unable to address all questions that were put to me, but will focus on the ones I found most pressing.

As a reminder, here are the three cases of incoherent attitudes from the paper (Schmidt, 2023, 7/8, 10, 12)¹:

6/49 Lottery

Lola participates in a lottery in which each player chooses six numbers from 1 to 49, and wins the jackpot if her numbers match the six numbers produced in the drawing. The probability that she will win the jackpot is 1 in 13,983,816. There is no guarantee that anyone will win the jackpot. The lottery has millions of regular participants, and as a matter of fact, the chance in any drawing that at least one player wins the jackpot is extremely high. To fix ideas, say that over the last five decades, there has been only 1 in 1000 drawings in which no ticket won the jackpot. Lola is aware both of the extremely high chances of losing of every single ticket and, by way of induction, of the extremely low chances that everyone’s tickets will lose (not win the jackpot). It is then extremely probable, from Lola’s point of view, for ticket 1, that it will lose; for ticket 2, that it will lose; …; for ticket n, that it will lose (call the respective propositions “p₁,” “p₂,” …, “p_n”). But at the same time, it is extremely probable for her that it is not the case that ticket 1 will lose and that ticket 2 will lose, …, and that ticket n will lose (call this proposition “p_□”).

Marple and Poirot

Miss Marple and Hercule Poirot team up investigating a murder. Master detective Miss Marple is first on the scene and takes in all the evidence, forming the (for once, mistaken) belief that the evidence indicates that the vicar did it (v), and she tells Poirot so. That Miss Marple provides this testimony is a sufficient reason for Poirot to believe that the evidence indicates that v, and he forms the belief for that reason. Next, Poirot himself takes in the evidence at the crime scene, which as a matter of fact indicates that the vicar did not do it; he therefore has sufficient reason to disbelieve v, and disbelieves v for that reason. Poirot now has incoherent doxastic attitudes, belief that the evidence indicates that v and disbelief that v. They are incoherent because by virtue of his belief about the evidence, he accepts that there is sufficient evidence and thus reason to believe that v, and thus that belief that v is the correct doxastic response; but nonetheless, he disbelieves v.

History vs. Philosophy

When Basna studies history, she comes to believe, based on her professors’ arguments, that the historical facts are relative to the historian who interprets historical sources. Basna later switches her focus to philosophy and, based on her philosophy professors’ arguments, she forms the belief that no facts are relative to anyone. Both beliefs are supported by sufficient evidence, which was provided, respectively, by Basna’s history professors and by her philosophy professors. Her evidence comprises both the professors’ expert testimony and the arguments with which she engages. At some point, Basna realizes that the two beliefs are inconsistent.

1 Response to Aleks Knoks

Aleks Knoks raises one principled worry for my attempt to spell out how epistemic reasons to suspend go beyond evidence and then presents several potentially troublesome cases. The worry is that my conceptual commitments imply on their own that all reasons to suspend are non-evidential (Knoks, 2023, 2/3). But this result would immediately undermine evidentialism, which would make my argument beside the point, and would also be unfair towards the evidentialist, who can be expected not to share my conceptual presuppositions. As Knoks points out, I start my investigation with the assumption that evidence indicates the truth or falsity of the content of a doxastic attitude. Further, in my discussion of the counterexample from the fact believed itself, I grant that evidentialists may endorse a (narrow) probability-raising conception of evidence (p. 7). Finally, I follow Friedman (2013) in holding that suspension is the attitude of remaining unsettled with regard to the question of whether p. If this picture is assumed, Knoks suggests, it’s no wonder that reasons to suspend come out as counterexamples to evidentialism. How could even clear candidates for reasons to suspend, such as the fact that the subject’s evidence is balanced, ever come out as factors that raise the probability of the content of a state of suspension, or that indicate the truth or falsity of its content?

I thank Knoks for giving me this opportunity to clarify the dialectic of my paper: I indeed start with a conception of evidence concerning p as what indicates the truth or falsity of p. In looking at cases of higher-order evidence, however, I deliberately move beyond that conception (because I agree that settling evidentialists with this overly narrow understanding of evidence is uninteresting) to the view that evidence, for the evidentialist, should be taken to include facts about first-order evidence. Call evidence according to this broader understanding “evidence_e.” In particular, as I claim on p. 5, “higher-order evidence that the subject’s evidence concerning p does not settle the issue [of whether p] is a reason to suspend on p.” Evidence_e is what I presuppose in the paper, and this conception explicitly allows that evidence bearing on p in the described indirect way is an epistemic reason to suspend. Indeed, it seems unproblematic that evidence that an issue is not settled can bear positively on the correctness of a question-directed attitude of unsettledness. When I grant that evidentialists may endorse the probability-raising conception of evidence to respond to the counterexample from the fact believed itself, I do not myself endorse this conception. It is open to evidentialists to deal with that counterexample in a different way, for instance by denying that the subject believes for a reason at all in cases where she believes that p directly in response to the fact that p. This is also why I don’t press evidentialism on potentially problematic implications of the probability-raising conception. I hope this makes it clear that I engage evidentialists on the intended point: on failing to account for incoherence facts as reasons to suspend.

The first case presented by Knoks is an intriguing example involving a subject who detects that his doxastic attitudes are incoherent and who also learns that he has taken a drug that causes him to hallucinate incoherencies between his doxastic states where there are none (Knoks, 2023, 5). In Marple and Poirot, imagine that Poirot discovers a genuine incoherence between his first-order disbelief about the vicar and his higher-order belief about what the evidence indicates about the vicar; but is additionally aware that this incoherence may be merely apparent and his impression induced by the drug. Knoks argues that in this case, Poirot can reasonably hold the following, itself incoherent combination of attitudes: disbelief that the vicar did it; belief that the evidence indicates that the vicar did it (this is the incoherent combination from the original case); suspension on whether these two doxastic attitudes are really incoherent. We might say that in virtue of his background knowledge about the drug, which affects his capacities, the incoherence fact has no normative impact as a reason to suspend on his conflicting attitudes.

In response, note first that I am not committed to saying that, in every case of incoherent attitudes of which the subject is aware, she is all-things-considered justified to suspend on these attitudes. My argument against (R-E) requires just one case of an epistemic reason that cannot be traced back to evidence. My proposal is compatible with different ways to make sense of Knoks’s particular case. (1) The fact that Poirot took the incoherency-hallucination drug is a disabler which affects his normative reason to suspend on the original belief and disbelief, i.e., the fact that these two attitudes are incoherent. The incoherence fact then plausibly remains a reason, but due to the doubt cast on the veracity of this reason by the fact that Poirot took the drug (which in its turn favors suspending on whether the incoherence fact obtains), the reason is blocked from actually supporting suspension. (2) I can give the account just provided but highlight the further fact that his attitudes appear to be incoherent. This appearance fact is still a reason the subject has to step back from this (now even bigger) mess of attitudes, so as to figure out whether something went wrong, and if so, where. He is unable to rely on his own judgment to find out whether something went wrong, but he could ask a friend who wasn’t drugged, or simply wait until its effect have waned and then assess the evidential status of his doxastic attitudes. The appearance fact is plausibly a weaker reason to suspend, since the subject cannot be sure that things are as they appear, but it could still outweigh his original evidential reasons to (dis)believe. Which approach is suitable will depend on the further details of the case, for instance on whether the subject knows or only suspects that he was drugged; on whether his attitudes are often incoherent or on whether this is highly unusual (as we can imagine with Poirot); or on how obvious the incoherence is.

The second problem case provided by Knoks (2023, 6) involves a subject who arrives at incoherent doxastic attitudes by way of wishful thinking. Marjorie disbelieves that the 2020 US presidential elections were manipulated. But as a Trump fan, she unconsciously relies on wishful thinking and forms the belief that her evidence indicates that the 2020 US presidential elections were manipulated.² Since this case involves incoherent attitudes just like Marple and Poirot, it looks like I have to accept that Marjorie has a sufficient reason to suspend judgment—the fact that her attitudes are incoherent. In another variant, Marjorie deliberately gets herself to adopt the higher-order belief that is incoherent with her ground-level disbelief that the elections were manipulated, just to have an excuse to suspend on that belief. Say, she is able to consciously activate her wishful thinking mechanism to do so. In either case, it is implausible that a subject can affect what normative reasons she has by improperly forming beliefs, or so Knoks argues.

I agree that deliberately adding incoherent beliefs to one’s belief set so that one can suspend on unwanted beliefs doesn’t generate reasons to suspend. For when Marjorie deliberately forms the higher-order belief that the evidence indicates that the elections were manipulated, she thereby acquires a reason to believe that this belief is improperly formed (viz. the fact that she formed it deliberately by wishful thinking). This is a reason to drop the belief. It further undermines the incoherence fact as a reason to suspend—given that she knows that she used wishful thinking in forming the belief, there is no reason for her to re-open the question of whether (the evidence supports that) the elections were manipulated, since she already knows where the error lies. By contrast, if Marjorie unwittingly ended up with this higher-order belief via wishful thinking, I think that my analysis from Marple and Poirot stands. Marjorie finds herself with incoherent beliefs, and it is not immediately obvious to her how she got there. To find out how she ended up with these incoherent beliefs, she should re-open inquiry, and this means at the same time that she has a reason to suspend on the incoherent beliefs. This will put her in a position to realize that all the experts say that the elections were not manipulated, that the voting results indeed favored Biden, and so on, and so that there is no support for the claim that the evidence indicates that the elections were manipulated.

2 Response to Keshav Singh

Keshav Singh pushes me to clarify three aspects of my proposal. First, he argues that an incoherence fact is a reason to suspend on all of the attitudes that give rise to the incoherence together, and not on the attitudes individually (Singh, 2023, 3/4).³ For instance, in 6/49 Lottery, the fact that Lola’s beliefs about the tickets are incoherent is one single reason to suspend on all these beliefs together. At the same time, Singh argues that I am committed to holding that the incoherence fact is not a reason to suspend on all these attitudes together, but on each of them individually. Accordingly—suggests Singh—I have to claim that the fact that Lola’s beliefs are incoherent is a reason for her to suspend on whether ticket 1 will lose, and a distinct reason to suspend on whether ticket 2 will lose, …, and a distinct reason to suspend on whether it’s false that all tickets will lose. He points out that this cannot be right. Not only does it sound odd to say that Lola has a huge number of reasons to suspend (one for each belief involved), this view also erroneously implies that Lola gets something right when she suspends only on some of her beliefs, but not on all of them. But the right way for her to respond is to suspend on all attitudes together, and her reasons don’t favor suspending on just some of them.

I agree with Singh. It can’t be that Lola has many distinct reasons to suspend, each relating separately to one individual belief. Lola has a reason to suspend exactly insofar has she has a reason to take a step back and reconsider her overall epistemic situation with respect to the 6/49 lottery, i.e., insofar she has a reason re-open simultaneously all the relevant issues that she previously took to be settled, i.e., concerning each ticket, whether it will lose, together with the issue of whether all tickets will lose. It seems exactly right that the fact that her beliefs are incoherent is one reason for her to suspend on all of the issues she took for settled together.

I deny, however, that my proposal commits me to the claim that the incoherence fact constitutes many individual reasons to suspend on the relevant issues separately. Rather, my view is naturally compatible with the attractive picture painted by Singh. When I discuss this issue in the paper (pp. 18/19), I do not object to the idea that incoherence facts are reasons to suspend on all the involved doxastic attitudes together. Rather, I reject the assumption that this means that they are not reasons bearing on attitudes at all, but only reasons to (mentally) act a certain way, in particular, to deliberate a certain way (see Worsnip, 2021). My argument there is that the correct mindset for genuine inquiry is to be suspending on the issues into which one inquires, and so that a reason to re-open inquiry is at the same time a reason to suspend on all the relevant issues or attitudes together. Singh and I are in agreement on how incoherence-based reasons to suspend function.

The second clarification requested by Singh concerns the scope of (R-E) (Singh, 2023, Sect. 3), which states that epistemic reasons for doxastic attitudes quite generally are exclusively provided by evidence. But now note that the counterexample of incoherent doxastic attitudes does not speak to the issue of whether reasons to believe or disbelieve are provided by evidence. Correspondingly, if the counterexample is successful, it shows at most that reasons to suspend are not due to evidence. So, evidentialists can evade my objection by endorsing a narrower principle:

(R-E_B) All reasons to (dis)believe that p are provided by evidence concerning p.

Moreover, my counterexample focuses on reasons to suspend understood as a transitional attitude. The cases are concerned with reasons to re-open inquiry and thus suspend, which is to say, not reasons to suspend in a way that terminates deliberation. So arguably, even a slightly different principle remains untouched:

(R-E_TERM) All reasons for terminal doxastic attitudes towards p, including terminal suspension of judgment, are provided by evidence concerning p.

As to the last point, I disagree. There are cases of terminal suspension of judgment that are favored by incoherence-based reasons. One of them is Lola’s case, after she has re-assessed her epistemic situation concerning the lottery propositions several times without being able to find out which belief should be dropped (p. 17). In light of this, she has come to realize that further deliberation is pointless. Given the fact that her attitudes are incoherent and that further deliberation is pointless, it is plausibly epistemically justified for her to terminally suspend judgment on all involved propositions for good. This fact is then a reason for her to suspend, in the sense of a terminal attitude, which ends all deliberation on the matter. So, my counterexample undermines (R-E_TERM) as well.

I agree with Singh, however, that I haven’t been able to attack (R-E_B) with my cases. Withdrawing from (R-E) to (R-E_B) is indeed a strategy available to evidentialists. However, this comes at the steep price of theoretical unity regarding epistemic reasons, and of failing to provide an account of epistemic justification that is purely in terms of evidence. If evidentialists⁴ could maintain the more general (R-E), they would have a nice unified picture of how epistemic reasons determine the overall justificatory status of doxastic attitudes concerning a proposition p. For the only items balanced against each other would be factors that indicate (or raise the probability of) the truth or falsity of p or of other propositions concerning p. The interplay of these purely evidential factors, as the balance of reasons, would determine the overall justificatory status of a subject’s doxastic attitudes regarding p, be it belief, disbelief, or suspension. They might even rely on a Bayesian formalization to express their view.

But now consider the picture of the (R-E_B) evidentialist. She needs to factor non-evidential epistemic reasons, in particular incoherence facts, into the determination of which attitude towards a proposition is justified all things considered, together with the truth-indicating factors that bear on the justification of belief and disbelief alone. The big stumbling block for the evidentialist picture is that, since the justificatory status of belief, disbelief, or suspension is fixed by the balance of epistemic reasons bearing on these three competing attitudes concerning a certain issue, purely evidential factors cannot determine all by themselves which of the three attitudes vis-à-vis the issue is justified. In other words, to accept (R-E_B) instead of (R-E) is to give up what is attractive about evidentialism anyway. It makes it impossible to account for even the justification of belief and disbelief in purely evidential terms.⁵

Finally, Singh asks me whether it isn’t better to conceive of my counterexample cases as right-kind reasons that aren’t epistemic. His motivation is that incoherence facts can be reasons to suspend and give up on attitudes beyond belief or disbelief, a worry I discuss in the original paper (pp. 18/19). Singh presents a problem case for my strategy there, which involves an irrational means-end belief. Here is his case:

Imagine a friend is having a bad day, and I intend to cheer them up. Bizarrely, I believe that the only way to cheer them up is to gift them a saucer of mud. But I don’t intend to gift them a saucer of mud. Given that the locus of my failing here is in my bizarre means-ends belief, my failing seems to be a theoretical one, not a practical one. (Singh, 2023, 7)

Singh argues that in this case, the failing is epistemic, since it concerns the obviously false means-end belief, whereas the intentions I have are in order. However, since (on my picture) this is a case where I fail with respect to standards of correct practical reasoning, my proposal entails that the incoherence fact here is a practical reason to suspend. I agree with Singh that I have an epistemic problem in the scenario, since my belief about what would cheer up my friend has taken complete leave of reality. However, this epistemic problem is independent of my incoherence: It concerns my means-end belief separately, which is off-track no matter whether considered together with my intentions or not. I make a further mistake by not bringing my intentions in line with my means-end belief. The resulting incoherence gives rise to my reason to take a step back and re-evaluate all three attitudes with a view to resolving the incoherence.⁶ As in the example about the climate crisis protest in my paper (p. 19), the incoherence-based reason I have in Singh’s example corresponds to a failure of mine to live up to standards of practical reasoning, which is a failing of practical virtue. I either fail to be motivated by practical reasons that I should be motivated by (my friend’s need, what would make them happy), or I fail to correctly represent what practical reasons there are for me (is my friend’s happiness really a reason for me to give them a saucer of mud?). Which of these it is I have practical reason to figure out, and so I have a reason of practical virtue to suspend on all involved attitudes.

Singh presents a further intriguing example, in which I hope to live forever, something I simultaneously believe to be impossible. My hope and belief are incoherent, since to hope for something is to believe it possible (Bloeser and Stahl, 2022). The incoherence between these two states is a reason I have to suspend on both of them. Is this a practical or an epistemic reason? Singh suggests that we cannot nonarbitrarily determine on which side this reason falls, and also that it does not really matter where it falls. I disagree. True, hope is a state that is constrained by both practical and theoretical rationality, but that hope is appropriate only towards the possible falls out of its epistemic/theoretical dimension (Bloeser and Stahl, 2022). By contrast, to hope for an event that I believe to be practically catastrophic (say, that I get murdered in the street) is practically problematic, and the failing pertains to hope’s practical dimension. Returning to Singh’s case, I submit that we are here faced with an epistemic incoherence-based reason, since what’s at issue is an incoherence that relates to a failure to live up to a standard of correct theoretical reasoning: Hoping that I will live forever implies belief that it is possible for me to live forever, but I simultaneously believe that it is impossible to live forever. So my beliefs are inconsistent, and it is intellectually virtuous of me to suspend on both mental states and to figure out where I went wrong.

3 Response to Sebastian Schmidt

In his comment, Sebastian Schmidt presents an elegant argument to the effect that incoherence facts are normatively and motivationally superfluous, that is to say, that they have no justificatory or motivational work to do; all justificatory statuses of or transitions to doxastic attitudes can be fully explained by evidence. His two-pronged strategy is to show that, first, in cases of irrational incoherence between (dis)beliefs, this is because the given evidence already favors suspending, so that incoherence-based reasons have no more work to do. Second, in cases of harmless incoherence, it is not irrational to have incoherent attitudes. At the same time, this is exactly what evidentialism predicts—the given evidential reasons do not give rise to a reason to suspend (and so again, the evidence is all that does any justificatory work).

I focus on the first part of Schmidt’s challenge, since I want to allow that there are cases of incoherent doxastic attitudes where it is all-things-considered justified to maintain these attitudes. Schmidt discusses two ways in which the evidence by itself gives subjects sufficient reason to suspend. The first way can be illustrated by cases of balanced evidence. For such cases, Sebastian Schmidt (2023, 4) argues “Basna has evidence that her evidence does not settle the issue: she has testimony from two experts and arguments from two experts that, it seems, roughly balance out. She therefore has a reason to suspend judgment that is provided by evidence that her evidence does not settle the issue.” The fact that Basna’s beliefs are inconsistent isn’t needed to explain the reason to suspend that she has—the real reason is the fact that the evidence backing both inconsistent beliefs is balanced. The incoherence fact is superfluous in our account of the normative features of the situation.

In response, not all versions of History vs. Philosophy are cases of balanced evidence. Consider the following scenario, where Basna’s evidence from her different professors is imbalanced: Basna’s history professor supported his claim, that the historical facts are relative, by no argument at all, and Basna formed the corresponding belief on the basis of nothing but the authority of a historical expert. It will, however, take Basna a little bit of effort to recover the fact that no argument was given from her memory. By contrast, Basna’s philosophy professor supplemented her claim that no facts are relative by a clear and compelling argument, which Basna is able to recall by concentrating for a bit. This belief is based on authority and a good argument. Consequently, Basna’s overall evidential situation is not balanced, but favors the belief that no facts are relative. Nonetheless, when she realizes that her beliefs are inconsistent, this is a reason she has to suspend on both beliefs, to impartially re-assess her situation, and to abandon the belief acquired in history class. Importantly, since Basna needs to invest some time and effort to recall what arguments were or weren’t given by her respective professors, she is unable to just drop the historical relativist belief immediately. That is why suspension is justified for her.

6/49 Lottery is such a case as well—the probabilities of the propositions concerning the individual tickets are much higher than that of p_□, the proposition that not all tickets will lose. So it is not a fact that the evidence is balanced. In such cases, Schmidt can fall back on the second way in which incoherence facts are normatively superfluous. Sebastian Schmidt (2023, 7) claims that “inconsistency is conclusive evidence that the inconsistent set of beliefs is false.” More fully, he argues that the inconsistency of a set of beliefs is evidence that the whole set of these beliefs is false, rather than evidence that any one individual belief in the set is false. Each individual belief is supported by the evidence in my cases, but the whole set is not, since it’s a fact that the whole inconsistent set cannot be true. In light of this, the fact that Lola’s beliefs are inconsistent, as evidence that the set of beliefs is false, is a reason for her to disbelieve this set. At the same time, Lola has strong evidence supporting each individual belief. That this is a confusing evidential situation is itself evidence that her evidence doesn’t settle the issues—and this evidence is her reason to suspend.

Schmidt’s claim, that the incoherence fact is a reason to disbelieve the whole set of beliefs by virtue of being evidence that the set is false, is not as unproblematic as one might think. Whether it might be true depends on how we understand his talk of a “set of beliefs.” First, if we take it literally and consider a set as a mathematical entity consisting of all the relevant beliefs, this is not the kind of thing that can be true or false. Only the individual beliefs that are its members can be true or false. Second, we might understand Schmidt’s claim more plausibly as saying that the conjunction of the beliefs (in the sense of believed propositions) is false. The conjunction of the beliefs in 6/49 Lottery is indeed false, since it is a conjunction of beliefs that, taken together, are inconsistent. However, this understanding of Schmidt’s claim faces another problem. For now, the claim is that the fact that Lola’s beliefs are inconsistent, as evidence that the conjunction of all these beliefs is false, is a reason to disbelieve the conjunction. But we can plausibly describe 6/49 Lottery so that Lola does not and cannot have the conjunctive disbelief. With the millions of conjuncts that the conjunctive belief concatenates, it seems to ask too much for an ordinary thinker like Lola to be able to form this belief in the first place. So it seems beside the point that Lola has a reason to disbelieve the conjunction (and, by “reason implies can,” it is questionable that she has a reason to form the disbelief).

Third, the best reading of Schmidt’s claim is the following: The fact that Lola’s beliefs are inconsistent, as evidence that their conjunction is false, together with her evidence for the truth of each individual belief, is evidence that her evidence doesn’t settle the issues, and as such is a reason for her to suspend on all the individual beliefs together. For the incoherence fact together with the evidence for the individual beliefs to be such a reason, and for it to become motivating in Lola’s reasoning towards suspension, she need not have the conjunctive belief. It seems good enough that she is able to appreciate that inconsistency means that the conjunctive proposition is false, even though she has great evidence for each conjunct. This evidential uncertainty of her situation rightly moves her to suspension.

I grant that this kind of reason is available to Lola. Note, however, that Schmidt’s proposal does not fit all cases. If we are dealing with cases of mere incoherence, not inconsistency, like Marple and Poirot, the conjunction of the relevant beliefs is not necessarily false. It can be true that the evidence indicates that v even though it is not the case that v. So, Poirot does not have conclusive evidence that a conjunction of individually well-supported propositions is necessarily false. For this kind of case, Schmidt presents a different story: Poirot’s evidence regarding v and regarding whether the evidence indicates that v, as far as he knows it, is insufficient. And this is a reason for him to suspend (Schmidt, 2023, 8). The very fact that Poirot should take a step back and reconsider his epistemic situation is higher-order evidence that his evidence, as far as he is currently able to appreciate it, does not settle the issues in question. And this is a fact about his evidence that favors suspension.

Schmidt here claims that quite generally, only the evidence insofar as the subject knows it fixes the justificatory status of her doxastic attitudes. In other words, a piece of evidence is normatively inert in all respects whose justificatory relevance the subject does not currently appreciate. This means that Poirot’s evidence does not support disbelief that v overall, even though, as a matter of fact, not-v is what his evidence indicates. (This is because he first has to think a bit before he will realize that his first-hand evidence is correct and that Marple was, for once, mistaken about what the evidence indicates.) I find Schmidt’s claim doubtful. I have no problem with the standard evidentialist claim that the epistemic status of doxastic attitudes is fixed by what evidence the subject has. For instance, imagine that the newspaper reports that the stating time of the film I want to watch at the local movie theater has changed. As long as I am not apprised of this fact, it has no bearing on the justificatory status of my belief that the film starts at eight (Feldman, 2003, 47). Moreover, I agree that it is possible that I know a certain fact f, which is evidence that p, but still f has no bearing on the justificatory status of my doxastic attitude regarding p. This can happen when it is beyond my cognitive capacities to appreciate f’s evidential import concerning p. For instance, say that p is a quantum theoretical statement, and f some fact whose relevance for quantum theory is beyond my abilities to appreciate.

However, Schmidt’s claim limits our evidence even more than this. Take a subject who will be able to appreciate the justificatory relevance of a piece of evidence, given a bit of time. On Schmidt’s proposal, this evidence is justificatorily inert even so, just because the subject doesn’t currently grasp its import. This view is in tension, for instance, with Lord’s (2018, 100) plausible claim that possessing reasons goes together with having a “route” that the subject “can take to an ex post rational reaction.” Subjects like Poirot or Basna do have routes to relevant rational responses; it’s just that these routes are not instantly available to them but require some thought. Evidentialists are well-advised to adopt a view like Lord’s. Given such a view, there will be cases where the fact that one’s doxastic attitudes are incoherent is a reason to take a step back and invest effort in figuring out more fully what one’s evidence supports—that is, where this fact is a reason to suspend.

My response so far is intended to show that incoherence-based reasons are not superfluous. Let me add that I don’t think it would be problematic even if rational motivation or epistemic status were overdetermined by evidence-based and incoherence-based reasons. I claim in the paper that Basna, say, rationally responds by suspending on all of the involved beliefs together, in light of their incoherence (p. 14). There may certainly be cases where she realizes that her evidence doesn’t settle the issue and this motivates her to suspend. My point is that she doesn’t need to go there. She can simply become aware that her beliefs are inconsistent, and from there, rightly and directly, move to suspension. That is to say, there is more than one rational route to suspension, and this is unproblematic. What about Schmidt’s claim that incoherence-based reasons are normatively idle? Note that normative reasons don’t simply add up, but interact in complex and sometimes surprising ways (see Dancy, 2004). For instance, assume that S infers that q from p and if p then q. Both p and if p then q are reasons to believe that q. But now take the fact that modus ponens is a valid inference pattern. It seems correct that this fact is a reason to believe q in this scenario, and that S can believe partly for that reason; but this doesn’t appear to be a reason whose normative strength can in any sense be added to p and if p, then q. So facts—including incoherence facts—don’t necessarily have to add anything to a situation, normatively speaking, to be normative reasons.

4 Response to Conor McHugh

Conor McHugh first presents some potential ways of tracing the subjects’ reasons to suspend back to evidence, before making trouble for my claim that incoherence facts provide reasons for the attitude of suspension via inquiry. I briefly comment on the first problem before focusing in more detail on the second challenge.

McHugh (2023, 2) points out that, where a subject’s belief that p and her belief that q are inconsistent, her evidence for p will ipso facto be evidence against q, and so a reason against believing q (and vice versa for the subject’s evidence that q). For instance, the testimony from Basna’s history professor is evidence that it is not true that no facts are relative, and the testimony from her philosophy professor is evidence that it is not the case that the historical facts are relative. So apparently, we can explain why Basna should drop both beliefs in virtue of her evidence. However, note that the reasons to which McHugh calls our attention are reasons against believing by virtue of being reasons to disbelieve, not by virtue of being reasons to suspend. The history professor’s statement that the historical facts are relative is a reason to disbelieve that no facts are relative; it is not a reason to suspend on whether this is so. So this suggestion doesn’t capture the intuitive idea that, in a situation with inconsistent beliefs, Basna has a reason to suspend.⁷

Next, McHugh raises the worry that I don’t have a good account of why it can be right to suspend in cases where the subject doesn’t yet believe the incoherent propositions—say, where Basna is reading a discussion of relativism in the historical sciences, and is still in the process of forming beliefs on the issue. Here, I cannot appeal to the fact that Basna’s beliefs are incoherent as the reason to suspend, since she lacks the beliefs. But I also cannot appeal to the inconsistency merely between the propositions involved. Any conceivable proposition is inconsistent with many other propositions, but that does not make for reasons to suspend. This is true. However, the fact that the propositions which Basna is considering to endorse are inconsistent is available as a reason for her to remain suspended on the relevant issue as she is trying to figure out which belief is best supported by the evidence. Similarly, if Basna already believes that the historical facts are relative and then hears from her philosophy professor that no facts are relative (but hasn’t yet formed the belief that no facts are relative), the fact that a proposition that she is considering to endorse is inconsistent with a belief she already holds is a reason for her to suspend on both.⁸

McHugh’s more pressing comments concern the relation that I hypothesize between suspension, inquiry, and incoherence. He considers, first, whether and how incoherence-based reasons to inquire transmit to suspension. Second, he questions whether we need a sui generis mental state of suspension in genuine inquiry. Finally, he argues that there are cases where incoherence does not provide us with reasons to inquire.

Starting with the first problem, McHugh (2023, 4, fn. 5) rightly insists that it is a bad idea to conceive of suspension as a means to proper inquiry. The claim I should have made more clearly in the original paper is that suspension is a necessary condition for genuine inquiry, not that subjects suspend with the end of inquiring properly. Given this, can it convincingly be argued that reasons to inquire always come down to reasons to suspend—is it sufficient for this claim that suspension is required so that we can genuinely inquire? McHugh points out that many “deny that reasons generally transmit to necessary attitudinal means” (McHugh, 2023, 4). Moreover, where a reason transmits from end to means, we end up with pragmatic or wrong-kind reasons. I agree; this cannot be the right picture of incoherence-based reasons to suspend. For instance, take the case where I realize in the evening that it’s my good friend’s birthday. This is a reason to mail her a package from the post office. As a means to doing so in time, I need to inquire how long the post office is open. That it’s my friend’s birthday is a reason that transmits to relevant means actions, and so, plausibly, is a practical reason to inquire how long the post office is open. Is it then also a reason for me to suspend on the post office hours, in order to properly inquire? If so, it could only be a practical reason to suspend. But then how can we avoid saying the same thing about incoherence facts that are reasons to suspend because they are reasons to inquire?

In my view, suspension of judgment is a constitutive element of proper inquiry. If you want to seriously figure something out or answer a certain question, you have to do this in the right mindset, one of not being settled on the question. We might say that the only proper way of inquiring into whether p is doing so open-mindedly or in a state of suspension on whether p. This is how reasons for inquiring are also reasons to suspend: because the attitude of suspending is part and parcel of proper inquiry. Compare: The only way of appropriately signing a contract is doing so in a sound state of mind. But this doesn’t mean that being of sound mind is a means to properly signing a contract, it’s just the only way of—it’s partly constitutive of—signing a binding contract. By virtue of this constitutive relation, a reason to properly sign a contract is a reason to remain of sound mind and not, e.g., to get extremely drunk. As a further indication that the relation between suspending and properly inquiring is not one between means and end, note that it sounds off to say that by suspending, Basna genuinely inquires whether all facts are relative. Rather, in suspending, Basna genuinely inquires (Alvarez & Hyman, 1998, 234). If this is right, a practical reason to inquire is a practical reason to suspend, at least if we grant that practical reasons can transmit to attitudes constitutive of the relevant actions. But what of the fact that Basna’s beliefs are inconsistent? Is this then a practical reason to suspend, on my account? No, for having incoherent beliefs is not fundamentally a practical matter, but a matter of failing with respect to standards of theoretical rationality. The reason to suspend arises out of what it takes to be intellectually virtuous. This is what distinguishes my counterexamples from cases of practical reasons to inquire.

As to the second challenge, in dealing with my cases, can’t we do away with a special attitude of transitional suspension of judgment, and explain what’s going on fully in terms of the attitudes of belief and disbelief? McHugh’s (2023, 5) proposal is that the subject has a reason to bracket belief and disbelief—as we might say, to isolate them from our reasoning for the sake of inquiry. Let’s grant for the following discussion that bracketing is a real phenomenon, and set aside worries about multiplying kinds of suspension beyond need. I take it McHugh is inspired by Edmund Husserl’s notion of epoché (Husserl, 1950, 65/66). Bracketing a claim in this sense does not involve dropping it completely; instead, the subject maintains the claim in a sense, while inhibiting its inferential import. Isn’t this just what the subject does in reconsidering her epistemic situation, after realizing her attitudes are incoherent? Once she knows where the mistake lay, she ends up revising or dropping some of her attitudes, while de-bracketing others.

McHugh’s suggestion is certainly intriguing, and touches on the fundamental question: Under what conditions are we dealing with a (sui generis) doxastic attitude, and under what conditions are we merely looking at some broader constellation of mental states? It goes beyond what I can do in this response to discuss this as fully as it deserves, and I hope I will be able to say more on another occasion. Instead, I want to merely push the burden of proof back on McHugh by pointing out that it is at least not obvious that bracketing is not a sui generis doxastic attitude. Just terminologically, the notion is not very far away from that of suspension of judgment, and Husserl (1950, 66) describes it as a kind of suspension. The term “suspension of judgment” suggests that there is a judgment that the subject maintains, and on which she takes a kind of meta-stance by disabling its inferential effects, just as sketched for bracketing. Moreover, as I read Husserl, for him, bracketing is a genuine attitude. He calls it an “Akt” (Husserl, 1950, 64) and describes it as a special mode of consciousness, which picks up and modifies the subject’s bracketed belief. Correspondingly, I propose that bracketing is a doxastic attitude after all, which is, as it were, superimposed on the incoherent states thereby inhibited. Given that this position is available, there is no obstacle for me to maintain that the incoherence fact is a reason for the subject to inquire and so to suspend on (here: bracket) the relevant attitudes.

McHugh’s final challenge concerns cases in which, intuitively, subjects have no reason to inquire despite having incoherent doxastic attitudes. Think of any utterly dull and tedious matter—such as the exact length of all the blades of grass in the garden (McHugh, 2023, 6)—and imagine that our subject has incoherent beliefs on this matter. It seems clear that, nonetheless, she has no reason whatsoever to inquire into this issue. And we might be undecided about whether to ascribe a reason to suspend to her. Let me respond. From the perspective of intellectual virtue, in having an incoherent combination of beliefs, the subject doesn’t live up to the ideal. In light of this, the subject has a reason to avoid the combination. In the cases I considered in the original paper, the involved beliefs (or disbeliefs) were not utterly dull and tedious. In such contexts, intuitively, the right way of avoiding the incoherent combination is to re-open the issues settled by the beliefs in question, which involves suspending. I stand by this proposal. However, I can naturally allow that in contexts involving utterly dull and tedious incoherent beliefs, the proper way to avoid incoherence is to drop the matter altogether. More generally, one available response is to rely on a context-dependent account of when the intellectually virtuous thing to do is to re-open inquiry, and when it is instead to turn to more important issues.

Once again, I thank my commentators for their challenging questions. I greatly appreciate that they have pushed me to think more thoroughly about my proposal and its implications and have provided me with impulses for future work on these issues. I hope I have been able to address at least some of their concerns satisfactorily.