Suppose that I’m an avid baseball fan, and I spend most of the regular season hoping that my team will make the playoffs. As the season winds down, however, my team is mathematically eliminated from playoff contention. But suppose that, even after I confirm this news with every reliable source I can find, I still find myself clinging to the hope that my team will make the playoffs. In this case, it seems that there’s something problematic about my hope. What’s more, it seems that the problem with my hope has to do with my epistemic position. If I lacked good evidence that my team had been eliminated, for instance, it could be entirely appropriate, in every sense, for me to hold on to hope.
Now, suppose that the next season unfolds differently. This time, I spend the entire season worrying that my team will not make the playoffs. And, happily, there’s better news this year: my team clinches a playoff spot. But suppose that, even after I confirm this news with every reliable source I can find, I can’t shake the worry that my team will not make the playoffs. In this case, again, something seems to have gone wrong, and again the problem seems grounded in my epistemic position.Footnote 1
What, precisely, is the problem with my hope and my worry in these cases? Well, there might be several. Norms of prudence, for instance, might counsel against my feeling worry. Or perhaps my hope and worry suggest a morally vicious, overweening concern for my team. In this paper, however, I want to set aside appraisals of the prudence, virtuousness, usefulness, or moral value of hope and worry. I want to focus, instead, on the fittingness of hope and worry. Fittingness is a distinctive normative status, and one that is notoriously difficult to define or analyze.Footnote 2 We first get a grip on questions about fittingness by setting aside certain so-called “reasons of the wrong kind,” including incentives for being in a given mental state.Footnote 3 The fact that I’ll be rewarded for hoping, for instance, might make hope desirable, or useful, or even the emotion that I have most reason to feel. But it is not the kind of fact that could make a difference to whether my hope is fitting.
In this paper, I defend a novel proposal about the epistemic conditions that must be met for hope and worry to be fitting. On this proposal, which I call Suspension Threshold, it’s fitting to hope that p or worry that p only when it’s also epistemically rational to suspend judgment as to whether p. If Suspension Threshold is the correct view of the epistemic conditions on fitting hope and worry, any possibilities that are too insignificant to make suspension of judgment rational are also too insignificant to make hope or worry fitting.
I make a case for Suspension Threshold in two stages. First, in section 1, I show that it has several important theoretical virtues; it is ecumenical, illuminating, elegant, and extensionally adequate. Second, in section 2, I show that the view can do important work: specifically, it can help to deflate a recent argument, due to Matthew Benton (Reference Benton2018), against fallibilism about knowledge. Cumulatively, these considerations show that theorists of all different stripes—and especially fallibilists about knowledge—have excellent reason to embrace Suspension Threshold.
1. The virtues of Suspension Threshold
The question at the core of this paper is a question about what sort of epistemic support is required in order for hope or worry to be fitting. But hope and worry are not the only attitudes for which this sort of question arises. Suppose, for instance, that I resent you for taking the last cupcake, but I don’t have good evidence that it was you who took it. Or suppose that I feel delighted about having won a prestigious award despite the fact that I have strong evidence that I haven’t in fact won it. In these cases, my resentment and delight are unfitting and, what’s more, they are unfitting precisely in virtue of the weakness of my epistemic position.Footnote 4 Cases like these provide some prima facie grounds for thinking that certain epistemic conditions must be met in order for resentment or delight to be fitting.Footnote 5
The question of how much epistemic support an attitude must enjoy in order to be fitting, then, arises for a great many attitudes. But we should not expect the precise degree of epistemic support necessary to be the same for all attitudes. Many attitudes require a fair bit of epistemic support in order to be fitting. If I resent you on the basis of evidence that makes it only 50 percent likely that you’ve wronged me, for instance, I’ve jumped the gun. Similar problems seem to arise in many cases where a person feels shame, jealousy, contempt, or pride on the basis of weak evidence. But, strikingly, some attitudes can be fitting even in the face of very weak epistemic support. Perhaps the clearest examples are hope and worry. I can fittingly hope that my aunt’s cancer treatment will succeed even when the likelihood of success, on my evidence, is much lower than 50 percent. A soldier could fittingly worry that she will be seriously injured even if the likelihood of serious injury, on her evidence, is much lower than 50 percent. Hope and worry, then, have a distinctive feature: unlike many other attitudes, they can be fitting even when directed at very slim epistemic possibilities.Footnote 6 Whatever we say about the epistemic conditions on fittingness for other attitudes, the epistemic conditions on fitting hope and fitting worry deserve attention in their own right.
In this paper, I defend the following view of the epistemic conditions on fitting hope and fitting worry:
Suspension Threshold: The epistemic conditions on fitting hope that p or fitting worry that p are satisfied if and only if it is epistemically rational for S to suspend judgment as to whether p.
Before enumerating the virtues of Suspension Threshold, I’ll pause to offer a few clarifications.
First, Suspension Threshold offers an account of the epistemic conditions on fitting hope and worry. There are surely also other conditions that must be met for hope or worry to be fitting. For instance, it is unfitting to hope for an outcome that is neither desired nor desirable, even if it’s rational to suspend judgment about whether that outcome will obtain. Suppose, for instance, that there would be nothing desirable, by your lights or anyone else’s, about your stubbing your toe. It would, then, be unfitting for you to hope that you will stub your toe. But this is not because the epistemic conditions on fitting hope are not met.
Second, it’s important to see that Suspension Threshold posits a normative connection, not a descriptive connection, between hope, worry, and suspension of judgment. Suspension Threshold should not be confused with the descriptive claim that a person does in fact hope or worry that p only if that person suspends judgment as to whether p. That descriptive claim is a view about the conditions under which hope and worry arise, not a claim about the conditions under which it would be fitting for a person to hope or worry. And some defenders of Suspension Threshold might want to reject this tight psychological connection between hope, worry, and suspension of judgment. They might want, for instance, to make room for the psychological possibility of irrationally worrying that one’s plane will crash even when one does not, in any sense, suspend judgment about whether one’s plane will crash. So, although Suspension Threshold can be endorsed by theorists who hold that hope and worry necessarily come along with suspension of judgment, it does not presuppose that view.
Finally, Suspension Threshold mentions suspension of judgment. But what is suspension of judgment? At least some answers to this question seem like a bad fit with the view that I aim to develop. McGrath (Reference McGrath2021, 467), for instance, understands suspending judgment as a way of putting off judgment about some subject/matter until a later time. This seems like the wrong state to look to when we aim to understand the fittingness conditions for hope and worry. Hoping that p and worrying that p seem importantly related to regarding p as an open possibility. And it might, in some cases, be entirely rational to ignore the question of whether p even though my evidence about p is totally conclusive, and therefore makes it irrational to consider p an open possibility.
In what follows, I’ll use the term ‘suspension of judgment’ in a more traditional way: to pick out a state that involves not just setting aside some proposition, but instead taking up an intermediate degree of confidence in that proposition short of both belief and disbelief. This is a fair dialectical move because, as McGrath (464) is keen to point out, there are multiple ways of being neutral toward a proposition. Even if one thinks that ‘suspension of judgment’ is not the right label for the state I have in mind, one should certainly acknowledge that state as one important variety of doxastic neutrality. McGrath, tellingly, does so, calling the state I have in mind ‘agnosticism.’ Readers who prefer McGrath’s terminology, then, should feel free to interpret my claims about the connection between hope, worry, and suspension of judgment as claims about the connection between hope, worry, and agnosticism.
With these preliminaries out of the way, I’ll now make a case in favor of Suspension Threshold. I’ll do so by showcasing four of its most notable virtues: Suspension Threshold is ecumenical, illuminating, elegant, and extensionally adequate.
1.a Ecumenical
Suspension Threshold is an ecumenical view; it retains its plausibility against a wide variety of different background theories about epistemic norms. Perhaps the easiest way to see this virtue of Suspension Threshold is to draw a contrast with another prominent account of the epistemic conditions on fitting hope and fitting worry:
Possibility Threshold: The epistemic conditions on fitting hope that p or worry that p are satisfied if and only if it is both epistemically possible for S that p and epistemically possible for S that –p. Footnote 7
Possibility Threshold says, loosely speaking, that any sliver of possibility that p, no matter how slim, is significant enough to make hope or worry fitting.Footnote 8 Possibility Threshold is not a competitor to Suspension Threshold in the sense that they contradict one another; in fact, given certain assumptions, the two offer coextensive verdicts about fitting hope and fitting worry. But, as we’ll soon see, there are some prominent approaches to epistemology on which Possibility Threshold and Suspension Threshold come apart.
One possible challenge for defenders of Possibility Threshold arises in cases where the possibility of some outcome seems too slim to make hope or worry fitting. The introduction offered two cases of this sort: if every reliable news source I can find clearly confirms that my team has made (or has failed to make) the playoffs and I have no reason to suspect any widespread error in reporting, it can seem unfitting to continue worrying (or hoping).
Now, some defenders of Possibility Threshold have a way to deflate this challenge: they can claim that once I’m faced with overwhelming evidence that my team has been eliminated, it is no longer epistemically possible for me, to any degree, that my team will make the playoffs. But, crucially, this move is available only against the backdrop of certain views about epistemic possibility. While some theories are compatible with the notion that an everyday body of evidence can make it downright epistemically impossible that my team has made the playoffs, many others entail that epistemic impossibilities are rare indeed. According to many fallibilists about knowledge, for instance, most items of everyday knowledge (including knowledge that a baseball team has been eliminated from playoff contention) come accompanied by a genuine possibility of error.Footnote 9 Many fallibilists, then, will be unable to use this straightforward denial of epistemic possibility to deflect the challenge I’ve just raised for Possibility Threshold. The only way for them to hold on to Possibility Threshold will be to make peace with the radical view that it is fitting for me to hold on to hope or worry about my team’s fate in the playoffs in the face of even overwhelming evidence.
Possibility Threshold, then, is difficult to square with some prominent, attractive approaches to epistemology. This does not mean that Possibility Threshold is false. But it does mean that Possibility Threshold is not ecumenical. It depends for its success on some highly controversial claims about the extension of epistemic possibility.
Suspension Threshold, by contrast, is nicely ecumenical. It allows epistemologists of many stripes to explain why, in the face of overwhelming evidence, it’s often unfitting to hold on to hope or worry. The explanation is a simple one: overwhelming evidence often makes it epistemically irrational to suspend judgment. This point could be made within a fallibilist framework; the fallibilist will simply have to say that, in at least some cases, a genuine possibility of error can be too slim to make suspension of judgment rational.Footnote 10 But Suspension Threshold is also friendly to infallibilist views on which everyday knowledge is not accompanied by the possibility of error. On these views, as we’ve already seen, overwhelming evidence that my team has been eliminated from playoff contention could make it epistemically impossible for me that they have made the playoffs. But it’s very plausible that, if an overwhelming body of evidence does away with any possibility of error, that body of evidence also renders it irrational for me to suspend judgment as to whether my team will make the playoffs. Given Suspension Threshold, that means that hoping my team will make the playoffs is also unfitting. So, theorists of many different stripes can use Suspension Threshold to explain why, in the face of overwhelming evidence, holding on to hopes and worries can be unfitting.Footnote 11
It’s important to note that defenders of Suspension Threshold might, but need not, take the epistemic threshold for rational suspension of judgment to be fixed at a particular degree of probabilistic support.Footnote 12 This is another respect in which Suspension Threshold is ecumenical. If the threshold for rational suspension of judgment is a flexible, sensitive one—if, for instance, it is responsive to factors having to do with salience, with practical stakes, or with the nature of one’s evidence—then Suspension Threshold suggests that the fittingness of hope and worry will be sensitive to those factors as well, and in just the same range of cases.
Some permissivists claim that it is rational to suspend judgment about any (or almost any) proposition because rationality for doxastic states is a permissive matter. Some skeptics might even claim that it is rationally mandatory to suspend judgment about all (or almost all) propositions. Suspension Threshold has a surprising result when conjoined with these views: it suggests that hope and worry can never (or almost never) be unfitting in virtue of a person’s epistemic position. But this is no problem for Suspension Threshold; to the contrary, it’s a result that fits nicely with the spirits of these views. A permissivist who thinks that I’m permitted to suspend judgment as to whether I’ll wake up tomorrow magically transformed into a perfectly happy unicorn, for instance, should not say that my epistemic position rules out the possibility of fittingly hoping that I’ll be so transformed. And a skeptic who takes rationality to require global suspension of judgment could very reasonably call it fitting for us to hope for any given possible good and to worry about any given possible evil. (Of course, our contingent constraints might keep us from meeting these rational requirements, and there might be good prudential reasons to avoid meeting them. But those are both familiar challenges for the skeptic about rational belief.)
Suspension Threshold is nicely ecumenical; it retains its plausibility against the background of many different views, including fallibilist, infallibilist, permissive, or skeptical ones. Suspension Threshold has a particular advantage over Possibility Threshold when it comes to accommodating the perspective of non-skeptical fallibilists. This is an advantage that matters; non-skeptical fallibilism is widely considered, by both friends and foes, to be the orthodox view within contemporary epistemology.Footnote 13 Ceteris paribus, we should prefer a view about hope and worry that pairs well with this orthodoxy.
1.b Illuminating
Suspension Threshold illuminates the nature of the connection between hoping, worrying, and keeping an open mind. This is a connection that needs illuminating. One of the most obvious ways to draw the connection—the claim that hope and worry constitutively involve suspension of judgment—faces prima facie challenges. As I’ve already noted, there seem to be cases in which a person worries (albeit perhaps irrationally) about some outcome that she staunchly believes will not obtain. But, even if the two states can in principle come apart, there does seem to be an important connection of some sort between worrying and keeping one’s mind open about whether something bad will happen. And, in the same way, there seems to be an important connection of some sort between hoping and keeping one’s mind open about whether something good will happen. How should we understand these connections?
It’s helpful to confront this question armed with an appreciation of a similar set of questions that arise in the literature on recalcitrant emotion. Imagine that you wake up feeling resentment toward your significant other after having dreamed that they wronged you. You might continue to feel resentment despite the fact that you firmly and wholeheartedly believe that your significant other has not, in fact, done you wrong. So, resentment seems psychologically possible even in the absence of belief about wrongdoing. But even if we grant this point, we should acknowledge that there seems to be a connection of some sort between resenting a person and believing that they’ve done wrong. How should we understand that connection?
On a currently popular proposal, the connection between resentment and beliefs about wrongdoing can be understood in terms of the conditions that make both fitting: it’s fitting to resent a person in all and only the conditions that make it fitting to believe that they’ve done wrong.Footnote 14 This proposal illuminates the connection between resentment and beliefs about wrongdoing, in other words, by pointing out that they share the same conditions for a certain kind of success.
Suspension Threshold illuminates the connection between hoping, worrying, and keeping one’s mind open in just the same way: it claims that the scenarios that make it fitting to hope and to worry are, necessarily, also scenarios that make it epistemically appropriate to suspend judgment. Again, the merits of Suspension Threshold become even clearer when we draw a contrast with Possibility Threshold; Possibility Threshold leaves obscure the connection between hoping, worrying, and keeping an open mind.Footnote 15 This will be particularly clear to theorists who hold that some epistemic possibilities are too insignificant to make suspension of judgment rational. On that view, there are situations in which a fully rational agent with fully fitting attitudes will hope (or worry) that p without suspending judgment regarding p. On this view, there is neither a robust descriptive nor a robust normative connection between hoping and keeping one’s mind open.Footnote 16
Now, there are views on which any sliver of possibility that p can make suspension of judgment rational; on these views, Possibility Threshold and Suspension Threshold are coextensive. They pick out the same fittingness conditions for hope and worry under different descriptions—one having to do with epistemic possibility, and one having to do with rational suspension of judgment. But even if a view of this sort is right, there are reasons to consider the latter gloss a more illuminating one. The term ‘epistemically possible’ is a philosopher’s term of art, and different theorists use it to pick out radically different properties. Some apply the term to all and only those scenarios that are not ruled out a priori by one’s evidence; others apply it to all and only those scenarios that are compatible with what one knows.Footnote 17 It’s easy to suspect, in disputes like this, that the disputants are merely talking past one another; perhaps there are radically different, but equally eligible, ways to understand the term ‘epistemic possibility.’ So even if Possibility Threshold does pick out the fittingness conditions for hope and worry, it picks them out with a slogan that obscures, rather than illuminating, the connection between hope, worry, and keeping an open mind.
Suspension Threshold does a better job. No matter one’s background theory, one can easily see the relevance of Suspension Threshold to questions about when to keep an open mind. In just the same way that we shed light on emotions like resentment when we specify the situations that make certain beliefs fitting, we shed light on hope and worry when we use Suspension Threshold to specify the epistemic situations that make suspension of judgment fitting.
1.c Elegant
If Suspension Threshold is false, then the epistemic thresholds for doxastic states and the epistemic thresholds for hope and worry are not coordinated. This gives rise to a kind of proliferation in epistemic thresholds for attitudes. When we’re asking whether it’s fitting to get into the game of hoping, we have to ask whether our epistemic position is strong enough to clear one threshold; when we’re asking whether it’s rational to get into the game of suspending judgment, we have to ask whether our epistemic position is strong enough to clear a different one.
There are reasons to avoid this proliferation in epistemic thresholds, if we can. Some of those reasons are most easily seen from the theorist’s perspective. Theorists should, ceteris paribus, prefer a principled, unified story about why thresholds for different attitudes are located where they are, and the more thresholds we fix in different locations, the less likely we are to find such a unified story. Some other reasons for avoiding proliferation are more easily seen from the perspective of the agent governed by these thresholds: if there are too many distinct epistemic thresholds for different attitudes, then a rational agent will have to be sensitive to each of a wide array of different benchmarks when forming and abandoning attitudes.Footnote 18
Suspension Threshold promises a more elegant, streamlined vision of epistemic thresholds. This is a boon to the theorist, who can offer a more unified, straightforward story about the epistemic conditions that license us in forming a wider range of attitudes. And it makes possible an attractively unified story about what deliberation and attitude formation could be like from the agent’s perspective: once we determine that our situation that makes it epistemically rational to suspend judgment as to whether p, we needn’t ask any further questions in order to determine whether p is likely enough to merit worry or hope.
Now, it’s important not to overemphasize the desirability of elegance in a theory. We should not pursue greater simplicity in our story about epistemic thresholds if that greater simplicity is purchased at the price of accuracy. So, no matter how elegant and streamlined Suspension Threshold is, we should not accept it if it forces us into implausible verdicts about the conditions in which hope and worry are fitting. In the next subsection, I’ll argue that Suspension Threshold does not have implausible verdicts of this sort. We can embrace its simplicity without compromising on accuracy.
1.d Extensionally adequate
I’ll conclude our look at the virtues of Suspension Threshold by arguing that it is extensionally adequate: that is, by arguing that it has plausible results about whether, in any given case, hope and worry would be fitting. I’ll consider three case types that seem most likely to be considered counterexamples to Suspension Threshold. I’ll then argue that Suspension Threshold gets the right results even in these cases.
First, note that there are a great many possible misfortunes about which it’s uniquely epistemically rational for me to suspend judgment. At some point in the next several years, my identity might be stolen, I might develop a debilitating illness, a natural disaster might seriously hurt someone that I love—the list goes on. I am not in a position to rationally form beliefs about whether these events will come to pass; if I take up any doxastic attitude as to whether they will, it had better be suspension of judgment. Suspension Threshold says, then, that I meet the epistemic condition on fitting worry toward these bad outcomes. Some will balk at this result; there’s something disturbing, and perhaps unfitting, about a life spent consumed by constant worry about each of an enormous array of possible bad outcomes.
These cases should not tempt us to reject Suspension Threshold. For one thing, they do not have the right form to straightforwardly disprove Suspension Threshold; Suspension Threshold does not claim that it’s fitting to worry about any bad outcome toward which it’s rational to suspend judgment. Suspension Threshold simply says that the epistemic conditions on fitting worry are met when it’s rational to suspend judgment. This leaves open the possibility that, when a significantly likely outcome does not make worry fitting, it’s not because that outcome is insufficiently epistemically likely, but instead because it falls short of another condition on fitting worry. (Perhaps, for instance, the badness of the relevant outcome is not severe enough for it to merit worry. Or perhaps an outcome must be related to one’s current decisions or projects in a certain way in order to merit worry.Footnote 19)
But even if we set this formal point aside, we should not be too concerned about the possibility that Suspension Threshold will force us into lives saturated with worry toward every possible misfortune. It’s true that Suspension Threshold says that there are a great many propositions that merit worry. But there are also a great many objects that merit other responses: aesthetic responses, emotional responses, intellectual responses, and more.Footnote 20 Given our limited cognitive resources, we will invariably miss out on some of these fitting attitudes—and the question of which ones we take up is partly determined by our interests and our character. So, realistic agents guided by Suspension Threshold should not expect to spend their lives mired in worry any more than realistic agents guided by standards of fitting aesthetic evaluation should expect to spend their lives in a never-ending state of aesthetically evaluation of every item they encounter.Footnote 21 The problem with a person whose life is full of constant worry (or constant aesthetic evaluation), in other words, is more plausibly understood as a problem with their character-level dispositions than as a problem with the fittingness of their individual attitudes. Suspension Threshold, then, is the wrong place to look when we seek to understand the problem in such a life.
Let’s move on to a second set of putative counterexamples to Suspension Threshold—ones in which it’s very difficult to see that one has conclusive epistemic reason to believe some proposition. Suppose, for example, that you’ve just taken a logic exam. The exam’s final question was a doozy—it required a great many complicated inferences, and you suspend judgment as to whether you got the right answer. Some will say that, assuming that there really is a uniquely correct answer to the question, epistemic rationality does not permit your suspension of judgment; after all, your body of evidence (just like every body of evidence) is only consistent with one answer to the question. But surely it’s fitting for you to hope that you got the question right.Footnote 22
This objection relies on an extremely austere view of epistemic rationality—one that does not permit suspension of judgment toward any logical truths. Any such view must account for the fact that there is some sense in which it can be appropriate for a limited agent to suspend judgment toward a truth of logic. One might distinguish, for instance, between what it’s epistemically rational to believe and what it’s reasonable to believe, or between ideal standards of epistemic rationality and non-ideal standards of epistemic rationality.Footnote 23 But once we draw this distinction, we have excellent reason to draw a similar one for standards of fittingness as well. When I hope that I got the tricky logic question right, we can assess my hope against multiple different standards. We can ask whether it is ideally fitting, in the sense that it perfectly matches every last fact no matter how obscure about my evidential position. Or we can ask a different question: we can ask whether my hope meets non-ideal standards for fitting attitudes. An attitude can be fitting in this latter sense as long as it matches, or fits, a reasonable approach to one’s evidence.
I do not commit myself to the view that standards of epistemic rationality are extremely austere. But if they are, no trouble arises for Suspension Threshold: we should simply distinguish between two ways of interpreting the claim. We might interpret it, first, as a claim about ideal rationality and ideal fittingness; read this way, Suspension Threshold says that the epistemic conditions on ideally fitting hope are met as long as one is ideally rational to suspend judgment. Or we might interpret Suspension Threshold as a claim about non-ideal standards; read this way, Suspension Threshold says that the epistemic conditions on non-ideally fitting hope are met as long as one is non-ideally rational to suspend judgment. Both of these claims are plausible, and neither is troubled by the logic-exam case.
I’ll close this section by considering cases in which very slim possibilities are associated with very high stakes. Consider, for instance, the following case from Jessica Brown:
… Liz knows that she was born in England. [She is offered] a bet with the following pay-offs:
Liz was born in England: Liz gains £1
Liz was not born in England: Liz loses her home (Reference Brown2008, 176)
Suppose that Liz (unwisely) accepts the bet. What should she think and what should she feel while she waits to learn whether she will lose her home? Some will be tempted to defend the following pair of verdicts. First, it is not epistemically rational for Liz to suspend judgment as to whether she was born in England, nor as to whether she’ll lose her house. And second, it is fitting for Liz to feel worry that she’ll lose her home. If these verdicts are right, then we have a counterexample to Suspension Threshold; according to Suspension Threshold, it can’t be fitting for Liz to worry that she will lose her home unless it’s also epistemically rational for her to suspend judgment as to whether she will.Footnote 24
My preferred approach to this case is to deny that it is epistemically irrational for Liz to suspend judgment. I think that it’s highly plausible that epistemic rationality at least permits Liz to keep her minds open about her birthplace.Footnote 25 Put yourself in Liz’s position; wouldn’t you start treating certain error possibilities—like the possibility that your parents lied to you about your birthplace—very seriously? And couldn’t you be entirely rational to do so? Now, some theorists will not share my preferred approach; they will say that it would be positively epistemically irrational for Liz to suspend judgment about her birthplace, even when offered with the bet. But even if we grant that point, it’s very hard to imagine a case like this in which it would clearly be epistemically irrational for Liz to suspend judgment as to whether she’ll lose her home. After all, she is interacting with a person who appears both able and willing to play a game that might result in her losing her source of shelter. Perhaps our judgments about fitting worry in this case are informed by a difficult-to-shake suspicion that such people are not trustworthy.
Perhaps this difficulty with imagining the case can be overcome. Some will certainly insist that they can imagine a version of Liz’s case in such a way that it is irrational to suspend judgment about the fate of her home. But once we commit ourselves to this verdict, there is no longer any obvious appeal to calling worry fitting. If suspension of judgment is not a rational option for Liz—if rationality, loosely speaking, forbids her to keep an open mind as to the fate of her house—then why would it be fitting for her to be emotionally open to the possibility that she’ll lose her house? Here’s one possible answer to that question: it might, generally speaking, be psychologically common, or prudent, or useful to be sensitive to error possibilities that are associated with high stakes. But this answer is unsatisfying; the commonality and the usefulness of a mental state can certainly make a difference to whether it’s good to be in that mental state, but we paradigmatically set these considerations aside when asking whether that mental state is fitting.
In short, it’s far from clear that high-stakes cases like Liz’s make suspension of judgment irrational. But if we insist that they do, the appeal of calling worry fitting is significantly diminished. There is no plausible approach, then, on which these high-stakes cases provide a counterexample to Suspension Threshold.
The challenge I’ve just considered, involving worry directed at slim possibilities, has a cousin: a challenge involving hopes invested in very slim possibilities. Such hopes are common: in the face of terminal illness, stubbornly persistent injustice, or looming environmental disaster, we frequently cling to the hope that everything will turn out all right—the illness will be cured, the injustice will be resolved, the disaster averted. Aren’t there some cases of this sort in which, even though it’s epistemically irrational to suspend judgment about what will happen, it’s nevertheless fitting to hold out hope for the best outcome?
Suspension Threshold says that there are not. This verdict may seem surprising, but it is entirely defensible for just the reasons I’ve mentioned above. For one, the mere fact that the possibility that p is very slim doesn’t entail that suspension of judgment is epistemically irrational. Defenders of Suspension Threshold needn’t say, then, that hopes invested in slim possibilities are always irrational. And if we do find a case in which it’s entirely clear that suspension of judgment would be epistemically irrational, there is no longer any obvious pressure to calling hope fitting. It may well be that there are some cases in which, even though suspension of judgment is irrational, it’s nevertheless virtuous (or healthy, or strategically rational, or admirable) to continue “hoping against hope.” But an attitude’s being virtuous (or healthy, or strategically rational, or admirable) does not make that attitude fitting.
We’ve now considered four virtues of Suspension Threshold: it is ecumenical, illuminating, elegant, and extensionally adequate. Another virtue of Suspension Threshold lies in its applications: it can shed light on debates within first-order epistemology. In the next section, I’ll illustrate this point by example: I’ll show that Suspension Threshold enables us to resist a recent argument for infallibilism offered by Matthew Benton (Reference Benton2018).
2. Hope and knowledge
Benton’s case against fallibilism starts from the observation that certain ascriptions of knowledge and hope seem problematic. Take, for instance, the following claim:
(3) # I hope that John is in his office, but I know that he is not. (Reference Benton2018, 3)
The problem with this sentence, Benton argues, is not merely a pragmatic infelicity, of the sort involved in Moore-paradoxical sentences like “I believe that John is in his office, but he is not.” Sentence (3), unlike a Moore-paradoxical sentence, remains problematic even when embedded within the antecedent of a conditional or under supposition:
(6) # If I hope that John is in his office but I know that he is not, then …
(7) # Suppose I hope that John is in his office but I know that he is not … (Reference Benton2018, 3)
By embedding (3) in these ways, Benton argues, we can screen off the possibility that the problem with (3) is an infelicity arises from the pragmatic effect of asserting it. After all, in (6) and (7), the content of (3) is not asserted, so no such pragmatic effects arise. Nevertheless, (6) and (7) are problematic. There are, Benton suggests, only two remaining viable ways to understand the tension involved in sentences like (3), (6), and (7): they either display a “semantic inconsistency between two conjuncts that could not both be true,” or the violation of “a requirement of rationality such that one may not rationally hope that p when one knows whether p” (2; emphasis mine). More briefly, Benton takes his linguistic data to show that “hope ascription is semantically or rationally (rather than pragmatically) incompatible with knowledge ascription” (4).
Benton argues, from this premise about the incompatibility between knowledge- and hope-ascription, that fallibilism is false. To do so, he appeals to one further premise: Possibility Threshold (5).Footnote 26 Why, according to Benton, do his two premises jointly make trouble for fallibilism about knowledge? Well, fallibilists hold that one can know that p even when there is a possibility for one that –p. Fallibilists who accept Possibility Threshold must further hold that when one’s knowledge that p comes with a possibility of error, it can be fitting to hope that –p. But if that’s right, then on the face of it, fallibilists must say that knowledge-ascription and hope-ascription are not rationally incompatible; to the contrary, knowledge and hope will sometimes coexist in the mind of an agent whose attitudes are all rational (and/or fitting). Fallibilists must also say that knowledge-ascription and hope-ascription are not semantically incompatible; we can sometimes truly say of an agent that she knows that p while hoping that –p. Those who accept both Possibility Threshold and Benton’s knowledge-hope incompatibility thesis, then, seem forced to reject fallibilism.
Some fallibilists will resist this argument by denying Benton’s first premise: that there is a semantic or rational incompatibility between knowledge- and hope-ascription. Any fallibilist who adopts this strategy must provide a compelling account of the apparent infelicity of claims like (3), (6), and (7). This task is more complex than it may seem at first; as Benton argues at length, some of the leading fallibilist strategies for addressing concessive knowledge attributions (like “I know that p, but p might be false”) cannot be smoothly applied to joint ascriptions of knowledge and hope (Reference Benton2018, 9). Fortunately for fallibilists, however, there is a way to deflate Benton’s argument without taking up this burden. Even if fallibilists grant Benton’s first premise for the sake of argument, they can comfortably reject Benton’s conclusion by denying Possibility Threshold and, in its place, endorsing Suspension Threshold.
To endorse Suspension Threshold while rejecting Possibility Threshold is to acknowledge that there are some genuine epistemic possibilities in which it’s unfitting to invest hope. And if that’s right, Benton’s challenge does not get off the ground: we can no longer infer, from the fact that my knowledge of p is fallible—that is, accompanied by a possibility of error—that the epistemic conditions on hope for –p are met. Consider an example that a fallibilist is likely to consider a case of fallible knowledge: my knowledge that Bigfoot has not discovered the cure for cancer. Fallibilists will say that there is a genuine possibility for me that this claim is false. Fallibilists who also accept Possibility Threshold would therefore have to claim that the epistemic conditions on fitting hope are met; I have a sufficiently strong epistemic position to fittingly hope that Bigfoot has discovered the cure for cancer. But fallibilists who reject Possibility Threshold and accept Suspension Threshold are not forced into that conclusion. As long as it would be epistemically irrational for me to suspend judgment as to whether Bigfoot has discovered the cure for cancer, hope for that outcome would be unfitting.
Benton suggests that there might be a distinct problem for fallibilists who reject Possibility Threshold: without Possibility Threshold, they will be hard-pressed to explain “why knowledge, though compatible with an epistemic chance of being wrong, might nevertheless be incompatible with hope” (Reference Benton2018, 5). Happily, Suspension Threshold puts fallibilists in a position to meet this explanatory challenge. To do so, they need one further commitment: that when one knows that p, one’s epistemic position does not make it rational to suspend judgment as to whether p. Footnote 27 If that’s right, then whenever I fallibly know that p, any chance that –p must be too insignificant to support rational suspension of judgment, and (by Suspension Threshold) also too insignificant to support fitting hope. So, fallibilists can appeal to Suspension Threshold to explain the clash between knowledge and hope, even in cases where knowledge comes along with a possibility of error.
Fallibilists can escape the force of Benton’s argument, then, by rejecting Possibility Threshold. But can they do so in a principled way? If Possibility Threshold were sufficiently well-motivated, this might be a difficult bullet for fallibilists to bite. So, what are the grounds for accepting Possibility Threshold? Well, Benton does not argue at length for Possibility Threshold; in fact, when he introduces it, he simply notes that it is “highly plausible” and “intuitive” (Reference Benton2018, 5). But this alone cannot entitle Possibility Threshold to be treated as a fixed point in our theorizing about hope. After all, it’s also highly plausible, and intuitive, that an overwhelming body of ordinary evidence can make hope unfitting (recall the example of the baseball team that’s eliminated from playoff contention). And it’s highly plausible, and intuitive, that even an overwhelming body of ordinary evidence cannot totally eliminate the possibility of error. Everyone should agree that one of these highly plausible and intuitive claims must be rejected.
Is there anything to be said for Possibility Threshold beyond its initial plausibility? Benton offers a more pointed line of support in a footnote, claiming that “lottery examples … suggest that there is no minimum confidence level required for hoping that p” (Reference Benton2018, 5). It does seem plausible that, in standard lottery cases of any size, it can be fitting to hope that one’s ticket will win.Footnote 28 But it’s too quick to conclude on this basis that Possibility Threshold is true. Instead of explaining the rationality of hope in lottery cases by appealing to the epistemic possibility of victory, a defender of Suspension Threshold can explain the rationality of hope in lottery cases by appealing to the rationality of suspending judgment in lottery cases.
This is a principled maneuver for the fallibilist; there are a variety of prominent attempts to explain why suspension of judgment is rational in lottery cases without simply appealing to the existence of an epistemic possibility that one’s ticket will win. On some theories, the distinctive feature of lottery cases is that they provide purely statistical evidence, which does not suffice to support rational outright belief.Footnote 29 On other approaches, lottery cases support suspension of judgment because they make the possibility of error salient, or because they make it impossible to have safe or sensitive outright belief that one’s ticket will lose.Footnote 30 If any of these theories is on the right track, then fallibilists can safely say that lottery cases make suspension of judgment rational without endorsing the bolder claim that any epistemic possibility that p makes suspension of judgment rational. And, given Suspension Threshold, the same can be said for hope: the fallibilist can say that lottery cases make hope fitting without endorsing the bolder claim (which is just Possibility Threshold) that any epistemic possibility that p makes hope that p fitting.
To sum up: by appealing to Suspension Threshold and rejecting Possibility Threshold, fallibilists can account for the tension between knowledge and hope. What’s more, there do not appear to be any significant, non-outweighed reasons that favor accepting Possibility Threshold while rejecting Suspension Threshold. So, once the independently well-motivated Suspension Threshold is in view, Benton’s argument for infallibilism loses its force.
Conclusion
In this paper, I defended a novel view of the epistemic conditions on rational hope and worry: in order to fittingly hope or worry that p, a person must be in a position to rationally suspend judgment as to whether p. This view has significant appeal across a wide range of different approaches to epistemology—and, as our discussion of Benton’s argument shows, fallibilists about knowledge have special reason to embrace it.
Kant’s question “What may I hope?” is an important and enduring one (Reference Kant, Guyer and Wood1997, 805). Traditional responses to this question have not drawn connections to a second important question: When is suspension of judgment rational? With this paper, I (fittingly) hope to encourage readers to see those two questions as intimately connected.
Acknowledgments
For discussions that helped to improve this paper, I’m grateful to Alexandra Bryson, Andrew Chignell, Sarah Colquhoun, Alex Gregory, Christopher Howard, Eugene Mills, Andrew Moon, Miles Tucker, Donald Smith, and Catherine Sutton. Special thanks to Ethan Brauer, Jesse Loi, Tristram McPherson, Mason Westfall, and Elizabeth Jackson Withorn for detailed written comments.
James Fritz is term assistant professor of philosophy at Virginia Commonwealth University. His research concerns ethics, epistemology, and emotion.
