2. Belief Accounts

Different doxastic-attitude accounts of suspending judgment about whether p agree on the following key claim: that suspending judgment about whether p is at least partly constituted by a doxastic attitude that represents, expresses, or constitutes one's neutrality or indecision about whether p. I want to begin by examining how exactly different doxastic-attitude accounts cash out this claim, with a focus in this section on belief accounts, in particular.

On belief accounts, it is partly constitutive of an agent's suspending judgment about whether p that that agent has some belief about their own epistemic position vis-à-vis whether p. (It is only partly constitutive, because belief accounts also require that the agent in fact neither believe that p nor disbelieve that p.) Different belief accounts differ primarily with respect to what the relevant belief about one's epistemic position is. On several belief accounts, the relevant belief about one's epistemic position is a normative one; the agent who suspends judgment believes that some aspect of her present epistemic circumstances – her evidential situation, for example, or her present cognitive limitations – does not warrant or permit believing or disbelieving that p. On a recent belief account defended by Raleigh (Reference Raleigh2021), for example, suspending judgment about whether p constitutively involves the belief that one cannot yet tell whether p, based on one's evidence. The agent who suspends judgment must also in fact neither believe that p nor disbelieve that p, and her neither believing nor disbelieving that p is in some way relevantly caused by her belief that she cannot yet tell whether p, based on the evidence.Footnote ⁵

Raleigh takes his view to be relevantly similar to views on which suspending judgment consists in believing that one does not know whether p. A view like this is hinted at in Russell (Reference Russell and Slater1997), on which an agnostic about p (in Russell's discussion, an agent who suspends judgment about whether or not God exists) is someone who believes that they cannot know whether p. Crawford (Reference Crawford2004) offers a similar account, on which suspending judgment about whether p involves some belief about one's own epistemic position. More specifically, Crawford's view holds that suspending judgment requires the belief “that one's epistemic perspective falls short of establishing whether p and thus that one does not know whether p” (Crawford Reference Crawford2004: 226). More recently, Masny (Reference Masny2020) has offered a belief account that he takes to be an amendment of Crawford's account. On Masny's account, suspending judgment about whether p constitutively requires (i) believing that one neither believes that p nor disbelieves that p; (ii) neither believing p nor disbelieving that p; and (iii) intending to judge whether p.Footnote ⁶

Recall that the intuitive difference between suspending judgment about whether p, and simply being belief-less about whether p, consists in something like being committed to being neutral or undecided about whether p, rather than simply being such that one is neutral or undecided about whether p. On belief accounts, we can cash out this commitment in the familiar terms of belief and its characteristic commitment to taking the believed proposition as true. To be committedly neutral or undecided about whether p is to take as true some proposition about one's deficient epistemic position with respect to whether p.

Perhaps the central objection to belief accounts is that the higher-order belief requirement is overly demanding. There are different ways of pressing this point. One objection that falls under this general heading is that taking up a higher-order belief requires a degree of cognitive sophistication that would preclude certain creatures who do not possess the capacity to represent to themselves their first-order doxastic states from being able to suspend judgment.Footnote ⁷ Some defenders of belief accounts, like Masny (Reference Masny2020: 5014–15) and Raleigh (Reference Raleigh2021: 2461), accept that implication and deny that it poses a problem for the view. For my part, I will say that I find it odd to think that dogs and young children can suspend judgment about certain matters (as opposed to merely being undecided about something), but disputes over intuitions only get us so far.

A related demandingness objection to belief accounts goes as follows. Friedman argues that the higher-order belief requirement is overly demanding, because it would preclude extremely epistemically cautious agents from suspending judgment. In an example from Friedman (Reference Friedman2013b: 155), an agent might refuse to judge that she knows when the train to Paris departs, and also refuse to judge that she does not know when it departs (perhaps because she doubts that introspection is a reliable basis upon which to determine what it is that one knows or doesn't know). Even so, it may still be possible for her to suspend judgment about when the train to Paris departs.

Masny (Reference Masny2020) rejects Friedman's claim that this case presents a counterexample to belief accounts. According to Masny, making this case successfully against belief accounts would require some further story about how it is possible for the epistemically cautious agent to not have beliefs about her first-order epistemic standing on the basis of some even higher-order belief to the effect that, say, introspection is an unreliable basis upon which to determine what it is that one knows. And Masny suggests that any such story would entail an implausible form of doxastic voluntarism. Just as an agent cannot simply make it the case that she does not believe something, on the basis of some further belief, so too is it implausible, Masny suggests, that an agent can make it the case she does not form any beliefs her epistemic standing, on the basis of some further belief (e.g., about the unreliability of introspection) (Masny Reference Masny2020: 5012–14).

But this is a precarious line of defense for a defender of a belief account to stake out. Belief accounts, after all, typically have the following basic structure: the agent who suspends judgment in fact neither believes that p nor disbelieves that p, and this absence of beliefs on her part is crucially sustained in part by some belief at the higher-order level about her epistemic standing.Footnote ⁸ Otherwise, these might only be accidentally related: an agent might form the belief that she cannot yet tell whether p (or the belief that she neither believes that p nor disbelieves that p), and then suffer a blow to the head, which results in her dropping any first-order attitudes she might have had about whether p. Surely that wouldn't be sufficient to make her count as suspending judgment about whether p. Belief accounts, then, seem to need some kind of story about the causal connection between the relevant higher-order belief and the agent's lack of belief at the first-order level. It's not clear why belief accounts wouldn't also suffer from the assumption that it is possible for a belief about one's deficient epistemic position vis-à-vis whether p to cause that agent not to form any p-related beliefs, if Friedman's objection suffers from a similar assumption.

We can also press the core of Friedman's objection in a different way, one that allows us to sidestep the sort of issues Masny raises. The central objection is that an agent who does not form any beliefs about her epistemic standing vis-à-vis whether p can nevertheless suspend judgment about whether p. Here's another argument for this claim, in brief. If suspending judgment about whether p requires the belief that one is in a deficient epistemic position vis-à-vis whether p, then if one has a reason to suspend judgment about whether p, then she has a reason (of the same kind) to believe that she is in a deficient epistemic position vis-à-vis whether p.Footnote ⁹ But there are cases in which an agent has a good reason to suspend judgment about whether p without having a reason (of the same kind) to believe that she is in a deficient epistemic position vis-à-vis whether p. And so, it's not the case that suspending judgment requires the belief that one is in a deficient epistemic position vis-à-vis whether p.

For concreteness, let's focus on Raleigh's (Reference Raleigh2021) view, on which suspending judgment requires the belief that one cannot yet tell whether p, based on one's evidence. The problem is that there are cases in which agents intuitively have reason to suspend judgment while also having reason (of the same kind) not to make any judgment about their epistemic position or the quality of their evidence. Consider, for example, the following case. Angie is a talented point guard who reliably makes three-point shots, and there is plenty of evidence to this effect (from her track-record, the testimony of her coaches, etc.). The problem is that forming any beliefs about her reliability at making three-point shots triggers crippling performance anxiety, making her less likely to make the shots. In this case, I take it that the fact that having beliefs about her reliability will frustrate her athletic goals provides good practical reason for Angie to suspend judgment about whether or not she's a reliable three-point shooter. But Angie does not have practical reason to believe that she cannot yet tell whether or not she's reliable, based on the evidence. Again, such beliefs about her reliability, which would include higher-order beliefs about whether or not she can tell that she's reliable, stand to trigger crippling performance anxiety. The practical reason she has to suspend judgment does not function as a practical reason to take up the higher-order attitude about her evidence or epistemic position with respect to that evidence.

We can generate other cases here. For example, there are many cases in which adding the relevant higher-order belief would only clutter the mind, which an agent plausibly has practical reason to avoid. There are countless many trivial whether p? questions that an agent might consider and suspend judgment on that are simply not worth the cognitive processing or storage space that a higher-order belief about one's epistemic position with respect to whether p might involve. There are other kinds of cases that might be relevant here, too – cases in which one has a practical reason to suspend judgment about whether p that is also a reason to turn away from the question of whether one's evidence makes it sufficiently likely or evident that p.Footnote ¹⁰ I consider a case of this sort below.

Return to the idea that suspending judgment about whether p involves, in some way or another, committing to neutrality or indecision about whether p. On belief accounts, this commitment constitutively requires taking as true some proposition about one's epistemic standing vis-à-vis whether p. But construing the commitment to neutrality or indecision in this way seems overly narrow. Suspending judgment about whether p, and so being committedly neutral or undecided about it, can manifest in different ways, some of which do not involve forming some belief about one's epistemic standing vis-à-vis whether p. Indeed, an agent can seemingly have good reason to suspend judgment without having any reason to form such a belief.

3. Sui Generis Attitude Accounts

There are other ways of thinking about what it might mean to be committedly neutral about whether p that don't centrally involve having some belief about one's epistemic standing. Let's consider Friedman's interrogative-attitude account, on which suspending judgment about whether p consists in taking up an interrogative attitude, one that takes a question, whether p?, as its content.

To get a sense of what it might mean to take up an interrogative attitude, let's consider what Friedman identifies as the key features of the functional profile of such attitudes. First, interrogative attitudes are goal-directed (Friedman Reference Friedman2017: 307). They aim, specifically, at answering the question it takes as its content. This feature of the attitude explains a second key feature: that having an interrogative attitude directed toward a question involves being oriented toward and sensitive to information that bears on that question (Friedman Reference Friedman2017: 308). The agent whose goal it is to close a question is disposed to take the means to achieving that end, which involves being alert to and responsive to new information that bears on the question. On Friedman's account, to suspend judgment about whether p is to be in an inquiring state of mind with respect to whether p. As such, dispositions associated with inquiry-guided behavior (in particular, sensitivity to information bearing on a question) are partly constitutive of the functional profile of suspension.

Friedman takes the functional profile of suspending judgment to inform the “normative profile” of suspending judgment – an account of when, and under what conditions, an agent ought to suspend judgment, has a reason to suspend judgment, and so on. According to Friedman, an agent has reason to suspend judgment when she has reason to take up other inquiry-related attitudes, like wondering about whether p, being curious about whether p, investigating into whether p, and so on. Reasons to suspend are, for Friedman, centrally reasons to inquire. And what kind of reasons are these? Friedman suggests that the sort of reasons that bear on suspending judgment will go beyond considerations that have to do with an agent's evidence and her epistemic circumstances. As she puts it, “inquiry looks to be a place at which the practical and epistemic collide. In inquiry we aim to know, and that aim can be prompted by pure epistemic interest or pure practical necessity, or anything in between” (Friedman Reference Friedman2017: 321).

But there are many cases in which an agent seemingly has reason to suspend judgment without having a reason to inquire. Note that the very sort of case we considered as an objection to belief accounts poses a problem for Friedman's interrogative-attitude account, as well. In the case we considered above, the agent appears to have a reason to suspend judgment without thereby having reason to form a belief about her epistemic standing. And she has no reason to form a belief about her epistemic standing because, intuitively, she has reason to be inattentive to the evidence about whether p, rather than sensitive to it. She has reason to turn away from inquiry into whether p, not to keep it open. If that is right, then we can reformulate the same core argument we gave above against belief accounts to apply to Friedman's interrogative-attitude account. If suspending judgment about whether p requires (or just is) taking up an interrogative attitude toward whether p?, then if one has a reason to suspend judgment about whether p, then one has a reason (of the same kind) to take up an interrogative attitude toward whether p?. But there are cases in which an agent has good reason to suspend judgment about whether p, without having a reason (of the same kind) to take up an interrogative attitude toward whether p?. And so, it's not the case that suspending judgment requires (or just is) taking up an interrogative attitude toward whether p?

Others have argued that there are cases in which the agent who suspends judgment about whether p does not also inquire into whether p. Raleigh (Reference Raleigh2021: 2467) notes that we can easily generate cases in which the relevant agent can intuitively suspend judgment about some trivial matter (e.g., the number of hairs in one's left nostril) without thereby inquiring into it. There are other cases, too, in which the relevant agent can suspend judgment on some question that she believes to be genuinely unanswerable. Archer (Reference Archer2019: 70–1), for example, offers a case in which an agnostic astronomer suspends judgment about claims about astronomical objects or events that take place outside of our Hubble sphere, while also believing that it is impossible to ever acquire the information necessary to settle questions about such claims. As a result, the astronomer gives up the aim of inquiring into these claims. Intuitively, the astronomer can suspend judgment about these astronomical claims without being in an inquiring state of mind with respect to them.

Friedman's response to these kinds of cases is to claim that either the relevant agent does not genuinely suspend judgment on the question (and that, instead, the agent is simply without any doxastic attitudes whatsoever, perhaps because she has no real interest in the matter); or that the relevant agent does suspend judgment on the question and is in some sort of inquiring state of mind with respect to it (even if the agent's desire or need to settle the question is rather weak). But for reasons that others have pressed against her account, I don't think that either of these interpretative options best captures what's going on in the cases we're interested in (see, e.g., Raleigh Reference Raleigh2021: 2468–70). As Archer (Reference Archer2019: 76–80) notes, the first option doesn't seem like the right way to interpret the agnostic astronomer case; she is certainly attentive to the astronomical questions on which she suspends judgment. Neither does the second interpretative option seem especially plausible. The agnostic astronomer genuinely believes that the relevant questions are not appropriate objects of inquiry, so attributing to her some (weakly) inquiring state of mind, one with the (not-very-pressing) aim of settling the question, would, as Archer notes, require interpreting her against her own sense of herself.

Lord (Reference Lord, Schmidt and Ernst2020) and Lord and Sylvan (Reference Lord, Slyvan, Simion and Brown2021) offer a way to amend Friedman's account in a way that would help block such cases. They introduce another sui generis attitude: namely, what they call an anti-interrogative attitude.Footnote ¹¹ On the view they defend, suspending judgment about whether p can manifest either as an interrogative attitude or as an anti-interrogative attitude. Interrogative and anti-interrogative attitudes commit the agent who has them to neutrality or indecision about whether p, though in importantly different ways. Whereas an interrogative attitude aims at resolving the question toward which it is directed, an anti-interrogative attitude aims at burying the question. Rather than involving an orientation toward or openness to information bearing on the question (as interrogative attitudes characteristically involve), anti-interrogative attitudes effectively insulate the agent from such information (Lord and Sylvan Reference Lord, Slyvan, Simion and Brown2021: section 6.3.2).

I agree that suspending judgment about something involves orienting oneself away from a question, or from evidence that bears on a question, in the ways that Lord and Sylvan describe. Cashing this out will be central to the account I sketch in the next section. The question is whether we need to add another sui generis attitude to our taxonomy of doxastic attitudes in order to accommodate this point. My view is that this move is unnecessary if we can accommodate the point their account seeks to accommodate with more familiar tools. If we don't begin with the assumption that suspending judgment is some kind of attitude that represents or constitutes one's neutrality – again, the starting point from which all of the accounts we've considered so far proceed – then the introduction of additional sui generis attitudes to capture the relevant data looks unmotivated.

Many philosophers have been too quick, in my view, to rule out the possibility of an account that does not begin with the assumption that suspending judgment is centrally a doxastic attitude of some kind. In the next section, I will sketch an account of suspending judgment that sees suspending judgment as a practical matter, rather than as a doxastic one.

4. Suspending as Refraining

Recall, from above, one of the observations that motivates the project of trying to locate more precisely the nature of suspending judgment: namely, the observation that suspending judgment about whether p is not the same thing as simply not making a judgment about whether p. My view is that this distinction between not judging and suspending judgment is just an instance of the broader distinction between (mere) omission and refraining. There are all kinds of things you're not currently doing that you're not refraining from doing. You're not currently watering my plants, for example, but you're not refraining from watering my plants. An agent's refraining from ϕ-ing involves something over and above that agent's not-ϕ-ing.

What marks the difference between refraining from ϕ-ing, and simply not-ϕ-ing? One distinguishing feature seems to be that when you refrain from ϕ-ing, your not-ϕ-ing is intentional, whereas not-ϕ-ing need not be. How should we understand, more broadly, the nature of intentional omissions? This question is connected to a much larger debate about the metaphysics of omissions.Footnote ¹² What I wish to do here is to draw on some recent work in this domain to provide a way of seeing how we might fill in the details of an account of suspending judgment that sees it as a kind of intentional omission. In this way, my account of suspending judgment offers a way of thinking more concretely about what suspending judgment as an intentional omission consists in, in a way that distinguishes it from a related view of suspending judgment in the recent literature by McGrath (Reference McGrath2021) that remains neutral on this question. I will return to some of the key differences between my account and McGrath's later in this section.

On my view, suspending judgment about whether p is essentially a matter of refraining from making a judgment about whether p. I understand judging that p (or judging that not-p) to entail that one forms the belief that p (or disbelief that p). So, refraining from making a judgment about whether p is incompatible with believing that p and disbelieving that p.Footnote ¹³ The agent who refrains from judging whether p omits to do something – namely, she omits to judge that p (or not-p) – and, like refrainings more broadly, this omission is intentional.

What makes the agent's omission to judge here intentional? As a first pass, we might say that the omission here is intentional in virtue of the fact that the agent's omitting to judge is caused, in the right kind of way, by that agent's intention not to judge.

But this simple account faces several challenges. One of these can be gleaned from a case that appears in Friedman (Reference Friedman2013a: 172), who offers it as an objection to accounts on which suspending judgment consists in simply withholding belief. Suppose that Sam intends not to judge whether p, which causes him not to judge whether p. Sam eventually forgets the matter entirely, and years later, cannot recall having ever intended not to judge, and has no memory of what his take on the issue was. At this later point, Sam is still such that he is in a state of not judging whether p, and this is preceded by his having intended not to judge at some earlier time. But he no longer appears to be suspending judgment about whether p. Friedman notes that, in such a case, we “seem to get stuck having to say that in virtue of ever having done the relevant cognitive act, S[am] will remain agnostic,” at least until he comes to make a judgment about whether p.

Here's how I think we can formulate the core challenge. We want to be able to maintain that agents who suspend judgment about whether p can at some point cease to suspend judgment, even if they never come to form a judgment about whether p. I might, for example, intentionally not judge whether contrastivism about knowledge attributions is true, having come to appreciate that a particular philosophical view I want to defend is consistent with it, but does not depend on it. Intuitively I can count as suspending judgment on the truth of contrastivism for now, but no longer count as suspending judgment years later, as I forget the matter and my attention shifts to completely unrelated projects. At this later point, I am simply belief-less about the matter. At the same time, we also want to be able to maintain that agents can suspend judgment about whether p over an extended period of time, not just at the moment immediately following the agent's having formed an intention not to judge whether p. A lifelong agnostic can, for example, continue to count as suspending judgment about whether or not God exists even after he no longer actively intends not to judge whether or not God exists. An account that analyzes suspending judgment about whether p in terms of intentionally omitting to judge whether p should be able to maintain that the agent's intention not to judge whether p can continue to guide or explain the agent's omission, without maintaining that the immediate causal role of the agent's intention permanently locks the agent into suspending judgment about whether p. But it's not clear that the simple account can do this. Once the agent's intention has done its causal work in bringing about the omission in judgment, how do we determine when that omission continues to count as relevantly intentional – and so, when that omission continues to count as suspending judgment?

We can look to some of the literature on the nature of intentional omissions more broadly for insight here. One way to improve on the simple account draws on Shepherd's (Reference Shepherd2014) account of intentional omissions more broadly, on which an intentional omission is grounded in the agent's intention not to ϕ, which appropriately causes the agent to be disposed not to ϕ. The agent's disposition not to ϕ is brought about by the intention's having made certain changes to that agent's cognitive and motivational systems (Shepherd Reference Shepherd2014: esp. 22–3). More specifically, an agent's intention not to ϕ typically causes that agent to form various beliefs and desires that help her execute that intention not to ϕ, in virtue of the roles that those beliefs and desires play in her planning behavior. In the case of the sort of intentional omission that we're interested in – intentionally not judging whether p – we might say that the agent who suspends judgment is someone whose intention not to judge whether p causes various changes to her cognitive and motivational systems which, in turn, dispose her not to judge whether p. For example, as a result of her intending not to judge whether p, the agent may want to avoid acquiring further information that bears on whether p, or aim at deliberately distracting herself with other matters, or engage in mindfulness exercises to take her mind off of the question, and so on.Footnote ¹⁴

Return to the case from above. The reason it seems inappropriate to describe Sam as suspending judgment later on is that Sam plausibly no longer retains, at this later time, the relevant disposition not to judge whether p, were the question of whether p? to become live for him again. To spell this out a bit further, the basic idea here is, roughly, that if Sam were to enter a context conducive to judging whether p, at this later time, it is not the case that Sam would not judge whether p. So even though Sam continues to be such that he does not judge whether p, his not judging whether p is not an instance of his suspending judgment about whether p, as he is no longer relevantly disposed not to judge whether p.

Of course, the question of how best to understand the nature of dispositions is contentious, and it would take us well beyond the scope of this paper to defend a particular account. But here is one way to think about the relevant disposition. To be disposed not to judge whether p, in the way that is distinctive of suspending judgment about whether p, is to be such that, across a certain range of contexts that are conducive to judging whether p, the agent does not judge whether p. What does it take for a context to be p-judgment conducive? To be clear, it need not be a context that is conducive to that agent's forming a correct or justified judgment. A context that is conducive to judgment is one that involves features that typically prompt the agent to form some kind of judgment, however initial or tentative. Whether or not a context is p-judgment-conducive may depend on features of the context itself, as well as on features of the agent. First, and most obviously, a p-judgment-conducive context must be one in which it is possible for an agent to form a judgment about whether p, where this might require certain things of an agent's cognitive abilities, among other things. Second, the fact that an agent has available to her some evidence that appears to her to bear on whether p, however tentative or partial that evidence is (and even if that evidence is radically misleading), can make her present context p-judgment-conducive. Finally, whether or not a context is judgment-conducive also depends on facts about the agent's motivations, desires, fears, anxieties, and so on. For example, an agent who is extremely anxious about whether p might be someone for whom a context with relatively little information bearing on whether p is judgment conducive. We might also think of agents who are indifferent to the question of whether p, and for whom a judgment-conducive context might involve quite a lot more information or evidence bearing on whether p. There are clearly many other features that affect an agent's inclinations to judge in particular different ways – the point is just the obvious one that what makes a context conducive to judgment depends, among other things, on what the agent is like, in addition to what the information available in that context is like.

It's also important to note that many cases of suspending judgment are cases in which the intention not to judge should be understood narrowly. After all, in many cases, agents who suspend judgment about whether p are not intending not to judge whether p in any future context. Sometimes, suspending judgment is strategic – one suspends judgment now in order to judge later. When Pat hears a bit of lurid office gossip about her friend Matt's activities and suspends judgment until she speaks with Matt himself, she intends not to judge in some, but not all, contexts that are conducive to making a judgment on the matter – namely, those that lack the particular kind of evidence she's after (namely, Matt's testimony). In suspending judgment, Pat intentionally insulates herself from certain kinds of evidence and other influences that may tempt her to form a judgment prematurely.

Should suspending judgment be essentially tied to an intention to judge later, if and when certain conditions obtain? Not in my view. This is a key difference between my account and McGrath's (Reference McGrath2021: 7) account, on which suspending judgment about whether p is to put off belief-forming judgment. Like my account, McGrath understands putting off judgment as a kind of intentional omission to judge, but on his view, this is an omission that is guided by the conditional aim of judging later, when (and only when) certain conditions obtain (which one does not yet believe obtain). McGrath distinguishes between this narrower notion of suspending judgment about whether p, and the broader notion of refraining from judgment about whether p, which he takes to be simply an intentional omission to judge that is not tied to any particular aim.

Why take care to distinguish between suspending judgment and refraining from judgment? McGrath (Reference McGrath2021: 5) motivates the distinction largely by appeal to the definitions and ordinary connotations of these terms: to suspend something – suspending a student from school, suspending train service, etc. – typically implies a temporary stoppage; that which is suspended is to be reinstated or resumed at some later point. Refraining from something, on the other hand, like abstaining or forbearing, is simply to prevent oneself from doing something; it does not carry the implication that one puts off the thing one refrains from. You might refrain from voting for the Republican candidate, or refrain from eating the fifth cookie, without returning to these at a later time. Refraining from judgment, for McGrath, is similar. You can briefly entertain a question that you don't know anything about or care about and put the matter to the side, not making any kind of judgment at all. You do not suspend judgment on the matter, but simply refrain from making a judgment about it at all. According to McGrath, these differences between suspending judgment and refraining from judgment, though subtle, illuminate genuinely different ways of being neutral about whether p, which we may be apt to miss if we were to collapse these into a single option.

I think we can acknowledge that there are different ways of being neutral about whether p, while also maintaining that a multiplicity of aims might guide suspending judgment in different cases. Sometimes suspending judgment is guided by the aim of arriving at a judgment later. But I don't think that we need to see all cases of suspending judgment as guided by the conditional aim of judging later, if and when certain conditions obtain. I think this comes out clearly in the Angie case, from above – we need not take Angie to be aiming to judge. Likewise in cases where one suspends judgment on trivial or junk subject matters. There are other, more interesting cases that are relevant here, too.Footnote ¹⁵ Imagine you're advising an undergraduate student on an honors thesis, a year-long project that requires the kind of intellectual commitment and persistence that you haven't observed in this student before. Her academic record is mixed, and you know that she has a tendency to set herself overly ambitious goals when it comes to writing papers that she's ultimately not able to meet. But she says she is committed to completing the thesis and promises that this time will be different. In the face of all the evidence you have of her track-record, let's suppose you decide not to make a judgment about whether she will complete the thesis. In this case, I think it's natural to describe what you're doing as suspending judgment about whether she will complete the thesis, and we need not see your suspending judgment in this case as guided by the aim of making a judgment later about whether she will complete the thesis.

Of course, by the final thesis deadline, presumably you will find out whether she completes the thesis, and so inevitably you'll arrive at some judgment here. But the aim of suspending judgment, in this case, is not (or need not be) to judge later, when the right conditions obtain, whether she will complete the thesis. One way to think about the aim of suspending judgment about whether she'll complete the project has to do with manifesting a certain kind of respect for her capacity to set an end for herself and to live up to it, something that you think might be precluded by making a prediction about the likelihood of her completing the thesis on the basis of the relevant evidence (Morton and Paul Reference Morton and Paul2018). As Morton and Paul (Reference Morton and Paul2018) put it, you resist taking up a diagnostic perspective toward her in assessing the likelihood of her completing the thesis, based on past evidence.

To be clear, the point is not that you ought to suspend judgment in this case. The point, rather, is to provide some support for the claim that a multiplicity of aims might guide an agent in suspending judgment – aims that need not include the aim of eventually making some judgment. Of course, one might respond that in these sorts of cases, the agent is not really suspending judgment, but rather refraining from judgment. My own view is that, unless the distinction is motivated by some other concern, it's not clear that we need to multiply categories of intentional omission.

5. Epistemic Reasons to Suspend Judgment?

On the view I've sketched, suspending judgment about whether p is fundamentally a practical matter, rather than a matter of taking up a doxastic attitude. This raises questions about how we ought to think about the nature of reasons to suspend judgment, many (if not most) of which seem to be epistemic. Indeed, one might argue that whatever my view gains in its ability to accommodate practical reasons to suspend judgment, it loses in its ability to accommodate epistemic reasons to suspend judgment. In this section, I want to say a little bit about how my account might handle epistemic reasons to suspend judgment.

Begin with a simple example. That there is an exactly even number of red and green jellybeans in the jar seems like a distinctively epistemic reason to suspend judgment about whether the next jellybean you will draw at random will be green. But it's not obvious how epistemic reasons can bear directly on intentional omissions, including intentionally not judging whether the jellybean will be green. Indeed, one might object that my account runs into the same basic problem I posed for the doxastic-attitude accounts above: namely, that there can be reasons (distinctively epistemic ones) for suspending judgment about whether p that are not reasons for intentionally not judging whether p.

To address this concern, we need to get clearer on what exactly epistemic reasons are. Suppose that epistemic reasons are just evidential considerations that bear on the likely truth or falsity of the object of a propositional attitude. So, for example, something is an epistemic reason in favor of believing that p just in case it makes p sufficiently likely. If that's what an epistemic reason is, then there are no epistemic reasons to suspend judgment on my account. But, by the same token, there also will not be epistemic reasons to suspend judgment on other accounts, including the doxastic-attitude accounts we considered above.Footnote ¹⁶ Recall that on Friedman's account, suspending judgment is a question-directed interrogative attitude that takes a question, whether p?, as its object, rather than a proposition. Questions cannot be true or false, so considerations that bear on the likely truth or falsity of the object of a propositional attitude cannot serve as reasons to take up an interrogative attitude toward whether p? There are problems here, too, for belief-based accounts of suspending judgment. Consider Raleigh's account, on which suspending judgment is reducible believing that one cannot yet tell whether p (based on one's evidence), and neither believing that p nor disbelieving that p on the basis of that higher-order belief. Suppose we grant that the fact that there is an exactly even number of red and green jellybeans in the jar makes it more likely that you cannot yet tell whether the next jellybean you draw will be green, and so serves as an epistemic reason for the higher-order belief that you cannot yet tell whether or not the next jellybean you draw will be green. Even so, it's not clear how this fact about your evidence functions as an epistemic reason for, at the first-order level, neither believing nor disbelieving that p.

What about other views about what epistemic reasons consist in? Suppose we adopt a broader evidentialist view, on which there are facts about which doxastic states are adequately supported by the total evidence (compared with the alternatives), and these facts can constitute epistemic reasons that count for or against being in these doxastic states. So, for example, the fact that some state is adequately supported by the total evidence constitutes a reason in favor of being in that doxastic state, just as the fact that some state is not adequately supported by the evidence constitutes an epistemic reason against being in that doxastic state. In the case we're considering, the belief that the next jellybean will be green is not adequately supported by the total evidence. In the same way, disbelieving that the next jellybean will be green is not adequately supported by the total evidence. You have epistemic reasons against both believing and disbelieving that the next jellybean will be green, which means that you have an epistemic reason in favor of the remaining alternative – neither believing nor disbelieving that the next jellybean will be green.

Now, simply neither believing nor disbelieving p is not the same thing as suspending judgment. How, then, do we get a reason to suspend judgment about whether p? One suggestion is that you have a kind of instrumental reason to suspend judgment. The idea here is that if you have an epistemic reason to neither believe nor disbelieve that p, and suspending judgment about whether p makes it more likely that you neither believe nor disbelieve that p, then you have a reason to suspend judgment. In other words, the epistemic reason to neither believe nor disbelieve that p transmits as a reason to take up a certain means (namely, suspending judgment), when the means is such that it is relevantly available to you (e.g., it is at least possible for you to suspend judgment) and the means makes it more likely that the end (in this case, neither believing nor disbelieving that p) is achieved.Footnote ¹⁷

Of course, there is more to say about this simple transmission principle and how it would cover other cases. For example, there are cases in which one currently has good, but not conclusive, evidence that supports believing that p, but there is much better (indeed, conclusive) evidence that bears on whether p that is soon forthcoming. In such a case, it seems that the fact that conclusive evidence is forthcoming seems to provide a distinctively epistemic reason to suspend judgment about whether p. Here, too, I think that we can use the transmission principle to generate epistemic reasons to suspend judgment, where suspending judgment is understood as a means to bringing it about that an agent realize an epistemically valuable end – namely, in this case, believing what the total evidence supports (or perhaps having a belief that qualifies as knowledge). The basic idea here is that epistemic reasons for suspending judgment are best understood as reasons for doing something that will help bring about certain epistemic ends.Footnote ¹⁸

One might worry that this account of how we can get epistemic reasons to suspend judgment fails to respect a key feature of doxastic deliberation: namely, that deliberation about what to believe is transparent onto deliberation about whether p.Footnote ¹⁹ When one deliberates about whether to believe that the next jellybean drawn from the jar will be green, one deliberates about whether the next jellybean drawn from the jar will be green. Finding that there is an exactly equal number of green and red jellybeans in the jar, one suspends judgment about whether the next jellybean drawn will be green. This entire episode of doxastic deliberation, which eventually concludes in the agent's suspending judgment about whether p, is focused on the question of whether p?, not on the question of which of one's doxastic options is best supported by the evidence.

First, for reasons that Rinard (Reference Rinard2017: 138–9) and others have pointed out, I think that the ‘transparency’ claim about doxastic deliberation is question begging. I do not think there is any interesting sense in which deliberation about whether to believe p? gives way to the question of whether p? unless we already presuppose that non-evidential considerations cannot bear on whether to believe or to suspend judgment. Second, to the extent that failing to respect transparency is a problem, it is not a problem that is unique to my account. As I mentioned above, any account of what reasons there are to suspend judgment will have to accommodate reasons to suspend that do not bear directly on the question whether p? – and that is so even if one denies that one can suspend judgment for what are clearly practical reasons. The fact that better evidence is forthcoming, for example, is not a reason that bears directly on the question whether p? – one must step outside of deliberation that is exclusively p-focused in order to recognize and respond to this sort of consideration. Similarly, learning that you've been given a drug that subtly alters your ability to reason about whether p can be a reason to suspend judgment about whether p, but this information about the drug is not information that bears directly on the question of whether p?.Footnote ²⁰ Given the apparent ubiquity of reasons to suspend judgment that do not bear directly on the question of whether p? – considerations that include not just the obviously practical, but also those that concern the nature of one's evidence – I think it would be distorting to try and fit all episodes of doxastic deliberation into a model that sees it as exclusively whether p?-focused.

Two final things to note about the transmission of epistemic reasons to suspend judgment. First, this principle of transmission is compatible with different views about what epistemic ends we have reasons to bring about. One need not think that our epistemic ends concern having doxastic attitudes that are best supported by the evidence – one could plug in a different conception of what one's epistemic ends are and still make use of the transmission principle. Second, to say that there is an epistemic reason to suspend judgment about whether p that is transmitted from an epistemic reason to be such that one neither believes nor disbelieves that p is not yet to say anything about the normative authority or the weight of that reason. (As a side note, I am skeptical that there is a distinctive domain of epistemic normativity that functions independently of practical normativity more broadly, but I don't want to rely on that controversial view here.) The transmission principle I've just sketched offers a way of understanding how distinctively epistemic reasons might bear on suspending judgment, but it is also compatible with views on which facts about evidential support for particular doxastic states do not have normative authority.

6. Concluding Remarks

We began by noting a common view that, within a full-belief framework for doxastic attitudes, there are three basic doxastic options available to an agent, with respect to any proposition p: to believe that p, to disbelieve that p, and to suspend judgment about whether p. I think this picture is misleading. Instead, the three basic options here include believing that p, disbelieving that p, and neither believing that p nor disbelieving that p. Suspending judgment about whether p is best understood, on my view, as a kind of practical means to maintaining or bringing it about that one neither believes nor disbelieves that p. In this way, suspending judgment is importantly different in kind from the other two basic doxastic options available to an agent.

The practical nature of suspending judgment raises several questions about the relationship between suspending judgment and epistemic norms and epistemic aims. On the relationship between suspending judgment and epistemic norms, I provided a sketch in the last section of how epistemic reasons might transmit as reasons to suspend judgment, in a way that would help to vindicate certain connections between having reasons in favor of suspending judgment and having reasons against believing that p and disbelieving that p. To be sure, further questions remain about how to determine the normative weight of transmitted reasons to suspend judgment, and whether (and under what conditions) such reasons contribute to determining when an agent epistemically ought to suspend judgment. The goal for this paper has just been to make space for a further story about how suspending judgment might be subject to epistemic norms, even if we conceive of suspending judgment as fundamentally practical.

I want to close by briefly touching on the relationship between suspending judgment and epistemic aims. It is a central feature of several accounts we canvassed above that suspending judgment about whether p is tied closely, if not constitutively, to keeping the question of whether p open, either so that one can inquire further into whether p, or simply so that one might return to the matter later when conditions for judging improve. By contrast, on the account I've defended, there is no essential connection between suspending judgment and these further epistemic aims. We can suspend judgment about whether p in order to return to the question later, or to allow inquiry into whether p to proceed. But we can also suspend judgment to shut down lines of inquiry, or to see to it that we do not (ever) make a judgment about whether p. In my view, flexibility with a variety of different aims an agent might take up in suspending judgment about whether p is something that an account of suspending judgment should aim for. And this sort of flexibility is worth aiming for not just so that we don't prematurely rule out cases, but also so that we allow for the possibility that I think many overlook: that suspending judgment about whether p, even if subject to epistemic norms, is perhaps not fundamentally an epistemic matter.Footnote ²¹