The Rationality of Emotions Across Time

Abstract

Philosophers have struggled to explain the mismatch of emotions and their objects across time, as when we stop grieving or feeling angry despite the persistence of the underlying cause. I argue for a sceptical approach that says that these emotional changes often lack rational fit. The key observation is that our emotions must periodically reset for purely functional reasons that have nothing to do with fit. I compare this account to David Hume’s sceptical approach in matters of belief, and conclude that resistance to it rests on a confusion similar to one that he identifies.

Résumé

Les philosophes ont longtemps peiné à expliquer le décalage entre les émotions et leurs objets à travers le temps, comme lorsque nous cessons de pleurer ou de ressentir de la colère malgré la persistance de la cause sous-jacente. Je défends une approche sceptique qui soutient que ces changements émotionnels manquent souvent d’adéquation rationnelle. L’observation clé est que nos émotions doivent périodiquement se réinitialiser pour des raisons purement fonctionnelles qui n’ont rien à voir avec cette adéquation. Je compare ce point de vue à l’approche sceptique de Hume en matière de croyance, et je conclus que la résistance à cette approche repose sur une confusion similaire à celle qu’il identifie.

1. Introduction

In this article, I offer a sceptical solution to a problem about the rationality of the emotions that has engaged philosophers. As I will be characterizing the problem, many of our emotions follow a pattern of spiking in reaction to some change that we encounter, and then fading, even as the object of our emotions remains invariant. Thus, when we have some great success, we feel elated, but quickly return to our baseline despite a lasting improvement in our condition, as if we walked along a hedonic treadmill. Or a spouse dies and we experience intense grief, only to return to our baseline a few months later, even though our loss persists unabated. What makes this a problem is that our emotions seem subject to a normative fittingness or reasons-responsiveness requirement, and yet as our emotions fade, they don’t seem to meet that requirement. We experience huge gains or losses, but quickly recalibrate in ways that don’t track the reality of our situation. ¹

Philosophers have addressed this conundrum variously, with some denying the premise, insisting that grief or happiness persists despite an empirical literature suggesting otherwise, while others argue that the object of our emotions itself changes over time, which in turn renders our changing reactions reasonable. ² Several writers propose conceptions of fittingness or reasonableness that are dynamic in nature, so that the overall trajectory of our emotions makes sense, even when they look out of kilter at a given instant (Goldie, 2012, pp. 56–75; Na’aman, 2021).

Here, I argue for another approach that might be described as a sceptical solution to the problem. In general, sceptical solutions are a genre of response to normative challenges to common sense, e.g., the claim that induction cannot be justified. The canonical instance is David Hume in the Enquiry Concerning Human Understanding. As I will understand it, a sceptical solution involves (a) agreeing with the sceptic that our practices cannot be given the “straight” justification that the sceptic demands; while (b) trying to offer a satisfying naturalistic explanation of our practices; that (c) is supposed to reassure us that justification isn’t really required in some sense, and that thinking that it is involves a confusion. ³

Analogously, I will be arguing that (a) our rapid return to baseline when those we love drop dead, or after we have a transformative improvement, is indeed not fitting or reasons-responsive; but (b) there is a satisfying explanation of this pattern built into the nature of the emotions, which must reset after spiking in order to perform the various tasks the emotions have in our psychology (anger would cease to be useful if we remained permanently angry at unrepentant friends and relatives); and yet (c) it is a confusion to imagine that we need a justification for our unruly emotions as they reset. The most commonly cited reason for worrying is that we are otherwise subject to an “unrealistic” or overly demanding norm that requires us to be angry or to grieve forever, but this is a mistake, as I show. Our psychology gives us little choice in the matter, making worries about demandingness moot, and, in any case, reasons of fit are counterposed by practical reasons that we are free to invoke whenever we wish. The various reasons we have for our emotions — causal-functional, normative reasons of fit, and practical — can come apart, and nothing in my view prevents us from drawing on practical considerations. The sceptical view that our emotions in time often make no sense suggests that we suffer from a kind of blindness, but not that we face impossible demands.

2. The Problem of Emotions Across Time

Some variations of the problem focus on particular emotions like grief or anger, or specific psychological processes like resilience, but the problem is more general. We begin with the assumption that:

1. (1) Some emotions are subject to a fittingness or reasons-responsiveness requirement, and when they don’t meet that requirement they are subject to rational criticism.

Perhaps some would wish to categorize moods like depression as emotions and deny that these are subject to rational criticism, but we can set moods aside. What is under discussion are states that take intentional objects, so that we can feel X at, about, or that Y, such as fear, anger, envy, jealousy, happiness, sadness, and grief (Greenspan, 1988, pp. 5–6; Kenny, 1963, p. 9). ⁴ These emotions can misfire when they are unsuited to their objects. We fear a fluffy bunny that poses no threat; we rationalize and feel happy when our hopes and dreams are crushed (“I didn’t want that job anyway!”).

There are different ways of thinking about the objects of our emotions. In one sense, we fear the bear that menaces us; in another sense, the object of our fear is the condition that our emotion is supposed to fit — the fact that we are in danger. It is the latter sense of “object” that will be relevant to us (Baier, 1990; Kenny, 1963, pp. 41–43).

There are also different ways of capturing the fundamental point that emotions are subject to normative criticism. In one idiom, the emotions sometimes don’t “fit” their objects. In another, they fail to respond to the reasons we have for or against certain responses. Although these idioms carry different nuances, those differences won’t matter here, and I will use these expressions interchangeably. ⁵ What matters is that it isn’t anything-goes when it comes to the emotions. If you panic and start screaming when you see fluffy bunnies (perhaps because of a hypnotic suggestion, or a tragic accident), something has gone wrong. If you persist, or if the reaction is strong enough, we’ll start saying you’re delusional; we may suggest medication (D’Arms & Jacobson, 2003).

Consider next the emotions across time. Sometimes the emotions neatly track their objects, spiking in reaction to the novel presence of something good, bad, or dangerous, and fading in response to its absence. A bear appears and we’re scared; it withdraws, and we return to baseline. But, of course, often things are more complex, as when the changes are long-lasting, and yet our emotions don’t follow suit. The change-state persists while our emotions spike and reset. When our loved ones die, they tend to stay that way; we get a promotion and keep it; we have a child who remains with us; we enter the trenches and fight under fire for years. And yet our emotions usually do not persist; they return to baseline. (At first we jump at every bang in the trenches, but later we just light a cigarette. ⁶ ) The puzzle arises from the mismatch between the emotional reset and the invariant changes to which we are responding.

Of course, sometimes the meaning of events changes due to our shifting circumstances or changes in ourselves. It’s no mystery why I don’t mourn the loss of something that I stop caring about, or that is no longer a good to me. In sixth grade, my friend Jeff Wheeler failed to invite me to his birthday party, a deeply traumatizing event at the time. For all eternity, Wheeler will have failed to have invited me to his party. But subsequent changes in my social context and my value-system mean that I no longer register this as a loss. It has been neutralized, we might say (cf. Schönherr, 2021, p. 241). Soldiers who attain a fervent belief in the afterlife may no longer fear death. Agnes Callard points out that there is something bizarre about eternal anger in response to the eternal fact of having been betrayed, but the neutralizing influence of later events explains at least some of these cases (Callard, 2018, pp. 123–124). Even without an apology, anger makes little sense if I no longer care about the wrongdoer, or if we’ve “moved on” and the current state of our relationship means I no longer care about the past injustice. (Outrage at a betrayal from long ago that no longer is relevant to our relationship doesn’t seem fitting, as with Jeff Wheeler.) Call these kinds of cases “neutralized changes,” and set them aside. What we’re interested in are cases in which it’s clear that we have gained something that remains wonderful to us, or have suffered a disastrous loss that remains disastrous, and yet, after the initial spike, our emotions nevertheless revert to their baseline. We cannot claim that the loss of a spouse or birth of a child is no longer relevant to us after a few months.

With some trepidation, I offer Figures 1 and 2 as highly schematic illustrations. They chart the intensity of our emotional reaction, E (our level of fear, grief, or happiness, say), across time, compared to the relevant change in the intentional object, O (how good, bad, or dangerous, say, we judge some event to be for us). Figure 1 represents a neat case, like our perception of the threat that a bear poses as it gradually approaches and then withdraws, as our level of fear increases and then subsides.

Figure 1. Neat case.

Figure 2. Messy case.

Figure 2 is the messier case of suddenly registering a long-term change, with downward slope to allow for an alteration in the meaning of events later on.

Perhaps O is the sense of loss due to the death of a loved one, where, after 50 years have passed, it might become difficult to view the loss as such at all. (You find it impossible to regret your new spouse and the family that you now have.) But, even then, Figure 2 illustrates that your emotions reset on a completely different schedule, returning to their baseline after weeks or months, not years. In a case like Figure 2, our emotions are wildly out of sync with the loss that befalls us.

We can now summarize these observations:

1. (2) Our emotions often follow a trajectory whereby they initially reflect the nature of some long-term change in our situation, but then dissipate, and, at a later stage, they no longer fit or rationally reflect the relevant changes, even though nothing subsequently has neutralized these changes.

Although I use the careful language of “some” and “often,” these aren’t weasel words. (1) and (2) capture the standard case when it comes to great events like having a child, the death of a loved one, or going to war: the very best and very worst things happen to us, and yet, for most people most of the time, after a short while, we recover our equilibrium. Of course, sometimes that is not the case. We lose someone and never recover; we so treasure the birth of a child that we are forever far happier. But this seems to be the exception, not the rule.

The problem that (1) and (2) generate is that when our emotions reset, they no longer reflect the changes that have taken place; we aren’t responding to the reasons we have and our emotions no longer fit the circumstances. We receive or are denied tenure, but after a few weeks, we no longer respond significantly to this important development; we get married or lose a loved one, but no longer register this to any great extent once we reset. ⁷ Making this point will seem eccentric, just as normative criticisms of induction seem odd by the standards of common sense. But just as it would be an important insight that induction does not proceed from an inference we can justify by appeal to reason, so too it would be an important insight were it true that our emotions systematically make no sense, since they periodically revert to baseline, making it inevitable that they do not fit our surroundings.

We can add, separately, that there is an inconsistency between our reactions when they spike and when they reset. Presumably, they can’t both be the fitting reaction, since they are quite different. It is hard to know what to make of the notion of appropriate fit if it turns out that any degree of fear, anger, or grief turns out to be fitting, whether it be the massive spike or the baseline level.

To summarize the problem:

Problem: Our emotions routinely fail to fit our circumstances or respond to the reasons we have to feel a certain way, because our emotional reactions to changes around us are generally temporary and follow a pattern of baseline-spike-reset. When this happens, we are in the same situation as one who panics at fluffy bunnies or greets his failures with joy. When we fail to respond to what remain huge losses or gains, we manifest a form of irrationality that appears pervasive.

The general point is that if we do accept (1) and the idea that screaming in horror at fluffy bunnies means something has gone wrong, then it’s hard to see how to avoid extending this observation to the everyday case of emotions reverting to baseline and failing to reflect our circumstances.

3. The Sceptical Solution

Let us invert the usual order, and take up the sceptical solution to the problem first, before turning to the straight solutions. The sceptical solution takes as its point of departure a natural feature of the mind, just as Hume invokes our psychological response to the “constant conjunction” of phenomena in addressing induction. The relevant feature is one I’ve alluded to several times already, the tendency of the emotions to reset after spiking. This fact is perfectly familiar, of course. When you get a parking ticket and feel down, you recover; when you get angry, you eventually let it go.

Psychologists have noted several proximate causes of resetting. They describe such mechanisms as habituation, rationalization of failures, substitution effects, changes in focus, and many others (Ekman, 1999, p. 54; LeDoux, 1996, pp. 261–262). (Often the context is that we tend to overlook these mechanisms, and as a result fail to recognize our tendency to reset; Gilbert et al., 1998, p. 619.) When it comes to happiness and sadness in particular, we seem to walk along a treadmill, though occasionally we can shift to a new baseline, if we escape or fall into poverty, say, or end or enter into a toxic relationship — situations that tend to generate a steady stream of new positive or negative experiences. Some researchers assert that few positive or negative changes affect us in significant ways for more than a few months, which is illustrative of the point, though there’s no need to take such a bold stance here (Suh et al., 1996).

But behind these proximate vehicles for spike and reset, there seems to be a more general phenomenon at work. The emotions have various jobs to do, functions to fulfill. There will be disagreement about how exactly to characterize these functions, but few people doubt that fear alerts us to danger and primes us for action, or that happiness registers positive changes and motivates us to produce more of the same. ⁸ Some researchers characterize the emotions as automatic programs that perform such tasks as increasing blood flow to the muscles to sustain a flight response, but we needn’t (and shouldn’t, in my view) take such a narrow view (Damasio, 1995, pp. 173–174). ⁹ We can also grant that some very abstract emotions like schadenfreude are mostly epiphenomenal; they don’t really do much apart from the faint pleasure we experience when making the relevant judgements. But few philosophers or psychologists think all or most emotions are like this. Anger, love, and jealousy seem to accomplish things for us. ¹⁰ They would be broken if they failed to alert and motivate us in various ways. And the central claim I wish to make is that many of the emotions cannot do their jobs without resetting.

Consider anger in the context of core relationships like friends, children, or parents, and assume that conflict isn’t resolved by apologies and forgiveness, perhaps because you just see things differently. Time goes by; days turn into weeks. Suppose that you start off furious with a parent and the emotion freezes in place; your intensity of anger flips to maximum and stays that way. The problem is that if your emotions don’t reset, you have deprived yourself of the capacity to register further changes in the relationship that would warrant anger and highlight the need for action. But this is fundamentally necessary for anger to work. Setting aside the fine-grained details of what exactly it is that anger does, we will be in trouble if once our parents annoy us at age 10, we remain angry at them forever. It would be like having an alarm that went off as soon as it was deployed and couldn’t be reset. Of course, sometimes we do stay angry for a long time and relationships are ruptured forever. But many relationships are ones that are not easy to get out of, and a lot of our anger is over tiny things like not washing the dishes. We need to respond to the ever-changing dynamics in our relationships without rupturing them, while still keeping track of whether we are being wronged and how badly. This is what resetting enables. ¹¹

This point generalizes. Remaining permanently fearful (of death, say), or sad, or happy would prevent these emotions from functioning properly. If my intensity of happiness is set at maximum once I have a child, say, I will no longer be able to enjoy the benefits of fluctuating emotions that alert me to changes to which I need to respond in various ways, say by reinforcing positive behaviour. If I grieve forever when a child dies, feeling that all is lost and nothing matters, I won’t have the benefits of adequately registering and responding to other adverse events. And, on reflection, it seems clear that it isn’t just a coincidence that the emotions both reset, and wouldn’t perform their various functions unless they did so. Resetting is a natural part of how the emotions work, since they wouldn’t work without resetting.

Another way of making the point is that the emotions are intimately connected with change. This isn’t a logical necessity, of course. Hume and others note the cool, calm emotions that sit in the background for years — the quiet love for a grown son or daughter, as opposed to “alarm bell” emotions (Damasio, 1995, pp. 173–174; Damasio, 1999, pp. 50–55; Hume, 2001, p. 181). But obviously registering changes and preparing us to face them is a core function of our emotional systems, even if it isn’t the only one. Are we in danger? Are we suffering setbacks and need a course-correction? Should keep up the good work? We would be poor agents if we didn’t respond to changes in the landscape. We might go so far as to say that emotions like happiness and sadness are more like derivatives, measuring the slope we’re on, as opposed to integrals telling us the absolute state of things. The emotions are often about improvement, not where things stand overall. ¹² But, for the emotions to respond to change, they must themselves reset. The baseline-spike-reset dynamic is inevitable if our emotions have jobs to do related to a dynamic environment that calls for constant updates in our perceptions and motivations.

Notice that the need for reset is completely independent of facts about fittingness. Our emotions are bound to reset whether or not the circumstances merit them. The fact that anger won’t do its job if it remains switched on is independent of facts about whether anger remains warranted — whether there has been an apology or change in the relationship. There is thus a perfectly sufficient, natural explanation for our emotional tendencies as depicted in Figure 2 that is orthogonal to what reasons we may have.

To summarize:

The Sceptical Solution: The emotions have important functions to fulfill, jobs to do for us. Without periodically resetting, they would no longer be able to fulfill those functions. This fact is capable of explaining why our emotions often follow a characteristic trajectory of baseline-spike-reset, independent of what type of response is fitting. Often our responses are in fact unfitting, as when we suffer horrendous losses or make fabulous gains but return to our hedonic treadmill after a few weeks.

On this view, there is a trade-off between perpetually registering where things stand in respect to the objects of our emotions, and being able to respond to frequent changes. The latter requires constant resetting, which is incompatible with the former. Unsurprisingly, humans have arrived at a point somewhere between the extremes (we typically grieve for weeks or months, not seconds or decades), but that point still gives substantial weight to tracking changes and allows for the state of affairs depicted in Figure 2, whereby we quickly stop tracking the relevant changes emotionally.

A comparison may help. The sceptical view says that we often suffer from a kind of emotional blindness, an incapacity to take in all that is happening to us. But, on reflection, we suffer from many, similar incapacities, and I am merely highlighting one of them. For example, our emotions are limited in their dynamic range, just as cameras or television screens are. Ten thousand deaths are far worse than 10 deaths, but our emotions don’t register the magnitude of this difference, just as cameras cannot capture pure black and the brightness of the sun. We don’t feel 1,000 times worse on hearing that 10,000 people have died than 10; we don’t feel 1,000 times more motivated to help, or 1,000 times more scared at the prospect. Of course, we maintain the ability to make the relevant discriminations, and sometimes we succeed in rationally responding to the difference, say in making appropriate resource allocations in a war. But our emotions quickly run out of headroom. We can only feel so awful. Once we feel pretty terrible about 10 people dying, there isn’t enough dynamic range left to feel another 1,000 times worse. Some of our systems of perception, like hearing, compensate for these limitations with tricks like logarithmic scaling, but even that has its limits (try listening to a rocket launch and then a whisper), and the range that is relevant to the emotions exceeds any plausible scaling technique once thousands or even billions of lives are at stake. Our emotions, like our ears or eyes, don’t work on very big or very small things.

The point of the comparison is that it would be silly to search for complicated, rational grounds for caring less and less about each marginal life that is lost. Once we see the natural explanation above, we can dismiss any suggestion that our emotions are accurately tracking a decrease in the marginal significance of death. The same applies to the sceptical solution. Given the need for reset, it was inevitable that we wouldn’t respond emotionally to persistent changes, and insisting on complicated rational grounds for the return to baseline is as misplaced as insisting that there are good reasons that our emotions fail to register the difference between 10 and 10,000 deaths.

4. A Straight Solution

We might prefer a straight solution to the sceptical view. Instead of superficially addressing all of the candidates, my strategy will be to investigate the most promising one at some length in the next section, but let me gesture toward a few other approaches.

One solution is to deny (1), which means saying the emotions are never inappropriate or unfitting. This might be possible on a pure Jamesian view of the emotions as physical responses, but even views in the spirit of William James are rarely this extreme. ¹³ To take this way out would require denying that panic isn’t the right response to fluffy bunnies. More plausibly, some view particular cases like grief as falling outside the scope of (1) because grief is said to be a complex activity comprising a multitude of emotions, not an emotion in itself (Cholbi, 2022, pp. 38–47). But we don’t want to quibble over particular cases here, since the problem is a wide ranging one that we’re unlikely to evade by pressing on the analysis of any one emotion.

We can reject (2) by denying the empirical claim that emotions tend to spike and rapidly dissipate. Here, too, we shouldn’t quibble over individual instances. If we focus on cerebral emotions like schadenfreude, (2) will seem less applicable, but that is why (2) isn’t universalized; more primal emotions like anger, happiness, or jealousy are more often relevant. Denying that these emotions tend to engage and then dissipate over time, even when the object of emotions persists, would come at a high cost, since we would have to explain why, for example, we aren’t all much more angry, happy, and jealous than we are.

Another option is to say that the object of our emotions changes over time. If the object of grief isn’t our loss but our own life circumstance in light of the loss — if grief is ultimately self-directed — then it’s easy to see why grief would dissipate, since over time we adjust to the new reality (Nussbaum, 2001, p. 82.). But apart from doubts about the solipsistic nature of this interpretation of grief, we are once again drifting into squabbling over particular cases.

I don’t pretend to have disposed of these straightforward solutions, but I don’t think they are particularly attractive either. They either don’t get at the generality of the problem, or else they incur very high theoretical costs. What we need is a straight solution that holds out hope of addressing the problem in its full generality.

5. Dynamic Fittingness

A better response says that fittingness is dynamic. On this view, there is no inconsistency between our initial and our later emotional states because what reaction is fitting depends on our earlier reactions. We cannot infer what the appropriate emotion is at present in isolation from how we have responded thus far. Making sense of anger at time T requires knowing things like whether we have we been angry for years or moments, or more exactly, whether we have undergone the various processes that anger involves, such as registering and examining an injustice. Grief will strike us differently depending on whether or not someone has absorbed their loss. Oded Na’aman writes:

[S]ome emotions […] are rationally self-consuming. Whether an emotion fittingly persists might not only depend on the persistence of the facts that constitute the fitting reason for it, but on the fitting evolution of the emotion over time […]. [T]he fact that an emotion has persisted for some time might itself render its dissipation fitting […]. (Na’aman, 2021, p. 251)

Berislav Marušić says that “reasons for grief expire: over time, as we grieve, it becomes not wrong to grieve less” (Marušić, 2018, p. 16). If our emotions are self-consuming and our reasons for having them expire, we might offer a straight solution to the problem. We can just insist that in the cases that might disturb us, our later reactions are reasonable conditional on our earlier reactions, since they are part of a fitting process.

Two kinds of ideas motivate this dynamic suggestion. First, there are particular cases like amusement or surprise that seem to call for a dynamic treatment. Laughing at a joke for hours on end looks misplaced, and shock at a surprising event makes sense only initially. Second, these authors emphasize (at least intermittently) that what matters isn’t the number of seconds, the mere passage of time, but the “temporally limited functional role” of the emotions (Marušić, 2018, p. 16). What counts is whether I have processed my grief, meditated on my culpability (diminishing feelings of guilt), and in general gone through certain steps in the right order (Na’aman, 2021, pp. 253, 259–260). In fact, my emphasis earlier on the jobs that the emotions do for us might be stood on its head. Whereas I claimed that these jobs cannot be performed without a reset, the dynamic view can be supported by insisting that once the emotions have performed their functions, they are no longer fitting. There will be disagreement, again, about how to characterize the work that is done in happiness at a childbirth, or in anger at an injustice, but as long as something is being accomplished, we might invoke that very fact to explain why there’s nothing unreasonable about emotions that don’t fit some fact in the past, conditional on the emotions having accomplished their tasks subsequently.

We can agree about cases like surprise or amusement, since neither of these involves a persistent change. Surprise is a reaction to an unexpected novelty, which ceases to be true of anything once it is registered. Amusement is much harder to explain in detail, but it, too, seems to have some relation to novelty and recall; the same joke repeated over and over is likely to seem less funny, at least until we’re distracted enough for it to hit us as fresh again. But clearly these are poor analogies for persistent change. What’s puzzling about the cases we have discussed is that your whole life may alter for the better or worse in a way that, by construction, remains better or worse, and yet your emotions fail to reflect that fact. No such puzzle presents itself in these other cases. It’s not even clear that something like surprise supports a dynamic interpretation. It doesn’t make sense to feel surprised once you’ve noticed the unexpected, but not because you felt previously surprised and your reasons for surprise have expired. There is just no reason at all to feel surprised by something of which you are already aware.

We can also agree that the emotions often perform tasks or play roles in the mental economy, always allowing for local disagreements. (For example, we might be sceptical that grieving involves going through a series of steps that produce “healing,” traditional lore for which there is little support; Bonanno, 2010, pp. 21–22.)

However, dynamic views run into serious problems. Christopher Howard argues that these views fail to make sense of the fact that the grounds of our emotions persist, even after the emotions have been engaged:

A fact can explain why a would-be fit-making fact does or doesn’t make something fitting to regret if and only if that fact can also explain why the would-be fit-making fact does or doesn’t make that thing regrettable […] [T]he fact that you’ve already felt a great deal of grief for the loss of your loved one can’t explain why the fact that you’ll never hear their laugh again no longer makes it fitting for you to grieve the loss, since this fact doesn’t explain why the fact that you’ll never hear their laugh again no longer makes the loss of your loved one grievable. (Howard, 2023, p. 90)

Perhaps this way of putting it sounds uncomfortably close to begging the question, since it is just a straightforward implication of dynamic views that this is false. If fittingness is conditional on prior emotion, then sometimes regret won’t be fitting even though something remains regrettable, i.e., when you’ve done the regrettable, regretted it, and thus (on the dynamic view) no longer have reason to regret. However, I think that Howard is on to something here. The basic problem is that dynamic views fail to address the persistence of the underlying fact-patterns to which we’re reacting, and yet these cannot just be ignored. The issue can be made vivid by highlighting cases in which those patterns are rendered salient again after we’ve reacted to them emotionally.

Suppose you have a child, and you experience an incredible high — this event feels like it changes everything, as if your life holds new meaning. For a few weeks, you are really happy, despite the sleepless nights. But then life returns to normal; your emotions reset. A few months later, the worst happens: an article is rejected. You feel crushed, depressed — why go on? But then you catch a glimpse of your baby and begin reflecting on what it means to be connected in such an intimate way to a whole new life that you get to nurture, shape, learn from. You reflect on how silly it is to feel down over some minor setback, and you once again feel intensely happy.

A bizarre implication of dynamic views is that your happiness now seems to be unfitting, since you already went through happiness once before and the fit of the emotion is supposed to be “consumed” by its previous manifestation. The same applies to the rare but real cases of non-resetting emotions — the mother who never stops intensely grieving a lost child, the persistent regret for sending someone under your command to their death. What has gone wrong is that the dynamic view doesn’t allow for the fact that the underlying facts occasioning the emotions (the facts making something regrettable or grief-worthy) are still there, and thus remain, in principle, available for us to react to, though, of course, we usually move on. The objection, then, is that neither the passage of time nor the activation of the emotion does in fact diminish our reasons for feeling a certain way, since the underlying facts that make the emotion merited persist in a way the dynamic view doesn’t address. That is why they remain available to us in the vignette above. The suspicion this engenders is that the dynamic view is just a way of affirming the status quo. Our emotions do in fact usually fade, and the dynamic view simply endorses this by announcing that the relevant reasons go away. But this looks a lot like announcing that our failure to discriminate emotionally between 10 and 10,000 deaths is the result of something about the nature of the relevant reasons, rather than just admitting that our emotional apparatus has inherent limitations.

The dynamic view might be revised in the spirit of Hume, who notes that our violent passions or emotions “may decay into so soft an emotion, as to become, in a manner, imperceptible” (Hume, 2001, p. 181). On the revised dynamic view, many of our emotions have an alarm bell quality initially, which inevitably resets. But as the violent manifestation of the emotion fades, it transitions into a quiet, background state. Violent anger, jealousy, happiness, or grief transitions into a quiescent state that is still available and perhaps marked on special occasions, but remains far less obtrusive, and generally stays out of sight when not prompted. The revised view differs from the original since it doesn’t claim that reasons expire or that the fittingness of the emotion (sans phrase) is conditional on whether the emotion has previously activated. It doesn’t leave us wondering how we can fail to have reason to regret something that remains regrettable. Instead, the revised view says that the particular manifestation or character of the emotion that is appropriate is conditional on past emotional engagement. In particular, the “decay” of our emotions from a state that grabs our attention and strongly motivates us into a quiescent background state is fitting (on the relevant occasions), since this transition reflects both the persistent fact-pattern and the fact that the emotion has successfully operated so as to perform whatever tasks were at hand (e.g., highlight an injustice or reinforce successful behaviour).

The revised view is in some ways closer to the sceptical solution. It acknowledges the spike and reset mechanism (in the form of decay), and that our emotions drastically diminish over time. But it still insists that there are good reasons for our emotional states after resetting, and thus for the gap between the persistent facts occasioning the emotion and the emotions themselves. An initial difficulty with this revision is that, while it acknowledges the persistent facts, it cannot account for renewed emotions in the child/article case. There, the point was that we recognize that strong emotions remain reasonable, as the underlying objects are still available — it’s still wonderful to have a child. But the Humean decay view renders these reactions unreasonable, since the dynamic process of the emotions is supposed to render the later, muted responses fitting. In other words, the dynamic view renders emotional decay rationally required rather than optional, and yet the child/article case shows that decay cannot be mandatory.

Another problem with the revised view can be seen in Figure 3, which is meant to suggest a short emotional spike followed by many years (say) of a quiescent state.

Figure 3. Revised view.

As before, the object O of the emotion persists, perhaps a loss or gain. While the revised view allows for the persistence of emotion, the gap between the intensity of loss and the intensity of emotion is often extreme, and thus the difficulty remains. It remains true that we often fail to have the emotions that fit our situations in respect to their character or intensity. ¹⁴ What befits great losses or gains, or terrible mistakes we’ve made whose significance persists, are correspondingly intense emotions, yet for reasons we’ve discussed, these are unavailable due to reset. Our situations are terrible, or wonderful, but even on the revised view, we remain incapable of fully registering these facts. The revised view suffers from the same fundamental problem as the dynamic view.

It may be replied that we do register these facts — fully. It’s just that the right way of registering them, once we have gone through a trajectory of grief or joy, say, is muted after a while. But consider some point in time after the decay. With the right case, it remains true then that your condition is the worst or best it has ever been in your life, and yet this is barely reflected at the time. You are in the position of saying, “I know that I am, right now, suffering the worst thing in the world, but I feel hardly anything, and that’s perfectly fitting.” There is something absurd about this if we genuinely accept (1), and that the emotions have rational fit at all.

Finally, the revised view raises a suspicion concerning timing. Why is decay after around three months normative for grief? Why not two hours or 10 years? It’s hard to see that there is anything normative to ground an answer to this question that isn’t merely pragmatic. Three months feels vaguely reasonable, but only because that is mostly what we observe. And as the sceptical solution reminds us, something like that period was completely inevitable, since a reset is functionally required. The suspicion is thus that “dynamically fitting decay” is just another way of describing functional reset, with nothing normative behind it.

6. Conclusion

Suppose we take the sceptical solution seriously, since its rivals face significant objections, as I’ve argued. In what sense is the solution actually a solution to anything?

The answer is twofold. On the one hand, the sceptical view explains why we so often fail to respond to our reasons by identifying a plausible naturalistic mechanism for this — resetting. This is a solution in the sense of providing a psychological explanation, just as Hume proposes to explain our tendency to project the past into the future by invoking custom or habit. On the other hand, this proposal may seem threatening. Hume notes that we may worry that scepticism imperils our “common life” by undermining the rational credibility of our inferences, but he assures us that “[n]ature will always maintain her rights,” and that there is thus “no danger” that we depart too far from everyday life (Hume, 1927, p. 41). This is a solution in the sense that it assuages a fear by assuring us that the fear rests on an overintellectualized picture of the role that inference plays in our lives.

Something similar applies in the case of the emotions, where some have worried that the sceptical view carries with it unattractive commitments. The fear is that it demands something absurd, like that widows remain permanently immersed in grief, that parents be perpetually ecstatic, or that we never let go of our long-ago regrets. “[P]ersistent grief would not be a mental disorder but the rational response to a loss,” on this view, which implies an “unrealistic moral psychology,” according to Marušić (2018, p. 10). Ignoring the modulating effects of distance on our emotions — distance in time but also along other dimensions — may seem to put us in an impossible position (Bykvist, 2009, p. 16). On one interpretation, the problem here is one of demandingness, since the sceptical view asks us to behave in ways inconsistent with our capacities or at least a good life. Talk of “unrealistic” points this way. But I think the deeper worry is that it isn’t reasonable to respond to persistent gains or losses with persistent emotions reflecting those facts. If anything, doing so seems pathological. The sceptical view may seem to get things backwards. People wringing their hands with regret over the distant past are in need of therapy, not commendation for their reasons-responsiveness. The same is often said of sceptics like Hume: “Refusing to eat your bread because you refuse to project forward from past experience doesn’t make you more rational. Starving epistemologists do not commend themselves to us as models of reasonable behaviour!”

But to the extent that our fear revolves around practical considerations and the reasons we have to promote happiness, the sceptical view can assuage our fears by allowing that these practical considerations (including demandingness) may often override reasons of fit. There’s nothing in the sceptical view that rules out sending people to therapy for fitting emotions, if we so wish. If this is what we are worried about, our fears do rest on a confusion. ¹⁵ We’re even welcome, if we like, to pathologize those who depart too far from patterns of ordinary life or custom as weird on practical grounds. (The DSM labels grief that exceeds “expected social, cultural or religious norms” a “grief disorder” that requires treatment (American Psychiatric Association, 2021).) This may seem inconsistent with (1) and my earlier dismissal of Jamesian views, since I am acknowledging we can have reasons besides those of fit for having or lacking an emotion. But (1) doesn’t say that we cannot have practical reasons for or against emotions; it just says that we do have reasons of merit or fit. For example, panic may well be fitting when going into mortal combat, but it would still be rational to take a pill that blocks the emotion so as to fight more effectively. It can be rational to induce unfitting emotions, just as it can be rational to persuade yourself that 2+2=5 to avoid torture. Fit, as a dimension of rational assessment, can come apart from what we have most practical reason to do, just as it comes apart from natural facts, like the need for reset. We see this in the case of the abused wife who has practical reason to sustain fitting anger that would otherwise reset too soon. There are functional reasons for her emotions to reset as a general matter (which we would explain in causal-scientific terms); there are fittingness reasons to be angry; and then there are the practical reasons. On the sceptical view, these will not always align, since the reasons involved are different, and sometimes our intuitive sense of what “makes sense” will be captured by practical considerations rather than ones of fit. In any case, we need not worry that the sceptical view demands an unrealistic psychology.

It is true that the sceptical view implies that we fail to have the appropriate emotions much of the time. But what of it? Our beliefs are frequently contradictory or fail to reflect our evidence in ways it would be “unrealistic” to expect us to correct; and our emotions frequently make no sense as it is — when we succumb to baseless fears and jealousies, or feel a surge in happiness when ego-defensive mechanisms activate. The view put forward here is just a stronger, more general version of what we already must admit of the human condition.

Opinions will vary as to how plausible or edifying these various forms of scepticism are. But Hume’s “solution” can be read as creating the space to accept a far-reaching critique of reason and its pretensions, its limits, by assuaging his critics’ fears. ¹⁶ The sceptical view of the emotions may, more modestly, remind us that we are apt to forget the extent to which we are emotionally blind to our situation since our emotions are subject to complex requirements imposed by our psychology. It also means that we should reflect more on the role that the emotions play in practical reasoning. We should be less fearful of outcomes whose badness consists partly in negative emotions that will reset, and vice versa. Scepticism, in this sense, can contribute to self-understanding.