On the rationality of emotion regulation

ABSTRACT

Much of the recent work in psychology (and affective science) has shown that humans regulate their emotions nearly constantly, sometimes well and sometimes poorly. I argue that properly regulating one’s emotions displays emotional rationality, and failing to do so displays emotional irrationality. If an agent feels an emotion that is obviously problematic for the agent to feel and she is aware that it is problematic, then the agent ought to regulate her emotions in future similar situations. To capture this aspect of emotional rationality, I introduce the concept of imprudence, which is meant to capture a familiar way that we assess each other’s emotions, despite the fact that it has yet to be a factor in the literature on emotions in philosophy, psychology, or affective science.

KEYWORDS:

• Emotion
• affect
• rationality
• emotion regulation
• imprudence
• emotion assessment

1. Emotional rationality

While emotions have a significant effect on our daily lives, they are deeply complex and notoriously difficult to understand. We frequently criticize other’s emotions. Even if we do not say anything, when someone’s emotion gets it wrong, we make note of it. While at times it seems obvious when an emotion is wrong (e.g., roughly, one’s fear is incorrect when one is not in danger), it is certainly less obvious when an agent counts as irrational with respect to her emotions. For example, we would not criticize an agent for feeling fear when first encountering a very realistic-looking rubber snake, but we would criticize her if she feels fear everyday as she walks by the same rubber snake in the same place. If one day the rubber snake is replaced in the same location with a nearly identical, but real and very dangerous snake, and the agent feels fear, her fear happens to be correct that day, but only accidentally.¹ Being rational is typically used as an assessment of an agent for something for which the agent is praiseworthy, but an agent is not praiseworthy for feeling an emotion that is merely accidentally correct.

I sketch a picture of emotional rationality based on exercising one’s emotional capacities well, and then provide an account of a way in which we assess agents as rational with respect to their emotions. We may sometimes say that an emotion is rational (or irrational), or a pattern of emotions is rational (or irrational), or an agent is rational (or irrational). Strictly speaking, according to my account, it is the agent who is irrational for failing to exercise well her emotional capacities in her practical endeavors. In other words, for the view I defend, the agent is the object of evaluation with respect to rationality assessments of emotions.² While on this account of emotional rationality there are many ways in which an agent can be rational with respect to their emotions, I focus on only one particular rationality assessment, which is related to emotion regulation.

One important way one could exercise one’s emotional capacities is through the regulation of one’s emotions. Emotion regulation is not currently a hot topic in philosophy, although it has been discussed.³ Much of the recent work in psychology (and in affective science) has shown that humans regulate their emotions nearly constantly, sometimes well and sometimes poorly. I argue that properly regulating one’s emotions is a way in which to display emotional rationality and failing to do so is a way in which to display emotional irrationality; here, I am interested in the latter, where one fails to properly regulate one’s emotions in a specific way. There is no established term we use to label an agent for failures of emotion regulation, so I introduce a new term: imprudence. I defend the idea that imprudence is a distinct assessment of an agent for feeling a particular pattern of emotions – imprudent emotion patterns have been neither adequately distinguished nor adequately appreciated in the emotion literature in philosophy or in psychology. Roughly, when an agent feels a pattern of emotions that is evidently harmful for her, and she does not regulate them to reduce or eliminate the harm, she is imprudent.

Thus, imprudence is a rationality assessment. On this account of rationality, I assume throughout that rational agents display excellence in some distinctive way, whether it is in belief formation, action, deliberation, or something else.⁴ With respect to one’s emotions, one can excel in the exercising of one’s emotional capacities. Individual states and events like beliefs, actions, and emotions are rational insofar as having them contributes to improving the agent’s rationality. Moreover, there are many ways in which an agent can be irrational (or rational) – imprudence is merely one.

In what follows, I first introduce a story of emotional deficiency in §2 – it involves an agent experiencing a pattern of emotions that are problematic but failing to properly regulate those emotions, despite knowing that they are problematic. I consider other emotion assessments and show how they cannot properly account for the difficulty that the agent faces. In §3, I discuss emotion regulation in connection to this account of emotional rationality. Finally, I introduce an account of imprudence as a kind of emotional irrationality in §4.

2. Nurse Dan

Consider an agent displaying some emotional deficiency. The following story is intended to spark intuitions that something is wrong, but it is only later that I discuss what has gone wrong. For now, I use non-technical terms like ‘problematic’ and ‘wrong’.⁵ We clearly assess one another’s emotions using terms like these, but I do not want initially to assume that there is an exact account of how it is problematic. I will argue that the story involves a particular deficiency: imprudence. Bear in mind that the problem with the character’s emotion is not due to some other disorder – the agent does not have a recognized mental disorder (a pathology of a different sort) such as schizophrenia, depression, or anxiety disorder.

Dan is a highly skilled and hard-working nurse employed in a hospital emergency room. He feels bouts of disgust on a daily basis (e.g., a homeless man removes his socks and shoes revealing an intensely smelly and rotting wound between his toes, when cleaning a tracheostomy’s mucous secretions, or when a patient coughs a phlegm-globber out and it lands on his shirt).⁶ At times, Dan’s disgust prevents him from properly caring for his patients (e.g., he cannot stop gagging from the smell or his disgust makes his patients feel uncomfortable) and he gets some poor performance reviews. Dan’s colleagues criticize him about his disgust. If Dan does not learn to regulate his disgust in these sorts of scenarios, then he will not succeed as a nurse (e.g., he will receive poor employment reviews and will eventually be fired from the hospital).

Traditionally, there are several different ways that one could assess Dan’s disgust.⁷ Even if the terminology differs, emotion theorists commonly assess emotions along three dimensions: fit, warrant, or benefit. I call this collection of assessments the traditional assessments. First, an agent’s emotion is fitting in a certain situation if and only if the emotion corresponds to the relevant features of the agent’s situation.⁸ A fitting emotion is often seen as roughly analogous to a true belief – the emotion has gotten the situation right. For example, I feel fear when, alone on a hike, a hungry mountain lion is stalking me. Because the mountain lion poses a genuine threat to my well-being, my fear is fitting in this situation. Second, an agent’s emotion is warranted in a certain situation if and only if the agent has sufficient evidence for the fittingness of the emotion in the situation.⁹ When I see that I am alone on a path with a mountain lion, I have good reason to feel fear; I have good evidence that I am in imminent danger. My fear is thus warranted.¹⁰ Third, an agent’s emotion is beneficial in a certain situation if and only if the emotion contributes to the agent’s well-being.¹¹ Imagine that my fear causes me to tremble, triggering a pot to slip from my grip onto a rock; the loud noise startles the mountain lion causing it to flee. By causing the lion to flee, any harm to me is prevented and thus, my well-being is protected. In this scenario, then, in addition to being both fitting and warranted, my fear is beneficial.

There is an important distinction to note between beneficial, unbeneficial, and harmful. Just as an emotion that fails to be fitting is unfitting, an emotion that fails to beneficial is unbeneficial. However, there is a significant difference between an emotion episode that is unbeneficial and one that is harmful. An emotion that fails to be beneficial could be merely neutral – it is not harmful to the agent. Likewise, there is a difference between an emotion episode that is beneficial and one that is not harmful. Indeed, an emotion that is not harmful may not actually be beneficial – it might also be merely neutral for the agent. For the purposes of the traditional assessments, theorists focus on the benefit (or lack thereof) an emotion. Undoubtably, when an emotion is harmful, it is also unbeneficial; though, the latter will not always be the case. The distinction is not always relevant to traditional assessments, but is important for the account of imprudence below.

Nurse Dan’s story does not focus on a particular instance of an emotional state. Rather, it focuses on a pattern of emotions (repeated bouts of disgust felt in similar situations). However, assessments of fit, warrant, and benefit concern a single token of an emotion. Thus, to apply these traditional assessments, we need to focus on a single emotion token. Note that the stories that follow contain fairly unpleasant details in order to bring salience to Dan’s disgust.¹²

First, use ‘disgust_1ʹ to refer to the disgust that Nurse Dan feels on the following occasion. On this occasion, Dan must clean a rotting wound, with an extremely foul and pungent smell, in between a homeless man’s toes. As Dan attends to the wound, he feels disgust₁. Dan’s disgust₁, on this occasion, is both fitting and warranted. It is fitting because Dan’s disgust₁ accurately corresponds to the relevant features of his situation. His disgust₁ is warranted because Dan has sufficient evidence for the fittingness of his disgust₁ in this situation. In order to finish attending to the wound, Dan must periodically take a moment to pause (perhaps to take a breath and calm his stomach). The man is terribly embarrassed by Dan’s disgust₁, and later complains to the hospital. Thus, while Dan’s disgust₁ is both fitting and warranted, it is not beneficial (harmful, even).

One might think that an assessment of benefit can explain what is problematic with respect to agents like Nurse Dan. However, this sort of case can be easily altered to reveal a difficulty. Let ‘disgust_2ʹ refer to the disgust that Nurse Dan feels on a different occasion. Dan assists a doctor with a 94-year-old female patient in the removal of a pessary (a plastic device placed in a vagina to hold a pelvic organ prolapse in place) that was adherent to her vagina because it was left in too long. When the doctor breaks suction of the pessary, a purulent foul-smelling fluid comes out. Both the doctor and Dan think that this could be the worst smell they have ever encountered. Dan’s disgust₂ in this scenario is both fitting and warranted. Dan gags a bit and nearly vomits. While the doctor is more successful at overcoming her own disgust, she does sympathize with Dan due to the extreme situation. So, when the doctor is done with the procedure, she requests a different nurse to finish with the patient. Thus, Dan’s disgust₂ in this scenario is beneficial because he is excused from cleaning up after the procedure.

Lastly, let ‘disgust_3ʹ refer to the disgust that Nurse Dan feels in the following occasion when Dan believes that he must clean a patient’s tracheostomy. Upon entering the patient’s room, Dan already feels disgust₃ because he is thinking about a tracheostomy. However, this patient does not have a tracheostomy. In this scenario, Dan’s disgust₃ is unwarranted because he did not actually have good evidence that this scenario was going to be worthy of disgust. His disgust₃ is unfitting as there does not happen to be anything disgusting in the patient’s room. The patient notices Dan’s disgust₃ and registers a complaint of unprofessionalism with the hospital. Thus, Dan’s disgust₃ is also unbeneficial (in addition to being also harmful).

To recap with respect to the traditional assessments: Dan’s disgust₁ is fitting, warranted, and unbeneficial; Dan’s disgust₂ is fitting, warranted, and beneficial; and, Dan’s disgust₃ is unfitting, unwarranted, and unbeneficial. These results are summarized below.

Table

	Fit	Warrant	Benefit
Disgust1	✓	✓	✗
Disgust2	✓	✓	✓
Disgust3	✗	✗	✗

Nurse Dan, recall, experiences many bouts of disgust in the workplace – far more than the three discussed above. But, across these three scenarios, there are instances of disgust that are fitting and those that are unfitting, warranted and unwarranted, and beneficial and unbeneficial. The traditional assessments do not reveal a single common thread within these situations despite their similarity.

One might then think that a diachronic assessment of benefit of Dan’s pattern of emotions might sufficiently explain what is problematic – perhaps what is needed is to understand the traditional assessment of benefit such that it uses roughly the same evaluative standard but has patterns of emotions as its object.¹³ When we look at the pattern of Dan’s tokens of disgust, it seems that this pattern is not good for Dan – it puts him at risk of losing his job. That is, a particular token of Dan’s disgust might be beneficial, but the pattern of disgust in these types of situations is not beneficial for Dan. So, his pattern of disgust is diachronically unbeneficial; indeed, his pattern of disgust is even diachronically harmful. An agent’s pattern of emotions is diachronically beneficial if the pattern of emotions contributes to the agent’s well-being.

In response to this objection, I argue that an assessment of diachronic benefit does not get at the heart of what is objectionable about Nurse Dan’s situation. The following new story focuses on an agent’s emotions that are diachronically harmful (they fail to be diachronically beneficial) without being irrational.

An older gentleman, Archie, lives alone in a neighborhood where the houses are quite close to one another. While Archie is not very handy, he is constantly trying to fix and build things around his house. On a day like many others, Archie is building a bookshelf. He gets angry when the wood he cuts does not fit together properly. Here, Archie’s anger is both fitting and warranted – he made a mistake when cutting the wood. Archie yells and curses as he tries to figure it out how to fix his mistake. A neighbor walks up to Archie’s front door to warn him of a very large tree limb that broke overnight and is balancing dangerously over his front porch, but she hears Archie’s cursing and hollering and is too nervous to interrupt. The following night in a windstorm, the branch tears down the roof over Archie’s front porch. If he had not been angry in that situation, the neighbor would have stopped by and warned him about the broken limb. He could have cut the limb down and saved his porch roof from harm. Archie’s anger is synchronically unbeneficial (and harmful) in this scenario, despite it being both synchronically fitting and synchronically warranted.

Imagine further that Archie frequently does projects around the house, frequently messes them up, and subsequently he habitually hollers out curses. And, as it so happens, neighbors try to offer to shovel his snow, bring him pies, tell him he left the lights on in his car, and so on. While they sometimes catch him in a good moment and succeed at offering their services, they do not stop by whenever Archie is angry and they hear him loudly cursing (which is fairly often). No one dares to tell him. So, Archie’s anger in these types of situations is also diachronically harmful; and yet, contra the objection, he is not irrational.

The reason that Archie is not irrational in this situation is that he is not in a position to know that his anger is harmful. It is important to understand that Archie could not have known that his anger in the privacy of his own home is problematic for him – despite the fact that it certainly does fit into a pattern that is undeniably diachronically harmful. In other words, we should not hold Archie accountable for the harmfulness of his anger in these situations. The pattern of Archie’s anger clearly harms him, but he is not at fault for this harm. I take this to be an important difference between Archie’s situation and the situations of Nurse Dan. Dan possesses evidence that his disgust is bad for him (e.g., patient feedback, performance reviews, comments from colleagues).

Assessments of emotions, as they are traditionally discussed, are static, not diachronic. While static assessments, those that concern only single tokens of emotions, are valuable, a static assessment of benefit (or of harm) does not explain the important features of Nurse Dan’s situations. Moreover, even diachronic benefit (or harm) is not sufficient to explain the problem that is present in Nurse Dan’s story. When one assesses an agent’s emotion as being harmful, even diachronically, the agent is not necessarily at fault or blameworthy. We often will not hold the agent accountable for feeling a harmful emotion – the agent might not have been in a good position to know that her emotion was problematic for her (e.g., when an agent does not possess evidence that a pattern of emotions is harmful). The important feature is not merely that an emotion is harmful and one knows it, but also that one fails to exercise one’s emotional capacities well (in a case where one is able to do so). The feature I am highlighting about Nurse Dan is not that his pattern of emotions is harmful to him (although, it is harmful to him), but rather that he is irrational for not changing this harmful pattern.

3. Emotion regulation

People often question how much control one actually has over the revision of one’s own emotions. Sometimes one is envious of a friend’s success, but wishes he did not feel envious. Or, one is angry with one’s partner, but wishes she was no longer angry. Sometimes it seems like our emotions happen to us and that we have no control over them. So, one might object that we cannot change our emotions merely because we want to change them. One might argue that it is not obvious that an agent, like Nurse Dan, can regulate his problematic pattern of emotions even if he has relevant evidential feedback that it is problematic.

The objection misfires. We certainly can do all sorts of things to change our emotions. For example, if I am angry and I do not want to be, I might take steps to get rid of my anger like take deep breaths or think about other, more positive things. Similarly, if Nurse Dan knows that he must attend to a patient with a particularly gruesome condition, he also might take deep breaths or think about other, more positive things before entering the patient’s room. These are methods of regulating one’s emotions. There are all sorts of things that we can do to change our emotions when we so desire. I freely acknowledge that this sort of command over emotions is not possible in every situation; clearly it is not. But, it does preliminarily answer the objection and puts the burden on its proponent to specify why, in the situations I have described, the agent does not have this sort of command. In what follows, consider the various ways in which an agent might have this sort of command.

A handful of philosophical theories focus on emotion regulation. For example, Aaron Ben-Ze’ev writes,

We have many strategies for regulating emotions: averting our gaze from a frightening sight, whistling a cheerful tune when afraid, concentrating on happy thoughts when we are sad, removing ourselves from a person who makes us nervous, reducing our attachment to other people, distracting ourselves, and so on. … In light of such variety, there is no doubt that we can regulate our emotions to a certain extent; the remaining issue of interest is how we manage to do it.¹⁴

Ben-Ze’ev’s examples illustrate the fact that emotion regulation is commonplace. He claims that, despite some types of emotion regulation appearing quite obvious, more must be added to establish properly what emotion regulation is and how do we do it. Emotion regulation may concern whether one has an emotion or not, when one has an emotion, how strong or weak an emotion is, how long an emotion lasts, and how one expresses an emotion. Emotion regulation can occur automatically and unconsciously or controlled and consciously.

Psychologists James J. Gross and Ross A. Thompson describe several kinds of emotion regulation: (i) situation selection, (ii) situation modification, (iii) attention deployment, (iv) cognitive change, and (v) response modulation.¹⁵ First, situation selection occurs when one determines which situations one should and should not be in such that one seeks out some situations and avoids others in order to feel or prevent feeling an emotion, e.g., “avoiding an offensive co-worker, renting a funny movie after a bad day, or seeking out a friend with whom we can have a good cry.”¹⁶ This is a forward-looking method of emotion regulation in that the agent takes action before the emotion is elicited.

Second, situation modification is when one changes the features of one’s situation in an attempt to control which emotion it elicits, e.g., “When conservative in-laws visit, situation modification may take the form of hiding politically incendiary art work [to avoid upset in others, but for present purposes, in oneself].”¹⁷ This method of emotion regulation involves acting on one’s external environment.

Third, attention deployment is when one directs one’s attention in a particular situation in order to elicit an emotion. There are four distinct methods of changing one’s attention. One can use distraction where one changes one’s focus to a different aspect of a situation. One can use concentration to emphasize a particular aspect of the situation. One can ruminate on an event – one attends repetitively to one’s feelings and their consequence. Or, one can withdraw attention, e.g., when one covers one’s ears and repeats, “I can’t hear you!” or simply puts on headphones.

Fourth, there are also several different kinds of cognitive change, where an agent changes how she appraises the situation she is in such that it adjusts the emotional significance of it. One can regulate one’s emotions through cognitive change by doing things like down-grading (e.g., telling one-self that it could be worse) or by re-appraising the emotional impact of the situation (e.g., rather than thinking of the weather as partly cloudy, thinking of it as partly sunny).

And, lastly, one regulates one’s emotions through response modulation by doing things like decreasing or changing the expression of one’s emotion (e.g., biting one’s tongue). Each of these five methods of emotion regulation can, at times, be carried out either involuntarily and unconsciously or voluntarily and consciously. While Gross and Thompson’s account is extensive, it is not necessarily exhaustive – there are likely other types of emotion regulation that do not fit cleanly within the categories that they list.

Although Gross and Thompson’s model dominates the literature,¹⁸ there are alternatives. For example, Philip Zelazo and William Cunningham introduce a theory of active or conscious emotion regulation that is associated with executive function.¹⁹ In Gross and Thompson’s model, emotion episodes consist of various automatic processes and emotion regulation takes place in the gaps between them. For Zelazo and Cunningham, by contrast, emotion regulation occurs more or less continuously perhaps even during the emotion episode itself. Executive function is essentially conscious regulation of emotion (or behavior) – it consists in a diverse set of higher cognitive processes, which are the central cognitive system responsible for conscious and deliberate control over other cognitive systems and processes; it contrasts with largely automatic or subpersonal processes. Zelazo and Cunningham provide a characterization of executive function at three different levels (the computational, the information processing, and the neurological) and they distinguish between motivational (or hot) aspects and non-motivational (or cold) aspects of executive function on the basis of neurological data. Using these tools, they describe how executive function regulates emotions at these three levels of abstraction. Moreover, the link between emotion regulation and iterative reprocessing permits their account of emotion regulation to fit well with Cunningham’s iterative reprocessing account of emotions in general. One often does these sorts of conscious and deliberate self-regulation in novel situations where it is difficult to act automatically. According to Zelazo and Cunningham one often employs these higher cognitive processes for deliberate self-regulation of one’s emotions.

These models of emotion regulation and emotional control draw attention to types of emotional capacities, but they leave something important out. We can and do assess one another for failures to exercise these emotional capacities well. In the following section, I discuss why this sort of failure is a matter of emotional rationality.

4. Imprudence

So far we have seen that the traditional assessments fail to explain what is problematic about Nurse Dan. What is distinctive of the Nurse Dan story is his failure to regulate his emotions properly, not the mere fact that he experiences a harmful pattern of emotions. The psychology literature on emotion regulation is clear – we have considerable control over our emotions and we regulate them almost continuously. Now, I will explain what is blameworthy about Nurse Dan’s failure of rationality. I will say that the way in which Nurse Dan’s emotions have “gone wrong,” which the traditional assessments cannot capture, is that they form a pattern that is imprudent. A key feature of imprudence is that Nurse Dan has failed to exercise well his emotional capacities in his practical endeavors, in particular, his emotion regulation capacities. Further, Nurse Dan’s failure is a rational failing, so imprudence is a kind of irrationality. In this section, I present the account of imprudence as a rationality assessment of emotions.

There is a trend in analytic philosophy of taking rationality to be a kind of virtue, which emphasizes doing certain things well.²⁰ In the case of emotions, the following is a plausible starting point: emotional rationality is excellence in exercising one’s emotional capacities in one’s practical endeavors. What are emotional capacities and what is it to exercise one’s emotional capacities well? I take emotions to have (or at least be sufficiently related to) several components: facial, bodily, and linguistic behaviors, physiological processes in body and brain, action tendencies, and appraisals of situations. Some examples of emotional capacities are²¹: the capacities to (i) experience these components, (ii) experience them coherently (i.e., the right kinds of components go together – e.g., the appraisal of danger goes with the facial expression of fear, the physiological response from the sympathetic nervous system, and the disposition to freeze or flee), (iii) discriminate between emotional states, (iv) regulate the development of individual emotional episodes, (v) regulate patterns of emotions, (vi) perceive emotions in oneself and others, and (vii) integrate emotions with cognitive capacities. To be emotionally rational is to excel in the exercising of one’s emotional capacities in a particular way. Because there are multiple emotional capacities, there are multiple ways in which an agent can count as emotionally rational. Here, I am particularly interested in the regulation of emotions because it is an important way in which an agent may count as emotionally rational.

Because of the diversity of emotional phenomena, it is unlikely that there is a short, comprehensive list of ways in which one can display excellence in exercising emotional capacities. If so, then, on this account of emotional rationality, there is more than one way in which to assess the rationality of an agent with respect to the agent’s emotions. Similarly, there are many ways in which someone can fail to display excellence in exercising emotional capacities – multiple ways to assess the irrationality of an agent with respect to the agent’s emotions. In what follows, I focus on a single assessment of emotional irrationality: imprudence.

A central excellence associated with emotional rationality concerns how we use information about the past to improve outcomes for the future. Two kinds of information about the past are especially important – harms and benefits. Agents should eliminate or diminish emotions when they have evidence to believe that feeling the emotions will bring harm that is not desirable for some other reason. Moreover, they should promote emotions they have evidence to believe will bring a benefit that is desirable for some other reason.

We are now in a position to define imprudence:

An agent A is imprudent with respect to a pattern of A’s emotion tokens of the same emotion type felt in similar situations if and only if A fails to take steps to regulate her emotion tokens properly in light of actual relevant feedback providing evidence that the pattern is undesirably harmful because it is contrary to A’s practical endeavours (and this failure is not due to some independently diagnosable psychological malady that renders her unable regulate her emotions in this way).

Imprudence occurs when an agent’s emotions are getting in the way of her relevant interests or harming her well-being, she has information about this fact, but she still fails to take steps to regulate her emotions properly. Recall that sometimes emotion regulation occurs consciously and sometimes unconsciously; sometimes emotion regulation occurs automatically in response to evidence, while other times emotion regulation requires a more concerted effort in response to evidence.

An important aspect of imprudence is that the agent continues feeling the same types of emotions in the same situations despite possessing sufficient evidence that her emotions are problematic, in particular, causing harm. If an agent unconsciously or automatically (in response to evidence) changes the problematic pattern of emotions such that she no longer feels them, she is not imprudent – she no longer feels tokens of the problematic pattern of emotions. An agent is imprudent when, despite being aware of the problematic pattern, she continues to feel tokens of the problematic pattern of emotions and does nothing.

The fact that the assessment of imprudence is a diachronic assessment is important – it concerns how an agent develops over time with respect to a collection of emotion episodes of a particular type.²² By contrast, when a single emotion episode turns out to be problematic, it is unlikely to have a great impact on an agent’s overall rationality. Accordingly, a synchronic rationality assessment is not likely to have a major impact on one’s overall rationality status. The significant impact on an agent’s rationality can be found in patterns of failure or excellence – when an agent frequently and reliably excellently exercises her emotional capacities, this will have a much greater impact on her overall rationality.

When a harmful pattern of emotions occurs and an agent has relevant evidential feedback that it is harmful, the agent should not continue in the same way. Actual evidential feedback ought to regulate (or influence) one’s emotions. The agent’s evidential feedback consists of signs, facts, and arguments that provide her with evidence that the pattern of emotion tokens in question is harmful to her. This actual evidential feedback is readily accessible to the agent (even in cases where an agent changes her emotion automatically or unconsciously). Sometimes just realizing that the pattern is harmful will automatically result in a change in the agent’s pattern of emotions – this may be an example of exercising her emotional capacities well in her practical endeavors. Other times, an agent might need to take conscious and deliberate steps to regulate her harmful pattern of emotions. If the agent merely continues feeling emotions in this pattern despite having evidence that the pattern of emotions is harmful (and does nothing), then the agent is imprudent. There are many different ways in which an agent can regulate her emotions, depending on the particular emotion, situation, and circumstances. Initially, regulation may be something like having the relevant emotion tokens be less intense (and eventually peter out). Not only should feedback from the environment influence the intensity of an emotion felt, but at times it should help to regulate whether one feels the emotion at all. The key is for the agent to respond properly to the relevant feedback by learning and by regulating her emotions – either consciously or unconsciously and either automatically or with more concerted effort.

Imprudence is a kind of irrationality, but it is easy to confuse with a distinct phenomenon. What marks the difference is the agent’s ability (either consciously or unconsciously) to change the relevant problematic emotions. There is widespread agreement that there is a connection between difficulties with regulating emotions and psychopathology.²³ Indeed, many theorists claim that emotion regulation problems are a central feature of (or lead to all, or at least to most) major forms psychopathology.²⁴ Many of the current diagnostic criteria for mental illnesses specifically refer to issues with emotion regulation.²⁵ A person who has a chronic inability to regulate her emotion associated with psychopathology would fall outside of this rationality assessment on this account; of course, this person still may be rational or irrational according to various other assessments.

There will be borderline cases where reasonable people disagree about whether a particular person is exhibiting imprudence; but there are cases that seem more obvious for our concerns. At one extreme, an agent receives feedback showing that she has a problematic pattern of emotions, and she can change her pattern of emotions relatively easily (perhaps she has done so successfully with similar problematic emotion patterns in the past), but she does nothing to change her emotional responses in similar situations. In this case, she is imprudent. On the other extreme, an agent has the evidence that her emotions are problematic, but even with years of intensive therapy and numerous medications she cannot change her problematic emotions at all. She is not imprudent. In order for someone to be imprudent, the agent must actually possess the ability to make these sorts of changes. When assessing an agent for imprudence, one must consider whether the agent’s emotions that are felt in relevantly similar situations actually fit into a pattern that is harmful.²⁶ I do not propose a method of determining whether there is a pattern of this type present. I take it that this is a vague issue – there will be obvious cases where there is a pattern (e.g., multiple occasions where I am angry at my partner for leaving the cap off of the toothpaste), obvious cases where there is no pattern (e.g., a situation where I am angry, a situation where I am happy, a situation where I am disgusted, and so on), and some cases where reasonable people might disagree.

One might object to this theory of imprudence by saying that it mislocates what ought to be assessed for rationality. If an agent must do things in order to change her emotions, then we should not assess the agent’s rationality with respect to her emotional capacities. Rather, the related rationality assessment should be about what the agent is doing to affect the emotional responses. Thus, according to the objection, agents are rational or irrational not with respect to their emotions, but rational or irrational with respect to their actions that impact their emotions.

In response, I defend the idea that an agent can count as rational or irrational with respect to her emotions, not just her actions or beliefs. There are cases in which the rational adjustment of emotional responses through action exhibits rationality, not only in the actions performed, but also in the emotional responsiveness to those actions. Imagine two women feel anger that neither wants to feel. They each try the technique of taking long, deep breaths to calm down and rid themselves of their anger, but only one of them succeeds – the first woman remains just as angry while the second’s anger dissipates. These two women have done the same thing. An assessment of their actions would not reveal a difference between the two. The difference between the two women must be in the excellence in exercising their emotional capacities. Only the second woman exhibits emotional rationality – she has exercised her emotional capacities well. This is not to deny that we can also assess the people for what they did. The two women are assessed in the same way on this matter. What they did probably contributes to their rationality (or lack thereof), but assessing their rationality in terms of choosing and implementing actions does not exhaust the rational assessments that one could make of these two women. The fact that the second woman exercised her emotional capacities well bolsters her rationality.

To be clear, I am not claiming that the only way in which one can exhibit emotional rationality is when one performs some sort of conscious action with respect to one’s emotions and successfully changes how one feels as a result. There are also plenty of cases where one exercises one’s emotional capacities without necessarily performing a conscious action. For example, suppose I feel fear of a snake that I see in the grass. Then, upon a closer look, I see that the object is actually a pretend rubber snake. Just upon seeing that there is no genuine threat, my fear goes away without any conscious action or effort. This too is an example of displaying emotional rationality.

It seems, then, that Nurse Dan is imprudent. He fails to regulate his emotion episodes in light of actual relevant feedback providing evidence that his pattern of disgust (in these sorts of situations) is contrary to his relevant interests. We hold a person accountable who is in a position like Dan’s – when a person has evidential feedback that his emotional pattern is bad for him and he does nothing about it, then he is doing something wrong. It seems that Dan is in a good position to see that his emotion tokens are bad for him (i.e., he has access to the relevant feedback) and to learn from this information. In other words, it is reasonable to hold him accountable for feeling this problematic pattern of emotions rather than learning from his situation and changing the pattern of emotions.

We do (and should) criticize people for these types of mistakes in their emotional lives. Why do we (and should we) assess one another in this way? When a person’s emotions fit into a problematic pattern, it is harmful for that person. When a person has already received actual relevant feedback providing evidence that this pattern is harmful, then the person should respond and change. We ought to learn from our mistakes. An important part of an answer to this question is this: the agent can change in light of this feedback. How? The way in which an agent changes or regulates her emotions in light of this sort of feedback or assessment is less direct than the sort of control one has in wiggling their left thumb, but more direct, perhaps, than the sort of control one engages in when one sees a therapist to overcome a fear.²⁷ Recall, though, that emotion regulation sometimes occurs consciously and sometimes occurs unconsciously. What is most crucial is that it occurs at all.

Imagine, instead, that Dan does not act in an emotionally irrational way. What would that look like? There are many suggestions from the emotion regulation literature. Rather than ignore the feedback that he has received that is evidence that his emotions fit into a pattern that is harmful for him, he does something to fix the situation – e.g., regulates his emotions. Dan would modify his emotions in light of this evidence. The change might not be immediate, but the change is not likely so difficult that it requires involving the help of a professional. Recall that in this story Dan does not have a mental disorder. Thus, he should be held accountable for these types of mistakes. He feels emotions that fit into a problematic pattern and he has evidence that this is so. In light of this evidence, consider the sorts of steps he might take to correct his situations.

We imagine that Dan acknowledges the feedback that he has received – his disgust of his various patients is jeopardizing his career. When one feels disgust of something, one has an aversion to that thing. Avoidance of situations is an important strategy for emotion regulation,²⁸ but this is not an option for Dan – he wants a career in nursing and he is a good nurse. Emotion regulation theorists also discuss the idea of simply regulating the expression of one’s emotion. This might work in some scenarios, but because of the intimate situations that a nurse and patient are often in, it is unlikely that Dan could successfully merely suppress the expression of his disgust to the extent that the patients could no longer detect Dan’s disgust. It is also important to recognize that merely regulating the expression of one’s emotion can sometimes have the opposite effect by actually intensifying the emotion.²⁹ Dan needs to regulate the actual disgust that he feels. Dan has many options for regulating his emotional response toward his patients.

From Gross and Thompson (2007), suggestions like attention deployment or cognitive change seem relevant to Dan’s situation. For example, Dan could do a type of attention deployment where he changes his attention (by attending to or ignoring certain features of the situation) whenever he encounters a particularly gruesome case with a patient. When Dan faces a patient with an especially foul condition, Dan could try to focus his attention on perfecting his craft of being a nurse.³⁰ Or, whenever Dan experiences a particularly gruesome case, Dan could consciously train himself to attend to the time he won the lottery to make himself smile.³¹ Dan could instead engage in cognitive change by doing something like down-grading such that he could think, “Well, it could be worse,” and then imagining the worst possible scenario with a patient.³² While the situations that nurses face do not get less disgusting, the disgust they feel tends to lessen in frequency and intensity as they learn to regulate it.³³

When Dan tries to regulate his emotion, he could work on overcoming this aversion tendency through cognitive reappraisal – changing how he sees the situation.³⁴ He could try to make himself just be okay with the gruesome or grotesque patient cases. We can imagine Dan prepping himself before walking into a patient’s room: this sort of thing happens to people’s bodies; the patient needs assistance, and I am trained to assist this patient; a patient will heal better and more quickly if they feel comfortable around me; I want my patients to feel comfortable with me, etc. Perhaps Dan might replay the scenarios that caused him to feel the strongest bouts of disgust. This sort of prepping could be Dan’s attempt to convince himself that his disgust is unfitting. Although fittingness considerations do not motivate Dan to change his disgust (he is not motivated to change his emotion because he is so concerned with his emotions getting it right), Dan appeals to fittingness considerations to help change his emotion.

Consider an application of an example from Zelazo and Cunningham where they discuss a case with down-regulating an emotion as the primary goal.³⁵ Imagine that Dan wants, at the very least, to make his disgust far less intense. First, Dan would consider his problem (he is feeling an intense disgust and he wants to feel very little disgust), then his various options (e.g., reappraise the situation, distract himself, etc.). Next, Dan ideally selects a plan that seems most promising. According to the view put forward by Zelazo and Cunningham, Dan’s initial emotional response to the situation is quick and to some extent automatic, but his executive function soon reprocesses the affective state to include all sorts of information. It is at this stage that Dan’s executive function can work to decrease the intensity of his disgust by reflecting on the initial response and reprocessing it according to one or more of these strategies. Over time, this new response can become more automatic.

In the original telling of Dan’s story, he has evidential feedback that his emotions fit into a problematic pattern, but he does nothing to correct it. When an agent does nothing with this feedback (does not learn) and simply goes on feeling the emotions that are indeed problematic, then we criticize the agent (for being imprudent). An agent can change the problematic pattern by employing any of these (or other) strategies for emotion regulation.

5. Conclusion

There is something in particular that is wrong with Nurse Dan’s pattern of disgust. He is imprudent. He has evidence that his disgust episodes form a problematic pattern. Dan is making emotional mistakes, and we ought to criticize him for these types of mistakes. This kind of mistake detracts from Dan’s emotional rationality.

Here I have captured only one particular way in which an emotion may count as irrational. This type of assessment has not been given the attention it deserves in philosophical discussions on emotion assessments or emotional rationality. It is difficult for traditional theories of emotions, which focus on emotions as they occur in a single moment, to explain this type of assessment of emotional rationality. This type of assessment has not been given the attention it deserves in psychological accounts of affect and emotion regulation. It is in light of the emotion regulation research in affective science that this type of assessment of emotions is so interesting. Agents are capable of regulating their emotions in all sorts of ways, and emotion regulation occurs almost continuously throughout one’s life. If an agent feels an emotion that is obviously problematic for the agent to feel and she is aware that it is problematic, then the agent ought to regulate her emotions in future similar situations. This ‘ought’ is supported by the degree to which we interact with and regulate our emotions. It is not just something we can do – it is something we are already doing.

Because the success or failure of a single emotion episode rarely has a significant impact on an agent’s overall rationality (i.e., it neither significantly bolsters nor significantly detracts from the agent’s overall rationality), diachronic assessments are of particular interest. Here, I have focused on the assessment of imprudence. An agent is imprudent when her emotions get in the way of her interests or detract from her well-being, she has information about this fact, and she fails to take steps to regulate her emotions properly. The central feature that the assessment of imprudence tries to capture is how it appears particularly objectionable when an agent is completely aware that something is bad for her and changing it is within her control, but she continues in the same way.

Acknowledgments

I am grateful to the audiences at EPSSE in Lisbon, the Mentoring Project workshop, the Philosophical Society at St Andrews, and the members of the Philosophy Department for the Reflectorium at the University of St Andrews. For helpful discussions on this project, thanks to Michael Brady, Brad Cokelet, Ley Cray, Morgan Currie, Will Kerr, Lindsey Mason, Stewart Shapiro, and the lovely members of the Awesome Writing Group at St Andrews. Thank you to Justin D’Arms, Clare Batty, Wil Cunningham, Owen King, Richard Samuels, Kevin Scharp, Sigrún Svavarsdóttir, and the anonymous referees provided by Philosophical Psychology, for very valuable comments on earlier versions of this paper.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Alison Duncan Kerr

Alison Duncan Kerr is a Research Fellow in the Philosophy Department at the University of St Andrews. She is the Founding Director of the St Andrews Institute for Gender Studies, and a member of Arché Philosophical Research Centre. Her research focuses on philosophy of mind, emotions, AI, and gender.

Notes

1. This relates to Gettier cases discussed in epistemology. See Gettier (1963) and Goldman (1976).

2. Nonetheless, in order to distinguish problematic emotions from non-problematic emotions (or problematic patterns of emotions from non-problematic patterns of emotions), sometimes it is easier to describe an emotion (or a patterns of emotions) as irrational. I consider this to be shorthand.

3. For example, many Stoics wrote treaties on how to control ones’ emotions, e.g., Seneca, 2010.

4. See Svavarsdóttir (2008); Wedgewood (2017).

5. I also use terms like ‘right’, ‘reasonable’, ‘appropriate’, and ‘understandable’ without any intention of using them as technical terms.

6. Thanks to nurses Daniel Scharp and Michelle Johnson, Dr. Anna Shope, and Dr. Jeanie Pacewic for stories and discussions concerning disgust (and other emotions) and emotion regulation in a work environment. And thanks to Wil Cunningham for conversations on this matter.

7. For an introduction to disgust, see Kelly (2013) and Strohminger and Kumar (2020).

8. This term comes from D’Arms and (2000). D’Arms and Jacobson use ‘objectively rational’ to mean the same thing. See also Kerr (2019). Other theorists endorse this sort of definition but use different terminology, e.g., Greenspan (1988) uses ‘applicable’, (1987) uses either ‘objectively correct’, ‘appropriate’, or ‘rational’, and Gibbard (1990) and Jones (2003) use ‘apt’.

9. This term comes from D’Arms and (2000), though they differ slightly in their definition. D’Arms and Jacobson use ‘subjectively rational’ to mean the same thing. See also Kerr (2019). Other theorists also endorse this sort of definition but use different terminology, e.g., Greenspan (1988) uses ‘appropriate’ and Jones (2003) uses ‘reasonable’.

10. Contrast this with someone merely hiking alone in a large forest where a mountain lion lives, but who has yet to see it. Even though this person is hiking in a forest where a mountain lion lives, the mountain lion does not yet obviously pose a genuine threat to his well-being. The likelihood of a fatal mountain lion attack is very low – there is an average of one death from a mountain lion attack per year in the United States. See Stevenson (2008). Because the possibility of being attacked is so low, this person does not yet possess a genuine reason to feel fear – his fear is not warranted.

11. In D’Arms and Kerr (2009), we use ‘prudential’ and gloss it as concerned with the self-interested advisability of the emotion; this use of ‘prudential’ is analogous to my use of ‘beneficial’ here and throughout the paper. My use of ‘prudential’ and the like is quite different here. For assessments similar to this use of ‘beneficial’, Greenspan (1988) uses ‘adaptive’, Gibbard (1990) uses both ‘advantageous’ and ‘pragmatic’, and Jones (2003) uses ‘strategically wise’.

12. I take this to be a trigger warning of sorts, “to allow those who are sensitive … to prepare themselves … and better manage their reactions,” Manne (2015).

13. See Samuels et al. (2002) for discussion on standards of evaluation.

14. Ben-Ze’ev (2000).

15. Gross and Thompson (2007), Gross (1998).

16. Gross and Thompson (2007).

17. Gross and Thompson (2007).

18. For criticisms of Gross’ theory of emotion regulation, see Barrett et al. (2006).

19. Zelazo and Cunningham (2007).

20. This understanding of emotional rationality draws on Svavarsdóttir’s (2006, 2008) theory of practical rationality; see also Wedgewood (2017).

21. The following are examples of emotional capacities, not features of a single emotional capacity.

22. See Na’aman (2019) for a different account of diachronic emotional rationality where, rather than considering patterns of emotions over time, he considers a single emotion token over a longer period of time.

23. In contrast, see Kring and Werner (2004) for the critique that many of the assumptions in this literature are lacking sufficient empirical support; despite this critique of current research, they still hold that there are important connections between difficulties with respect to emotion regulation and psychopathology.

24. See Koole (2009), Bradley (2000), and Werner and Gross (2009).

25. Consider, though, someone who has bipolar disorder. Many people with bipolar disorder like their manic states, despite the fact that when one is in a manic state one’s emotions are typically not socially appropriate. In light of this latter claim, one would think that a person with bipolar disorder ought to regulate their emotions better (at least those that are not socially appropriate). It would be interesting to see if someone with bipolar disorder, in the midst of mania, is actually able to regulate their emotions but simply does not want to do so. See Kring and Werner (2004, pp. 372–373) for discussion and a call for research on people with bipolar disorder and the connection to emotion regulation. See also Johnson (2005).

26. The qualification of “relevantly similar situations” is important. For example, imagine that a person has an extreme phobia of snakes (such that she flatly refuses to walk in the tall grass, on the off chance that she brushes up against a snake, even if her life depends on it) but a normal fear of lions (such that she does not feel fear when safely observing the lion at the zoo but does feel some fear when in the open plains lion habitat in Africa); her fear in snake-situations is largely unreasonable whereas her fear in lion-situations is largely reasonable. The qualification of “relevantly similar situations” allows us to assess this person’s fear as two separate things – fear-in-snake-type-situations versus fear-in-lion-type-situations.

27. For example, see cognitive behavioral therapy, e.g., Dobson (2002) and rational emotive behavioral therapy, e.g., Ellis and (2005).

28. Zeelenberg and Pieters (2007).

29. Lanzetta et al. (1976); see also Frijda (2007, p. 167) for discussion.

30. Ben-Ze’ev (2000, p. 389) discusses how avoiding looking at something or thinking about it is a very good method in particular for reducing or eliminating disgust.

31. Zelazo and Cunningham (2007) discuss this sort of method of distraction for emotion regulation. Frijda (2007) talks about diversion as a method of emotion regulation.

32. This is similar to what Ben-Ze’ev (2000, p. 231) calls “broadening our perspective” by comparing our misfortune to that of others in order to make our own misfortune seem less grave which results in a reduction of emotional intensity.

33. The way that medical professionals describe these sorts of scenarios (as being undeniably disgusting despite regulating their response to not feel disgust), conflicts with claims from D’Arms and (2005, p. 5): “Similarly, the conviction that a given circumstance merits a given emotional response that we do not tend to give it has some tendency, over time, to dispose us to feel the way we think appropriate.”

34. Some theorists talk about a regulation of experience or rumination but only seem to do so in a negative context such that doing so is problematic (i.e., produces longer lasting and more intense emotions); see Niedenthal et al. (2006). For discussions on the difference between attention deployment and cognitive reappraisal and on cognitive reappraisal’s potential for successful emotion regulation; see Urry (2010) and McRae and Ciesielski (2012).

35. Zelazo and Cunningham (2007: 147).