2. Emotions and reason-giving considerations

Those who think that emotions have a role in rational agency tend to take as their starting point the idea that emotions can track and be responsive to putative reason-giving considerations for action (as examples, see Arpaly Reference Arpaly2002; Jones Reference Jones and Hatzimoysis2003; Döring Reference Döring2007). In this section, I cash out why this is thought to be the case.

Let us focus on paradigmatic emotions, understood as conscious episodic affective experiences of the agent.Footnote ¹ I follow the current trend in philosophy in endorsing the claim that such emotions are weakly cognitive states, but not by being forms of, or constituted by, beliefs or judgements.Footnote ² We need not, however, commit to any one account of emotions here, so long as we have some key features of conscious episodic emotions to hand.

These key features include the following. Emotions are current experiences of the subject. They have a distinctive phenomenology, where part of that distinctive phenomenology is an action readiness or tendency. And emotions are intentional – they are directed at objects and represent those objects in their intentional content as being certain evaluative ways.Footnote ³ For instance, I am afraid of the bear. My fear is directed at some target object, the bear, and represents the bear as dangerous. The intentional content of my fear will be expressible as ‘the dangerous bear’ or ‘the bear as dangerous’. All cases of an emotion, it is claimed, represent their object in the same evaluative way, and emotions can be assessed for accuracy in terms of whether the object really is that evaluative way. For example, all instances of fear represent their objects as dangerous, and my fear is accurate if the bear really is dangerous.

Further, emotions can play an epistemic role in our lives in at least three ways (Brady Reference Brady2013, 9). They can make things salient for us by drawing our attention to certain aspects of our environment. They can inform us about value, such as when we reflect on our emotions and learn, say, that the noise in the house is indicative of a dangerous intruder. And they can inform us about ourselves, such as when we reflect on our emotions and learn, say, that we have implicit racial prejudices underscoring our fear of a certain person.

Their intentional nature, the way emotions are assessable for accuracy and the epistemic role of emotions all support the claim that emotions can track and be responses to ‘real and important reason-giving considerations’ (Jones Reference Jones and Hatzimoysis2003, 181). The relevant reason-giving considerations, what I call the ‘emotional considerations’, are what make up the emotion’s evaluative intentional content.Footnote ⁴ For instance, my fear is responsive to danger. In the example where I am hiking in the mountains, it is responsive to the danger of the bear. The danger of the bear is a reason-giving consideration for me to act to avoid it.Footnote ⁵

Our emotions may sometimes more accurately track reason-giving considerations than our beliefs do, and it seems possible that we could act for genuine reason-giving considerations when acting emotionally. Insights drawing on the intentional nature of emotions, like these insights, motivate the challenge to provide an account of how an agent could act rationally when acting emotionally, by responding to the emotional considerations. In the next section, I explain in more detail the particular version of the challenge that I shall take on, before showing how it can be met in the remainder of the paper.

3. Acting for reasons seen as reasons

Agents act for reasons, where a reason for acting is a consideration that counts in favour of acting.Footnote ⁶ This idea is central to our understanding of what it means to be a rational agent.Footnote ⁷ Reasons, however, do not simply cause a rational agent’s actions in any which way. Rather, to say that an agent acts for reasons is to say that the agent acts for reasons seen as reasons (Jones Reference Jones and Hatzimoysis2003, 189).Footnote ⁸ What this means is that the agent sees the reason as counting in favour of and justifying the action choice, and guides her action because of the reason seen in this way.

Requiring that the agent sees the reason as justifying the action choice is not, here, the requirement that the agent judges that, all-things-considered, the reason is the best reason for which she should act. Rather, as Jones writes, it is the requirement that ‘if one has a reason to do A rather than B, and one has a reason to do both … then if one is to choose A rather than B that must be because [the reason for A] is in this context seen as a better reason than all its competitor reasons’ (Jones Reference Jones and Hatzimoysis2003, n. 11). So, seeing the emotional considerations as justifying an action choice, in the sense relevant here, is a comparative notion requiring that the agent sees the considerations as pro tanto reasons she has in favour of acting, but also that she sees them as better than competing considerations. In that way, she sees them, prima facie at least, as the best reasons, or as the reasons for which she ought to act.

Cashed out like this, the question of whether an agent acts for reasons is a question that relates to her subjective rationality, where subjective rationality is understood as a coherence amongst one’s mental states and actions, and is independent of whether what the agent sees as reasons really are reasons. With this in mind, an account resting solely on the way emotions track reasons, such as an account like Arpaly’s (Reference Arpaly2002) account, is not sufficient as an account of how we can act for reasons when acting emotionally. Even if emotions track genuine reasons, which is in itself an important insight, it does not follow that we see the tracked reasons as reasons, rather than simply passively react to them. And our conception of ourselves as rational agents is not as mere reason-trackers that passively react to reasons, however well-functioning we may view our reason-tracking mechanisms as being (Jones Reference Jones and Hatzimoysis2003; Döring Reference Döring and Goldie2010). The challenge is to show how the agent guides her actions by those considerations that are tracked, seen as reasons.

One step towards meeting this challenge is to examine our conception of rational agency in order to identify what is required of a rational agent, qua rational agent, for her to guide her actions by reasons, when acting emotionally. This is the approach Jones (Reference Jones and Hatzimoysis2003) takes.

Our understanding of a rational agent is of someone who acts for reasons: she guides her actions by reasons seen as reasons. Being able to guide her actions by reasons seen as reasons requires, at the very least, that the rational agent has the ability to critically reflect on her reasons and actions, has the disposition to do so in situations when it is needed and has the disposition to allow the results of those critical reflections to control her behaviour (Jones Reference Jones and Hatzimoysis2003, 23; see also Scanlon Reference Scanlon1998, chap. 1). If the agent is not only able, but is committed, to being a rational agent and to guiding her actions by reasons, then she must also be committed to the ongoing cultivation, monitoring and exercising of the abilities and mechanisms that allow her to have, and to exercise, those capacities and dispositions necessary for responding to reasons. Amongst these abilities and mechanisms are the reason-giving mechanisms, such as emotions. So, as Jones argues, the rational agent must be committed to the ongoing cultivation and monitoring of her emotions, and she must accept that her emotions are generally reliable reason-giving mechanisms. When a rational agent acts emotionally, unlike a non-rational subject, the thought goes, she can act rationally because she has these capacities and dispositions, and she regulates and monitors her emotions. The rational agent, on this approach, can guide her capacities and dispositions by her capacity of reason, and can thereby be seen to be committed to guiding her actions by reasons.

While this is a step towards meeting the challenge because it gives a detailed picture of the kind of being a rational agent is, what she must be committed to doing and how she can monitor and regulate her capacities by reason, it is insufficient as an account of how the rational agent can guide her actions by reasons seen as reasons, when acting emotionally. Examining our conception of rational agency only tells us about the rational agent; it does not tell us what is going on at the level of individual actions such that the agent can be seen to be guiding her action by reasons seen as reasons. While these other requirements are by no means unimportant or irrelevant, Jones was aiming to show how we can guide our action by reasons when acting emotionally, and this she has not done.

Consider a case of akratic action, where the agent acts emotionally against her all-things-considered judgement. In such a case, we can be rational agents who monitor and cultivate our emotions. It can even be the case that acting against our best judgement is more coherent with our beliefs, desires and values than acting in lines with it, and we may be more rational in terms of overall coherence than if we were to act in lines with our judgement. We will, however, be failing to guide an individual action by what we see as reasons in the given situation, and our action itself will not be rationally guided.

To see this, take Jones’ example of a person at a meeting who raises a topic because of her feminist anger, despite having decided that she would not do so. The agent regulates her emotions and accepts them as generally reliable. On reflection, she also realises that her original decision not to raise the topic was a function of cowardice rather than an assessment of the actual merits or demerits of raising the topic. If so, she was not self-monitoring her deliberative reasoning as she should have, else she should have come to a different judgement. On Jones’ account, such akratic action will be rational (Jones Reference Jones and Hatzimoysis2003, 196–97).

However, even if the agent regulates her emotions and would have distrusted her judgement on more critical reflection, it remains the case that such a scenario involves a conflict of putative reasons arising from two different reason-giving mechanisms, emotion and belief (Döring Reference Döring and Goldie2010). With such a conflict, the agent must somehow choose which considerations from which mechanism to act for at the time of acting, and simply regulating her reason-giving mechanisms does not provide a way in which to do so. In fact, resolving the conflict is in part what her judgement does, even if it is ultimately imperfect. Nevertheless, without first resolving the conflict in favour of the emotion, the agent cannot act for the emotional considerations seen as a reason.

We therefore see that, while looking at qualities that the rational agent must have is important, it only takes us so far. It is not sufficient to show that the actions of the rational agent are themselves rationally guided, even though this is what Jones was aiming to do. What we need is an analysis of what is happening on the level of individual actions, such that a rational agent can be seen to guide those actions by what she sees as reasons.

Now, one way in which we could understand an agent as guiding her action by reasons seen as reasons at the level of individual actions is by requiring that the agent forms beliefs about her reasons and how she ought to act, and acts in line with those beliefs. This is how the idea that rational agents act for reasons seen as reasons has traditionally been cashed out. For instance, if the agent is to act for the emotional considerations seen as reasons, she must believe that things are as they appear emotionally, and guide her action by a normative judgement about what to do. This belief may be a non-inferentially justified belief, and it may only be necessary that the agent would authorise the action with a normative judgement if she were to deliberate about what she has reasons to do (Döring Reference Döring and Goldie2010).Footnote ⁹

Traditional understandings of rational agency that centre on belief, however, are increasingly being questioned, such as by challenging the centrality of belief and deliberation in our conception of rational agency (Arpaly and Schroeder Reference Arpaly and Schroeder2012), or by arguing for other kinds of psychological states, like imagination, to play a belief-like role in action (Velleman Reference Velleman1996). Indeed, accepting that emotions could have a role in rational agency in the first place is to challenge the more traditional conceptions of rational agency (Arpaly Reference Arpaly2002; Pacherie Reference Pacherie2002; Jones Reference Jones and Hatzimoysis2003, Reference Jones2006).

Take the example with which I opened this paper: I am afraid of a bear and climb through a gap because of my fear. Actions like these do not necessarily involve believing that the bear is dangerous or forming the intention to so act. Nevertheless, the emotion is causally involved with putative rational action without requiring a belief, and I am responding to the danger of the bear as a reason to climb through the gap (Pacherie Reference Pacherie2002). Cases like these cannot be straightforwardly accommodated by traditional understandings of rational agency, and are exactly why the challenge to account for a role for emotion arises.

Rather than starting from a traditional understanding of rational agency that centres on belief, the challenge for providing an account of how an agent can guide her actions by reasons when acting emotionally requires that we approach the topic bottom-up, so-to-speak. By that, I mean that we should approach the challenge by looking first at the emotional actions to see if there is a way in which the agent can be understood to be guiding her action by a reason seen as a reason, rather than by starting with the claim that a belief is required.

So, as our starting point, let us accept Jones’s picture of the rational agent and accept that our rational agent has the relevant capacities and dispositions required for rational agency, and also that she monitors and cultivates her reason-giving mechanisms, including her emotions. Let us also accept that acting for a reason seen as a reason requires that the agent sees the reason-giving emotional considerations as pro tanto reasons favouring an action choice, and also that she sees them as the reason for which she should act. How can this be the case when the agent is acting emotionally?

In the rest of this paper, I propose a way in which we can understand the agent to be aware of the emotional considerations as pro tanto reasons favouring an action choice and, indeed, as the reason for which she should act. This can be the case without requiring or centring on belief. My argument highlights the fact that the emotional agent is a caring agent. Through being a caring agent, the emotional agent can see the emotional considerations in the right kind of way in order to guide her action.

4. Cares and caring

My proposal turns on the interconnection between emotions and caring, so before turning to how we can guide our actions by reasons when acting emotionally, it is necessary to say more about what the interconnection is. This is the task of the current section.

A platitude about emotions, as Brady writes, is that they ‘constitute reactions to objects, events, and states of affairs that are potentially significant or important to us’. Further, ‘what is potentially significant or important to us is a matter of what we care or are concerned about. Without such concern, it is puzzling how there could be emotional responses in the first place’ (Brady Reference Brady2013, 10–11). In this section, I take this platitude as a starting point to lay the ground for my subsequent argument that the notion of caring can help us see how an agent can guide her action by what she sees as reasons, when acting emotionally.

I begin by elucidating the role of cares in emotion, a role shared by the emotions of both agents and non-agents. As I am interested in actions performed by agents, I next ask whether there is a special agential way of having cares and responding to them that may help distinguish between mere behavioural responses and emotional actions. For this reason, I turn to the philosophically rich notion of caring.

So, let us start with cares and concerns, and the idea that they are crucial for making sense of an emotional response.

Amongst emotion theorists, having cares and concerns is a prerequisite for experiencing an emotion. Appraisal theorists in psychology, for instance, typically express this as the subject’s having a goal, which is appraised as being relevant and affected in a particular situation. Without such cares and concerns, ‘nothing matters to the person’ (Clore Reference Clore, Ekman and Davidson1994, 188). We also find a similar idea amongst philosophers. Roberts (Reference Roberts2003), for instance, argues for an account of emotions as ‘concern-based construals’, and Döring agrees with the general picture, in that we emotively construe a situation ‘in terms of what we care about’ (Döring Reference Döring, Roeser and Todd2014, 132).

Suppose that I am afraid of an upcoming talk. Because I care about my reputation, the talk is significant and I am afraid of it as a potential threat to my reputation. Or, suppose that I am angry at the mechanic for not adjusting my handbrake despite my suggesting that there was a problem. Because I care about being a credible source of information, his not listening to me is significant and offensive, and I am angry.

Further, through having cares, our emotions draw our attention to the object of the emotion as that which is putatively affecting ourselves in good or bad ways.Footnote ¹⁰ In this way, an emotional response draws the subject’s attention to things of putative significance for her, relative to her cares. And because of that significance, the emotional considerations are taken seriously as being the case.

Compare fearing the bear with imagining or wishing that the bear is dangerous. When I encounter a bear while hiking in the Canadian Rockies and experience fear, the emotional considerations of the bear as dangerous appear far more compelling to me as being the case than they do if I were to imagine a dangerous bear or to wish it were dangerous. Part of why the dangerousness of the bear appears more compelling to me is that I take it seriously because it is currently significant to myself and my safety, something I care about. This makes sense because emotion plays a very different role in our lives to imagination or wishing, in part because emotions are grounded in our cares and concerns. For instance, if someone were truly faced with danger and her cares under threat, then she may need to respond quickly without first establishing whether there is a genuine danger.Footnote ¹¹ The fact that the significance is grounded in the subject’s having background cares and concerns helps explain why she takes the situation seriously: she has a vested interest in it.

This is the case even if the agent reflectively rejects the way things appear emotionally.Footnote ¹² Consider an example of a conflicted liberal who believes firmly in the value of all persons, and desires to live her life in accordance with her values. Despite her best intentions, however, she still finds that she has an implicit racism. On encountering an individual from another racial group, she experiences fear and is faced with the emotional considerations of the person as dangerous. Nevertheless, she reflectively rejects that the person is dangerous and does not take her emotional considerations seriously as being the case. But this does not undermine my claim, which is simply that the subject experiences things as being that way when she experiences an emotion, without necessarily endorsing the way she experiences things as being.Footnote ¹³ Indeed, the subject takes the emotional considerations seriously as being the case because of their significance, not because of their truth, and she need not endorse the emotional considerations for her to experience things as being a certain way.

Now, other things being equal, if someone has certain cares, she will be disposed to respond to the ways in which those cares are affected. Brady captures this when he writes that ‘[t]he idea that emotions are reactions to matters of apparent importance or significance, and grounded in our cares and concerns, suggests that emotions involve or motivate a behavioural response to such things’ (Brady Reference Brady2013, 11–12). So, not only do cares and concerns influence emotional responses, they also influence behavioural responses.

Non-agents, such as non-human animals and human infants, also experience emotions and have behavioural responses. Having background cares and concerns is thus also necessary for them. This means that simply having cares and concerns is not enough for an emotional action to be a candidate for a rational action, as opposed to simply being merely a behavioural response. Cares do, however, still hold the key.

In the case of non-agents’ emotions, the motivated behavioural response may be just that: a behavioural response. But if we are dealing with agents, we can ask whether there is a special agential way of having cares and concerns that may make the motivated behavioural response something more like a rational action. And it seems that there may be: we do not simply have background cares and concerns that mediate our interaction with the environment; we also actively care about those things. But what does it mean to say that we ‘actively care’?

In the philosophical literature on caring, caring is typically understood as being ‘constituted by a complex set of cognitive, affective and volitional dispositions and states’ (Frankfurt Reference Frankfurt1982, 262). These can be described as a pattern of ‘dispositions to perform certain actions and dispositions to make certain long-term commitments, particularly commitments that conduce to your protecting, promoting, and cherishing the things you care about’, emotional vulnerabilities to feel certain ways in response to how the objects of your cares are being affected, and possessing standing attitudes and feelings, like feeling fondness or concern (Mendlow Reference Mendlow2014, 303). Caring also requires a degree of persistence because someone ‘who cared about something just for a single moment would be indistinguishable from someone who was being moved by impulse’ (Frankfurt Reference Frankfurt1982, 261).

We therefore see that caring is constituted not just by dispositions that manifest as emotional responses, but also by a range of other propensities and states. As part of a complex set of these propensities and states, we expect the propensities and states that underwrite the same cares to cohere with one another, as they are manifestations of the same cares.Footnote ¹⁴ For example, if I love my parents and my love is one manifestation of the dispositions that constitute my caring for them, then I will experience a range of emotions in reaction to how they are being affected. If I care about my parents and they are harmed, then I will feel sadness; if they achieve some goal, then I will feel elation. I am also disposed to act in certain ways in line with how I care for them, dispositions that also constitute my caring for them. I seek ways to look out for them when they are harmed, for instance; or I phone them to congratulate them when they succeed.

A further aspect to caring is that the person needs to identify with her cares because it is possible that some of her cares, and the dispositions, desires or states that would constitute her caring, are ‘external’ to her. What this means is that the person can be alienated from them, experiencing them as something alien or as arising from an external force (Frankfurt Reference Frankfurt and Frankfurt1988). They may be ‘hers’ in the sense that she finds them occurring in her in much the same way that our bodies can be the locus of bodily movements, but it does not follow that they are more than mere occurrences in her. Take the conflicted liberal: she does not identify with her emotion as a manifestation of a genuine or relevant care for her in the situation, and she experiences its continuation as distressing and not ‘hers’. So, for the care, the disposition, desire or state to be ‘hers’, and thereby for it to be part of her actual caring, it is necessary that the person identifies with the care, and the dispositions that constitute her caring.

This identification need not be self-reflexive and explicit. As Jaworska (Reference Jaworska2007) argues, healthy agents are able to explicitly identify with their psychological states, but it is not necessary for them to do so. We can usually tell if someone’s desire, care, etc., is ‘hers’ because the ways in which she is disposed to respond themselves represent her perspective, provided that she does not otherwise reject them.

So, caring is constituted by a pattern of dispositions and states. In addition, caring involves identification. Experiencing an emotion does not: it is possible for someone to respond emotionally, motivating behavioural responses, without having a complex capacity to care or without identifying with the cares and concerns of which the emotion is typically a manifestation. Our focus, however, is on the actions of rational agents, and our aim is to flesh out how an agent can guide her actions by reasons when acting emotionally. As such, our focus is on agents who are capable of responding to reasons seen as reasons, who are capable of critically reflecting on their emotional capacities and responses. The emotional agent thus has the complex capacity to care.

As the emotional agent can have this complex capacity to care, she is capable of identifying with manifestations of the dispositions that constitute her caring, including her emotions, and one way the agent can identify with her emotions is by reflectively endorsing them as generally reliable reason-tracking mechanisms: by reflectively endorsing them, she accepts that they are generally manifestations of genuine cares for her. If she accepts her emotions in this way, she need not explicitly identify with each experience in order for it to be internal to her in a way relevant for it to be a manifestation of her caring. Indeed, this elaborates on the force of an account of rational agency such as Jones’.

Because the rational agent is a caring agent, I shall now argue, she can guide her actions by reasons when acting emotionally. In the next section, I argue that, through being a caring agent, the agent sees the emotional considerations as pro tanto reasons favouring an action choice and, in the following section, I argue that she can also see the emotional considerations as the reason for which she should act.

5. Seeing the emotional considerations as favouring an action

As discussed in the previous section, a prerequisite for responding emotionally for both agents and non-agents is that they have background cares and concerns. Through having cares, the agent takes the emotional considerations seriously as being the case because of their significance. And, as rational agents are capable of responding to reasons seen as reasons, the emotional agent has the complex capacity to care, including identifying with the manifestations of dispositions that constitute caring. In this section, I take the first steps to showing how this understanding of caring can help us meet the challenge of showing how we guide our actions by reasons seen as reasons, when acting emotionally. In particular, I argue that the emotional agent, as a caring agent, is able to see the emotional considerations as favouring a behavioural response as a potential action choice.

Let us start with stating a truism: emotions have a close connection to actions. Paradigmatic emotional experiences are typically states that motivate action, and part of their distinct phenomenology is characteristically an action readiness. The action readiness of emotion, it is widely accepted, is a readiness to act on what is presented by the emotion, where it is a tendency or a readiness to establish, maintain or disrupt some relation that the subject stands in with her environment (Frijda Reference Frijda1986, 70–71). The relevant kind of relation is in terms of the subject’s cares and concerns; remember: other things being equal, if someone has certain cares, she will be disposed to respond to the ways in which those cares are affected. For instance, an action tendency associated with an emotion like anger will be defined by the intent of the action, such as the removal of an obstruction to one’s cares, and may be for attacking or for spitting, depending on what the situation allows.

So, emotions not only have a close connection to action in general, they have a close connection to actions of certain types, those actions that are associated with establishing, maintaining or disrupting some relation that the subject stands in with her environment, and which is typical of an emotion type. In this way, the subject’s emotion motivates a behavioural response to the ways the subject’s cares are being affected.

The emotions of non-agents, however, also share this feature of characteristically involving an action readiness. If the agent is to act and not merely respond behaviourally, it is thus necessary for an agent, as an agent, that she is able to see her emotional considerations as favouring the behavioural response as an action choice. This is where caring becomes important: for an agent, emotions are typical manifestations of dispositions that constitute caring. Because of this, emotional agents are able to see their emotional considerations as pro tanto reasons favouring action choices – as I shall now argue.

If the agent reflectively endorses her emotions as generally reliable reason-tracking mechanisms, as our rational agent does, and if emotional responses are typical manifestations of dispositions that constitute caring, the agent will, as a default, experience her emotion as a manifestation of her having certain cares. I am not claiming here that all emotions are manifestations of dispositions that constitute genuine cares for the agent because someone may not identify with the cares or with the emotion. The point is just that the emotional experience is the same kind of experience as a typical manifestation. It is also not a claim about the content of the emotion, that it somehow represents how the cares are faring in the circumstances. Rather, it is the claim that the kind of experience is the same as experiences that are manifestations of dispositions that constitute caring.

Further, as caring is constituted by dispositions other than those that manifest in an emotion, and if the agent experiences the emotion as a manifestation of her caring, she will be disposed to find other ways to express her caring. Because of the way she emotionally experiences things as being, in relation to her cares, she will therefore favour actions that are also manifestations of the dispositions that constitute her caring. In other words, she sees the emotional considerations as favouring action choices that promote her cares, because of their connection through caring, and, if reasons for action are considerations that favour an action, she sees the emotional considerations as pro tanto reasons favouring those action choices.

This can be brought into relief by considering a case of conflict. Take the example of the conflicted liberal again who, despite her best efforts, experiences fear when encountering an individual from another racial group. She is faced with the emotional considerations of the person as dangerous and, just experiencing fear, she may be motivated to move away from the person. As both her emotion and her inclination to move away are two typical manifestations of dispositions that would constitute a care for her safety, by default she sees the person as dangerous and, because she is a caring agent, she sees the person’s being dangerous as a pro tanto reason favouring the action choice of moving away. This is the case even if she does not identify with her emotion and rejects the emotional considerations. In fact, the conflicted liberal believes firmly that the person is not dangerous, and she dismisses the option to move away as a justified action choice. She nevertheless continues to feel motivated to move, and sees the emotional considerations as favouring such a choice. Part of what is so distressing about her experience is that simply not identifying with the emotion or rejecting the emotional considerations is not enough to change the nature of her experience – she has to work at not performing the action she prima facie favours, but rejects.

As should be evident in the above example, the claim that the agent sees the emotional considerations as pro tanto reasons favouring an action choice is a very minimal claim about the agent’s experience, independent of whether she reflectively endorses the emotion or the emotional considerations, or accepts that the action is justified. The claim is simply that, as an agent who is capable of caring and who experiences an emotion that is typically a manifestation of her caring, she experiences the emotional considerations as pro tanto reasons favouring some behavioural response that promotes those cares as an action choice.

Because my proposal is so far only made up of this minimal claim, it is not sufficient to show how an agent can guide her actions by reasons seen as reasons, when acting emotionally. My proposal does not yet show how the agent sees the emotional considerations as the best reason, or as the reason for which she should act. The conflicted liberal, for instance, may see moving away from the person of another racial group as a potential action choice because of her emotional considerations of the person as dangerous, but she nevertheless rejects that the person is dangerous and that she should so act. If she does still move away on the basis of her fear despite her all-things-considered judgement, she does not guide her action by what she sees as the reason for which she should act. In the next section, I therefore argue that the emotional agent can be aware of the emotional considerations, not just as a pro tanto reason favouring an action choice, but also as the reason for which she should act.

6. Seeing the emotional considerations as the reason for which she should act

I have argued that the agent, as a caring agent, takes the emotional considerations seriously because of their significance and, because of her emotion’s being part of a pattern of caring, she sees the emotional considerations as pro tanto reasons favouring action choices that promote her cares. But does she actually see those considerations as justifying the action? As discussed in Section 3, the relevant sense of justification here is that the agent sees the considerations comparatively: she sees them as pro tanto reasons but also as the prima facie best reason, or the reason for which she should act, because she chooses one action over another. So, for the agent to see the emotional considerations as justifying and not merely as favouring the action, she must in some way see the considerations as the reason for which she should act. In this section, I propose a way in which we can test if this is so, when the agent acts emotionally.

We first need to identify what must be the case if the agent sees the emotional considerations as the reason for which she should act in the right kind of way. In order to do so, let us work with the example of Emily the PhD student that both Arpaly (Reference Arpaly2002) and Jones (Reference Jones and Hatzimoysis2003) discuss.

Emily is a PhD student in chemistry who believes firmly that the PhD is right for her and that she ought to continue. She nevertheless feels ‘restless, sad, and ill-motivated’ (Arpaly Reference Arpaly2002, 49), feelings she views as groundless. Yet, despite her all-things-considered judgement to continue with the programme, she quits. She later comes to realise that her feelings were actually responses to the fact that the programme was not right for her. In such a case, it seems that Emily’s feelings are tracking reasons better than her judgement, and her action is a response to genuine reasons, that the PhD programme is not right for her. From this perspective, quitting the PhD may be more coherent with who Emily is, and ultimately she may be more rational, as a whole, in quitting. But is her action itself rational in the sense that her action is guided by reasons she sees as reasons?

It is not clear that Emily is acting on the considerations of the PhD not being right for her. Remember that Emily dismisses her ill-feelings as groundless. This could be in one of two ways: either Emily traces her ill-feelings back to the PhD programme, that it is not right for her, but dismisses the possibility that this really is the case; or, Emily is simply not able to identify why she feels the way she does, a common-enough scenario.

If the latter, then Emily may act because she is aware that something is not right, but cannot identify what, or she acts simply because she feels bad. But neither of these considerations directly favours quitting the PhD programme, and it is a push too far to suggest that she quits the PhD programme because she sees something’s not being right or her general ill-feeling as the reason for which she should quit the PhD. At the very least, she needs to make the connection between the considerations and the action, but then she is not acting purely on the basis of her emotion. So, if she is not aware of the considerations that the PhD is not right for her, she cannot be acting for them seen as a reason.

Suppose instead that Emily dismisses her ill-feelings as groundless because she traces the ill-feelings back to the PhD programme, that it is not right for her. She nevertheless dismisses the possibility that this is really how things are. It remains the case, however, that she is still aware of the emotional considerations of the PhD not being right for her as favouring quitting the PhD and, in that sense, as pro tanto reasons. But she does not see the considerations as her best reason amongst the competing considerations, and she does not see it as the reason for which she should act – or even as a reason all-things-considered at all. When she nevertheless quits, she is thus not acting for what she sees as the reason for which she should act, given all the competing considerations of which she was aware at the time.

By looking at the case of Emily, we therefore see that, if the agent is to act for the emotional considerations seen as the reason for which she should act, it must be the case that the agent is aware of the emotional considerations as favouring the particular action, and that she does in fact see them as the reasons for which she should act. To test if this is the case when the agent acts emotionally, we can ask two questions. First, if the agent were to focus only on the considerations that she took to favour so-acting at the time of acting, would she focus on the emotional considerations? If yes, then it is plausible that she acted for those considerations, and we can rule out the considerations she only identifies at a later stage. Second, if the agent were to take into account all the competing considerations of which she was aware at the time of acting as having a bearing on her action choice, would she still have endorsed the action? If yes, then it would seem that she did act for what she saw as her best reason for which she should act or, at least, that she acted for considerations she saw as pro tanto reasons favouring the action choice and had no competing considerations to cause doubt.

Let us ask these questions of Emily. If Emily were to focus on all the considerations that favour her quitting the PhD programme at the time of acting, would she focus on the emotional considerations that the PhD is not right for her? If she was unable to identify the source of her ill-feelings, then the answer is No. Perhaps she would focus on the way something is wrong or on the way she feels bad, but these are not the considerations that directly favour the action. If she was able to identify the source of her ill-feelings as the way the PhD is not right for her, then she would focus on the emotional considerations. So, it is plausible that Emily could have acted for those considerations at the time of acting. However, when we turn to the second question and ask, if Emily took into account all of the competing considerations she was aware of at the time of acting as having a bearing on her action choice, would she have endorsed the action? The answer is No. Emily in fact rejected the emotional considerations as the reason for which she should act. As a result, we can establish that Emily did not see the emotional considerations as the reason for which she should act, and she did not guide her action by reasons seen as reasons.

What is important about this analysis is that it is possible that Emily could have acted rationally when acting emotionally, so long as she would have endorsed the action. In a default case where there is no conflict, therefore, emotional actions can come out as guided by reasons. For instance, when I am afraid of the bear while hiking in the Rockies, I am aware of the dangerous bear and hide amongst the rocks to escape it. In such a case, if I were to focus on all the considerations I was aware of at the time as favouring hiding, I would focus on the dangerousness of the bear. And given all that I was aware of at the time of acting, I would endorse the action. My action is guided by reasons. But rather than just accepting that the agent is a rational agent who monitors and cultivates her emotions, and accepts them as generally reliable reason-giving mechanisms, we now have a story for how her individual actions are guided by reasons.

On my proposal, we also see that in cases of akratic action the agent is not guiding her action by reasons in the relevant way. In Jones’ example of the person who raises a topic at a meeting because of her feminist anger despite her judgement not to do so, for instance, the person does not see her emotional considerations as the reason for which she should act. My proposal thus fleshes out the details missing from an account like Jones’. Nevertheless, it does not replace it, and we need the two together to begin to give a satisfactory picture of our rational agency as a whole. Let us return to the conflicted liberal so see why.

Suppose that the conflicted liberal accepts that her emotions are generally reliable reason-giving mechanisms, has the capacity to critically reflect and the dispositions to do so when needed and to act on the results of those reflections. She is a rational agent. She is not, however, aware that she has an implicit racism – it is implicit, after all. Now, suppose that she experiences fear when approaching a person of another racial group. In this particular case, she has no current awareness of her reflective commitments, de re or otherwise, but she is aware of the emotional considerations, namely the person as dangerous. She crosses the road to get away from the person on the basis of her fear. In such a case, if she were to focus only on her considerations in favour of crossing the road, she would focus on the emotional considerations. In addition, because she was not consciously aware of other competing considerations at the time, she would endorse her action as being guided by what the pro tanto reason was that she took to favour the action, given what she was aware of at the time, namely the person as dangerous. The action may be at odds with her endorsed values, but it is not at odds with the fact that she acts in line with seeing the person as dangerous.

As a caring agent, however, this is an undesirable situation to be in. When the agent reflects on her action and why she took the person to be dangerous, and reflects on her values and interests, she would not endorse the action. In fact, she will feel distressed about her action and herself. Distress like this indicates that there is still a failing on the part of her rational agency, even if her individual action was guided by putative reasons.

There may nevertheless be ways to avoid such failings. Suppose now that, through her distress and her reflection on her action, the liberal realises that she has an implicit racism, and actively seeks to tackle it by cultivating more admirable attitudes. As a result, she is now disposed to feel shame, say, whenever she experiences fear of a person of another racial group. Or, perhaps she rejects that her fear reliably tracks danger in such a situation and cultivates a de re suspicion of such fear. If so, she may not be actively deliberating about her reasons when she experiences fear of a person from another racial group. Nevertheless, her dispositions are such that she will have a set of conflicting considerations with a bearing on the same action choice. These conflicting considerations would ideally provide a check on her acting on emotions she does not trust. If she does still act on her fear, she no longer rationally guides her action. While not ideal, the rational nature of her action will nevertheless be aligned with her overall commitments and she can, at least, take comfort in the fact that she has internalised her expressed values so that they can have an input in guiding her actions.

So, cultivating and monitoring her emotions is not just important for regulating reason-giving mechanisms like her emotions, it is also a potential way for the agent to develop a check on the emotions she does not trust, and for the rational nature of her actions to be aligned with her overall commitments.

Funding

This work was supported in part by the National Centre for Competence in Research (NCCR) ‘Affective Sciences’.