1 Introduction

In everyday discourse, being wise can mean a variety of different things: from being a good decision-maker, or being farsighted, to being an expert advisor or mentor. In particular, the ability to be empathetic seems – at least on an intuitive level – to be an implicit common denominator of all characterizations of wisdom that consider it essential for deliberating well in situations that involve other people beyond the subject herself. As the literature on practical wisdom attests, being empathetic, i.e., successfully feeling what people are going through, or understanding how they see the world, or how they will act, seems to be a valuable asset for the wise person to possess, especially when their decisions directly affect other people (Gallagher 2007).

This is also consistent with the fact that it is common to think of empathy as one of the building blocks of our interpersonal lives. Empathy, in some cases, is even considered a synonym for a good person (Simmons 2013). In moral philosophy, the connection between empathy and morality has been abundantly explored with a shared tendency to recognize empathy as a positive contribution to our moral lives: Empathy is thought to be able to motivate moral action (Slote 2016; Baron-Cohen 2011), to be necessary for virtue when properly cultivated (Peterson 2017), and to be able to gain an understanding of other people’s inner states (Goldman 2006; Stueber 2009).

In this sense, it is not difficult to see how the temptation to relate practical wisdom and empathy can be strong, even when discussing the Aristotelian concept of phronesis (Warren 2021). Apart from this general conceptual assonance, however, there have not been many systematic attempts to conceptualize their relationship. The aim of this paper is to address this rather unexplored topic by investigating, on the one hand, what role empathy is thought to play in phronesis and, on the other hand, whether there is a legitimate place for empathy in Aristotelian (or neo-Aristotelian) accounts of phronesis. Specifically, in the first part, after a brief overview of Aristotle’s account of phronesis, I will try to define three different ways in which empathy is thought of to contribute to phronesis according to the existing literature. My taxonomy will be based on which characterization of the empathic relationship – affective empathy, cognitive empathy, or sympathy – the authors have chosen to focus on. In the second part, I will challenge each of the three ways in which empathy is supposed to play a role in phronesis, questioning their compatibility with the Aristotelian conception and, more generally, whether the involvement of empathy in practical wisdom is necessary or whether it might lead to morally problematic if not immoral results.

2 Empathy as an Ally to Phronesis

As Aristotle argued in Book VI of Nicomachean Ethics (NE), phronesis is one of the five intellectual virtues together with craft, scientific knowledge, philosophic wisdom, and intuitive reason. Phronesis and craft knowledge (techne) belong to the rationally calculating (logistikon) part of the soul while scientific knowledge (episteme), philosophic wisdom (sophia), and intuitive reason (nous) belong to the scientific part of the soul (epistemonikon) (NE VI 1138b19; 1139b18)Footnote 1.

In particular, Aristotle emphasized that phronesis must be neatly distinguished from both scientific knowledge and craft knowledge. Whereas scientific knowledge is concerned with first principles that cannot be otherwise, phronesis is concerned with human action that does not follow necessary principles because it can be otherwise (NE VI 1140b 3–4). Moreover, phronesis is inherently different from the act of production (poiesis), which has its end in the thing the agent produces, whereas the end of phronetic action is “acting well itself” (NE VI 1140b8).

After clarifying what phronesis is not, Aristotle defined it as “a state grasping the truth, involving reason, concerned with action about things that are good or bad for a human being” (NE VI 1140b5-7). To fully understand what phronesis is, however, we need to look at people who possess it: Pericles is one of those people since he is “able to study what is good for themselves and for human beings” (NE VI 1140b10-11). Crucially, the proper activity of phronesis is “deliberating well” (bouleusis) about what is good and bad in human actions and behavior (NE VI 1141b10-11). It follows that the concept of decision (proairesis) is necessarily involved in what phronesis is and how it works and, for a decision to be good, “the reason must be true and the desire correct” (NE VI 1139a24-25).

By defining decision as “deliberative desire” (NE VI 1139a24), Aristotle makes it abundantly clear that, in order to cultivate phronesis and deliberate well, it is necessary to balance both the rational and the desiderative parts of the soul. Good decisions are the product of a desire informed by reason, that is, a desire that desires what practical reason has determined to be the most appropriate course of action in the situation. The wise person who is able to deliberate and act well embodies this inextricable blend of understanding and desire (NE VI 1139b5-6), which is not only the essence of what decision is but also exemplifies what human nature ultimately is.

Importantly, although phronesis is an intellectual virtue, it is closely related to the moral good. In fact, phronesis and ethical virtues are inevitably intertwined: to act virtuously, excellence of character (ethical virtues such as courage, temperance, justice, etc.) and excellence of intellect must work in concert to set the right end and choose the right means to achieve it (NE VI 1144b32-33). Phronesis, then, is the intellectual virtue that enables the agent to deliberate about the good course of action, i.e. to act virtuously, in a given situation by grasping the relevant (moral and non-moral) aspects of that situation.

It is within this Aristotelian account that empathy has been thought by some to interact fruitfully with phronesis. Unfortunately, as is often the case with empathy, there is some ambiguity about its precise definition, which can greatly affect how the interaction between these two concepts will take shape. Thus, in an effort to shed some light on this aspect, I suggest that some conceptual housekeeping about the definition of empathy is needed to better understand how the relationship between empathy and phronesis has been conceptualized in the literature.

One possible way to characterize empathy, which seems to enjoy some consensus in the literature, is as an umbrella concept consisting of three different phenomena (Decety and Jackson 2004; Maibom 2014): affective empathy, cognitive empathy, and sympathyFootnote 2. Cognitive empathy is thought to be the mental activity of perspective taking, which aims to understand the other person’s internal states (both cognitive and affective) (Spaulding 2017). Affective empathy is defined as the phenomenon of affectively sharing the other person’s emotional states with varying degrees of isomorphism (Maibom 2017). Finally, sympathy is thought to consist of concern for the welfare and general well-being of the other person (Darwall 1998)Footnote 3.

The distinction between the three phenomena seems justified on conceptual grounds since cognitively taking the perspective of others is a conceptually distinct operation from experiencing the emotional states of others or feeling concerned about their condition and alleviating their suffering. The conceptual distinction is also reflected at the phenomenological level since these three operations feel qualitatively different from the subject’s point of view. Moreover, this tripartite distinction seems to be supported by the results of several empirical studies, especially neuroscientific ones (Guo 2017; Singer 2006), which attest to the involvement of different neural pathways when it comes to cognitive and affective empathy-related phenomena (Decety and Cowell 2014). Relevantly, cognitive empathy, affective empathy, and sympathy also have inherently different goals: while cognitive empathy has the ambition to gain epistemic access and thus properly understand the other’s inner states, affective empathy has the sole goal of sharing the other’s feelings, which can only provide an epistemically superficial grasp of the other’s condition, as well as sympathy, which is specifically interested in the other’s welfare.

At the same time, however, the claim that empathy consists of three distinct phenomena does not mean that they cannot or do not overlap in our everyday experience: perspective-taking and affective sharing can occur simultaneously as a result of the same interaction with the other person. Alternatively, they can be in asynchronous relationships: cognitive empathy can lead to affective empathy or vice versa. Moreover, according to some scholars (Batson 2011; Denham 2017), empathy, especially affective empathy, can cause, motivate, or lead to sympathetic actions, which can be characterized in terms of altruistic and/or moral actions. In contrast, for others (Battaly 2011; Prinz 2011b), cognitive and affective empathy have nothing to do with sympathy or morality.

According to this definition, empathy is a broad conceptual label used to refer to the very general phenomenon of being in an empathic relationship with the other, which can be characterized either cognitively as perspective taking, affectively as emotional sharing, or sympathetically as concern for the other’s welfare. The distinction between cognitive empathy, affective empathy, and sympathy is particularly relevant to this discussion because empathy seems to play a different role in phronesis depending on which characterization of empathy different authors focus on. In this sense, applying this tripartite definition to the current literature on empathy and phronesis, there are three possible ways in which empathy can be a part of phronesis: affective empathy is involved in phronesis (EP1 from now on), cognitive empathy and sympathy are involved in phronesis (EP2 from now on), and cognitive empathy is involved in phronesis (EP3 from now on).

For example, Svenaeus (2014) argued for a version of EP1. According to him, “empathy is (…) the feeling component of phronesis” (Svenaeus 2014, p. 296). The conception of empathy at work in Svenaeus’s EP1 is – although not explicitly mentioned – compatible with the definition of affective empathy: the way to empathize with the other person is to share what she feels. Since affective empathy provides the subject with this ability to affectively grasp what is going on with the other person by sharing the other person’s feelings, it is necessary to set the phronetic process in motion and guide the phronetic agent in her deliberations about the specific situation in which the target of her empathy is involved.

So empathically sharing the other person’s emotional state should both ignite phronesis and signal to the agent extremely valuable aspects of the situation at hand that might otherwise be ignored and compromise the outcome of the deliberative process. Without empathically sharing that a student is panicking during the exam, the wise professor might not be able to alert herself that the wise decision is to give the student a breather so that she can relax and perform at her best. Without affective empathy, the professor would have been blind to the need to even begin a deliberative process and exercise phronesis, resulting in simply failing the student or giving her a poor grade. Importantly, following Aristotle (NE VI 1144b32-33), phronesis and affective empathy are not sufficient, though they are necessary, to guarantee the morality of the action: ethical virtues must be involved. To act virtuously, the professor must not only be empathetic and wise but also possess the virtues of character.

According to Svenaeus, the role of affective empathy in phronesis is justified by the affective nature of phronesis itself: “Phronesis is not devoid of feelings, it is rather based in feelings that help the wise person to see and judge what is at stake in the situation” (Svenaeus 2014, p. 295) In his view, this is proven by a locus classicus in Aristotle’s Book II of NE: “Having these feelings at the right times, about the right things, toward the right people, for the right end, and in the right way, is the intermediate and best condition, and this is proper to virtue” (NE II 1106b21-24). In short, to be wise and virtuous, it is necessary that these “right feelings, toward the right people, and so on” (“the right feelings” from now on) accompany the action. Note that this version of EP1 is particularly strong: there is no phronesis without (affective) empathy (and ethical virtues).

In contrast to Svenaeus, Darnell and colleagues (2022) defended a version of EP2 by assigning a role to cognitive empathy and sympathy in phronesis. They argued that phronesis has four functions: the constitutive function, the integrative function, the blueprint function, and the emotional regulative function (Darnell et al. 2019). In their empirical study (Darnell, Fowers, and Kristjánsson 2022), cognitive empathy and sympathy were used as measures to assess the ability of phronetic subjects to regulate emotional states. Specifically, following the Interpersonal Reactivity Index (IRI), subjects were asked to self-report their perspective-taking ability and empathic concern, defined as “adoption of others’ viewpoints”, i.e., cognitive empathy, and “compassion and concern for others”, i.e., sympathy (Darnell, Fowers, and Kristjánsson 2022, p. 5).

According to their neo-Aristotelian model of phronesis (APM) (Darnell et al. 2019), part of being wise – along with the other three functions – is being able to use reason to shape one’s emotional responses so that they are appropriate to the situation at hand. Gaining insight into the situation by imaginatively taking the other person’s point of view and caring for her should enable the wise person to understand and evaluate what the appropriate emotional response is and, if inappropriate, to correct and precisely regulate the agent’s affective responses. For example, cognitive empathy and sympathy should lead the wise person to recognize that happiness, rather than indignation, is the appropriate emotional response to a friend’s bragging about a promotion, given the wise person’s genuine compassion and understanding of the hardships she has experienced in the past year. So, one aspect of being wise requires being empathetic. Interestingly, to justify the emotional regulatory function of phronesis, they used the same famous Aristotelian passage about the need to have the “right feelings” when acting virtuously that Svenaeus used to justify his version of EP1.

Another way to characterize the role of empathy in phronesis is to refer to the Aristotelian concept of sungnome. As Russell pointed out (2009), Aristotle understood phronesis as a virtue composed of specific practical capacities, among which is sungnome. Russell interpreted the concept as the “ability to see things from another’s point of view in deliberating about what is reasonable and appropriate”, i.e., cognitive empathy (Russell 2009, p. 21). The idea is that the person possessing sungnome will be a particularly fair and reasonable judge of the behavior and conduct of others because she will be able to take their perspective and gain a much better understanding of their overall situation and condition. As Aristotle put it, “Considerateness is the correct consideration that judges what is decent, and correct consideration judges what is true” (NE VI 1143a24-25). It follows that having sungnome is a particularly useful skill for the wise person to possess (NE VI 1143a27-28). It gives the phronimos access to epistemically valuable information about other people involved in the situation that she needs to deliberate about. For example, putting oneself in the position of someone who has just lost their job should allow the person with sungnome to fairly consider the decision not to donate to the annual charity raffle and wisely deliberate what would be the virtuous thing to do in that situation.

In these three accounts, not only is empathy seen as certainly having a role in wisdom, but always a positive one. Affective empathy, cognitive empathy, and sympathy are seen as valuable and indispensable allies of the wise person. In the next section, I will try to put some pressure on these two propositions: I will question whether empathy has a legitimate conceptual place in Aristotle’s account of phronesis and, more in general, whether the wise person should rely on empathy in order to deliberate and act well.

3 Is Empathy the Right Match for Phronesis?

3.1 Affective Empathy and Phronesis

According to Svenaeus’s version of EP1, affective empathy constitutes the feeling component of phronesis. Two issues: On the one hand, having affective empathy does not seem to guarantee having the “right feelings” necessary for acting virtuously, and on the other hand, its specific role in phronesis is not well defined.

If I understand Svenaeus’s position correctly, affective empathy consists of feeling what others feel, which in turn should allow us to initiate the phronetic process and ultimately to act virtuously and feel the right feelings while doing so (otherwise it would not be a proper virtuous action). Svenaeus also mentioned that affective empathy has an instrumental and preliminary role since it is a feeling that makes us see the other person’s feelings and focuses our attention on the emotionally salient aspect of the situation. However, this is where the problem arises: he used the “right feelings” argument not only to justify the inclusion of an affective phenomenon such as affective empathy in phronesis, but also to imply that affective empathy is responsible for having the “right feelings” that accompany virtuous actions.

In contrast, while the idea that phronesis also involved an affective component is undeniable (Kosman 1980), I want to argue that affective empathy does not amount to the “right feelings” that Aristotle considered necessary for virtuous action. Recall the example of the wise professor: she empathically shares her student’s fear, which preliminarily alerts her that something is wrong and that a decision is needed. Her deliberative process begins, and thanks to her phronetic skill and other ethical virtues, she decides to pause the exam and let the student reduce her anxiety. However, as the professor makes this decision and acts on it, she does not feel the initial feeling of anxiety that she felt through affective empathy. If this were the case, we – including Aristotle – would be hard-pressed to judge her action as virtuous: feeling anxious while postponing the exam would be inappropriate, just as feeling resentment, anger, or other negative emotions would be. Instead, for this action to be considered virtuous, the professor should be willing to allow the student to take a break. Thus, there is no guarantee that the feelings experienced through affective empathy are the appropriate feelings to accompany virtuous actions. In fact, I would argue that there is a greater chance that these feelings will be inappropriate since they simply mirror the emotional state of the other person involved in the situation without taking into account any of the specific factors of the situation at hand.

Moreover, defining affective empathy as the “feeling component” of phronesis might create some confusion about at what point affective empathy intervenes in phronesis. It implies that empathically acquired emotional states are active throughout phronetic deliberation, not just as its starting point. This tension emerges clearly in Svenaeus, since he began by saying that “empathy is the starting point of phronesis, without which it cannot be performed,” but then immediately adds that “the wise deliberation of phronesis must be guided by an emotional discernment of the ways other people feel and think and other moral virtues”, i.e., by affective empathy (Svenaeus 2014, p. 297, my emphasis).

It is not entirely clear whether empathy is only the starting point of phronesis or whether it guides the wise person throughout the deliberative process. In either case, its definition as the “affective component” should be revised. If it is the former, then “component” is not the right term: affective empathy should be characterized as a preliminary tool for identifying the emotional aspect of the situation about which the wise agent must deliberate, not as a constitutive part of phronesis: phronesis only comes after affective empathy. If it is the latter, then qualifying the component in terms of “feeling”, i.e. as affective empathy, risks compromising the phronetic process itself: to reiterate, sharing the other person’s emotional state might lead the empathetic agent not to have the “right feelings” when deliberating and acting, and thus not to be wise.

Leaving aside the specific aspects of Svenaues’s account, I find that even a more general version of EP1 – i.e., affective empathy is involved in phronesis – can raise some issues. A standard argument for this position might be that since Aristotle argued that “thought by itself moves nothing” (NE VI 1139a35-b4), the wise person must be moved by desire which, in EP1, is understood as the empathically experienced feelings of the other people involved in the situation. According to Aristotle, in order to choose wisely and act virtuously, the agent must desire to pursue the right course of action, otherwise she will not actually perform a virtuous act. Affective empathy could provide this “desiderative push”. This may seem plausible in cases where the other person is suffering: empathically feeling her pain might motivate the agent to deliberate wisely about the best course of action to alleviate the other’s pain and, consequently, to take it.

To be sure, desiring to act virtuously and having the right feelings while doing so are two different, though related, aspects of phronesis and virtue. For an action to be virtuous, the agent must desire to choose that course of action and must have the appropriate feeling while performing that action. For example, the agent must not only desire to spend Sunday feeding the people in need instead of going to the movies, but she must also be happy while doing so, not hate or resent every minute of it. Accordingly, to argue that affective empathy constitutes the feeling component of phronesis in the sense of being the source of these right feelings (Svenaeus’s version of EP1), or to argue that affective empathy is the desire to deliberate wisely about and perform virtuous actions (general version of EP1), albeit interconnected, are two different theoretical positions.

Contrary to the general version of EP1, I do not think that affective empathy is the right conceptual candidate to explain what the desiderative part of phronesis is. To see why, we need to be more precise in our translation of the term desire. The kind of desire involved in decision and, therefore, in phronetic deliberation, is not what Aristotle called emotion (thumos) or appetite (epithumia) (NE III 1111b16-17). Instead, “decision is the result of wish (boulesis) for the end” and, most importantly, “wish and decision are concerned with the good (NE 1113al5-16, 11lbl7-18) whereas emotion (thumos) and appetite (epithumia) are not (NE 111Ib16-17)” (Irwin 1976, p. 570)Footnote 4.

It is the connection between boulesis and moral good that makes affective empathy incompatible with phronesis: affective empathy has absolutely no relation to the good, since there is nothing inherently good about it. Sharing what others feel is not in itself a morally good operation. As empirical evidence attests (for an overview see Batson 2014), it is entirely plausible and it often happens that merely sharing the suffering of others does not lead to performing a moral action, nor to anything resembling a moral feeling toward the sufferer, such as concern or care. In short, affective empathy does not automatically or necessarily lead to the good, that is, to moral or even simple altruistic action (Prinz 2011b; Ugazio, Majdandžić, and Lamm 2014). If it did, we would spend most of our lives performing altruistic or moral acts. Only what Batson (2011) called “empathic concern”, which is already a morally colored feeling of care for the other person, seems to be associated with altruistic actions, but not necessarily with moral actionsFootnote 5. Conversely, when it comes to affective empathy, we can share the plight of someone begging on the street, have a quick but distinct feeling of suffering, but at the same time not help that person and not care about his or her fate. Whether we end up helping that person does not necessarily depend on whether we share his or her plight, and even if we do share his or her plight, it does not mean that we will be motivated to help him or her.

Not only can affective empathy be insufficient or completely ineffective in motivating the good, but it can also be used for morally evil purposes (Prinz 2011a). Consider sadists: they do what they do precisely because they enjoy sharing the pain of others (Scheler 1913). But even without looking at the pathological realm, we all still sometimes use affective empathy to achieve our own not-so-noble ends. When we are angry, we sometimes say things that we know will hurt the other person, and the ability to empathize with the target of our anger is crucial to identifying what that thing might be. For example, only if we empathically share the other person’s embarrassment are we able to use it, if we so choose, to make it worse. Moreover, as Batson and colleagues famously showed, in resource-allocation situations when instructed to empathize with the target, the subjects “forsake justice in the interest of benefiting the person whom they felt empathy, even though they perceived doing so to be less moral and even though they knew this person’s need was not as great as the need of others” (Batson et al. 1995, p. 1052). The implications of these findings are quite powerful: empathy-based motivation can override justice-based motivation, leading to unfair actions (Batson et al. 1995, p. 1051), such as prioritizing helping someone just because they are the target of empathy while ignoring more deserving people who do not elicit empathy.

So, empathy seems suspiciously “compatible” with actions that we would generally consider to range from problematic to downright immoral (Prinz 2011b, p. 221). To be clear, I do not intend to deny that sometimes empathy does motivate moral actions (Masto 2015). My point is simply that we are (and should be) more concerned with potential sources of immorality than with potential sources of morality, which means that the attention to the potential immorality of empathy should trump the attention to the potential morality of empathy, especially if we are determined to assign it a role in phronesis. It follows then that it is precisely because of the relation of affective empathy to moral good and, especially, to moral bad that the general version of EP1 does not seem justified.

3.2 Sympathy, Cognitive Empathy, and Phronesis

At this point, one might think that EP2 would have a better chance than EP1, since it points to sympathy and cognitive empathy, rather than affective empathy, as playing a role in phronesis. In Darnell and colleagues’ version of EP2, sympathy and cognitive empathy are responsible for regulating the wise person’s emotional responses, that is, for ensuring that the wise person has the “right feelings” when acting virtuously. As with EP1, I am concerned with both the proposal itself and its compatibility with Aristotelian phronesis.

First, it seems reductive to assign the role of emotional regulation exclusively to sympathy and cognitive empathy. Is concern for the other person’s well-being and the ability to imaginatively take the other person’s perspective sufficient to ensure that the empathizer has appropriate emotions while acting? In my view, there are many more tools we need to get thereFootnote 6.

For example, it might be important for the wise person to have some understanding of the socio-cultural context in which the deliberative process is taking place. This is especially useful in cases where the people involved in the deliberative process are not “near and dear” and the situation at hand involves broader dynamics that go beyond the person’s limited life experience. Sympathy and cognitive empathy do not seem to have much to say in such situations (Throop and Zahavi 2020; Hollan 2017).

Suppose a wise person wants to support the environmental cause and decides to donate some money to an NGO that is involved in environmental protection. As part of their activities, some members of the NGO organize public protests to raise awareness about the issue. The wise person dislikes these protests because she thinks they are too aggressive. However, she still donates money, but in doing so, she feels contempt for those who participate in these protests. Her donation will not count as a virtuous action unless – according to Darnell and colleagues – her negative feeling is regulated by sympathy and cognitive empathy: concern for the welfare of those involved in the NGO and for the people in general, combined with the ability to take the NGO member’s perspective and better understand their choice, should be enough to change the wise person’s emotional response to, say, joy and hope, i.e., the appropriate feelings to accompany the donation.

To me, it is not enough. In order for the wise person to regulate their emotional response, they need to understand the bigger picture. In this case, it is crucial to understand that, for instance, politicians and institutions are reluctant to discuss environmental issues because they are inherently long-term, and the media is only interested in them when something catastrophic happens or when protests are particularly loud, hence the need to organize such protests. The point is that this kind of valuable contextual understanding does not seem to be achievable through sympathy or cognitive empathy which are merely active on an individual level.

Moreover, empirical evidence suggests that both sympathy and cognitive empathy work better when a relatively small number of people are involved in the situation, and even better when these people are perceived as in-group members (Gutsell and Inzlicht 2010; Zaki and Cikara 2015; Hein et al. 2010). It therefore seems phenomenologically unrealistic to expect a wise person to care about a large number of people, especially strangers, and to be able to successfully take their perspective. As it turns out, empathy and sympathy are limited resources: we cannot (and do not want to) empathize and sympathize with everyone all of the time, so we select our targets according to various factors, most importantly group membership. In Paul Bloom’s words: “Empathy is a spotlight focusing on certain people in the here and now” (Bloom 2016, p. 9). It follows that when the wise person has to think about a more complex situation with a broader scope and involving many people who are usually unidentifiable (Jenni and Loewenstein 1997) – as in the case of the NGO –, sympathy and cognitive empathy seem to be less effective and thus lose their power to regulate one’s emotional response.

Not only do sympathy and cognitive empathy appear to be insufficient, but they also do not appear to be necessary for the subject’s ability to perform the emotional regulative function. This seems to be implicitly confirmed by the authors’ own definition of emotional regulation as “the ability to infuse one’s emotional experience with reason to appropriately shape those emotional responses” (Darnell, Fowers, and Kristjánsson 2022, p. 3). There seems to be no necessary causal connection between sympathy and cognitive empathy, on the one hand, and the infusion of reason into emotional responses, on the other. Caring about someone and understanding their perspective does not necessarily guarantee that the wise person’s emotions will be appropriate to the situation by infusing them with reason. Sometimes the initial inappropriate emotional reaction is only reinforced and entrenched by taking the other perspective or caring for her wellbeing. And even in successful cases, it is reason, not sympathy or empathy, that does the regulating work, i.e. will concretely infuse and regulate the agent’s emotions. Crucially, recognizing reason as responsible for emotional regulation does not imply excluding or suppressing emotion in phronesis, which is by nature and, in fact, must be affectively charged, but it does prevent assigning this task, or at least not exclusively, to cognitive empathy and sympathyFootnote 7.

Finally, I am not convinced that the concept of emotional regulation can legitimately be considered one of the functions of phronesis, at least in the strict Aristotelian sense. The “right feelings” that must accompany the wise person in her deliberations and actions are not something that should require an intentional act of regulation carried out through sympathy and cognitive empathy. The wise person should be able to feel effortlessly and organically what is appropriate in that particular situation precisely because she is wise: the wise person should already react with happiness, not indignation, to her boastful friend’s announcement of a promotion, without the need for cognitive empathy or sympathy. Phronesis is not responsible for regulating these feelings, because if there is phronesis and virtue, there should be nothing to regulate. In fact, “in the fully good person excellent thinking, desiring, feeling, and action all coalesce” (Gottlieb 2021, p. 52).

The wise person no longer needs emotional regulation. If she does, it means that she is still in the learning stage of virtues and phronesis (NE II 1103b8-13). In the learning stage, the person is able to adjust the way she acts and feels based on “feedback from the world and from other people,” which in turn results in a habitual disposition (Gottlieb 2021, p. 55). Importantly, this adaptation process can go either way: if the feedback reinforces inappropriate actions and feelings, the person will develop a bad character. On the other hand, if the feedback leads the person to correct her actions and feelings so that they are appropriate to the situation, the person will develop a good character and thus phronesis (NE II 1103b14-20).

Undoubtedly, emotional regulation can be very useful to the learner in the acquisition of virtues and phronesis. In particular, sympathy and cognitive empathy can sometimes help the learner acquire and better understand the feedback that other people, especially role models, give her. However, once the learning stage is completed and the virtuous character is developed, emotional regulation seems to be an unnecessary step for the wise person, who, by virtue of being wise, already displays feelings “at the right times, about the right things, toward the right people, for the right end, and in the right way” (NE II 1106b21-24). In short, emotional regulation can sometimes help achieve phronesis, but it is not one of the functions of phronesis.

3.3 Cognitive Empathy and Phronesis

I have raised some doubts about the involvement of empathy in the affective and desiderative aspects of phronesis. A less problematic approach might be to consider empathy’s contribution to the evaluative aspect of phronesis. For example, in Russell’s version of EP3, sungnome is one of the practical capacities that help the wise person to deliberate. Although Russell – following Louden (1997, p. 114 − 15), who in turn referred to Stewart (1892, p. 88) – translated the term sungnome with the concept of “sympathy” (Russell 2009, p. 21), his subsequent definition of the term – i.e. seeing things from another’s point of view – seems to indicate that the concept of cognitive empathy would better describe the ability the commentators had in mind rather than sympathy, which is often associated with concern for the status or compassion for the plight of another person. While I agree with both the inclusion of sungnome in phronesis and with its general definition as the ability to consider other people’s situations and thus to judge them fairly, I am not convinced that the concept of cognitive empathy, i.e., the ability to take the other person’s perspective, accurately describes how sungnome achieves this and thus its overall contribution to phronesis.

Specifically, what I find problematic is the assumption that perspective-taking is responsible for achieving “the correct consideration” of other people’s conditions (NE VI 1143a24-25). Stewart’s comments on the concept of sungnome are the source of this interpretation, which was later accepted by the majority of commentators (Russell 2009; Louden 1997; Rowe and Broadie 2002). He argued that sungnome “means properly thinking and feeling with others” which allows the person who displays this ability to think and feel with others to “make allowance for their difficulties in his formal or informal verdicts” making them more just than “a rigid interpretation of the law would warrant” (Stewart 1892, p. 88). He, then, insisted that the person who thinks and feels with others is a person “of social sympathy, who enters into the thoughts and feelings of others” (Stewart 1892, p. 88) which led to translating sungnome directly with the ability to have “sympathetic consideration”, i.e., to take the other’s perspective, and sungnomon (the person with sungnome) with “sympathetic judge” (Ross and Brown 2009; Rowe and Broadie 2002; Greenwood 1909)Footnote 8.

To say that to think and feel with others is to enter into their thoughts and feelings seems to conflate two phenomenologically and conceptually distinct abilities. Thinking with others does not mean thinking what others are thinking or thinking as if you were the other person: I can think with my daughter about what is the best college for her to apply to without assuming my daughter’s perspective about what she thinks is the best college for her. Similarly, feeling with others does not mean either feeling what they are feeling, i.e., sharing their emotions, or feeling for them, i.e., caring about them: I can be aware of and understand the players’ sadness at losing the final game of the World Cup, and in that sense feel with them, without either sharing that sadness first-personally or worrying about the players’ well-being. These operations – sungnome as thinking and feeling with others on the one hand, and all the components of empathy, including cognitive empathy, on the other – are qualitatively different not only from a theoretical standpoint but also in how they are concretely experienced by the subjects who perform them: being involved in the other person’s thought process and feelings is not the same as taking his or her perspective or sharing those feelings first-personally.

Moreover, there does not seem to be strong textual evidence to justify the overlap between the sungnome and perspective-taking. Nowhere in the NE does Aristotle directly or indirectly refer to any mechanism that might resemble cognitive empathy. Imagining being the other person in their situation is a very specific and defined cognitive process that does not seem to be mentioned in Aristotle’s discussion of sungnome (NE VI 114320-24). Similarly, the related concept of gnome, usually translated as “consideration” (Irwin 1999), does not seem to overtly imply any perspective-taking activity. This concept is also mentioned in the Athenian juror’s oath, and most commentators assume that this is further evidence – external to Aristotle’s work – that to judge someone fairly is necessary to put oneself, in this case, the judge, in the position of others, in this case, the citizens on trial (Russell 2009; Louden 1997; Gottlieb 2021). However, even in the oath, there seems to be no reference to the necessity of this operation but simply to the act of judging others’ actions fairly: “Bear this in mind, too, gentlemen of the jury: you have entered court today sworn to judge according to the laws … and, of those matters on which the law is silent, to decide by your own lights, in the fairest manner possible” (Demosthenes, Against Leptines 118). A more straightforward interpretation might be that the prefix sun- in sungnome, rather than confirming or reinforcing the sympathetic nature of gnome, simply indicates, as Louden himself first noted, its “other-regarding” character (1997, p. 115): sungnome implies considering the actions of others not our own, rather than, contra Louden, indicating that perspective-taking is responsible for carrying out this activity.

Additionally, what is usually translated as sympathy, i.e. eleos, in Aristotle’s works is quite different from what some commentators have in mind when they characterize sungnome as “sympathetic”Footnote 9. In the Rhetoric, Aristotle defined eleos as “a certain feeling of pain, destructive and painful, undeserved, which we might expect to befall ourselves or some friends of ours, and moreover to befall us soon” (Rh. 1385b13-16). Sympathy (eleos) is thus triggered by the pain of others, but only insofar as this pain is undeserved and is likely to befall the sympathizer or his close friends in the near future. Conversely, sungnome – precisely because it is related to phronesis – should be able to work in all situations, regardless of the people involved and whether their condition is deserved or not. Sungnome also has the specific goal of formulating a judgment after considering the other person’s situation, while sympathy does notFootnote 10.

Thus, for both conceptual and textual reasons, neither perspective-taking nor sympathetic consideration seems to accurately capture the Aristotelian concept of sungnome. I suggest, therefore, that it might be more appropriate to understand sungnome as other-regarding consideration, that is as an openness to and active engagement with other people in order to judge them fairly and, ultimately, to deliberate wisely about the situation in which they are involved. This broader definition of sungnome has two kinds of advantages: first, on a folk-psychological level, it better captures what we actually do when we try to consider and thus understand other people’s situations. Second, it seems more consistent with the concept of phronesis as a whole.

Aristotle aside, identifying consideration for others solely with the mental operation of perspective-taking does not accurately reflect our experience of it. Although consideration for others is only one specific aspect of interpersonal understanding, I still resist the idea of equating it exclusively with perspective-taking or, for that matter, with any of the components of empathy, because it seems to trivialize the complexity and richness of our attempts to understand one another. When we try to consider and understand other people’s situations to attempt to judge them fairly, we systematically use a variety of tools and strategies.

It is also worth noting that in order to fairly judge the actions of others, the information obtained through perspective-taking must be accurate, otherwise, the fairness of the judgment would be compromised. Unfortunately, there is no guarantee that cognitive empathy will provide epistemically accurate information about other people and thus fair judgments about their behavior. Based on empirical evidence (Eyal, Steffel, and Epley 2018; Ickes 1993)Footnote 11, perspective-taking seems to be less accurate than we usually take it to be which warns us once again of the danger of identifying other-regarding consideration solely with perspective-taking.

Incidentally, like affective empathy, perspective-taking has no necessary connection to moral goodness (or fairness) per se: the information gained through cognitive empathy can be used by the agent to harm the other person instead of judging her actions fairly: sharing information gained through cognitive empathy without consent, or using that information to manipulate the other person, are examples of this (Hollan 2017). Of course, this would not be the case if cognitive empathy were exercised by a wise person, but it reminds us that equating other-regarding consideration with perspective-taking ignores that the former is related to the moral good (fairness) and the latter is not.

For these reasons, openness and active engagement with others seem to better capture how we are considerate of the conditions of others as they can manifest themselves in many different ways that are unrelated to cognitive empathy. For example, the ability to listen to other people’s testimonies and find them epistemically valuable (Fricker 2007), combined with epistemic humility (Whitcomb et al. 2017), especially when confronted with experiences very different from our own, is essential to be considerate of others’ situations. When we consider a Black person’s reaction to a racial slur with the intention of judging it fairly, taking his or her perspective could be not only useless but harmful, especially if we are not Black (Jones 2022). In fact, from an epistemic standpoint, it seems extremely difficult – if not impossible – for a non-Black person to accurately reproduce a Black person’s perspective when it comes to experiencing racism (Liebow and Ades 2022), which is not to say that fair consideration and understanding cannot be achieved. It is much more reasonable to use the other person’s direct account of the experience as the most epistemically accurate source of information when making a judgment about their situation. In sum, being considerate of others does (and should) involve a lot more than perspective-taking.

Finally, on a more general note, since other-regarding consideration is one of the practical capacities involved in phronesis, reducing other-regarding consideration to perspective-taking or sympathetic consideration might result in phronesis itself being confined to too narrow a conceptual territory. What is so compelling about the concept of phronesis is the flexibility of its actualization: the mark of the wise person is to choose the most appropriate tools for the specific situation at hand, to craft “made-to-measure” strategies for deliberating the best course of action without losing sight of the good end. If we force other-regarding consideration, as one of the tools available to phronesis, to be a “one-trick pony” and merely a version of cognitive empathy, we also risk deflating the flexibility potential of phronesis itself. To avoid this, it should be left open how other-regarding consideration is pursued, from which it should follow that the toolbox of phronesis will also be unrestricted.

4 Concluding Remarks

I have raised some doubts about the compatibility between empathy and an Aristotelian or Aristotelian-inspired account of phronesis. It turns out that justifying empathy as a part of phronesis is much harder than one might think. Contrary to EP1, affective empathy does not seem to be able to guarantee either the right feelings or the desiderative impulse necessary to be wise and act well. Contrary to EP2, sympathy and cognitive empathy seem to fail at effectively regulating the agent’s emotional reactions by themselves. I have also stressed that emotional regulation is something that is only necessary at the learner stage and not for the wise person. Finally, contrary to EP3, cognitive empathy also seems an inadequate concept for translating Aristotle’s sungnome, since considering and fairly judging the actions of others does not require perspective-taking.

Moreover, I have tried to show that empathy, far from being a perennial source of morality, has some problematic tendencies that seem to conflict with phronesis. Because of its inherent features, such as modulation based on group membership, lack of accuracy, or inability to work with large numbers of targets, empathy, when incorporated into the deliberative process as part of phronesis, can be counterproductive, even compromising the morality of deliberation itself. Conversely, the connection between phronesis and morality is built into the definition of what it means to be wise. As Aristotle explained, phronesis without the moral good (i.e., the ethical virtues) would simply cease to be phronesis, since it would be mere cleverness (deinotes) (NE VI 1144a29-31).

For this reason, my skepticism about the involvement of empathy in phronesis is motivated not only by conceptual, phenomenological, and textual reasons but also by something deeper: there is an ontological difference between how empathy and phronesis relate to morality and immorality. Only for phronesis is morality a conditio sine qua non for its existence. Empathy, on the other hand, need not pursue the moral good in order to exist: it can be used to pursue immoral actions.

To conclude, empathy, rather than being its ally, seems either useless or even detrimental to phronesis. To be clear, I am not arguing for the exclusion of empathy from our moral life tout court, nor am I suggesting that empathy is incapable of contributing to morality at all. Sometimes empathy can play a role in the development of our virtuous traits, as in the case of the agent who is learning to be wise and needs empathy to receive feedback from others. However, empathy is not and should not be considered a component of phronesis, or a quality or capacity that the fully formed wise person needs to exercise this virtue. In other words, the wise person may empathize and sympathize with others, but these empathic or sympathetic acts are not the reasons why she is wise. In fact, under certain circumstances, being wise may precisely require not being empathetic.