II. The Appeal of Accusation as Education Accounts

Strawson presented the free will debate as one between pessimists who worry that determinism undermines any justification for deserved blame, and optimists who argue that blaming can simply be justified on the basis of the good consequences of blaming someone (Strawson, Reference Strawson1962, section 1 and 2). What both of these positions left out, according to Strawson, are the feelings we experience when we perceive others to act in ways that are morally objectionable, and the conditions under which we would, and would not, be willing to modify or suspend those emotional responses (Strawson, Reference Strawson1962, section 3 and 4).

In light of this observation, the post-Strawsonian literature generally concurs that these feelings matter. Being deserving of blame is now often referred to as being an apt target of accusatory anger. Of course, not all forms of blame can be reduced to accusatory anger. One may allocate fault to a person in ways that are devoid of anger. Nor can anger be completely reduced to accusatory anger, because anger may be characterized by appraisals that do not find fault with a person, whereas accusatory anger does. Still, angry responses are often accusations and blaming often falls within an angry emotional register.

Strawsonians believe that one will typically experience such accusatory anger towards a person who is perceived to be at fault, and that one would (or at least should) suspend or modify their anger when it turns out the person was not really, or completely, at fault for what they did. On these accounts accusatory anger can only be apt when the object of anger was indeed at fault.

Metaphysical skeptics worry that the truth-conditions for this requirement can never be met (e.g. Pereboom, Reference Pereboom2014). Epistemic skeptics add that it is difficult to impossible to know whether a person is really blameworthy (e.g. Rosen, Reference Rosen2004). But the majority of philosophers insist that the conditions under which a person may be aptly appraised to be at fault both can be met and can be known.

A person may for example be considered at fault in this way, when they ‘could reasonably be expected to have acted in the required way, to have grasped the facts bearing on the practical situation, and to have understood the moral significance of her behaviour’ (Fricker Reference Fricker2016, p.168). Grave injustices will be among these cases, but they also include more mundane transgressions like neglecting to tell the staff that they did not charge me for the cocktails I’d had. In these cases, accusatory anger can be considered apt, provided that it is also proportional to the fault and properly motivated by the fault.

Whether apt accusatory anger is intrinsically valuable is contested (Paytas, Reference Paytas2022; Srinivasan, Reference Srinivasan2018). I will bracket the debate about the intrinsic value of apt accusatory anger. I also put aside some other ways in which accusatory anger may be valuable. Anger may be a source of energy and empowerment, it may be a source of knowledge, it may help mobilize people to combat injustice, or it may be good for other reasons (See Carman, Reference Carman2022; Cherry, Reference Cherry2021; Lorde, Reference Lorde1981; Srinivasan, Reference Srinivasan2018; Tessman, Reference Tessman2005; Thomason, Reference Thomason2024). But note that even if apt accusatory anger is valuable, this need not in itself provide a person with sufficient reason to maintain and express their accusatory anger to the person who angered them. I here focus on the additional reasons that would speak in favor of expressing accusatory anger towards the person one is angry with.

Accusation as education accounts argue that accusatory anger is an important conversational move or form of address that helps make the blamee genuinely or fully understand their wrongdoing from the perspective of the speaker. They furthermore maintain that the cultivation or scaffolding of moral agency is a typical or ‘normal’ downstream effect of this process (Fricker, Reference Fricker2016; McGeer, Reference McGeer, Shoemaker and Tognazzini2014; McKenna, Reference McKenna2012; Shoemaker, Reference Shoemaker2015; M. Vargas, Reference Vargas2013). According to, for example McGeer, accusatory anger plays an essential role in sustaining the capacities that make for responsible agency, which on her view provides a plausible defense for treating the apt targets of anger as being deserving of these responses (2018, p. 301). Manuel Vargas presents a similarly optimistic notion of our responsibility practices in his book Building Better Beings (2013). According to Vargas, our praising and blaming practices may ‘for the record’ be justified on the basis of backward looking considerations (aptness conditions), but find their true justification in the fact that blaming others enhances their moral considerations-sensitive agency (M. R. Vargas, Reference Vargas2015, p. 2621).

This take on the effects of reactive attitudes is appealing to a number of contemporary compatibilist approaches to responsibility. New consequentialist approaches to responsibility have now found a forward-looking justification for blame, whilst they can avoid some of the earlier objections to their position, like the wrong kind of reasons problem. They concur with deontologists that the reasons for being angry are the person’s deservingness (the aptness of their anger), but find a forward-looking justification for that in the productivity of this anger.

The purported communicative and cultivating value of reactive attitudes is also appealing to those compatibilists who do maintain that apt accusatory anger has intrinsic value. When the worry arises that the value of an aptly expressed reactive attitude is not, in itself, a very strong or overriding reason for getting angry with someone, the communicative and cultivating value of these attitudes helps further vindicate our blaming practices (Fricker Reference Fricker2016). To fully justify our practices of expressing apt anger, such compatibilists add that apt accusatory anger has communicative and cultivating effects. This is how, for example, Duff’s communicative approach to criminal justice conceptualizes punishment as a form of appropriate address that combines backward looking justice with forward looking benefits (Duff, Reference Duff2003).

To summarize, while the aptness of accusatory anger provides a contested or possibly insufficient justification for holding onto and expressing accusatory anger towards the person one is angry with, the purported communicative and cultivating value of accusatory anger would vindicate the practice and alleviates these worries. This highlights the dialectic appeal of accusation as education accounts and might go some way towards explaining their popularity. But if I am right, philosophers help themselves to these purported advantages too easily. Once we move to other disciplines, these assumptions about the effects of accusatory anger on the blamee are not widely shared.

III. A Therapeutic Approach to Interpersonal Discord

Hanna Pickard (Reference Pickard2011) was one of the first to point out that blame, even when appropriate, can be counterproductive when it is communicated or expressed to the blamee. She observes that, within clinical psychiatric settings, blame is detrimental to recovery, even when the service user is in fact responsible for their harmful behavior (Pickard, Reference Pickard2013). In the light of that, clinicians should refrain from affectively blaming service users, and hold them responsible in other ways (Pickard, Reference Pickard2011).

McGeer acknowledges this, but insists that Pickard’s approach of responsibility without blame is only a fitting response for hard cases, and does not apply to ‘psychologically normal’ day-to-day interactions (McGeer, Reference McGeer2018, p. 320). Miranda Fricker similarly refers to Pickard’s approach as an ‘invaluable resource for those situations in which the normal mechanisms have stopped working’ (Fricker, Reference Fricker2016, p. 174). Pickard herself would not recommend a blanket adoption of the clinical stance either. But she provides a different reason for that. She worries that a focus on the other person’s interest at the expense of how we naturally feel may stand in the way of having a ‘real and genuine relationships’ with them (Pickard Reference Pickard2013, p1149).

Nevertheless and more broadly, the therapeutic approach to interpersonal discord that I am interested in cautions against expressions of accusatory anger in general. The worry is that maintaining and expressing accusatory anger towards a person will typically fail to secure mutual understanding and moral growth. Whatever the ‘psychologically normal’ day-to-day interactions in which one can rely on ‘normal mechanisms’ might be, this particular therapeutic approach worries that in those day-to-day interactions too, accusatory anger does not tend to help resolve interpersonal discord and may damage one’s relationship with the other person.

Instead mediators and conflict resolution strategists will often recommend translating such anger into, or redirecting one’s attention to, an explanation of why one feels the way one does in the light of one’s own values and needs (Rosenberg, Reference Rosenberg2015; Burton, Reference Burton1990; Fischer, Reference Fischer1997; Rothman, Reference Rothman1992; Arieli & Abboud Armaly, Reference Arieli and Abboud Armaly2023). Another difference between Pickard’s clinical stance and this approach to interpersonal discord is that the clinical stance is primarily concerned with the recovery of a service user whereas the therapeutic approach aims to attend to the needs of everyone involved. Therefore, the latter approach should not come at the expense of how we naturally feel. On the contrary, the aim is to investigate those feelings so as to better understand and communicate them.

This therapeutic approach to conflict resolution is based on the principles of ‘compassionate communication’ developed by Marshall Rosenberg (Reference Rosenberg2015).Footnote ¹ He worries that blaming those who do not act in harmony with our values renders us less attentive to our own needs, jeopardizes successful communication with the person we blame, and is likely to damage our relationship with them (Rosenberg, Reference Rosenberg2015, pp. 15–25). On this approach, angry accusation is considered to render one less attentive to the reasons behind one’s own anger, and renders it less likely that the other person will genuinely understand those reasons and empathize, which in turn makes it less likely the discord will be resolved. Rosenberg’s account of compassionate communication is widely endorsed, and similar objections to blame can be found in popular contemporary therapeutic approaches to interpersonal conflict. Esther Perel and Brene Brown, for example, suggest that—if one aims is to resolve the discord in one’s relationship, and aims to maintain, restore, or build a good relationship—accusatory appraisals are best bracketed, and both parties should instead focus on acknowledging and understanding their own feelings and needs, and empathizing with the perspective of the other person (Brown, Reference Brown2021; Perel, Reference Perel2021).

The philosopher and the mediation therapist, may at first sight be considered to have different interests: the mediator cares about interpersonal discord resolution. The philosopher cares about something akin to moral agency cultivation, especially in those who fall short of moral norms and expectations. But upon a closer look there is a lot of overlap between the central concerns of each approach.

The mutual understanding and commitment that’s conducive of interpersonal conflict resolution implies an understanding of how the other person felt and why they did so. And it implies an understanding of how to (begin to) relate differently to one another in the light of what has happened and the commitment to do so. This means that interpersonal conflict resolution also facilitates the development of moral agency. A shared understanding of interpersonal discord, makes for the kind of knowledge that can help one avoid similar harm in the future and help one treat one another with more consideration. A good solution to an interpersonal conflict implies that the persons involved will grasp, acknowledge and learn from their own faults.

Sometimes a good relationship is not in the offing. Severe disrespect, harm, or oppression provide one with good reason to walk away from a relationship. But in those cases, educating the other person through angry communicative engagement does not seem advisable either. More generally, in bad relationships, protecting oneself from harm and unjust treatment should take priority over both moral agency cultivation and compassionate discord resolution. Under other circumstances, the aim of resolving interpersonal discord coincides with the aim of moral agency cultivation. In light of this overlap, it seems odd that the therapist and the philosopher come to such diverging verdicts about the value of expressing accusatory anger within interpersonal relationships.

On my reading, the therapeutic approach to discord resolution questions the productivity of angry accusations, without thereby deeming anger to be entirely pointless or irrational, and without denying the existence of (some sense of) accountability. Therapists and mediators see anger as an emotion which has a signalling function, because it alerts us to our own beliefs, values, and needs (Palmieri et al., Reference Palmieri, Fernandez, Cariolato, Kleinbub, Salvatore and Gross2022; Rosenberg, Reference Rosenberg2015, p. 144; Saccaro et al., Reference Saccaro, Giff, De Rossi and Piguet2024). As such it provides us with an opportunity for better self-understanding. When one feels angry with someone it is important to allow for and investigate this feeling without being swept away by it, or repressing and avoiding it (Blackledge & Hayes, Reference Blackledge and Hayes2001; Rosenberg, Reference Rosenberg2015, pp. 25–49, 148–149). Though the anticipated result is that the anger then also subsides, this investigation is primarily valuable because it provides one with a better understanding of what one perceived and why one is angry. By asking oneself how one’s anger is tied up with what one values and needs, one gains better self-knowledge (Rosenberg, Reference Rosenberg2015, pp. 49–67). This in turn can be used to get clear on what one would request from the person one was angry with (Rosenberg, Reference Rosenberg2015, pp. 67–91). So, angry feelings are a valuable part of compassionate communication and can be used for good.

Therefore, this therapeutic approach is different from the anti-emotion concern discussed by McGeer in her paper ‘Civilizing Blame’ (Reference McGeer, Coates and Tognazzini2013). The anti-emotion concern suggests that a shared normative review of wrongdoing is best carried out without the coloring of angry (or any) emotions (McGeer, Reference McGeer, Coates and Tognazzini2013, p. 181). But the therapeutic approach does leave room for one’s emotions and even for some forms of expressing them. Feelings matter, and the entire approach aims to acknowledge and better understand them by means of attending to them and talking about them in a particular manner. This also distinguishes the approach from Pickard’s clinical stance, which recommends a more detached response (2013, p1145).

But the way in which this therapeutic approach makes room for investigating and communicating feelings, does not encourage sustaining and expressing an accusatory appraisal (Rosenberg, Reference Rosenberg2015, pp. 141–161). Accusatory anger lashes out and places the fault with the other person. This particular accusatory appraisal is what, on my understanding of it, the therapist would ask you to bracket and investigate so as to arrive at a more nuanced understanding of how your perception, values, and needs explain why you feel angry with the other person. The aim is not to replace one’s feelings with a detached response without any emotional coloring. The aim is to be less reactive and to pause and investigate so as to make sure you and your interlocutor can more constructively respond to the needs and feelings you have (Rosenberg, Reference Rosenberg2015, pp. 148–154).

One might, like McGeer, worry that it isn’t feasible to always regulate one’s accusatory anger in this manner. I agree. But the therapeutic approach can accommodate this worry. In those situations, it would be better to engage in a more empathetic exchange at a later point in time. One might admit that one is too angry to talk and ask to address the discord later. And even if one does happen to lash out, a discord can still be resolved at a later point by acknowledging this all too human, but not very productive, reactive response, and by engaging in a more responsive and compassionate form of communication. It should also be noted that, on the receiving end of accusatory anger, one can try to translate such angry accusations into what would be a more compassionate form of address so as to more constructively respond and resolve discord (Rosenberg, Reference Rosenberg2015, pp. 91–113).

Of course, all of this is conditional on the aims of resolving the discord at hand and maintaining, restoring, or building a good relationship. I am open to the possibility that accusatory anger may be valuable for other reasons. And there may be circumstances in which these other reasons take priority. One would hope that achieving, say, social justice related aims can be made to be compatible with the principles of compassionate communication. But there can be tragic cases in which it is not feasible to do so or reasonable to request this (also see Lorde, Reference Lorde1981; and Tessman, Reference Tessman2005).

Furthermore, on my reading of it, this therapeutic approach to interpersonal conflict-resolution does not deny the existence of (a sense of) accountability. This distinguishes it from traditional blame skepticism(e.g. Pereboom, Reference Pereboom2009). On the therapeutic approach, every person is accountable for managing their own feelings and needs in ways that are compatible with those of others. Compassionate communication stresses this responsibility and one envisioned outcome of a compassionate exchange is for the speaker and receiver to take responsibility (Rosenberg, Reference Rosenberg2015, pp. 49–67, 154–159). As a speaker, one becomes better aware of one’s needs and feelings by rephrasing one’s own anger, which facilitates taking accountability for those feelings and needs. You take accountability for your own feelings and needs when you understand that it is up to you to respond to them and to consider when, how, and to what extent it is reasonable to request from others that they take these feelings and needs into account (Rosenberg, Reference Rosenberg2015, pp. 57–61). One possible outcome of a compassionate conversation may be that you should do so differently.

Compassionate communication also aims to make it clear to the person one was angry with that they have a responsibility to think about how they should manage their own feelings and needs in a way that takes the needs of others into account. For example, imagine a friend did not return a loaned sum of money on the agreed upon date, and did not contact you about this either. You may tell your friend that when they did not keep their promise to you, you felt angry because you need trust and transparency in a friendship. You may share that you now struggle to trust them and request from your friend that they keep their promises or take responsibility when they find out they cannot meet them. Your friend may respond by sharing that they were scared or overwhelmed when they found out they could not pay you back, which may help you understand why they did not return the money or contact you. But this insight should also help your friend understand that they were attending to their own needs in ways that came at the expense of your needs and that they should regulate them differently. So, understanding and taking accountability is one envisioned outcome of compassionate communication.

To sum up, the therapeutic approach to interpersonal discord does not deny the existence of accountability nor imply that anger is pointless or irrational. The approach is critical of holding onto and expressing reactive attitudes within interpersonal relationships for other reasons. The critique is that 1) holding onto and expressing accusatory anger obstructs discord resolution, and does not help to develop or scaffold moral agency. And 2) even if angry accusation were to lead to moral agency cultivation, compassionate communication would be a more beneficial way of achieving these same aims.

Why do accusation as education accounts, and therapeutic approaches to interpersonal discord arrive at such divergent verdicts about the value of accusatory anger? I believe much of this disagreement is explained by distinct empirical assumptions about human psychology.

IV. Accusation as Education Accounts from the Armchair

On accusation as education accounts, accusatory anger is understood to be a form of moral address that alerts the addressee to relevant moral considerations and moves the addressee to, from now on, better respond to those moral considerations. Like any other skill, the responsiveness to moral reasons develops by means of practice or repetition, and can get rusty or decay through disuse (McGeer, Reference McGeer2018, Section 4). This means that it takes work to sustain such capacities. And the responsiveness to moral reasons requires specifically social feedback in order to be sustained and developed, just like, for example, language skills do (McGeer, Reference McGeer2018, p. 312).

So far so good, but how does accusatory anger enter into this picture? On these models, reactive attitudes (like anger) make for the sort of social feedback that helps sustain and develop the reason-responsiveness skills of the wrongdoer. They attribute fault, and call on the wrongdoer to justify, explain, or apologize in response to this accusation (Fricker, Reference Fricker2016; McGeer, Reference McGeer, Langdon and Mackenzie2011, Reference McGeer2018; McKenna, Reference McKenna2012; Shoemaker, Reference Shoemaker, Cherry and Flanagan2018). A blameworthy wrongdoer is, on these accounts, already sufficiently sensitive to the moral reasons at stake, but this sensitivity is at the same time, maintained and strengthened by means of accusatory anger. In this way moral understanding and self-guidance involves and is upheld by a ‘susceptibility to the scaffolding power of reactive attitudes, experienced as a form of moral address (McGeer, Reference McGeer2018, p. 315).’

One might at this stage wonder how a sensitivity to moral reasons is supposed to be improved by accusatory anger. If a wrongdoer can, on some level, already grasp the reasons at stake, why not simply remind them of those reasons and sustain their responsiveness in that way? Why add angry accusation into the mix? By comparison with the language analogy: an ability to speak is sustained and developed by means of, for example, correcting a mistake, reminding a person of words they may be looking for, or by simply illustrating how to communicate something. So, wouldn’t similar responses (see for example Zheng, Reference Zheng2021) be equally or even more effective ways to sustain and cultivate a sensitivity to socio-moral reasons?

The typical line of response to this question is that a person will be more motivated to do the right thing if responses to wrongdoing are somewhat unpleasant. Accusatory anger sustains and develops our sensitivity to moral reasons via or combined with our sensitivity to our reputation or standing in the eyes of others, a fear of rejection by others, or just our general need ‘to be liked’ by others (McGeer & Pettit, Reference McGeer, Pettit and Shoemaker2015; McGeer, Reference McGeer, Langdon and Mackenzie2011; See also Bagley, Reference Bagley2017). The psychological mechanism at work when one is on the receiving end of accusatory anger should then be understood as a sensitivity to the social sting that goes hand in hand with blame, which in turn feeds into an improved sensitivity to moral considerations.

Accusation as education accounts aim to provide a functional and also naturalistic account of blame as moral anger. The account is considered to be naturalistic because ‘the attitudes we experience towards one another in the light of normative transgressions, have been shaped by selective pressures because of their aptness in performing the primary function of (directly) eliciting better norm governed behavior from conspecifics’ (McGeer, Reference McGeer2018, p. 320). In other words, accusatory anger is an evolved response to norm-violations which renders human communities more compliant and cooperative, which in turn enhances their chances of survival. An evolutionary approach could indeed suggest that angry responses elicit compliance and help secure norm-governed behavior. But the problem is that there is no evidence that this also feeds into moral agency cultivation. Accusatory anger may help secure compliance within an in-group, but it is one step further to say that it tracks the right moral norms and secures compliance for the right moral reasons. More generally, an evolutionary argument for accusatory anger’s cooperation inducing function would not generalize to exchanges between in-group and out-group members, which are common in contemporary society(Haidt, Reference Haidt2012). McGeer admits that her story is speculative at this point and falls short of a fully worked out account (McGeer, Reference McGeer2018, p. 315).

Yet, there is good reason to believe that accusatory anger in fact fails to naturally track the right moral norms for the right moral reasons. There is ample evidence that humans demonstrate a strong bias for the norms of an in-group, and tend to punish out-group members more harshly for similar violations (Yudkin et al., Reference Yudkin, Rothmund, Twardawski, Thalla and Van Bavel2016). This suggests that the norms that accusatory anger tracks, and the ways in which accusatory anger secures compliance with them, are biased at best. Others have pointed out that reactive attitudes are subject to power dynamics and can perpetuate oppression (Ciurria, Reference Ciurria2023; Mackenzie, Reference Mackenzie2021).

It is also far from clear that the recipients of accusatory anger respond to it for the right reasons and in the ways that are suggested by McGeer. Fricker partly recognize this when she warns that blame is a social power that can also be used for ill. She highlights that blame may motivate people to respond to bad reasons for action. For example, a religious fundamentalist patriarch may blame his daughter for wanting an education. It is conceivable that the daughter wants to avoid this blame and remain in her father’s good graces. She may even end up rationalizing this desire by claiming that the ‘likes of them are not due an education’ (Fricker, Reference Fricker2016, p. 182). Accusatory anger may move people to do what someone wants them to do because they want the other person to like them, where this process is not mediated by moral reasoning or is at best only rationalized by it.

But Fricker, McGeer, and others nevertheless maintain that emotional blame paradigmatically tracks culpable moral transgressions (is apt) and moves persons to do what’s right for the right reasons (Fricker, Reference Fricker2016, p. 181). Yet why would we believe this positive function to be the typical or common one? On an evolutionary approach, it would make sense to think of angry accusation as paradigmatically self-serving or in-group serving, and therefore biased. Second, even if accusatory anger were to be appropriate most of the time, it may tend to move the addressee to do what is right for bad reasons like fear or a wish to please. These worries resonate with the therapeutic approach, which provides us with a different account of human psychology and interpersonal discord.

V. The Psychology of Interpersonal Discord

The therapeutic approach that is central to this article maintains that accusatory anger obfuscates a full understanding of the discord at hand and typically fails to motivate someone to do better, or will only motivate them to do so for bad reasons. This means that the co-reactive exchange presented by accusation as education models is an idealized or unrealistic understanding of how human beings resolve discord and achieve mutual alignment.

A first worry about the angry accusation model is of an epistemic nature: accusatory anger obfuscates (or at least does not facilitate) a thorough understanding of the discord at hand. To illustrate this point, let us consider Fricker’s example of finding fault with your neighbor for not walking your dog as promised when you were away for the weekend (2016). Upon learning this, you find yourself in a state of accusatory anger. The therapist worries that relishing in the accusatory appraisal that is central to it may mean missing out on some of the things we might learn from this anger. It may be that you are angry because you need your dog to be safe and feel responsible for him. You may also be angry because you feel disrespected by your neighbor. Or you may be angry with yourself for entrusting your dog to someone who is not reliable (or some combination). Reflection on how your own anger is connected to what you value and need (as above) would help you acquire self-knowledge. It would also help to ensure that the feeling of anger stays within its remit and that one understands it in a more fine-grained way—a way that facilitates effective emotion regulation and communication (L. F. Barrett et al., Reference Barrett, Gross, Christensen and Benvenuto2001; Blanke et al., Reference Blanke, Bellingtier, Riediger and Brose2022; Starr et al., Reference Starr, Hershenberg, Shaw, Li and Santee2020). One is less likely to acquire this fine-grained self-understanding when in a state of other-directed accusatory appraisal (Also see Pettigrove, Reference Pettigrove2012).

Second, and more importantly for our purposes here, when you do decide to accuse your neighbor, this form of address tends to compromise your neighbor’s thorough understanding of the discord as well. Upon being accused of neglecting your dog, your neighbor may understand that you are angry with them and even that they should have kept their promise. But other than that, this accusation does not facilitate attention to further relevant details of the discord, and may even draw attention away from them.

In contrast, compassionate communication does direct attention to these details. To illustrate, in order for the neighbor to understand their own sociomoral failure in a way that can lead to future improvement, they need to understand why they failed to keep their promise to walk the dog. They need to get insight into how they—in doing so—failed to manage their own feelings and needs in a way that is compatible with those of others (you and the dog). Now, the neighbor may have failed to walk your dog because they struggle to say ‘no’ and over-commit as a result, or because they got distracted and simply forgot, or because they were scared of the dog. These reasons matter, not because they excuse your neighbour, but because they help your neighbor understand how they can and should go about changing their behavior in the future. Yet if you ring your neighbor’s doorbell and angrily say ‘how could you neglect my dog like that?!’ or simply freeze them out, or in some other way accuse them of fault, your neighbor may not be helped to acquire this deeper understanding of their own actions. The sting of this accusation will alert them to a risk of social rejection, or punishment (Gausel & Leach Reference Gausel and Leach2011) but does not facilitate an understanding of why they failed to walk your dog and how they may avoid doing so in the future.

Another reason to be sceptical about the communicative and capacitating effects of accusatory anger is provided by research about the motivational impact of blame. Where accusation as education accounts believe that the ‘sting’ of communicative blame helps to make people more responsive to moral reasons, the therapeutic approach worries that this social sting is more likely to have the opposite effect. Pickard points out that in clinical contexts, blaming attitudes commonly trigger feelings of rejection, anger, and self-blame, which in turn trigger disengagement, distrust, relapse, and possibly self-harm (Pickard, Reference Pickard2013). As I mentioned before, McGeer and Fricker do not believe these responses generalize outside of clinical contexts and refer to them as psychologically recalcitrant, or hard cases. But there is good reason to believe that these responses are more common than they think.

Matt Stichter writes that when we are confronted with the claim that we have engaged in some form of wrongdoing, the distress prompted by this ‘can often cause defensiveness, rather than attempts to redress the wrong and work on changing oneself for the better’ (Stichter, Reference Stichter2020, p. 341). He, like McGeer, argues that moral self-regulation is a skill but adds that learning from one’s own moral failures does not come naturally to us (2020). The cognitive dissonance that arises when a person is confronted with moral criticism can go either way; one might correct for it or disengage from this criticism so as to avoid self-castigation (Gausel & Leach, Reference Gausel and Leach2011; Lorde, Reference Lorde1981; Stichter, Reference Stichter2007).

There is evidence that people are not capacitated by moralized blame. Lee et al. demonstrate that moral evaluations of the behavior of others induce guilt, which in turn diminishes perceived coping-abilities and further intensifies guilt. They also found that by comparison, competence based evaluations have a more empowering effect (van der Lee et al., Reference van der Lee, Ellemers and Scheepers2016). Other studies confirm these or similar findings (Does et al., Reference Does, Derks and Ellemers2012; Ellemers, Reference Ellemers2017; Rösler et al., Reference Rösler, van Nunspeet and Ellemers2021; Snoek et al., Reference Snoek, McGeer, Brandenburg and Kennett2021).

Therefore, accusatory anger may commonly fail to have the effects that accusation as education accounts envision it to have. Furthermore, withdrawing or disengaging in response to blame and anger is equally consistent with an evolutionary explanation. The explanation would be that blaming attitudes do not only function to enforce cooperation or compliance but also function to reject and exclude those conspecifics with whom one cannot cooperate. Similarly, for the recipient, blame and anger may prompt this person to avoid the blamer and to seek out conspecifics with whom one can better cooperate.

That accusatory anger also functions to exclude uncooperative conspecifics would explain blame’s polarizing ‘you are either with us or against us’ dynamic that we see in online and political discourse. And even if accusatory anger has both an exclusive and inclusive effect, this dual nature renders it less reliably productive than non-violent forms of communication. Compassionate communication pre-empts the excluding outcome by decidedly engaging the other person as an in-group member with whom one plans to continue to cooperate, starting with the conflict at hand.

All in all, the therapeutic approach finds support in empirical evidence where accusation as education accounts, to the best of my knowledge, do not. This means there is good reason to doubt that angrily accusing those whom we perceive to do wrong will have the purported communicative and moral agency cultivating effects on them. Note that some of these worries may generalize to non-emotive accounts of fault-finding. Footnote ²

VI. A Possible Objection: Talking Past Each Other

Across different disciplines concepts can be used in different ways. For example, therapists often urge someone ‘not to judge’, or to be ‘less judgmental’. I take it, they mean to suggest that one should not form an opinion about others too quickly, and on the basis of insufficient information. Their advice is also to pause before one does so, or to change one’s object of focus altogether. Of course, what they seem to really mean, if you ask me (as a philosopher), is that one’s judgments about others should be careful and well informed, and that there is a place and a time for them. Judgment entails attributing predicates to a subject, and it would be very hard, let alone inadvisable, to entirely refrain from doing so!

Similarly, a therapist and a philosopher may be talking past each other when they discuss the nature and value of blame. In a footnote, McGeer mentions that persons who are psychologically recalcitrant to reactive attitudes should be, in her words, blamed in the way that Pickard proposes (McGeer, Reference McGeer2018, p. 320). This is puzzling, because Pickard (Reference Pickard2013, p. 1142-1146) suggests that those persons should not be blamed. This suggests a conceptual disagreement about the sort of address that should count as blame, where McGeer’s conception is a permissive one. In this same vein, a possible objection to my argument could be that compassionate communication is in fact just another way to express or communicate one’s own reactive attitudes to a person. Think of a person who would address someone whom they were angry with in one of the following ways:

Although these forms of address align with the principles of compassionate communication, one could argue that they communicate why you have the reactive attitudes that you do. And in some sense the emotions one feels and shares here could be conceived of as an accusatory attitude. Maybe this is in line with Fricker’s suggestion that one may soften an accusation to better bring home one’s point (2016, p 172). So, are compassionate forms of address just a specific way of communicating or expressing one’s own reactive attitudes?

There is one reason to think that these examples of compassionate communication are not an accusation: when one addresses a person in the above manner one does not (yet) assume one’s own appraisal to be accurate. Maybe the person arriving late had good reasons to do so and did not fail to respect the speaker. Similarly, maybe one hears aggression where there is none. These forms of address are not meant to be a fault-finding verdict.

Instead, they are reports of how one feels and why, which then allow the interlocutors to investigate how they should manage their emotions and needs in the future. Fault finding may be a shared outcome of this conversation when both parties agree on the blameworthiness of the addressed person, but it need not be.

Accusation as education accounts might accept that non-violent forms of address are among the sorts of expressions they had in mind as opening moves in a scaffolding and capacitating moral dialogue—what McGeer calls—a ‘co-reactive exchange’. However, this would require an alteration of their position. For, it is not plausible to refer to these non-violent opening moves as ‘blaming someone’ or as ‘expressing a reactive attitude’ because they are a form of self-report that exactly involves the bracketing or rephrasing of accusatory appraisals. Furthermore, it is not the emotion of accusatory anger in itself that secures the claimed morally productive effects.

So, to the extent that accusation as education accounts maintain that angry accusations are in and of themselves capacitating, there remains a genuine disagreement with the therapeutic approach discussed in this article. To illustrate, on these accounts, phrases like ‘how could you be so selfish!’, ‘you do not respect me!’, ‘this is your fault!’, or a furious glance which says as much without words, would express an attitude that typically helps to scaffold and capacitate the addressee. The therapist would disagree, and this disagreement is not just a verbal one but relies on fundamentally different empirical claims about how human psychology usually works.

VII. Continuing the Debate

The empirical evidence discussed in this article is far from conclusive. It, for example, focuses on short-term effects. The long-term effects of accusatory anger might be different. It is also an open question whether the undesirable effects of accusations on the blamee generalize to bystanders and third parties or not (Moody & Nojoumian, Reference Moody and Nojoumian2024) and whether they generalize to the institutional level. These further considerations concern empirical questions about human psychologies that cannot be resolved from the armchair alone.

Questions about the effectiveness of blame and accusatory anger, and about how to scaffold and capacitate the moral skills of others are of an empirical nature. This article hypothesizes that the value of accusatory anger does not lie in its purported moral agency cultivating effect on the accused party. And, pending further research, the available evidence also suggests that the therapeutic approach can do a lot of the work that accusatory anger is purported to do, without the added risk of obfuscating mutual moral understanding, enforcing compliance for the wrong reasons, and polarizing the other party.

The ethical implications of this hypothesis need further attention. Questions that arise are: when would engaging in compassionate communication be something we may expect of a person, and when would it be supererogatory or even inadvisable to do so? When your neighbor fails to walk your dog as promised, you may object it is not your job to initiate or accommodate compassionate communication and thereby cultivate moral agency. Although it may be nice of you and even ultimately good for yourself too, it may not be something we ought to expect of you. But it is unclear of whom we may expect that they initiate and accommodate compassionate discord resolution and why. What would that depend on and how to allocate these therapeutic tasks in ways that are fair and feasible? This question needs to be addressed, with special attention to how duties of care are currently disproportionately allocated to members of already disadvantaged groups.

Further ethical analysis is needed with regards to combatting social injustice and escaping one’s own oppression through anger. The therapeutic approach to accusatory anger does not deny that anger can have value. But when anger features an accusatory appraisal, the worry is that communicating such anger alienates the blamee and complicates mutual understanding, discord resolution, and fails to cultivate their moral agency. This would also contribute to polarization and in-group versus out-group thinking. But would it be feasible to rise against one’s own oppression and combat social injustice without sustained accusatory anger? And can anger can be the apt and beneficial response to injustice it is considered to be, without also featuring an accusatory appraisal and the expressions thereof (Bell, Reference Bell and Tessman2009; Cherry, Reference Cherry2021; Silva, Reference Silva, Falcato and da Silva2021)? If so, how might one ask of a person to redirect or translate the accusatory appraisal, without resorting to a problematic form of victim blaming (McRae, Reference McRae, Cherry and Flanagan2017; Srinivasan, Reference Srinivasan2018)? In order to take this last worry seriously, one should at least prioritize the encouragement of compassionate communication in the agents of injustice, before asking this of the victims of injustice. How might we organize our institutions and relationships in ways that would facilitate this? If my hypothesis is correct, these questions become pressing and merit (even) more academic attention.