The idea that anger is a central threat to decent human interactions runs through the Western philosophical tradition—as do various claims about its usefulness and value. Nonetheless, recent philosophers, at least, spend little time analyzing the emotion. Typical, and highly influential, are Peter Strawson's reference to a class of ‘reactive attitudes’, including guilt, resentment, and indignation, all of which track the relation of another's will to us (Reference Strawson1968);Footnote ¹ and R. Jay Wallace's highly abstract, albeit valuable, characterization of a class of ‘reactive emotions’ (Reference Wallace1994). But anger is a specific emotion, distinct from disgust and guilt, and it seems crucial to analyze it closely, examining its general cognitive content and distinguishing its varieties.

Agreeing with most traditional philosophical definitions of anger, I shall argue that the idea of payback or retribution—in some form, however subtle—is a conceptual part of anger. I shall then argue that the payback idea is normatively problematic, and anger, therefore, with it. There are two possibilities. Either anger focuses on some significant injury, such as a murder or a rape, or it focuses only on the significance of the wrongful act for the victim's relative status—what Aristotle calls a ‘down-ranking’. In the first case, the idea of payback makes no sense, since inflicting pain on the offender does not remove or constructively address the victim's injury. In the second, it makes all too much sense—payback may successfully effect a reversal of positions—but only because the values involved are distorted: relative status should not be so important. In the process of defending these contentions, I shall recognize a borderline species of anger that is free from these defects, and I shall describe, and recommend, a transition from anger to constructive thinking about future good.

Let me begin by simply stipulating three parts of my framework for which I argue in other writing. Like all the major emotions, anger has a cognitive/intentional content, including appraisals or evaluations of several distinct types. Often, it involves not simply value-laden appraisals, but also beliefs.

Second, the appraisals and beliefs involved in anger are what I call ‘eudaimonistic’: as is the case with all the major emotions, they are made from the point of view of the agent, and register the agent's own view of what matters for life, rather than some detached or impersonal table of values.

Third, anger is typically accompanied by a wide range of bodily changes and subjective feeling-states. But these bodily changes and subjective feelings, though important in their way, have too little constancy for them to be included in the definition of anger as necessary conditions of that emotion.Footnote ² For one thing some anger isn't felt at all, like a fear of death that lurks beneath the surface of awareness.

What is anger's distinctive content? A good starting point is Aristotle's definition. Although it will turn out to be too narrow to cover all cases and varieties of anger, it helps us dissect its elements, as contemporary cognitive psychologists acknowledge.

Anger, Aristotle holds, is: ‘a desire accompanied by pain for an imagined retribution on account of an imagined slighting inflicted by people who have no legitimate reason to slight oneself or one's own’ (1378a31–3). Anger, then, involves:

1. 1. slighting or down-ranking (oligôria)

2. 2. of the self or people close to the self

3. 3. wrongfully or inappropriately done (më prosêkontôn)

4. 4. accompanied by pain

5. 5. and linked to a desire for retribution

By twice repeating ‘imagined’ (phainomenês), Aristotle emphasizes that what is relevant to the emotion is the way the situation is seen from the angry person's viewpoint, not the way it really is, which could, of course, be different.

Anger is an unusually complex emotion since it involves both pain and pleasure: Aristotle shortly says that the prospect of retaliation is pleasant. He does not clarify the causal relationships involved, but we can easily see that the pain is supposed to be produced by the injury, and the desire for retaliation somehow responds to the injury. Moreover, anger also involves a double reference—to a person or people and to an act. To use non-Aristotelian terminology that makes explicit an issue that remains implicit in his discussion: the target of anger is typically a person, the one who is seen as having inflicted damage—and as having done so wrongfully. ‘I am angry at so-and-so’. And the focus of anger is an act imputed to the target, which is taken to be a wrongful damage.

Injuries may be the focus in grief as well. But whereas grief focuses on the loss or damage itself, anger focuses on the act that inflicted the damage, seeing it as wrongfully inflicted by the target. Anger, then, requires causal thinking, and ideas of right and wrong. The damage may be inflicted on the person who, as a result, feels anger, or it may be inflicted on some other person or thing within that person's circle of concern.

From the vantage point of contemporary intuitions, the least puzzling parts of Aristotle's definition are its emphasis on pain and its emphasis on wrongful damage. How exactly does the wrongful act of another cause pain to the self? Well, says Aristotle, the person sees (or believes) that something about which she cares deeply has been damaged. In other words, the item damaged has to be seen as significant and not trivial, and that is why pain is a consequence. This pain is, up to a point, not dissimilar to the pain felt in grief. It tracks the perceived ‘size’ of the damage. Nonetheless, the pain of anger typically also makes internal reference to the (believed) wrongful act of another person: the pain of seeing one's child murdered just feels different from that of losing a child to accidental death. In numerous texts, Aristotle emphasizes that pleasure and pain themselves have an intentional content: the pain, then, is pain at the injury that has (as the person believes) been inflicted. It's that specific sort of pain.

As for wrongful injury: even though we experience frustration when someone inadvertently damages us, we only become angry when we believe (rightly or wrongly) that the damage was inflicted by a person or persons, and in a manner that was illegitimate or wrongful. Contemporary psychologist Richard Lazarus gives the example of a store clerk who ignores a customer because he is busy talking on the phone. The customer will feel wrongly slighted. But if she learns that the reason for the phone call was a medical emergency involving the clerk's child, she will no longer be angry, because she will see that it was legitimate to give the phone call priority (Lazarus Reference Lazarus1991: 219, 223). We aren't always so reasonable, of course, but what matters is how we see the situation: we are angry only if we see the damage as illegitimate or wrongful.

Notoriously, however, people sometimes get angry when they are frustrated by inanimate objects, which presumably cannot act wrongfully. This sort of behavior was reported already by the Stoic philosopher Chrysippus, who spoke of people biting their keys and kicking their door when it doesn't open right away, all the while ‘saying the most inappropriate things’ (Stoicorum Veterum Fragmenta III.478). In 1988, the Journal of the American Medical Association published an article on ‘vending machine rage’: fifteen injuries, three of them fatal, as a result of angry men,Footnote ³ kicking or rocking machines that had taken their money without dispensing the drink. The fatal injuries were caused by machines falling over on the men and crushing them (see Tavris Reference Tavris1982: 164, cf. 72; cf. also Averill Reference Averill1982: 166). Do such familiar reactions show that anger does not require the perception that there is wrongful damage? I see no reason to think this. We irrationally think that we have a right to expect ‘respect’ and cooperation from the inanimate objects that assist us. So we react as if they were bad people, since they clearly are not doing ‘their job’ for us. We quickly realize that this doesn't make sense—most of the time.

More problematic, at least initially, is Aristotle's restriction to ‘oneself or one's own’; surely we may have anger when a cause or principle we care about has been wrongfully assailed or when a stranger is the victim of an unjust aggression. Yes, indeed, we may, but that (claims the Aristotelian) is because in that case the cause or stranger has become part of our circle of concern. In other words, ‘oneself or one's own’ is just a way of alluding to the eudaimonistic structure that anger shares with other emotions. This response seems correct: just as we grieve not about every death in the world, but only the deaths of those who are dear to us, so we get angry not at any and every instance of wrongdoing in the world, but only those that touch on core values of the self. As with other emotions, a vivid episode may jump-start the response by moving a distant object into the circle of concern. If, instead of Adam Smith's tale of an earthquake in China, which jump-starts compassion, we hear a vivid tale of genocide in a distant country, then we may be aroused to anger on behalf of the slaughtered people, even if they were not antecedently of concern. But Smith's point holds: the emotion lasts only as long as those people are of concern to us. If the concern ceases (because, for example, we are diverted by pressing concerns closer to home), so does the emotion.

Far more problematic, at least initially, is Aristotle's reference to a ‘slighting’ or ‘down-ranking’. We immediately associate that emphasis with the values of an honor culture, in which people are always ranking themselves against one another and in which the central case of wrongdoing is a down-ranking. Surely, we are inclined to say, many cases of wrongdoing involve cherished projects without being seen as diminutions of status.

Has Aristotle simply made a mistake here? I shall argue that he has, but not as large a mistake as one might think: he has captured a style of thinking that is very common in anger though not omnipresent.

The narrower sense of oligôria as involving down-ranking proves more explanatorily fertile, however, than we might at first suppose. There is something comical in the self-congratulatory idea that honor cultures are in another time or at least another place (such as, putatively, the Middle East), given the obsessive attention paid by Americans (and perhaps Europeans, too) to competitive ranking in terms of status, money, and other qualities. Empirical psychologist Carol Tavris's wide-ranging study of anger in America finds ubiquitous reference to ‘insults’, ‘slights’, ‘condescension’, ‘being treated as if I were of no account’ (Reference Tavris1982: 72, 94). People remain intensely concerned about their standing, and they find endless occasions for anger in acts that seem to threaten it.

From now on I shall call this sort of perceived down-ranking a status-injury. The very idea of a status-injury already includes the idea of culpability, for, as Aristotle notes, diminution of status is always voluntary: if someone acted accidentally, I won't perceive that as diminishing my status. (Remember the store clerk who had an urgent phone call.) Anger is not always, but very often, about status injury. And status-injury has a narcissistic flavor: rather than focusing on the wrongfulness of the act as such, a focus that might lead to concern for wrongful acts of the same type more generally, the status-angry person focuses obsessively on her own standing vis-à-vis others.

In connection with such injuries, both Aristotle and Lazarus emphasize the relevance of personal insecurity or vulnerability: we are prone to anger to the extent that we feel insecure or lacking in control with respect to the aspect of our goals that has been assailed—and to the extent that we expect or desire control. Anger aims at restoring lost control and often achieves at least an illusion of control (Tavris Reference Tavris1982: 152–3). To the extent that a culture encourages people to feel vulnerable to affront and down-ranking in a wide variety of situations, to that extent it encourages the roots of status-focused anger.

What is anger's aim? The philosophical tradition concurs in holding that there is a double movement in the emotion; this double movement, from pain inflicted to striking back, is so prominent that ancient taxonomies classify anger as an emotion that looks forward to a future good, rather than as one that responds to a present bad—although, once they say more, they acknowledge that anger has both aspects. Aristotle emphasizes that the outward movement characteristic of anger is pleasant, and that anger is in that sense constructive and linked to hope. The imagined retaliation or payback is seen as somehow assuaging the pain or making good the damage.

But how exactly does this work? How does pain lead to the sort of lashing out, or striking back, that we associate with anger in many, if not all, cases? And why would someone who has been gravely wounded look forward with hope to doing something unwelcome to the offender? If we had a noncognitive account of anger, there would be nothing further to say: that is just the way hard-wired mechanisms work. But ours is not that type of account, so we must try to understand this puzzle. For it is a puzzle. Doing something to the offender does not bring dead people back to life, heal a broken limb, or undo a sexual violation. So why do people somehow believe that it does? Or what, exactly, do they believe that makes even a little sense of their retaliatory project?

First, however, we had better make sure that the philosophical tradition is correct in holding that a wish for payback is a conceptual part of anger. It is pretty impressive that so many first-rate thinkers, from Aristotle and the Stoics to Butler and Smith to recent empirical psychologists such as Richard Lazarus and James Averill should agree on this. They have thought long and hard about the concept, and it would be surprising if they had made an obvious error. Still, let us think again. Anger is not the only emotion that contains a double movement. Many emotions involve a backward-looking appraisal of what has occurred, as well as associated action tendencies oriented toward a future goal. Grief contains pain at a loss but also often involves a wish for restoration. The grieved person fantasizes about bringing back the loved one. Despite the fact that this is impossible if the person is dead (rather than lost or merely ill), the fantasy can be very persistent and can organize long stretches of the bereaved person's life. When the person is not, or not known to be, dead, the restoration idea is even more central to grief. Parents of an abducted child often respond with obsessive recreation of the child's room, clothing, etc., and obsessive pursuit of any hope for restoration. As grief runs its course, the fantasy of restoration is typically transmuted into a dream of substitution, which can be enacted by finding a new lover to replace the lost one, having another child to replace the lost one.

But even though these action tendencies are closely associated with grief, it is interesting to observe that no standard philosophical or psychological analysis of grief makes them an intrinsic part of grief, a necessary element in its definition. And this corresponds, I believe, to our usage. We typically think that grief and mourning can take people in many directions, even if restoration is a powerful element in many of them.

Compassion too has an associated future-directed action tendency, which has been the focus of a lot of psychological research. When I feel compassion for a person who is suffering, I often imagine helping that person, and in many cases I do it. C. Daniel Batson's empirical research shows that this tendency toward helping is powerful if the helpful action is ready at hand and not very costly.Footnote ⁴ But that connection is typically understood as contingent and causal, rather than conceptual, even if the causality is pretty robust. I think this is probably correct.

With anger, however, the future-oriented aim is standardly thought to be part of the emotion, something without which there is pain of some sort, but not anger. (Butler holds that anger's internal goal is the misery of our fellow humans.) We must figure out, first, whether this is correct—whether there really is a conceptual connection in this case, and not simply a causal connection as in the others. Second, we must figure out precisely how the pain is connected to the strike-back response.

First, let's be clear about what the claim is. The claim is not that anger conceptually involves a wish for violent revenge; nor is it that anger involves the wish to inflict suffering upon the offender. For I may not want to get involved in revenge myself; I may want someone else, or the law, or life itself, to do it for me. I just want the doer to suffer. And the suffering can be quite subtle. One might wish for a physical injury; one might wish for psychological unhappiness; one might wish for unpopularity; one might merely wish for the perpetrator's future (your ex's new marriage, for example) to turn out badly. And one can even imagine as a type of punishment the sheer continued existence of the person as the bad and benighted person he or she is; that is how Dante imagines hell. All I am investigating here (and ultimately accepting, with one significant exception) is that anger involves, conceptually, a wish for things to go badly, somehow, for the offender in a way that is envisaged, somehow, however vaguely, as a payback for the offense.

So let's investigate this further, considering a range of different cases. And let us start from a basic scenario: Offender O has raped Angela's close friend Rebecca on the campus where both Angela and Rebecca are students. Angela has true beliefs about what has occurred, about how seriously damaging it is, and about the wrongful intentions involved: O, she knows, is mentally competent, understood the wrongfulness of his act, etc. (I choose rape rather than murder in order to leave Angela with a wider range of possible actions and wishes than would typically be the case with murder. And I choose a friend in order to give Angela more latitude about how to position herself.)

Case 1. Angela feels pain at Rebecca's rape. She feels that her circle of concern, what she deeply cares about, has been severely damaged, and she believes, correctly, that the damage was wrongful. She now takes steps to mitigate the damage: she spends time with Rebecca, she makes efforts to support her in therapy, she devotes a great deal of energy to mending Rebecca's life—and thus to mending the breach in her own circle of concern. So far, Angela's emotion appears to be grief and/or compassion, and I think the standard definitions are correct when they suggest that it is not anger, even though the occasion for the grief is a wrongful act. We should notice that in this case the primary focus of Angela's emotion is the loss and pain caused to Rebecca, rather than the criminal act itself, and to that extent her emotion would seem to have Rebecca, not the rapist, as its target.

Case 2. Angela feels pain at Rebecca's rape, etc. She does all the things that she did in Case 1, thus expressing her compassion. But she also focuses on the wrongfulness of the act, and her pain includes a special pain directed at the wrongful act—to some extent distinct from her pain at Rebecca's suffering. This additional pain leads her to want to do something about that wrongfulness. So Angela forms a group to support rape victims, and she gives money to such groups. She also campaigns for better public safety measures to prevent rape. Should we call Angela's emotion anger because it focuses not only on Rebecca's pain but also on the wrongfulness of the act, and has an outward movement aimed at something like a righting of the wrong? It is an interesting case, but I think that we typically would not call Angela's emotion anger. I am inclined to see it as a type of morally inflected compassion—not very different, really, from compassion for one hungry acquaintance that leads me to campaign for better welfare support for all. As in Case 1, the emotion does not have the offender as its target; its target is Rebecca, and other women in Rebecca's position. The offender comes into it only because stopping similar harms is Angela's goal for the future, so she will want to deter or incapacitate O and people like O.

Case 3. Angela feels pain, etc., as in Cases 1 and 2. As in Case 2, she focuses on the wrongfulness of O's act, comforts Rebecca, and she may campaign for general measures to prevent that sort of damage in future. But this time she also focuses on O. She seeks to mend the damage by making the offender suffer. Because her circle of concern is damaged, she wants something to happen to O (whether through legal or extralegal means). Here we finally have arrived at anger as the philosophical tradition understands it: a retributive and hopeful outward movement that seeks the pain of the offender because of and as a way of compensating for one's own damage.

The question now is, why? Why would an intelligent person think that inflicting pain on the offender assuages or cancels her own harm? There seems to be some type of magical thinking going on. In reality, harsh punishment of the offender rarely repairs a damage. Adding O's pain to Rebecca's does not do anything to ameliorate Rebecca's situation, as far as one can see. In a TV interview after his father's murder, Michael Jordan was asked whether, if they ever caught the murderer, Jordan would want him executed. Jordan sadly replied, ‘Why? That wouldn't bring him back’. This eminently sensible reply is rare, however, and perhaps only someone whose credentials in the area of masculinity are as impeccable as Jordan's would dare to think and say it.Footnote ⁵ The fantasy that payback restores is magical thinking, abetted by ideas of cosmic balance that are deeply engrained in many cultures, but not the less irrational for that.

This brings us back to Aristotle's idea of down-ranking, which, it emerges, is a likely abettor of this type of magical thinking.

Case 4. Angela is pained, etc. She believes that O's bad act is not only a wrongful act that seriously damaged someone dear to her, but also an insult or denigration of her. She thinks something like, ‘This guy thinks that he can insult my friend's dignity with impunity, and, insofar as he thinks this, he thinks that he can push me around—that I’ll just sit by while my friend is insulted. So he diminishes me and insults my self-respect’. Here, the connection between pain and retaliation is made through the Aristotelian idea that the damage O has inflicted is a kind of humiliation or down-ranking. No matter how implausible it is to read O's act as a down-ranking of Angela (given that O doesn't know Angela, or even Rebecca), Angela sees O's harm to her friend as an ego-wound that lessens Angela's status. She therefore thinks that lowering O through pain and even humiliation will right the balance.Footnote ⁶

Modern western cultures think this way all the time. In most major sports we find an emphasis on retaliation for injury, and players are thought unmanly if they do not strike back to the extent the rules permit (and a little beyond). Even though it is obvious that injuring one player does not take away the injury to another, it is a different story if one focuses not on injury but on ranking and humiliation: the retaliatory hit is plausibly seen as taking away the humiliation of the first hit. Slighting in the sense of diminution reaches a broad class of cases, even if not all cases, where anger is involved. It is very easy for people to shift mentally from a eudaimonistic concern (this is part of my circle of concern, what I care about) to a narrower status-focused concern (this is all about me and my pride or rank). In such cases, a retaliatory strike back is thought to restore the balance of status, manliness, or whatever. And often it does.

Jean Hampton, whose analysis is very close to mine, puts it this way: if people are secure in their dignity, they won't see an injury as a diminishment; but people are rarely this secure. They secretly fear that the offense has revealed a real lowness or lack of value in themselves, and that putting the offender down will prove that the offender has made a mistake (Murphy and Hampton 1988: 54–9). I feel her account does not cover all the cases: more straightforwardly, people may simply care a lot about public standing, and they can see quite clearly that to be pushed around has indeed diminished that. Even in Hampton's cases, the fear she describes is much more plausible if the value people care about is status, which is easily damaged, than if it is human dignity, which is not.

Now the retaliatory tendency makes sense and is no longer merely magical. To someone who thinks this way, in terms of diminution and status-ranking, it is not only plausible to think that retaliation atones for or annuls the damage, it is often true. If Angela retaliates successfully (whether through law or in some other way, but always focusing on status-injury), the retaliation really does effect a reversal that annuls the injury, seen as an injury of down-ranking. Angela is victorious, and the previously powerful offender is suffering in prison. Insofar as the salient feature of O's act is its down-ranking of Angela, the turnabout effected by the retaliation really does put him down and her (relatively) up.

Notice that things make sense only if the focus is purely on relative status, rather than on some intrinsic attribute (health, safety, bodily integrity, friendship, love, wealth, good academic work) that has been jeopardized by the wrongful act. Retaliation does not confer or restore those things. It's only if Angela thinks purely in terms of relative status that she can plausibly hope to effect a reversal through a strike-back that inflicts pain of some type on the offender. Thus, for example, people in academic life who love to diss scholars who have criticized them and who believe that this does them some good, have to be focusing only on reputation and status, since it's obvious that injuring someone else's reputation does not make one's own work better than it was before, or correct whatever flaws the other person has found in it.

It's clear that Angela need not think that the injury she has suffered is a down-ranking. That is why Aristotle's definition is too narrow. Indeed, in this case it seems odd for her to think in those terms, given that O is a stranger who does not know her connection with Rebecca. But this way of seeing injury is very common, and it is very common even in cases where people are eager to deny that this is really what is going on.Footnote ⁷ That is why Aristotle's definition is helpful.

Suppose Angela does not think this way, but stops at Case 3. Then, insofar as her emotion is anger and not simply some combination of grief and compassion, she does initially wish some sort of bad result for the offender, and she does initially think (magically) that this will set things right, somehow counterbalancing or even annulling the offense. It is human to think this way. However, if she is really focusing on Rebecca and not on her own status-injury, she is likely to think this way only briefly. Magical fantasies of replacement can be very powerful, but in most sane people they prove short-lived. Instead, Angela is likely to take a mental turn toward a different set of future-directed attitudes. Insofar as she really wants to help Rebecca and women in Rebecca's position, she will focus on the responses characteristic of Cases 1 and 2: helping Rebecca get on with her life, but also setting up help groups, trying to publicize the problem of campus rape and to urge the authorities to deal with it better.

One of these future-directed projects may well involve the punishment of O. But notice that insofar as Angela is thinking rationally about what will make the world a better place for rape victims, she will view the punishment of O very differently from the way she viewed it in case 4. There she saw punishment as ‘payback’ or retribution—or, more specifically, as a down-ranking of O, which effected a reversal of positions between her and O: women (and Angela above all) on top, bad men (and O in particular) on the bottom. Now, however, she is likely to view the punishment of O in the light of the future good that could be achieved by punishment. This can take several forms: specific deterrence, general deterrence, and, possibly, the reform of O. But it might also take the form of creating a better society with better educational institutions and less poverty.Footnote ⁸

In short, an Angela who is really angry, seeking to strike back, soon arrives, I claim, at a fork in the road. Three paths lie before her. Either she goes down the path of status-focus, seeing the event as all about her and her rank, or she focuses on payback and imagines that the offender's suffering would actually make things better, a thought that doesn't make sense. Or, if she is rational, after exploring and rejecting these two roads, she will notice that a third path is open to her, which is the best of all: she can focus on doing what would make sense in the situation and be really helpful. This may include the punishment of O, but in a spirit that is ameliorative rather than retaliatory.Footnote ⁹

What is really wrong with the first path, the path of status? Many societies do encourage people to think of all injuries as essentially about them and their own ranking. Life involves perpetual status-anxiety, and more or less everything that happens to one either raises one's rank or lowers it. Aristotle's society, as he depicts it, was to a large extent like this, and he was very critical of this tendency on the grounds that obsessive focus on honor impedes the pursuit of intrinsic goods. The error involved in the first path is not silly or easily dismissed. Still, the tendency to see everything that happens as about oneself and one's own rank seems very narcissistic, and ill-suited to a society in which many aspects of human welfare have intrinsic value. This way of seeing things loses the sense that actions have intrinsic moral worth: that rape is bad because of the suffering it inflicts, and not because of the way it humiliates the friends of the victim. If rape were primarily a down-ranking, it could be rectified by the humiliation of the offender, and many people, certainly, believe something like this. But isn't this thought a red herring, diverting us from the reality of the victim's pain and trauma, which need to be constructively addressed? All sorts of bad acts—murder, assault, theft—need to be addressed as the specific acts they are, and their victims (or the victims’ families) need constructive attention. None of this will be likely to happen if one thinks of the offense as all about relative status rather than injury and pain.

There is an instructive exception. Discrimination, for example, on grounds of race or gender, is often conceived as an injury that really does consist in down-ranking, and there is truth to this, just in this special sense: discrimination involves a denial of a special status of equal dignity, and this status has intrinsic value. But the idea that denials of equal dignity can be rectified by bringing the injurer low is a false lure. What is wanted is equal respect for human dignity. What is wrong with discrimination is its denial of equality. Reversing positions through payback does not create equality. It just substitutes one inequality for another. As we shall see shortly, Martin Luther King Jr. wisely eschewed this way of framing the racial issue.

So the first path, the path of status, makes payback intelligible and useful, but it seems morally flawed. This path converts all injuries into problems of relative rank, thus making the world revolve around the desire of vulnerable selves for domination and control. Because this wish is at the heart of infantile narcissism, I think of this as a narcissistic error, but we can also ignore that label and just call it the status error. If Angela takes the first path, then, payback makes sense, but she commits a (ubiquitous) moral error.

If Angela chooses the second path, by contrast, the path of payback, she remains focused on the intrinsic good of bodily integrity, but thinks the suffering of the offender somehow counterbalances or assuages the damage to that intrinsic good. In focusing on this good, she does not make a moral error, but in thinking that payback helps, she engages in magical thinking, which is normatively objectionable in a different way since we all want to make sense to ourselves and to be rational. If she cares about rationality, she will soon see little point in payback, and she will soon backtrack and shift, very likely, to a third path—a focus on promoting future welfare.

This third path, which I recommend, seems, and is, very Utilitarian, and this may be surprising. But sympathy with the Utilitarian idea of punishment arises as the more or less inexorable conclusion of some thoughts about why anger is problematic—irrational in some cases, morally objectionable (because hooked on one's own status) in others. I began working on anger with little sympathy with Utilitarian views of punishment, having criticized them in print numerous times. I find it hard to avoid the conclusion that Bentham had a deep insight about the defects of his society, suffused as it was with status-consciousness and a virulent payback mentality.

I am hereby renouncing a range of things I said in earlier work about the constructive role of anger, and I am now saying something very radical: that in a sane and not excessively anxious and status-focused person, anger's idea of retribution or payback is a brief dream or cloud, soon dispelled by saner thoughts of personal and social welfare. So anger (if we understand it to involve, internally, a wish for retributive suffering) quickly puts itself out of business, in that even the residual focus on punishing the offender is soon seen as part of a set of projects for improving both offenders and society—and the emotion that has this goal is not so easy to see as anger. It looks more like compassionate hope. When anger does not put itself out of business in this way—and we all know that in a multitude of cases it does not—its persistence and power, I claim, owes much, even perhaps everything, to an underlying competitive obsession, which is the only thing that really makes sense of retribution as ordinarily conceived.

To put my radical claim succinctly: when anger makes sense, it is normatively problematic (focused on status); when it is normatively reasonable (focused on the injury), it doesn't make good sense, and is normatively problematic in that different way. In a rational person, anger, realizing that, soon laughs at itself and goes away. From now on, I shall call this healthy segue into forward-looking thoughts of welfare and, accordingly, from anger into compassionate hope, the Transition.

I have imagined the Transition in personal terms, and there is much more to say about these cases.Footnote ¹⁰ But to clarify further what I mean by the Transition, let us consider a case in which it takes a political form. For it has often been thought (including by me, in many earlier writings) that anger provides an essential motivation for work to correct social injustice. Let us look carefully at just one case, the sequence of emotions in Martin Luther King Jr.'s speech ‘I Have a Dream’ (see my longer analysis in Nussbaum Reference Nussbaum2013: ch. 8). King begins, indeed, with an Aristotelian summons to anger: he points to the wrongful injuries of racism, which have failed to fulfill the nation's implicit promises of equality. One hundred years after the Emancipation Proclamation, ‘the life of the Negro is still sadly crippled by the manacles of segregation and the chains of discrimination’.

The next move King makes is significant: for instead of demonizing white Americans, or portraying their behavior in terms apt to elicit murderous rage, he calmly compares them to people who have defaulted on a financial obligation: ‘America has given the Negro people a bad check, a check which has come back marked “insufficient funds”’. This begins the Transition: for it makes us think ahead in non-retributive ways: the question is not how whites can be humiliated, but how can this debt be paid, and in the financial metaphor the thought of humiliating the debtor is not likely to be central. Indeed, humiliation looks counterproductive, for how will such a debtor be in a position to pay?

The Transition then gets underway in earnest, as King focuses on a future in which all may join together in pursuing justice and honoring obligations: ‘But we refuse to believe that the bank of justice is bankrupt. We refuse to believe that there are insufficient funds in the great vaults of opportunity of this nation’. No mention, again, of torment or payback, only of determination to ensure payment of what is owed, at last. King reminds his audience that the moment is urgent, and that there is a danger of rage spilling over, but he repudiates that behavior in advance. ‘In the process of gaining our rightful place, we must not be guilty of wrongful deeds. Let us not seek to satisfy our thirst for freedom by drinking from the cup of bitterness and hatred . . . . Again and again, we must rise to the majestic heights of meeting physical force with soul force’.

Here, the payback is reconceived as the paying of a debt, a process that unites black and white in a quest for freedom and justice. Everyone benefits: as many white people already recognize, ‘their freedom is inextricably bound to our freedom’.

King next repudiates a despair that could lead either to violence or to the abandonment of effort. It is at this point that the most famous section of the speech, ‘I have a dream’, takes flight. And of course, this dream is one not of torment or retributive punishment but of equality, liberty, and brotherhood. In pointed terms, King invites the African-American members of his audience to imagine brotherhood even with their former tormentors:

There is indeed anger in this speech, and the anger summons up a vision of rectification, which naturally takes a retributive form initially. But King gets busy right away reshaping retributivism into work and hope. For how, sanely and really, could injustice be made good by retributive payback? Only an intelligent and imaginative effort toward justice can do that. This is what I mean by the ‘Transition’.

We notice something else: once the Transition gets underway, there is no room for forgiveness as classically conceived in transactional terms, namely, as a waiving of resentment because of an expression of contrition.Footnote ¹¹ The payback mentality wants groveling. The Transition mentality wants justice and brotherhood. If what we want is a racially just society, it would do no more good for Governor Wallace to moan and grovel than for him to burn in hell: these things do not produce justice, and they are restorative only in the magical thinking characteristic of anger's initial pre-Transition phase. In the Transition, one comes to see that the real issue is how to produce justice. Rituals of forgiveness might possibly be thought useful to this end. But King has no room for them: he wants reconciliation and shared effort.

It is here that I introduce a major exception to my thesis that anger always involves, conceptually, a thought of payback. There are many cases in which one gets standardly angry first, thinking about some type of payback, and then, in a cooler moment, heads for the Transition. But there are at least a few cases in which one is there already: the entire content of one's emotion is, ‘How outrageous! Something must be done about this.’ I shall call this emotion Transition-Anger, since it is anger, or quasi-anger, already heading down the third fork in Angela's road. One might give it some ordinary-language name, such as Hampton's ‘indignation’, but I prefer to segment it cleanly from other cases, since I think a lot of cases of ‘indignation’ involve some thought of payback. So I prefer the clearly made-up term. Transition-Anger does not focus on status; nor does it want, even briefly, the suffering of the offender as a type of payback for the injury. It never gets involved in that type of magical thinking. It focuses on future welfare from the start. Saying ‘Something should be done about this’, Transition-Anger commits itself to a search for strategies, but it remains an open question whether the suffering of the offender will be a strong candidate.

Is Transition-Anger a species of anger? I really don't care how we answer this question. Such special borderline cases are rarely handled well by conceptual analysis. It's certainly an emotion: the person is really upset. And it appears distinct, though subtly, from compassionate hope, since the focus is on outrage and the target is the offender. The person says, ‘How outrageous’, not ‘How sad’, and entertains forward-looking projects focused on diminishing or preventing wrongful acts. What is important is how rare and exceptional this pure forward-looking emotion is. Angry people very rarely think in this way from the start. It is much more common to get angry first and then head to the Transition, than to be there already, focused on social welfare, because the retaliatory instinct is, as Butler observed, deeply human, no doubt through both evolutionary tendency and cultural reinforcement. It is only exceptional individuals who are there already, in major issues affecting their welfare. Such presence of mind typically requires long self-discipline. Thus, one could imagine that King's own emotion was Transition-Anger, while the emotion constructed in his speech, for his audience, is brief (standard) anger and then a turn to the Transition.

How might someone become less prone to the errors of anger, and more likely to make the Transition? Aristotle offers a suggestive insight. He says that the person who manages anger well is likely to be good at sympathetic understanding of the positions and motives of other people. How does this work? The idea is, I think, that seeing the situation from the other person's viewpoint helps steer one toward a balanced focus on harm and correction of harm, rather than toward the often unbalanced wishes and motives of anger. To put things in my terms: if you see the other person's point of view, by that very act you are no longer exclusively focused on your own status, and therefore you are less prone to make the status error. You are also less prone to make the payback error, for you will see the future as one involving other people, and your tendency to think of welfare in general social terms will be assisted.

What good can be said of (garden-variety) anger, in the end? First, it may serve as a signal that something is amiss. Anger embodies the idea of significant wrongdoing targeting a person or thing that is of deep concern to the self. While one could have that idea of significant injury without anger—with, and through, grief and compassion—those two emotions do not contain the idea of wrongfulness, which is anger's specific focus. It is for that reason that Bishop Butler, for all his animadversions against the passion, nonetheless concedes that it is ‘one of the common bonds, by which society is held together; a fellow feeling which each individual has in behalf of the whole species, as well as of himself’ (Sermon VIII). Nor, importantly, do those two emotions contain the thought that something needs to be done, which, as I’ve argued, is a conceptual part of anger. The signal anger sends is pretty misleading, since it embodies an idea of payback that is primitive. So it is a false lead to that extent, and the angry person is always well advised to begin moving beyond anger as soon as possible, in the direction of the Transition. Still, anger can be a useful wake-up call. We see this in King's speech, where he does express anger at the behavior of white America, and urges his audience to feel anger as well, acknowledging the magnitude of the wrongs done and the way in which they affect everyone's well-being. But then he immediately turns the audience away from the payback thought that inevitably surfaces, toward a different picture of the future. Managed by such a skillful entrepreneur, anger can be useful, and King always conceived of his project as active and militant, pitted against complacency. Perhaps it's even more useful in cases where the wrongdoing might have slid along barely noticed, beneath the surface of daily life, and only the emotion directs people's attention to its presence.

Anger can also be a source of motivation. The Greek Stoics were often charged with robbing society of motives to pursue justice by their insistence that anger is always mistaken. This charge is even more pertinent to my own view, since, according to my non-Stoic view of damage and loss, anger is often appropriate in its underlying values: the loss or damage can be major, and something really ought to be done about it. The problem comes with the idea of payback. The payback idea is, I argued, a conceptual part of anger (except in the borderline case of Transition-Anger), and no doubt it is part of what motivates people, at least initially. The intensity of the emotion, and perhaps, too, its magical fantasy of retribution are part of what get people going when otherwise at least some people might simply fail to act (or, without anger's signal, even fail to notice the wrongdoing). So in fact Dr. King acknowledged in other writings.

But once people get going, they had better not follow anger's lure all the way to fantasized retribution. King's audience might have imagined a future of payback, in which African-Americans would attain power and inflict retributive pain and humbling on white Americans. Society abounded with such ideas. King's altogether superior stance was that the Transition is only a heartbeat away, since only cooperation will really solve the nation's problems. Still, anger was a useful motivational step along the road—for a very brief time, and carefully managed.

Anger has a very limited but real utility, which derives, very likely, from its evolutionary role as a ‘fight-or-flight’ mechanism. We may retain this limited role for anger while insisting that its payback fantasy is profoundly misleading and that to the extent that it makes sense, it does so against the background of diseased values. The emotion, in consequence, is highly likely to lead us astray.

Finally, anger may be a deterrent. People who are known to get angry often thereby deter others from infringing their rights. Here one can only say that the way anger deters is not likely to lead to a future of stability or peace; instead, it is all too likely to lead to more aggression. And there are many ways of deterring wrongdoing, some of which are much more attractive than inspiring fear of an explosion.

The tendency to anger is deeply rooted in human psychology. Believers in a providential deity, like Butler, find this fact difficult to explain, given its irrationality and destructiveness.Footnote ¹² For those who do not share Butler's framework, however, it is much less difficult to understand. Anger brings some benefits that may have been valuable at one stage in human prehistory. Even today, vestiges of its useful role remain. As Aeschylus notes, however, forward-looking systems of justice have to a great extent made this emotion unnecessary, whether in personal or in public life. Like Athena's citizens, we are now free to attend to its irrationality and destructiveness, and we should do so, focusing first on intimate personal relations, and then on the political realm.