Outrage and the Bounds of Empathy

On January 18, 2015, in the early hours of the morning, two cyclists on the Stanford campus see a man on top of a motionless, half-naked woman behind a dumpster. The cyclists chase the man off the woman and, when he tries to flee, hold him down until the police arrive. The woman, Chanel Miller, wakes up in the hospital with no memory of what has happened to her. She is sore, her hands are caked with blood, and she finds pine needles in her hair. She is only able to piece together the details of what happened ten days later when she reads about her own assault in a newspaper. She sees herself described as “the unconscious woman,” a nameless victim, whereas the man who assaulted her, Brock Turner, is described as a promising young athlete with dreams of making the Olympic trials (Miller, 2019).

Turner pleads not guilty and is eventually convicted on three felony sexual assault charges. He is given a remarkably lenient sentence of six months in prison with the judge explaining that a longer sentence would have a severe negative impact on Turner; in the end Turner only serves ninety days. Throughout the trial, Chanel Miller faces overwhelming pressure to feel sorry for him and to see him as the victim. And indeed, early in the trial, she shows a remarkable amount of empathy for Turner:

I wanted accountability and punishment, but I also hoped he was getting better. I didn’t fight to end him, I fought to convert him to my side. I wanted him to understand, to acknowledge the harm his actions had caused and reform himself. (Miller, 2019, p. 91)

She describes obsessively worrying about the pain he must be in, she tries to find out whether he has been seeing a therapist, and she acknowledges that she would feel responsible if he were to cause harm to himself or anyone else. A turning point for her is when she sees her sister take the stand and comes to understand how much the defense is trying to build a case on which she is responsible for her own assault:

He knew there was a part of us that was self-conscious, the lingering voices that told us we were wrong. … He found it, hooked into it, injected it, grew it until the guilt was all-consuming. Until we became so inundated with self-blame, so blinded by the pain, we lost the ability to see. It was happening to her and it was happening to me, the two of us fed distorted realities, our words twisted until we became uncertain, discredited, writing ourselves off as flawed and broken. We willingly rammed our heads against the walls, confused, apologizing, unsure of what right we had to speak. (Miller, 2019, p. 187)

In her memoir, she describes in this moment being flooded with rage. She recognizes how much she has come to internalize guilt and blame for her own assault, and she resolves to resist the narrative that is being forced on her. When she eventually writes her now famous victim statement, she describes an anger so overwhelming she has difficulty typing: “I typed furiously until my fingers curled” and “I physically could not handle it, could not funnel my anger into the petty tap-tap of keys, could not worry about syntax when my body was breaking down” (Miller, 2019, p. 221). In the victim statement, she describes feeling an anger so intense she thought one of her organs might implode. The victim statement is addressed to Brock Turner, but she no longer displays any empathy for him. She is no longer interested in helping him understand what he has done, in having him apologize, repair the damage, or reform. Her statement is in the form of a systematic refusal of his narrative of the events. And, in the end, she makes clear the statement is ultimately for the sake of other victims of assault whose stories are not believed: “I hope that by speaking today, you absorbed a small amount of light, a small knowing that you can’t be silenced, a small satisfaction that justice was served” (Miller, 2019, p. 357).

I’m interested in Chanel Miller’s anger and how it differs from the sort of anger most commonly valorized in political contexts, the sort of anger I will refer to as reform anger.¹ Miller’s anger is not a demand for repair or reform. However, it is also not hatred. It is not an attempt to destroy the offending party. In fact, I will argue, Miller’s anger is not really at the offending party at all. Her anger is outrage anger: a kind of anger at the state of affairs in which the offense against her is unintelligible as the offense that it is. I’ll argue that outrage anger has an important role to play in the context of political injustice and that it is often maligned because it is mistaken for reform anger.² Before I defend this view, it will be helpful to have a fuller picture of reform anger by way of contrast.

By reform anger, I have in mind the broadly Strawsonian view of anger as a reactive attitude. For Strawson (2003), emotions like anger are the fundamental way that we hold one another morally accountable; anger treats someone as capable of recognizing that they have done something wrong and capable of reintegrating into the moral community.³ As he argues, reactive attitudes like resentment are “essentially reactions to the quality of others’ will towards us, as manifested in their behavior: to their good or ill will or indifference or lack of concern,” and they “rest on, and reflect, an expectation, and demand for, the manifestation of a certain degree of goodwill or regard on the part of other human beings towards ourselves; or at least on the expectation of, and demand for, an absence of the manifestation of active ill will or indifferent disregard” (pp. 8–9). To see someone as the appropriate target of a reactive attitude is to see them as responsible for meeting the expectations and demands of goodwill that constitute our relationships to one another. We can have reactive attitudes to someone we are directly involved in a relationship with, as a response to the quality of will they manifest to us and in light of the expectations and demands appropriate to the relationship we have with them. We can also have reactive attitudes towards someone we are not directly interacting with, in virtue of the quality of will they manifest towards someone else. All reactive attitudes involve taking what Strawson called a “participant” stance towards others; they involve us taking ourselves to stand in moral community with one another, where our communal bonds are constituted by the demands and expectations we have of one another. As Wallace (2019) similarly argues, although anger can give rise to violent and aggressive behavior, “the core behavioral manifestation of this attitude is not so much the desire for payback as the inclination to confront the target person, to lodge protest against what that person has done, and to focus attention of both parties on the action protested as something that needs to be dealt with” (p. 542).

This broadly Strawsonian view of anger as a reactive attitude has been more recently taken up by Myisha Cherry. Cherry (2019b), who is particularly interested in anger at racial injustice, defends the view that politically valuable anger is not only compatible with love for the abuser but, indeed, also an expression of love. She argues that anger “expresses active concern for the moral improvement of those not only who are resistant to racial injustice, but also those who perpetuate it, or are complicit in it” (p. 165). This sort of anger is concerned not only with the treatment of those who are victims of injustice but also with the moral well-being of the perpetrators of injustice. For Cherry, the kind of anger that is morally valuable is not concerned with destroying the abuser or making them suffer but, instead, wanting them to do better. She argues that moral anger at injustice is a form of criticism, and it is fundamentally constructive, not destructive: “it shows that the object of anger has work to do” (p. 165).

Even critics of anger like Nussbaum (2016) and Flanagan (2016) acknowledge a kind of anger, which they refer to as transitional anger and righteous indignation, respectively, that is an important and apt response to injustice. And even Callard (2017), who is critical of what she describes as the “overly sanguine picture” of anger defended by others, defends a view of anger where anger is a response to the violation of the norms of a relationship and seeks a return to the co-valuing of the relationship. For her, anger is not so much an attempt to repair a damaged relationship but an expression of its being in need of repair. Once there is a shared understanding of the wrongdoing that occasioned the anger, it is possible to move past the anger and return to a co-valuation of the relationship.

The similarity I want to emphasize in all these accounts is that anger is an expression of care for the abuser. Anger as a reactive attitude expresses seeing one another as members of a shared moral community or relationship and seeks to demand some kind of repair or reform. To see anger as an expression of care is to contrast it with what Strawson (2003) describes as the “objective” stance. To view someone in this way is to view them as an object of social policy, or a subject of “treatment.” That is, it is to see them as something other than a full-fledged morally responsible agent; it is to see their behavior as in some way not reflecting the sort of poor quality of will that would justify resentment. Reactive attitudes, even when painful and scary, are a reflection of our continued care and investment in being in moral community with one another.

We might think of the reform anger described above in terms of both its object and its function. The object of reform anger is the person or people who have committed some moral violation and done so as an expression of a poor quality of will. Its function is to hold the abuser to account and to demand that the injustice be addressed. Typically, this demand involves a call for the abuser to apologize, to repair the damage or relationship, and to reform. So, as Cherry (2019b) argues:

If the object of moral anger is change, then anger is never just a call to “leave me alone.” More importantly, it asks “How might we live together?” Moral anger at any political injustice sounds the alarm that a disruption in the community has occurred and expresses the need for repairing the damage. (p. 167)

For these defenders of the political importance of anger, anger is criticizable when it is no longer consistent with benevolence towards the abuser. Baldwin (1998) criticizes anger when it is tied to hate and therefore incompatible with love. Lorde (1997) insists that hatred “is the fury of those who do not share our goals and its object is death and destruction. Anger is the grief of distortion between peers, and its object is change” (p. 282). This sort of anger aims to eliminate the hated person, instead of aiming at reconciliation and change. Moreover, Cherry (2019b) argues, hate is not focused on understanding the abuser and instead focuses only on the abuser’s negative actions and “feeds off false information and ignores the person’s whole character” (p. 164). This sort of anger is incompatible with any empathetic understanding of the abuser and so also incompatible with change or reconciliation and bringing the abuser back into the moral community.

Chanel Miller’s anger seems to differ from reform anger both in its object and its function. The rage Miller experiences is not, centrally, about the assault. Again, immediately after the incident, she is highly empathetic to Turner. She wants him to understand what he has done and reform himself. Perhaps she is angry then, too, but her anger takes on a new character during the experience of the trial as she watches her own narrative consistently dismissed, ignored, or undermined. She becomes angry, not at Turner, so much as that Turner cannot, and will not, understand what he has done. Indeed, the object of her subsequent anger is not really any particular person or set of people. It is, instead, the entire system—the defense attorney, the probation officer, the judge, Turner’s family and friends, the media reporting—that privileges his narrative over hers, that allows women to be sexually assaulted without consequence, and that demands her erasure.

Likewise, the function of her rage is not, principally, to demand that Turner apologize or reform. It becomes clear to her over the course of the trial that he is unlikely to ever understand what he has done. To the extent Miller cares about Turner being held to account, it is only as a symbol of something broader. What Miller cares about is helping to build a world in which the testimony of victims is not dismissed or undermined. Notice though that although Miller’s anger is not the reform anger described by Cherry and others, it is also not indiscriminate hatred. On the one hand, its object is not the person who has harmed her, and its function is not to hold that person to account, to demand repair or reform. On the other hand, it is also not merely aimed at revenge or destruction.

To better understand Miller’s anger, it is helpful to look at a distinction María Lugones draws between two kinds of anger.⁴ Lugones (2003) distinguishes between what she calls “first-order anger” and “second-order anger.” For her, first-order anger is anger that “has a communicative intent and does or does not succeed in getting ‘uptake’ within a particular world of sense” (p. 91), whereas second-order anger presupposes “worlds of sense against which the anger constitutes an indictment or a rebellion, worlds of sense from which one needs to separate. These angers also presuppose or establish a need for, or begin to speak from within, separate worlds of sense. Separate, that is, from worlds of sense that deny intelligibility to the anger” (p. 89). Roughly, first-order anger is something like anger as a reactive attitude; it is a form of protest that demands acknowledgment of a harm done and some kind of repair or reform. First-order anger seeks to communicate, or to make itself intelligible, to the offender. By contrast, second-order anger is anger that is appropriate in contexts where first-order anger is not intelligible. Second-order anger involves a kind of refusal to make itself intelligible in the dominant “world of sense.”

Here, Lugones is drawing on Frye’s (1983) analysis of anger as an “instrument of cartography” (pp. 94–95), where anger can be instructive in part by showing us the domains in which our anger is intelligible. Frye explains that women’s anger is only treated as intelligible, or given uptake, within particular “worlds of sense.”⁵ Lugones wants to suggest that first-order anger is the sort of anger that demands uptake or intelligibility within the “official” world of sense.

Imagine a woman in a fairly progressive workplace who protests against sexual harassment from a co-worker. She recognizes that harassment is not consistent with the respect she deserves, and so she expresses anger, making a claim upon that respect. In Lugones’s framing, she inhabits a world of sense in which she makes sense to herself, in which she is respectable. But the “official” world of sense is one in which she is being denied this respect. If, in expressing her anger, she is believed and the co-worker is disciplined or punished for his misbehavior, her anger is treated as intelligible; it is given uptake within the dominant world of sense. In fact, even if her complaint is treated with defensiveness, or disagreement about what respect involves, her anger is still being given uptake, treated as intelligible to the person to whom she is expressing it. If her boss, instead of immediately disciplining the co-worker, suggests that she mistook her co-worker’s friendliness for a sexual advance, her boss is still treating her complaint as intelligible. Her boss believes that, if she were being sexually harassed, this would be inappropriate. In succeeding in being given uptake, the woman is able to expand the scope in which her anger is intelligible and to bring the “official” world of sense closer to the world she inhabits, in which she is respectable.

However, Lugones (2003) notes, this first-order anger can pose a risk to the person expressing it:

The demand for respectability can become a trap against inhabiting worlds of sense that are entirely antagonistic to the social and collective scheme that determines the limits of the intelligible, because from within them one may not be able to intend to present oneself as respectable to others who are outside of these limits. (pp. 93–94)

If the world in which a woman imagines herself to be respectable is entirely antagonistic to the “official” world of sense, Lugones thinks anger might turn to second-level anger. This anger contains the recognition that there is more than one world of sense. It does not make a claim upon respect in the “official” world of sense: instead, it “decries the sense of the world that erases it precisely since that world of sense stands in the way of its possibility. It recognizes this world’s walls. It pushes against them rather than making claims within them” (p. 93). This anger, which often takes the form of rage, is “a way of isolating her self, of making space for her self” (p. 95).

Imagine the woman described above, but now imagine that instead of her being met with defensiveness or disagreement, she is met with incredulity or ridicule. Imagine, that is, that instead of her anger being given uptake, it is treated as ridiculous or unintelligible. Instead of being told that she is overreacting, or that she misread the behavior towards her, imagine that she is treated like she is crazy for protesting the behavior at all or for believing she might deserve a different kind of treatment.⁶ Here, Lugones seems to be suggesting, the world of sense in which this woman believes she is not merely an object for male sexual pleasure is antagonistic to the world of sense of the workplace environment, in which it appears ridiculous for her to protest her harassment. The woman cannot simultaneously inhabit both worlds: she cannot at once understand herself as being worthy of respect and seek to make herself intelligible to the world in which she is not given respect. She cannot, at once, make sense to herself and attempt to make herself intelligible to a world radically opposed. Second-order anger is a way of closing herself off to the official world of sense, refusing to try and make herself intelligible in this world.

Although, for Lugones, second-order anger does not aim to communicate within the established world of sense, it is also not devoid of communicative content. Lugones distinguishes between the communicative and non-communicative dimension of second-order anger. While it “decries” the dominant world of sense that “erases it,” it also “echoes or reverberates across worlds.” She goes on to explain:

This anger speaks its sense within the official world of sense in enraged tones without the intention to make sense to those within it. Its harshness attests to the hardness of the walls against which it echoes. Its intimidating power indicates that it does echo. Its inspiring power indicates that it does echo. This is a separatist anger. (p. 93)

For Lugones, second-order anger communicates across worlds through a kind of “emotional echoing.”⁷As Medina (2020) puts it, this emotional echoing “calls upon us to recognize the failures of intelligibility within existing worlds of sense and to engage in the collaborative construction of new worlds of sense” (p. 225). In creating distance or estrangement between oneself and the dominant world of sense, second-order anger has an important constructive role. Lugones thinks it has the power to create new worlds of sense in which one does not find oneself unintelligible. Second-order anger, then, has the power to create coalition or community. Lugones echoes Lorde’s (1996 idea that anger requires “peers meeting upon a common basis to examine difference, and to alter those distortions which history has created around our difference” (pp. 129–130).

I want to suggest that Miller’s anger is the second-order anger that Lugones describes. This might not seem immediately obvious. After all, unlike in the case of workplace harassment described above, Miller isn’t ridiculed or treated as insane. In some sense, her testimony is believed. Few deny that she was harmed by Turner, and Turner is ultimately convicted of assault. That is, the dominant world of sense is one in which sexual assault is intelligible as a harm. It is a world of sense in which Turner should have known better and is morally responsible for his behavior. Why, then, does Miller feel her pain is unintelligible?

What Miller experiences is a dominant world of sense that understands her assault and the subsequent trial, in abstraction from patriarchy and rape culture. Her actions, and those of Turner, are understood outside of a context of unequal power dynamics. For example, Turner insists that Miller communicated her consent. Even though no one in the trial denies that Turner penetrated her while she was unconscious, the incident is treated more like an unfortunate misunderstanding than predation. Throughout the trial, Miller is told both implicitly and explicitly that she is partly responsible for her assault, that she should not have been at the party, should not have been drunk, that her making these choices invited the encounter with Turner. What Turner interprets as consent is the fact that Miller at one point rubs his back: “The night after it happened, he said he thought I liked it because I rubbed his back. A back rub. Never mentioned me voicing consent, never mentioned us even speaking, a back rub” (Miller, 2019, p. 345).

Miller’s assault is treated as an unfortunate misunderstanding because it takes place in a background context where women have a limited ability to meaningfully refuse sex, where the slightest gesture of affection is treated as enthusiastic consent, and where any decision they make that increases their vulnerability to assault is understood as an invitation. Miller (2019) also lacks the ability to meaningfully advocate for herself after the assault. As a woman, she is wary of being too assertive when she asks that Turner face consequences for his actions. Moreover, she is genuinely concerned about Turner’s mental health after the assault and the violence he might inflict on other women. She describes how the concern she expresses for Turner is taken as a reason to give him a lighter sentence:

[A]s a woman I’d tried asserting my opinion without coming off as self-serving or overcontrolling. So I repressed pissed-off victim. Now I wondered if I had handled it too gracefully, my composure a signal that what he’d done was of little consequence. When I’d advocated for him to take classes and be in therapy, she mistook it as a nurturing passivity, gentle absolution. What I meant was take note of his mental health, because in my experience, when men were upset, lonely, or neglected, we get killed. (pp. 218–219)

Miller here describes an oppressive double bind.⁸ If she is too assertive—if she refuses to comply with the oppressive norm that women be endlessly gentle and forgiving—she will be dismissed as an angry, vindictive victim. If she is not assertive enough—if she cooperates with the oppressive norm that women be endlessly gentle and forgiving—her concern for Turner will be taken as absolution. No matter what she does, she cannot accomplish her goal of having him be held properly accountable.

Miller (2019) experiences a similar gendered pressure at the sentencing. Throughout the trial, she faces overwhelming pressure to feel sorry for Turner, to imagine what things feel like from his perspective, and to accept his frustration as contrition. At the sentencing, the judge implies that Miller is being unreasonable for not accepting Turner’s remorse as sincere:

And so you have Mr. Turner expressing remorse, which I think, subjectively, is genuine, and Chanel not seeing that as a genuine expression of remorse because he never says “I did this. I knew how drunk you were. I knew how out of it you were, and I did it anyway.” And that—I don’t think that bridge will, probably, ever be crossed. (p. 234)

Like Miller, the judge recognizes that her perspective and that of Turner’s will remain irreconcilable. However, the judge locates the problem in Miller, implying that she is failing to be appropriately compassionate and forgiving.

Part of the irreconcilability of their perspectives is that Turner’s behavior is also interpreted outside of the context of oppression and gendered power dynamics. Turner is understood by the judge as expressing genuine remorse for his actions. But as Miller points out, he never arrives at an understanding of the action for which he is expressing remorse. Turner’s gesture of taking accountability for his actions is to speak out against drinking culture and sexual promiscuity on college campuses. But for him to identify drinking culture and promiscuity as the background conditions that made the assault possible is for him to miss the gendered dynamics at play. What Turner or the judge don’t see is how normalized it is for men to go to college parties with the expectation they can have sex with drunk women and face little consequence. As Turner himself acknowledges, he went to the party with the intention of finding a woman to have sex with and attempted to kiss several women, including Miller’s younger sister, before he found Miller. The background conditions that made Miller’s assault so inevitable are not a culture of promiscuity and college drinking, but rape culture.⁹ Indeed, at the sentencing, the judge reads a statement from a friend of Turner’s that reflects how difficult it is for the dominant world of sense to accommodate the idea that Miller was raped or seriously violated:

I don’t think it’s fair to base the fate of the next ten+ years of his life on the decision of a girl who doesn’t remember anything but the amount she drank to press charges against him. … This is completely different from a woman getting kidnapped and raped as she is walking to her car in a parking lot. That is a rapist. These are not rapists. These are idiot boys and girls having too much to drink and not being aware of their surroundings and having clouded judgement. (p. 273)

In some ways, what Miller encounters is more pernicious than the case I described above of workplace harassment. Miller’s experience is of being gaslit:¹⁰ she is, on the face of it, believed. But the way her experience is understood is dramatically different from the reality of her assault. The way her experience is understood takes what happens to her and distorts it, presenting her with a version of the events where she is responsible, where she should be more forgiving, where Turner’s pain is what gets centered.¹¹ As she says in explaining why she decided to write her memoir, throughout the trial she felt “flattened, characterized, mislabeled, and vilified”; she explains that the “saddest thing about these cases, beyond the crimes themselves, are the degrading things the victim begins to believe about her being” (Miller, 2019, pp. xi–xii).

Miller’s outrage, then, is a response to hermeneutical injustice, that is, the “injustice of having some significant area of one’s social experience obscured from collective understanding owing to hermeneutical marginalization.”¹² As Jenkins (2017) argues, there are three aspects of this definition that are relevant to narratives around rape: the experience is significant, the experience is obscured from collective understanding, and the subject is hermeneutically marginalized (p. 197). The experience of suffering sexual assault or rape is significant. It is also very often obscured from common understanding, and very often even the victim is not in a position to understand their experience as rape. Finally, Jenkins notes, victims of rape are “hermeneutically marginalized by the stigma associated with rape and abuse, including the tendency to blame the victims for their own rape or abuse” (p. 197).

It is worth emphasizing here that the failure of the dominant moral community to fully understand the nature of Miller’s experience is, in Medina’s (2012) language, a blameworthy hermeneutical ignorance. Turner and the dominant moral community exhibit the epistemic vice of actively avoiding external epistemic friction and the resistance of epistemic others (p. 56).¹³ As Medina argues, this active ignorance is “supported by psychological structures and social arrangements that prevent subjects from correcting misconceptions and acquiring knowledge because they would have to change so much of themselves and their communities before they can start seeing things differently” (pp. 57–58). For Medina, active ignorance is a form of meta-ignorance that involves inadequate second-order epistemic attitudes including vicious resistance to epistemic challenge. Individuals who exhibit this kind of willful hermeneutical ignorance do not live up to their epistemic responsibilities in confronting epistemic friction or challenge. The abuser’s individual failing is embedded in a much broader economic, sociopolitical, and cultural set of patterns that keep members of certain groups in subordinate epistemic positions: “the social imagination can, in this sense, be a negative epistemic force that silences subjects and renders certain things invisible, making it very difficult—and in the extreme case, impossible—for subjects to even register epistemic friction, much less handle it responsibly” (Medina, 2013, p. 71).

So far, I’ve described the circumstances in which first-order anger turns to second-order anger. Whereas the object of reform anger is the person who has caused an offense, the object of outrage anger is the state of affairs or the hermeneutical space that makes that offense unintelligible as the offense that it is. Even though Miller’s testimony is believed, in some sense, her assault is not understood in the context of oppression and gendered power dynamics. She is reduced to a nameless victim who is partly responsible for her own assault.

A note on terminology. In what follows, I will generally use the term “outrage” rather than “second-order anger.” The term “second-order anger” is potentially misleading for my purposes, since a second-order state tends to have as part of its object some first-order state. Lugones argues that first-order anger turns into second-order anger under the appropriate conditions; that is, she suggests that second-order anger is always generated from first-order anger. As such, on her view, second-order anger might always have as part of its object the first-order anger from which it originates. I am not committed to this, however. It seems plausible to me that one could feel outrage anger without having felt reform anger first. The term “outrage anger” is itself potentially misleading and should not be confused for what is sometimes referred to as “moral outrage.” Psychologists and philosophers sometimes talk about moral outrage as anger produced at the perceived violation of a moral standard, where this is sometimes contrasted with “personal anger,” which is anger produced when one is harmed. This distinction cuts across my distinction between reform and outrage anger; on my view, both reform and outrage anger can be felt either about a personal harm or about a violation of a moral standard that does not directly harm oneself.¹⁴ Similarly, Cherry uses the term “rage” to refer to the Lordean anger that I have been treating as a paradigm case of reform anger, the sort of anger that, on Cherry’s view, is potentially powerful in fighting racial injustice (Cherry, 2021). These worries aside, and in the absence of a widely shared taxonomy of different forms of anger, I think the term “outrage” here still best captures what I am after: a particularly intense form of anger that involves, in part, a refusal to accept the world as it is.¹⁵

Turn now to the function of outrage anger. Again, Lugones argues that second-order anger functions to decry “the sense of the world that erases it” and that it “pushes against” that world’s walls “rather than making claims within them”; it is a way of “isolating her self, of making space for her self.” And, as Lugones argues, second-order anger makes sense in contexts where there is a serious risk to the person who tries to make a claim of respect in the world of sense in which her claim of respect is not intelligible.¹⁶

We can see this in Miller’s case. After the assault, Miller tries to make her pain understood, to advocate for herself, and to bring Turner to justice. When she tries to make this claim of respect in the dominant world of sense, she is gaslit, made to feel doubt, shame, and guilt. Miller describes anger as what protects her and allows her to assert her own narrative and to refuse the narrative of guilt and blame that is being forced on her. The function of her anger is to close herself off from the dominant world of sense, to refuse to try and make herself intelligible to Turner or others. Her anger functions as a kind of emotional boundary setting.

Notice how this differs from reform anger. I suggested that reform anger is an expression of care—its function is to demand that an offense be acknowledged so that the relationship can be repaired. Second-order anger is not an attempt to repair the relationship; it is an attempt to remove oneself from the world of sense in which that relationship existed. We can think about this in the language of Strawsonian reactive attitudes. Reform anger, as a paradigmatic reactive attitude, is a temporary distancing that has as its ultimate goal that the offender be brought back into the moral community; even when the anger is painful or scary, it is a reflection of one’s continued investment in the relationship. By contrast, outrage anger is a kind of stepping out of the dominant moral community; it is a refusal to try to make a claim of respect within that community.¹⁷ In fact, the attitude that Miller ultimately takes towards Turner is more like the objective stance than the participant stance.¹⁸ She sees him more as an object of social policy, the subject of “treatment,” than as someone who can really come to see and reform his behavior. As Strawson (2003) notes, we might take the objective stance to someone when we seek to exit a moral relationship: such a stance can provide a strategic refuge from “the strains of involvement” in a relationship, and such a stance forbodes the relationship’s likely demise (p. 6). Notice however, that she takes this stance towards Turner only because, in some sense, she has already been pushed out of the dominant moral community. Insofar as the violation against her is treated as unintelligible as the violation it is, she is not given the same respect as Turner; she is not treated as an equal member of the dominant moral community.

The main function of outrage then, is a kind of self-protection function. In Miller’s case, part of what this self-protection function seems to involve is blocking her ability to feel empathy for Turner. Not only does her empathy for Turner ultimately undermine the outcome she wants—the probation officer uses the empathy she expresses as justification for a trivial sentence—it also, more seriously, begins to undermine her own self-conception. Her anger in part involves a refusal to engage with his narrative insofar as doing so undermines her ability to see herself in certain ways. What she learns over the course of the trial is that she cannot believe that she deserves respect, and believe that she was seriously harmed, while also empathizing with his perspective according to which she is, in part, responsible for what happened because she was drunk and according to which what happened to her was just an unfortunate but trivial mistake.

Consider what empathy involves. Much of the recent work on empathy emphasizes, as a key feature of empathy, taking on another person’s perspective. Coplan and Goldie (2011) describe empathy as “a complex imaginative process in which an observer simulates another person’s situated psychological states [both cognitive and affective] while maintaining clear self-other differentiation” (p. 4).¹⁹ Coplan and Goldie identify three essential features of empathy: affective matching, where one takes on the same affective states as the target of empathy; other-oriented perspective taking, where the observer imagines the target’s situation, experiences, and characteristics as though they were the target; and self-other differentiation, where the observer continuously represents themself as different from the target (p. 6).

Stueber (2006) defends a similar view, distinguishing between basic empathy, which allows us to recognize another person’s emotional states, and reenactive empathy, which allows a complex understanding of another person’s behavior via reenacting or imitating in our own mind the thought processes of another person. The latter allows us to see an individual’s complex social behavior as the behavior of a rational agent who acts for a reason (pp. 20–22).

Bailey (2022) similarly takes as central to empathy the ability to imaginatively take up another person’s perspective, to feel their reasons as reasons:

To humanely understand another’s emotion is to have a first-hand appreciation of the emotion’s intelligibility. Securing that appreciation means imaginatively taking up the other’s first-personal perspective and seeing the world as calling for the emotional response that the other is experiencing. And this form of seeing achieved within the context of this imaginative effort, just is empathy itself. (pp. 8–9)

L. A. Paul (2021) describes empathy of this sort as the cognitive analogue to visual representation; by imaginatively representing someone’s situation, their experiences, feelings, and beliefs, you simulate their cognitive “point of view” or what it is like to be them (pp. 4–5). Doing this, Paul argues, allows you to “understand their point of view, in a way that explains their point of view.” However, Paul insists, empathy can have the potential to epistemically corrupt, especially in cases where there is no universal arbiter of truth:

Opening yourself up to receiving a new kind of information or to assessing a new kind of potential evidence might be epistemically dangerous. If so, you might reasonably fear that what you need in order to assess potential evidence could change you in a damaging way. In particular: it could corrupt you, changing your ability to make evidential assessments by your current lights. (p. 2)

It is perhaps no surprise that empathy for Turner would come at the cost of Miller’s seeing herself as having deserved a better treatment: to have a first-hand appreciation of his frustration is to imaginatively take up his perspective, a perspective according to which she is being relentlessly cruel and uncompromising, trying to ruin his life for a stupid mistake for which she is in part responsible. This perspective, although part of the dominant hermeneutical space, is incompatible with seeing herself as she actually is, a victim of a serious violation who is subsequently pressured to blame herself. There is, in Lugones’s terms, too big a gap between the “official world of sense”—the misogynist culture that doubts and blames victims of sexual assault—and the “world of sense” in which she inhabits, one where her pain is an appropriate response to a serious moral harm.

Notice, though, I am not suggesting that Turner does not deserve empathy. We might think that everyone deserves empathy. And we might think that Turner is himself a product of misogyny, shaped by social forces that make it very difficult for him to appreciate what he has done. Indeed, to dismiss Turner as a mere anomaly, an evil actor, is to ignore how pervasive the sorts of beliefs and behaviors he exhibits are. And to ignore this is to ignore the deeper, structural forces that make sexual assault so prevalent, and it is to ignore how damaging the effects of misogyny are on all genders. I am not here saying that Turner is not responsible for what he does to Miller, nor that he should not be held responsible; we hold people responsible all the time for beliefs and behaviors that are largely the product of oppressive forces. But this is compatible with thinking it is rationally permissible to feel empathy for him. After all, such empathy is compatible even with reform anger. I think the intuition many of us will have is that, although it might be reasonable for some people to have some empathy for Turner, Chanel Miller need not, and perhaps should not, feel empathy for him.

What I’m suggesting instead is that even if Turner deserves some empathy, it might not be possible or healthy for Miller to feel it for him. It is better, with a view to protecting her own self-conception, for Miller to be closed to Turner and to any facts about Turner that might warrant empathy. Likewise, in the language of reactive attitudes, it is better for her to adopt the objective stance towards him even if, in some sense, the participant stance is more appropriate; he is morally responsible for his actions and was capable of acting other than he did.

Compare the function of outrage here with the function of forming a resolution in the face of weakness of will.²⁰ A weak-willed academic might form a resolution not to go on social media for the day while she completes a referee report. This resolution involves an intention not to go on social media, combined with a second-order intention not to let that first-order intention be deflected. We are cognitively and emotionally limited beings, and in areas where we are weak-willed, it is often too exhausting, and too risky, to allow ourselves to continually reconsider our intentions. So, too, I want to suggest, Miller is in a situation where she is confronted with her own cognitive and emotional limitations. The dominant hermeneutical space, and the persistent gaslighting she experiences, functions in something like the way that weakness of will might. In the absence of a resolution, weakness of will forces the academic to constantly question her decision not to be on social media. So, also, in the absence of outrage, the dominant world of sense forces Miller to constantly question her own value and whether she was significantly harmed by Turner.

Why, we might wonder, is outrage anger the functional equivalent of a resolution in this case? Why can’t Miller take on a different attitude, something more like apathetic withdrawal? Why isn’t it enough for her to understand the ways that patriarchy is operating and refuse to engage? Again, unlike in cases where reform anger is appropriate, Miller is in a situation where there is very little possibility of her anger being given uptake or treated as intelligible given the entrenched misogyny and victim-blaming narratives in the community she finds herself in. In this context, a cool intellectual understanding that there is injustice isn’t enough because the “official” world of sense the agent inhabits is one that tells her that her experience isn’t authoritative, that she is deluded or crazy or irrational. In a situation like hers, outrage anger might not just be a fitting response to the way the world is, it might be the agent’s only real path to trusting or really knowing what the world is like. The conviction that comes with her felt experience of anger, and the accompanying refusal to seriously consider the competing perspective, is what protects her self-conception.

So much for the first, central function of outrage. There is also, I think, a second function of outrage anger that again maps onto Lugones’s account. Again, for Lugones, second-order anger has a communicative intent across worlds, even though it does not have communicative intent within the official world of sense. For Lugones, second-order aims to speak across worlds and plays a role in both self-creation and world creation outside of the dominant world of sense. In making space for new worlds, outrage functions to create solidarity amongst oppressed groups. Miller’s explicit aim in writing both her victim statement and her memoir is to assert her own narrative, in solidarity with other victims of sexual assault. As Lugones argues, while first-order anger is backward looking, aiming at redress from harm, second-order anger is forward looking, aiming to change what constitutes the official world of sense. Miller (2019) wants to expand the “space” in which her experience, and the experience of those like her, is intelligible, and she wants to separate this space from the distorting influence of the misogynist world in which Turner inhabits: “The statement had created a room, a place for survivors to step into and speak aloud their heaviest truths, to revisit the untouched parts of their past” (p. 252). Like reform anger, outrage anger is a “grief of distortion amongst peers” and is about change, but the change it aims at is not via an appeal to the wrongdoer, but instead it is about nurturing power and solidarity from within the world of sense occupied by members of oppressed groups.

So far, I have described outrage anger and explained how it differs from reform anger, as well as from attitudes like hatred and apathetic withdrawal. Outrage anger is a second-order anger at the state of affairs in which the harm one has endured is not intelligible as the harm that it is. The function of outrage is, primarily, to distance oneself from the dominant world of sense. It is to close oneself off from empathy for the abuser and from the hermeneutical space that trivializes or undermines the character of the harm one has experienced. A second function, at least in political contexts, is to help build a new hermeneutical space in which one’s experience is intelligible, to create a sense of solidarity amongst victims of abuse. My hope in identifying outrage anger, and distinguishing it from the more commonly valorized reform anger, has been to show that outrage anger has an important and distinct role to play in the context of political injustice. It is easy for outrage anger to come across as unreasonable or unconstructive and even to be mistaken for hatred; after all, I have argued, it is not the sort of anger that aims at uptake or reform, and it is incompatible with love or empathy for the abuser.

Consider other examples of outrage anger. When Black Lives Matter protestors damaged property in the wake of the murder of George Floyd, we need not think of this action as an expression of reform anger, that is, as a way to draw the abuser’s attention to a violation and to demand repair or reform. Instead, we might think of this expression of anger as outrage: as an expression of refusal, the refusal to engage constructively with a system that treats the harm done to Black lives as insignificant or unintelligible. Consider also social media movements like #MeToo. Much of the anger and frustration that women expressed during this movement was not to constructively engage with abusers but instead to harness women’s collective anger against a system that denies there is a persistent pattern of sexual abuse against women. This anger was constituted in part by a refusal to have empathy for men who have enacted sexual violence against women. Consider, too, slogans like ACAB (All Cops are Bastards). Slogans like this are not a demand that police, or the police system, reform themselves in light of criticisms. Instead, such slogans, which involve a blanket condemnation of a group, constitute in part a refusal to treat cops as capable of the relevant kind of repair or reform; slogans like this condemn a system, not individual actors. What is relevantly similar in all of these cases is that the anger is not in the form of a classic Strawsonian reactive attitude. That is, the anger being expressed does not take a participant stance towards the abuser. Instead, the person feeling outrage occupies something like the objective stance: their anger is constituted in part by a refusal to treat the abuser as a member of the same moral community.

This has all been by way of vindication of outrage anger. But, of course, I do not mean to offer a full-throated vindication of all outrage. Sometimes, outrage is not warranted. When protestors stormed the Capitol after the 2020 election, their outrage was not warranted; it was based on a false belief that the election had been wrongfully stolen. Sometimes, even when outrage has warrant, it is disproportionate to its object. When Natalie Wynn, the creator of the popular YouTube channel ContraPoints, was flooded with death threats for featuring a controversial trans actor and filmmaker on an episode of her show, the outrage against her was arguably disproportionate to any harm that resulted from her choice.²¹

But even aside from these sorts of cases—even when outrage is both warranted and proportionate—outrage is costly. In fact, I want to suggest, each of the positive functions of outrage is tied to some significant cost. The first function of outrage, that it protects the agent’s own self-conception against the dominant world of sense, is tied to an epistemic cost—namely, that it is epistemically obscuring in some ways. And the second function of outrage, that it allows for the creation of solidarity amongst victims, is tied to a potential political cost, specifically that it entrenches the behavior of abusers.

Take the first function first. There is a long tradition of thinking of anger as epistemically clarifying.²² But, recall, the way that outrage anger works to preserve Miller’s self-conception is by blocking her ability to feel empathy for Turner. Once she starts to sympathetically imagine what things are like for Turner, she has difficulty, at the same time, maintaining the perspective that she is not to blame for what happened to her. Miller, in virtue of her rage, is able to see some things better, but only by limiting the sort of information she is taking in.

In Miller’s case, this seems to be epistemically responsible: after all, she clearly is a victim of sexual assault and is not blameworthy for what happened to her. The perspective she is able to maintain through her anger is truth-tracking, and Turner’s is not. But it seems plausible to me that very often outrage anger is epistemically clarifying in certain ways by being epistemically obscuring in other ways. In the case of Miller and Turner, matters are fairly clear-cut: Turner committed an egregious offense, one for which he shows no real understanding or remorse. Even in the case of sexual harassment or assault, however, matters are not always so simple.

A common criticism of the #MeToo movement is that it failed to discriminate carefully amongst different types of offenses.²³ It lumped in serial rapists and horrific abuses of power alongside clumsy or unsuccessful communication around sexual boundaries. Part of the problem is that anger is a powerful weapon precisely because of what a dull instrumental it is. It is indiscriminate, incapable of nuance. Again, in the case of outrage anger, it functions precisely by closing off deliberation or consideration on behalf of the abuser. As we saw earlier, Cherry worries that hatred focuses only on the negative aspects of a person’s character, that it “feeds false information and ignores the person’s whole character,” and that it is incompatible with reconciliation and bringing the abuser back into the moral community. I think much of these worries hold for outrage anger. Not everyone who was called out and punished during the #MeToo movement was like Turner, guilty of an egregious offense and unwilling or unable to recognize what he’d done. Inevitably, some of the people who were called out in this moment were punished in a way disproportionate to what they had actually done or to the quality of their will. It would be naive to think there is no such collateral damage in movements driven by outrage anger. Indeed, as I suggested in the last section, even in the case of Turner, we might think some degree of empathy is warranted, the sort of empathy I’ve argued that is incompatible with outrage anger.

There is also a worry about how outrage shapes our response to harms. Miller takes Turner’s light sentence as a symbol of the ways in which sexual assault is not treated as a serious crime by the legal system. Her assessment seems correct; the judge in her case seems more concerned with Turner’s well-being than Miller’s. But the attention her victim statement draws to this issue is taken as an argument to increase the mandatory minimum sentences in California.²⁴ These laws are unlikely to seriously decrease sexual assault on college campuses and are likely to disproportionately harm marginalized communities.²⁵ Miller’s outrage is warranted, but the legal response to her outrage is likely detrimental to social justice outcomes.

The worries about outrage I am describing here are borne out by the empirical research. Research has demonstrated the negative effects that anger can have on our reasoning processes, making us less likely to be sensitive to mitigating or exonerating evidence, more likely to blame third parties who are unrelated to the initial cause of anger, and more likely to act in punitive ways. The stronger the intensity anger, the higher the likelihood and severity of these responses.²⁶

Let’s turn now to the second function. Again, I’ve argued that outrage anger can function to create solidarity amongst victims of an injustice. It does this by refusing to try and make itself intelligible to the abuser; outrage anger does not aim at uptake. Unlike reform anger, it is not principally about holding someone to account, having them understand their behavior and change. In fact, it is plausible that in many instances, it has the opposite effect. It is the sort of emotional response that, we can imagine, often entrenches the beliefs and behavior of the abuser. After all, it involves, in part, a refusal to see the abuser as part of the same moral community. It is, in part, a refusal to see them as capable of change or growth.

Consider the disagreement over the last few years amongst Democrats about how to engage with Trump supporters. On the one hand, there were those who insisted that, without an appreciation of the political concerns of rural working-class white people, and constructive engagement with them, Democrats would undermine their own goals. On the other hand, there were those who expressed frustration at the expectation that their time and energy should be devoted to sympathetically engaging with people whose political and moral commitments were antithetical to their own. One way to think about this disagreement is in terms of a disagreement about what political outcome should matter more: changing the hearts and minds of Trump supporters or the symbolic and material costs of making Trump supporters the center of moral concern (and whose moral concerns get displaced as a result). And the assumption underlying this sort of disagreement is that the energy devoted to either group is a finite resource.

The upshot is that while outrage anger is likely to be productive with respect to one politically important outcome—the creation of solidarity amongst victims of injustice—it plausibly undermines a different politically important outcome, one that is the target of reform anger: constructively engaging with the perpetrators of injustice with the aim of having them change their behavior. Of course, whether it is worth constructively engaging with a perpetrator of abuse depends in part on their willingness to grow and confront epistemic challenge. As I suggested at the end of §2, the intelligibility deficit that exists between victims of abuse and the dominant moral community is very often the result of active ignorance, or willful hermeneutical ignorance. As such, I want to be clear that, even if there are epistemic costs to the victim closing herself off from the abuser’s perspective, it is ultimately a result of the culpable epistemic failing of the abuser and the dominant moral community.

I’ve argued that outrage anger allows for a kind of psychological boundary setting. The costs I described in the last section are the costs that come with this boundary setting: they are the costs of emotionally and epistemically closing yourself off to another person or to a moral community. A question that naturally arises, and a question that is much discussed in conversations about political anger, is when the benefits of anger are worth the costs. One strategy to answer this question is to employ consequentialist reasoning, to weigh the positive outcomes against the negative ones, and to determine in any particular case, or type of case, whether outrage anger is on balance likely to produce better outcomes. A different strategy is to resist the idea that the value of outrage can be settled on purely instrumental grounds. For example, in talking about political anger more generally, Srinivasan (2018) argues that, in addition to the prudential reasons for and against political anger, we should consider the reasons of aptness and the intrinsic value of feeling an emotion that is an appropriate evaluative response to one’s circumstance. Indeed, she argues, the conflict that members of oppressed groups feel between reasons of aptness for their anger, and prudential reasons against it, is itself a kind of injustice, affective injustice.

I won’t weigh in directly on this debate. Instead, I want to point out how, in the case of outrage anger, even when there are prudential reasons that come into conflict with reasons of aptness, there are also always prudential reasons that align with reasons of aptness and count in favor of apt outrage. What this means is that even if the conflict Srinivasan (2018) describes often holds in the case of reform anger, it does not hold in the cases of warranted outrage. Again, I’ve argued that outrage is warranted in circumstances where the harm one has undergone is unintelligible in the dominant world of sense. Outrage anger is appropriate precisely in those cases where reform anger isn’t enough; it is appropriate in cases where there is no realistic chance of the anger being given uptake and of there being a constructive path forward involving some kind of apology or repair. In these circumstances—in which outrage is apt—outrage is always playing an indispensable prudential role. As we saw, Miller’s outrage is necessary for her to protect her own self-conception against the dominant narrative. Outrage anger is not just a fitting response to the way the world is; it is her only real path to trusting or really knowing what the world is like.

The lesson I want to emphasize here is that what gives Miller’s outrage warrant is also what makes her outrage anger prudentially valuable for her. Srinivasan (2018) suggests that affective injustice occurs when an agent is faced with a conflict between self-preservation and living a full emotional life. The Black Lives Matter protestor should not have to choose between doing what will best protect his community and appreciating at a felt level how unjust the world is. But in cases where outrage anger is apt, this is not how the reasons divide. It is true that, in some sense, Miller’s anger might threaten her self-preservation; after all, her outrage makes her the target of threats and abuse. But her anger also plays a role in her self-preservation, specifically at the level of her conceiving of herself as having moral status. Her felt appreciation of the injustice against her is a matter of self-preservation but at a psychological not merely physical level. What this means is that even if there are prudential reasons that count against outrage anger, there are also stronger prudential reasons that count in favor of it, reasons that are closely tied to what makes the outrage apt. The critic of anger insists that victims of injustice should give up their anger if it is counterproductive to the goal of resisting oppression. What I am suggesting is that, in cases like Miller’s, this advice misses the productive role that outrage anger plays even when it comes with serious costs: for Miller to resist misogyny more broadly, she first needs to resist her own internalized misogyny and to convince herself that she has been harmed and that her pain matters. These prudential reasons take priority over the prudential reasons that count against her anger.

So far, I have been focusing on the case of Miller because, I suspect, most of us will share the intuition that she is entitled to her outrage. She is the victim of a moral violation, in a hermeneutical space in which her sense that she has been violated is treated as unintelligible; in these conditions, her outrage serves as a mechanism for her own self-preservation at the level of her self-conception.²⁷

We might wonder at this point who else is entitled to outrage? After all, very often, when we feel outrage, we do so on behalf of others. Are we entitled, for example, to feel outrage on behalf of Miller? I think there are no easy answers here. If outrage is sometimes necessary to protect one’s self-conception, and if it’s being necessary depends in part on what one’s hermeneutical space is like, and in part on what one’s emotional resources are like,²⁸ there will not be a one-size-fits-all answer. Not all outrage is warranted, and, as is perhaps especially clear in our current political moment, the costs of outrage are high. Given the costs of outrage, including to victims of oppression themselves, the bar for justified outrage should be high. However, there are also significant risks to policing outrage. Not everyone who feels entitled to outrage is in fact entitled to it, but making this judgement is often difficult from the outside. So, the bar for criticism of outrage should perhaps also be quite high.

One general principle I think holds is that proximity to injustice matters. A sexual assault survivor watching Miller’s trial is entitled to outrage. What the survivor witnesses is likely to threaten their self-conception in the ways that the trial threatens Miller’s own self-conception. Miller’s experience is not unique to her; instead, it is a reflective of the dominant culture’s dismissal of the experiences of victims of sexual assault. Likewise, I think any woman witnessing Miller’s trial is entitled to some outrage. The dominant narrative, that Turner is the real victim and that Miller invited her own assault, is predicated on thinking of women as objects for male sexual pleasure and on thinking that when women protest their assault, they are being manipulative or lacking in compassion. It is reasonable for women feeling pressure to accept this narrative to channel outrage as a way to resist. In cases where one’s identity is bound up with the injustice that one is witnessing, outrage might be a necessary mechanism of self-preservation.

There is, however, also a lesson here about allyship. Sometimes, for some people, outrage is a necessary tool for self-preservation. However, outrage can come with serious costs, including to the victims of oppression themselves. When we see people express outrage at injustice, and when we see that their anger is warranted—when we share their sense that there is an injustice that needs to be rectified—it is tempting to feel outrage on their behalf. Again, anger is contagious. Moreover, it is natural to feel as though our feeling outrage is a kind of expression of solidarity. In cases where we do not need outrage to protect our own self-conception, however, I think defaulting to outrage is a mistake. The hard work of emotional compression, and of constructive engagement with perpetrators of injustice, should fall to the people who do not need outrage for their own self-preservation.²⁹ A similar lesson applies to people who feel wrongly targeted by the outrage felt by members of oppressed groups but are sympathetic to the broader social justice cause that underlies the outrage. If you are a cis man who hears women expressing indiscriminate anger at “all men,” it can feel alienating and frustrating. It can be tempting to respond by policing the outrage, encouraging a more nuanced or civil discussion, or reacting defensively. In cases where the outrage is playing an important role, however, these may be the wrong impulses. Instead, the work of allyship might be to absorb some of this indiscriminate anger, to allow the victim’s pain to be centered, and to allow for the space to be created in which that pain is fully intelligible.³⁰