Survivor guilt

Abstract

We often feel survivor guilt when the very circumstances that harm others leave us unscathed. Although survivor guilt is both commonplace and intelligible, it raises a puzzle for the standard philosophical account of guilt, according to which people feel guilt only when they take themselves to be morally blameworthy. The standard account implies that survivor guilt is uniformly unfitting, as people are not blameworthy simply for having fared better than others. In this paper, we offer a rival account of guilt, the relational account of guilt, which makes sense of survivor guilt and other forms of guilt without self-blame while preserving the intelligibility of guilt about culpable wrongdoing. According to this account, guilt involves the feeling of being unable to justify ourselves to others, and we lack self-justification when we (however blamelessly) stand on the positive side of an undesirable asymmetry with them. When someone survives something that those around her do not, the disparity in outcome constitutes an asymmetry that is often undesirable, because it arises from luck or violates a requirement of solidarity. Thus, survivor guilt is often fitting.

1 Introduction

Suppose your cancer goes into remission, but the person who sat next to you in your support group is not so lucky. Or suppose you land a coveted tenure-track job during an application year that leaves most of your friends unemployed. Or perhaps the very same virus that killed thousands of people in your city leaves you with nothing more than a mild cough.

We often feel survivor guilt when we find ourselves on an island of good fortune in an ocean of bad luck.¹ We may be grateful to be on that island, but something holds us back from feeling unqualified joy; our good luck seems somehow to have been tainted by others’ misfortunes. In recognizing its taintedness, we may find ourselves feeling guilty about having been lucky at all. ‘What’s so special about me?’ we may ask ourselves, ‘Why should I thrive while others languish?’ These feelings often strike us as appropriate. Simply put, in feeling survivor guilt, we often seem to be responding correctly to our good fortune. In the same way that Bernard Williams (1981) thought it appropriate for a lorry driver who had faultlessly killed a child to feel a special connection to the child’s death, we would expect the sole survivor of a tragedy to feel a connection to the misfortunes of the victims that a bystander would not.

Although survivor guilt may seem intuitively appropriate, it is also philosophically puzzling: why do we feel bad about ourselves for simply having fared better than others? Despite the pervasiveness and puzzlingness of survivor guilt, there has been little written about the phenomenon in the philosophical literature.² This paper aims to provide an account of survivor guilt that renders it both intelligible and, typically, fitting. In doing so, we’ll offer a novel account of guilt, which we will call the relational account of guilt. On this account, we feel guilt not only when we take ourselves to be blameworthy, but more generally when we take ourselves to be unable to justify ourselves to others in situations that require justification. We suggest that people can be unable to justify themselves to others for one of two reasons in cases of survivor guilt. First, the disparity in outcome between them and others might be unjust because it results from luck; this is the basis of a type of survivor guilt that we call ‘luck guilt’. Second, the disparity might violate a requirement of solidarity that exists within particular groups; this is the basis of what we will call ‘solidarity guilt’.

Narrowly, this paper will offer a defense of the fittingness of one specific form of guilt. More broadly, however, our account aims to serve as a defense against the commonly held view that there is something unfitting about guilt in the absence of moral blameworthiness. In doing so, we hope to offer a more general strategy for vindicating the guilt that often accompanies cases of bad moral luck, moral dilemmas, and past injustices. Guilt in the absence of moral blameworthiness need not be indefensible, but may instead be a fitting response to a lack of self-justification.³

2 Guilt and justification

We begin by rehearsing the standard account of guilt in the philosophical literature and by arguing that survivor guilt, along with some other forms of guilt, makes trouble for it. In the place of the standard account, we will offer a novel account of guilt, the relational account of guilt.

2.1 The standard account of guilt

We take, as our point of departure, the idea that when someone feels guilty, she construes herself as being unable to justify herself to others to whom she owes justification. T. M. Scanlon (1998, p. 271), for example, writes that in feeling guilt, the agent is aware that his “mode of self-governance has ignored or flouted requirements flowing from another person’s standing as someone to whom justification is owed.” Similarly, although he does not endorse the standard account, David Velleman (2003, p. 235) writes that guilt involves a feeling of “normative vulnerability,” which he understands as “the sense of being somehow unjustified, or having nothing to say for oneself.” This connection is especially evident in cases of survivor guilt. As Primo Levi (1988, p. 78) writes, drawing from his own experiences in the Holocaust, “Consciously or not, [the survivor] feels accused and judged, compelled to justify and defend himself.” And Robert Lifton (1967, p. 35) summarizes the experiences of survivors of the atomic bombing of Hiroshima thus: “From the moment of atomic bomb exposure, the hibakusha [survivor] experienced a need to justify his own survival in the face of others’ deaths.”

What does it take for someone to be able to justify herself? One natural answer is that one has not acted blameworthily: that is, one has not violated a moral requirement in a culpable way. More precisely, the items that call for self-justification are one’s own actions, and self-justification obtains when those actions were not blameworthy.⁴ What we will call the standard account of guilt conjoins the claim that guilt has to do with thoughts of lacking self-justification with this condition for self-justification. So according to the standard account, when someone feels guilt, she construes herself as having acted blameworthily, and her guilt is fitting just in case she really has acted in such a way. T. M. Scanlon (1998, p. 271), endorsing the standard account, writes that “it is appropriate to feel guilt only when one believes that one has violated principles specifying what one owes to other people,” violations of which render one’s actions unjustifiable to others. Similarly, Stephen Darwall (2009, p. 71) writes, “To feel guilty is to feel as if one is appropriately blamed (to blame) and held responsible for something one has done.”⁵

The standard account deals nicely with many garden-variety cases of guilt, which we will call cases of ‘guilt about wrongdoing’, and which typically have to do with harming someone intentionally, knowingly, or negligently, all of which are mental states that render us blameworthy for the harm. If you cheat on your partner, make gratuitously hurtful remarks about a colleague, or injure a pedestrian while speeding, you have acted in a blameworthy way. Given the standards for justification adopted by the standard account, this means that you cannot justify yourself to those whom you’ve harmed, so that it would be fitting for you to feel guilt.

The standard account also naturally suggests a story about the function of guilt. Given that guilt is a painful emotion, it is natural to think of guilt functioning as a kind of self-punishment: we punish ourselves by inflicting the pain of guilt on ourselves when we think that we have done something wrong. After all, punishments are fitting if and only if they are morally deserved, and they are deserved only as retribution for culpable wrongdoing. As Martha Nussbaum (2006, p. 207) writes, “Guilt is a type of self-punishing anger, reacting to the perception that one has done a wrong or a harm.” Indeed, some accounts of guilt build thoughts about desert into its cognitive content: on Douglas Portmore’s (2019, p. 12) account, for example, guilt requires “the thought that one deserves to feel bad given one’s failure to live up to some legitimate demand.”

Despite the attractions of the standard account, however, many cases of guilt do not fit neatly within its contours.⁶ Real-life ‘lorry drivers’, for instance, often describe themselves as feeling profound guilt at having ended a life, however blameless they may take themselves to be.⁷ In much the same way, we can imagine a teacher at a disadvantaged school feeling guilty about not having more resources to provide to her students, or a new property-owner feeling guilty upon learning that the land he just bought to build his home on was unjustly expropriated from farmers ten years prior. In such cases, it is implausible to think that the subjects construe themselves as being blameworthy for the bad outcome. And yet, their guilt is perfectly intelligible; the teacher feels guilty because she can’t help her students, and the land-owner feels guilty because he has benefitted from ill-gotten gains.

Similarly, in many cases of survivor guilt, it is quite a stretch to suggest that survivors construe themselves as blameworthy for others’ deaths. Consider the case of Jeff, a cancer survivor (Glaser et al., 2019). During treatment, he befriended another cancer patient, Allen. When Allen died from the disease, Jeff was struck with survivor guilt. As his social worker reports, “Jeff talked about the unfairness of the situation. He described Allen’s life in positive terms and compared what Allen had had to his own life circumstances… [He didn’t] understand how a ‘good person’ such as Allen could die while he got to live” (Glaser et al., 2019, p. 3). It is implausible to suppose that Jeff blamed himself for Allen’s death; he never, for example, accused himself of failing to save Allen. Rather, his guilt seems to be based on the simple thought that he survived while his friend died.

Now, someone committed to the standard account of guilt might try to accommodate these cases in one of two ways. First, he might argue that the subjects in these cases really do construe themselves as blameworthy, despite being mistaken (Rosebury, 1995; Sinnott-Armstrong, 2005). But if this is the response, then the proponent of the standard account must be able to explain how subjects do this in cases where it is obvious that they are not to blame. Perhaps we can offer such an explanation in cases of agent-regret, where there is often room to question one’s blamelessness, but it’s hard to see how survivors like Jeff could construe themselves as blameworthy.

Second, the proponent might try to argue that, despite appearances, these cases are not really cases of guilt: at least, not guilt in the moral sense. Perhaps, for instance, the cases that we describe as ‘survivor guilt’ are actually cases of another emotion, like shame.⁸ Or perhaps there is a distinction to be drawn between a moral and a non-moral sense of guilt, one of which is a response only to wrongdoing on the agents’ part, and the other of which responds to conditions that are not explicitly moral (Morris, (1987).

We think that this response is also implausible. If we decide to treat cases of survivor guilt as instances of another emotion, like shame, then we are simply using the term ‘guilt’ to mean something different from what ordinary speakers use it to mean. After all, ordinary speakers do describe what survivors feel as ‘guilt,’ despite recognizing that these feelings occur without self-attributions of moral responsibility. If the standard account cannot provide an explanation for why speakers are systematically in error about the term, then the fact that it does not account for the patterns in our ordinary usage of ‘guilt’ means that it is not really an account of guilt.

And the strategy of treating “guilt” as ambiguous between a moral and a non-moral sense seems poorly motivated and inelegant. Note that cases of moral guilt and non-moral guilt are similar in terms of their phenomenological and behavioral dimensions: both are painful emotions, and both motivate apologies, attempts to repair personal relationships that have been damaged, and other kinds of submissive behavior.⁹ The only motivation for wanting to distinguish between moral and non-moral guilt would be an existing commitment to the standard account. Furthermore, the resulting account is disunified, positing two separate emotions where we would normally recognize only one. For these reasons, we think that a unified treatment of the cases that ordinary people label “guilt” would, all else being equal, be preferable to one that draws distinctions where we ordinarily see none.

These considerations are not meant to be knockdown arguments against the standard account, as a conclusive refutation of that account isn’t the aim of this paper. Rather, given that there is a wide range of cases that make trouble for the standard account, we want to show that a more expansive account of guilt is better-positioned to make sense of them.

2.2 The relational account of guilt

We accept the standard account’s contention that guilt, at its core, involves the feeling that one can’t justify oneself to others, but we disagree about what it is to justify oneself to others. In contrast to the standard account, we suggest that the items that call for justification are not the actions that one performs, but are rather the asymmetric relations that one stands in to others. And one has justification when these relations are morally innocuous: that is, when there is nothing bad about the moral quality of how one relates to others. We will call our account the relational account of guilt.

Obviously, the account needs some unpacking, so here is what we have in mind. For starters, many asymmetric relations are morally neutral or even good: consider the asymmetric relation that obtains between two people when one benefits the other. On the other hand, many asymmetric relations are morally bad, or at least seem to be: perhaps two people stand in an asymmetrical relation because one has harmed the other, or benefited from a harm done to the other, or has been unduly benefited relative to the other. When Alice lets slip a secret that Bill has told her, for example, she is on the positive end of an asymmetric relation between them in that she has affected Bill negatively, despite his not having affected her negatively. Similarly, if Alice was fired from her job, and Bill was promoted only because of the opening that was left, there is now an asymmetric relation between them that is skewed toward Bill: he benefits from some harm done to her. These relations are prima facie morally bad: in the absence of further factors, it is bad that Alice divulged what Bill told her confidentially; in the absence of further factors, it is bad that Bill benefits from a harm done to Alice.

The existence of an apparently bad asymmetry can be detrimental for the relationship between the people on the different ends, suggesting that things are morally amiss between them. It also opens up the person on the positive end to negative reactive attitudes of resentment or indignation from those on the other end. For these reasons, when there is such an asymmetry, it is incumbent on the person on the positive end of it to show that the asymmetry is actually morally innocuous rather than morally undesirable.¹⁰ In doing so, the agent justifies herself in the face of such an asymmetry, and undermines others’ warrant to negative reactions like resentment toward her.¹¹ On the other hand, if the asymmetry really is morally bad, then the agent is unable to justify herself, and she feels guilt. This guilt then motivates her to recompense those on the other end of the asymmetry, restoring the balance between them. According to our account, the function of guilt is not to punish the agent for her own wrongdoing, but to motivate the agent to perform actions that restore the moral balance, or at least lessen the moral imbalance, between herself and others. This is true not just in cases of guilt about wrongdoing, but also in other cases of guilt, like survivor guilt. That a survivor may feel motivated to compensate those who fared worse or even to deprive herself of what they were deprived of, we think, can be explained by her desire to restore the moral balance between herself and those others.

To demonstrate how the relational account of guilt works, imagine that your aunt wills her cottage to you and her porcelain clown figurines to your sister. This creates an asymmetry between the two of you: specifically, there is a disparity between what each of you inherited. The disparity has apparent moral relevance. In the absence of justificatory factors, the division in the will is unfair, and it would be understandable if your sister felt resentment. Even though you may not be to blame for the unequal inheritance, there is nonetheless the appearance of a moral imbalance between the two of you that you must dispel to restore your relationship. Justifying yourself, in our sense, amounts to dispelling this appearance. To do this, you must show that the asymmetry that you stand on the positive side of is not morally undesirable, and thus that you do not merit your sister’s resentment. For example, perhaps your aunt was closer to you than your sister, or perhaps she knew that your sister was much wealthier than you. Sometimes, however, no self-justification is forthcoming. If you and your sister were equally close to your aunt, and neither was wealthier than the other, then perhaps the fact that you were left so much and your sister left so little is morally bad, because it is unfair. In such a case, guilt would be fitting, and it would motivate you to compensate your sister, thereby alleviating the moral imbalance between the two of you.

The relational account of guilt can explain paradigmatic cases of guilt, in which the subject takes himself to have acted blameworthily. After all, having acted blameworthily toward someone—having acted with malice toward him or negligence of his wellbeing—constitutes standing in a genuinely undesirable asymmetry toward him, which rules out self-justification.¹² But the account also explains cases of guilt in which an agent harms someone blamelessly. A morally bad asymmetry can be constituted not just by having acted with malice or negligence, but also by having harmed someone in general. And we can harm someone without having been malicious or negligent, or even having acted voluntarily. The blameless lorry driver, for example, stands in an asymmetry to the child and the child’s parents that has moral significance. After all, he needs to justify himself in the face of his having killed the child. And whereas the standard account implies that he has justification—he was not blameworthy, after all—our account implies that he does not. Regardless of whether the killing was intentional or foreseeable, it is morally terrible that he killed an innocent person. There is nothing that he could say to show that the asymmetric relation between him and the child is morally innocuous.

One might worry, however, that in capturing the fittingness of guilt in cases like the lorry driver’s, the relational account inadvertently overshoots its target and renders too much guilt fitting. If non-culpable harm is sufficient to create a morally bad asymmetry, how else might such asymmetries arise? Perhaps we should feel guilty for being well off when so many are struggling. Or perhaps we should feel guilty at the fact that, given how contingent the existence of each person is, we very likely would not have existed had history gone even slightly differently: for example, had millions not died in the Holocaust. In that way, we might seem to have benefitted from historical monstrosities.¹³ While we will not be able to fully resolve questions about guilt’s scope here, we think that they are not as pernicious as they may initially look. For instance, the fact that we wouldn’t have existed but for some terrible historical event doesn’t imply that the event was itself a benefit to us, unless we accept the controversial suggestion that existence is itself a benefit.¹⁴ And, as we suggest in the following section, demands for self-justification tend to be fairly local: we have to justify ourselves only to those with whom we share relationships, those with whom we share important social identities, those whom we encounter face-to-face, and those with whom we experience significant events. If this is so, then there may be no worry that we cannot justify ourselves toward unseen strangers who fare worse than we do.

The relational account also has a story to tell about what intensity of guilt is fitting in a particular situation. On the relational account, the question of how much guilt is fitting depends on the moral badness of the asymmetry that one stands in with others. Very small or very transient asymmetries will warrant less intense guilt than very large or very sustained asymmetries. This can also capture the thought that cases of guilt without wrongdoing will, on average, warrant less intense feelings of guilt about wrongdoing. Wrongdoing tends to constitute more severe asymmetries: when we wrong someone, the badness of the asymmetry is exacerbated by our having expressed a bad will toward her. Relational repair will thus generally be more difficult. Nevertheless, there will be exceptions to this rule. It makes sense that someone would feel more guilt at walking away unscathed from the car crash that paralyzed her best friend than at having cheated on a pop quiz. While only the latter expresses a morally bad will, the former results in a much more serious and sustained asymmetry.

At this point, a natural objection might arise. Someone might worry that, although the relational account of guilt seems to capture first-personal intuitions about when guilt is fitting, it nevertheless conflicts with third-personal reactions. Take the case of survivor guilt. While the survivor might feel guilty at having survived, those around her are often quick to try to talk her out of her guilt by reminding her that she did nothing wrong. And in those cases wherein a survivor feels no guilt at having survived a tragedy that has taken the lives of others, those around her are not likely to try to talk her into feeling guilt.

We grant that third party reactions to survivor guilt (and other forms of non-culpable guilt) often do take the form of trying to talk survivors out of feeling guilty. But we think that there are reasons for this that don’t imply that survivor guilt is unfitting. First, even if third-party reactions express the belief that survivor guilt is unfitting, this isn’t strong evidence that it really is. Rather, it may simply speak to the fact that the standard account is not a philosophical artifact, but is rather the view of guilt that many of us have internalized. In other words, regardless of when guilt is actually fitting, we may mistakenly believe that it is fitting only when the agent has acted wrongly.

Second, it’s unclear that such third parties really believe that survivor guilt is unfitting. After all, survivor guilt is a psychologically painful emotion, and we may want others not to feel it even in cases where it is fitting. And so, utterances like ‘you shouldn’t feel guilty’ might represent gentle lies told in an effort to soothe rather than sincere judgments on the appropriateness of a survivor’s guilt. In much the same way, you might tell a lie to a friend about how egregious her drunken behavior at the party actually was not because you think she acted appropriately, but because you can’t stand to see her so mortified.

And perhaps there is actually some truth to utterances like ‘you shouldn’t feel guilty’. We do not always have strong reasons to feel fitting emotions, as the familiar distinction between fittingness and moral, prudential, or all-things-considered appropriateness reminds us (D’Arms & Jacobson, 2000). Other reasons—such as reasons relating to the survivor’s well-being—may very well pack a greater normative punch. In trying to get survivors to move past their guilt, third parties may be responding to the fact that those survivors have a stronger prudential reason not to feel guilt. So even if it is true that survivors should not, all things considered, feel guilt, that does not imply that guilt is unfitting.¹⁵

In the rest of the paper, we’ll show how the relational account of guilt applies to survivor guilt. Here is a brief plan. Survivor guilt concerns a disparity between how one has fared and how others have fared. We’ll suggest that there are at least two ways in which such a disparity can constitute a morally undesirable asymmetry, each of which we associate with a particular species of survivor guilt. In Sect. 2, we discuss the first kind of survivor guilt, luck guilt, which occurs when one finds oneself on the positive side of a disparity that arose purely through luck; in Sect. 3, we discuss the second kind of survivor guilt, solidarity guilt, which occurs when the disparity violates a requirement of solidarity among the members of a group. We stress that these kinds of survivor guilt should be thought of as ideal types, and real-life cases will typically have elements of both. And survivor guilt might arise for other reasons, as well: someone might feel survivor guilt about not having done everything one could have done to help others survive, or about being alive in someone else’s place, or about having failed in one’s role-based duties toward them. Nonetheless, we believe that luck and solidarity guilt capture survivor guilt at its most distinctive.

3 Luck guilt

First, consider a kind of survivor guilt that we will label luck guilt, guilt about having fared better than others purely through luck. To get a handle on this form of guilt, imagine the guilt that someone might feel about having been one of the few passengers to survive a plane crash; or the guilt that a resident of Hiroshima might feel about having escaped the atomic bombing of his city simply because he happened to be away on business; or the guilt that a resident of a town might feel when a hurricane leaves her home relatively unscathed while destroying the homes of her neighbors.

In these cases, which include many of the paradigmatic cases of survivor guilt, it is natural for the survivor to be haunted by the question, “why me and not them?” This is exactly the thought that haunted Jerry Schemmel, a sportscaster who survived the United Airlines Flight 232 crash that killed over 100 passengers. In the aftermath of the crash, Schemmel found himself fixated on a toddler who had been sitting close to him on the plane: “I asked myself the question every day—why did I survive the crash? Why did I survive and this little boy in front of me die [sic]?”¹⁶ Or take the case of Elise O’Kane, a flight attendant who, but for a computer scheduling error, would have been on one of the planes that crashed on 9/11. Reflecting on the event years later, she asked, “Why me—out of all those wonderful people? What have I done? I’m not a saint or angel.”¹⁷

The question ‘why me and not them?’ is not simply a demand for a causal explanation. Rather, the survivor seeks a justification for the disparity between how they fared and how others fared: what makes it all right that I survived, given that so many died? A satisfactory answer to this question would thus constitute a self-justification in our sense: while it’s tragic that others died, here’s why the asymmetry between the victims and survivors is not itself a morally bad thing. We suggest that guilt arises in these cases because the survivors recognize that there is no satisfactory answer to the justificatory question.

It might seem puzzling to suggest that it is bad that some survived while others died. After all, even though it is tragic that others died, surely some survivors are better than no survivors. We agree that a world in which everyone dies is, all things considered, worse than a world in which a few survive. But things are not morally unequivocal; there is a way in which it is pro tanto worse if some survive than if all die. If this is so, then the asymmetry between surviving and dying in these cases is undesirable, and survivors can fittingly feel guilt about surviving while others died purely as the result of luck.

Why is this true? Our answer has two components. First, the fact that the disparity arose purely through luck means that, typically, survivors in cases like these are not more deserving of survival than the victims: the survivors were not necessarily better people, nor did they necessarily act more prudently than the victims. (O’Kane’s remark, that she is not a saint, suggests that if she somehow were more deserving of survival, like by being a moral saint, it would be less bad that she survived while others died.) Second, there is a tight connection between justice and desert, usually expressed in terms of the idea that an inequality is unjust to the degree that those on the positive end do not deserve that outcome more than those on the negative end do.¹⁸ Shelly Kagan, (2012, p. 349), for example, endorses something like the following principle:

It is (pro tanto) unjust that A should be at a certain level of welfare and B should be at a lower level of welfare if A does not deserve to be at that level more than B does.

It follows from this principle that if someone fares significantly better than others simply as a result of luck, and so is undeserved, then the disparity between how well he does and how well the others do is (pro tanto) unjust. Because the life-and-death disparity between survivors and victims is typically the result of luck, then, it is typically (pro tanto) unjust, and for that reason morally undesirable.

Of course, such a principle of distributive justice is vulnerable to Derek Parfit (1997)’s famous leveling-down objection. After all, the principle implies that if A and B are equally deserving, then it is more just, hence better, if both die from a plane crash than if A dies and B survives. And the objection is that this conclusion seems perverse. But we should be careful to distinguish between principles of absolute desert—what someone deserves to have on the basis of his own merits—and those of comparative desert—what someone deserves to have given facts about what others have and facts about his being more or less deserving than they are of those things. These two kinds of desert can pull in different directions. For example, if all of your students do mediocre work for your course, then they deserve Cs on an absolute basis; however, if you give some of your students As, then the other students deserve As comparatively. As a matter of absolute desert, it is just if you assign all of them Cs; but as a matter of comparative desert, it is unjust if you assign one student a C while assigning everyone else an A. Likewise, as a matter of absolute desert, it is just if everyone survives, since no one absolutely deserves to die. But this is compatible with the claim that, as a matter of comparative desert, it is unjust if some survive while others die. And since we have said nothing about how to weigh these different principles of justice, our view is consistent with the claim that it is all-things-considered more just if some survive. (And even if it is all-things-considered more just if all die, it still might be all-things-considered morally better overall if some survive.) Our point is only that there is something pro tanto undesirable about someone’s surviving while others die if he does not deserve to survive more than others.

Now, recall that according to the relational account, guilt arises when the agent cannot provide self-justification in a situation that calls for it, which is when the appearance of a morally bad asymmetry between her and others is veridical. What we have said so far does not imply that every luck-driven inequality will call for self-justification. First, some inequalities are so small as to be trivial. If Anne finds a $5 bill on the ground, while her friend Barbara does not, the difference in how they fare is so small that it does not rise to the level of requiring Anne to justify herself; it would be extremely petty of Barbara to ask, “How come you found $5 and I didn’t?”.

And second, perhaps there are people whom we do not owe self-justification. Given that the relational account says that we feel guilt when we are unable to justify ourselves to others whom we owe self-justification, constraints on the latter will impose constraints on the former. And here, we have the intuition that demands for self-justification are fairly local. Consider the fact that self-justification might be demanded from the survivors of a plane crash, but not from those with no connection to the crash. One could easily imagine the family of the toddler who died in the plane crash resentfully wondering, “How come Schemmel didn’t die in the plane crash, while our child did?” But it would be strange if they also wondered, “How come that random person on the street didn’t die in the plane crash, while our child did?”.

Of course, some might disagree with our intuition; perhaps anyone may call upon us to justify ourselves. If this is true, then a great deal of luck guilt may be fitting: it might be fitting for random passersby to feel guilt about “surviving” crashes that they have no real connection to, just as it might be fitting for people in developed countries to feel guilt about having better lives than most people around the world. If this is so, then the peculiarity of someone feeling guilt about “surviving” a plane crash that he had no connection to has less to do with its unfittingness than with the psychological difficulty of imagining the disparity between oneself and some identical other vividly enough.

Our point here is not to say that we have settled these questions, but simply to show that our account has the resources to address them.¹⁹ Regardless of how we answer them, when the unjust inequality between how the agent and others fare is large enough to constitute a disparity, and when the person on the other end of the disparity is someone to whom the agent owes self-justification, the agent will have no justification to provide: given that it was unjust that he fared much better than the others, there is nothing that he could say to show that the disparity is not itself a bad thing. According to the relational account of guilt, because there really is a morally bad asymmetry between him and the others, guilt is fitting.

The idea of luck guilt might elicit some objections. First, one might push back against this account by arguing that it falsely implies that undeserved benefits, like lottery wins, ought to elicit survivor guilt. We grant that our account has this implication, but deny that the implication is implausible. Indeed, there is evidence that suggests that people who win the lottery or inherit wealth feel guilt at their undeserved relative good fortunes (Hedenus, 2011; Jaffe & Grubman, 2007).²⁰ Consider what Lesley Herbert, a British woman who won a lottery drawing worth £4 million along with three of her friends, had to say about her experience: “Winner’s guilt is something we all struggle with and still do because we know there’s so many people who deserve to win.” Herbert’s guilt also motivated her towards the same sorts of reparative actions often undertaken by survivors: just as survivors often feel moved to perform reparative, altruistic activities in response to their guilt (Pethania, Murray and Brown 2018; Wang et al., 2018), so too was Herbert driven by guilt to ‘give back’ by volunteering at a local charity.²¹ Lottery winners are not special in this regard: more generally, there’s evidence that people feel guilt simply for being benefited more than others undeservedly (Austin et al., 2009). So guilt may be fitting even when we are benefitted by an outcome that harms no-one.

Second, don’t we feel guilt about some just inequalities as well? In the next section, we will discuss another species of survivor guilt that does not require the disparity between how one fared and how others fared to be unjust. Putting aside these cases for the time being, however, we find that there is empirical evidence that perceiving an inequality as just actually decreases feelings of survivor guilt, at least in the context of job layoffs (Brockner et al., 1986). As well, depending on how large the inequality is, the bar that it will have to clear to be just may be quite high. The fact that you were the only one in the car wearing your seatbelt might be sufficient to justify some disparity between you and the other passengers. If it was the reason that you walked away from the crash unscathed, while another passenger suffered a broken arm, for instance, then it would make sense that you would feel no guilt at your relative good fortune. But the fact that you were wearing a seatbelt does not seem sufficient for justifying a disparity as large as the one between life and death. And so, luck guilt may be fitting even in cases where one has a partial justification for their asymmetrical good fortune.

4 Solidarity guilt

Our account of luck guilt will capture many of the paradigmatic instances of survivor guilt, in which survivors are no more worthy of life or good fortune than victims. While these disparities call out for justification, luck, by its very nature, gives us nothing to say. But the account of luck guilt may not capture other cases of survivor guilt, in which survivors feel guilt despite recognizing that there is some sense in which it was just that they fared better. To see what we mean, consider the survivor guilt that Sally Racket, an academic, reported feeling upon finally landing a coveted tenure-track job after three rounds on the academic job market:

it seemed incredibly callous to yelp, dance, or run through the halls when I knew that more than 100 other exhausted candidates would get a form letter from a search committee or human resources.²²

By the time she landed her first tenure-track job, Racket really was more impressive on paper than many of those hundred other candidates: she had a PhD in hand, numerous publications, and a strong teaching record. Assuming that she recognized herself as more accomplished than many of her competitors, then her guilt will not be completely assimilable to luck guilt.

Instead, we want to suggest that, in cases like Racket’s, the subject feels survivor guilt about having done better than others—even though she deserved to—because of a norm governing inequalities in a group that she identifies with. Specifically, we suggest that within many close-knit groups, there is a requirement to “share the fates” of the other members of the group, so that any disparity—any significant difference—between how the members of the group fare in certain ways violates that requirement, and thus is morally undesirable.²³ When a disparity arises, it has to be eliminated in one of two ways: either those on the positive side of the disparity have to make whatever they have and the others lack available to those others, or, if that is impossible, they have to forgo what the others lack themselves. Of course, such fate-sharing requirements may be circumscribed in scope, and their scope may depend on the nature of the group. But what is important is that agents feel moral pressure to share in each others’ fates in certain ways. To be sure, the mere fact that someone is on the positive side of a relevant disparity does not mean that she is blameworthy. Nevertheless, insofar as she accepts the requirement, she will take the disparity between how she fares and how the other group member fares to be morally bad. She will thus take herself to be unable to justify herself toward the other members of the group in the face of the disparity, and will feel guilt.

While the idea of a fate-sharing obligation might seem strange—like the principle of justice we appealed to in the last section, it might even seem like a perverse requirement to level down—we claim such obligations are common within certain groups. We might feel that we’ve done wrong by our siblings if we are the only one who manages to escape our podunk home town, or we might find ourselves wishing that we could get sick with the same serious illness that ails our child simply so that we can share in their suffering. Sometimes, commitments to fate-sharing might move us to stay ‘rooted’ in our particular town, country, or job even when opportunity beckons us elsewhere. Finally, we often see a commitment to fate-sharing within groups that have historically suffered oppression. For instance, first-generation students of color often experience survivor guilt toward their relative academic success (Austin et al., 2009; Piorkowski, 1983). Crucially, in cases of survivor guilt that involve failures of fate sharing, the survivor need not think that the disparity between her and others is unjust in order to feel guilt. The first-generation student, for instance, might feel guilt despite recognizing that she actually did work harder than her siblings to make it to college, and that such facts make her more deserving of attending college. Rather, the mere fact that a disparity of the relevant kind exists between a survivor and those with whom she shares a normatively robust relationship seems to be enough to generate feelings of survivor guilt.

What these groups have in common is that they exhibit a strong degree of solidarity.²⁴ “Solidarity” is a protean term, but we’ll use it here roughly to mean a communal way of relating to the other members of the group, such that one treats the interests and experiences of other members of that group almost as if they were one’s own. The prototypical group that exhibits solidarity, for instance, might be a family: here, responding sympathetically to the interests and experiences of the other members is basically a sine qua non for the successful functioning of the group. But as suggested above, many other groups exhibit solidarity as well: we can, for instance, feel solidarity towards people who share our profession, religion, or nationality. Indeed, that we sometimes describe people with whom we share such ties as ‘brothers’ and ‘sisters’ speaks to the existence of solidarity within these relationships. In contrast, more diffuse groups, like humanity as a whole, do not typically generate feelings of solidarity, or commitments to fate-sharing. We won’t offer a full account of which groups feature a strong degree of solidarity, but we will note that they tend to be organized around a shared identity that matters a great deal to its members.

Because the cases of survivor guilt elicited by violations of fate-sharing commitments typically take place in the context of groups that exhibit solidarity, we’ll call this form of survivor guilt solidarity guilt.²⁵ To understand how this form of guilt can be fitting, one must first appreciate the connection between solidarity and fate-sharing obligations. We suggest that such obligations exist in order to maintain a level of group cohesion necessary for group functioning. After all, there are at least two reasons why a baseline level of sameness in the experiences of the members of a group may be necessary for members of the group to identify with one another. First, shared experiences may often be required to establish a group. Second, shared experiences may be required to maintain a group. Even if the group identifies with one another on other grounds, like shared blood, a lack of commonality in their experiences might make those other grounds for identification feel too thin. If the members of an extended family live vastly different lives around the world, for instance, then they will likely find it difficult to maintain a strong sense of familial identification.

To the extent that people accept fate-sharing requirements toward those with whom they identify, they will take any relevant disparity between how they are doing and how the others are doing to be morally undesirable. The existence of these disparities might not render agents blameworthy, but it will nevertheless morally impair their relationship. After all, because such relationships require shared experiences, failures to share experiences imperil them. For this reason, in a relationship of solidarity, there is no way to justify oneself when one is on the positive side of a disparity. When Racket runs into an old friend of hers from graduate school who has not done well on the job market, and who has had to string together various gigs adjuncting at different universities to make ends meet, what could she say for herself that would make the disparity between them all right? Even if we think that the disparity between Racket and her friend is just because of factors that make her more deserving of academic success than the friend, such facts do not provide her with a self-justification in the face of the disparity.

This gives us the following story about solidarity’s fittingness: when we stand on the positive side of a disparity with some other member (or members) of a group that grounds fate-sharing requirements that concern that kind of disparity, we may fittingly feel survivor guilt. And so too do we now have a story about solidarity guilt’s unfittingness. Specifically, solidarity guilt will be unfitting in cases where no fate-sharing obligation has been violated. There are various ways in which this might occur. First, solidarity guilt will not be fitting in cases where the relevant disparity does not actually violate the specific fate-sharing obligations of a group. While Racket might feel guilt toward other members of her field for landing a tenure-track job, it would be intuitively odd for her to feel guilt toward them for having living grandparents when many of them do not, because this disparity is unrelated to the nature of the relevant relationship. Second, solidarity guilt will not be fitting in cases where the relationship does not actually generate any fate-sharing commitments. While Racket may fittingly feel solidarity guilt towards struggling academics, it makes less sense for her to feel solidarity (and solidarity guilt) towards the struggling dentists, dancers, and data scientists of the world.

We’ll close by noting one thing that does not follow from this account. Specifically, the fact that solidarity guilt may often be a fitting response to even just disparities within relationships of solidarity does not imply that, when all is said and done, those in the relationship who fare better should relinquish their relative good fortune. It does not imply, for example, that Racket should have turned down her tenure-track job offer in order to more fully share the fates of her academic comrades. Fate-sharing commitments matter, but so does job security. What we want to suggest here is that the fact that Racket violated a fate-sharing commitment leaves a ‘moral residue’ on her good fortune, and thus renders fitting her feelings of guilt. She recognizes that even if it is permissible, all things considered, for her to accept the job offer, her job market success still violates an important requirement of solidarity.

5 Conclusion

Our feelings of guilt do not neatly track self-ascriptions of blameworthiness, as the case of survivor guilt makes apparent. Survivors need not construe themselves as blameworthy for the misfortunes of others to feel guilt. Rather, their guilt responds to the simple fact that they have no self-justification in the face of surviving or flourishing while others died or languished. Even though they are not blameworthy for others’ misfortune, they still may have nothing to say for themselves that would make the disparity between how they fared and how others fared all right. This creates a difficulty for the standard account of guilt, according to which someone feels guilt if and only if she construes herself as morally blameworthy. According to our novel account of guilt, the relational account, we feel guilt instead when we take ourselves to stand on the positive side of a morally undesirable asymmetry with someone to whom we owe justification. In the case of survivor guilt, this asymmetry consists in a disparity between how we have done and how others have done that is undesirable either because it is unjust, or because it is a violation of a requirement of solidarity.

We want to conclude by briefly noting one broader implication of our account. Although this paper has been interested in vindicating the fittingness of survivor guilt, survivors are not the only people who feel guilt without self-blame. People may feel guilt about atrocities committed by their ancestors, about benefiting from past and current injustices, and about the unintended and unforeseeable consequences of their actions. By exploring how these things can create morally bad asymmetries between us and others can arise, we may be able to vindicate other forms of guilt that are at once philosophically puzzling and deeply human.