1. Constraints and Enablements in Accounts of Social Ontology

Many prominent accounts of the ontology of social kinds explicitly hold that being a certain kind of social entity is, in some sense, a matter of being subject to certain social constraints and enablements. Probably the most well-known example of this is Searle's account of institutional facts (Reference Searle1996; Reference Searle2011). Searle focuses on those social facts that are the products of human institutions, such the facts that certain people are husbands, judges, and presidents. He holds that in order for it to be a social fact that some individual is a husband, a judge, or a president, that individual must have certain ‘conventional deontic powers’. Conventional deontic powers are socially created permissions, duties, and so on that an individual has in dependence upon collective intentionality. If members of a society collectively recognize people who have been through a certain ceremony (a marriage, a swearing in) as having the powers associated with being a husband, a judge, or a president, then this is what makes it the case that certain people are husbands, judges, and presidents.

Another view that explicitly focuses on constraints and enablements is Ásta's (Reference Ásta2018) account of social properties as conferred properties. Ásta's view is somewhat different from Searle's, since she is concerned with not only institutional kinds but also more informal social kinds, which she terms ‘communal’ kinds (2018: 180). According to Ásta, to be a particular sort of social entity is to have a certain status conferred upon you by relevant other social agents, where this status consists of particular constraints and enablements. For example, you might have the status woman conferred on you in a cocktail bar, where this means that you can expect that men will buy you a drink if they wish to make conversation with you and that you are expected to dress in a certain way and to respond politely to overtures from strange men (unless you are already talking to a man). Having the status woman conferred on you at a family party or in a workplace would involve different constraints and enablements.

Moreover, when other agents confer the status on you, they typically take themselves to be tracking something about you. This ‘base property’ (Ásta Reference Ásta2018: 2) can also vary from context to context. In some social locations, it may be genital status, in others, legal sex, and in others, self-identification as a woman. As should be apparent, on this view social kind properties are radically context-dependent. On Ásta's view, the social property consists of a conferred social status, and the social status in turn consists of constraints and enablements. Thus, being a member of a certain social kind consists of being subject to the relevant constraints and enablements.

There are other accounts of the ontology of social kinds that do not explicitly equate being a member of a social kind with being subject to social constraints and enablements. Nevertheless, many of these accounts implicitly allocate a similar role to constraints and enablements. For example, some accounts focus on equilibria in social behavior (Calvert Reference Calvert, Knight and Sened1998; Greif and Kingston Reference Greif, Kingston, Schofield and Caballero2011). Finding oneself in a situation of social equilibrium, in which no one can unilaterally deviate from a pattern of behavior without making themselves worse off, necessarily means being constrained: the available options and outcomes have been altered by the social coordination, and some possibilities (deviating and benefitting from doing so) are ruled out.

Other accounts focus on causal patterns and predictive power. Here, too, we find constraints and enablements playing an important, though tacit, role. Consider Ron Mallon's (Reference Mallon2016) account of certain human social kinds, which equates these kinds with homeostatic property clusters, described by Mallon as ‘clusters of nonaccidentally co-occurring properties that sustain explanatory and practical projects including prediction, explanation, and intervention’ (Reference Mallon2016: 69). These clusters are the result of coordinated behaviors prompted and organized by commonly understood conceptions of social roles. Races and genders are, Mallon argues, human social kinds of this sort: shared ideas about race and gender prompt people to behave in patterned ways, and this makes certain outcomes more or less likely for people who have been categorized a particular way with regard to race or gender. Here, too, constraints and enablements are implicitly in the picture, because in order for there to be patterns of outcomes that can support predictive power, it must be made harder or easier for people to do certain things. For example, if gender serves to predict earnings, then it must be the case that social practices surrounding gender roles have made it harder (in some sense or other) for women than for men to earn high wages.

I do not claim that all accounts of the ontology of social kinds entail that social constraints and enablements constitute (at least partly) membership in social kinds. Rather, I claim that there is an implicit consensus to this effect: most accounts of the ontology of social kinds either directly hold or indirectly imply that social constraints and enablements at least partly constitute social kind membership. In what follows, I assume a view of the ontology of social kinds that accords with this consensus.

2. Defining Ontic Injustice

The phenomenon that I identify with the label ontic injustice concerns the constraints and enablements that constitute social kind membership. The thought is that in some cases, these constraints and enablements may be wrongful, in the sense that they are in contravention of the individual's moral entitlements.

As an example of this phenomenon, consider the situation of wives in England and Wales prior to 1991, which is when the marital rape exemption was ended in these countries. Prior to 1991, if a husband had sex with his wife without her agreement, including by force, this did not constitute rape in the eyes of the law, as it was deemed that in getting married the wife had consented in perpetuity to sex with her husband. Accordingly, an individual socially constructed as a wife lacked the social entitlement to refuse to have sex with her husband.

According to many accounts of social ontology, including Searle's and Ásta's, what it was to be a wife in England and Wales prior to 1991 was, in part, to be someone who was not socially entitled to control fully sexual access to one's own body. This state of affairs is clearly morally wrong: everyone is morally entitled to control fully sexual access to their own body. Accordingly, there is a profound mismatch between the moral entitlements of the individuals who were wives and the social constraints and entitlements that constituted the social kind wife. It is this mismatch that I think is a distinctive type of moral wrong affecting those socially constructed as wives at that time, and which I aim to capture with the concept of ontic injustice.

It is important to be clear from the beginning about what ontic injustice is not. Ontic injustice is not the harm that occurs when individuals actually act in accordance with the relevant constraints and enablements. For instance, in the example of wives in England and Wales prior to 1991, ontic injustice is distinct from the harm of a wife actually being raped by her husband. Furthermore, ontic injustice is not the same as the psychological damage that may arise from a person's being aware of the wrongful constraints and enablements, which is part of what has been termed psychological oppression (see, for example, Bartky Reference Bartky1979). Staying with the example of wives, we may imagine that a wife who was aware of the fact that she had no legal protection against being raped by her husband might well have found this thought distressing and might also have suffered, for example, damage to her self-esteem. She may even have come to believe that she did not deserve to be fully able to control sexual access to her own body. Any of these psychological effects constitutes a serious harm, but none of these harms is what I am aiming to capture with the idea of ontic injustice.

Rather, the wrong in which I am interested is the mere fact that an individual is a certain kind of social being, where what it means to be that kind of social being is to be subject to morally inappropriate constraints and enablements. In the example concerning the marital rape exemption, the ontic injustice is the mere fact that someone is a wife, where being a wife consists, at least in part, of being someone who is not entitled to control fully sexual access to one's own body. A wife who is not raped by her husband and who suffers no psychological harm from knowing about the marital rape exemption still suffers ontic injustice. There is a parallel here with accounts of freedom as nondomination: people are dominated if they are vulnerable to the arbitrary interference of another, even if that interference never takes place (Pettit Reference Pettit1996). Ontic injustice, then, is a wrong at the level of being that affects every member of the social kind in question, regardless of their specific experiences.

Here is my definition of ontic injustice:

For instance, in England and Wales prior to 1991 the social kind wife was a site of ontic injustice because the marital rape exemption meant that being a member of the social kind wife consisted in part of lacking the social entitlement to control fully sexual access to one's own body. This constraint was wrongful to those individuals who were socially constructed as wives because it removed their social entitlement to do something that they were morally entitled to do (namely, to control fully sexual access to their own body).

Note that the definition specifies that the set of constraints and enablements must be wrongful to the individual in question. This is important, because social constraints and enablements are relational: my being under a certain social constraint, say, not to take your wallet from your pocket means that you have a certain social enablement, in this case an enablement to retain your wallet in your pocket. It follows that a person being subject to a wrongful constraint will mean that others are subject to corresponding enablements that are likewise wrongful. These corresponding enablements are wrongful, but they are not wrongful to the individual who has them; rather, they are wrongful to other individuals. For example, in England and Wales prior to 1991, a husband had the social enablement to have sex with his wife whenever he wanted regardless of whether she agreed to the sex. The husband's set of constraints and enablements was therefore wrongful, but it was not wrongful to the husband; rather, it was wrongful to the wife. Therefore, the category of husband, although it involved a wrongful set of constraints and enablements, was not a site of ontic injustice.

One question that might arise about the definition of ontic injustice that I have offered is this: why does the definition refer to constraints and enablements when surely only constraints can prevent people from doing or receiving what they are morally entitled to do or receive? The answer is that I take it that sometimes enablements can indirectly prevent individuals from doing or receiving what they are morally entitled to do or to receive. One way this can happen is if the enablements create a pressure to act in a certain way, depriving the person of choices to which they were entitled. For example, consider a case in which mothers, but not fathers, are entitled to take time off from paid work to look after children. This might result in mothers feeling obliged to take time off from paid work, because there are no good alternatives, and in their being judged negatively if they do not take time off from paid work. As a consequence, mothers’ ability to choose how to arrange their lives is compromised and they may have reduced access to goods associated with paid work. Now, on some accounts of social ontology, this case would count as one in which mothers are under a constraint regarding their careers. If we confined our attention to those views, then we would be able to account for this case purely in terms of wrongful constraints. However, on other views, such as Searle's (Reference Searle1996; Reference Searle2011), mothers would not count as being under a constraint, because there is no formal, collectively recognized duty for mothers to take time off from paid work—merely a prerogative to take time off (which is not shared by fathers). Therefore, the possibility of wrongful enablements must be kept open in order to make the definition of ontic injustice fit with any account of social ontology. Note, too, that the definition refers to a ‘wrongful set of constraints and enablements’. This allows for the possibility that a certain combination of constraints and enablements may prevent an individual from doing or receiving what they are morally entitled to do or to receive, even if no single constraint or enablement would have this effect by itself.

3. The Wrong of Ontic Injustice

The idea of a wrong that consists of being subject to morally inappropriate constraints and enablements serves to give an initial fix on the phenomenon that I want to identify as ontic injustice. However, the question then arises: what exactly does this wrong consist of? Being under a morally inappropriate set of constraints and enablements, such that other agents are socially licensed to treat one in morally impermissible ways, will typically raise the risk that such treatment, with the attendant material and psychological harms, will occur. For example, the existence of a marital rape exemption will usually mean that wives are at greater risk both of being raped by their husbands and of suffering psychological harms as a result of their awareness of this vulnerability. We tend to think that being placed at risk of harm can be wrongful in itself, as when someone is on the receiving end of irresponsible medical practice but, due to sheer luck, does not sustain physical injury. Risk thus offers one way of understanding the wrong of ontic injustice. Yet I suggest that being placed under a morally inappropriate set of constraints and enablements additionally involves a further harm, one that Jean Hampton (Reference Hampton1992: 1666) has termed ‘moral injury’, and which is understood roughly as an ‘affront to value or dignity’.

Hampton notes that the ways in which agents behave toward each other can convey meaning: by turning my back on you when you try to speak, for instance, I convey something about my attitude to you and to your attempt to speak in this context. Sometimes, the meaning of a certain behavior enacts what Hampton terms ‘diminishment’, which is the appearance that a person's value has been lowered (1991: 1674). For example, violence inflicted on a person in a deliberately humiliating way not only harms the person materially and psychologically but also conveys the impression that the person lacks the kind of value that should properly be attributed to people. ‘Diminishment’, Hampton writes, ‘is an “objective” phenomenon, by which I merely mean … that it is not something that can be identified with any psychological experience of victimization. Instead, it is something that we “read off” the effects of immoral behaviour’ (1991: 1674) Moreover, Hampton stresses that diminishment is not an actual reduction in moral worth, but rather the appearance of a reduction in moral worth. The person who is subject to a deliberately humiliating violent attack does not actually come to have lower moral worth as a consequence; rather, the way they are treated gives the false impression that they have lower moral worth than they really do.

Diminishment in turn, Hampton argues, causes damage to either the realization or the acknowledgment of a person's value. Damage to the realization of value occurs when actual treatment takes place that is contrary to the entitlements someone has in virtue of her moral value. Damage to the acknowledgment of value occurs when someone is represented as having less moral value than she in fact has; for example, an unsuccessful attempt to inflict humiliating violence on someone still conveys the impression that this form of treatment would have been appropriate, and hence that she has a lower moral value than she in fact has.

If either form of damage to value occurs, the victim has suffered moral injury. Thus Hampton defines moral injury as ‘damage to the realization of a victim's value, or damage to the acknowledgement of the victim's value, accomplished through behaviour whose meaning is such that the victim is diminished in value’ (1991: 1679, italics in original).

In articulating the idea of moral injury, Hampton relies on the claim that most of us do in fact operate with some conception of the moral worth of human beings, which is to say, some ‘beliefs about how human beings are to be valued, and how to appraise each individual's value’ (1991: 1666). For this reason, Hampton contends, we should care about the damage to the realization or acknowledgment of a person's value that is caused by diminishment:

Thus, the concept of moral injury does not supply an account of the moral worth of human beings that we should be concerned to protect from damage—of what such worth is based on, why we should care about it, and so on. Rather, the concept of moral injury identifies a form of injury or wrong that anyone who already subscribes to an account of the moral worth of human beings should acknowledge, basing her judgments about putative cases of moral injury on her prior commitments concerning the moral worth of human beings. The concept of moral injury therefore relies on a modest metaethical commitment: one must not hold a conception of the value of different sorts of beings according to which an individual's moral worth is entirely set by how they are socially represented because such a conception would imply that it is impossible for an individual's moral worth to be wrongly represented as lower than it in fact is. There needs to be some gap between actual moral worth and social representation of moral worth in order for the concept of moral injury to get off the ground.

Importantly for our purposes, although moral injury is related to the risk of material or psychological harm, it is not identical with such risk. For example, noting that books advancing racist and sexist worldviews can cause moral injury, Hampton writes, ‘such books morally injure not one individual, but a whole class of individuals, leaving them sitting ducks for treatment lower (perhaps much lower) than that which they deserve’ (1991: 1679).

However, moral injury is defined not merely as risk (being made into a ‘sitting duck’), but as actual damage. Moral injury brings with it a risk of damage to material and psychological well-being, but what it consists of is actual damage to the realization or acknowledgment of value.

The concept of moral injury captures the wrong of ontic injustice. When someone is subject to constraints and enablements that license treatment that falls far short of the sort of treatment to which she is morally entitled, she suffers a moral injury, one that takes the form of damage to the acknowledgment of her value. This damage is an additional wrong over and above the wrong of being placed at risk of material and psychological harm. Someone who is socially constructed as a wife in the context of a marital rape exemption has been represented as having less moral value than she in fact has, suffering damage to the acknowledgment of her value that constitutes a moral injury. However, there are two ways in which I slightly adapt Hampton's account of moral injury presentation to make it better fit my purposes.

First, Hampton emphasizes ‘behavior’ and seems mostly to be thinking about cases where an agent A behaves a certain way toward another agent, B, causing B to suffer diminishment. This makes sense within the context of Hampton's larger project, which concerns theories of punishment. The problem here is that it does not make sense to think of an individual's being socially constructed as a member of a certain kind as something that is enacted through the behavior of one single other agent. Even in cases where a particular agent plays an important role—the officiant in a marriage ceremony, say—the maintenance of a social ontology requires the participation of many others as well, and this participation is often cashed out in terms of their attitudes rather than specific actions. However, there is nothing in the notion of moral injury itself that necessitates an interpersonal focus. The wrong that a person suffers when they suffer a moral injury is the actual damage to the realization or acknowledgment of their value. From the victim's perspective, this damage can be the same, whether it is inflicted directly by one individual or via more diffuse mechanism. I therefore think of moral injury as accomplished through treatment than through behavior, where interpersonal behavior is just one form of treatment. Someone who has been socially constructed as a member of a certain human kind has been on the receiving end of a form of treatment, even though this treatment cannot usually be equated with the behavior of a specific individual. What matters is that the treatment is the product of human social arrangements and therefore falls within the scope of considerations of justice rather than being the product of natural forces. This last point is important. Categories that are not socially constructed cannot be sites of ontic injustice because they are not the sort of thing that can be appropriately considered to be just or unjust.

The second adaptation I make to Hampton's account of moral injury concerns her focus on human beings. Again, this makes sense in the context of a discussion of punishment. But, in fact, there is no general reason to limit the scope of moral injury to human beings. The concept of moral injury relies on a prior notion of a conception of the moral worth of human beings, and it is true that such conceptions loom particularly large in our moral landscape. However, it is perfectly possible to have a conception of the moral worth of beings of other sorts—sentient beings, for example—and many people do seem to hold such conceptions. Such conceptions can function as the basis of a judgment of moral injury just as well as can conceptions of human value. I do not claim that everyone, in fact, holds conceptions of the moral worth of broader classes of beings beyond human beings, nor do I defend the claim that some such conceptions are warranted. I simply note that such conceptions will serve perfectly well as support for judgments about moral injury, and for this reason, I do not take the notion of moral injury to apply solely to human beings but treat it as applicable to any sort of being that may be the subject of a conception of value. (This distinguishes my use of the concept of moral injury from, for example, Mari Mikkola's (Reference Mikkola2016) use of it in her account of dehumanization.)

There is a close connection between Hampton's account of moral injury and Stephen Darwall's account of ‘recognition respect’ (Reference Darwall1977: 38). According to Darwall, there are two quite different attitudes that go by the name of respect. On the one hand, ‘appraisal respect’ is a positive evaluation of a person; this is the sort of respect involved in respecting someone as a sportsperson or as a philosopher. Appraisal respect can be withdrawn, for example, if one learns that one's favorite sportsperson has been taking performance-enhancing drugs or that one's favorite philosopher is a serial plagiarist. Recognition respect, on the other hand, is a matter of recognizing certain properties of someone or something and being disposed to respond appropriately to these. This is the attitude involved in respecting someone as a person or in respecting a work of art. We recognize that the person has certain morally relevant properties that mean that we ought not to harm her, exploit her, and so on. Similarly, we recognize that the work of art has aesthetic value, and we ought not to wantonly destroy or damage it. Unlike appraisal respect, recognition respect cannot legitimately be withdrawn; even someone who has forfeited our appraisal respect as a sportsperson still has a claim to our recognition respect as a person.

Darwall further identifies a subcategory of recognition respect, moral recognition respect, in which the properties in question are morally relevant. The case of respecting the person is a case of moral recognition respect, whereas the case of respecting the painting is not. There is a strong relationship between failures of recognition respect and moral injuries. When a failure of moral recognition respect occurs, the individual toward whom the respect was owed suffers a moral injury, because damage has been done to the acknowledgment of their value. Moral injury is a more direct way of spelling out the wrong of ontic injustice because it places the focus on what has happened to the individual who has been wronged, not on the attitude by means of which the wrong was enacted. However, the notion of recognition respect enriches our understanding of how moral injury is enacted and, in particular, is useful in that it highlights the idea of a mismatch between moral entitlements, on the one hand, and social entitlements, on the other. When ontic injustice occurs, there has been a social failure to respond appropriately to the morally relevant properties of the individual in question, and it is this failure that enacts diminishment, the impression that the person's moral value is less than it really is. In identifying cases of ontic injustice, then, we should be on the lookout for instances of mismatch between moral and social entitlements as highlighted by the idea of recognition respect. It is worth mentioning here that the notion of recognition respect has clear connections to more overtly political conceptions of recognition (Honneth Reference Honneth1996; Fraser Reference Fraser2014). It seems to me that the account of the wrong of ontic injustice that I have offered here could easily be developed in that direction if one were so inclined, yielding a version of the concept that was suited to use in a certain political framework. While I have no objection to this, I am here seeking to offer a version of the concept that can be incorporated into as many different political and ethical frameworks as possible.

Not everyone will be friendly to notions such as moral injury and recognition respect. For those who do not accept these notions, the wrong that I seek to pick out with the concept of ontic injustice must be understood more simply as the wrong of being placed at risk of material and psychological harms. If a particular wife never suffers marital rape and experiences no distress at knowing that she would have no legal recourse if she were to suffer marital rape, well and good; but she has still been placed at risk of these harms in virtue of being socially constructed as a wife, and that is wrong. For someone who rejects the concepts of moral injury and recognition respect, the wrong of ontic injustice will perhaps not seem as deep as it does to one who accepts those notions, but the pragmatic benefits of having the concept of ontic injustice are exactly the same; I discuss these benefits in section 5 below.

The severity of an ontic injustice is the severity of the moral injury it involves, which in turn is determined by the strength of the diminishment, that is, the strength of the impression that the individual's value has been lowered. The question of what determines the strength of diminishment is a complex issue, beyond the scope of the article. I suspect, though, that it is a function of at least three factors. The first factor is extent to which the social treatment falls short of that to which the individual is entitled. For example, an act of extreme and humiliating violence enacts a greater diminishment than does a social slight or snub. The second is the importance of the relevant moral entitlements—how central they are to the conception of value (such as of human value) that is in question. Bodily integrity and autonomy are central to many conceptions of human value, so violations of these entitlements enact particularly strong diminishment. The third factor is the interpersonal context in which the diminishment takes place. For example, the meaning that is conveyed by someone being subjected to wrongful constraints and enablements is affected by who else is also subjected to those constraints and enablements. Imagine a society of total surveillance: everyone is constantly monitored, perhaps by some rotating system of mutual monitoring, and no one is exempt. Let us further suppose that this monitoring is justified by reference to some conception of what it is to be a citizen; it is a citizen's duty to submit to monitoring for the collective good. Qua citizen, one loses the social entitlement to maintain privacy; but since everyone is socially constructed as a citizen, everyone loses this entitlement. Now, assuming that people do have a moral entitlement to greater privacy than they receive here, there is indeed a moral injury in the picture here, one that is suffered by all citizens. However, the diminishment that is enacted for any given individual is much less severe than it would be if the monitoring only applied to a small group of individuals picked out by some characteristic such as race, religion, or sexual orientation. If that were the case, it seems to me that the impression of lowered value for those individuals would be much more pronounced, and the diminishment would accordingly be stronger. There may well be further factors relevant to the strength of diminishment beyond the three I have mentioned here, but a fuller exploration must wait for another occasion.

4. Using the Concept of Ontic Injustice

Can the concept of ontic injustice earn its philosophical keep? In this section, I explain how the concept functions as a philosophical tool, and in the next section, I say more about what we stand to gain from adding it to our conceptual repertoire.

In order for the concept of ontic injustice to yield detailed judgments about actual cases, it must be combined with two further components: firstly, an account of the ontology of the social kind in question (which must draw on an account of the ontology of social kinds in general); and, secondly, an account of the moral value of the individual(s) in question and the moral demands to which this value gives rise.

With regard to the first component, making judgments about ontic injustice requires us to know whether or not a certain constraint or enablement is partly constitutive of a social kind in question. This is a point on which people can and do disagree, based in part on the account of social ontology to which they subscribe. Consider, for example, the phenomenon of bisexual exclusion, whereby bisexual people are often, to varying extents, denied membership in queer communities and denied access to queer spaces and queer cultural resources (Eadie Reference Eadie, Bristow and Wilson1993; Serano Reference Serano2013). Bisexual exclusion is wrong: bisexual people have a genuine interest in having access to queer communities and cultural resources, and their exclusion is unjustified. Now, is this a case of ontic injustice? Well, that depends on whether there is a wrongful constraint that partly constitutes the social kind bisexual. According to Ásta's conferralist account of social ontology, bisexual is here functioning as a communal kind, a social status conferred informally by other social agents in the queer community. When those agents behave toward people who they understand to be bisexual in ways that convey the impression that those people are unwelcome in queer spaces, then we can say that the social status of bisexual includes a constraint that makes it harder to be present in queer spaces. Therefore, a conferralist would say that the social kind bisexual is a site of ontic injustice. However, applying Searle's account to this case would give the opposite result, because Searle holds that conventional deontic powers are the only type of constraint or enablement that contributes to constructing social kinds. Since conventional deontic powers depend on the collective recognition of entitlements, in order for the wrongful constraint generated by bisexual exclusion to partly constitute the social kind bisexual there would need to be an explicit recognition—that is at least widely shared—that bisexuals are not entitled to access queer communities and cultural resources. Since (I take it) bisexual exclusion typically does not work like this, because it is usually a much more subtle phenomenon, Searle's account entails that the wrongful constraint generated by bisexual exclusion is not part of what constitutes the social kind bisexual, and so this is not a case of ontic injustice. This shows how judgments about putative cases of ontic injustice depend on one's understanding of the ontology of the social kind in question.

With regard to the second component, what someone will think about putative cases of ontic injustice will depend on their views about what kinds of moral demands are created by different features of agents. For example, whether or not one thinks that the social kind nonhuman animals is a site of ontic injustice will depend on one's views on the moral entitlements of different sorts of beings. If you hold a conception of the moral value of sentient beings according to which they ought not to be killed simply for the expedience of others, then you will probably think that the social kind nonhuman animal involves a wrongful constraint, because nonhuman animals are sentient beings and it is socially sanctioned to kill them simply for the expedience of others. Recall that the explanation of the wrong of ontic injustice as a moral injury does not purport to supply a general account of moral demands, because applying the concept of moral injury always involves an appeal to some independent conception of the moral value of different sorts of beings and the demands this value generates.

Although it is beyond the scope of this article to address the question of how ontic injustice may best be combatted, it is important to recognize that the idea of ontic injustice does not entail that the best or only response to ontic injustice is to abolish the social kind in question. The best response could equally well be to change social ontology so that the kind is no longer constituted by wrongful constraints and enablements. Whether this is possible in a particular case depends on whether or not the wrongful constraints are a necessary part of the social kind. There is no general answer about whether this is possible; it depends on the kind in question.

For example, the marital rape exemption in England and Wales was ended in 1991: rape within marriage is now legally considered rape. Thus, the institutional kind wives has changed. Whether or not all wrongful constraints have been removed is another question, but it certainly seems plausible to think that all wrongful constraints could be removed without transforming the category so much that the label wife is no longer appropriate. We may thus think that the social kind wife is only contingently a site of ontic injustice. In a similar spirit, Chike Jeffers holds a cultural theory of race according to which racial kinds could exist in the absence of racial oppression, since what is essential to racial kinds is not relations of dominance and subordination, but shared culture. On this view, ‘a situation in which racial groups persist but in a state of equality rather than socioeconomic and Eurocentric cultural hierarchy, respecting and mutually influencing each other while remaining relatively distinct, is a coherent and admirable goal’ (Reference Jeffers2013: 421). This kind of position, on which racial futures are possible and desirable, is perfectly compatible with the claim that races are currently sites of ontic injustice. By contrast, consider the social kind slave. Although the constraints that constitute the kind slave are different in different institutions of slavery that have existed at different times and places, all of these institutions involve such a serious limitation of autonomy (besides frequently licensing the most extreme and brutal physical and mental harms) that it is impossible to imagine a social kind that merited the term slave that did not involve wrongful constraints. Even the least vicious form of slavery is very seriously wrongful. Therefore, it is natural to think that the social kind slave is necessarily a site of ontic injustice, and that the only appropriate response to this injustice is to abolish the kind completely. Thus, what one thinks about the response that ontic injustice merits will depend on the particular case of ontic injustice and on one's favored account of social ontology and, in particular, of the ways in which social categories persist through time.

5. The Theoretical Value of Ontic Injustice

We do not need a concept of ontic injustice to see that some constraints and enablements can be wrongful; what, then, do we gain by having a concept of ontic injustice that draws our attention specifically to the role these constraints and enablements play in social ontology? In other words, we already knew that the marital rape exemption, for example, was an injustice; what do we gain by thinking of it as an ontic injustice in particular? There are three main reasons for incorporating the concept of ontic injustice into our conceptual repertoire.

The first is that attending to the construction of social kinds highlights some specific ways in which wrongful constraints and enablements can function. This is because social kinds (as understood here) function to transmit constraints and enablements. If I am made into a wife, I come under the constraint of not being able to refuse sex with my husband. Nothing specific needs to be done to place me personally under that constraint in particular; indeed, I may not even know about it. Yet, in becoming a wife, I become subject to this constraint, because to be so constrained is part of what it is to be a wife. The operation of social ontology is thus a way by which many people come to be placed under constraints and enablements that are wrongful. Moreover, because many social kinds involve constraints and enablements that are valuable to people as well as ones that are wrongful, it can be the case that people cannot become members of social kinds that matter to them without placing themselves under wrongful constraints and enablements. This is because social kinds ‘bundle’ constraints and enablements, so to speak. Staying with the marriage case, suppose that I want to become a certain person's legal next of kin. Under certain social conditions I can do this only by becoming a wife and hence at the cost of placing myself under an obligation to have sex with my husband whenever he chooses. Although one or both of us might protest at this—John Stuart Mill's 1851 (Reference Mill and Elliot1910: 1:158) statement preceding his marriage to Harriet Taylor comes to mind—nothing can be done about it without changing the nature of the social kind wife, which is something that we, as two individuals, are not in position to do directly. Taking this thought even further, I might value membership in a social kind because of the meaning that kind has to me, independently of any particular constraints and enablements; perhaps I want to be a wife because of the religious significance this category holds for me. Yet I cannot access this good without placing myself under the wrongful constraint that partly constitutes the kind wife in my particular social setting. Thus, noticing how wrongful constraints and enablements can partly constitute membership in a social kind enriches our understanding of the sorts of difficulties they create for actual and potential members of the kind in question.

The second benefit of the concept of ontic injustice is that acknowledging that being made into a member of a certain kind can be wrongful in and of itself can bring clarity and insight when it comes to debates about the nature of specific social kinds. For example, it is sometimes claimed that in order to do justice to the experiences of trans people, theorists of gender ought to adopt a constraint on descriptive accounts of the metaphysics of gender kinds to the effect that any acceptable account must entail that all trans people are members of their identified gender. In other words, it is sometimes claimed that if a descriptive account of the metaphysics of gender kinds does not entail that all trans women are members of the social kind women and that none are members of other gender kinds (and the same, mutatis mutandis, for other genders), it should be rejected. Call this the metaphysical constraint. The metaphysical constraint has been advocated by Stephanie Julia Kapusta in an interview by Nathan Eckstrand (Kapusta Reference Kapusta2018) and by Rachel McKinnon in a contribution (January 26, 2016) to the blog ‘PEA Soup’ (2016). Note that the metaphysical constraint is different from the claim that prescriptive accounts of gender terms ought to have this implication. The latter concerns how we ought to use terms like woman within the context of an emancipatory political project. One proposal here is to use gender terms to refer to something other than gender social kinds in the sense usually understood in social ontology—for example, to gender identity (Jenkins Reference Jenkins2015).

For someone who believes, as I do, that misgendering someone (treating them as a gender other than that with which they identify) is seriously morally wrong and plays a major role in oppression (Bettcher Reference Bettcher2007; Kapusta Reference Kapusta2016), the metaphysical constraint has undeniable appeal. If the nature of gender kinds is such that everyone is a member of the gender with which they identify, then all instances of misgendering are straightforwardly factually inaccurate. This seems to offer a satisfying way of capturing the sense that something is going very wrong when misgendering occurs. However, offering an account of the ontology of gender kinds as they currently exist that meets the metaphysical constraint rules out the possibility of applying most accounts of social ontology to gender. For example, a proponent of conferralism would have to say that because, in a transphobic society, gender is frequently conferred in ways that depart from self-identification, many people are often made into members of gender kinds with which they do not identify. Other accounts fare similarly, because they take actual patterns of social behavior to determine (in one way or another) the boundaries of social kinds, and current actual patterns of social behavior frequently position trans people as members of the genders to which they were assigned at birth. Accepting the metaphysical constraint therefore gives rise to serious theoretical challenges.

Recognizing the phenomenon of ontic injustice helps to open up a different avenue: instead of accepting the metaphysical constraint, we should acknowledge the possibility that gender kinds at present are set up in ways that exclude some trans people from membership in the gender kind with which they identify (and include them in other gender kinds) and that this social construction is itself wrongful (Dembroff [Reference Dembroffforthcoming] adopts a similar stance on cases such as this). For example, in many contexts at present, genital status is treated as determining people's gender for social purposes, meaning that some trans people are prevented from navigating gendered social spaces in a way that is congruent with their gender identity (Bettcher Reference Bettcher2007). I offer the fictional case of Alex, a trans woman, as an example of this situation as giving rise to ontic injustice.

Because Alex lives in a transphobic society in which genitals are taken to be the decisive determiner of gender, and because of the way that her body appears, she is socially constructed as a member of the social kind men. This social construction is at least partly constituted by Alex's being subject to a set of constraints and enablements that includes, for example, not being permitted to use women's bathrooms (alongside many other wrongful constraints and enablements). However, I would argue, and proponents of the metaphysical constraint would agree, that this social constraint is in violation of Alex's moral entitlements. Thus, we can say that Alex is wronged by being subject to the constraints and enablements that constitute membership in the social kind men. She suffers an ontic injustice.

Having the concept of ontic injustice to hand helps us to articulate the claim that it is a social fact that Alex is a man and not a woman—but it ought not to be. Alex is wronged by being socially constructed as a man because the constraints and enablements that partly constitute kind membership are at odds with her moral entitlements. By bringing this wrong into focus, the concept of ontic injustice allows us to say that an account of the ontology of gender kinds, according to which Alex is a member of the social kind men (and is not a member of the social kind women), may be, not an incorrect or wrong account of ontology, but rather a correct account of a wrongful ontology. This does justice to our sense that there is something profoundly wrong going on when Alex is categorized as a man while avoiding taking on the theoretical difficulties to which the descriptive constraint gives rise. Having the concept of ontic injustice to hand thus helps us identify an alternative to the metaphysical constraint that is worthy of investigation. (Of course, recognizing ontic injustice does not prevent us from endorsing accounts of the ontology of gender kinds on which all trans people are members of their identified genders; it merely enables us to entertain accounts that do not have this consequence.)

The third reason that ontic injustice is a valuable philosophical tool is that it aids our understanding of various kinds of systematic oppression, such as gender-based oppression and race-based oppression. Although specific definitions vary, oppression is generally understood to be a systematic, significant, and pervasive phenomenon (Frye Reference Frye1983; Young Reference Young2011; Cudd Reference Cudd2006). In some cases of ontic injustice, the set of constraints and enablements that (partly) constitute kind membership impose limitations that are sufficiently systematic, significant and pervasive to qualify as oppressive. I use the term ontic oppression to describe this type of ontic injustice (see also Dembroff [Reference Dembroffforthcoming] for exploration of a similar idea). Ontic oppression should be understood as a subspecies of ontic injustice: not every case of ontic injustice is an instance of ontic oppression, for some cases of ontic injustice will feature only trivial limitations in a very narrow domain, which cannot be said to amount to oppression. In a similar vein, ontic oppression is not synonymous with oppression as such because it may well be the case that some of the limitations or disadvantages that feature in a certain instance of oppression are not constitutive of the oppressed kind. Rather, I understand ontic oppression as capturing one facet or aspect of oppression.

There is a close connection between the concept of ontic oppression and a family of accounts of race and of gender/sex according to which race and/or gender/sex is the product of oppressive social relations. Monique Wittig, for example, writes,

Similarly, Catharine MacKinnon (Reference MacKinnon1991) holds that gender is constituted by relations of sexualized domination and subordination. Charles Mills's work on the ‘racial contract’ (Reference Mills1999) expounds a roughly parallel view about race categories, according to which race categories are the product of an agreement between European people to treat non-European people as ‘subhuman’ (having an inferior set of rights), thereby producing the category of ‘white’ and various categories of non-white people. Sally Haslanger (Reference Haslanger2012a) defends accounts of both race and gender according to which these kinds are defined by systematic subordination. Each of these views posits that ontological categories such as woman or black person are constituted by oppressive constraints. The concept of ontic oppression brings into sharper focus the claim that this social construction is wrongful in and of itself.

One reason that this sharpening of focus is useful is that recognizing ontic oppression has the potential to affect how we approach emancipatory projects. The claim that social kinds are wrongful in and of themselves can function as the basis for a demand for intervention to change the nature of these kinds. If people constructed as women, for example, are wronged by being so constructed, then justice demands that the process of social construction be disrupted. The exact form that this disruption would need to take depends, of course, on the specific account of the ontology of social kinds that one favors; as Haslanger puts it, we need to understand social ontology in order to identify the ‘levers for change’ (Reference Haslanger2012b: 215). However, it is certainly possible that the necessary disruption may turn out to involve intervening in social processes which are not themselves obviously wrongful aside from their role in maintaining the construction of a kind that is a site of ontic oppression.

The concept of ontic oppression, and of ontic injustice more broadly, can help us to make a strong case for this sort of intervention by getting a clear fix on the wrong that we are seeking to prevent—ontic oppression—and laying bare its relation to the specific social processes that we are targeting in our efforts to bring about emancipatory social change.