JUSTIFYING ANTISUBORDINATION

In Mandla v. Dowell-Lee,¹ the UK House of Lords decided that the Park Grove School’s uniform policy unjustifiably discriminated against Gurdiner Singh, a Sikh boy, because of his “ethnic origin,” in violation of Britain’s Race Relations Act. The statute prohibited “direct” discrimination, where a person’s motivating reason for disadvantaging someone is that the victim has some property protected as a legally prohibited ground of discrimination. It also prohibited “indirect” discrimination, where, regardless of a perpetrator’s motivating reasons, an act disadvantages people who have such a property worse than people who do not. The uniform policy was indirectly discriminatory. By requiring all boys to cut their hair at shoulder length and wear a designated school cap while attending classes, it disadvantaged Sikh boys worse than non-Sikh boys because it prevented them from growing their hair and wearing it in a turban as required by their religious convictions and communal identity.

According to the “antisubordination” view of antidiscrimination law, the law’s purpose is to reduce caste-like hierarchies of social status between social groups.² Examples of status inequality include social stratification between masters and enslaved people, lords and serfs, nobles and peasants, or Brahmins and untouchables. In these examples, social norms create levels of rank or station between groups, with some having superior or dominant social status over others with inferior or subordinate status. An antisubordination theorist might hold that the reason to legally prohibit Park Grove’s uniform policy is that although there was no motivation to disadvantage Sikhs, it exacerbated norms that portray European cultural traditions as normal or standard and juxtapose the traditions of non-European ethnic groups as aberrant deviations from this standard of normalcy. By privileging the former and marginalizing the latter, it perpetuated a social hierarchy wherein non-Europeans are situated beneath Europeans. More broadly, an interpretation of statutory antidiscrimination law as pursuing antisubordination is attractive because it seems to fit and justify the law’s main features. The reason to include certain properties (race, sex, gender identity, sexual orientation, religion, disability, and age) in the legally prohibited grounds of discrimination is that they characterize historically subordinate groups (people of color, women, transgender persons, same-sex couples, Muslim or Jewish people, disabled people, and older people). Protecting members of these groups from being disadvantaged, directly or indirectly, in employment, housing, education, and the provision of goods and services (the usual spheres of life that antidiscrimination law regulates) furthers the goal of dismantling status hierarchies.³

This paper has two interrelated objectives. The first is to clarify the moral justification for the antisubordination interpretation of antidiscrimination law’s purpose. The second is to respond to a powerful objection to the claims of fit outlined above.

First, the dominant view of the justification for antisubordination is “non-welfarist.”⁴ On this view, we have a moral duty to refrain from exacerbating the subordination of the social group to which a person belongs (by discriminating against her, for example), and this duty has a basis that is independent of how sustaining status hierarchies is bad for people’s well-being. The non-welfarist view, while not denying that we should not contribute to status inequality because it is bad, holds that subordinating someone also wrongs or disrespects her by violating a moral claim or right she has that is grounded in her special moral status, such as her status as an equal. A standard rationale for this view is that there appear to be cases where we can intuitively wrong someone by violating a duty to avoid subordinating them without setting back their interests. I argue, however, that these cases do not provide a compelling motivation for the non-welfarist view. If we grant that there is an antisubordination duty in these cases, we are morally required to comply with it only if violating it is bad for its beneficiary’s well-being and produces no compensating benefits for the beneficiary or third parties. It follows, I claim, that we should accept a fully “welfarist” view of antisubordination’s justification.

Second, it has been objected that because antidiscrimination law offers “symmetrical” protection to members of privileged and subordinate social groups, it cannot be interpreted as pursuing antisubordination. It is not clear how prohibiting discriminatory acts against privileged groups reduces social subordination. I argue that extant responses to this objection posit the wrong kind of reason for symmetry. I then offer a novel reply according to which the reason for symmetry is that it reduces subordination by preventing privileged groups from becoming subordinate in the future. This reason is much weaker than the reason to protect subordinate groups being to prevent their historical marginalization from being perpetuated into the future. But it has some weight. This helps explain why affirmative action initiatives, which commit discriminatory acts against members of privileged groups to benefit members of subordinate groups, are generally permissible yet generate so much controversy.

My two objectives are connected. It might be thought that the non-welfarist view of antisubordination has an advantage over the welfarist view because only it can explain antidiscrimination law’s symmetry in (what I say is) the right way. But the novel explanation I offer does not rely on the non-welfarist view. Therefore, this view is inessential for properly explaining symmetry.

In part 2, I explain the nature of social subordination. In part 3, I argue that it harms members of social groups with lower status by setting back their interest in secured social bases of self-respect. In part 4, I describe cases of objectionable, harmless subordination that are thought to support the non-welfarist view. In part 5, I argue that if in these cases we must comply with a duty to refrain from contributing to social status hierarchy, that requirement is ultimately rooted in how violating the duty sets back its beneficiary’s well-being without any compensatory benefits, not in respect for the beneficiary’s status. In part 6, I explain the symmetry objection to the antisubordination interpretation of antidiscrimination law. I critique existing replies and defend a novel one.

“Social subordination” has been defined as “the state of affairs in which one social group has a standing in society as a whole that is lower than that of another social group.”⁵ A “social group” is characterized by a property its members share to the exclusion of nonmembers. That property, it is frequently thought, must be “socially salient” or taken “to have implications for the character and behavior of members of the group, and for the social roles that they are capable of occupying.”⁶ Paradigmatic socially salient properties include one’s race, sex, disability, sexual orientation, and gender identity. But a group need not enjoy an existence beyond that of its members, its members need not be homogenous in all respects, and they need not regard their group membership as an important part of their identities.

Antisubordination as I shall understand it is analogous to resisting “structural” injustice or pursuing “patterned” principles of justice.⁷ Subordination is a state of affairs that, when it is objectionable, is so regardless of whether its existence is historically attributable to the conduct of any agent(s). Discrete acts that accumulate over time to cause subordination, such as discrete discriminatory acts, are objectionable too. But that is due to how they link up to objectionable societal states of affairs, patterns, or structures. They contribute to, exacerbate, or perpetuate objectionable social norms, conventions, beliefs, or practices under which some groups have social status that is lower than that of others. So, antisubordination signals resistance to both the conduct in question and the state of affairs of social status hierarchy it links up to.

Subordination is, moreover, a comparative problem. It obtains where a social group has lower social status relative to another. Antisubordination thus aims to reduce these gaps in social status, not just to make people absolutely better off in a manner that is specifiable without reference to relations between groups. In this respect, it is a “strongly egalitarian” goal.⁸

Members of a socially subordinated group usually have less political and social power than others. They also have less “de facto authority” to get others to voluntarily obey them when they issue orders, to listen to their testimony when they voice grievances, or to demand redress.⁹ Further, the characteristic properties of subordinate groups attract less “consideration” than those of corresponding privileged groups. “Consideration” denotes attitudes of “deference and respect” from others.¹⁰ These attitudes signify that the people who receive them belong to “a higher or better class” and that their interests deserve priority over others’. Groups with superior social status can invoke their high status to ground the legitimacy of demands that others display positive attitudes toward them regardless of whether their characteristic properties make those attitudes warranted.¹¹

Groups with lower status are, conversely, subject to negative attitudes of “censure,” which identify them “with patterns of action or dispositions of behavior that are perceived as worthless, or worse, as vices.”¹² These attitudes are fueled by widely held generalizations that ascribe negative dispositions or social roles to people simply because they possess a given group-characterizing property. Just as people with superior social status can invoke their high status to rationalize claims for preferential treatment, stereotypes about subordinate groups are perceived as rationalizing subordinate groups’ lack of power, lack of authority, and deservingness of censure.¹³

Sophia Moreau has proposed that these facets of subordination are tied together by what she calls “structural accommodations.” These are social structures or norms that transform the properties that differentiate privileged from subordinate groups into status hierarchies. While being apparently neutral and not generated by malign motivations, they cater to the needs and interests of members of privileged groups, making them seem standard, normal, or natural while making the differing needs of members of subordinate groups seem exceptional, deviant, or invisible. They are distinct from conscious attitudes of deference toward higher-status groups in being constituted by tacit and diffuse norms. The way they normalize the needs of privileged groups supports those attitudes as well as privileged groups’ greater power and authority, and stereotypes about the behavioral dispositions and proper roles of members of subordinate groups.¹⁴

For Moreau, discriminatory acts are morally wrong because they “contribute to,” “perpetuate,” or “play an important causal role in sustaining” social subordination.¹⁵ A directly discriminatory act can constitute an expression of censure toward the victim’s social group. And an indirectly discriminatory act “can mark out certain groups as inferior to others by rendering them invisible in certain contexts.”¹⁶ For example, in Mandla, Park Grove’s indirectly discriminatory uniform policy constituted a norm that accommodated European traditions of dress and depicted them as normal and non-European traditions as exceptions to the norm. This marginalized Sikh persons and rendered their needs invisible within the British education system.

As I shall explain in what follows, we can develop a plausible “welfarist” justification for antisubordination if we can ascertain the precise sense in which social subordination is a state of affairs that harms members of subordinate groups by setting back their interests or well-being. It will prove instructive to consider a first pass at this task that fails.

One conspicuous harm of subordination is that it deprives members of subordinate groups of certain goods and opportunities for well-being by excluding them from areas such as employment, education, and housing that are critical for supporting their interests. But this cannot be the complete story. As others have pointed out, if antisubordination were limited to preventing this harm, it would be only a “weakly egalitarian” goal urging no more than “prioritarian regard.”¹⁷ That is, it would amount to a distributive imperative to prioritize increases to the absolute well-being of subordinated groups by securing their access to well-being-enhancing opportunities, which can in principle be realized without reducing social status hierarchies through the distribution of material resources. But antisubordination seems to be a strongly egalitarian goal. It recommends not just alleviating deprivation and suffering for the worst off but reducing disparities in status between privileged and marginalized groups.

We can begin to see how these disparities are distinctively bad by seeing that relative social status inequality between groups is a state of affairs that sets back the interest of members of subordinate groups in living in a social environment that is conducive to cultivating self-respect, i.e., in the “social bases of self-respect.”¹⁸ For people to exercise their capacity to autonomously author their lives by choosing and pursuing their own conception of the good, they must have “a secure sense of [their] value, a firm conviction that [their] determinate conception of the good is worth carrying out. Without self-respect, nothing may seem worth doing, and if some things have value for [them], [they] lack the will to pursue them.”¹⁹ Thus, an essential condition of autonomy is to live in a social climate that does not make it challenging for us to have self-esteem or self-confidence that our projects and aspirations are worthwhile and that we are worthy of realizing them. Being a member of a group with inferior social status relative to others is bad because it is inconsistent with the materialization of such a climate.²⁰

If social subordination deprives subordinated groups of the social bases of self-respect, we might conclude that it harms members of those groups mainly by frustrating fulfillment of their autonomously chosen plans. This constitutes a setback to well-being because human flourishing plausibly includes succeeding in one’s self-determined projects.²¹ However, it cannot be the main harm of subordination. Some people realize their aspirations even when deprived of the social bases of self-respect. And securing these bases cannot guarantee success in our projects (we are always susceptible to bad luck). The negative psychological experience of a loss of the self-respect that one needs to autonomously pursue a conception of the good is also not the main harm of subordination. A person need not suffer a felt loss of self-respect when deprived of the social bases of self-respect. They may happen to not view the social status of the group of which they are a member as influencing their identity. Or they may possess a particularly robust sense of self that makes them highly resistant to threats to their subjective self-worth. That, after all, is a critical role self-respect plays in our psychological lives.²² Thus, although the badness of insecure social bases of self-respect is intimately related to the importance of self-respect and autonomy for individual well-being, it is not the badness of not actually having self-respect or not actually being able to succeed in one’s aspirations. The main harm of subordination must be a setback to one’s interest in living in a social atmosphere that facilitates the psychosocial prerequisites for one’s successful autonomous pursuits and does not erect obstacles to securing them.²³

The process through which social subordination undercuts this atmosphere for marginalized groups reveals antisubordination’s strongly egalitarian credentials. It is plausible that “our sense of our own value, as well as our self-confidence, depends on the mutuality and respect shown us by others.”²⁴ Our ability to have a positive outlook on our identities and see ourselves as capable of conferring value on our chosen projects psychologically depends on reciprocal recognition by others in society that we have this ability and that our projects are valuable. Members of social groups who have inferior status in a status hierarchy do not receive this recognition. When social norms are constructed, consciously or not, to privilege and standardize the identities and interests of some groups, the ways of being of members of marginalized groups who are rendered abnormal, unnatural, or invisible are not reciprocally recognized as valuable. Nor is their capacity to select what is valuable to them. Instead, their conception of the good is censured and portrayed as inferior or vicious, they are rendered powerless to pursue it, and their roles and behavioral dispositions are stereotypically defined for them by privileged groups in juxtaposition to what is conventionally viewed as standard and natural. At worst, if they regard positive social valuations of the group to which they belong as important for their sense of personal identity, they may internalize negative societal valuations and come to see themselves as unworthy and incapable of investing value in their identities and choices.

Thus, when some social groups have higher status than others (in being socially constructed as the normal measure according to which we determine who gets power and authority, what counts as aberrant and warranting censure, and the expectations and capabilities of subordinated groups), our concern about the impact of this state of affairs on the social environment within which members of subordinated groups can develop self-respect and live autonomously is a concern not just about the absolute well-being of those groups but about the effects on their well-being of their social position relative to privileged groups. On this story, the goal of antisubordination amounts to reducing disparities in social status to prevent those harms.²⁵

However, it is often argued that what is objectionable about social subordination cannot be only its harmfulness to people’s well-being.²⁶ Moreau, for example, holds that discrimination is morally wrong because, by subordinating its victims to others, it is one way of failing to treat some people as equals to others.²⁷ When we fail to treat someone as an equal by discriminating against her, we do not just “act wrongly” by causing an outcome that is on balance bad; we “wrong” her specifically by acting inconsistently with her status as an equal.²⁸

Moreau maintains that we have the status of being one another’s equals because we collectively find ourselves as having among our moral beliefs a preexisting commitment to living in a “society of equals.” In such a society, everyone is treated as equal and independent agents with the same opportunity to participate in social institutions and political governance and live life in accordance with their conceptions of the good, insofar as this is compatible with recognizing others’ status as equals.²⁹ We have this commitment not because we accept some empirical fact about human beings from which we get our status; certain empirical facts about people have been said to bestow our status as one another’s equals, such as our rational faculties, but there are always some people, such as people with profound cognitive impairments, who lack these characteristics but who we still believe are equal to others. Our commitment is very difficult to abandon. So, rather than being grounded in nonmoral facts, it must be grounded in other, more foundational moral facts, namely, “the fact that our lives matter, and that each person’s life matters just as much as, and no more than, every other person’s life. In other words, we are all of equal moral value.”³⁰

On a “non-welfarist” view of the justification for antisubordination, then, when we subordinate others, we wrong them by disrespecting their moral status, such as their personhood or equality, violating a right or a claim they have on us that is grounded in that status. For example, we might violate what Samuel Scheffler calls an “egalitarian deliberative constraint” that requires our internal practical deliberations to give equal consideration to everyone’s equally important interests.³¹ Or our conduct might express a disrespectful social message of acquiescence to social status hierarchies that are inconsistent with persons’ moral equality.³² Violating someone’s moral status amounts to a wrong toward them that is independent of whether it is bad for them.

The non-welfarist view need not deny that we should resist social subordination, and acts contributing to it, in part because of subordination’s harmfulness. Indeed, the view that subordinating others is contrary to the victim’s moral status likely fails to explain all our concerns about subordination. If X contributes to Y’s social subordination to Z, Y has a grievance on her own behalf, but, plausibly, Z also has a grievance on his own behalf. Assuming that Y’s grievance is that X wrongs her by disrespecting her moral status, it is hard to see how Z’s grievance can be that he has been likewise wronged. It is unclear how we disrespect people with superior status when we contribute to the subordination of others to them. More plausibly, Z’s grievance is based on how status hierarchies set back the well-being of members of high-status groups. For example, it has been said that status hierarchies encourage privileged people to live in falsehood, depriving them of the good of knowledge, by adopting a false belief that they can draw self-respect from their high social status when in fact having such status is not a proper basis of self-respect.³³

Can a non-welfarist view capture concerns about subordination that a welfarist view cannot? It might be argued that if status hierarchies have bad effects on the social bases for self-respect for marginalized groups, the groups’ members must have a prior, independent concern about social inferiority as being inconsistent with their moral status, or else the lower status of their group would not be psychologically debilitating. Their exposure to the disrespectfulness of conduct perpetuating their social subordination must be what makes their awareness of their social marginalization harmful to their self-respect.³⁴

But it does not follow from the psychological fact that a person takes the subordinate status of the social group to which they belong as burdening her self-respect that she must regard subordination as objectionable for reasons that are independent of how it sets back well-being. She could believe that the reason is that subordination has bad effects other than harm to its members’ self-respect. For example, it may decrease their representation in a lucrative profession and make them materially worse off. As a result, she may view herself as unable to pursue a career in that profession and obtain resulting material benefits. Or subordination may diminish overall social cohesion and solidarity, which could heighten the risk of violent civil unrest.³⁵ As a result, fear and apathy may come to permeate the populace and make all citizens view their conceptions of the good as lacking in value. We can also be concerned about the bad effects of social subordination on self-respect without being concerned for the same reasons that members of low-status groups believe their lower status debilitates their self-respect, even if we assume that their reasons rest on considerations of their moral status. There does not seem to be anything problematic in an analogous position according to which, for example, we can be concerned about the bad effects of environmental degradation on future generations without being concerned about the reasons those generations would object to it, even if we assume that their reasons concern their moral status.

A different motivation for the non-welfarist view is the claim that there are cases where the intuitive problem with perpetuating social subordination is irreducible to causing harm. Here is an example that Andrew Mason calls “Local Shop”:

One tempting response is that even if the refusal is harmless on its own, it is still objectionable because it combines with a pattern of similar conduct to produce a hierarchical state of affairs that has harmful effects. Nonetheless, Mason suspects that “there is something objectionable about the individual acts involved independently of their role in generating harms. … These acts would seem to be morally problematic … even if they did not play such a role.”³⁸ This independent concern is said to be how the would-be patron’s act is inconsistent with “relations of mutual care and respect.”³⁹ If it somehow does not play a role in generating accumulative harm by participating in broader harmful patterns, it still expresses a disrespectful meaning toward the store owner’s status.⁴⁰ In stemming from stereotypical beliefs that disrespect her by being indifferent to, or outright rejecting, her moral status as an equal and the equal importance of her interests, it wrongs her in a non-welfarist sense by violating a special claim she has that others respect her status.

However, in the next part, I argue that the non-welfarist justification of subordination is not well motivated by examples of harmless acts that perpetuate social status inequality. These cases fail to show that such acts’ objectionableness is irreducible to their harmful effects.

My argument against the non-welfarist view rests on two principles:

• Defeasibility: If there is a duty to avoid contributing to social subordination grounded in respect for persons’ moral status as equals, it must be defeasible, and it is all-things-considered permissible to violate it if doing so produces an outcome that is sufficiently good for people’s well-being.

• Strength: To determine whether an outcome is sufficiently good such that the duty lacks the strength required for it to be all-things-considered impermissible to violate it, we must balance the good for people’s well-being that the violation produces against the violation’s badness for the duty’s beneficiary.

As I shall explain below, it follows from these principles that whether we must comply with a duty to avoid subordination depends fundamentally on how we weigh or balance effects on well-being. If the duty impinges on our practical reasoning via its strength, and its strength is a function of how important observing it is for its beneficiary’s well-being relative to competing considerations, what makes it ultimately matter is the importance of people’s well-being, not the importance of people’s moral status. Having offered this preview, I will proceed by defending each principle before drawing out their implications.

Suppose there is a duty to refrain from perpetuating social subordination grounded in respect for our moral status, such as our status as an equal. It would be hard to accept it if it were absolute. We would need to comply with it even when the costs of doing so measured by other values were extremely high.⁴¹ Hence the duty to treat others as equals is defeasible. We do not act wrongly when we wrong someone if we are justified in doing so.⁴² But how should we analyze this idea of defeasibility?

Sometimes, some indicia of status inequality obtain, including differences in authority and consideration, but acts that reinforce these differences do not appear to be all-things-considered wrong. For example, relations between employers and employees, teachers and students, parents and children, military sergeants and privates, and prison guards and prisoners involve differences in authority and power of one party over another but are justified overall. We justifiably afford greater consideration and esteem to people with extraordinary technical skills, job qualifications, athletic or intellectual talent, or moral virtue. We justifiably censure, blame, and punish people for moral viciousness and withhold opportunities from the unqualified.

Some say that subordination is not wrong when “differences in status hold only within a particular organization, rather than across a number of different social contexts, and only to the extent that those with the higher status require this status in order to fulfil the legitimate demands of their particular institutional roles.”⁴³ Differential authority structures are worse when they permeate society as a whole and are not restricted to isolated relationships. But your boss cannot tell you what books to read after you punch out, and your teacher cannot tell you where to eat dinner after class. And it is less problematic to show differential consideration and censure for certain qualities when these attitudes do not spill over to influence appraisals of the entirety of a person in all dimensions of their lives. A violinist’s musical skill may be esteemed, but that does not mean that others must give up their parking spots for her, and a person convicted of a crime may deserve proportionate punishment, but that does not mean that she has a wicked character that is forever incapable of rehabilitation.⁴⁴

Following Richard Arneson, I suggest that in the circumstances just described, the antisubordination duty is defeated because status hierarchies are all-things-considered justified by the beneficial effects for people’s interests.⁴⁵ Some institutions are so important for promoting well-being that, if they need to be structured by differences in power and authority to function well, we must tolerate the wrong of sustaining hierarchy within them, even if this disrespects people’s moral status, so long as the hierarchy is confined to those contexts and does not permeate society more widely. Thus, even if an employer or sergeant wrongs an employee or private when wielding authority, the benefits created by organizing businesses to maximize efficiency through hierarchical workplace relationships, and the national security benefits created by hierarchical military organizations, justify the wrong, all things considered. But since these benefits do not obtain outside these specific institutions that hierarchies enable to function effectively, they cannot justify hierarchies outside those contexts.

Similarly, sometimes showing elevated esteem for certain excellences people may exhibit, and condemning vices, may have good effects, making it permissible to breach the antisubordination duty, so long as one’s positive and negative appraisals do not spill over to totalizing judgments about others’ entire characters. Even if it disrespects some job applicants’ moral status to allocate greater opportunities to others with better education and experience, doing so is all-things-considered justifiable by the benefits accruing from getting the job done well. And even if it wrongs those who lie to chastise them for their dishonesty, this is justified by the beneficial effect of deterring them from dishonest behavior in the future. But these benefits would not accrue if attitudes of esteem or disesteem were shown to aspects of these individuals’ characters other than the specific socially beneficial or detrimental qualities they exhibit.

It might be objected that when differences in authority and esteem are permitted in the circumstances outlined above, this is not due to how a person lacking them is wronged by being disrespected, with the wrong justified by beneficial outcomes. For there is no wrong to justify in the first place. When subordination is restricted rather than permeating society as a whole and does not involve totalizing judgments of a person’s whole character, reinforcing it does not disrespect people’s moral status as equals.

If you find this objection attractive, you must explain why contributions to some hierarchies of authority or esteem (say, between students and teachers or differently skilled chess players) are respectful and compliant with persons’ moral status when contributions to others (say, between men and women or white and Black people) are not. For example, discrimination may wrong its victims when it subordinates groups characterized by a “socially salient” property, which is taken throughout society to have implications for beliefs about the entire character of the people they characterize. So, perhaps status hierarchies are not disrespectful when they are not constructed around such properties, but they are disrespectful when they are so constructed.

But it remains unclear why maintaining hierarchies between socially salient groups is disrespectful while maintaining them between non-salient groups is not. It has been said that “grounds for according esteem” are disrespectful when they pick out certain groups, such as minority racial groups, who have experienced a history of widespread social oppression.⁴⁶ This suggests, however, that the socially salient groups that constitute the forbidden grounds of subordination are groups whose members have historically been harmed the worst by status inequality. Members of these groups have had their interests in the social bases of self-respect set back the most given how they have suffered from inferior social status across many social contexts, instead of only within specific institutional contexts, and given how censorial attitudes toward them have been long taken to influence judgments about their entire being. If this is correct, what distinguishes, for example, a benign teacher/student status hierarchy from a malignant racial hierarchy is not that it is not disrespectful to the student. It is that it is less harmful to the student than the racial hierarchy is to members of the disfavored race. But then the notion of socially salient traits fails to tell us why, when subordination is not troubling, it is not disrespectful. Hence, it does not help establish that some differences in authority and esteem, i.e., those that do not revolve around socially salient traits, do not breach an antisubordination duty at all instead of justifiably breaching it. We have yet to witness a viable alternative to the analysis of the duty’s defeasibility offered earlier according to which subordination is not troubling although it violates an antisubordination duty, when the violation is justified by beneficial effects for people’s interests.

I think it favors that analysis that, when status inequality is not all-things-considered wrong, it often still seems to be intuitively lamentable in one respect, even if only slightly. Simply from a narrow egalitarian point of view at least, it would be best if, for example, businesses could operate efficiently without employees having to lack power relative to employers; children came into the world with fully developed physical and psychosocial abilities enabling them to be free from parental authority even from nonage; nobody benefited from competitions in esteem, skill, or beauty; and nobody ever had to be praised for moral magnanimity or denounced for moral viciousness. The fact that certain parties are in a position of regrettable vulnerability in virtue of their lack of power and authority is why legal systems impose strict rules on certain otherwise desirable relationships, like employer-employee, parent-child, and teacher-student relationships, to protect that party from abuse.⁴⁷ By analogy, the fact that unequal authority between state and citizen appears to be inconsistent with respect for our status as free is what generates the puzzle of how to justify state power and gives defenses of anarchism traction.⁴⁸ Moreover, it is often recognized that the disesteem of punishment has a lamentable cost for an offender’s social status, whether it is effectuated through institutionalized criminal proceedings or social criticism and calls for “cancellation” in culture and media. That is why, despite its benefits, there are strict procedural requirements to ensure the fairness of punishment in the legal context and why there are demands for similar “due process” protections outside that context.

So, often when we depart from conditions of equal authority or consideration, something is lost, sacrificed, or left on the table. We cannot explain why such a cost is generated, even a negligible one, unless members of lower-status groups are wronged by having an antisubordination duty owed to them violated and having their moral status disrespected, with the wrong being all-things-considered justified. If there is no violation that is amenable to justification, nothing is left behind to feel compunctious about.

If, as I have claimed, a putative antisubordination duty grounded in people’s moral status must be defeasible, we need to know how to determine its strength compared to countervailing benefits of breaching it so we can evaluate whether it can be justifiably overridden. There are at least two options here. First, the duty’s strength depends on how severely its violation disrespects the beneficiary’s moral status, where this is separate from how much it sets back her well-being. Second, it depends on how bad violating it would be for its beneficiary’s interests.

One worry for the first option is that the notion of a moral requirement having a weight, which may or may not be outweighed by other values, that is irreducible to the degree to which breaching it is bad for people seems mysterious. Consider how W. D. Ross described how to deliberate in cases of conflict between “prima facie duties.” Denying that a conflict “should be resolved by asking ‘by which action will most good be produced?’” Ross said, “[I ought to] study the situation as fully as I can and form the considered opinion (it is never more) that in the circumstances one [duty] is more incumbent than any other; then I am bound to think that to do this prima facie duty is my duty sans phrase in the situation.”⁴⁹ But it seems mysterious how a prima facie duty can have even a rough degree of “incumbency” compared to that of another without appealing to how bad it would be to violate that duty compared to violating the other. Referring to the duty’s incumbency seems to amount to nothing more than stating that it has decisive weight or strength without explaining why it has it. It seems similarly mysterious how the antisubordination duty could have a degree of incumbency without comparing the burdens of violating it for its beneficiary’s interests with any benefits for other people’s interests.

Another worry is that if the duty’s strength is not measured on the same scale as benefits to people’s well-being that could justify its violation, it is mysterious how trade-offs between its importance and the importance of those benefits are possible. Suppose by analogy that we have a duty to refrain from destroying a certain sculpture that is grounded in aesthetic value, not in any person’s well-being. One villain asks us to destroy it as a condition of paying for a poor student’s college tuition. Another offers to pay the same tuition only if we break another student’s nose. If we want the tuition to be paid, we must compare the strength of the aesthetic duty to conserve the sculpture to the good of the tuition payment for the poor student and then the badness of the broken nose to that good. Whatever we decide to do, the latter comparison, where the comparators are both changes to well-being, seems less mysterious than the former, where the comparators are not the same. No doubt, comparisons between changes to well-being rest on the difficult task of finding a measurement scale shared between quite different aspects of well-being, such as physical pain and deprivations of resources needed for human flourishing. But my point is that if it is hard to construct a shared measurement scale within the singular order of well-being, it must be ever harder to construct such a scale across the multiple orders of well-being and non-welfarist values, such as aesthetic value. The latter is the scale presupposed by the defeasibility of the antisubordination duty if the duty’s strength is rooted in the importance of the beneficiary’s moral status in a non-welfarist sense.

It might be argued that it is not mysterious how disrespect to a person’s moral status as an equal comes in degrees. Some breaches intuitively have a worse impact on a victim’s status than others. For example, giving less credence to the opinion of a woman in a workplace and explicitly offering her a promotion in exchange for sexual favors both disrespect her status. But the latter contribution to women’s social subordination seems to be more disrespectful to women’s status than the former. Or if it is disrespectful for an airline to contribute to the diminished social esteem of unskilled pilots by not hiring them, it is more disrespectful for it to contribute to the social subordination of Black people by not hiring them. Thus, to evaluate the strength of the antidiscrimination duty, we simply evaluate how disrespectful it is to breach it compared to the benefits for well-being from the duty’s breach.

However, degrees of disrespect are not really doing the work in intuitively differentiating between the severities of violations of the antisubordination duty. What does the work is how badly a violation sets back its beneficiary’s well-being.⁵⁰ For example, given the historical and socially widespread oppression of Black people, exacerbating a social status hierarchy between Black and white people is more harmful to Black people’s interest in secured social bases of self-respect than the harm to the same interest of unskilled pilots caused by contributing to a hierarchy of consideration between them and skilled pilots. And even though epistemic disregard and overt sexual harassment at work each contribute to the subordination of women, harming their interest in the social bases of self-respect, the greater importance of sexual autonomy for one’s well-being relative to one’s opinion being taken seriously at work explains why sexual harassment is comparatively worse.

This leads us to embrace the view that the duty’s strength is a function of the harmfulness of violating it to its bearer’s interests. On this view, to determine whether it is all-things-considered permissible to violate the duty, we multiply the size of the base value harm done to the victim by some fixed amount to generate a figure that can be compared to the expected good results.⁵¹ To illustrate, we can explain why, to save five lives, it is all-things-considered justifiable to violate a duty to not kick someone but it is unjustifiable to violate a duty to not kill her, if the latter is stronger than the former in virtue of being kicked is not as bad for the person as being killed. The badness of being kicked is outweighed by the goodness of saving five lives, whereas the badness of being killed is not. This approach leaves open how much better the good brought about by a violation must be compared to the badness of the violation for the beneficiary (depending on what factor the multiplier dictates we must apply to the base value harm). But it demystifies the question of the duty’s strength, and the method for comparing it to the interests promoted by the violation, by locating each comparator on the singular scale of well-being.

By extension, we can determine whether the duty to refrain from contributing to social subordination is strong enough that compliance with it is required notwithstanding any good of violating it by comparing that good to how bad for the duty’s beneficiary it is to violate it. Suppose that the violation sustains the lower status of a group characterized by a socially salient property, which is taken throughout society to influence judgments about the abilities and roles of the group’s members. Then, it is very harmful to these persons’ interest in the social basis of self-respect. It would be more harmful than it would be if it were only to sustain a localized status hierarchy that does not involve totalizing judgments about the characters of members of lower-status groups. Conversely, if the benefits of a violation for overall societal well-being are very great, they are more likely to justify the violation when it is least harmful to the social bases of the beneficiary’s self-respect, such as when it does not contribute to the subordination of a socially salient group to which the beneficiary belongs.

I have argued that if there is a duty to avoid contributing to social subordination grounded in respect for the moral status of persons, it must be defeasible, and it is all-things-considered permissible to violate it if doing so produces an outcome that is sufficiently good for people’s well-being. Furthermore, to determine whether an outcome is sufficiently good, we must balance the violation’s badness for the duty’s beneficiary against the goodness it produces. With these principles on the table, we can now revisit Mason’s Local Shop case.

The storeowner, we are supposing, has not been harmed. The man’s refusal to shop at her store has not set back her interest in the social bases of self-respect. Suppose that he owed her a duty to not perpetuate social subordination grounded in the imperative of respect for her moral status. If this duty’s weight is determined by a function mapping harm to strength through a multiplier, it follows that if the harm caused by its violation is nil, the right’s strength is nil. Being weightless, the right can be justifiably overridden if any benefit is thereby produced, however small.⁵² For example, it might be argued that the shopper has experienced an improvement in his well-being because he has been able to exercise his freedom of association. If so, his conduct was all-things-considered permissible. If he wronged the storeowner, it was not wrong for him to do so.

Now, even when the duty’s strength is nil in virtue of its violation not being bad for its beneficiary, violating it might still seem intuitively impermissible if no gains result for anyone affected by it. Thus, if we deny that any benefit accrues from the man’s act in Local Shop, he can still owe the storeowner an all-things-considered duty to not perpetuate her ethnic group’s subordination. This duty, it might be said, must be grounded in the storeowner’s moral status.

But this argument presupposes that disrespecting a person’s moral status is irreducible to setting back her well-being. Ignoring any other respects in which a person may be harmed, it does set back her interests to treat her in a way that is inconsistent with her status insofar as her status requires that her interests ought to be appropriately accounted for in others’ practical deliberations. That is often what proponents of the non-welfarist justification of antisubordination seem to believe when describing what we must do to respect people’s status by refraining from contributing to status hierarchies.⁵³ If respect requires us to account for everyone’s interests appropriately, we have a second-order interest in being treated in accordance with our status by having our interests properly accounted for in others’ deliberations, including our interests in secured social bases of self-respect.⁵⁴ If we have such an interest, and in Local Shop the man disrespects the storeowner’s moral status in violating an antisubordination duty he owes her without creating any benefits, we can maintain that the man still ought to comply with this duty because doing so protects her well-being. For she does, after all, suffer some distinctive harm when the duty is breached. This is the same harm that anyone suffers whenever a right they have that is grounded in their status is violated, even if the violation is not bad for any other aspect of their well-being.

Suppose, then, that we think about the strength of a right as a function that maps harm to the bearer onto strength through an additive factor. After multiplying the base value harm, the function adds to it some standardized value that reflects the severity of the setback to one’s second-order interests that anybody suffers whenever their status is disrespected.⁵⁵ On this approach, even when the base value is nil because the violation of a right is not otherwise bad for the bearer, the function can yield an above-nil strength that will be equal to the standardized value. So, there can be an above-nil-weighted antisubordination duty correlating with the right when violating the duty does not appear to be otherwise bad for the person and no benefits ensue from disregarding the duty. If the duty is fairly said to be grounded on the importance of respecting a person’s moral status, the importance of respect gives the duty significant strength only because of how disrespecting a person’s status is bad for the beneficiary’s well-being; it sets back a second-order interest in having one’s first-order interests properly accounted for.

In sum, Local Shop is an example either of where it is permissible to violate an antisubordination duty grounded in people’s moral status or of where we are required to comply with the duty because of how it protects its beneficiary’s second-order interest in having her moral status respected. Neither reading enables the case to provide a compelling motivation for a non-welfarist justification of antisubordination. The case either highlights a moral consideration that does not ultimately matter compared to others or supports a welfarist justification as truly animating the non-welfarist one.

More generally, it follows from the principles established above that the duty to not perpetuate subordination, which Local Shop might be taken to support, does not play a role in guiding our practical deliberations apart from identifying some aspect of the storeowner’s well-being that we must not set back without producing a sufficiently good outcome. Even assuming that the duty is grounded in the storeowner’s moral status, whether we are required to comply with it depends on what input the interest it protects has into the process of balancing these burdens and benefits. In other words, what makes the duty matter for our practical reasoning lies not in whether noncompliance with it wrongs the storeowner by disrespecting her moral status. It lies in the influence that its beneficiary’s well-being exerts on our deliberations about whether noncompliance with the duty is all-things-considered permissible. In the end, if we agree that our moral status as equals is rooted not in some nonmoral fact about human beings but in some further moral fact, that further moral fact is not that we all have equal moral value, if this is given a non-welfarist elaboration. Rather, it is that what fundamentally matters is well-being and preventing harm to it.

There is, however, a potential problem confronting the welfarist view. This problem concerns the view’s fit with the main features of antidiscrimination law and the potentially better fit of the non-welfarist view. The view that the law aims to reduce states of affairs of subordination that are harmful to members of groups with historically inferior social status seems to conflict with how, in addition to protecting members of social groups with historically inferior social status from discriminatory acts, the law symmetrically protects members of groups with historically superior status.

More precisely, antidiscrimination law typically describes the prohibited grounds of discrimination in terms of “universal” properties, such as “race,” “sex,” and “age.”⁵⁶ These apply to all people. Everyone has a race, a sex, and an age, just not the same race, sex, or age. The universal properties divide into more “particular” properties instantiating them, such as being white or Black, male or female, or under sixty-five or over sixty-five. Usually, one of these particular properties characterizes a social group that has been historically subordinated to others, such as Black people, women, and elderly people. But the law symmetrically prohibits discriminatory acts based on any particular property that the universal properties listed in the prohibited grounds divide into, regardless of whether it defines a marginalized or privileged group. It prohibits discrimination because of, for example, a person being white, a man, or any age.⁵⁷

It seems puzzling why the law has this structure if it aims at alleviating the harms of social status hierarchies for subordinated groups. To achieve this goal, it would make sense for the law to asymmetrically prohibit discrimination only because of particular properties that define subordinate groups. It is unclear how discriminatory acts against members of privileged groups exacerbate their social inferiority and set back their interest in the social bases of self-respect. So, it is unclear how symmetry reduces status hierarchies.⁵⁸

Now, notwithstanding first impressions, compelling arguments for symmetry have in fact been advanced. First, in protecting members of high-status social groups, the law aims to prevent “backlash,” or sociopolitical resistance against initiatives to reduce the subordination of low-status groups. It conveys a message of evenhandedness that bolsters the law’s legitimacy as perceived by privileged persons and thus mitigates the potential for these persons to believe that they have been victimized by the state’s preferential treatment of subordinate groups.⁵⁹ Second, discrimination against the privileged can reinforce the subordination of marginalized groups. So, preventing it helps combat the latter. For example, discrimination against men based on stereotypes that they are the standard familial breadwinners and not caregivers to children can reinforce stereotypes that women are not proper breadwinners and that their main social role is to care for children. Requiring employers to give parental leave to men in addition to women thus attacks the social norm that women are “mothers first and workers second” by attacking the “interlocking stereotype” of men as “workers first and fathers second.”⁶⁰ These arguments for symmetry concentrate on how symmetry is a tool for reducing the status inequality of socially inferior groups. A welfarist justification can be given for this approach: symmetry helps prevent setbacks to the interests of members of subordinate groups in secured social bases of self-respect.

Having said that, we can also give a non-welfarist explanation for symmetry by holding that discrimination against members of privileged groups should be prohibited for a reason other than to prevent harm to marginalized groups. Recall Local Shop. Suppose that the storeowner in that case belongs to a majority rather than a minority ethnic group. If the man’s refusal to patronize the store because of the owner’s ethnicity does not contribute to the harmful social subordination of a minority ethnic group, we still might find it objectionable because the man violates a right the storeowner has based on her moral status as an equal. Members of privileged groups presumably have such a status. The man disrespects it, whether the storeowner is marginalized or privileged, by being motivated to act on a belief in her inferior status.⁶¹ Hence, it might be thought that antidiscrimination law’s symmetry reflects the need to vindicate everyone’s status as equals.⁶²

This non-welfarist justification differs from the justification considered above that focuses on preventing harm to marginalized groups in postulating that discriminating against members of privileged groups generates a grievance that members of those groups have on their own behalf. That might give the non-welfarist view an advantage over the welfarist view. We can appreciate this advantage by observing how symmetry is implicated in affirmative action. Affirmative action aims to reduce social subordination by benefiting a person because she belongs to a marginalized group. In the process, it often involves the commission of discriminatory acts that disadvantage other people who belong to privileged groups.⁶³ This has been called “collateral” discrimination.⁶⁴ Collateral discrimination is usually prohibited by the law’s symmetry. Most jurisdictions treat it as regrettable, even if only slightly so, or as involving a cost, even if it is negligible. This cost can be all-things-considered justified by the importance of benefiting the worse off. But it manifests in how, to be legally valid, affirmative action must meet strict proportionality criteria, such as advancing a valid purpose using measures that are rationally connected to and necessary to achieve it.⁶⁵ If collateral discrimination did not leave behind a cost to be compunctious about, there would be no need for these criteria to rigorously delineate the conditions under which affirmative action is acceptable.

How should we explain this cost? We might claim that, when we permit collateral discrimination to pursue affirmative action, we thereby forgo the gain that symmetry offers for reducing the subordination of groups with inferior status and preventing harm to them. But this seems to be the wrong kind of explanation. Controversies over affirmative action are intuitively not best viewed as debates over which of two methods for advancing antisubordination we should select: an oblique tactic of symmetrically prohibiting discrimination versus straightforwardly conferring benefits on members of subordinate groups. If that were the case, it is hard to see how there would be a controversy at all; the straightforward approach seems to be obviously preferable. Rather, the controversies seem to revolve around how collateral discrimination generates a grievance that privileged people have on their own behalf, not just on behalf of marginalized people. That is how the cost of collateral discrimination is framed in legal discourse. For example, the Supreme Court of the United States has written that affirmative action in employment contexts that aims to benefit marginalized racial groups or women must not “unnecessarily trammel the interests of” white people or men.⁶⁶ A non-welfarist justification of symmetry that explains it in terms of how discriminating against members of privileged groups disrespects their status seems to be uniquely capable of capturing the grievance that members of these groups have on their own behalf and therefore why committing collateral discrimination to pursue affirmative action is controversial.

One might argue that it is only a myth that members of privileged groups have a grievance on their own behalf when they are discriminated against. The myth is useful to propagate because if privileged groups accept it, they are less likely to resist antisubordination for the worst off and more likely to bring lawsuits that deter discrimination against them, which discrimination, as we saw, can reinforce the lower status of subordinate groups. Legal actors only need to believe, even if falsely, that the grievance exists for controversies over affirmative action to revolve around the putative clash between it and the importance of benefiting members of subordinate groups. But by attributing widespread error to legal actors, this defense of the welfarist view explains the controversies as rooted in systematic deception. It makes them opaque by holding that the reasons people think they are engaging in them are not real. A more transparent defense of the welfarist view would avoid this attribution of error.⁶⁷ I shall present, in what follows, an alternative that does.

Recall the vocabulary we used above to describe how discriminatory acts can contribute to social subordination. They “contribute to,” “perpetuate,” or “sustain” subordination. I suggest that the notions of contribution, perpetuation, and sustenance are unified in having what we might call both “retrospective” and “prospective” senses.

By analogy, consider a baseball team that perpetuates a ten-game losing streak by losing an eleventh game on a particular night. The problem with the eleventh loss is retrospective. It piles onto a past string of ten losses and exacerbates an already bad situation. But it is also prospective. It relates to that past string by keeping the streak alive into the future and further dimming the prospects of it ending. The team has been losing, it continues to lose, and this bodes ill for its future chances of success.

We can invoke a welfarist framework to conceptualize the problem with perpetuating social subordination as similarly both retrospective and prospective. It is bad for the well-being of members of groups with lower status. It sets back their interest in the social bases of self-respect. Retrospectively, it adds to and compounds prior harms of subordination that these groups have suffered. Prospectively, it simultaneously adds a new ingredient into the mix that worsens those harms and makes them persist into the future when they otherwise might not have.

These observations allow us to see why discriminatory acts against privileged groups can be viewed as contributing to a state of affairs of social status hierarchy that is bad for the well-being of members of these groups. True, they cannot do so retrospectively. The privileged groups we are familiar with—white people, men, heterosexual people, able-bodied people—have never suffered the harms of inferior social status in the past. Discriminatory acts against them cannot compound their past subordination. But they can contribute to subordination prospectively. They can play a causal role in the materialization of now-privileged groups’ social status inferiority in the future, which might not otherwise come into existence. This potential state of affairs can one day set back the interests in the social bases of self-respect of members of these groups. So, antidiscrimination law’s symmetry is a prophylactic that guards against this potential harm.

One might object that it is possible to contribute to social status hierarchies prospectively only by simultaneously contributing to it retrospectively. One must, that is, compound, exacerbate, worsen, or carry forward preexisting harmful subordination. Thus, it is possible to contribute to the harmful social subordination of only historically marginalized groups. But recall the baseball team’s losing streak. Even without the benefit of hindsight, it seems plausible that the problem with the first loss in the team’s streak, at the time of that loss, was that, prospectively, it contributed to the future materialization of a losing streak even though there was no extant losing streak that it could have exacerbated. The bad streak had to get off the ground somewhere. It got off the ground with that first loss. That loss gets its badness from how, at the time, it had the potential to launch a bad streak.

Similarly, suppose, hypothetically, that we could think back to the very first discriminatory act against a member of a given historically low-status social group. Even without the benefit of hindsight, it is plausible that the concern about that first act, at the time it was committed, was how, in a purely prospective sense, it contributed to the manifestation of a prospectively harmful state of affairs of status inequality for these people even though there was no prior status inequality it could have worsened. The current subordination of these people had to get off the ground somewhere. It got off the ground with that hypothetical first act, which, at the time, got its badness from its potential to inaugurate a harmful state of affairs of social status inferiority.

If discrimination against privileged groups makes a prospective contribution to social subordination that is bad for the well-being of these groups’ members, the grievance it generates is on behalf of these people themselves and not only members of groups with inferior status. This allows us to explain the nature of controversies over committing collateral discrimination to pursue affirmative action without having to resort to how collateral discrimination disrespects its victims’ moral status in a non-welfarist sense. Of course, the fact that symmetry furthers antisubordination for privileged and marginalized groups alike is compatible with how it is much more urgent to pursue that goal for marginalized groups. Because of how discrimination against the privileged contributes to subordination only prospectively, the harm it causes to people’s interest in secured social bases of self-respect is much less serious, much less immediate, and much more speculative than the harm to these interests caused by how discrimination against people with lower status contributes to status inequality both retrospectively and prospectively. That is why, although affirmative action might generate controversy and a slight cost in the form of the minimal badness for members of high-status social groups of collateral discrimination, it is all-things-considered justifiable insofar as the relevant proportionality criteria are satisfied.

One might worry that the account of symmetry I have defended proves too much. Presumably, discriminating against any person because of any property whatsoever, whether it characterizes a historically marginalized social group or not, such as having green eyes or being a Scorpio, could contribute to the subordination of the group in a purely prospective sense. If that justifies symmetry, why would antidiscrimination law’s prohibitions be restricted to the prohibited grounds we are familiar with? Shouldn’t it prohibit discrimination because of any property whatsoever?

In reply, an account of symmetry is not implausible just for entailing that antidiscrimination law could protect even groups that are unfamiliar or surprising. It might be prudent to protect groups that have never been subject to social subordination because it can be difficult to predict what groups may become subordinated in the future and thereby exposed to harm to the social bases for their self-respect. Out of an abundance of caution, protecting certain groups serves as a prophylactic against this potentiality. This design technique has been called “adaptive breadth.”⁶⁸ It makes the law adaptable to unforeseen contingencies while avoiding the political inexpediency of having to amend it in the future.

It is easier to predict that some social groups who have not had inferior status require protection against discrimination more than others. These are the privileged groups defined by a particular property (white people or men) that instantiates a universal property (race or sex) that is also instantiated by a different particular property that does define a historically marginalized group (Black people or women). History teaches that social status hierarchies that are organized around the universal properties typically found in the legally prohibited grounds, with groups defined by one particular property subordinated to groups defined by another, are special threats to people’s interests in secured social bases of self-respect. This allows us to predict that the privileged groups under these historical hierarchies, such as white people or men, might be particularly susceptible to becoming subordinated in the future and that that subordination will be bad for them. But this prediction is unavailable when it comes to groups who are defined by a particular property instantiating a universal property, such as eye color or astrological sign, that is not also instantiated by a different particular property that defines a historically marginalized group.

It turns out, then, that discrimination against privileged groups that receive symmetrical protection under actual antidiscrimination law contributes to subordination in an attenuated retrospective sense as well as a prospective sense. For example, discrimination against men perpetuates the historical existence of harmful sex-based social status hierarchies at a universal level, although evidence of the harmfulness of this state of affairs that warrants guarding against them comes from historical evidence of its harmfulness for women. But discrimination against Scorpios, for example, does not perpetuate any such universal-level hierarchy, so legal protection for Scorpios is unwarranted.

In this article, I have argued that the prevailing non-welfarist view of antisubordination is unmotivated. We should instead view the goal as justified only because of how social subordination is bad for people’s well-being. I have also replied to the symmetry objection against the welfarist account of the antisubordination interpretation of antidiscrimination law. Along the way, I have offered clarifications of the harmful impact of social status hierarchies on people’s interest in the social bases of self-respect, the strongly egalitarian nature of antisubordination goal, how to distinguish justifiable from unjustifiable differences in social status, and controversies over affirmative action.