Equal and ashamed? Egalitarianism,anti-discrimination,and redistribution

“Around the corner from my office at Ranchi University in eastern India is a wide avenue that leads into the city, called Kanke Road. If you go there at the crack of dawn, you will see something you may never forget: hundreds of young men, many emaciated and disheveled, pushing bicycles loaded with more than two hundred kilograms of smuggled coal. … These young men are known as koilawalas (coal guys).”

However, distinguishing between inequalities which are objectionable in this sense and inequalities which are less objectionable is a difficult matter. Closer reflection on this distinction raises concerns. In principle, the state would need to find out whether any of the koilawalas “drank or gambled away their land.” ² Implementing this would require the state to adopt an attitude of mistrust toward the koilawalas. Even worse, as Drèze's statement makes clear, hardly any of the koilawalas bear responsibility for their fate. Yet, all would be put under suspicion. In proving that it is not their fault, they would have to portray themselves as unable to earn a decent living. This admission, that one's skills are insufficient to make a living, can be shameful (Wolff, 1998). In unflattering terms, the process brands you as untalented, low-skilled, stupid, or the like. As I understand it, shameful revelations are a concern about the practice of luck egalitarian justice.

There is a second concern about the theory of luck egalitarian justice. It suggests that the state adopts a problematic judgment toward those whom it helps via redistributive measures. Redistributive policies toward those at the lower rungs of the income distribution are devised because these citizens are stupid, socially undesirable outcasts, or otherwise defective, and not because they are equal citizens. Feeding this suspicion Thomas Nagel (1979: 104) contends that “[w]hen racial and sexual injustice have been reduced, we shall still be left with the great injustice of the smart and the dumb, who are so differently rewarded for comparable efforts.” Such an attitude by the state is insulting and humiliating even if the state never communicates disparaging messages (Anderson, 1999: 304–307). The state would nonetheless pass judgment on the recipients of redistribution and deem them as inferior. The reasons that luck egalitarianism identifies as the basis for redistribution are humiliating.

To clarify my terminology: What I shall call the humiliation objection is concerned with the theory of luck egalitarianism. It holds that luck egalitarianism should be rejected as a principle of justice. What I shall call the shameful revelation objection is concerned with egalitarian public policy. It holds that a straightforward application of luck egalitarianism into policy is problematic and that, therefore, public policy should not seek to emulate pure luck egalitarian justice.

Some critics of luck egalitarianism believe that a common diagnosis underpins these two objections. Both objections stem from a failure to adequately recognize interpersonal wrongs. Luck egalitarianism, for these critics, is caught in an impersonal distributive paradigm. Ultimately, for luck egalitarians what matters is the equalization of some specified “currency.” Yet in the examples mentioned, the problem is that the state is wronging citizens by requiring them to demean themselves or by communicating humiliating messages (Anderson, 2010a; Axelsen and Nielsen, 2020). More specifically these wrongs constitute failures among citizens to relate to each other as equals.

This suggests a change in focus. We should focus less (directly) on material inequality and its causes. Instead, we should focus on building a society in which everyone can relate to others in a position of equality. This position has been called relational egalitarianism (Anderson, 1999; Fourier, Schuppert and Wallimann-Helmer, 2015; Miller, 1997; Nath, 2020; O’Neill, 2008; Scheffler, 2003). ³ Sometimes inequality in living conditions is grave enough that it translates into status differences, social hierarchies, or makes it otherwise impossible to relate as equals. But oftentimes it does not. When it does not, there is no egalitarian reason to be concerned with such inequality.

In this article, I challenge the idea that concerns about humiliation or shameful revelations speak against luck egalitarianism and in favor of a sole focus on relational equality. My main thesis is that relational egalitarianism, as such, is guilty of the same charge. ⁴ My argument relies on an intricate connection between relational egalitarianism and a particular design of anti-discrimination laws which I lay out in the first three sections. In the next four sections, I argue that the kind of humiliation and shameful revelations inherent in anti-discrimination law is on a par with the humiliation and shameful revelations inherent in luck egalitarianism. The last two sections discuss what this should mean for concerns about humiliation and shameful revelations respectively. In the case of humiliation, this leads me to conclude that the objection disappears. In the case of shameful revelations, it leads me to conclude that what matters is the design of public policy.

The bottom line is the following: In an unequal society, either approach to egalitarianism needs to confront and recognize inequality before addressing it. This process of recognizing inequality can be shameful for those identified as underdogs in society. Sometimes this shame gives us reason to modify our policy response, but oftentimes it is a price we will have to pay in pursuit of a more equal society. While my argument serves as a defense of luck egalitarianism, this should not undermine the concerns of relational egalitarians. An egalitarian society requires both reducing the gap in living standards due to bad brute luck and quashing systems of social hierarchy. ⁵

Relational egalitarianism and non-discrimination

The idea of relating to others as equals is highly abstract. A complete theory of relational egalitarianism needs to spell out more concretely what it means to relate to others as equals. Different proponents of relational egalitarianism will inevitably disagree and give different accounts of what constitutes equal social relations. However, in one sense it may not be necessary to explain what characterizes equal social relations. This is because we know paradigm cases of unequal social relations. Relational egalitarians can start by focusing on these paradigms of inequality. This approach also helps to understand relational egalitarianism as such, as opposed to a specific interpretation of it, because all accounts of what makes social relations equal should condemn certain paradigmatic unequal social relations. For some relational egalitarians the focus on unequal social relations may even be more fundamental. One proposal for how to understand relating to others as equals defines it negatively. You relate to others as equals if you do not oppress, stigmatize, discriminate, and so on (Wolff, 2015). ⁶

The important point is that all relational egalitarians agree that their theory identifies problematic forms of unequal social relations. One such paradigm is discrimination. Discrimination involves treating people worse than others on account of some social group membership. Discrimination is also an interpersonal wrong. Different accounts of what makes discrimination wrong can be seen as describing unequal social relations and how discrimination wrongs its victims by treating them as inferiors. Deborah Hellman, for example, argues discrimination is wrong when it demeans. Demeaning in Hellman's sense requires a demeaning message and the ability to put down the recipient of the message (Hellman, 2008; also Anderson and Pildes, 2000: 1533–1544). Sophia Moreau argues that one of the ways in which discrimination is wrong is by causally contributing to status differences which are linked to people's de facto authority and deference they receive from others (Moreau, 2020a: ch. 2; also Kolodny, 2014; 2023: esp. ch. 5). Benjamin Eidelson argues that discrimination is wrongful when it disrespects. He understands disrespect as a lack of recognition respect in which the discriminator does not give the interests of the discriminatee appropriate weight (Eidelson 2015). ⁷ Each account picks out and highlights a different way in which discrimination treats people as less than equals.

This strengthens the conclusion that the core idea of relational egalitarianism is closely related to the aims of anti-discrimination law. In this spirit, Sophia Moreau (2020b: 172) suggests that spelling out “relational equality sounds very much like a description of the aims of discrimination law.” Since relational egalitarianism focuses on relations between citizens and not just on relations between the government and citizens, the legal regulations will have to encompass the actions of private citizens as well. Anti-discrimination laws in fact usually cover a range of private interactions. They cover economic transactions, housing, privately provided education or health care, and so on. Only some private actions, like invitations to your birthday party or marriage decisions, invoke powerful privacy concerns that justify exemptions from anti-discrimination laws. ⁸ The focus on private conduct should not obscure the fact that relational egalitarians also focus on the state. Discrimination by the state, as much as discrimination by private persons, wrongs citizens by failing to treat them as equals.

Anti-discrimination law and (a)symmetry

Relational egalitarianism tells us not only that anti-discrimination laws are justified, but also something about the content of such laws. One approach to anti-discrimination laws is concerned with the classification of individuals (for a defense see Brest 1976; for criticism see Fiss 1976). This approach suggests that the problem with discrimination is that individuals are judged on the basis of their race, sex, age, caste, disability status, or the like when benefits or burdens are distributed. To avoid discrimination, we should insist on treatment that is color-blind (and sex-blind, caste-blind, and so on). The aim of this anti-classification approach is laudable. Relational egalitarians, too, wish for a society in which people will be judged “by the content of their character and not by the color of their skin” to quote Martin Luther King. But King was acutely aware that society had not reached this point. It still has not reached this point as Sylvia Law (1999: 2) makes clear: “Black people invariably note their race and white people almost never do. … It is a privilege not to have to think about race.” The point is not limited to race in the United States. In a poem by Indian poet Akhil Katyal (2018) a ten-or-twelve-year-old asks his mother what his caste is. The mother replies “if you don’t know it by now, it must be upper.” ⁹

Relational egalitarians should reject color- or caste-blindness in present, non-ideal circumstances of societies marred by racism and caste. The mere fact that a policy or action uses race or caste as a criterion does not establish that there has been wrongful discrimination. Excluding African Americans from law school or setting a ceiling on the number of African Americans allowed into law school wrongfully discriminates against them. There is no justification for such a ceiling which does not involve invidious prejudice or bias. But a Caucasian or Asian American candidate could not raise the parallel objection. ¹⁰ A formal or informal quota system which effectively sets a ceiling on such applicants can be justified without reference to invidious prejudice or bias. It could be justified, for example, as a means to provide better legal services to African Americans in non-ideal conditions in which Black lawyers are most likely to do so. There is no wrongful discrimination in this case (Dworkin, 1977: ch. 9).

Consider another example: statistical discrimination in law enforcement. Law enforcement often resorts to statistical discrimination. Some examples create little concern. Focusing law enforcement priorities on young men to reduce violent crime raises few eyebrows. Racial profiling, on the other hand, is a cause of grave concern. One reason, much in line with relational egalitarianism, is that racial profiling in racially unjust societies creates a reasonable sense of inferior political status (Hosein, 2018). Racial profiling creates unequal social relations in the political realm. Profiling young men, by contrast, does not implicate any inferior political status. The problem with profiling then is not with racial or gender categories as such. Context about social hierarchies matters to the assessment of whether an individual has been wronged.

The idea that we need to consider the role of social hierarchies in determining the interpersonal wrong of discrimination is also supported by the theories of discrimination mentioned in the previous section. Hellman argues that discrimination is wrong when it demeans. Demeaning is an expressive feature of the act. Hellman also argues that “one needs a degree of power or status to demean another” (Hellman, 2008: 35). Hellman's expressive theory presupposes a difference in power or status for wrongful discrimination to occur. Moreau's approach straightforwardly locates the wrong of discrimination in its contribution to social subordination and status differences (Moreau, 2020a: ch. 2). Eidelson's approach as an example of a mental-state view of discrimination may, at first sight, appear to be different. However, even on a view of discrimination which focuses on mental states, it is useful to have legal rules which give greater protection to members of subordinated groups than to members of dominant groups. Relational egalitarians should at the very least agree that (intentional) discrimination against subordinated groups occurs much more frequently. While classifications which burden members of subordinated groups usually wrongfully discriminate against them, classifications which burden members of dominant groups are often vital parts of egalitarian affirmative action. Color-blind discrimination laws which focus on the classification of the individual without considering social context, as they exist in the United States, have ended frustrating policies to tackle racial injustice.

Prominent relational egalitarians agree with this analysis. T.M. Scanlon highlights that his egalitarian concern is with people being treated as inferior not with people treated on the basis of their race or sex. He clarifies that “[n]on-discrimination rules out race-based decisions only when these involve exclusion and attitudes of inferiority” (Scanlon, 2018: 47–52, quote at 49, my emphasis; also Scanlon, 2003: 204–205, 212–218). Niko Kolodny in arguing for the special importance of political equality for equal social relations accepts that non-ideal circumstances are special. In such circumstances we may practice, compatible with relational equality, a form of affirmative action for the political influence of those whose acceptance as equals is under threat (Kolodny, 2014: 309). Elizabeth Anderson rejects color-blindness as a moral principle. She argues that color-blindness confuses different ways in which race is used by policymakers (Anderson, 2010b: 155–168). Policies focusing on biological races because of presumed character differences between them is unjust. But policy measures which focus on racialized groups, that is groups which are stigmatized and perceived as inferior by a process of racialization, are different. She strongly supports integration measures which use race, understood as racialized groups, as a criterion in policymaking (Anderson, 2002; 2010b: chs. 6–7.)

Consequently, anti-discrimination law should not adopt a color- or caste-blind approach. Such an approach creates obstacles for policies which seek to transition an unequal status quo toward the relational egalitarian ideal. Such affirmative action measures often require benign racial or caste preferences which should not count as unlawful discrimination. Relational egalitarians should reject the anti-classification approach and adopt what (American) lawyers call an anti-subordination approach (Fiss, 1976; Siegel 2003). ¹¹ This approach takes its name from a concern with subordination, a quintessentially relational egalitarian concern. For the anti-subordination approach, facts about broader social hierarchies matter in deciding whether a classification amounts to unlawful discrimination. Laws that are sensitive to such facts are called asymmetric.

Designating the asymmetry

If relational egalitarianism requires an asymmetric discrimination law, then relational egalitarianism requires a determination concerning the social status of groups in society. This is not because relational egalitarians treat social status as a distributive “currency.” Rather, the social status of groups is relevant to assessing whether preferential treatment amounts to unlawful and wrongful discrimination or whether it is permissible affirmative action. Facts about social status are relevant to assess whether the interpersonal wrong of discrimination is committed. The law needs to specify the groups for which preferential treatment is permissible as a form of affirmative action. This, I shall argue, carries with it the same humiliation and need for shameful revelations as redistributive policies carry for luck egalitarians.

I make my argument by looking at one central example: India's affirmative action programs based on caste. Caste is a paradigm form of social hierarchy relational egalitarians are concerned about. It is deeply historically entrenched with a history of strong caste endogamy going back to around 2000 to 4000 years ago (e.g., Reich, 2018: 140–146). The caste system is a form of graded inequality in which there are sophisticated subcategorizations and hierarchies which strengthen each other (Ambedkar, 2014a). One's status in the hierarchy is determined by birth and there is no means by which one can change one's caste status. Even religious conversion to religions which ostensibly do not practice caste does not remove the stigma associated with one's inherited caste. ¹² In addition, caste practices are particularly objectionable in their intensity. They include practices of untouchability which are still followed today. In among the most extreme form of untouchability, upper-caste people consider food “polluted” when the shadow of a Dalit (“untouchable”) falls on their food. ¹³

While there is a variety of anti-caste measures, including laws prohibiting caste-based violence and untouchability, my focus is on affirmative action. The Indian government has an extensive system of affirmative action by means of quotas (called reservations) for higher education, public employment, and elections. ¹⁴ I argue that the reservations system on a whole is justified and should be endorsed by relational egalitarians. The reservations system pursues a variety of key anti-caste and relational egalitarian aims. I briefly focus on two aims.

First, reservations serve to de-segregate and integrate Indian society. Historically only upper-castes were permitted access to education. Because of inherited educational capital and continued discrimination, these castes would, in the absence of affirmative action, continue to dominate academic institutions. This would reinforce and sustain the divisions of the caste system by keeping the de facto upper-caste monopoly on education and jobs which require higher education. Upper-castes would continue to be the intellectual class and manual work would remain the exclusive domain of the oppressed castes. Given the persistence of caste, merely formal desegregation would be insufficient. If society is to break the patterns of caste segregation, affirmative integration measures are needed. Integration efforts would also be needed to increase social connections between castes. Ambedkar (2014a) famously decried the caste system as a “division of laborers” and called Hindu society a myth. Integration through affirmative action means bringing together Indians across caste divides to study and work together. Such integration is aimed to weaken caste networks and thereby weaken the significance of caste. The integration argument resembles Anderson's (2010b) argument for integration in the context of American racial inequality. In both cases integrative measures are justified as a way to eradicate the conditions which cause unjustly unequal social relations. ¹⁵

Second, reservations are targeted toward positions of political power. They cover elections, the administrative machinery of the state, and higher education needed to obtain these positions. Affirmative action serves as a way to prevent the concentration of political power in the hands of upper-castes (Ambedkar, 2014a: 260; 2014b; Omvedt, 1990). They break what used to be the monopoly of upper-castes to positions of power. This monopoly continues in those arenas in which affirmative action is not in place. A recent study found that not a single leadership position in the Indian media is held by a Dalit or Adivasi (“Scheduled Tribe”)—which together make out a quarter of India's population. All identifiable leadership positions in newspaper and television were held by upper-castes whose share in the population is estimated at only around 15 percent (Oxfam and Newslaundry, 2019). The situation is not limited to media but extends to leadership in business and civil society (e.g., Aggarwal, Drèze and Gupta, 2015; Ajit, Donker and Saxena, 2012). The only exemptions to this dominance are those parts of the public sector covered by affirmative action. Such concentration of power threatens one's status as equal democratic citizen. Concentration of power means the ability to control and interfere with the lives of the powerless. Since politics determines the allocation of power in other domains, relational egalitarians have rightfully considered the political domain as central to their theory (Anderson, 1999; Bengtson, 2024; Kolodny, 2014; Viehoff, 2014). This idea features prominently in the justification for caste-based reservations. Placing Dalits in positions of power is intended to provide a level of protection and ideally support from the state machinery. For example, Dalits in leading positions in the police can use their powers to more resolutely tackle caste atrocities. This rationale coheres well with the relational egalitarian idea of reducing domination by dispersing political power. The reservations system shares political power in a way which mitigates this threat to equal citizenship.

Given the realities of caste and its significance in Indian society, relational egalitarians should support the reservations system. It challenges the caste system, a paradigm case of unequal social relations. It does so through means of integration and the dispersion of political power, both of which are relational egalitarian aims. Furthermore, it aligns relational egalitarians with the real-world struggles of egalitarians, a connection which is key for some relational egalitarians (Anderson, 1999: 312; Schemmel 2012, 134). Reservations are the key policy of egalitarian anti-caste movements in India. ¹⁶ It is virtually impossible to find an Indian egalitarian activist who rejects reservations. At the very least, rejecting reservations would render relational egalitarians strangely disconnected from an egalitarian movement working against a paradigm of relational inequality.

The question then turns to the implementation of reservations. The implementation of the reservations system is highly complex because of the complexity of the caste system itself. The caste system is composed of more than 4000 caste groups, called jatis, with a close occupational link. Jatis are arranged hierarchically, but the precise position of each jati in the caste hierarchy can be somewhat ambiguous. Brahmins, a collection of several jatis traditionally associated with priesthood are clearly at the top. Dalits who bear the burden of untouchability are clearly at the bottom. But for many jatis their social standing is partially contested. The social standing of a jati is determined partly by ritual practices, partly by social power and political and economic influence. An example of this complexity and contestation provides the concept of a dominant caste. Sociologists and anthropologists understand a dominant caste to be the caste which is numerically, politically, or economically dominant in a social setting, for example a village (Srinivas, 1959). ¹⁷ Landowning castes can become dominant even in regional contexts. Yet landowning castes often come from ritually lower backgrounds. ¹⁸ While in a traditional sense the landowning caste is at the lower end of the caste hierarchy, the caste exerts influence politically and economically. A second example of the complexity is the phenomenon of caste groups, not particular individuals, to shift the social standing of their group by imitating habits of castes of higher social standing. This process, called Sanskritization, is only possible because caste positions are ambiguous at the margins (Srinivas, 2003). ¹⁹

India's discrimination law needs to decide which of the over 4000 caste groups is eligible for affirmative action. The partial ambiguity of caste positions makes this task non-obvious. Designating castes is necessary for any feasible system. The Indian government designates castes in different groups. “Scheduled Castes” are Dalit castes which are listed in a specific schedule for the purposes of inclusion in affirmative action policies. The relevant criterion for a classification as Scheduled Caste is whether the caste is subject to socially practiced untouchability. “Other Backward Classes” (OBCs) are a category for non-Dalit castes deemed to be of lower social standing or other disadvantage. The category also includes some religious minorities. Some Indian states subdivide OBCs further. Bihar, for example, uses the terminology of “Extremely Backward Classes.” The term “backward” also appears in the Indian Constitution's anti-discrimination provision (Articles 15 and 16). In carving out its asymmetric character, the Constitution states that special provisions are permissible for “any socially and educationally backward classes of citizens.” Since caste is not transparent to strangers, certification is needed. Individuals are given official caste certificates which certify their membership in one of these castes. If an individual wants to avail themselves of affirmative action, they need to provide their official caste certificate to show that they are of a “lower” caste and of socially inferior standing. In light of the complexity of the caste system, the non-transparency of caste to outsiders, and the sheer number of applicants for jobs or higher education, certification of individual and collective caste status is necessary. Anything else is unfeasible.

Both parts of this policy, the message sent by the state and the way individuals interact with the state, seem very similar to the objections levelled against luck egalitarians. Take first the official communication by the state. The state is officially certifying an individual's position in the social hierarchy. It designates some as “backward” castes. The state looks at caste groups and then determines their social status. There is an official determination of this status for the purposes of administrating justice. In the case of Dalits, a certificate as “Scheduled Caste” means nothing less than an official certificate that one is regarded in society as untouchable, regarded as a stain so severe that others conceive of one as being polluted by coming into contact. In discussing luck egalitarianism, Anderson, argued that it is one thing for society to have such discriminatory preferences, it is quite another thing for these preferences to be officially certified. She puts her point as follows:

“Even if everyone thought that A was so … socially unappealing that they preferred socially attractive B's personal qualities, it is none of the state's business to attach an official stamp of recognition on such private judgments. If it is humiliating to be widely regarded by one's associates as a social clod, think how much more degrading it would be for the state to raise such private judgments to the status of publicly recognized opinions … for purposes of administering justice.” ²⁰ (Anderson 1999: 305–306)

How does one receive a caste certificate? An intricate administrative process has evolved in which an individual's caste status is determined. In many cases the caste status will be unambiguous, but in some cases, there is doubt. One example are people whose parents do not have caste certificates, perhaps because they never intended to avail the benefits of affirmative action. Another example are religious conversions. Christianity is comparatively uncontroversial given the practice of baptism as an initiating ritual. But cases of children of converted Christians, converting to Hinduism are more difficult. A third case is that of children of inter-caste couples. Their caste status is inherently ambiguous since the caste system is predicated on the practice of endogamy. In all these cases, administrative authorities need to make a determination about the caste status of the child. Some children of inter-caste marriages grow up without the stigma and discrimination of being of “lower” caste. They are, effectively, members of an upper-caste. But other children of inter-caste marriages are stigmatized and experience caste the way a member of an oppressed caste does. The caste determination will have to be based on the social conditions of the person to verify whether they in fact belong to the caste they proclaim to be member of. ²¹

Troubles continue, furthermore, if individuals do not have the caste certificate of the caste they identify with. This brings us to the relationship between the individual and the state and the topic of shameful revelations. In India's harshly competitive society access to affirmative action is extremely valuable. What should an individual do if they do not have the caste certificate that they are entitled to? Let us consider the case of inter-caste parents. The best option is to sue. Indian courts have to decide which caste status the child should have. Their approach is to consider the social context of the child to determine group membership. What is their social status? Are they regarded as “one of theirs” by some caste group? How are they regarded by others? Do they suffer from social stigma, humiliations, ridicule in society? What are the living conditions of the child? These factors determine their caste status. ²²

The approach of the Supreme Court is in line with relational egalitarianism. Relational egalitarians would want to know how the child is perceived in society to see whether the child is still perceived as “untouchable” or as a member of a “lower” caste. This is important because, as I explained earlier, for relational egalitarians facts about social hierarchies are relevant to determining whether wrongful discrimination occurs. Preferential treatment of those privileged in the caste system would amount to wrongful discrimination. If, on the other hand, the beneficiary is oppressed by the caste system, their inclusion in the affirmative action program does not raise any issue of discrimination.

But this means that, in order to be issued the desired caste certificate, children or their parents need to approach the court. These petitioners need to stand up in court and explain how badly off they are in the social hierarchy. They have to narrate their personal experience, their upbringing, discrimination suffered, social exclusion, their living circumstances, their poverty and economic suffering, and so on. For example, it would be very helpful for their case if their parents work in a profession associated with a Dalit or OBC caste. One of these is manual scavenging, a Dalit profession, whose job it is to manually clean human excreta. The profession carries with it great stigma, but it would be advantageous to narrate this in court. It would also help to narrate how part of their family who belong to a so-called upper-caste abandoned them and cut off contact with them on account of their perceived inferiority.

I take it that these constitute shameful revelations. It is one thing for such facts to be known locally by oneself and one's close associates, but it is quite another thing for such things to be revealed to third parties and made part of one's public record. Besides shame involved, there is also an aspect of mistrust, not too dissimilar from the mistrust involved in overzealous attempts to catch social security fraud (Wolff, 1998: 108–109, 111–112). Petitioners need to approach a court and their court submissions narrate in detail their situation. They are not given any benefit of doubt but must show that they are in the social position they are in.

The caste example then brings out my central argument. Relational egalitarianism requires that the state officially recognizes, in its design of anti-discrimination law, some groups as socially inferior. It needs to tell people that they receive benefits on the ground that they are social underdogs or outcastes. In India, the government makes a literal designation of which castes are classified as “untouchables” and which castes are merely “backward.” If individuals want to avail themselves of the benefits of affirmative action, they need to declare themselves to belong to these groups. In cases of doubt, they may need to narrate detailed histories of themselves filled with embarrassing and shameful moments. In the arena of affirmative action, relational egalitarianism is passing as much humiliating judgment on its citizens and requires (at least) as many shameful revelations from them as luck egalitarianism does in welfare programs.

Humiliation, affirmative action, and social discrimination (I)

Relational egalitarians will want to deny this analogy. They would want to suggest that the two situations are, in fact, disanalogous to one another. Why might this be the case? In this and the next section, I will discuss a first suggestion. Relational egalitarians may suggest that it is easier to admit and less shameful to say that one has been discriminated against by others. In admitting that one has been the subject of discrimination and oppression, one admits that others wrongly saw a flaw in oneself. By contrast, admitting oneself to be stupid or ugly is to admit a flaw in oneself. The determination that a group has been discriminated against does not entail anything unfavorable about that group in itself.

The first thing to point out in reply is that this deviates from how Anderson set out her humiliation objection. In Anderson's narration it is the fact that the state certifies others’ judgments of someone's inferiority which is humiliating. She believes that this means that the state is passing judgment on citizens by elevating such judgments to reasons for the provision (or denial) of benefits. However, this, as we have seen, equally happens in decisions with respect to who is entitled to affirmative action policies.

Furthermore, re-casting the humiliation objection in this respect gives luck egalitarians a way to meet the objection. They can reply that they, too, are primarily concerned with judgments about the opinions and values of others and not with judgments of natural or inherent talent. A luck egalitarian can reply that there is nothing like “inherently untalented” or even “untalented” simpliciter. Every person has a different profile of talents and abilities. Some of these talents are valuable in a market economy and others are not. Being academically gifted is a talent which is rewarded in our current economy. Being physically strong is a talent which is currently less rewarded, but highly rewarded in economies in which manual labor takes up an important role. Being emotionally caring is a talent which, too, is currently less rewarded but may be highly rewarded in the future if professions like caregivers will be valued and paid more. The fact that caregivers are paid less than academics reflects the outcome of political, economic, and social processes, not any differential inherent talent. Thus, it is the combination of one's endowment with the preferences and values of others which determines who will be (economically) left behind (Dworkin, 2002: 116–117). The reply is even more forceful when it comes to “the ugly,” another example Anderson uses. While there may be objective beauty in nature or art, it is more contentious that some humans are “inherently ugly.” Beauty standards are commonly attached to a particular social context. Is Robinson Crusoe handsome or ugly? I do not think that this question has an answer.

A relational egalitarian may concede this point and respond that while this is true in principle, it remains the case that in present-day societies people will feel shame in being called untalented, but they will not feel shame to admit they are Dalit. ²³ The first thing to point out is that this is not the correct comparison. “Dalit” is a terminology deliberately used to avoid the offensive term “Untouchable.” Using the terminology of “Untouchable” would be perceived as humiliating. For the “untalented” we lack affirming terminology—although for some learning disabilities the use of “neurodiverse” is aimed at doing exactly this. Thus, part of the problem is that society has not adopted better linguistic conventions in the case of those with talents less rewarded in the market.

Second, it remains the case that luck egalitarian theory is in the same boat as relational egalitarian theory. For both it is the combination of personal attributes and social attitudes which creates the need for egalitarian interventions. The reason why luck egalitarianism is redistributing toward those with less marketable talents is not a flaw in them but the interplay of their abilities and social circumstances. The concern about people actually experiencing shame when deemed “untalented” is rather a concern with the practice of luck egalitarian justice. Insofar as it is a practical concern, I agree that a policy which called recipients “untalented” would create feelings of shame while simply calling recipients “Scheduled Caste” would not cause shame. But real-world policies create shame in more subtle ways. No redistributive policy calls its beneficiaries “stupid” or “untalented.” And while affirmative action unproblematically calls recipients “Scheduled Caste,” my concern about shameful revelations was not directed at this terminology but rather at how the process of determining caste status causes shameful revelations.

Humiliation, affirmative action, and social discrimination (II)

There is a second reply to the purported disanalogy between admitting one has been subject to discrimination and oppression (in the case of affirmative action) and admitting one has had bad brute luck (in the case of redistribution). Recall that the defender of the humiliation and shameful revelation objections wants to establish the following: It is less humiliating to admit that one has been subject to discrimination in society and therefore is now eligible for affirmative action than it is to admit that one is untalented and therefore eligible for redistribution.

This argument assumes that groups which are eligible for affirmative action are also victims of discrimination and that groups which are ineligible for affirmative action do not face widespread discrimination in society. There are two distinctions at play: first between groups eligible for affirmative action and those ineligible for it, and second between those subject to widespread discrimination in society and those not subject to it. The objection assumes that these two distinctions coincide. But this is incorrect. Affirmative action policies are not only permissible when they benefit groups whose disadvantage is due to social discrimination.

To see why we should first distinguish between two questions which operate at different levels. First, there is the constitutional level at which anti-discrimination law operates. At this level the question is whether putative affirmative action wrongfully discriminates. Relational egalitarianism applied to the constitutional level would hold that such policies can permissibly be pursued as long as they do not wrongfully discriminate and do not lead to undesirable social hierarchies. Sketching out this answer determines the asymmetrical character of anti-discrimination law. This asymmetry is what I pointed out raises concerns about humiliation for relational egalitarianism. This asymmetry exists regardless of whether an actual policy is pursued simply by determining what is legally permissible.

Second, there is the policy level at which we need to ask whether affirmative action should be in fact pursued. A relational egalitarian may give different answers regarding a specific policy depending on the level in question. For example, a relational egalitarian may believe that a given policy is inefficient or that it does not pursue relational egalitarian aims. But she may add that pursuing this policy would not wrongfully discriminate despite the use of racial, caste or other allegedly problematic classifications. ²⁴ Or a relational egalitarian may believe that a policy of preferential treatment neither reduces nor exacerbates social hierarchies. If so, the policy should be constitutionally permissible and be decided on grounds other than relational egalitarianism.

The asymmetry and the concern about humiliation arise even if only the constitutional level is involved, but they are more visible if the policy is also endorsed by relational egalitarians on the policy level. Hence, in the following my argument only needs to establish that relational egalitarians would endorse the proposals’ constitutional permissibility. For ease of exposition, I use examples in which I believe relational egalitarians should endorse them on the policy level, too.

Let me begin then with my first example of an affirmative action policy which is (at least) (constitutionally) permissible for relational egalitarians, but whose need is not due to past or present discrimination. The example is disability, a trait commonly protected in anti-discrimination law. Disability is protected in a particularly strong asymmetric manner. Disability discrimination law only applies to the disabled; no able-bodied person can claim disability discrimination. Disability covers a wide array of conditions. It covers various physical impairments, like mobility impairments or sensory impairments. It also covers people with psychological illnesses. Some legislations, like the UK's Equality Act, even include people who have recovered from a severe mental illness, such as depression.

Affirmative action policies for people with disabilities are (at least) permissible for relational egalitarians. Such a preference would not constitute wrongful discrimination against able-bodied people nor create any problematic social hierarchy. Affirmative action policies for people with disabilities are also in the interest of relational egalitarianism. They help to foster a society in which people are not judged by their physical or mental disabilities. The policies are integrative and aim to foster an inclusive society that reflects the ideal of relational egalitarianism.

To maintain their earlier objection, the relational egalitarian would need to claim that this is so because the disadvantage inherent in disabilities is entirely due to social discrimination. My imagined objector would need to embrace a particularly strong interpretation of the social model of disability (Barnes, 2014; 2016; Oliver, 2013; Shakespeare, 2017). This raises the stakes for the social model of disability. If the social model cannot explain a certain disability, then we have no grounds for allowing affirmative action for this group.

To assess this suggestion, we should disambiguate between different disabilities (Anastasiou and Kauffman, 2013; Andrić and Wündisch, 2015; Lippert-Rasmussen, 2022). There is no a priori reason to think that visual impairments, mobility impairments, schizophrenia, and depression are all alike. Indeed, it appears that various disabilities occupy different places on a spectrum. On one side of the spectrum are those disabilities whose disadvantages are mainly (or even exclusively) due to discrimination. On the other side of the spectrum are those disabilities whose disadvantages are mainly (or exclusively) due to the badness of the disability in itself. If we want to claim that the eligibility for affirmative action tracks social discrimination, then we should think that affirmative action for those disabilities on the inherent side is impermissible.

Severe depression is a good example of a disability on the “more inherent than social” side. Severe depression obstructs the ability of individuals to lead their life. For some it makes everyday tasks difficult and makes it difficult to function in society. It makes it difficult to form goals and pursue them. In addition, there are also adverse effects due to social stigma and prejudice against people with mental illnesses. But it seems safe to say that a large part of the disadvantage due to severe depression is because it is bad to be depressed. Depression is not just a mere difference, a different way of being. To make the point more starkly, let us then imagine that efforts to reduce social stigma are successful, but that people who suffer from depression in their teenage years nevertheless suffer setbacks. A preference for students with depression would not foster a social hierarchy in a way that a preference which compounds the disadvantage of depression would. Affirmative action for students with depression would also promote the aim mentioned earlier of creating a more inclusive and welcoming society for people with disabilities. By contrast, a converse preference for non-depressed students would, by contrast, be disability discrimination. We should locate people with depression on the favorable side of the asymmetry despite them not being subject to discrimination in our reformed society. We identified then an asymmetry in discrimination law, but ex hypothesi this asymmetry is not explained by facts about discrimination suffered by society.

A second example strengthens this point. A public university that is entirely reliant on taxpayer money decides to give preferential treatment for first-generation university students. It recognizes barriers for aspiring students whose parents did not go to university. These barriers, let us suppose, are not related to social discrimination. Rather, parents without a university education are less likely to be able to provide academic support for their children. Children who grow up in highly educated households received a different background in which they started their school career. The university now aims to level the playing field by installing a preferential access program for first-generation students. Such a preference would not count as wrongful discrimination.

Now consider the opposite preference. The university considers a preference for “legacy admissions,” that is for students whose parents already went to university. Such a preference would exacerbate inequality of opportunity, reduce social mobility, and rigidify a class system. All of this speaks against such preferences and indicates that a policy of legacy admissions contributes to social hierarchies. Relational egalitarianism might not require equality of opportunity as a positive ideal. But it is opposed to a rigid class system without intergenerational social mobility. The problem is that a rigid class system in which people cannot change their economic fortunes spills over from economic differences into social and political differences. This would cause unequal social relations and make it difficult to relate as equals with those outside your social class. Wolff (2015: 211–213), for example, is acutely concerned about societies in which rich and poor live in parallel worlds. Iris Marion Young (1990) counts class as one of the five faces of oppression. ²⁵ The same considerations also counsel positively in favor of affirmative action for first-generation university students. Notably in this case there is no discrimination involved at all. The descent into a rigid class system due to lack of social mobility can occur solely in virtue of the transmission of educational advantages within the family. This should not decrease the urgency of affirmative action measures for first-generation university students.

These examples show that the asymmetry in discrimination law which relational egalitarians should endorse is not based on the distinction between those who suffered from social discrimination and those who have not. Eligibility for affirmative action need not track discrimination faced in society. Asymmetric anti-discrimination law, as demanded by relational egalitarianism, permits affirmative action for some groups not subject to discrimination in society. In these cases, the concerns of humiliation and shameful revelations inherent in relational egalitarianism extend to groups not subject to discrimination in society. Thus, even if it was true that admitting oneself to be a victim of discrimination is easier, this does not apply to such groups—for example those suffering with severe depression. For affirmative action beneficiaries with severe depression concerns about shameful revelations are not mitigated by the fact that it is easier to admit that others wrongly discriminated against them.

Shame in ideal and non-ideal theory

Here is another way to deny the analogy between the shameful revelations inherent in the welfare system (demanded by luck egalitarianism) and asymmetric anti-discrimination law (demanded by relational egalitarianism). A relational egalitarian may argue that shameful revelations are inherent in the ideal theory of luck egalitarianism whereas they only appear in the non-ideal theory of relational egalitarianism.

It is true that in an ideal society of equals there would be no castes and no social hierarchies. Consequently, there would be no reason to have asymmetric discrimination laws. By contrast, it seems that even in an ideal society of equals there would be inequalities between rich and poor. Luck egalitarianism distinguishes itself from equality of outcome insofar as it accepts that some inequalities can be justified. However, it is less clear that those with less marketable talents would feel shame in an ideal society. It is the impetus and motivating idea of luck egalitarianism that we understand our talents as morally arbitrary and a matter of chance. Shame is an emotion of feeling deficient in some respect or of not living up to a standard. But if we internalize the motivation of luck egalitarianism, there would be no grounds for feeling ashamed if one's talent yield less in the market (Firth, 2013; Preda and Voigt, 2023).

Wolff admits as much when he writes that shame is a socially relative and contingent feature of societies. He emphasizes that this does not make shame any less real (Wolff, 1998: 114–115). Wolff is correct on this point, but it shifts the focus. His interest is clearly in the real (and non-ideal) world. Shameful revelations are a problem because we are not dealing with an egalitarian wonderland. We must confront social attitudes and background conditions that are marked deviations from equality. However, if that is the case, then we can surely be equally concerned about how relational egalitarianism confronts these attitudes and background conditions.

A related objection to my argument is that the case of luck egalitarianism is more troubling because it necessarily requires shameful revelations (Wolff, 1998: 113–118). The whole point of luck egalitarianism is to distinguish inequalities by their origin, and this requires us to investigate their cause and to find out whether people are unemployable. If shameful revelations only occur for some traits of discrimination law and in some places, then perhaps it is a confluence of numerous unfortunate circumstances that relational egalitarianism requires shameful revelations. But it would not be necessarily required by the theory.

This suggestion, however, is mistaken. In Wolff's statement luck egalitarianism requires shameful revelations in the case in which there is nearly full employment. Only in this situation it would become clear that an unemployed person is not voluntarily unemployed. It is this combination of factors which makes shameful revelations necessary. Similarly, shameful revelations are necessary if membership in a subordinated group is either not fully transparent or if there can be indeterminate borderline castes. Caste groups are of this kind. In some social contexts one can pass as a member of a different caste (e.g., Dutt, 2019). In addition, in cases of inter-caste couples the caste status of their children is indeterminate.

Physical attractiveness and lookism as a parallel

My core argument has been that both the humiliation objection and the concern about shameful revelations apply to relational egalitarianism, too. I made this argument through an analysis of affirmative action programs based on caste. Relational egalitarianism requires affirmative action to tackle caste. The complexities of the caste system require a process of determining caste status (both collective and individual) which leads to concerns about humiliation and shameful revelations. In this section, I provide another brief illustration, drawn from the discrimination literature, of the parallel between luck and relational egalitarianism with respect to the two objections.

One example which Anderson uses to criticize luck egalitarians is their stance toward “the ugly.” In aiming to compensate for disadvantages due to people's perceptions of physical attractiveness, luck egalitarians would brand some citizen as ugly. In discussions on discrimination law, the idea of lookism, that is discrimination based on one's perceived attractiveness, has gained some traction. Take the example of the U.S. restaurant chain Hooters known for hiring only female waitresses fitting a very particular beauty standard. Such lookism raises obvious concerns for how it interacts with sex discrimination and perpetuates an impression of women in service jobs as objects of admiration (Hellman, 2008: ch. 2; Rhode, 2014: ch. 7). But lookism goes beyond an extreme case as Hooters. Conventionally physically attractive women (and men, too) experience advantages on the labor market. One suggestion to combat this tendency is to have physical attractiveness a protected trait of discrimination law (Mason and Minerva, 2022; Mason, 2023: chs. 4–5; Rhode, 2010; see also Liu, 2018).

If affirmative action measures were designed, then this would require the state to make the same designation between “the beautiful” and “the ugly.” Of course, we should refer to these groups both more respectfully and more accurately as those fitting (or not fitting) a conventional beauty standard. However, even if there are no affirmative action measures, the mere operation of anti-discrimination law can be humiliating. Imagine an unsuccessful applicant to Hooters were to sue the company for appearance discrimination. Hiring decisions are complex and a variety of factors may explain the rejection of the applicant. Only if the applicant was rejected on the basis of her looks, she could succeed in court. The lawyers from Hooters therefore have a strong incentive to claim that the applicant was rejected for reasons other than her appearance. One way to strengthen this case is to point out in court that the applicant is physically attractive. The lawyers for the applicant have the reverse incentive. They can strengthen their case by convincing the judge (or jury) on behalf of their client that their client is unattractive. This would support the case that her unattractiveness was the factor which led to her application being rejected. In the case of lookism, the mere operation of anti-discrimination law can be shameful for the person involved.

The argument has gone full circle. Starting out with a concern about humiliation we moved on to identify social hierarchies. We identified that physical attractiveness can be such a hierarchy given widespread “beauty bias” in the population. The proposed remedy of anti-discrimination law ends up raising fresh concerns about humiliation. Relational egalitarianism is no better off than luck egalitarianism.

The humiliation objection, redux

What should we make of this conclusion? Remember that the humiliation objection was an objection targeted against luck egalitarianism as a theory. It held that because luck egalitarianism needs to endorse humiliating reasons for benefiting the underprivileged, we should reject it. Now if both forms of egalitarianism are equally subject to a humiliation objection, there are two possible conclusions to draw from this. Either we extend the objection to relational egalitarianism, too, or we conclude that luck egalitarianism is not subject to this concern after all.

I think my arguments contain grounds for taking the latter route. Designating who is society's underdog reflects social realities rather than endorses them. Such determinations can be politically problematic because the state needs to make a judgment in difficult cases. ²⁶ But nevertheless the state is not making anyone an underdog. Both forms of egalitarianism operate in an unequal society and seek to reduce inequalities. The first step is to acknowledge this situation and identify who is currently treated as less than equal. Thus, something similar is going on in welfare policies that are aimed at redressing the effects of bad brute luck. Determining who is on the unlucky side reflects a social reality and does not endorse it. It reflects the social reality that their skills and talents are not currently profitable.

This suggests that neither justification is insulting. What gives rise to this impression is only the form of putting the reason for compensation. Only Anderson's uncharitable formulation that the luck egalitarian speaks to “the stupid” is insulting and humiliating. Similarly, only phrasing affirmative action benefit as given to those “backward” or to “officially certified untouchables” would be insulting.

Implications for public policy

This insight leads to the second concern about redistributive welfare programs, namely that they require shameful revelations. My example of caste determination cases in India has given a vivid description of the shame that can be involved in claiming affirmative action. Either sets of policies can then require shameful revelations. Unlike the humiliation objection, I do not think this conflict can be fully avoided.

Eligibility for welfare programs and eligibility for affirmative action have to be defined in relation to furthering the aims of redistributive justice and of social equality. Means-testing is one way to set eligibility for welfare payments. It has the advantage of being, in ideal circumstances, narrowly tailored to the group of legitimate beneficiaries. But it comes at the cost of being intrusive and requiring shameful revelations. Reducing the amount of shame in dealing with welfare bureaucracy will come at the cost of a mismatch with the aims of distributive justice.

Similar remarks apply to affirmative action and discrimination law. Rigorous testing of whether people belong to a specific group tailors the group of beneficiaries as narrowly as possible. But this may require intrusive and shameful procedures. Caste determination is such an example, medical testing for disabilities is another example. Allowing affirmative action by means of self-identification would reduce this cost but will come at some mismatch with the aims of the policy.

In this sense, shameful revelations remain a concern, but they are a concern for both forms of egalitarianism. Shameful revelations are a concern to be addressed not at the level of moral theory but at the level of public policy. Both forms of egalitarianism need to address the uncomfortable and potentially shameful fact of inequalities and differences between people. Public policy should be designed to strike the right balance between making policies that efficiently further egalitarian justice and policies that reduce shame in dealing with bureaucracy. In some sense these two aims are linked. Shameful revelations act as a disincentive to apply for affirmative action or social welfare benefits. If the application process is too burdensome and is perceived as embarrassing, then some people who we want to include will decide not to apply. The aim of only including those who ideally should be included would be defeated.

Conclusion

I started out by describing two sets of egalitarian concerns. One concern connects with equality in social relations, another with equalizing the effects of bad brute luck. Some believe that the luck egalitarian concern is insulting or humiliating to the beneficiaries of their policy. I argued that parallel concerns exist for the social egalitarian concern. Relational egalitarianism requires an asymmetric anti-discrimination law and thereby needs to designate who is the top dog and who is the underdog in society. The concern about humiliation is equally forceful in this case. I illustrated this argument by looking at caste, a paradigm of the type of inequality relational egalitarians are concerned about.

I also rejected suggestions that there are relevant differences between the putative humiliation and shameful revelation inherent in redistribution (demanded by luck egalitarianism) on the one side and asymmetric anti-discrimination law (demanded by relational egalitarianism) on the other side. Unlike first appearances, in both cases we are concerned with the interplay of social attitudes and natural endowments. And in both cases shameful revelations exists because of features of our present-day inegalitarian societies.

Both egalitarian aims, to rectify unequal distributions and to rectify unequal social relations, require us to acknowledge existing inequalities. This social and public acknowledgment can be embarrassing to those who are disadvantaged. While a society should strive to lessen this embarrassment and shame, it is difficult to avoid it entirely. In some cases, the gains to equality are too limited to justify the resulting embarrassment and shame. But in many cases, the shame in confronting inequalities will be well worth the gain of a society in which resources are distributed more equally and people relate to one another on a more egalitarian footing.