Racial Integration and the Problem of Relational Devaluation

Abstract

This article argues that blacks should reject integration on self-protective and solidarity grounds. It distinguishes two aspects of black devaluation: a ‘stigmatization’ aspect that has to do with the fact that blacks are subject to various forms of discrimination, and an aesthetic aspect (‘phenotypic devaluation’) that concerns the aesthetic devaluation of characteristically black phenotypic traits. It identifies four self-worth harms that integration may inflict, and suggests that these may outweigh the benefits of integration. Further, it argues that, while the integrating process may reduce stigmatization, there is less reason to think that it can do the same for phenotypic devaluation.

Résumé

Cet article soutient que les noirs devraient rejeter l'intégration pour des raisons d'autoprotection et de solidarité. Il distingue deux aspects de la dévalorisation des noirs : un aspect de « stigmatisation » qui a à voir avec le fait que les noirs sont soumis à diverses formes de discrimination, et un aspect esthétique (« dévalorisation phénotypique ») qui concerne la dévalorisation esthétique de traits phénotypiques typiquement noirs. Il identifie quatre préjudices à l'estime de soi que l'intégration peut infliger et suggère que ceux-ci peuvent l'emporter sur les avantages de l'intégration. En outre, il soutient que si le processus d'intégration peut réduire la stigmatisation, il y a moins de raisons de penser que ce processus puisse faire de même pour la dévalorisation phénotypique.

In recent years, a number of commentators have issued renewed calls for racial integration (e.g., Adams, 2006; Anderson, 2010; Cashin, 2004; Patterson, 1997, esp. pp. 171–203; Stanley, 2017).¹ Their arguments usually involve a descriptive aspect in which segregation is identified as a cause of one or more social problems, and a prescriptive aspect in which integration is proposed as the solution. Michelle Adams (2006), for instance, claims that segregation “structures, maintains, and perpetuates inequality across virtually every indicia of social, political, educational and economic well being” (p. 275), and Elizabeth Anderson argues that segregation not only causes unjust group inequality, but expresses stigmatizing ideas, engenders stigmatizing stereotypes and discrimination, and even undermines democracy (Anderson, 2010, Chs. 2–3). Having identified segregation as a cause of these problems, these ‘new integrationists’ go on to propose integration — especially residential integration — as the solution.² As Anderson (2010) writes, “If racial segregation is the problem, it stands to reason that racial integration is the remedy” (p. 112). In this paper I will argue that these new integrationists are wrong. Even if segregation is a problem, integration may not be the solution. Integration, I will argue, threatens black self-worth, and for blacks a concern to safeguard their self-worth should outweigh the justice and other benefits that integration is supposed to bring.

Although this paper opposes integration in terms of values that are race-neutral, it focuses exclusively on the case of blacks.³ There are good reasons for doing so. As some have argued, the emerging pattern of race relations in the United States, Canada, and elsewhere appears to be one in which the main division is not between whites and non-whites, but between blacks and non-blacks. For example, George Yancey (2003) has argued that, like the barriers to the assimilation of European immigrant groups before them, barriers to Asian American and Hispanic American assimilation are receding and will continue to recede over time, thereby allowing these groups to join the mainstream. But the same, he insists, is not true for African Americans, who he believes are “destined to remain an outcast race” (p. 13). And in their research on what they call “the new diversity” in the US, Jennifer Lee and Frank Bean (2010) find a consistent pattern of “black exceptionalism” in the marriage market and multiracial identification, suggesting, they say, that “cultural boundaries” are dissolving more quickly for new immigrant groups like Asians and Hispanics than for blacks (see also Alba et al., 2018; Gans, 1999; Sears, 2015; Warren & Twine, 1997; for Canada, see Kymlicka, 1998, Ch. 5).

I believe that these facts reflect what Roy Brooks (1996) describes as the “devaluation of the Negroid phenotype” (p. 205). One aspect of this devaluation surely concerns the fact that the phenotypic traits that historically have in part defined blackness — darker skin, broader or flatter noses, etc. — have been ascribed certain meanings that have been used as the basis for holding blacks to be of lesser worth and for justifying various forms of discrimination against them. At the same time, however, many seem to regard these traits as contributing to a person's attractiveness, with characteristically white phenotypic traits, for example, thought to be more attractive than characteristically black ones.

This raises the prospect that there may be two aspects of black devaluation: what we may call a ‘stigmatization’ aspect, and an aesthetic aspect. The stigmatization aspect has to do with the fact that blacks are subject to negative stereotypes, interpersonal rejection, and various forms of discrimination because of their race (on stigma, see Crocker et al., 1998; Goffman, 1963; see also Loury, 2002). The aesthetic aspect concerns the widely shared judgement that characteristically black phenotypic traits (e.g., dark skin, etc.) are less attractive than non-black ones, with the result that many consider blacks to be singularly unattractive.⁴ I call this aesthetic-based devaluation phenotypic devaluation.

I will argue that stigmatization and phenotypic devaluation tend to depress the ‘relational value’ of blacks, or how much blacks are affectively (or non-instrumentally) valued by others. Of course, both phenotypic devaluation and especially stigmatization produce other harms that have little or nothing to do with relational value. I focus on the harms of devaluation because I believe that such harms may endure even in an integrated society. I understand the harms in question in terms of self-worth, and identify four of them: one having to do with self-esteem, another with self-respect, a third with feelings of inferiority, and the fourth with the acceptance of inferiority.

Although this paper has much to say about the problem of stigmatization, it is ultimately more interested in the problem of phenotypic devaluation. The problem of stigmatization is, to be sure, a very serious and urgent one: it contributes to — and militates against reforms in — biased policing, injustice in the criminal justice system (including mass incarceration), and inequalities in various other domains (e.g., education and health care). In my view, while one cannot plausibly deny that there has been remarkable progress in racial attitudes over the past half-century, it would be a mistake to assume that we are on an inexorable march of progress. What progress there has been has at times been halting, reverses have not been unknown, and there are signs that more reversals are possible.

But this paper was written with an eye toward the future, including the distant future, and it is hard to be confident about what will happen over the long term. It may be that we will eventually solve the stigmatization problem. But I argue that the decline of stigmatization may put into sharper relief a different problem (phenotypic devaluation) with which it has always co-existed — and one that produces similar self-worth harms.

If the stigmatization problem cannot be solved, that strengthens the argument, for the two problems together are bigger threats to black self-worth than the problem of phenotypic devaluation alone. Indeed, one might reject integration for purely stigma-related reasons. But such arguments invite a particular kind of reply: that the very integrating process itself will reduce stigmatization. For example, one leading theory of intergroup relations holds that contact between majority group members and minority group members can reduce ‘prejudice,’ especially if certain conditions obtain (Dovidio et al., 2017; Pettigrew & Tropp, 2011). I argue, however, that even if such contact can reduce prejudice, there is less reason to believe that it can significantly reduce phenotypic devaluation.

In the US, integration has largely fallen out of favour among African Americans. In part, this is because of the persistence of segregation. As Adams (2006) has noted, many African Americans reject integration “precisely because segregation continues to be an enduring feature of American life” (p. 287). But in part it also reflects African American desire to retain the identities and institutions developed over the course of their long oppression. Tommie Shelby (2012) has written:

The victims of white domination have often rejected the tenets of racial ideology that demean, misrepresent, and stigmatize them … But instead of discarding the race idea altogether, many have cultivated positive racial identities and racial solidarity that have enabled them to survive and to resist their oppression … [They] are often proud of their collective identities and sometimes seek to celebrate them publicly. On the basis of these identities, they have formed race-based organizations and institutions for political, religious, economic, and educational purposes. (p. 349)

In other words, African Americans have formed their own institutions and forged distinct identities in the face of racial oppression, and as a result many are reluctant to give them up in the name of integration. Some critics, and even some advocates, of integration have made much of this reluctance. For example, in a recent book, Sharon Stanley (2017) is at pains to develop a model of integration that does not demand assimilation into white norms and institutions or the dissolution of black institutions or cultural and social practices.

In my view, however, it may be better not to let our thinking about integration be much influenced by the attachment of blacks to longstanding cultural practices and institutions. Wanting to keep such practices and institutions may be a perfectly good reason to resist integration, but there may be better reasons still. I think that there are such reasons, and I suggest that we may gain a clearer appreciation of them if in considering the case for integration it is assumed that blacks are willing to integrate. If the case against integration is strong even in the case of a willing group, the case against it where the group is unwilling must be stronger still. So I will assume that the black population is willing to integrate, and thus will abstract from black attachment to longstanding cultural practices and institutions.

In addition, if integration no longer appeals to many African Americans, this may not be true of other blacks. The question of integration is not just an American one; Black British people, African Canadians, and other blacks must also confront it. This suggests that, in thinking about its merits, it may be better not to restrict our attention to the American context. If we must think of any single context, a better one may be a place like Canada, where most blacks are either recent immigrants or the descendants of recent immigrants, and so should not have the same attachment to longstanding (Canada-shaped) cultural practices and institutions (on blacks in Canada, see Winks, 1997; see also Mensah, 2010). There is also some reason to think that such immigrants may have a different attitude toward integration.⁵

Though this paper does not have much to say about the concept of race, some readers may wonder if there is not a working conception that is implicitly being employed. Given this, I will note here that I regard its argument as being compatible with the two main positions in the debate about the nature of race: biological racial realism and social constructionism. Biological racial realists hold that race is biologically real, typically holding that it is either a natural kind in biology or a real biological kind (Spencer, 2012). Social constructionists, by contrast, hold that race is a social construct, a product of our social practices. On this view, race is real, but real as a social kind. I believe that the ideas defended in this paper are compatible with both of these views, and I regard this as a virtue of the account.⁶ These views are varieties of racial realism. Against racial realist views are anti-realist views, which deny that race is real in any sense (e.g., Appiah, 1996). My argument may be in greater tension with anti-realist views (see Mallon, 2017). But I believe — though I cannot argue for this here — that realist views have more to be said for them.⁷

Finally, this paper makes two assumptions that should be explicitly noted. First, I will assume that the black group is a numerical minority with a non-black majority. There is some reason to think that the assimilative effects of integration will be attenuated if the black population is too large. So I assume that blacks constitute no more than 20% of the population. Second, I will initially assume that the black population is a relatively homogenous group phenotypically, with phenotypes that include those features commonly associated with blacks (dark skin, more tightly coiled hair textures, etc.).⁸ Later I will relax this assumption and consider the phenotypic variation that in fact characterizes black people.

The paper proceeds as follows. Section I further describes phenotypic devaluation, distinguishes it from racism, and explains one way that it raises a problem of justice. Section II presents the self-esteem argument, and Section III explains the connection between integration and assimilation. The following section, Section IV, discusses the self-respect-based argument and the feelings/relations of inferiority arguments. The claimed prejudice-reduction benefits of integration are examined in Section V. This section also considers objections as well as the difficulty that phenotypic variation presents for my argument. Section VI provides a conclusion.

I.

Phenotypic devaluation, I have said, is the relative aesthetic devaluation of characteristically black phenotypic traits. In this section, I explain why it should be distinguished from racism, but may nevertheless give rise to racial injustice.

Before proceeding any further, we should be a little clearer about what phenotypic devaluation involves. Three points are important to emphasize. First, it can involve not just the mere mental devaluation of certain phenotypic traits, but the external expression of this. In fact, I will be especially interested in its external expression in interpersonal relations. All else being equal, people not only prefer to be around those who they consider physically attractive, they also prefer not to be around those who they consider sufficiently unattractive. Moreover, they are likely to treat the former better than the latter. Phenotypic devaluation can of course be expressed in other media (e.g., movies and television) as well, and when this is the case, it will have more harmful effects than when it is expressed in interpersonal relations alone. Admittedly, some of the research I cite below relies on these other expressions of it, but I will assume, perhaps unrealistically, that there are relatively easy measures that can be adopted to significantly reduce them. In the case of interpersonal relations, however, its expression — where the underlying aesthetic judgements exist — is ineliminable short of the elimination of interpersonal relations itself.

Second, phenotypic devaluation can be a purely aesthetic devaluation, but some expressions of it are not ‘pure’ because the aesthetic judgements involved are not ‘pure.’ Impure aesthetic judgements can be of different kinds. One such kind is when the group identity of the person being judged negatively affects ostensibly aesthetic judgements of bodily and facial beauty.⁹ If I have a negative attitude toward Asians, for example, this very fact can affect my judgements of phenotypic traits that Asians are more likely to possess. Although much current phenotypic devaluation may involve such ‘impurities,’ I assume that it can be free of them. I'll return to this issue below.

Finally, note that what is most significant about phenotypic devaluation is not that some people are judged to be unattractive, but that some people are judged to be unattractive below a certain minimum. It is of course a commonplace observation that some individuals are considered more attractive than others, and this is no different if we restrict our attention to blacks. But what is significant about phenotypic devaluation is the possibility that many blacks may fall below a threshold the satisfaction of which secures a person a sufficient degree of relational value.

Now the history of anti-black racism, it is clear, is bound up with the reactions and meanings attributed to phenotypic traits such as skin colour and nose shape. Nevertheless, it is possible to imagine progress in race relations so dramatic that these reactions and meanings are eliminated.¹⁰ My claim is that the aesthetic judgements that in part constitute phenotypic devaluation may still remain. In this scenario, no black person need ever be hated, held to be of less moral worth, judged less intelligent, less law-abiding, etc. because of race. If this is so, phenotypic devaluation should be distinguished from racism.¹¹

To be sure, in such a scenario, phenotypic devaluation would still lead to a kind of inequality of esteem between individuals who tend to fall into familiar racial groups. Persons with lighter skin, certain nose shapes, hair textures, and so on may be judged to be more physically attractive than persons with darker skin, more tightly coiled hair, etc., and the persons judged more favourably for these characteristics would be more relationally valued as a result. These facts may lead some to balk at my suggestion that phenotypic devaluation should be distinguished from racism. In a paper on the problem that inequalities of esteem present for relational egalitarians, Carina Fourie (2015) writes:

Besides the practical objection that in reality the only people who would esteem white skin seem indeed to be motivated by racism … this form of esteem seems worrying. We can categorize this concern as being about the grounds for esteem. Certain kinds of reasons for esteem would be objectionable from a socially egalitarian perspective … [I]n a world characterized by racial oppression and discrimination, “esteeming” race seems far too close to a violation of respect for comfort. Esteeming people because of their skin color does not seem to gel with the notion that we should treat people as equals as it seems strongly associated with discrimination and with treating certain races as inferior and superior, even if, as a “genuine” case of esteeming, no injustice is really indicated. (p. 98; see also Runciman, 1967)

But this argument may be faulted on a number of grounds. First, and as suggested, we might have a world in which characteristically white features are esteemed and characteristically black features are disesteemed that is not characterized by racial oppression. To the extent that “esteeming race” appears troubling because it takes place in a world characterized by racial oppression — and arguably what is being esteemed is not race per se, but certain race-linked phenotypic traits — we can simply imagine a world not so characterized, and the troubling appearance should disappear (Cf. Fourie, 2015, pp. 98–99). This, to be sure, is not to say that such a world could not itself be oppressive in any way. After all, inequality of race-linked esteem can certainly make individuals’ lives miserable. My point is simply that a world where phenotypic devaluation is widespread may not be one that is otherwise oppressive (and so be troubling in some way because of that).

Second, even if inequalities of esteem are “strongly associated with discrimination” (Fourie, 2015, p. 98), this wouldn't make them racist. Though differential treatment of persons because of phenotypic traits can properly be spoken of in the language of ‘discrimination,’ we should not assume that discrimination — even racial discrimination — must in some way involve racism. While it is possible to defend a behavioural conception of racism, so that ‘racism’ is primarily attributed to actions and acts of racial discrimination always come out as ‘racist’ (see Flew, 1990; Philips, 1984), this sort of view should be rejected. It is better if we distinguish between racial discrimination and racist discrimination, so that while every instance of the latter is an instance of the former, the reverse is not true: not every case of racial discrimination need be a case of racist discrimination (see Matthew, 2017; see also Garcia, 1996, pp. 16, 28). If a black man looking for a wife refuses to consider any woman who is not black, he is surely discriminating by race in his search.¹² But it is another question whether this discrimination is racist.¹³ Whether an act is racist seems to depend in some way on the mental state of the person performing it, but it is not clear if the same is true, in the same way, for whether an act is racially discriminatory.¹⁴

As for Fourie's claim that esteeming phenotypic traits unequally distributed by race is “associated” with “treating certain races as inferior and superior,” there is a sense in which this is true, but that sense does not justify the claim that the treatment is racist. For to the extent that persons with different phenotypic traits are unequally esteemed because of this, they will tend to be treated unequally, and this unequal treatment will of necessity produce an inequality between the preferred and the dispreferred. One might choose to describe this result in terms of treating some races as “inferior” and others as “superior.” But given that blackness is not associated with anything like (overt or covert) anti-black animus or inferiorizing beliefs in the society we are imagining, it is not clear that we should think of this in terms of ‘racism.’

To be clear, even if it is true that a society where phenotypic devaluation is widespread need not be either racist or otherwise racially oppressive, this need not put it in the clear as far as race is concerned. Being ‘racist’ or ‘racially oppressive’ are simply two dimensions along which something may be evaluated in the racial domain. There may be other ways that thing — whether an act, a feeling, a situation, etc. — offends against ideals in this domain (see Matthew, 2017; see also Blum, 2002, Chs. 1 and 3). Thus if a society characterized by inequality of phenotype-based esteem is one in which it can truly be said that some races are treated as inferior and some as superior, this fact may be of normative significance from a race-related point of view, even if it is not racist. It may also, of course, be of normative significance from a non-race-related point of view. For example, it may be troubling from a relational egalitarian perspective, as Fourie suggests.¹⁵

So even if not racist, race-linked, phenotype-based inequalities of esteem may be objectionable for other race-related reasons. One possible argument here deserves mention. It may be suggested that given that characteristically black phenotypes have in the past been judged to be markers of inferiority, negative aesthetic judgements of them will express stigmatizing messages and so be disrespectful to blacks (Mitchell & Wells, 2018). It might be claimed that such stigmatizing messages would carry over even in our imagined society. Now though I do not want to deny that, given our history, such negative judgements may express stigmatizing messages, I do not think that any such messages will persist indefinitely, even if they won't fade away immediately. Here we need to think on an appropriately long-term basis (in terms of hundreds of years, and not simply years or even decades), and recognize that while the past can powerfully affect the future, it is not infinitely powerful.¹⁶

I have argued that it is not the case that a society characterized by unequal phenotypic esteem need be one that is characterized by racism. But now I want to argue — consistent with the above analysis — that it may still be a racially unjust society. Even if not themselves unjust, inequalities of esteem might still harm those who are disfavoured, and such harms might figure in a theory of justice (even apart from relational egalitarian considerations, which are often regarded as considerations of justice). For example, they may figure in virtue of the opportunity (formal or substantive) costs that they involve. Thus research has shown that persons who are judged to be more attractive are more likely to be voted for, hired, and to receive favourable outcomes in court (Frevert & Walker, 2014). All of this reduces the opportunities (to be elected, hired, and to receive a favourable judicial outcome) of those judged to be less attractive. Unequal phenotypic esteem may also lead to injustice through social capital effects: given that information about jobs is often provided informally through social networks, as Anderson (2010) emphasizes, racially segregated social networks can disadvantage members of excluded groups by cutting them off from such information (pp. 33–34; see also Roithmayr, 2014, Ch. 6).

But suppose that this line of thinking is wrong. Suppose that inequality of phenotypic esteem shouldn't be regarded as giving rise to injustice (or otherwise being wrong). I maintain that phenotypic devaluation may still pose a significant obstacle to integration. For whether or not it gives rise to injustice, being relationally devalued because of one's phenotypic traits may still undermine one's self-worth. From the perspective of those who are so devalued, it should not matter whether the devalued treatment to which they are subject is morally in the clear. They may have prudential and perhaps even moral reasons to resist being in situations that threaten to compromise their self-worth.

To show that phenotypic devaluation should be distinguished from racism, I have imagined a society that is free from some of the reactions and social meanings that dark skin, etc. have tended to attract. I want now to come back to the real world, where these reactions and meanings are very much still with us. Blacks remain a highly stigmatized group, and in what follows I do not abstract away from this fact. The main point of this section was to suggest that it may be possible for anti-black racism to decline while phenotypic devaluation persists — both now exist, and are linked, but this link may not be robust.¹⁷ Now if integration is to be viable at all, its advocates should have to make the case that it will eliminate or at least significantly reduce anti-black racism and discrimination. But I will argue that even if integrationists can make such a case, phenotypic devaluation may still remain. Before I get to these matters though, we need to first understand how integration may intensify the self-worth harms of stigmatization and phenotypic devaluation. I begin with self-esteem.

II.

In the most general sense, self-esteem is a form of self-worth. But it is not a coldly evaluative assessment of worth; it is an affectively laden one. As Mark Leary and Roy Baumeister (2000) put it, “At its core, self-esteem refers to how we feel about ourselves” (p. 3). Nevertheless, it does have an evaluative dimension, and psychologists have distinguished distinct sources or bases of the self-evaluations relevant to it (Leary & Terry, 2013). I begin by explaining why two of these appear to suggest that African Americans should have low trait self-esteem. These sources roughly correspond to the sources identified by reflective appraisal theory and social comparison theory — ‘roughly’ because I simplify considerably, and neither theory, in fact, straightforwardly predicts that African Americans should have poor self-esteem (see Corcoran et al., 2011; Wallace & Tice, 2012).

First, our self-evaluations seem in part to reflect the appraisal of others. The perceptions of others, in other words, seem to influence our self-perceptions, and so in turn our self-views. So by this source of self-evaluation, if we seem in general to be positively appraised by others, we should in turn positively appraise ourselves as well, and similarly for negative appraisals. A second basis for self-evaluation is social comparison. If we want to come to some judgement about ourselves on some particular dimension, one way to come to it is to examine how we fare relative to others. The results of these should then inform our self-views. Given both stigmatization and phenotypic devaluation, it would seem that both of these sources would predict that African Americans should have low self-esteem.

A leading theory of self-esteem also seems to suggest the same. Sociometer theory holds that the self-esteem system is one that “monitors the quality of an individual's interpersonal relationships and motivates behaviors that help the person to maintain a minimum level of acceptance by other people” (Leary & Baumeister, 2000, p. 9; see also Holden et al., 2016, Leary, 2005). Self-esteem, on this view, is essentially a marker of our relational value. Now if self-esteem is a subjective monitor of a person's eligibility for social inclusion, then it should depend on the factors that influence inclusion. In addition to the standardly identified characteristics, membership in a stigmatized group is one such factor. Stigmatized persons have a ‘discredited’ or devalued social identity, and this tends to make them less eligible for social inclusion (Goffman, 1963; see also Crocker et al., 1998, pp. 505–506). As for the standardly identified characteristics, they include likeability, competence, trustworthiness — and physical attractiveness (Leary & Baumeister, 2000, pp. 17–18; see also Leary & Allen, 2011). As one researcher reports, “self-evaluations in the domain of physical appearance” have been found to be “inextricably linked to global self-esteem” (Harter, 1993, p. 117). The reason is obvious: people prefer to be around, and treat better, those who they regard as physically attractive. Because this is so, persons who are generally judged as attractive should receive more offers of inclusion than persons who are judged as unattractive, and so should tend to have higher self-esteem (Leary & Baumeister, 2000, pp. 17–18). And this is something that should hold true not just in certain special (e.g., dating) contexts, for the attractive are likely to be perceived as more relationally valuable in a way that is not context-dependent. As Leary and Ashley Batts Allen (2001) argue, “even when one of these sets of characteristics [i.e., likeability, competence, etc.] is objectively irrelevant in a particular relational context, it can nonetheless influence other people's perceptions of the person and, thus, his or her relational value” (p. 39). All else being equal, this means that African Americans should generally have a lower relational value than persons who are not African American, and so lower self-esteem. Critical to this is the potentially chronic character of negative feedback about relational value.

So different considerations appear to predict that African Americans should have low self-esteem. In study after study, however, researchers have consistently found that African Americans have self-esteem that is equal to or higher than that of other Americans. In a meta-analysis of well over 300 studies of self-esteem, Jean Twenge and Jennifer Crocker (2002) found that African Americans had higher self-esteem scores than white Americans, who themselves had higher scores than other racial groups (Hispanic Americans, Asian Americans, and American Indians).

What explains African American resilience? Social identity theory points toward at least a partial answer. Positing that people are motivated to enhance and maintain a positive self-concept, and that the positive evaluation of the social groups to which they belong is key to a positive self-concept, social identity theory suggests that individuals from low status groups may strive to achieve a positive social identity (Tajfel & Turner, 2004; see also Ellemers & Haslam, 2012). It claims that when a group's low status is perceived to be illegitimate and the boundaries around it relatively impermeable, its members will attempt to secure a positive social identity using group strategies such as differentiating their group from outgroups, and engaging in social comparisons with relevant outgroups. According to Henri Tajfel and John Turner (2004), key to the use of these strategies is subjective identification with the in-group; for individuals to be able to use them, they write, they “must have internalized their group membership as an aspect of their self-concept: they must be subjectively identified with the relevant in-group” (p. 284).¹⁸

Social identity theory appears to explain African American psychological well-being in the face of devaluation. Research by Michael Hughes and colleagues (2015) found that, for African Americans, strongly identifying with their racial group and evaluating it very positively promoted “self-worth, self-efficacy, and good mental health” (p. 42). Social identity theory also seems to explain, at least in part, why African American women are able to maintain positive self-evaluations about their appearance. One particularly noteworthy study, by Jalmeen Makkar and Michael Strube (1995), compared the responses of black women exposed to a white standard of beauty and a black standard of beauty, both in the form of colour photographs of women pre-rated as attractive. They were then asked to complete ratings of their own attractiveness. A black self-consciousness scale was administered to measure the strength of black identity. As Makkar and Strube predicted, following exposure to the white models (but not the black ones) those blacks with a stronger black identity rated their own attractiveness higher, and that of the white models lower, than did those with a weaker black identity. Makkar and Strube suggest that these results “indicate that women who do not embrace their own ethnic identity may be more susceptible to comparison to White standards of beauty, resulting in lower self-acceptance. By contrast, women with high Black identity are relatively less susceptible to the potentially damaging comparison to unrealistic standards of beauty offered by the White culture” (Makkar & Strube, 1995, p. 1560).

So it seems that a positive group identity can buffer some of the negative feedback of others and so protect black psychological well-being. Some researchers have suggested specific psychological strategies that members of stigmatized groups might use. Perhaps the best-known work of this kind is that of Jennifer Crocker and Brenda Major (1989). They suggest that “the more an individual has structured his or her self-concept around membership in a group that is devalued, deprived, or discriminated against, the better that individual feels about him or herself in terms of global self-esteem … [for] identification with the stigmatized group allows the individual to use [certain] group-level self-protective strategies” (p. 620; see also Major et al., 2003).¹⁹ They identify three such strategies: i) attributing negative outcomes to group membership (instead of stable, internal personal attributes); ii) devaluing those attributes or dimensions on which the members of the group fare poorly, relative to others; and iii) using members of the ingroup for comparison purposes, which Crocker and Major (1989) say “allow[s] the stigmatized to avoid the self-esteem threatening consequences of outgroup social comparisons” (p. 614). In each case, integration — if it leads to assimilation, as I will argue it will — should reduce the ability and/or willingness of blacks to employ these strategies. Take the devaluing strategy, for example. It will presumably be more difficult for the members of a group to devalue an attribute or dimension valued highly by the larger culture if it is too immersed in that larger culture.

Once its implications are clarified, sociometer theory is also able to explain why African Americans do not have low self-esteem. As Leary and Baumeister (2000) explain, our need for belonging is not boundless: “people need only a certain level of belongingness and are not motivated to be valued and included by everyone. Once the person's need to belong has been fulfilled by relationships with some relatively circumscribed group of people, relational devaluation — even outright rejection — by others may have little effect on self-esteem” (p. 37). But once again, there is reason here to doubt whether the circumstances that protect black self-esteem will carry over to an integrated society, for Leary and Baumeister go on to suggest that “members of discriminated-against groups should suffer a decrement in self-esteem only to the extent that they either do not otherwise have an adequate social network or desire to be accepted by the out-group members who reject them” (p. 37; see also Leary, 2005). Since in an integrated society blacks would presumably not place racial restrictions on those seen as eligible to fulfill their belonging needs, sociometer theory also seems to suggest that they should have lower self-esteem in such a society.

So it seems that black self-esteem would be better protected in a segregated setting than in an integrated one. This explains one little-known fact about segregation in the Jim Crow past. In coming to its decision in its landmark ruling striking down the de jure segregation of public schools, Brown v. Board of Education of Topeka (1954), the US Supreme Court famously cited some doll studies by psychologists Kenneth and Mamie Clark. In a series of studies, the Clarks presented black children between the ages of 3 and 7 with otherwise identical white and brown dolls and asked them, among other things, to indicate which doll they wanted to play with and which doll they thought looked “bad” (Clark & Clark, 1947, p. 169). One of their last papers on the subject revealed an interesting fact: Southern — and so more segregated — children showed less of a tendency than Northern children to prefer the white doll over the brown one (Clark & Clark, 1947, pp. 177–178; but see also Cross, 1991). As Brooks (1996) notes, this “tends to indicate that racial isolation (even when tied to de jure segregation) provides a better environment than does racial integration for the development of a positive African American identity” (p. 21).²⁰

To be sure, I do not claim that blacks will necessarily have low self-esteem in an integrated society. To see one important reason why, consider the fact that a person may have high self-esteem in one domain but low self-esteem in another: high self-esteem in athletics, for example, but low self-esteem in academics. Because this is true, having low self-esteem in one domain far from guarantees that a person will have low global esteem, for even if her low self-esteem reflects her poor appraisal in that domain, her self-esteem in other domains may be more important to her global self-esteem. But notice that a domain assigned a low value may be lowly valued precisely because of the person's poor appraisal in it; thus if someone is poorly appraised as a student, but highly appraised as an athlete, she may come to value athletics more than she values academics (Crocker et al., 1998, pp. 528–530; Harter, 1993, pp. 101–109). This suggests that even if blacks are devalued in the domain of physical attractiveness, they may come to place more value on other domains, which will then become more important to their self-esteem.²¹

Now though I have claimed that African Americans have had some success in protecting themselves against stigmatization and phenotypic devaluation, I do not want to claim that they have been wholly successful in this regard. In particular, phenotypic devaluation — as expressed in any context or media, not just interpersonal relations — seems to have already had a great impact on them, influencing their judgements of physical attractiveness, especially for women, who are more likely to be judged on their physical appearance (see, e.g., Neal & Wilson, 1989; Parmer et al., 2004; Russell et al., 2013). As one team of researchers has bluntly stated, “females with lighter skin, long silky hair, and White European facial features are considered as the standard for physical attractiveness among African Americans” (Parmer et al., 2004, p. 234). We can perhaps see the clearest evidence of this in the tendency of many African American women to straighten their hair.

My claim, then, is that racial integration threatens black self-esteem because it dissolves what we have reason to believe has been protective of African American self-esteem: group boundaries. It is the strong sense of themselves as members of a distinct group, which they positively evaluate, that seems to allow African Americans to have high self-esteem in the face of stigmatization and phenotypic devaluation. Moreover, since high self-esteem is intrinsically valuable, and studies have shown it to predict important life outcomes, concern for self-esteem provides one reason for blacks to resist integration (Orth & Robbins, 2014).

This argument, to be sure, assumes that there will be a need to protect black self-esteem even in an integrated society. After all, if there is no or little stigmatization or phenotypic devaluation in an integrated society, then it hardly matters if blacks will not be in a position to employ the psychological strategies that have been protective of their self-esteem in a segregated one. Some advocates of integration have in fact defended it, in part, because of its claimed prejudice-reducing effects (e.g., Anderson, 2010). Below I will argue that while integration may neutralize the threat to black self-worth posed by stigmatization, there is more reason to doubt whether it could neutralize the threat to it posed by phenotypic devaluation.

III.

I have argued that integration threatens black self-esteem because it will dissolve the positive group identity that has been protective of it. In the next section, I explain three other ways that integration threatens black self-worth. Before turning to these arguments, however, there is an issue that needs to be discussed first. It will be noted that the above argument assumes that integration will somehow involve assimilation, but that so far I have said nothing to defend this assumption. I do so in this section.

Before I begin, two points of clarification are in order. First, while my argument does identify some negative consequences of assimilation, it does not claim that assimilation — even black assimilation — per se is objectionable. Others have made such claims. For example, some critics of integration argue that it ‘deprives’ blacks of culture, and through this, of self-worth (McGary, 1999).²² Other critics demand that blacks be respected as equals “from a distinct and valued subculture,” and reject integration as incompatible with this demand (Adams, 2006, p. 294).²³ My argument is different. Though it places great importance on a distinct group identity, it does not claim that this distinct identity is valuable for its own sake.²⁴

Second, we should be clear about what exactly the dispute regarding integration and assimilation is about. One reason integration has fallen out of favour among many African Americans is precisely its association with assimilation. As Adams (2006) explains, “Customarily, integration is seen as synonymous with assimilation, and successful integration necessarily entails the obliteration of a unique black identity” (p. 276). In light of this, recent defenders of it have been at pains to deny that their vision of it involves assimilation, emphasizing that it does not require it (Adams, 2006; Anderson, 2010; Stanley, 2017). For example, Anderson (2010) has stressed that, on her account, integration should not be conflated with a colourblind ideal that would “abolish” racial identities (p. 114). On her view, whether distinct racial identities would remain would be left to individual choice. This is all well and good, but the most pressing question regarding assimilation is not whether any particular vision of integration would require it, but whether in practice it would lead to it.²⁵ Thus my discussion focuses on this question.

Integration, as I have been understanding it, is the joining together of the members of distinct groups into some form of enduring association despite their differing group membership. It can be more or less comprehensive in scope, and can take place in different domains. I begin by listing some of these. First, civic integration includes both legal residency and “a disposition to engage with fellow citizens on matters of mutual concern” (Merry, 2013, p. 8). School integration is integration in schools, both public and private. Economic integration is the integration of all sectors of an economy so that, in a maximally integrated economy, there would be no sector associated with some racial groups and not others. Residential integration is integration in residential neighbourhoods. Finally, psychological integration is the psychological orientation, on the part of members of some group, of feeling a part of some other group.²⁶ It essentially involves the dissolution of the group boundaries that I have argued are so important to black self-esteem. In fact, it is the domain most crucial, in my view, for protecting black self-worth more generally.

I claim that assimilation in conditions of phenotypic devaluation or stigmatization makes it likely that blacks will internalize their devaluation. I understand ‘assimilation’ to mean the process, or end result of the process, whereby a person or group of persons loses the distinctive behaviour, outlook, and values of one group and gains those of another. Now integration in different domains will exert assimilative pressures of differing strengths because of differences in the closeness of the interracial contact that they involve. Economic integration, for example, involves closer sustained contact than civic integration, and so exerts stronger assimilative pressures. Assimilative pressures will also be greater when the group is integrated in multiple domains at once since this means more sustained close contact. In my view, given the closeness of the contact that it involves, residential integration — especially when combined with integration in other domains, like school integration — significantly increases black vulnerability, and it does so because it increases the risk of psychological integration.

Key to my argument is the fact that we are dealing with members of a group who are (as we are assuming) a small numerical minority. When the members of a small numerical minority integrate with a numerically larger group, they will most commonly be in contact with members of the other group. This is important because people's cultural orientation is most influenced by those with whom they have had the most sustained close contact. This means that if blacks are devalued by a non-black numerical majority, sustained close contact with them will mean assimilating into their culture, and possibly internalizing the group's devalued status.

Now it may be wondered, if residential integration significantly increases black vulnerability, why may we not limit integration to domains that involve less close contact? This raises a point that some of the new integrationists have stressed, i.e., that integration does not require perfect mixture in all spaces and institutions (Anderson, 2010, pp. 113–114; Stanley, 2017). So some allow that racial clustering in residential neighbourhoods, for example, may continue to exist.

The problem with this is that it raises questions about whether an integration that is limited in this way can provide the benefits that integrationists have argued it would bring. For example, one of the most important selling points of integration is that it will reduce black stigmatization. But it may be that nothing less than full residential integration, for example, can have this effect. (I discuss this issue more in Section V.) However, even such halfway measures may be enough to effect significant assimilation. Although residential integration increases the risk of psychological integration more than any other (single) domain, it is not the only route to this result. In short, in-between forms of integration jeopardize the benefits that integration is supposed to bring, but without ensuring that blacks would be spared the harms of assimilation.

Notice that I do not claim that residential integration (even when combined with school integration) will inevitably lead to assimilation. For one thing, blacks may simply want to maintain a distinct identity, and this desire can help serve to limit its assimilative effects. My argument doesn't deny this. I do not claim that a racial group, however unwilling it may be to fully integrate, will still inevitably lose its distinct identity. Instead, and as I noted in the introduction, I am dealing with the special case of a group that is willing to integrate.

Another possibility is that continuing discrimination may lead blacks to adopt practices that have the effect of limiting their assimilation. Adams (2006) at one point appeals to just this possibility: “individuals might identify as black even outside predominantly black communities because of an appreciation for past discrimination and the ongoing challenges that blacks face” (p. 299). But this is actually troubling since it suggests that blacks might still be stigmatized in an integrated society. If integration is to be viable at all, it must be compatible with black self-worth, and if there is any hope for it to be compatible with black self-worth, it must be true that blacks would generally — if not immediately, then eventually — not be stigmatized in an integrated society. So if this is what limits the assimilative effects of integration, it is reason to reject integration itself. We should also keep in mind the possibility that its assimilative effects may at some point become irreversible, so that efforts to hold on to a distinct identity will be unsuccessful, a little too late.

A common strategy of integrationists is to contrast assimilation with what they call ‘mutual transformation,’ and to claim that their ideal of integration involves the latter, not the former (Adams, 2006, pp. 302–304; Stanley, 2017, Chs. 2 and 3). The idea is that instead of blacks simply assimilating into white spaces and institutions, they can “contribute meaningfully to the shape and content of integrated spaces and institutions,” so that integration would “defy assimilation and reflect the aspirations of both blacks and whites” (Stanley, 2017, p. 74). But if blacks are a numerical minority, it is hard to see why even reconstituted or transformed spaces and institutions would not still mostly reflect the aspirations of non-blacks. And even if they didn't, it is hard, for the same reason, to see why they would avoid phenotypic devaluation, at least, with all of its implications for self-worth. Blacks may generally still be relationally devalued because of their phenotypes.²⁷

IV.1

I have argued that, given the assimilative pressures that it will exert, integration may undermine black self-esteem. I will now argue that it may compromise black self-respect. Self-respect and self-esteem are clearly distinct. When a person is accused of lacking self-respect, the accusation is not necessarily that he lacks self-esteem (Thomas, 1995, pp. 256–257). In addition, a person can have poor self-esteem while having a robust self-respect, and vice versa. For example, a person may feel bad about himself, but still adamantly refuse to do things that he considers beneath him. Both self-respect and self-esteem are forms of self-worth, but as Laurence Thomas (1995) says, self-respect is “a more fundamental sense of worth than self-esteem” (p. 261; see also Dillon, 2013; Sachs, 1981). The following self-respect-based argument thus provides a stronger reason to oppose integration than the self-esteem argument.

For many, the kind of worth that self-respect involves has to do with one's moral status or standing as a person, in particular with the idea that one's moral status is equal to that of others, so that one is no less entitled to basic rights (Boxill, 1992, Ch. 9; Hill, 1995a; Thomas, 1978, 1995). So for Thomas (1978), for example, self-respect “consists in having the belief that one has and is worthy of full moral status. And to believe this is to believe that one is as worthy as any other person of the set of moral rights to the recognition of which persons are entitled” (p. 309).²⁸ And since, in his view, this belief can only be sustained when the social institutions of one's society affirm this status, self-respect requires the equal distribution of rights and liberties (Thomas, 1978; see also Hill, 1995a; but see also Hill, 1995b, where he retreats from a purely rights-based account of self-respect). In my view, however, while it is certainly true that self-respect involves recognition of one's equal moral status, there is more to it than that.

Self-respect is a robust sense of one's value or worth (moral and otherwise) that indisposes one to act in certain ways or tolerate certain forms of treatment. The self-respecting person regards the relevant treatment and acts to be in some sense beneath her, and where feasible refuses to perform or tolerate them. Of course, there may be circumstances when a person can permissibly (i.e., consistent with self-respect) tolerate objectionable treatment or act in ways normally inconsistent with self-respect; cases of threats to life (to take the most extreme case) are presumably one such example. These actions and inactions do not imply an insufficient appreciation of worth, or insufficient steadfastness. However, a person can't act in certain ways or tolerate certain forms of treatment in order to avoid some minor setback to well-being (Hill, 1995a, pp. 84–85). These points suggest that our duty of self-respect should be understood as at least moderately demanding, and possibly very demanding, but that there is surely a point at which we can make it too demanding.

Self-respect is a robust recognition and appreciation of one's worth, and what follows from this with regard to how one acts and what one tolerates. Even when one duly appreciates one's worth, the appreciation may not be sufficiently robust, and if this is the case, the person will fail to be self-respecting. As Christian Schemmel (2019) has recently emphasized, a person's self-respect must be robust, for “being able to deal with adversity is a constitutive part of having self-respect” (p. 633). This means that the self-respecting person doesn't crumble in the face of adversity. She does not buckle at the first appearance of unfavourable circumstances. Rather, she remains steadfast.

Now it may be wondered why a person's self-respect, as opposed to her self-esteem, should be affected by how she is treated by others (Cf. Margalit, 1996, pp. 24–25). Howard McGary (1999) has written:

A crucial aspect of self-esteem is the fact that the opinions of others are crucial to a person's assessment of whether or not he feels that he is worthy of esteem. Self-respect, on the other hand, does not necessarily depend upon the assessment of others. In fact, the person who stands steadfast in spite of the negative assessment of others might be thought, under appropriate conditions, to epitomize the self-respecting person. (p. 49)

McGary is surely right that self-respect does not necessarily depend on the assessment of others, and that it is the person who stands steadfast in the face of negative assessments who epitomizes the self-respecting person. Self-respect, as I have said, is robust. But one can of course affirm that self-respect should be at least somewhat resistant to the negative assessments of others, while also affirming that it can be negatively impacted by such assessments, at least some subset of them.

At this point, however, we must consider a more uncompromising view of the demands of self-respect. It can be denied that negative assessments can rationally impact self-respect, for it is possible to claim, as Colin Bird (2010) has claimed, that however often people typically let the assessment of others undermine their self-respect, in doing so they act irrationally, for “what others do cannot constitute a good reason to lose one's self-respect; it is always possible to resist outside assaults on one's worth” (p. 19; emphasis added). This uncompromising stoical view regards threats to a person's self-respect — no matter the circumstances that constitute them — as always in principle capable of being resisted, and so any failure of resistance as irrational. To be sure, those who endorse this kind of view need not indiscriminately condemn all those who let the assessments of others compromise their self-respect; they might accept that in some circumstances we may not properly blame those who succumb to negative assessments (Bird, 2010, p. 19).

This position is flawed. To see this, note that there are surely at least three kinds of positions that one might take on the relation of a person's self-respect to the assessments (or treatment) of others: first, one might claim, like Bird, that it is always the agent's responsibility to maintain her self-respect in the face of others’ negative assessment, no matter how poor, widely shared, or rudely expressed; second, one might claim that any negative assessment, by however few people, in whatever circumstances, and however expressed, can provide a good reason for a person to lose her self-respect; and finally, one might take a more middle ground position, which holds that it is neither always the agent's responsibility to maintain her self-respect nor should it be so vulnerable that even a single person politely conveying a mildly negative assessment can undermine it.

I take it that the first position is too demanding, the second is insufficiently demanding, and that the correct view is some version of the third. But Bird (2010) strangely overlooks it. The only alternative to allowing any negative assessment undermine a person's self-respect, he appears to think, is the stoic one. This is why he can claim that it is an advantage of that alternative that “it doesn't hold self-respect hostage to the utopian idea that these forms of abuse will easily disappear from the earth” (p. 38). Of course, there are a variety of positions that one might take consistent with making self-respect neither completely vulnerable nor completely invulnerable; within these broad constraints, we can make self-respect more or less demanding. I hold that self-respect is a duty that makes relatively strong demands on individuals; we should not let just any negative assessment undermine our recognition or appreciation of our worth. But negative assessments can surely be sufficiently poor and widespread that they can reasonably undermine our self-respect. That is, they can be such that a person whose self-respect is undermined by them does not thereby show himself to be irrational or insufficiently steadfast.

The conception of self-respect that I have described is one that has to do with what Stephen Darwall (1977) has called “recognition self-respect” (p. 38). This is a form of what Robin Dillon (2018) calls “status worth,” which she defines as “worth that derives from such things as one's essential nature as a person, membership in a certain class, group, or people, social role, or place in a social hierarchy.” As she notes, most discussions of self-respect focus on a person-based status. I assume that all persons, as persons, should recognize and appreciate themselves as having worth, but I believe that in thinking about self-respect it is helpful to think beyond general personhood and consider how persons, as distinct, concrete individuals, live this status.

The reason for this is that we are not simply persons but persons with distinct group identities: we have racial, sexual, and other identities. This is to say that though we all have status worth as persons, we also have status worth as persons who are black, as persons who are women, as persons who are gay/lesbian, and so on. To be clear, I am not claiming that this identity-based status worth is a worth over and beyond our worth as persons; it is a worth that we have simply in virtue of being persons. So it trivially follows from our status worth as persons, who have various group identities. I highlight it because some persons face threats to their self-respect because of their identities; for example, some threats to person-worth are distinctive because a person is a woman, or is gay. Thus if many feminists are right that a woman who is unduly deferential to her husband lacks self-respect, this is arguably not simply because a married person shouldn't be unduly deferential to his or her spouse, but rather is closely connected to the fact that she is a woman (Dillon, 1992; Hill, 1995a; Holberg, 2017; Stark, 1997). In a similar way, what is non-self-respecting about the behaviour of the figure of ‘Uncle Tom’ is arguably closely connected to the fact that he is black (Dillon, 1992; Margalit, 1996, pp. 35–39; Stark, 1997; Thomas, 1978, 1995). So though I am not claiming that a person's race, sex, sexual orientation, and so on ever provides what we may call a basic ground of self-respect, a person's self-respect may be threatened in certain ways because of her identity. Moreover, when a person is able to continue living in light of a robust appreciation of her worth in the face of these distinctive identity-based threats, it seems fitting to speak of that person as having racial self-respect, sexual self-respect, and so on. My claim, then, is that racial integration threatens blacks’ racial self-respect.

Before proceeding, however, let me say something about how self-respect-inconsistent actions relate to the state of lacking self-respect. There are two kinds of relations. One, what I call the ‘constitutive’ relation, holds that some acts are such that they constitute the person performing them as lacking self-respect; if someone tolerates a certain act of abuse in order to achieve some minor good, for example, we may view this not merely as evidence that she fails to appreciate her worth, and so evidence that she lacks self-respect, but as constituting her, at least at that time, as non-self-respecting. On the other, which I call the ‘evidential’ relation, the relevant acting is merely inconclusive evidence that the person lacks self-respect. This evidence might be inconclusive because i) it is possible that it does not in fact flow from an appropriate (self-respect-compromising) source (i.e., a lack of steadfastness), and ii) self-respect is usually character-trait-like in that it is relatively stable over time, but just as a person with, say, the character trait of honesty may occasionally act dishonestly, so a self-respecting person may occasionally, at least for some actions, act in a self-respect inconsistent fashion.²⁹ Self-respect-inconsistent actions can relate to the state of lacking self-respect in both ways: some acts count as evidence that the person performing them is non-self-respecting, and some acts constitute the person as non-self-respecting.

Implicit in my account thus far is that self-respect-inconsistent-seeming actions are of two kinds: they can be the result of a failure to recognize or appreciate one's worth, and they can be the result of a failure of steadfastness. Self-respect-inconsistent acts that are failures of steadfastness always constitute the person as non-self-respecting. If one fails to be sufficiently steadfast in the face of threats to self-respect, one has by definition failed to be sufficiently resolute, and so acts that flow from insufficient steadfastness always constitute one as non-self-respecting. By contrast, self-respect-inconsistent-seeming actions that are failures of recognition or appreciation of worth can either constitute the person as non-self-respecting or be evidence that he lacks self-respect.

Let us turn, then, to how integration can compromise black self-respect. Now when individuals sense that they are relationally devalued, this will likely motivate efforts to enhance their relational value (Cf. Leary & Allen, 2011; Leary & Baumeister, 2000, pp. 38–39). Actions taken in this regard raise questions of self-respect. For example, being devalued because of one's race may lead a person to act in a servile manner, as she recognizes the devalued status of her group, and attempts to ingratiate herself in the out-group by acting in this way. Such a person, it is worth adding, may sincerely reject any suggestion that she is not of equal moral status, and so not equally deserving of moral rights. This suggests that lacking self-respect need not involve the belief that one is not of equal moral status (Cf. Hill, 1995b).

So servilely ingratiating oneself with an out-group is one way that one can act in a self-respect-inconsistent manner, and so may be non-self-respecting. Another is trying to pass as a member of an out-group — that is, hiding one's membership in a group. A person can pass with regard to race, ethnicity, religion, sexual orientation, and more, and the passing may be temporary or permanent. With regard to racial passing, in the nineteenth century and the first half of the twentieth century, many light-skinned African Americans passed as white as a way of escaping first slavery, and later the stigma and restrictions of the American caste system (Hobbs, 2014; Kennedy, 2003, Chs. 7 and 8).

Now given my assumptions about the black population whose integration we are considering, it is clear that passing for white is not an option for them. But there is a kind of passing that even phenotypically stereotypical blacks may be tempted to try. Suppose, for example, that knowledge of racial self-identification by itself — that is, quite apart from a person's phenotypic traits — affects how individuals are treated.³⁰ In that case, some stereotypical blacks may be tempted to try to pass as members of a racial category defined in part by lesser stereotypicality (e.g., mixed race). Something like this seems to happen in Brazil, where sociologist Elizabeth Hordge-Freeman (2015) finds that black Brazilians tend to “whiten themselves by using intermediate color terms” (p. 205). As one of her informants says, “We call ourselves morenos so that we do not have to say negro” (pp. 146–147; see also Twine, 1998, Ch. 6). We may want to say that in doing so such blacks constitute themselves as non-self-respecting. Although blacks remain highly stigmatized, we might demand that they be more steadfast than that.³¹

At this point, we should consider the question of self-respect and blame. Should we blame people for not having self-respect? Not always. Here we return to the distinction between self-respect-inconsistent actions that are failures of recognition or appreciation of worth, and those that are failures of steadfastness. I believe that failures of steadfastness are always blameworthy, but that failures of recognition/appreciation are not. Circumstances can be such as to make it very difficult to recognize and appreciate one's worth. If one is subject to widely shared and frequently expressed negative assessments, for example, this can make it difficult to develop and/or sustain an appreciation of worth, and so make it difficult to maintain one's self-respect. Blame in such cases seems inappropriate. I think that this can be the case in circumstances of widespread devaluation.

The discussion so far offers general lessons about the potential dangers of integration for blacks. But it also suggests a corresponding ideal that societies should strive to achieve: if a society's ‘racial arrangements’ make blacks want to be something other than black, then they are inconsistent with black self-worth and so should be unacceptable to them.³² We should think of this as a particular application of a general race-neutral test that societies must pass in order to be acceptable to the members of the various racial groups: these members should not want to be some other race. So my claim is that, in conditions of stigmatization or phenotypic devaluation, there is reason to doubt whether integration can pass what we may call the same-race test.

Another kind of act that is inconsistent with black self-respect concerns dating and marriage. For many, dating and marrying across racial lines are signs of admirable open-mindedness (Cashin, 2017). Yet I will argue that they may also be signs of the internalization of one's devaluation (Cf. Fanon, 1967, Ch. 2). Before proceeding, however, two points should be kept in mind. First, the meaning of instances of interracial dating and marriage may be different depending on the racial identity of the party being considered, especially if there are group status differences between the groups to which the parties belong. Second, people date and marry interracially for a variety of reasons, and not all of these raise questions of self-worth.³³ Just as it seems that an act that is permissible if done for one reason can be impermissible if done for another, so an act that is self-respect-consistent if done from one motive may be self-respect-inconsistent if done from another. Thus the interracial relationship of a black person with no self-worth issues should not be judged the same way as one in which the black person has such issues.

Now if you are a member of a devalued group, and you want to increase your relational value by disassociating oneself from the group, one way of doing so is to find some way of signalling to out-group members that, despite one's group identity, one in fact has high relational value. Dating and marrying interracially are ways of doing this. I claim that if one dates or marries interracially for this reason, one constitutes oneself as non-self-respecting.³⁴

IV.2

A third self-worth harm of relational devaluation is that it can lead to feelings of inferiority. Some philosophers have noted that one reason for objecting to inequalities of various kinds (status, esteem, etc.) is that they can generate feelings of inferiority. For example, T. M Scanlon (2003) has claimed that, among other reasons, we may object to inequality because “it is an evil for people to be treated as inferior, or made to feel inferior” (p. 204; see also Fourie, 2015; Scanlon, 2018, Ch. 3). This is clearly because it is intrinsically bad to have such feelings. I claim that being a member of a devalued group in an integrated society may generate them.

Inequality of phenotypic esteem may also be objectionable if it gives rise to relations of inferiority and superiority, even if this is not accompanied by feelings of inferiority. Above, in discussing Fourie's work, I noted that a society characterized by inequality of phenotype-based esteem is one that relational egalitarians may have reason to object to. I explore this further now. Relational egalitarianism is best explained by contrasting it with distributive egalitarianism, which was traditionally the dominant approach to equality. According to the distributive view, justice requires the equal distribution of something (e.g., resources or welfare), and the task for the egalitarian theorist is to figure out what this is. Like the distributive view, the relational view accepts that there is an ideal of equality that is required by justice, but understands it as one that in the first instance applies to certain kinds of interpersonal relationships among the members of society. As Samuel Scheffler (2015) puts it, it requires the establishment of a “society of equals” in which its members “relate to one another on a footing of equality” (p. 21). This kind of equality may have distributive implications, but they are indirect, flowing from a broader ideal.

There is much to be said in favour of relational egalitarianism. However, accounts of it often leave it unclear why individuals should want to avoid the unequal relations that it condemns. These accounts seem to suggest that these reasons bottom out in considerations of justice: relating on a footing of equality is good because it is what justice requires.³⁵ But I believe that there are reasons beyond justice for establishing a society of equals, and, moreover, that individuals have reasons beyond justice for wanting to avoid unequal relations. I have in mind reasons of self-worth. We all have an interest in achieving and maintaining a sense of our own worth. But unequal relations may make this impossible. So, from this point of view, the main problem with unequal relations, at bottom, is that the parties in the unequal position may come to accept the inferior place that they are assigned in the relationship.

To see that integration in conditions of stigmatization and phenotypic devaluation can lead to objectionably unequal relations for this reason, consider the following passage from Hordge-Freeman's (2015) study of Brazil:

As I sit in on an art course for children, I observe Afro-Brazilian girls stroking the straight hair of their white and straight-haired peers during instructional time, during classroom breaks, and during lunch. Even as they participate in conversations with peers and teachers, they continue running their fingers through the straight hair of their white classmates.

During breaks, these young Afro-Brazilian girls run to the bathroom and douse their hair with water in order to achieve a look that is wet and wavy, rather than dry or “bad,” and then they return to class. Their straight-haired peers, the objects of their attention, do not respond to the attention, and everyone acts as though this is a normal part of their daily activities. (pp. 91–92)

Part of what makes the scene depicted so tragic is that the Afro-Brazilian girls evidently accept the inferiority attributed to them by Brazilian society: whites and near-whites have ‘good’ phenotypic traits, but they do not. Note, moreover, that they may accept this inferiority without feeling inferior. Indeed, one characteristic feature of accepting one's inferiority is that one is often blind to this fact.

IV.3

Importantly, the self-worth harms of relational devaluation can be brought about even when the devaluation is rooted in phenotypic devaluation, not just stigmatization. This is perhaps clearer with regard to self-esteem and feelings and relations of inferiority: a person's self-esteem can be negatively impacted by phenotypic devaluation, and the same can also give rise to a feeling of or acceptance of inferiority. Phenotypic devaluation can also lead to acts that suggest that the person is, or that constitute the person as, non-self-respecting. It can lead persons to act in servile ways. It can also lead to passing. Entering into an interracial relationship may also constitute a person as non-respecting when done because of phenotypic devaluation. A person might date or marry interracially not to raise her relational value, but that of her children. She may desire to reposition her family along what Hordge-Freeman (2015) calls the “phenotypic continuum” by having phenotypically less stereotypical (e.g., lighter-skinned) descendants — something, it is worth nothing, that can also be pursued endogamously with less stereotypical members of the group.³⁶ (Thus the same reasons that we have for objecting to some interracial unions, when entered into for certain reasons, also apply to some intraracial unions, when entered into for the same reasons.) This act is in serious tension with black self-respect, and indeed is among the most serious ways that a black person may manifest a lack of it.³⁷ Acting servilely and hiding one's identity are both ways that a black person might fail to live in light of a robust appreciation of worth, but acting on a desire — whether for oneself or one's children, or both — to be some other race, or some reduced version of one's race, is perhaps the quintessential way of doing this.³⁸

Consider also this. Above, I noted that one important way of testing whether integration is consistent with black self-worth is to apply what I called the ‘same-race test’; in this case, if a significant number of blacks would not want to be black in an integrated society, integration should be considered inconsistent with black self-worth, and so rejected. Consider now that phenotypic devaluation can make blacks not want to be black. If one is devalued because of one's phenotypic traits, a natural reaction is to wish that one had some other traits. If this is right, then the self-worth harms of relational devaluation may be rooted in phenotypic devaluation, and not just stigmatization.

However, we should recognize that even when phenotypic devaluation makes persons want to have different phenotypic traits, it may not always make them want to be a different race. If one is devalued because of one's phenotypic traits, one may wish to have different traits, but this need not mean that one wishes to be of a different race. This may be the case, but it may not. In her study of Brazil, many of Hordge-Freeman's (2015) interview subjects clearly seem to wish that they were closer to the white end of the phenotypic continuum, but some expressly deny that they want to be white; as one nine-year-old girl says, “If I could make any wish, I would wish for long blonde hair and blue eyes. But I don't really want to be white” (p. 119).³⁹

But even if phenotypic devaluation makes blacks merely want to be less stereotypical phenotypically, rather than being some other race, this too is objectionable.⁴⁰ It cannot be that an arrangement of society is unacceptable if it would make persons wish to be some other race, but is acceptable if it makes them want to be less stereotypical members of their own race — in our present case, unacceptable if it would make blacks not want to be black, but acceptable if they want to be less black.⁴¹ The intuitions that tell against the one seem also to tell against the other. Further, the same-race test seems apt to focus attention on self-worth harms less likely to endure in an integrated society and to fail to notice those unlikely to fade away. For if blacks are no longer stigmatized in such a society, as integrationists must argue, then blackness as a racial category would not be something marked as deviant and blacks should not want to be something other than black. Indeed, blackness might be something that is viewed with considerable pride. But such a view of blackness is compatible with more stereotypical black phenotypic traits being regarded as misfortunes, if only for aesthetic reasons.⁴²

All of this suggests that the same-race test should be supplemented with an additional test more particularly concerned with phenotypic traits. This additional test would require that for a society's racial arrangements to be acceptable to the members of the relevant racial groups, they should not want to have different phenotypic traits; in particular, they should not want to have less stereotypical traits. Let us call this the phenotypic test. My claim, then, is that in conditions of phenotypic devaluation, an integrated society won't pass the phenotypic test, even if it can pass the same-race test.

IV.4

Now the self-esteem argument developed in Section II and the feelings of inferiority argument of Section IV.2 work by providing individual blacks with prudential reasons to avoid racial integration — integration is counter to their good or well-being. The self-respect-based argument, however, proceeds differently (so too does the relations of inferiority argument). It holds that individual blacks have moral reasons to resist integration. The idea is that we have a moral duty to robustly appreciate our own worth, and to act in light of this appreciation (Cf. Hill, 1995a, pp. 89–91). Moreover, since we have a moral duty to maintain our self-respect, we have moral reasons to avoid that which tends to undermine self-respect.⁴³ Hence blacks have moral — in addition to prudential — reasons to avoid integration.

To be sure, some deny that we can have duties to ourselves, and if they are right, we of course can't have a duty to respect ourselves (Singer, 1959). This is surely not the place to take up this issue. I therefore make my claim conditional: if we can have duties to ourselves — and duties of self-respect are surely plausible candidates for such duties, if they exist — then blacks may have duties to resist integration (Cf. Hill, 1995a). Further, since some acts may be morally valuable whether or not they are imposed by moral duties, so that we may have moral reasons to perform them even if these reasons fail to amount to duties, blacks may have moral reasons to resist integration even if these reasons do not amount to duties.

These arguments, it should be noted, do not question the benefits of integration for which integrationists like Anderson (2010) have argued; for all that I argue, integration may indeed be justice/democracy-enhancing, even if it is not quite the imperative that she has claimed. Rather, I hold that self-worth is so valuable, and consequently the potential self-worth harms of integration so significant, that they may outweigh the justice/democracy benefits it is said to bring. Self-worth is a great good. Self-esteem, as I have suggested, is intrinsically valuable, and it has been linked to important life outcomes; avoiding feelings of inferiority is an obvious and significant good; self-respect is an extremely important form of self-worth (Cf. Rawls, 1999, pp. 386–391), the absence of which precludes a truly flourishing life; and, as the precondition for the sense of agency needed to pursue so much of what is valuable in life, not accepting that one is anyone's inferior may be the most important form of self-worth of them all.⁴⁴

And yet, as great as the good of self-worth is, and as bad as self-worth harms can be, self-worth does not have absolute overriding weight such that it cannot be outweighed by anything else. So it is possible for the self-worth harms of integration to be outweighed by its benefits. Suppose that if blacks integrate, a protracted and deadly civil war, in which millions — blacks and non-blacks alike — lose their lives, is avoided. If so, the harms of integration are probably outweighed by its benefits. And this is so even if a worst-case scenario obtains as far as these harms are concerned. But such fantastic scenarios are unlikely to obtain in the world in which we live. In this world, the benefits are almost certain to be much more modest. In addition, the example has a special feature — great benefits to non-blacks — that is also unlikely to obtain. In this world, any benefits to others are likely to be negligible.

The benefits to blacks, however, may be great, as the example itself suggests. But again, I think that we are entitled to set aside such unrealistic scenarios. If we do that, it is very hard to see how the benefits of integration could be so great as to outweigh the harms, assuming something at least reasonably close to a worst-case scenario. So if self-worth is as valuable as I have argued that it is, even the fullest realization of the benefits that we can realistically expect could not outweigh the harms that we can reasonably expect. In comparison to these (realistic) benefits, it does have absolute overriding weight. An increase in an individual's economic standing, for example, could never outweigh the harm of a severely damaged sense of self-worth.⁴⁵

So I argue that blacks have prudential and moral reasons to resist integration, even if it is true that integration is justice/democracy-enhancing. Now it could be argued that this pitting of self-worth, and in particular self-respect, against justice wrongly assumes that a society can rightly be said to be just while some of its members — defined by a group identity such as race no less — lack a form of self-respect. This kind of objection appears most at home in a view that sees self-respect as exclusively threatened by social institutions, at least if it is combined with the view that the only subject of social justice is social institutions.⁴⁶ For if social institutions are the only subject of justice, it seems plausible to hold that they should be so arranged that they secure for individuals a good as important as self-respect. But my argument shows that this narrow view of the threats to self-respect is mistaken; various forms of interpersonal interaction, in addition to social institutions, can undermine it. Given this expanded view of the threats to self-respect, there is correspondingly less reason to think that justice is inconsistent with all actions that tend to undermine it.⁴⁷

V.1

I have argued that integration will intensify the self-worth harms of stigmatization and phenotypic devaluation, and that these harms may be so significant that they may outweigh the benefits that integration is often said to bring. But for all that has been said, my argument may seem to simply assume that an integrated society will not in certain ways be much different from a segregated one. After all, and as I noted in Section II, one of the central virtues of integration, according to its champions, is precisely its prejudice-reducing benefits. So this issue requires examination.

To its advocates, integration is not short of virtues. It will, they say, overcome unjust group inequality, end mutual mistrust, and more fully realize a democratic society (Adams, 2006; Anderson, 2010). But what interests us here are claims that it can improve interracial relations by reducing black devaluation. As Anderson (2010) says, “Integration should reduce antiblack prejudice, stigmatization, discrimination, and anxiety” (p. 123). Let us call these benefits ‘prejudice-reduction benefits.’ In arguing that integration has prejudice-reduction benefits, Anderson relies heavily on the theory of intergroup relations mentioned in the introduction: contact theory (pp. 123–127; see also Cashin, 2017, pp. 162–168). I will argue that there are limits to the benefits of contact.

Contact theory makes a simple claim: contact between members of different groups can reduce ‘intergroup conflict,’ especially if the following conditions obtain: i) the groups have equal status, ii) the groups meet in pursuit of common goals, iii) there is cooperation between groups, and iv) there is institutional support for contact.⁴⁸ It is a general theory, and so the beneficial effects of contact are supposed to hold across groups defined by race, ethnicity, sexual orientation, age, and more. Moreover, the benefits are supposed to be generalizable with regard to out-group members not involved in the contact, other situations, and even other out-groups. In their meta-analysis of over 500 studies, Thomas Pettigrew and Linda Tropp (2011) examine all this, and much more. They find that intergroup contact does indeed have prejudice-reduction benefits that generalize broadly, and they convincingly defend the theory against many objections.

I want to concede the basic claim of contact theory, as applied to race relations: interracial contact can reduce prejudice. But this concession leaves us with two important questions. The first is whether prejudice reduction is a sufficient goal to aim for in the realm of race — and if it is not, whether contact can secure whatever would be sufficient. The second is whether intergroup contact has any unintended side effects. I will argue that prejudice reduction is not sufficient to address phenotypic devaluation, that interracial contact by itself seems incapable of addressing this, and that the most effective forms of interracial contact, as far as prejudice reduction is concerned, are just those that we have reason to believe will make blacks most vulnerable to the self-worth harms of phenotypic devaluation.

To substantiate these claims, more needs to be said about the theory. In particular, we need to ask exactly how intergroup contact is supposed to lead to prejudice reduction. Surmising that intergroup relations for groups meeting for the first time is liable to be marred by ignorance, anxiety, and lack of empathy, researchers have been primarily concerned with examining whether these three variables — knowledge, anxiety, and empathy — mediate contact and prejudice reduction. As Pettigrew and Tropp (2011) explain, “The straightforward idea is that successful intergroup contact will provide you with useful knowledge about the outgroup, reduce your anxiety in intergroup encounters, and help you to take the perspective of outgroup members and empathize with their concerns” (p. 77). Their work suggests that empathy and anxiety reduction are most important, and that knowledge is less so (Ch. 6). Their research also suggests that contact is more likely to reduce the affective components (dislikings, negative emotions) of prejudice than the cognitive components (beliefs and negative stereotypes) (Ch. 7).

Let us turn, then, to the first question, which is whether prejudice reduction is a sufficient goal. As I have been claiming, relational devaluation of blacks has at least two distinct sources: stigmatization and phenotypic devaluation. If its advocates are right, and interracial contact can reduce anti-black prejudice, stereotyping, and (race-based) discrimination, contact will be able to address relational devaluation rooted in stigmatization. But relational devaluation rooted in phenotypic devaluation may remain untouched. As I argued in Section I, there is no reason in principle why phenotypic devaluation might not remain after devaluation rooted in stigmatization is eliminated or significantly reduced. What's more, it is hard to see how interracial contact can even in principle reduce phenotypic devaluation through the familiar mediators of contact theory. It is hard to see how greater knowledge about blacks, more empathy with regard to them, or less anxiety in encounters with them can lessen it. Nor is it clear how other possible mediators could have this effect, or how contact itself could have it — our judgements of attractiveness do not seem to be a function of mere exposure or familiarity (see, e.g., Burke et al., 2013).

But there is one possibility that cannot be dismissed entirely. It has to do with the fact that mathematical averageness has been found to contribute to judgements of attractiveness (see, e.g., Rhodes, 2006). This raises the question: if average is attractive, what characteristics make a face average? The answer would seem to be the characteristics of the population of faces that individuals experience. With this idea in mind, Gillian Rhodes and colleagues (2003) performed experiments that showed that exposure to distorted faces can shift, at least temporarily, what perceivers perceive as average and consequently what they judge as attractive. They thus suggest that “people's preferences can be rapidly calibrated to match whatever physical characteristics are typical of the population of faces that they see” (p. 558). This suggests one way in which increasing interracial contact may reduce phenotypic devaluation: by increasing exposure to black faces, it may effect a change in what non-blacks perceive as average and so what they perceive as attractive.

Rhodes and colleagues’ study may seem to offer powerful evidence in favour of a possibility that my argument denies, but there is a very good reason to question it. The theory on which it is based would seem to predict that persons living in societies with same-race majorities should have a same-race preference in judgements of attractiveness, since they will be mostly exposed to persons of the same race. But considerable evidence confounds this prediction. There is no general same-race preference in judgements of attractiveness (Burke et al., 2013; Lewis, 2011; Rhodes et al., 2001). Participants in several studies exhibit an other-race preference, including individuals who live in societies where there are few other-race individuals.

V.2

There are at least three objections that my argument will prompt. First, it may be argued that, contrary to what I have suggested, phenotypic devaluation is in fact closely linked to stigmatization such that to the extent that the latter declines, so too will the former.⁴⁹ The suggestion is that negative aesthetic judgements about blacks depend on their stigmatized status, so that if blacks are no longer stigmatized, characteristically black phenotypic traits will no longer be negatively evaluated. Recent trends in popular culture may be offered in support of it. One historian, for example, has pointed out that since the 1980s an African American has won the Miss USA pageant once every three years (Watson, 2010).⁵⁰ And, in his recent book on black aesthetics, Paul Taylor (2016) suggests that there appears to be a narrowing of what he calls the “beauty gap” between African Americans and White Americans (Ch. 4; Taylor does recognize that things are more complicated than they appear). It may be argued that these changes reflect the fact that blacks are less stigmatized than they used to be.

In order to properly evaluate this argument, we need to be clear about how we should understand it. Recall first the distinction between pure and impure aesthetic judgements; aesthetic judgements, I suggested, can be infected by non-aesthetic impurities that taint them. Now if the objection is understood as claiming that eliminating or reducing black stigmatization will lead to revisions of judgements of black attractiveness because these are currently infected by the low social status of blacks and so are not pure, there may be something to it. It is not hard to believe that the negative aesthetic evaluation of a low social status group has been influenced by the group's low status. But the argument goes beyond this to claim that phenotypic devaluation is entirely a function of non-aesthetic impurities, and this is harder to believe. We do not generally think that social status is all that there is to judgements of attractiveness. And, in the case of blacks, many people can and do find many individual blacks to be physically attractive despite their low social status (though this point, by itself, is admittedly far from conclusive). And judgements of black unattractiveness seem to have been widely shared prior to the various ways in which stigmatization would later manifest itself (Barker, 1978; Jordan, 1968, pp. 8–11). But the main thing that I will say here is that the burden of argument should surely fall on those who would reject the moderate view that social status can influence judgements of attractiveness and embrace the more extreme view that such judgements are entirely a function of such factors. So, until that burden is fulfilled, I conclude that though it is likely that the history of stigmatization has influenced phenotypic devaluation, it is doubtful that it is wholly a function of it.

The evidence that I suggested may be offered in support of the objection should also be questioned. First, to the extent that there is indeed a narrowing of the so-called ‘beauty gap,’ it may more reflect changes in black grooming practices (such as hair processing techniques) and black phenotypes (through reproductive exogamy) than anything else. And, though it can be argued that the popularity of certain curvier African American female entertainers shows that there has been a change in body ideals, it is more likely that African Americans have created an alternative ideal that competes with, but is still subordinate to, a still-dominant mainstream ideal that valorizes thinness. Again, the case of Brazil is instructive. In Brazil, there is a racially mixed ideal defined in part by a curvier body shape. But this exists alongside a thinner ideal that some maintain is “associate[d] with a more distinguished taste” (Jarrín, 2015, p. 547). More importantly, it is far from clear that there has been any change at all for ideals of facial beauty, with unmixed blackness still considered the most unattractive.⁵¹ As Donna Goldstein (1999) explains about the Brazilian context, “The generally accepted equation is that a particular combination of white and black characteristics creates mulata beauty, but it is white characteristics alone that can also qualify in another (higher) category of beauty. Purely African characteristics with no mixture of white characteristics are considered ugly” (p. 569).

Consider now a second objection. As the reader will have noticed, I have repeatedly used Brazil as a real-life example of the potential self-worth harms of integration in circumstances of stigmatization and phenotypic devaluation. It may be claimed that this use of Brazil is unfair, for it suggests that in evaluating the permissibility and wisdom of integration its benefits have to be weighed against self-worth harms of an order of magnitude that all but assures the conclusion that integration must be rejected. In other words, it may be argued that the self-worth harms of integration can outweigh the benefits of integration only if something like a worst-case scenario obtains, but this is very unlikely to be the case at this point. We have simply progressed too far for another Brazil to be a real possibility.

This objection invites us to consider the facts about anti-black racism and phenotypic devaluation that obtain in the societies that concern us. I have already indicated that no detailed examination of this issue can be undertaken here. Still, I want to concede that Brazil is indeed something close to a worst-case scenario that may be unlikely to be fully repeated elsewhere. But the self-worth harms of integration do not have to reach Brazil-levels for them to outweigh the (realistic) benefits of integration, and if what I have argued about their absolute overriding weight compared to these benefits is right, likely harms will outweigh them.

Turn now to a third and final objection. It may be argued that in rejecting integration in order to protect black self-worth, I ensure that blacks will continue to suffer from the stigmatizing effects of a segregated racial inequality, which itself imposes its own self-worth harms. After all, it will be argued, I do not deny that such inequality is stigmatizing, though I reject the necessary means for ending it. Now I concede that segregated racial inequality can express stigmatizing messages, and that integration can help reduce this. But I do not believe that racial inequality cannot be ended or reduced without racial integration. Part of my thinking here is that I believe that there are a number of areas where action neutral with regard to integration might be taken to improve black opportunities. For example, there are ways of improving the education of black children that do not involve integrating their schools (Robbie & Fryer, 2011; Tough, 2008). But a bigger part of it is that I believe that we should not underestimate what a committed and determined black population can achieve despite the significant obstacles that it faces. Crucial here will be an honest reckoning with the legacies of past oppression.⁵² Also likely of some importance is some significant degree of class integration within black communities. Space constraints limit how much detail can be provided here, but I do not believe that racial inequality is inevitable. Moreover, without it, segregation — especially if voluntary — need not be stigmatizing.⁵³

V.3

The argument thus far suggests a future where the increasing interracial contact that integration involves reduces black stigmatization, but not phenotypic devaluation. Things may in fact be more complicated, but not for the better. For it is quite possible that insofar as stigmatization is reduced, this will increase the saliency of phenotypic variation among blacks, and heighten the devaluation of those blacks who rate high in phenotypic stereotypicality.

At this point, we must revisit an important assumption with which we have been working — something that will require conceding an important limitation of the overall argument. I have been assuming that the black population whose integration we are considering is a relatively homogenous group phenotypically. We must now relax this assumption. Individual blacks in fact differ significantly in the extent to which they are stereotypical phenotypically: some blacks are darker than others, have wider noses, and so on. There is reason to think that blacks who are more stereotypical will be more subject to phenotypic devaluation and so relational devaluation. For example, in three studies Michelle Hebl and colleagues (2012) found that “being perceived as more physically stereotypical was associated with having a smaller group of outgroup friends and fewer outgroup social interactions” (p. 1333). In one particularly important study, they altered photographs of moderately stereotypical blacks (one male and one female) to make them more and less stereotypical, and used these to create Facebook profiles, which were then used to send out friendship requests to randomly selected profiles. They found that “an overture of friendship from a Black person to a non-Black person was less likely to be accepted when the friendship requester looked more (vs. less) stereotypically Black. Further, even among those requests that were accepted, more time passed before acceptance when the requester looked more (vs. less) stereotypically Black” (p. 1331).⁵⁴

It may well be that the best kind of integration we can realistically hope for is one that will differentially incorporate blacks in the way suggested by Hebl and colleagues’ study. Lighter-skinned blacks and blacks with other less stereotypical features (e.g., narrower noses) may come to enjoy the full benefits of integration, while more stereotypical blacks are denied the same. Thus the fact that blacks as a group are characterized by significant phenotypic variability presents a problem for integrationists. But it also presents a problem for my overall argument, for the individual arguments that I have made apply entirely at the level of the individual, and yet individual blacks vary in their stereotypicality. So, for all that I have argued so far, it is not clear why less stereotypical blacks should not integrate. As members of a still-stigmatized group, to be sure, they will experience some relational devaluation rooted in stigmatization, but over time — as the integration process unfolds — this may be expected to decrease. Moreover, they would not be relationally devalued because of phenotypic devaluation. As a result, integration for them would not inflict, to the same degree, the self-worth harms that we have discussed.

A different possibility must be considered. It is possible that the integration of less stereotypical blacks, increasingly on terms of equality, will pave the way for the later integration of more stereotypical blacks. It may be argued that less stereotypical blacks might constitute a kind of ‘gateway’ group, exposure to which will improve general anti-black attitudes, the benefits of which will redound to all blacks. In a series of studies on this issue, Aharon Levy and colleagues (2019) found that the presence of bi-racial individuals reduced prejudice through the reduction of symbolic threat. But even if it is true that less stereotypical blacks can reduce the prejudice directed at more stereotypical blacks, it is a separate question whether they can reduce their phenotypic devaluation. I see no reason to draw this inference, based on this study.

So, if such gateway hopes should not be relied on, how do I propose dealing with the problem of phenotypic variation? At this point, I turn to an ideal of black solidarity. Solidarity among the members of a group can be effective in achieving important goals (such as justice for the group). But solidarity may also be valuable because of the value of the goods that it embodies (e.g., mutual trust, sense of special concern, and shared fate). Because this is so, it may motivate a principled stance toward some more or less permanent adversity, faced by some but not all members of the group, where there is no clear prospect in the foreseeable future that collective action will be effective in overcoming it. In such cases, solidarity may demand that more fortunate members of the group decline certain benefits that other members can't share. Acceptance of these benefits may be particularly destructive of solidarity. I believe that something like this is true for blacks. In particular, I suggest that solidarity demands that less stereotypical blacks decline opportunities for social inclusion that are denied to more stereotypical blacks. Included here are opportunities to integrate.⁵⁵

To be sure, it is not always inconsistent with black solidarity (or rather, with an ideal of it worth defending) for less stereotypical blacks to accept benefits denied to other blacks. Take the economic domain. It would seem that a high-income earning business executive can keep her high income for herself even when her lesser stereotypicality was a factor in — even a necessary condition of — her career ascent. Phenotypic variation makes a difference not only with respect to phenotypic devaluation, but stigmatization as well: blacks who are more stereotypical are more likely to be subject to negative evaluation (e.g., judged as less intelligent) and stereotyping (Ben-Zeev et al., 2014; Hannon, 2015; Maddox, 2004).⁵⁶ Thus lighter-skinned African Americans have been found to earn higher wages than darker-skinned African Americans, with the wage gap widening as the skin colour of workers darken (Goldsmith et al., 2007). It also has an impact in the domain of electoral politics (Hochschild & Weaver, 2007, pp. 650–652). Enjoying success here too would seem to be consistent with black solidarity. But social situations, I claim, are different. Here the acceptance by less stereotypical blacks of benefits denied to more stereotypical blacks is uniquely destructive of black solidarity, and so of the goods that it embodies.

As this suggests, it is an implication of my argument that while some blacks should reject integration on self-protective grounds, some should reject it on solidarity grounds. Interestingly, on my view, blacks who should reject it on solidarity grounds may have stronger — more absolute — reasons to do so than those who should reject it on self-protection grounds. Say that a society has some stereotypical blacks, some less stereotypical blacks, and that there are benefits to integrating. My view is that the benefits of integration can be given more weight by those blacks who, for their part, would evaluate it from the perspective of its self-worth harms than those who would evaluate it from the perspective of the demands of solidarity.

So phenotypic variation limits my argument in that it requires recourse to an ideal of solidarity that cannot be fully defended here. But the very facts that give rise to this limitation also raise the prospect that the self-worth harms of integration may be more significant than I have so far suggested. An integrated society would presumably contain more interracial unions, and this will over time increase the number of blacks who are less stereotypical. It may be that once the population balance of blacks who are more and those who are less stereotypical shifts in favour of the latter, this will give more licence to open expressions of phenotypic devaluation of the former. Some of us may think that our societies have progressed beyond such open denigration, but once it is understood that the denigration is feature- and not category-based, old norms against category-based denigration may no longer seem to apply. Moreover, this denigration may seem different if it is done by persons who are also considered to be black.

V.4

Let us turn now, more briefly, to the second question about intergroup contact, which concerns whether it has any unintended and unwelcome side effects. The problem here is that research on interracial contact indicates, not surprisingly, that it is close cross-racial group friendships that are “especially effective in reducing prejudice and promoting positive intergroup outcomes” (Pettigrew & Tropp, 2011, p. 119). The idea is that it is such friendships that can reduce intergroup anxiety and trigger the affective processes that are so important to prejudice reduction.

This is concerning because, as we have seen, the problem with integration is that it will tend to dissolve group boundaries, and what effects this result are the assimilative pressures that I have claimed it will inevitably exert. Close personal friendships would seem to be a main conduit for these pressures. So it seems that the very thing that may be most effective in reducing prejudice is just what will most expose blacks to the self-worth harms of integration.

If this is right, it shows that if integration is to fully produce its claimed beneficial effects, it may require forms of interracial contact that will likely lead to psychological integration, which I have claimed significantly increases the risk of assimilation. This in turn suggests that integrationists are in a dilemma: either integration will fall short with respect to close personal relationships, thereby possibly avoid assimilation but risk failing to sufficiently reduce even stigmatization, or it will achieve the requisite level of close personal relationships, achieve de-stigmatization, but lead to assimilation, and so expose blacks — especially those who are more stereotypical — to self-worth harms rooted in phenotypic devaluation.

So, even if integration can be effective in reducing stigmatization, it is not only doubtful whether it can also reduce phenotypic devaluation, but it may in fact make it worse for some blacks. But something must also be said about the ‘transition costs’ — transition self-worth harms — that integration will involve, for de-stigmatization, even if successful over the long term, can't happen overnight. The self-worth harms of relational devaluation rooted in stigmatization will not end as soon as the integration process begins. This is potentially quite significant, if one places great weight on avoiding these harms, as I do. Now if we add to this the uncertainty of the long-term effects of integration, on the one hand, and its very likely immediate self-worth harms, on the other, it would appear that integrationists want to risk sacrificing current and immediately succeeding generations in the name of a good whose attainment is far from guaranteed. For these reasons, I suggest that even if we put phenotypic devaluation aside, concerns about stigmatization alone seem to provide enough reason for blacks to reject integration.

VI.1

Discussions of the pros and cons of integration have not given enough attention to the issue of self-worth. In this paper I have argued that integration will likely have self-worth costs, and that these costs may be so significant that blacks should reject it on self-protective and solidarity grounds. It will, I argue, inevitably exert assimilative pressures, and assimilation may intensify the self-worth harms of stigmatization and phenotypic devaluation by leading blacks to internalize racial and/or phenotypic hierarchies. In making this argument, I have sought to stress the importance of considering phenotypic traits in thinking about integration. Unless we attend to the way that racial features differentially affect the black experience with stigmatization and especially phenotypic devaluation, I suggest, we may fail to notice what may be integration's greatest threat. The problem is that in conditions of phenotypic devaluation it may make blacks aspire, not to whiteness, but to a lesser kind of blackness. And we may fail to notice this problem if we focus exclusively on racial categories.⁵⁷

Now what does my argument mean for specific forms of integration? The answer to this is less clear than it may be appear. To be sure, I certainly maintain that assimilation should be avoided, and the most vulnerable blacks in this regard would seem to be black children. But integration in which domains will lead to the assimilation of black children? Though I earlier singled out residential integration as posing a special threat, I do not believe that it makes black internalization of racial or phenotypic hierarchies inevitable. For certain individual blacks, a sufficiently robust sense of racial identity may be compatible with residential integration, even when paired with school integration — there may be a certain psychological orientation that can block some of the effects of residential integration. Crucial here are parental practices of racial socialization. The skilled use of such practices may protect the self-worth of even children raised in predominately white environments where blacks continue to be devalued, especially if blacks form a sufficient critical mass. But this should not generally be relied on. Some black parents may not have the requisite knowledge and skills to appropriately socialize their children; these must be learned, but we should not presume that all blacks have done so. This is especially true for black immigrants who have no history of living in multiracial societies. By contrast, the racial socialization of black children is less important in a context of residential segregation.

At the same time, more innocuous-seeming forms of interracial contact may come with surprising risks, even in a context of residential segregation. The problem here has to do with the fact that, as Michael Merry (2013) has written, “one can live one's entire life within a segregated neighborhood and in a number of crucial ways still be integrated” (p. 170). For example, a person might live in a black neighbourhood while having a mostly white circle of friends. This kind of situation will exert assimilative pressures too. Thus, if my argument is right, it should be avoided.

In her book on integration, Stanley (2017) distinguishes between a temporary resistance to integration and a more permanent rejection of it (pp. 187–188). If my argument is right, it is true that it suggests a resistance that may be more than temporary: as long as phenotypic devaluation persists, blacks should resist integration. And yet my argument doesn't quite provide reason to favour a final rejection of it either. For if it becomes clear that phenotypic devaluation is not quite the problem that I believe it is, then one main reason for blacks to resist integration can be put aside. Until then, however, blacks should consider it off the table.

VI.2

Now it may be wondered why I should devote so much energy to arguing against something that I have already indicated has fallen out of favour among most African Americans. The primary answer has to do with something that I mentioned in the introduction: I intend my argument to have international application. I suspect that when we turn our attention to non-African American blacks — African Canadians, Black British people, and others — we will find that there are a great many who still find integration appealing.⁵⁸

And yet if integration is more appealing to other blacks, they generally may be more at risk of suffering the self-worth harms I have described. I have suggested that blacks are differentially subject to phenotypic devaluation based on their stereotypicality — in particular, that more stereotypical blacks are more subject. If this is right, then integration will not inflict the same degree of harm on all blacks. But now notice that this is true not just at the level of the individual, but also at the level of the state. Different black nationalities differ in the extent of their general stereotypicality; African Americans, for example are less stereotypical than African Canadians, largely because most African Canadians are recent immigrants or the descendants of recent immigrants from countries with a less significant history of intermixing. If all of this is true, then the blacks who are most likely to still find integration appealing are just those who are most likely to be harmed by it.