1. A Very Brief History of the Race Idea

François Bernier, writing in Reference Bernier, Bernasconi and Lott1684, has been described as the first to employ ‘race’ as a word ‘in something like its modern sense to refer to discrete human groups organized on the basis of skin color and other physical attributes’ (Bernasconi and Lott Reference Bernasconi, Lott, Bernasconi and Lott2000: 1). He does not claim to distinguish races through differences in natural abilities and dispositions. His originating, thin idea of race—which reduces to difference in visible markers of continental ancestry—did not supply much fodder for debate.

Races, for Bernier, are only slightly more than skin deep:

Bernier indicates no concrete use for this racial division. His classification scheme is limited to patterns of physical difference. While he does observe that physical differences of race are heritable, his rationale for identifying Africans as one of the ‘species or races’ is not about setting up a racial hierarchy.

Race becomes a less innocuous idea during the Enlightenment. Kant's scientific spin on the race idea is perhaps most responsible for this development. He introduces a reasonable distinction between species and race: ‘Negroes and whites are clearly not different species of human beings (since they presumably belong to one line of descent), but they do comprise two different races’ (Kant Reference Kant, Bernasconi and Lott1777: 9). This could have corrected loose, interchangeable talk of ‘species’ and ‘race’ while retaining the spirit of Bernier's classification of peoples by continental ancestry. Instead, Kant pursues deeper truths that would reveal there is more to race than patterns of mere physical difference. After describing purportedly repugnant physical traits of blacks, he asserts, ‘[The Negro] is well-suited to his climate, namely, strong, fleshy, and agile. However, because he is so amply supplied by his motherland, he is also lazy, indolent, and dawdling’ (Kant Reference Kant, Bernasconi and Lott1777: 17). Kant thus inaugurated the tradition of giving speculatively naturalistic backing to antiblack stereotypes. He goes on to admit, ‘[M]y opinions in these matters are only preliminary, and I offer them only for the purpose of stimulating further investigation’ (Kant Reference Kant, Bernasconi and Lott1777: 19)—an ostensibly cautious approach to voicing judgments about the innate deficiencies of blacks that has proven visionary. (Kleingeld [Reference Kleingeld2007] argues that during the 1790s Kant gave up his hierarchical conception of race. Having reviewed the textual evidence, I find only that Kant moderated and restricted his hierarchical views.)

Fast forward to the beginning of the twenty-first century. Experimental psychologist Steven Pinker eagerly predicts that ‘the dangerous idea of the next decade’ will be that ‘groups of people may differ genetically in their average talents and temperaments’ (Reference Pinker and Brockman2007: 13). The hypotheses he highlights include an economist's gesture toward research that women and men have different ‘cognitive abilities and life priorities’, a biologist ‘rebutting the conventional wisdom that race does not exist’, and a rehashed argument that ‘average racial differences in intelligence are intractable and partly genetic’ (Pinker Reference Pinker and Brockman2007: 13–14). Pinker is not necessarily endorsing these views. He believes that ‘the evidence for gender differences is reasonably good, for ethnic and racial differences much less so’ (Pinker Reference Pinker and Brockman2007: 14). But he dismisses resistance to his dangerous idea as the ideological, antiscientific byproduct of a worthy commitment to the moral and political equality of persons. Critics, he imagines, are ‘unwilling to grasp’ that innate group differences ‘pertain to the average or variance of a statistical distribution’, not to individuals: ‘Large swaths of the intellectual landscape have been reengineered to try to rule out these hypotheses a priori (race does not exist, intelligence does not exist . . .)’, at a time when ‘genetics and genomics will soon enable us to test hypotheses about group differences rigorously’ (Pinker Reference Pinker and Brockman2007: 14). Sophisticated science finally could confirm, according to Pinker, that blacks and women by nature are not as intelligent, on average, as white men.

The histories of philosophy and science are littered with discredited, often preposterous claims about race and gender differences. Now, we are told again that science truly might discover innate group differences—though few critics would deny that there are such differences in certain domains, depending on where group lines are drawn. Group-average differences in height and physique, for instance, will contribute to group-average differences in performance for sports like basketball and swimming. By contrast, intelligence hardly seems as promising a candidate for investigation into innate group differences, given the lack of an adequate explanation for the factors yielding group differences in IQ test results (see, e.g., Neisser et al. Reference Neisser1996; for trenchant critiques of IQ testing, see, e.g., Gould Reference Gould1996 and Block Reference Block1995). In vigorously defending such investigation, Pinker mentions nothing specific about how a revolution in genomics could break new ground.

I am here mainly concerned with the criteria by which science would rebut the view that races do not exist. Since no one disputes that certain physical characteristics (e.g., skin color, facial features) are highly correlated with continental ancestry, decoding genetic determinants for these physical characteristics would not settle the contested question of whether there are races. If this were all race amounted to, the uncontested answer would already be that races exist. Evidently, race scientists have believed that something else or something more is at stake in exploring whether there are innate group differences that substantiate the existence of race. This something else or something more usually leads to the cognitive domain, where Pinker would have us focus. Why might any innate cognitive differences between supposed racial groups be so important?

Evolutionary biologist Stephen Jay Gould suggests why race theorists have been preoccupied with innate group differences in intelligence. ‘Biological determinism’, as he describes it, ‘holds that shared behavioral norms, and the social and economic differences between human groups—primarily races, classes, and sexes—arise from inherited, inborn distinctions and that society, in this sense, is an accurate reflection of biology’ (Gould Reference Gould1996: 52). The implicit upshot is that intervening to offset natural inequality could well be an unfair, inefficient, or futile exercise in social engineering. This blurs the boundary between observation and justification of certain outcomes in a society. There could be innate group differences marginal to explaining why societies are organized as they are. But if the differences were limited to quite superficial (e.g., physical) or narrow (e.g., genetic predispositions to certain diseases) domains that do not have profound social significance, there would be nothing particularly controversial about the claim that these differences support an idea of race. Cognitive differences, by contrast, seem better suited to ground a strong thesis about the social consequences of racial nature.

In effect, then, the race debate has depended on a thicker idea of race. Scientific racialism, which is a version of biological determinism, raises the stakes in the debate. A prerequisite for racialism is the notion that racially heritable uniform traits, or racial ‘essences’, distinguish human groups from each other. Kant's racialism won the early debate against the likes of Blumenbach's misappropriated, non-racialist anthropology (Blumenbach Reference Blumenbach, Bernasconi and Lott1795). While I am roughly in agreement with Michael Hardimon (Reference Hardimon2003: 455) about the race idea's historical development, I am puzzled by his view that this history yields a master concept that is critical for recognizing ‘striking’ patterns of physical difference and analyzing racism. The race idea was and is contested, which is why there has been never-ending controversy over the word's meaning. For Kant, race warrants conceptualizing because he has a thick idea of racial nature: certain patterns of physical difference are a sign of uniform traits that yield cognitive differences. Kant is a founding father of racial essentialism. I contend that racial essentialism is where the core idea of race originates and often still resides.

2. Renewed Race Science

More than a few scientists and philosophers of science defend the idea of race because they believe that, biologically speaking, there are major human races. Few of these race theorists openly subscribe to the core, essentialist idea of race as I have described it: generally, they have no commitment to the view that ‘race’ refers to human groups determined through continental ancestry, some of which by nature are significantly less capable cognitively, a condition that might justify lesser socioeconomic outcomes for less capable groups. The renewed science of race mostly disowns the biological determinism aspect, though scientists and philosophers do not always doubt that there might be innate significant cognitive differences between races (on the philosophical front, see, e.g., Sesardic Reference Sesardic2010 and Levin Reference Levin2002). Our question, then, is what would scientifically vindicate the idea of race?

Evolutionary biologist Ernst Mayr insists that the answer is uncomplicated and should be uncontroversial. Persons of the opinion that ‘“there are no human races”’, he complains, are ‘obviously ignorant of modern biology’—since races ‘occur in a large percentage of species of animals’, including the human species (Mayr Reference Mayr2002: 89). For Mayr, the word ‘race’ gets its meaning in taxonomy, the scientific practice of classification: ‘A subspecies is a geographic race that is sufficiently different taxonomically to be worthy of a separate name. What is characteristic of a geographic race is, first, that it is restricted to a geographic subdivision of the range of a species, and second, that in spite of certain diagnostic differences, it is part of a larger species’ (2002: 90). He remains unperturbed, 60 years after fellow biologists pointed out the problem, that his taxonomic definition of ‘race’ circularly introduces the perspective of what scientists will count as differences ‘worthy’ enough to merit a division of peoples into races.

Mayr cites the sizable overrepresentation of ‘contenders of African descent’ in Olympics sprint finals as ‘surely not an accidental percentage’ and evidence that ‘each human race consists of individuals who, on average and in certain ways, are demonstrably superior to the average individual of another race’ (Reference Mayr2002: 91). Molecular biologist James D. Watson (Reference Watson2007) makes race claims in a similar spirit, with the stark implication that blacks are of athletically superior body and lesser mind (Hunt-Grubbe Reference Hunt-Grubbe2007). Apart from the tired fallacy of drawing broad conclusions about innate racial differences by citing sports performance, we should notice that Mayr's sprinters are mainly Black American and Afro-Caribbean—diaspora peoples mostly traceable to West Africa, who are conventionally recognized as ‘black’ and assigned to the ‘African’ racial category despite a considerable rate of non-African ancestry among them. (Native Africans, mainly from East Africa, in fact are overrepresented among world-class distance runners, not sprinters.)

The general argument that E. O. Wilson and W. L. Brown Jr. make in their critique of Mayr's ‘subspecies concept’ is this: the notion of a ‘genetically distinct geographical fraction of the species’, following ‘whatever “diagnostic” characters are chosen to delimit races’, is ‘subjective and arbitrary in taxonomic practice’ (Reference Wilson and Brown1953: 109). There are no natural, objective criteria for drawing boundaries between human races or any other subspecies of animal. Which characteristics are significant enough for a subspecies designation will have to be a judgment call—given the aims of taxonomy and the perspective of taxonomists who believe that certain patterns of difference warrant a subspecies designation. For the human species, patterns of physical difference tied to continental ancestry fall short of nonsubjectively indicating natural races, contrary to the conventional wisdom.

Population geneticist L. Luca Cavalli-Sforza and colleagues sum up the quandary:

Mayr thinks he has an objective stopping level for his taxonomy: ‘the geographic races of the human species [were] established before the voyages of European discovery and subsequent rise of a global economy’ (2002: 90). Actually, over the past 500 years, this global economy (initiated through colonization and the transcontinental slave trade) greatly complicates the geographic isolation that he supposes—especially in the Americas, where there has been extensive reproductive mixing from African, indigenous American, and European sources. His story also ignores longer-standing genetic differentiation of peoples within continents as well as short travel distances between continents. All of this is to say that Mayr draws taxonomic divisions where he already sees human races—namely, wherever he finds morphological, physiological, or genetic differences that can mark purportedly common-sense racial groupings.

Philosophers of science have tried to improve upon this conceptually naïve taxonomy of human races. Philip Kitcher and Robin Andreasen, for example, independently propose a ‘phylogenetic’ approach that ‘defines races in part as lineages of reasonably reproductively isolated breeding populations’ (Andreasen Reference Andreasen, Matthen and Stephens2007: 468). These breeding populations are supposed to be genetically distinctive enough to support individuating them as races. Kitcher does not take his ‘minimalist notion of race’ to reflect much more than ongoing reproductive patterns that continue to generate ‘three major races’—albeit ‘in highly qualified form’ in locales, such as the United States, that saw extensive reproductive mixing (Reference Kitcher and Harris1999: 103–4; see Kitcher [Reference Kitcher2007] for his ‘pragmatic’ turn). Phenotypic differences or ‘looks’ that distinctly indicate African, Asian, or European ancestry can be all there is biologically to this notion. Andreasen offers a similarly minimalist ‘cladistic race concept’ that designates racial groupings ‘solely in terms of common ancestry’ (2007: 471–72). Yet, this strong condition of genealogical differentiation leads her to speculate that human races ‘are likely on their way out’, since ‘reasonable reproductive isolation’ has been subjected to ‘too much gene flow’ among human breeding populations to support useful taxonomic racial divisions going forward (Andreasen Reference Andreasen, Matthen and Stephens2007: 472–73; for resistance to ‘cladistic subspecies [as] a genuine biological candidate for race’, see Spencer [Reference Spencer2012: 203]).

Anyway, the question is not whether reproductive patterns can create and sustain genetically differentiable groups of people: this is plainly possible, as the case of genetically differentiable ‘white’ peoples of Europe demonstrates (see, e.g., Cavalli-Sforza et al. Reference Cavalli-Sforza, Menozzi, Piazza and paperback1994: 268–72). Racial classifications originally were made on the basis of physical differences associated with continental ancestry. But we have no reason to presuppose that these differences mark a thick natural basis for designating human groups. We need not doubt that once certain race lines have been conventionally drawn, some feature of scientific or social interest could be found that rationalizes drawing the lines accordingly. This does not mean that we will have discovered racial difference in any deep, objective sense. Andreasen, like Kitcher, acknowledges that a phylogenetic approach is quite removed from the essentialist idea of race: phylogenetic accounts ‘are often taken to be improvements over their predecessors, in part, because they are relatively minimalist in comparison with previous biological conceptions’, by virtue of not being partial to racialist views about cognitive abilities and dispositions (Andreasen Reference Andreasen, Matthen and Stephens2007: 468).

If innate cognitive differences are not invoked to vindicate the race idea, then the question of scientific vindication might already and too easily be answered—by an appeal to skin-deep differences tied to continental ancestry. Genetic predispositions to certain diseases and other, narrow genetic characteristics do not support a classification project that sustains inquiry about deep racial difference—as the effort by Neil Risch and colleagues (Reference Risch, Burchard, Ziv and Tang2002) to vindicate the race idea through appeal to epidemiology, for example, has shown. The most credible candidates for a biological race concept have not been thick and thus do not capture the ordinary interest driving the preoccupation with the idea of human races. To the best of current scientific knowledge, even a conceptually sophisticated racial taxonomy would win a Pyrrhic victory that tends to perpetuate essentialist myths and confusions about racial peoples.

3. Constructing a Meaning of ‘Race’

Many race theorists, including those who do not believe there are natural races, seem unwilling to give up on the race idea. These theorists believe that since race—apart from its biological standing—is thickly real ‘as a social kind of thing’, we should recognize social reality that lends practical meaning to ‘race’ talk (Glasgow Reference Glasgow2009: 5). The historian, sociologist, and civil rights activist W. E. B. Du Bois inaugurated this line of thought over 100 years ago.

In ‘The Conservation of Races’, Du Bois sketches an argument for the social reality of race: ‘Although . . . the grosser physical differences of color, hair and bone go but a short way toward explaining the different roles which groups of men have played in Human Progress, yet there are differences—subtle, delicate and elusive, though they may be—which have silently but definitely separated men into groups. . . . [These differences] have divided human beings into races, which . . . perhaps transcend scientific definition’ (Reference Du Bois, Bernasconi and Lott1897: 109–10). This is a prototype of the contemporary view that race is mainly, if not entirely, a social construction. While biology explains physical differences that can distinguish groups of people, these groups become races, in a thick sense, when innate yet somehow nonbiological differences are tied to the physical differences. As we have seen, the essentialist idea of race had been employed to rationalize hierarchy and subordination. Against this tradition, Du Bois proposes a reconceptualization of race that allows for political equality, multicultural value, and cross-cultural respect (for elaboration and defense of races as distinct cultures, see Jeffers Reference Jeffers2013).

Specifying the idea of race remains a challenge for social constructionists. In philosophy, a standard approach for investigating what ‘race’ refers to is to call upon some preferred theory of meaning. To paraphrase Maslow's dictum: when all you have is a hammer, every problem looks like a nail. Sally Haslanger gives a concise summary of competing theories of meaning in the philosophy of language as they apply to race:

Our subject of inquiry has become hazy. The race debate had been about whether there are thick natural races—not whether the word ‘race’ can be shown to refer to anything biologically or socially real. Some race theorists might be tempted to reframe the long-standing debate as a question of whether races exist in any sense. Since social phenomena (e.g., teams, in-laws) exist as part of the social world, racial groups as social constructions could be said to exist objectively as a social phenomenon. However, the long-standing debate concerned whether races exist as part of the natural world, independently of their conceptualization as what ‘race’ is taken to mean.

Haslanger's social constructionist approach presupposes a determined effort to make sense of ‘race’ talk—which comes close to dictating that races must exist in some sense. On her preferred theory of meaning, the meaning of ‘race’ arises through ordinary use of the term today: ‘The rational improvisation model [of pure reference externalism] invites us to consider the social dynamics, collaboration, and reflective practice required for shared meanings’ (Haslanger Reference Haslanger2010: 183). Our idea of race is not, she contends, ‘“non-negotiably committed” to a biological basis for race that we know does not obtain’, let alone to a racialist one (Haslanger Reference Haslanger2009: 5). There is no such nonnegotiable commitment, of course, on the assumption that related propositions are widely believed to be false. But Haslanger's improvisation model leaves a lot of room for negotiating ad hoc the content of the race idea. Moreover, lack of public consensus does not inspire confidence in what others think they know about race. Ordinary ‘race’ talk will indeed reflect some commonplace beliefs. At the same time, improvising the meaning of ‘race’ gives the appearance of reaching into a rationalizing grab bag: unremarkable beliefs can be kept (e.g., persons identified as black typically have visible African ancestry) while more controversial beliefs that also reflect ordinary use can be discarded (e.g., some racial groups innately and nontrivially differ in average intelligence).

In earlier work, Haslanger tacitly acknowledged this type of rationalizing move when she implied an anti-essentialist, revisionist shift of the meaning of ‘race’ (Reference Haslanger2000). Her updated view holds that as long as people think, act, and talk as if things called ‘races’ exist, there are races in some familiar enough sense: ‘[I]t is misleading to suggest (as I myself have sometimes done) that social constructionist accounts of race are revisionist; the issue is what counts as the important features of our past practice of using the term “race” as we move forward’ (Haslanger Reference Haslanger2010: 181). A rationally improvised meaning would arise from a low common denominator that does ‘justice both to the historical collective practice and the worldly facts’, enabling us to ‘co-refer’ when ‘engaged in a shared project of understanding our representational tradition in using the term “race”’ (Haslanger Reference Haslanger2010: 181). When the facts remain in dispute, a philosopher might settle the matter as she sees fit, given what enlightened people are supposed to know. Such a model for explaining the meaning of ‘race’ does dictate a revisionist shift in meaning—insofar as the essentialist orientation of race ideology is left behind by fiat. Humpty Dumpty returns in social constructionist garb.

4. Racial Pragmatism

According to Appiah, trying to conserve ordinary use of the race idea is almost bound to reflect or sustain critical errors. His eliminativist critique of Du Bois's racial thinking (for circularity) broadly concludes, ‘The truth is that there are no races: there is nothing in the world that can do all we ask “race” to do for us. The evil that is done is done by the concept and by easy—yet impossible—assumptions as to its application’ (Appiah Reference Appiah and Gates1986: 35). This goes beyond the claim that racial essentialism is false—that there are no major races generally distinguishable by innate, significant cognitive differences. There are no races in any interestingly familiar sense, period, Appiah believes; ‘race’ talk tends to be counterproductive in social practice, and racial identities are suspect as a mode of social identity. His eliminativism cautions that any race idea supporting a color-conscious division of peoples—even a thin idea reducible to little more than certain visible markers of continental ancestry—risks distortion by the legacy of belief in races as thick natural kinds (Appiah Reference Appiah, Appiah and Gutmann1996: 73n).

Appiah believes that any account of substantial racial difference is lost without racial essences. On his view, ‘you can get various possible candidates [for a race concept], but none of them will be much good for explaining social or psychological life, and none of them corresponds to the social groups we call “races” in America’ (Appiah Reference Appiah, Appiah and Gutmann1996: 74). Consider ‘the one-drop rule’, formerly recognized in American law, whereby a person counts as racially black if and only if she has at least one traceable African ancestor. As Appiah observes, ‘[M]ost people who are African-American by the one-drop rule are, are regarded as, and regard themselves as white. Most people in the United States have a social conception of the African-American identity that entails that this is not so’ (2002: 284; for a critical response to Appiah's use of this apparent conundrum, see McPherson and Shelby [Reference McPherson and Shelby2004]). Rather than attempting to qualify and revise our ‘race’ talk so that it conforms to actual practices of assigning people to racial groups, we might come to doubt the value of racial thinking. This version of racial eliminativism is skeptical about the prospects for reconstructing the concept of race as a social kind untainted by its history of essentialism.

There has been heated resistance to Appiah's overall approach to race. Critics charge that his eliminativism uncharitably renders racial identities hostage to belief in racial essences. In a Du Boisian key, these critics argue that racial identities can be socially important, apart from an essentialist (un)reality of race (for critical responses to Appiah's critique of Du Bois's ‘Conservation’, see, e.g., Gooding-Williams [Reference Gooding-Williams2011]; Outlaw [Reference Outlaw, Bell, Grosholz and Stewart1996]; and Taylor [Reference Taylor2000]). Appiah is not wrong to highlight and reject an ambiguous kind of racial essentialism that remains in popular circulation. Yet, critics believe he misconstrues Du Bois's principal motivations and aims—and obscures the possibility of a coherent, reasonable interest in racial identities. Cornel West places Du Bois squarely in the philosophical tradition of American pragmatism, citing the influence of William James on Du Bois's sense that pragmatism's grounding in social reality is particularly relevant to ‘the Afro-American predicament’ (Reference West1989: 139). West and other contemporary pragmatists reject eliminativism: they accept that races are not thickly natural but insist that races are thickly social.

As Paul Taylor construes racial pragmatism, there are thick social races because ‘a racial description of a person effectively locates that person in a racialized scheme of social interaction’, exerting pressure on persons to self-identify accordingly (2004: 112; see also Gooding-Williams Reference Gooding-Williams and Boxill2001). Such a scheme is produced by the enduring and mutable practice of classifying people as members of major racial groups. ‘[Western races] are the probabilistically defined populations that result from the white supremacist determination to link appearance and ancestry to social location and life chances’, Taylor claims (Reference Taylor2004: 86). This would explain how races are social kinds and why race is still an important mode of social identity. Whether there are somehow natural races becomes secondary to the question of whether people have acted as if there are races in ways that have brought the corresponding groups into social existence. Recognizing the social reality of race would make sense of ordinary ‘race’ talk in our time.

Racial ontology gains greater clarity through a distinction Lawrence Blum draws between natural races and racialized groups: ‘For several centuries Americans believed that the groups we now call “whites,” “blacks,” “Asians,” and “Native Americans” were [thick natural] races’, he observes; ‘and as a result of this belief the groups have been treated so’ (Blum Reference Blum2002: 147). The process of ‘racialization’ that Blum elaborates is ‘the treating of groups as if there were inherent and immutable differences between them; as if certain somatic characteristics marked the presence of significant characteristics of mind, emotion, and character’ (2002: 147). These racialized groups have persisted after belief in essentialist racial hierarchy has been widely renounced. Blum would have us shift from ‘race’ talk to ‘racialized’ talk.

For racial pragmatists, though, racialization validates talk of races and racial identities. Taylor argues that racial identities, especially for members of racially subordinated groups, can be a source of cultural bonds, social competence, and antiracist consciousness—given that racialized group and personal histories continue to have serious political, economic, and interpersonal consequences (2004: 113–14, 85–86). I do not disagree that racial identities can play this role. But the idea of racialized groups does not fully disambiguate ‘race’ talk: we get a conceptual understanding of these groups that derives from what races were falsely supposed to be, which is shadowed by stubbornly popular belief in thick natural races. Moreover, the notion that ‘race’ talk is central to the social dynamics of color seems to arise from losing sight of the fact that visible continental ancestry, not the race idea, is the root of the social reality of color consciousness.

On the limited question of whether races exist, I am sympathetic both to the spirit of racial pragmatism and the letter of racial eliminativism. However, I favor adopting a deflationary approach: plausible or implausible views of race depend on what ‘race’ is taken to mean. Racial pragmatists turn the race debate into primarily an issue of philosophical methodology: they come to the defense of talk of races and sideline the foundational question of whether there are thick natural races. My perspective does not privilege the reality of natural kinds over social kinds. Rather, I have argued that racial thinking in general remains unstable. Yet, I have no interest in rejecting all possible meanings of ‘race’ or in turning skepticism about the race idea into skepticism about color-conscious social identity.

On the foundational question, there is no substantial dispute between racial eliminativists and racial pragmatists. My deflationary pluralism would defuse their contest over racial ontology. We could say that races exist, depending on the meaning of ‘race’ different partisans use to capture their respective ideas. Or we could say that the question of whether races exist is empty until particular meanings of ‘race’ are specified. In neither case is a master meaning on the horizon, and this reflects a lack of clarity and consensus about what people have been talking about.

5. Introducing Socioancestry

Skepticism about the race idea does not entail rejecting all forms of color-conscious thinking. No less a skeptic than Appiah acknowledges that racial identities can be instrumental in resisting racism. But he is convinced that many people counterproductively regard ‘racial identity as a species of cultural identity’ (Appiah Reference Appiah, Appiah and Gutmann1996: 83). A stock example is Black American kids purportedly risking the peer accusation of ‘acting white’ for being serious students, speaking Standard English, and not listening to rap music. ‘[I]t is not that there is one way that blacks should behave’, Appiah claims, ‘but that there are [reputedly] proper black modes of behavior’ (1996: 97). His objection is that racial identities are linked with cultural norms believed to naturally express a person's being of a certain race. He suggests that Black Americans, individually and collectively, might be better off with a much weaker attachment to their racial identity, which he contends is often incoherent (Appiah Reference Appiah, Appiah and Gutmann1996, Reference Appiah and Peterson2002).

This line of objection becomes less compelling when color-conscious social identity is distinguished from racial identity. Set aside whether Appiah's skepticism about the color consciousness of Black Americans is overinflated. Our question is whether color-conscious social identity is almost inevitably burdened by assumptions about racial nature. The answer, I argue, is no. Moving on from the race idea would help to free color consciousness from lingering assumptions about racial nature and its influence on mind and culture. Skeptics and pragmatists alike seem under the misimpression that some concept of race is integral to color-conscious social identity. The confusion is a product of the ambiguity surrounding ‘race’ talk—which can refer to continentally identified groups simply or continentally identified groups that supposedly constitute natural races. Neither the race idea nor racial identity is needed for a reasonable color consciousness.

To help remedy this confusion, I propose reframing racial thinking as socioancestral thinking. We can recognize socioancestral groups that give rise to socioancestral identities—rather than continuing to recognize racial identities or refusing to recognize color-conscious social reality. Consider the ‘major races’: Caucasoid/white, Mongoloid/yellow, and Negroid/black, which are tied to the continental regions of Europe, Asia (not including the Indian subcontinent), and (sub-Saharan) Africa, respectively. The standard membership of each of these continentally identified groups consists of persons who have the full associated ancestry; that is, their ancestors, roughly dating from 1500, come from only one of these continents. In this case, there is supposed to be no reproductive mixing and so no classification problem. As we have seen, however, many persons identified as white, black, or (in parlance not everywhere defunct) yellow do not have the full associated ancestry: they nonetheless have been assigned to a single racial group or ascribed a single racial identity. This is where social conventions figure in the idea of socioancestry. Socioancestral groups are a function of globally prevalent, locally variable, color-conscious social dynamics that reflect facts about the full, or some component of the partial, continental ancestry of a group's members.

Facts about continental ancestry do not necessarily translate into a ‘one-drop’, majority, or plurality ‘blood’ quantum criterion for socioancestral group membership. In this respect, socioancestral thinking parallels racial thinking's near lack of a global principle for group assignment when traceable continental ancestry is mixed. For race, though, the logic of descent does generally support a global principle when both biological parents are assigned (not incorrectly per the operative scheme) to the same race and thereby produce children who also are supposed to be of that race. Socioancestral thinking in this case would diverge from racial thinking: the social dimension of the former can allow enough conceptual flexibility to reflect literally color-conscious social status or reception. Biological parents of the same socioancestry might have a child who was placed for adoption at birth and later is of a different socioancestry—if, say, she were to grow up in a family of that different socioancestry, had some of the associated continental ancestry, and ‘looked’ and accepted the socioancestral identity. For socioancestry, as compared to race, the logic of descent is more readily defeasible and plainly anti-essentialist.

Let's explore a standard case. Afro-Caribbeans, Black Americans, and (sub-Saharan) Africans or African peoples can be said to belong to a black socioancestral group comprising Africa-identified peoples. Traceable African ancestry has only been a necessary, but not also a sufficient, condition for being identified as racially black. The same holds for black socioancestral identity. My deflationary pluralism about the existence of race does not imply that socioancestral blackness, as it were, is possible without some component of traceable African ancestry. Nor does deflationary pluralism imply that if there are no blacks in some thick natural sense, blacks cannot constitute a real, conceptually sound group. But the idea of socioancestry is not very revisionist about the grounds for recognizing socioancestral group membership. Afro-Caribbeans, Black Americans, and Africans are different socioancestrally black subgroups whose members typically have visible African ancestry and have been thought to share distinguishing, socially significant features with other persons of whatever nonnegligible quantum of traceable African ancestry.

Socioancestral thinking is subject to local variation. For example, Americans who formerly would be recognized as racially black have a reasonable foundation for identifying with the particular mode of socioancestral blackness constitutive of, I submit, a Black American social identity: namely, they have some traceable African ancestry and (a) strongly identify as descendants of slaves in the United States, or (b) when not descendants of slaves in the United States, strongly identify with them via distinctively African physical features and the social status or reception that accompanies those features. This type of strong identification, as I am understanding it, is more historical and sociological than cultural—which is why socioancestral identity should be distinguished from ethnicity and its emphasis on a shared cultural tradition.

Americans who have visible African ancestry—that is, who ‘look black’—are not normatively bound to identify strongly as or with descendants of slaves in the United States and thereby to claim a Black American social identity. Socioancestrally non-Black Americans who have visible African ancestry would include persons from immigrant backgrounds who disclaim—via appeal to their particular history, culture, or homeland conventions for color-conscious identity—strong identification with descendants of slaves in the United States. Some of these persons might not strongly identify with any Africa-identified peoples. An example would be Brazil's ‘pardos’, who are of mixed African and non-African ancestry: they do not identify as usually darker-skinned ‘pretos’ (i.e., ‘blacks’), who tend to have some greater degree of African ancestry mixed with their non-African ancestry. Self-identifying pardos, whether in Brazil or as immigrants to the United States, would not have a socioancestrally black identity of any kind.

Persons of mixed African and non-African ancestry whose African ancestry is visible might well continue to be regarded as ‘black’ in the United States and beyond—regardless of whether they accept a socioancestral identity as black. Would the persistence of a global, nonelective conception of blackness demonstrate that the idea of race is almost indispensable to color consciousness, including the idea of socioancestry? I think not. Color and continental labels such as ‘black’, ‘white’, and ‘Asian’, apart from their use as designations for natural races, are often used simply to represent visible continental ancestry (in Western convention, relative to a baseline of visible non-European ancestry). The same color and continental labels could be easily enough repurposed for socioancestral thinking, notwithstanding local variation we already find in their application as ancestral, racial, and ethnic markers.

The notion of ‘mixed race’ is sometimes invoked against assigning persons of mixed continental ancestry to a single racial group (see, e.g., Zack Reference Zack and Zack1995b). Linda Alcoff proposes an expansive Latino conception of mixed-race identity to serve as a model: ‘Mestizo consciousness is . . . a conscious articulation of mixed [European and Native American] identity, allegiances, and traditions’, which could include ‘[a]ll forms of racial mixes . . . thus avoiding the elaborate divisions that a proliferation of specific mixed identities could produce’ (1995: 277; see Alcoff [Reference Alcoff2006] for her somewhat less sanguine vision of mixed race). But reality on the ground resists compliance. In much of Latin America, non-monoracial schemes appear to entrench, not diminish, an essentialist notion of racial difference. ‘Part of the reason for a multiplicity of descriptions for nonwhite Brazilians, particularly for those whose African descent is visible’, political scientist Michael Hanchard claims, ‘is [that] such categorizations attempt to avoid the [racial] mark of blackness’ (Reference Hanchard1994: 177; see also Gates Reference Gates2011). This explicit or implicit distancing from racialized blackness is an instantiation of racial hierarchy, an updated version of the mulatto/quadroon/octoroon system for mixed-race assignment in slave societies in the Americas. By contrast, socioancestral thinking in effect acknowledges color-conscious social identities while conceptually countering the essentialism that lingers around racial identities.

Unlike the idea of race, socioancestry comes with no pretense of an objective, comprehensive classification scheme. There are no supposedly fixed necessary and sufficient criteria for membership in any socioancestral group. If pressure is put on the fact that the boundaries of continents are not always well-defined (e.g., between Europe and Asia)—indeed, continental boundaries are largely understood as a geographical convention—this is not a problem for the socioancestral account. Do people of the Southern Caucasus (of what are now the countries Armenia, Azerbaijan, and Georgia) have European ancestry, Asian ancestry, or both, since the Caucasus is regarded as a border region of Europe and Asia? For the socioancestral account, such gray-area cases raise an empty question: we know the region the people are from and the facts of their traceable ancestry. Generally, people native to the Southern Caucasus are categorized as European (after all, they are the root people, as it were, of the Caucasoid/white racial designation). The geographical gray area poses no puzzle for this categorization.

More broadly, the idea of socioancestry does not strain to assign everyone to a standard color-conscious group. There may be no globally or locally prevalent social conventions that cover people from certain regions (e.g., India, Iran, the Middle East, North Africa) or of certain mixes. A people might identify as a socioancestral group based on their ancestry and history as a certain color-conscious ‘other’ relative to ‘major’ racialized groups. Some peoples or individuals might not identify as members of any socioancestral group, since they might not have an affirmative color-conscious identity. These contingencies allow for meaningful categories of the socioancestrally unassigned, nonstandard, or undeclared—which would be roughly akin to the informal category ‘racially ambiguous’—and is hardly a problem for socioancestral as compared to racial thinking.

In sum, socioancestry is a type of social group formation—one that develops when persons accept (or are ascribed) a social identity because they share a component of continental ancestry that distinctively shapes color-conscious social reality—for which facts about continental ancestry alone are not sufficient and a (thick) shared culture is not necessary. Recognizing that Black Americans, for example, constitute a socioancestrally black subgroup is compatible with the observation that many Black Americans, like many members of other continentally identified groups, still think they belong to the same racial group, whatever they might mean by ‘race’. But the possibility of a Black (American) or a black socioancestral identity neither depends on nor presupposes the idea of race. We already can and do use color labels first and foremost to describe people on the basis of some component of their continental ancestry. Furthermore, I claim, we could abandon use of color labels as racial designations, without making a radical leap in color-conscious thought or practice. To reiterate, visible continental ancestry, not any notion of race, is the root of the social reality of color consciousness.

6. Enough Already

Human beings can be classified on the basis of features they share with some in contrast with others. When such classifications are made, the criteria for the sorting scheme—e.g., biological sex, eye color, dominant handedness, nationality—usually are transparent: we know what we are talking about and looking for. Classification practices become sketchy when a sorting scheme has a name that lends an impression of transparency even as its criteria remain conceptually obscure, esoteric, or unsound. This is the case with racial classification. Various criteria—most prominently, continental ancestry, visible continental ancestry, or genetic makeup—have been proposed as the basis for sorting human beings by race. The problem is that, no matter which criteria are adopted, the results seem bound to depart significantly from the prior understandings of racial group membership that would support adopting that very criterion.

If the criterion for race is strictly continental ancestry, the sheer number of persons of mixed continental ancestry, especially in the New World, confounds standard classification practices that have assigned most such persons and their descendants to monoracial groups. If visible continental ancestry is invoked instead, siblings with the same biological parents could be of different races, and biological parents of the same race could have a child of a different race—scenarios that run contrary to the standard logic of descent governing racial assignment. If the criterion is genetic makeup, the genetic markers will be tied to some visible or nonvisible component of continental ancestry—yielding similar anomalies due to a mismatch between the genetic markers and either appearance or the logic of descent. And a combination platter of criteria will not deliver a stable, nonrevisionist idea of race.

No matter which criterion is favored, we are left with a recipe for confusion and misunderstanding about the nature of human beings as classified by race. Thus, the race debate persists, as believers in some reality of race go forth like desert wanderers who, long convinced they have neared their destination, continue to walk in circles or, eventually exhausted, redraw the map and declare arrival. At this late stage in the race debate, believers will discover no new facts about human beings that could objectively resolve controversy in favor of a master concept of race. The real controversy—despite the pretensions of some current scientific and philosophical research programs—turns on which scheme of human categories might manage to seize title to the word ‘race’, not some prior facts that would establish the existence and the nature of race and the proper boundaries of racial group membership.

We’re long past time to stop wandering around ‘race’. The idea of socioancestry represents a move onward.