1. Introduction
A dominant view about races today is the so called “social constructionist” view. Social constructionists propose that the concept of race—i.e., the belief that a classification based on skin color and other skin-deep properties like body shape or hair style maps onto meaningful, important biological kinds—is a pseudo-biological concept that has been used to justify and rationalize the unequal treatment of groups of people by others.Footnote 1
Social constructionism became prevalent mainly because from the 1970s on, it has been widely recognized that the biological concept of subspecies, that is, of populations of conspecifics that are genetically and morphologically different from each other, could not be applied to humans. For one thing, it has been shown that there is more genetic variability within human racial groups than between them (Lewontin Reference Lewontin1972; Brown and Armelagos Reference Brown and Armelagos2001). Moreover, assigning an individual to a race does not buy the inferential power you are usually warranted to expect from a biological kind term.Footnote 2 Finally, classifications based on different phenotypic traits (skin color, body shape, hair, etc.) usually cross-cut each other (Brown and Armelagos Reference Brown and Armelagos2001). Thus, the racialist tenet that skin color and other skin-deep properties pick up different biological groups has been assumed to be false.
Biology has thus fuelled the recent racial skepticism of social constructionists, that is, the view that races do not exist.Footnote 3 But social constructionists about race are not mere skeptics. They usually underscore the instability and diversity of human beings’ concepts of races. For instance, Omi and Winant note that an “effort must be made to understand race as an unstable and ‘decentered’ complex of social meanings constantly being transformed by political struggle” (Reference Omi, Winant, Essed and Goldberg2002, 123; see also Root Reference Root2000). Others suggest that the notion is a modern invention, rooted in the eighteenth-century taxonomies of Linnaeus and Blumenbach. For them, there were times or places where people did not have any concept of race (Banton Reference Banton and Zubaida1970).
The constructionist contribution to the understanding of racialism is important (for a critical review, see Machery and Faucher Reference Machery, Faucher, Cohen and Lefebvre2005). It rightly suggests that human beings’ concepts of race do not occur in a social vacuum: social environments are important to explain the content of our concepts of race. It also correctly emphasizes the diversity of human beings’ concepts of race across cultures. Any account of racialism has to be consistent with these facts. However, it is not without difficulties either. First, it does not explain why many cultures have developed some concept of race and some classification based on phenotypic features. Moreover, the social constructionist approach does not explain the commonalities between the culture-specific concepts of race, e.g., the concepts of race in contemporary North America, in nineteenth-century France, in Germany during the Nazi era, and so on. Some aspects of the folk concepts of race vary little across cultures (Hirschfeld Reference Hirschfeld1996), while others vary much more. This should be explained.
In recent years, there has been a growing literature in evolutionary psychology and evolutionary anthropology about racialism. Although no consensus has yet emerged, several proposals have recently attempted to describe the underlying cognitive mechanisms responsible for the production of racial concepts (e.g., Hirschfeld Reference Hirschfeld1995, Reference Hirschfeld1996, Reference Hirschfeld1997, Reference Hirschfeld2001; Gil-White Reference Gil-White1999, Reference Gil-White2001a, Reference Gil-White2001b; Kurzban et al. Reference Kurzban, Tooby and Cosmides2001; Cosmides et al. Reference Cosmides, Tooby and Kurzban2003; Machery and Faucher Reference Machery, Faucher, Cohen and Lefebvre2005). Researchers agree that racialism has not been selected for: it is a byproduct of an evolved cognitive system, which was selected for another function. However, they disagree on the nature of this system.
The cognitive and evolutionary approach to racialism is a needed supplement to the social constructionist approach. The recurrence of racial classification across cultures and the commonalities between them suggest that racial classifications are the product of some universal psychological disposition. However, evolutionary theorists face a challenge that is symmetric to the challenge faced by social constructionists. Since they posit a species-typical cognitive system to explain racial categorization, they have a hard time explaining the cultural diversity of the concepts of race. It has to be shown that the claim that a species-specific human cognitive system underlies racialism is consistent with the evidence that racial concepts vary across cultures and times and are influenced by culture-specific beliefs.Footnote 4
Thus, we are confronted with two explanatory approaches to racial categorization that are symmetrically incomplete. This point has been recognized by several evolutionary-minded researchers. Indeed, they have paid lip service to the project of integrating the constructionist approach and the cognitive/evolutionary approach in the domain of race (e.g., Hirschfeld Reference Hirschfeld1996). However, in the domain of race, few have walked their talk.
In this paper, we propose that the theory of cultural evolution is the proper framework for integrating both approaches to racialism. In line with the social constructionists’ emphasis on the social environment, we claim that the concept of race—how race membership is thought of—is culturally transmitted: one acquires the concept of race from one's social environment. However, we insist that social learning is determined by several factors. Following Gil-White (Reference Gil-White1999, Reference Gil-White2001a, Reference Gil-White2001b), we emphasize particularly the importance of an evolved, canalized disposition to think about ethnies in a biological way. We argue that our proposal accounts for the similarities between culture-specific concepts of race as well as for their differences.
Our strategy is the following. In Section 2, we distinguish three kinds of groups, kin-based groups, small-scale coalitions, and ethnies. Following Gil-White (Reference Gil-White1999, Reference Gil-White2001a, Reference Gil-White2001b), we propose that ethnies have raised specific evolutionary challenges that were solved by an evolved cognitive system. The concept of race is shaped by this mechanism. We thereby meet the challenge faced by the social constructionist view: we account for the similarities between concepts of race. In Section 3, we build on Boyd and Richerson's theory of cultural evolution (Boyd and Richerson Reference Boyd and Richerson1985; Richerson and Boyd Reference Richerson and Boyd2004) in order to integrate social constructionists’ insights and cognitive/evolutionary theorists’ insights. We thereby meet the challenge faced by the cognitive/evolutionary approach: we account for the differences between concepts of race.
2. Ethnic Cognition and Racialism
2.1. The Ethnic Cognition Hypothesis
There is now a large body of evidence that small coalitions were not the only important social groups during human evolution (Bettinger Reference Bettinger1991, 203–205; Rodseth et al. Reference Rodseth, Wrangham, Harrigan and Smuts1991; Richerson and Boyd Reference Richerson, Boyd, Eibl-Eibesfeldt and Salter1998, Reference Richerson and Boyd1999, Reference Richerson, Boyd and Nesse2001; Richerson et al. Reference Richerson, Boyd, Henrich and Hammerstein2003, 369). Besides their kin-based groups and small coalitional groups, our ancestors also belonged to larger groups, often called ‘tribes’ or ‘ethnies’.
Ethnies are large groups—from 500 members to some thousands. They are divided into smaller units, sometimes called ‘bands’. An essential property of ethnies is that they are cultural units. The members of a given ethnie share many culturally transmitted beliefs, preferences, and norms, including norms of cooperation, and these beliefs, preferences, and norms tend to differ from those that prevail in other ethnies (Richerson and Boyd Reference Richerson, Boyd, Eibl-Eibesfeldt and Salter1998, Reference Richerson and Boyd1999). Finally, ethnies are characterized by a normative endogamy. The Nuer in Sudan and the Iroquois in North America illustrate this form of social organization. Ethnies are also specifically human. There are clear traces of ethnies in the archaeological record 50,000 years ago (Klein Reference Klein1999) and ethnies may have existed earlier (but see Knauft Reference Knauft1991, 392).Footnote 5
We follow Boyd, Richerson, and colleagues in hypothesizing that this form of social organization has created sui generis adaptive pressures. According to them, besides the cognitive mechanisms that evolved to deal with kin and small-scale coalitions, Mother Nature has endowed us with specific cognitive mechanisms whose function is to commit us to respect the norms of our own ethnie (particularly, the group-beneficial norms). More generally, it is plausible that this social organization put enough selective pressures on humans that we evolved a cognitive system dedicated to various dimensions of the ethnic life. With Gil-White, we would like to suggest that this is the key for understanding racialist cognition.
2.2. An Adaptive Scenario: Ethnic Cognition and the Exaptation of Human Folk Biology
Gil-White has suggested the following adaptive hypothesis (Gil-White Reference Gil-White1999, Reference Gil-White2001a, Reference Gil-White2001b).Footnote 6 Humans are disposed by evolutionary design to perceive ethnies as biological species. They apply their evolved folk biology to them. Our folk biology contains the innate knowledge about biological species and the reasoning heuristics that are generally applied to them (Atran Reference Atran1990; Medin and Atran Reference Medin and Atran1999). Essentialism, that is, the belief that categories are defined by essences, is supposed to be an important element of this system (Atran Reference Atran1990; Gil-White Reference Gil-White2001a; but see Machery and Faucher Reference Machery, Faucher, Cohen and Lefebvre2005). We are thus disposed to believe that ethnic membership is an essential property, which is transmitted at birth from one's parents and which determines people's behavior.
During human evolution, folk biology was applied to ethnies, because ethnies and species shared several important characteristics (Gil-White Reference Gil-White2001a, 518–519). Ethnies are characterized by clusters of stable, culturally transmitted behavioral norms, and different ethnies have often different norms. Thus, like conspecifics, coethnics behave similarly, and members of different ethnies behave differently. Besides, when members of two different ethnies interact, the interactions whose success requires shared behavioral norms often remain fruitless. Humans are sensitive to such costs. Hence, norm boundaries tended to coincide with many social interactions. This is particularly true of mating. Finally, ethnies are often distinguished by external markers (McElreath et al. Reference McElreath, Boyd and Richerson2001). Our ancestors tended to broadcast their ethnic membership and to pay attention to these signals (dress, body marks, etc.). Parents and children usually display the same markers. To summarize, ethnies share the following characteristics with species: coethnics have a distinctive morphology (dress etc.), coethnics behave in a characteristic way, ethnic membership is based on descent, and reproduction is endogamous.
Gil-White's hypothesis (Reference Gil-White2001a, 518, 530–532) is that our folk biology has been exapted to be applied to ethnies: that is, thinking biologically about ethnies was adaptive and was selected for. This is good epistemology (though certainly bad science), for it promotes inductive generalizations on the basis of limited contacts. Since members of a given ethnie tend to behave similarly because they share the same norms, such generalizations tend to be true. More important, a biological view of the ethnic world plausibly reduces the frequency of fruitless interactions across ethnic boundaries, particularly mating across ethnic boundaries.Footnote 7 It may underlie a preference for interactions with coethnics and a reluctance to interact with members of other ethnies.
2.3. Racialism: The Misfiring of Our Ethnic Cognitive System
According to this hypothesis, races trigger by mistake our folk biology-based ethnic cognitive system. That is, people tend to assume erroneously that humans with a given skin color or a given phenotype form an ethnie. The reason is that the physical properties that define race membership are similar to ethnic markers. And, like ethnic markers, they are shared by parents and children. Thus, skin color, body type, and other properties are taken to be ethnic markers.Footnote 8
2.4. The Ethnic Cognitive System
The evolutionary importance of ethnies suggests that Mother Nature has predisposed us to pay attention to people's ethnic membership (Gil-White Reference Gil-White2001a; McElreath et al. Reference McElreath, Boyd and Richerson2001). Hence, ethnic or racial membership should be a primitive of our encoding of people's characteristics (with gender and age, for example). This is supported by the literature about race categorization. Psychologists generally agree that race is automatically encoded by adults (but see Kurzban et al. Reference Kurzban, Tooby and Cosmides2001). Notice that this does not imply that humans are always paying attention to ethnic or racial membership to the same degree. Ethnicity's or race's salience may vary from context to context.
Given that ethnic information may be important early in life, for example to determine which individuals to imitate during childhood or youth, ethnic and racial encoding should be active quite early in life. Hirschfeld (Reference Hirschfeld1996) shows indeed that young children spontaneously encode race information (particularly, when it is presented verbally).
In order to determine people's ethnic membership, one has to pay attention to the physical cues that signal it, i.e., to ethnic markers. Thus, we should be disposed to pay attention to ethnic markers. Gil-White (Reference Gil-White2001a, 548–549) also suggests that children are predisposed to pay attention to specific types of ethnic markers, like clothes or body marks.
We should also be endowed by design with a domain-specific mechanism to learn concepts of ethnies, that is, beliefs about ethnic markers, behaviors, etc., of members of specific ethnies. Clearly, we are not predisposed to entertain any specific ethnic concept, say NUER. But we may be predisposed to learn ethnic concepts in a specific way. Let's call the hypothetical domain-specific mechanism through which we learn concepts of ethnies the “ethnic concepts acquisition device” (ECAD).
Gil-White's evolutionary argument suggests that the ECAD is based on our folk biology. Now, concepts that are formed by a domain-specific cognitive system have a default content. For, when these concepts are formed, the domain-specific system fills them in with default beliefs. This idea has been applied to religious concepts by Pascal Boyer. No culture has developed a concept of a god that exists only on Sunday (Boyer Reference Boyer2001). Boyer suggests that the cognitive systems that create religious concepts provide the default assumption that, like any other individual, gods exist continuously. The same is plausibly true of our ethnic and racial concepts. That is, when our ethnic cognitive system forms an ethnic or a racial concept, it fills it in with some default assumptions. Hence, by default, ethnic and racial concepts should be similar to animal species concepts. This idea is supported by Hirschfeld's developmental studies and by Gil-White's cross-cultural data. From an early age on, and in several cultures, children reason about ethnies and races in a biological manner. That is, in several respects, they reason about ethnies and races as if they were species. For instance, they believe that some racial properties are transmitted at birth and constant over life. We call this disposition `biologism’.
Hirschfeld's studies cast some light on the ECAD (Hirschfeld Reference Hirschfeld1996). Particularly, they show that this system is on very early. Moreover, its inputs are not necessarily, and maybe not primarily, visual. Linguistic inputs, for example names, say ‘Nuer’, may be sufficient to acquire an ethnic concept. McElreath et al.'s and Gil-White's arguments about the importance of ethnic markers suggest that visual cues, for example bodily and behavioral characteristics, are likely to trigger the ECAD as well. Other perceptual cues, for example auditory cues like a foreign language or a specific accent, may also be important.
3. Culture and Evolved Cognition: Toward an Integrated Account of Racialism
3.1. Cultural Transmission
Modern theories of cultural transmission provide the proper framework for integrating the two main traditions in the study of racialism (Boyd and Richerson Reference Boyd and Richerson1985; Richerson and Boyd Reference Richerson and Boyd2004; Sperber Reference Sperber1996). The core idea is that many beliefs, preferences, reasoning patterns are socially learned: like in traditional social learning theory, they are acquired from one's social environment—from one's cultural parents. However, Boyd, Richerson, and their colleagues emphasize that several forces determine which information is culturally transmitted (Boyd and Richerson Reference Boyd and Richerson1985; Richerson and Boyd Reference Richerson and Boyd2004; Henrich and McElreath Reference Henrich and McElreath2003). In other words, cultural transmission is shaped by several biases. Two kinds of biases can be distinguished, the content biases and the context biases (Henrich and McElreath Reference Henrich and McElreath2003). The context biases favor the acquisition of beliefs, concepts, etc., from specific cultural parents. For example, in some situations, cultural transmission is conformist: people tend to acquire the beliefs, etc., that are held by most of their cultural parents (Boyd and Richerson Reference Boyd and Richerson1985; Henrich and Boyd Reference Henrich and Boyd1998). In other situations, cultural transmission is prestige-dependent (Henrich and Gil-White Reference Henrich and Gil-White2001): people acquire the cultural variants that are held by prestigious individuals. The content biases correspond to the psychological systems that favor the transmission of specific beliefs, etc., instead of others (“attractor” in Sperber's terminology, “cognitive track” in Boyer's). Beliefs, etc., that fit with these psychological systems are easily memorized and easily applied by cultural learners; those that do not fit with them tend to be forgotten.
3.2. How Children Learn Racial Concepts?
We propose that concepts of race are culturally transmitted. This is in line with social constructionists’ reliance on traditional theories of social learning, that is, with the idea that the concept of race is acquired from one's social environment. This explains why within a culture, at a time, people tend to have the same concept of race. This also explains why different cultures at different times have endorsed different concepts. We add to social learning theory the idea that the two context biases mentioned above, namely, conformism and prestige-dependent imitation, affect the transmission of the concept of race. Thus, the acquisition of the concept of race by a cultural learner depends on whether successful individuals or most cultural parents classify people into races.
However, whereas social learning theory suggests that the mind has no disposition to think about races in a particular way, we propose that human ethnic cognition creates a cognitive track for the cultural transmission of racial concepts. That is, it favors the acquisition of concepts of race that are consistent with the default assumptions provided by our folk biology. Concepts of race that are inconsistent with these assumptions are less easily memorized. Thus, we propose that by default, humans tend to think biologically about groups of individuals that share superficial properties like skin color or body shape.
This provides a framework for integrating most theses and pieces of evidence of the social constructionist approach and of the evolutionary/cognitive approach to racialism. Importantly, this is a mere framework, not a psychological hypothesis. Within this framework, several detailed hypotheses can be formulated.
For instance, one could propose that children spontaneously classify people into races (Hirschfeld Reference Hirschfeld1996). Skin color (or other physical properties) or race names could trigger the ethnic concepts acquisition device. As a result, concepts of specific races would be created and would refer to the classes of individuals that have these physical properties or are referred to by these names. These classes would be assimilated to ethnies and default beliefs, based on children's folk biology, would be assumed to be true of them. This hypothesis fits well with the recurrence of racialism across cultures.
Instead, one could argue that children do not classify spontaneously people into races. They are primed to draw racial distinctions when their cultural parents use racial classifications. The ECAD would then influence the way children think about races, selecting for biological concepts of race. Children's concept of race and their concepts of specific races would also be influenced by the way cultural parents think of races.
The racial distinctions that are made by children can be consistent or inconsistent with the distinctions that are made by cultural parents. Detailed empirical hypotheses have to specify what happens when the ECAD, conformism, and prestige-dependency pulls in different directions. For instance, if children spontaneously classify people into races, one wants to know what happens when their classification is inconsistent with the classification that is made by their cultural parents. Detailed empirical hypotheses and new empirical evidence, maybe based on longitudinal studies, are needed.
3.3. Solving the Integration Challenge
The framework presented above integrates the social constructionist approach and the evolutionary/cognitive approach. The concept of race is socially learned, as social constructionists would have it. However, our evolved ethnic cognition creates a psychological bias in favor of biological concepts of race.
Moreover, this framework explains the most striking aspects of racialism. Since the cultural transmission of the concept of race is assumed to be primed by the ethnic cognitive system, the cross-cultural recurrence of racialism is to be expected. Moreover, the fact that races tend to be thought of biologically is thereby explained.
Social constructionists have rightly emphasized those aspects of racialism that vary across cultures and times. The framework proposed here predicts which aspects vary across cultures and how they vary. Whereas biologism should tend to be cross-cultural, aspects of our concepts of race that are not based on the evolved components of our folk biology should vary across cultures.
Conformism and prestige-dependent imitation are important to explain cultural variation. Concepts of race that are held by prestigious cultural parents or by most parents should be easily culturally transmitted. More historical work is needed to find out whether some conceptualization of race membership has spread within a population because of the influence of some prestigious individuals.
Finally, we propose that acquired, culture-specific content biases favor some concepts of race over others. Not all content biases are innate. Some are culturally acquired. Concepts of race that are consistent with these culture-specific folk theories may be transmitted more easily than other concepts of race within the corresponding cultures. Since racial concepts are filled in with default values derived from folk biology, the concept of race within a given culture should be strongly influenced by the culture specific aspects of people's folk biology in that culture. Evidence suggests that the concept of race is influenced by culture-specific theories of contamination and by the culture-specific aspects of folk biology (Hirschfeld Reference Hirschfeld1996, Chapter 2). More historical and ethnographic work is needed to determine whether across cultures, racialism tends to rely on culture-specific folk theories.
4. Conclusion
Social constructionists and evolutionary-minded social scientists avoid interacting with each others. This is detrimental and unjustified. For, cognition is shaped by culture and cultural transmission is an evolved aspect of our mind.
Racialism, that is, the belief that groups of human beings made on the basis of skin color (etc.) map onto biological kinds, illustrates this point. Any good theory of racialism has to take into account the main points of the social constructionist approach, including the fact that people's concepts of race vary across cultures. However, without a cognitive cum evolutionary background, the social constructionist approach is incomplete.
Our position aims at accounting for the similarities and for the differences between culture-specific concepts of race. Instead of simply illustrating this diversity, as constructionists do, and instead of neglecting it, as many evolutionary/cognitive scientists do, we try to explain it. Concepts of race are culturally transmitted. The cultural transmission of these concepts is shaped by several biases. It is biased by conformism and prestige-dependent imitation. These two biases are supplemented by an evolved ethnic cognitive system that is misapplied to races. This system, the ethnic concepts acquisition device, results from the exaptation of our folk biology. Together, these biases determine whether skin color and other superficial properties are treated as ethnic markers.
The study of the interaction between culture and our evolved cognition is still in its infancy. We are conscious that our approach is just a small step. But, we believe that the theory of cultural transmission is currently our best hope for unifying the social sciences.
