Human Kinds in the Making: Race and the Mind

The Sciences
March/​April 1997

Review-Essay
By David Berreby


The old joke puts the Lone Ranger and Tonto side by side, out of ammunition, surrounded by thousands of angry Indian warriors.

"Looks like we're done for, Tonto."

"What mean 'we,' white man?"

What mean we, indeed? Almost everyone carries a set of identities like a pack of cards, playing each one as appropriate-Methodist, parent, Democrat, Lakers fan, member of the art department. But in the past couple of decades, more and more of the cards have been declared to be trumps, which complicates the game. Traits considered deviant or even despicable, and enthusiasms regarded as silly or quaint, have been promoted to what the sociologically minded call master categories-valuable essences that define their lucky owners and can be understood only by them.

Most of us still condescend to people who take a Saturday afternoon identity, such as sports fan or Trekkie, as if it were serious. And the category error of leaning too hard on a supposedly frivolous marker is a staple of comedy, as in Seinfeld's "soup Nazi" or The New Yorker cartoon of the husband kissing another woman who tells his surprised wife, "You wouldn't understand. It's a drunk thing." But the general rule of identity politics nowadays is: If you say so. . . . Barbara Adams, an alternate juror at the Whitewater trial, was thrown off the jury for violating a gag order, not for wearing her Star Trek uniform to court. Different lives teach different lessons, and it seems presumptuous to doubt another adult's testimony about herself. (A Trekkie-or Trekker, to be polite about it-once told me she had kicked her cocaine habit by following the code of the Klingon Warrior, not by submitting to any of that wishy-washy Twelve Step stuff.)

After the civil rights movement, after feminism and gay rights and disability rights, Americans have a standardized playbook for creating a political-cultural identity. You start with the conviction that being a member of your group is a distinct experience, separating you from people who are not in it (even close friends and relatives) and uniting you with other members of the group (even if you have never met them). Sisterhood is powerful. Second, you assume that your own personal struggles and humiliations and triumphs in wrestling with your trait are a version of the struggles of the group in society. The personal is political. Third, you maintain your group has interests that are being neglected or acted against, and so it must take action-changing how the group is seen by those outside it, for instance. The once insulting "black" becomes respectable, then de rigueur. The supposedly female concern for feelings and relationships gets redefined as a different kind of moral consciousness. "Queer" becomes a term of pride. Once denigrated facets of the group-soul food for African-Americans, quilting circles for women, sign language for the deaf-become emblems to be celebrated. Historical forebears, brave Virginia Woolf and gay Shakespeare, for instance, are claimed and feted.

Of course, noticing that the promotion of group consciousness proceeds by analogy is not to argue that the process of new identity-formation is bad, even though it may divide communities, polarize families and challenge traditional values. That the old ways of the South were turned upside down by the civil rights movement was not a bad thing. All new affiliations disrupt old ties. "If any man come to me, and hate not his father, and mother, and wife, and children, and brethren, and sisters, yea, and his own life also," Jesus said, "he cannot be my disciple."

But as more and more statuses have taken on the trappings of a culture, a race, a gender, the game has become more complicated. When identity is a pack, no one can play one card all the time. And as more and more categories have their advocates claiming master status, there is more chance for conflict-not only between members of a group and the people outside it, but also within each group and even within the heart of each member.

The identity explosion makes all the more urgent questions that should be accessible to science: What does it mean to have a sense of Us as opposed to Them? What parts of that mentality are universal to the mind, and what parts are the products of particular cultures and histories? If, as seems obvious, some aspects of affiliation with one's kin and allies are innate, where is the border between that natural endowment and the workings of culture?

What psychologists and philosophers call naive theories-the common sense of how the world works that people seem to carry by default unless trained out of it-have been much studied in both children and adults. So has the working of the brain in various cognitive and emotional states. Perhaps, then, the time has come for a science of groups and group hatred. "There is no comprehensive field of study devoted to the most serious of human conditions: hate," Kenneth S. Stern of the American Jewish Committee has written. "There is no college from which you can graduate today with a degree in hate. But there should be."

The Concept of Race


Lawrence A. Hirschfeld, a psychological anthropologist at the University of Michigan in Ann Arbor, takes a step toward that direction in this startling, subtle and challenging book. He rejects the idea that all categories of Us and Them are alike and equal. Instead, he says, his studies of children's thinking show that children have a pervasive tendency to think racially-that is, to see race as natural and important in predicting what people will be like and how they will behave. The argument suggests that Hirschfeld would not have been surprised about a fact that, according to The New Yorker writer Jeffrey Toobin, confounded the prosecutors in the O. J. Simpson case: black women, responding to reports of a black man abusing his white wife, felt their sympathies drawn not to a fellow woman but to a fellow black.

Many studies have documented the human habit of categorizing people into Us and Them. Strikingly, the habit does not seem to be attached to any particular distinction. As Hirschfeld reports, studies have found that people will develop a preference for their own group even if the group is unheard of and arbitrary: one study found the bias even among people who had been told their group was formed on the basis of the last digit of their social security numbers.

But race, Hirschfeld says, is different. It may not be built into our brains, but what is built in is a way of sorting people into categories that makes race a particularly easy category to use. "We (and our children) mentally represent racial groups with remarkable facility, easily invest them with significance, and readily communicate our ideas about them. . . . Class, for example, is as potent an explanatory concept. Many would argue that it is more explanatorily potent. Yet class is not easy to think."

"The concept of race," Hirschfeld goes on, develops out of a psychological propensity to learn about a specific phenomenon: the world of human kinds. Most important, I suggest that this propensity is best viewed as a special-purpose cognitive facility. Our notion of race is unlike our notion of other things. . . . Racial thinking represents a specific sort of thought more than a specific sort of shared experience.

Such a counterintuitive and politically unpalatable position (Hirschfeld calls it "disquieting") is all the more extraordinary because, as he notes, race is a changeable fiction. The naive "biological theory," as he puts it-that people will act, by virtue of their race, in this way and not in that-has no basis in reality. The naive theory regards race as immutable, heritable, essential. What is the reality? Item: People seem better at distinguishing individuals of their own race, but that ability has to be learned; before adolescence, there is no difference by race in accuracy on recognition tests. Item: In the early twentieth century most Americans denied that Italians were white. Item: Preschool children in at least one survey seemed to think racial features were temporary and caused by tanning or changes of clothes.

That people come in varying strains is of course undeniable. But there is a sharp difference between real biology, in which different genes or phenotypic traits might be traced through populations, and folk biology, which invariably draws borders between populations that do not exist in the real world. In other words, not only are people's ideas of race flexible and arbitrary, but also they do not correspond to such divisions among humanity as actually exist. In his book Human Biodiversity: Genes, Race and History, Jonathan Marks, an anthropologist and geneticist at Yale University, notes: "There can thus be no genetic test to perform in order to determine whether or not one is 'Caucasian,' 'Alpine,' or 'Hopi.' The reason is simple: populations are constantly in genetic contact with one another."

In sum, according to Hirschfeld, race seems to be an idea that comes to people easily and shapes how we order our social lives, even though it has no basis in the real biology of the species. Most of what people believe about race is demonstrably false; but to say so is to say not much. People's beliefs have enormous practical consequences. You might as well tell people that romantic love is an illusion of the hormones and a cultural inheritance from the troubadours. The fiction of race is a fact; it is just not the fact we think it is.

Hirschfeld's approach usefully highlights that point by noting that the scientific falseness of race is the bare beginning of an argument, not the end-that the important matter is not the content of beliefs about race but their role in structuring psychic and communal life. Race is not only "a bad idea," he writes, but also a deeply rooted bad idea. [Thus] race may be as firmly grounded in our minds as it is in the politics of our day. Many people, perhaps understandably, prefer to believe that this is not the case.

But racial thinking, Hirschfeld argues, arises in children as young as three, and they come to the conclusion that race is an essential aspect of a person's identity not because "you've got to be taught," but on their own.

The evidence for these assertions comes from a series of experiments Hirschfeld and his coworkers have done in the past fifteen years. In one experiment, for instance, he presented sets of color-wash line drawings to 109 children from Ann Arbor, Michigan, in groups of three-, four- and seven-year-olds. Each set included drawings of an adult and two children of the same gender. The adults in the pictures varied according to three kinds of obvious, outward cues: body build (thin versus stocky); occupation (wearing tools, a stethoscope or a police uniform) and race (black versus white). Thus a child might be shown a picture of a thin black policewoman along with two pictures of children: a stocky white girl wearing a play police uniform, and a stocky black girl wearing ordinary children's clothing.

Each child subject was then asked: Which child in the drawing is the adult as a child? Which is the adult's child? Which one is most like the adult? Hirschfeld and his colleagues reasoned that if the children simply relied on outward appearances for making their judgments, they should be as likely to rely on one form of outward appearance as on another. Accordingly, they should choose at random.

They did not. Asked which picture of a child depicted the adult as a child, the children showed a distinct tendency to match racial types instead of body build or occupational types. The children showed that same tendency when asked to choose which child was the adult's child. The third question strongly suggested that the choices were not based on similarity alone. When asked which child was most "similar" to the adult (a question that was presumably interpreted to mean which child bore a resemblance to the adult but no intrinsic connection), the subject children were almost as likely to use occupation or body build to make their judgments as they were to use race.

Why would such findings convince anyone that race is a particularly salient category? Why would anyone think that body build or occupational dress is anywhere nearly as salient to children as, say, the color of a person's skin? As a matter of fact, Hirschfeld notes, earlier studies have suggested those categories are quite important to young children. Both body build and occupational type are named categories, marked by conspicuous visual cues and known to be associated with stereotypes that develop in early childhood. Psychologists have shown that knowledge of all three kinds of category develops at about the same age. And body build (as distinguished from weight) is, objectively speaking, a particularly stable feature of people as they grow from child to adult.

Classifying Race


The investigators were also able to show that the children's tendency to classify by race was not simply done by skin color alone-which, as Hirschfeld points out, is "a real possibility in view of the salience color has for young children's sortings." To rule out that possibility, Hirschfeld and his colleagues asked the children to choose which of two pictures of cars "belonged to" the adults. Thus a picture of, say, a stocky black woman might be shown to the children along with two pictures of cars: a sleek black car and a "fat," bulbous-looking white one. The children showed no color bias at all in making those choices.

A second experiment shows how even very young children associate "race" with certain individual characteristics that go beyond family relations, such as native language. Hirschfeld and his coworkers presented thirty-six three-, four- and six-year-old children with pairs of colored drawings accompanied by speech samples. Each pair depicted two adults who contrasted either in race (black versus white), clothing (Western versus non-Western ethnic apparel), style of dwelling (Western versus non-Western) or orientation (forward-facing or backward-facing). Each speech sample was a short sentence spoken by an adult in either English or Portuguese. The experimenters simply asked the children which of the two people pictured was talking.

All three groups of children predominantly chose white or Western images for the speech samples in English. To confirm the result, Hirschfeld's group repeated the experiment by showing eleven three-year-olds pictures of people whose only obvious difference was that some were much older than others. Here the children's choices were random. That young children would associate non-English speech with nonwhiteness is striking, since, as Hirschfeld writes, the vast majority of blacks encountered by midwestern preschoolers in the United States are native English speakers-and, conversely, there is no reason to expect that the percentage of non-native English speakers is higher among U.S. blacks than among U.S. whites. This observation suggests that children construct these beliefs to some extent on their own, in spite of the lack of any empirical evidence to support them.

The findings, Hirschfeld writes, contradict the general view in psychology that children's racial thinking is a matter of applying "general learning mechanisms" to superficial differences

Hirschfeld proposes a rather elaborate model of the mind to account for his findings-a model that nonetheless lies in the mainstream of what many call evolutionary psychology or behavioral ecology. According to that model, the mind is not a general-purpose problem-solving device, free to induce answers with raw computing power. Instead, he writes, it is "an assembly of domain-specific devices specialized to handle specific types of information." Each device, or module, evolved because it helped our evolutionary forebears to overcome some crucial challenge. "A modular mind," writes the evolutionary biologist David Sloan Wilson of the State University of New York at Binghamton, is required because a single general purpose cognitive structure could not efficiently (or even conceivably) carry out such diverse tasks as language acquisition, face recognition, adaptive social exchange behavior and so on.

As Hirschfeld notes, the most convincing example of such a module is the "language acquisition device," first proposed forty years ago by the linguist Noam Chomsky of the Massachusetts Institute of Technology. Chomsky's LAD hovers over all the arguments for "domain-specific devices." According to him, children acquire language early in life because they bring to the task innate, ready-made conceptual tools, such as the concepts of verb and noun, without which they would never be able to sort out the complexities of syntax. Nowadays Chomsky compares the LAD to a switchbox, present in every child's brain by an early age, whose switches are primed to take on certain settings, depending on the language into which the child is brought up. The child's linguistic environment moves the various switches to the appropriate language-Chinese here, English there-but the wiring of the switchbox is much the same in every brain.

Chomsky's switchbox is not often mentioned in the literature of evolutionary psychology. Yet it is hard not to invoke his model when a mental module is being proposed. After all, as some evolutionary psychologists will concede, fundamental questions about mental modules are still being debated: Which ones exist? How do they interact? How much does a module vary from person to person? And the switchbox metaphor focuses questions that ought to be asked of any modular theory: What exactly is the preset architecture of the mind that is being proposed? What is left open, later to be determined by a person's environment?

Those questions need to be asked of Hirschfeld, because in Race in the Making he boldly proposes to link two fields that seldom communicate-cognitive studies and anthropology-by means of a Chomsky-like theory. Human beings, Hirschfeld argues, possess a race module: an innate and universal propensity for noticing racial differences and ascribing importance to them. It is a large assertion and, he acknowledges, an uncomfortable one. Its fuzzy edges are easy to attack. But it is also a fruitful claim, because even if the model is flawed, its basic shape and premise-a mental structure whose innate tendencies interact with social and cultural surroundings-may help to enlighten the deeply entangled issues of racial consciousness.

Current orthodoxy suggests that children develop concepts of race by noticing differences among people and by reasoning from analogy: there are different kinds of animals, so there must be different kinds of people. How, Hirschfeld asks, could thinking as stable and consistent as children's beliefs about race arise from so many different kinds of circumstances and sources of information? Race, after all, is a peculiar belief. It is certainly not held, for instance, exclusively on the basis of appearance. That people may look "white" but be "black" was a commonplace of American racial thinking, leading to the self-lacerating drama of passing. And which people are assigned to which races for political and cultural purposes is a matter that changes with time. My aunt, in her seventies, lives in an ethnic map I scarcely recognize, in which the Germans are distinct from the Irish, and the Irish are distinct from the Scandinavians. To me, born in the late 1950s, all those people are just "white Americans."

But racial thinking is too consistent, Hirschfeld maintains, to be derived from so many different environmental cues-unless it has a component that is innate. The argument is a racial version of the "poverty of stimulus" argument from Chomsky's linguistics: Children in any culture all end up speaking the same language, even though they hear speakers of varying quality and the syntactic examples they do encounter are much too limited for them to infer what they come to master about syntax. Hence, Chomsky concludes, syntax could not come from the environment. Hirschfeld draws the analogous conclusion:

Direct learning has less to do with the way racial thinking develops than is often imagined. Substantial aspects of children's racial cognitions do not appear to be derived from adult culture.

The faults of a book this brief and this bold are not hard to find. Hirschfeld's is a weighty theoretical edifice to build with the scaffolding of a few experiments. That is doubly the case for a group of experiments conducted, as Hirschfeld readily concedes, only on children in the U.S. and earlier (in a somewhat different form) in France-in other words, on children of Western, industrialized cultures that have much history in common. Like most other "modules" and "content-specific domains," the borders of Hirschfeld's module are not clearly elucidated. But lest you be tempted to think there is only an innate Us-versus-Them reasoning capacity, which can be molded to be racial here or gender-based there, Hirschfeld explicitly rejects the idea. He insists there is more innate specificity to racial thinking than that. Yet he also explicitly refuses to concede that thinking about race is itself innate.

So is the mechanism a general sorter for human kinds or is it specifically focused on race? If the innate module is a built-in capacity to theorize about what kinds of people live in the world, as Hirschfeld writes, then it is certainly conceivable that a child could grow up using the switchbox part of the module with nonracial settings-organizing the neighboring tribes by sexual preference, or gender, or occupation. There is nothing in Hirschfeld's studies of young American children to contradict the idea that his mechanism for racial thinking could serve other kinds of group distinctions. American children construct race-based identities with little prompting. It remains to be seen whether Ghanaian children, or Japanese children, or South African children, would do the same.

A more profound doubt, which Hirschfeld tries briefly to address, has to do with the theory of the modular mind. Clearly not everything human beings can do has evolved over hundreds of thousands of years. Many people drive cars, but that does not mean, as the linguist Philip Lieberman of Brown University notes, that one ought to posit a universal driving capacity, innate to each person, to explain driving. Moreover, as Lieberman has written, one hallmark of a trait that is subject to natural selection is that it must be heritable. Heritability implies some variation from person to person in the way the kind-distinguishing module operates-just as some people are double-jointed, and just as hair color and skin tone vary within even a single family. What variations might arise in the "kind-sorting" module? If such variations exist, what is left of the idea that a kind-sorting module is innate to all children?

On those questions, and others, the model is silent. Perhaps that deficit arises because the model needs further development. Or perhaps no model will ever account for the mysterious flexibility of racial thinking: the way a person can begin life "white" and end it "black"; the way a person can hate all blacks but not good old Dan, who is black; the way distinctions of merely academic interest (he is a Serb, his wife is a Croat) can be turned into matters of life or death. But Hirschfeld's book has the great merit of suggesting a beginning for a real discipline that might replace 5,000 years of futile platitudes about why it is that We hate Them.