Authors: Paul Bloom
The psychologist Katherine Kinzler and her colleagues have looked at the consequences of linguistic preference for how babies come to navigate their social worlds.
In one experiment, they tested ten-month-olds in Boston and Paris. The babies listened to both an English speaker and a French speaker, and then each speaker held out a toy. The Bostonians tended to reach for the one offered by the English speaker; the Parisians went for the one offered by the French speaker. Other studies found that
twelve-month-olds would rather take food from a stranger who speaks their language than from one who speaks a different language;
two-year-olds prefer to give a gift to a speaker of their language; and
five-year-olds prefer a child who speaks their own language as a friend.
Such choices make sense. It is easier, after all, to be friends with someone who speaks the same language, and, other things being equal, someone who speaks the same language is more likely to share one’s preferences for toys and food. What’s more interesting, though, is that we see the same effect with
accents.
Babies prefer to look at a speaker without an accent, even if the speaker with an accent is perfectly comprehensible.
When choosing friends, five-year-olds are more likely to choose children who speak American English as opposed to French-accented English, and when learning about the function of a new object,
four- and five-year-olds trust a native speaker more than an accented speaker. This suggests that children’s preferences are driven by some degree of cultural identification,
conveyed via language, just as predicted by the coalitional theory.
A
S YOU
can imagine, there is plenty of
research into the development of racial bias in children. The first experimental procedure was developed in the 1930s. An adult showed children pairs of dolls—a white doll and a black or brown doll—and asked questions like “Who would you like to play with?,” “Which looks bad?,” and “Which has the nice color?” In the 1970s, an expanded version was developed. Researchers showed children a picture of a white boy and a black boy and tested them with questions like “Here are two boys. One of them is a kind boy. Once he saw a kitten fall into a lake and he picked up the kitten to save it from drowning. Which is the kind boy?”
It is perhaps not so surprising that white children were drawn to the white child for the good things and the black child for the bad things. But what was shocking to many was that the first studies, done by the psychologists Kenneth and Mamie Clark, found that black children also tended to favor the white child. This study, which was cited in the
Brown v. Board of Education
decision that ended school segregation in the United States, might well be the most important developmental psychology finding in American history.
These studies have their critics.
The psychologist Frances Aboud points out that there is something absurd about the demands being put on participants. Children are forced to choose, and there is only one dimension of difference—race.
The only options are to favor one’s own group (which would be racist) or favor the other group (which would be racist in a different sense, as well as perverse). Children have no opportunity to opt out and say race doesn’t matter.
But better-designed experimental methods confirm that racial biases are established by the age of six. Consider the research of the psychologists Heidi McGlothlin and Melanie Killen, who presented white children between the ages of six and nine with images of ambiguous situations, such as a picture of a child in a playground sitting in front of a swing with an expression of pain, with another child standing next to her. Sometimes the standing child was black and the sitting child was white; sometimes the races were reversed. Other scenes involved acts that could be interpreted as cheating and stealing. Children were asked to describe the scenes and to answer questions about them. In these studies, unlike the earlier ones, they were not forced to take race into account. But they did: white children were more likely to describe these ambiguous situations as corresponding to bad acts when the white child could be seen as the victim and the black child could be seen as the perpetrator. Importantly, though, this held only for children in all-white schools. White children in racially heterogeneous schools weren’t influenced by the race of the characters.
Other studies find that children often favor peers of the same race and think that they are better people—but again this holds mostly in racially homogeneous schools.
When the studies are run in heterogeneous schools, children don’t care about race. Such results provide some support for what
social psychologists call
the “contact hypothesis”—the notion that, under the right circumstances, social contact diminishes prejudice. Apparently, the mixed-race schools provide the right circumstances.
What about younger children?
Studies with three-year-olds find that when they get to choose whom to accept an object from or engage in an activity with, gender matters—boys tend to choose the male, girls the female—and age matters, with children tending to choose a child over an adult. And as we’ve just discussed, language matters as well: children tend to choose individuals who speak the same language and don’t have a foreign accent.
But race doesn’t matter for the three-year-olds: white children don’t choose whites over blacks, for example. Only later do racial biases start to creep in, and only for children raised in certain environments. We might have natural biases to favor some groups over others, but apparently we are not natural-born racists.
And even for the older children who do take race into account, it’s not as important as language. For instance, when white five-year-olds were asked to choose between a white child and a black child as a friend, they tended to prefer the white child. But when asked to choose between a white child with an accent and a black child without one, they chose the black child.
N
EITHER
race nor language is necessary to sort people into coalitions. There is a large body of research showing that it
takes very little to make a coalition that really matters: to establish group loyalty, to pit people against one another.
The most famous studies here were developed independently by two European social psychologists. Muzafer Sherif was born in 1906 in Turkey and was as a young man nearly murdered by the Greek army; he later spent time in prison in the 1940s for his opposition to the Nazis. Henri Tajfel, born in 1919 in Poland, was a Jew who fought with the French against the Nazis and spent five years as a prisoner of war. To put it mildly, then, both men had personal experience with coalitions.
Sherif and Tajfel were both interested in what it takes to form an Us that clashes with a Them. Now, one possible way of exploring this would have been to look at real-world conflicts, but these reflect long and complex histories—an Israeli might hold many legitimate grievances against a Palestinian and vice versa—and Sherif and Tajfel wanted to determine the
minimum
that it takes to divide people. Instead of examining conflicts with long historical records, then, they each conducted experiments designed to create social divisions where none had previously existed.
In 1954, Sherif invited twenty-two fifth graders—white middle-class boys from what he described as “established families”—to attend a summer camp at Robbers Cave State Park in Oklahoma. The boys were split into two groups, each housed in a separate cabin; neither group knew of the other’s existence. During the first week of the camp, each group explored the area, played games, and had an
all-around good time. And each group named itself: “the Rattlers” and “the Eagles.”
Then the experimenters set up first contact. Sherif, who posed as the camp janitor to observe the interactions, noted that one of the boys, upon hearing but not seeing the other group, called them “the nigger campers.” The researchers arranged tournaments between the groups, and relations slowly went from cautious animosity to something quite a bit worse. These small societies began to emphasize their distinct customs: the Rattlers would cuss; the Eagles would take pride in their clean language. They made flags. They objected to eating together in the mess hall. They continued to use racial epithets, though everyone was white—it seems as if these terms were used as all-purpose expressions for “others.” In written tests, the boys from each group said that members of their own tribe were stronger and faster than their foes.
After the Rattlers won a few competitions, the Eagles stole their flag, set it on fire, and put up its charred remains. The Rattlers retaliated and destroyed their rivals’ cabin while the Eagles were at dinner. The Eagles won a tournament and the Rattlers stole their prized trophy—knives that had been given to them by the psychologists.
Sherif then moved to the next phase of the experiment, which was figuring out how to bring the groups together—in other words, seeking world peace in a test tube. Many attempts, such as shared meals and shared movies, failed, but the researchers finally succeeded by introducing a problem that threatened both groups’ existence:
a mysteriously cut water pipe. The factions were brought together by a common cause, perhaps a common enemy.
The Robbers Cave experiment demonstrated that you could create warring communities in a couple of weeks. Still, the situation did encourage individuals to identify with their group: not only did the psychologists facilitate competition between the groups, but each boy spent a week with his own group before he even knew about the other group, and it does seem reasonable for a boy to trust his friends more than strangers. Could coalitions emerge without all of this social support?
This was Tajfel’s question. He designed a simple experiment in which he asked adults to rank a series of abstract paintings. He then randomly told half of them that they had shown a preference for the works of Paul Klee and told the other half that they preferred the works of Wassily Kandinsky. This was enough to make people feel a sense of group membership. When later asked to distribute money to other Klee lovers and other Kandinsky lovers, participants would give more to the group that they belonged to—even if they themselves didn’t profit from doing so.
These findings have been replicated many times; some studies find that you can divide people using the Platonic ideal of randomness—the toss of a coin.
Such “minimal-group” studies have been done with children as well. The psychologist Rebecca Bigler and her colleagues did a series of experiments in which children in summer programs were divided up arbitrarily—some got blue T-shirts and others got red T-shirts. They found that
if the children’s teachers mentioned these distinctions and used them to divide children into competing teams, robust in-group preferences emerged—the children preferred children of their own color (of shirt, that is) and allocated more resources to their group.
Other researchers found that explicit cues from a teacher weren’t even necessary; they could create group preferences just by giving the children different-color T-shirts or categorizing them based on the toss of a coin. Children in these experiments gave more money to their own group, predicted that their own group would behave better, and were more likely to remember bad acts that were done by an out-group member.
Now, people won’t seize on just
any
distinction. If someone is sitting at one side of a crowded table, she can divide the group into those who are sitting on her side versus those who are sitting on the other side, or those who are on her right versus those on her left—but neither of these divisions would be the basis for psychologically natural groups. That would be
too
minimal. Rather, children and adults glom on to differences that matter to the other people around them. We are social creatures, so distinctions that are as arbitrary as heads versus tails, red shirts versus blue shirts, and Klee lovers versus Kandinsky lovers can matter to us, but only when we see that others take them seriously. It’s not quite right, then, to say that we form groups solely on the basis of something as arbitrary as a coin toss. It’s not the toss per se but the fact that the toss is done in a social situation where the outcome clearly matters to other people.
As another illustration of the social nature of categories,
recall that even babies are capable of distinguishing people according to the color of their skin. But children display no early-emerging bias to choose friends on the basis of skin color: preschoolers don’t care about race, and neither do older children in certain mixed schools. If skin color is socially relevant—if black kids sit at one table and white kids at another—children will pick up on this. If not, they won’t. We start off prepared to make distinctions, but it’s our environments that tell us precisely how to do so.
M
ANY
of the generalizations that we make about social groups have some basis in reality.
The science writer David Berreby begins his book
Us and Them
with the observation that on the streets of his neighborhood in New York he sees people, almost always women, pushing children in strollers; when he sees a white adult with a nonwhite child, he assumes that the adult is the parent, but when he sees a nonwhite adult with a white child, he assumes that the adult is a nanny.
Berreby asks rhetorically if there’s something wrong with him for thinking this. The answer might be yes if he thought that this pattern had no exceptions—if the idea of a nonwhite adult being the parent of a white child was
impossible.
But Berreby knows full well that it is a generalization, not an absolute rule. As a different example, one might notice that there are a lot of Jewish university professors.
Jews make up between 1 and 2 percent of the total American population and 4 percent of the population in New Haven, Connecticut, the city where I live and teach.
I haven’t seen any statistics, but I can assure you the proportion of my colleagues who are Jewish is a lot higher than 4 percent.
The origins of these generalizations are better understood through history and sociology than through psychology, neuroscience, or evolutionary biology. It would be absurd to explain the gross disparities between whites and blacks in America, for instance, without reference to the legacy of slavery and Jim Crow.