The Nurture Assumption: Why Children Turn Out the Way They Do by Judith Rich Harris, Steven Pinker (Foreword by)

The sweet, kind daughter was, of course, Cinderella. Unlike the Disney movie, this version depicts the (unnamed) stepsisters as beautiful. It was only their personalities that were ugly. In this respect, they closely resembled their mother. Cinderella presumably inherited her sweet nature from her mother, who was dead. Dead mothers were not a rarity in the old days; as many families were broken by death as are broken today by divorce.

In a fairy tale, events are compressed. Cinderella must have suffered years of abuse from her stepmother and stepsisters. She had no recourse: her father was unwilling or unable to stand up for her, and there were no laws or agencies in those days to protect children against mistreatment. She must have learned early on that it was best to remain as inconspicuous as possible, to do what she was told, and to accept verbal and physical insults without protest. And then -- then came the ball, and the fairy godmother, and the prince.

The folk who gave us this tale ask us to accept the following premises: that Cinderella was able to go to the ball and not be recognized by her stepsisters, that despite years of degradation she was able to charm and hold the attention of a sophisticated guy like the prince, that the prince didn't recognize her when he saw her again in her own home dressed in her workaday clothing, and that he never doubted that Cinderella would be able to fulfill the duties of a princess and, ultimately, of a queen.

Preposterous? Maybe not. The whole thing works if you accept one simple idea: that children develop different selves, different personas, in different environments. Cinderella learned when she was still quite small that it was best to act meek when her stepmother was around, and to look unattractive in order to avoid arousing her jealousy. But from time to time, like all children who are not kept under lock and key, she would slip out of the cottage in search of playmates. (They couldn't keep her locked in the cottage -- there was no indoor plumbing.)

Outside the cottage things were different. Outside the cottage no one insulted Cinderella or treated her like a slave, and she discovered that she could win friends (including the kindly neighbor whom she would later refer to as "my fairy godmother") by looking pretty. Her stepsisters didn't recognize her at the ball not just because she was dressed differently: her whole demeanor was different -- her facial expressions, her posture, the way she walked and talked. They had never seen her outside-the-cottage persona. And the prince, of course, had never seen her inside-the-cottage persona, so he didn't recognize her when he called at the cottage in search of the girl who dropped the shoe. She was quite charming at the ball, though admittedly lacking in sophistication. But that, he figured, could be easily remedied.

The Two Faces of Cinderella?

Perhaps it sounds like I am describing someone with a "split personality," like the protagonist of The Three Faces of Eve. But what made Eve abnormal was not the fact that she had more than one personality, or even that the alternate personalities were very different. Eve's problem was that her personalities appeared and disappeared unpredictably and didn't have access to each other's memories.

Having more than one personality is not abnormal. William James, brother of the novelist Henry James, was the first psychologist to point this out. Over a hundred years ago, William described multiple personalities in normal adolescents and adults -- that is, in normal male adolescents and adults.

Properly speaking, a man has as many social selves as there are individuals who recognize him and carry an image of him in their mind....But as the individuals who carry the images fall naturally into classes, we may practically say that he has as many different social selves as there are distinct groups of persons about whose opinion he cares. He generally shows a different side of himself to each of these different groups. Many a youth who is demure enough before his parents and teachers, swears and swaggers like a pirate among his 'tough' young friends. We do not show ourselves to our children as to our club-companions, to our customers as to the laborers we employ, to our own masters and employers as to our intimate friends. From this there results what practically is a division of the man into several selves; and this may be a discordant splitting, as where one is afraid to let one set of his acquaintances know him as he is elsewhere; or it may be a perfectly harmonious division of labor, as where one tender to his children is stern to the soldiers or prisoners under his command.

In other words, to put James's observations into current terminology, people behave differently in different social contexts. Contemporary personality theorists do not dispute this. What they argue about is whether there is any "real" personality under all these masks. If a man can be tender in one context and stern in another, which is he really? If three different men all are tender with their children and stern with their prisoners, isn't it the situation that determines personality and not the man?

The passage from William James comes from his book The Principles of Psychology -- America's first psychology textbook, published in 1890 (I own a copy of it, too tattered to be valuable). Because psychology was just beginning, James had it pretty much to himself for a while, and he stuck his finger in every pie. He talked about personality, cognition, language, sensation and perception, and child development. James was the one who said -- incorrectly; as it turned out -- that the world of the newborn infant is "one great blooming, buzzing confusion."

Today, these fields of psychology are entirely separate, presided over by specialists who seldom read articles outside their own field once they've made it through graduate school. Arcane arguments about adult personality are unlikely to attract the interest of socialization researchers. The word "selves" is not in the vocabulary of most behavioral geneticists.

Which is a pity, because I think it's relevant. Indeed, I think James's observation that people behave differently in different social contexts, and the subsequent discussions about why this happens and whether there is a "real" personality underneath, contain important clues to one of the big puzzles of personality development.

Here is the puzzle. There is evidence (I told you about it in Chapters 2 and 3) that parents cannot modify the personality their child was born with, at least not in ways that can be detected after the child grows up. If that is true, how come everyone is so certain that parents do have important effects on the child's personality?

Different Places, Different Faces

Unlike Eve of the Three Faces, most people do not have multiple personalities that lack access to each other's memories. Normal people may behave differently in different social contexts, but they carry along their memories from one context to another. Nonetheless, if they learn something in one situation they do not necessarily make use of it in another.

In fact, there is a strong tendency not to transfer the knowledge or training to new situations. According to learning theorist Douglas Detterman, there is no convincing evidence that people spontaneously transfer what they learned in one situation to a new situation, unless the new situation closely resembles the old one. Detterman points out that under-generalization may be more adaptive than overgeneralization. It is safer to assume that a new situation has new rules, and that one must determine what the new rules are, than to blithely forge ahead under the assumption that the old rules are still in effect.

At any rate, that is how babies appear to be constructed. Developmentalist Carolyn Rovee-Collier and her colleagues have done a series of experiments on the learning ability of young babies. The babies lie in a crib, looking up at a mobile hanging above them. A ribbon is tied to one of their ankles in such a way that when they kick that foot, the mobile jiggles. Six-month-old babies catch on to this very quickly: they are delighted to discover that they can control the mobile's movements by kicking their foot. Moreover, they will still remember the trick two weeks later. But if any detail of the experimental setup is changed -- if a couple of the doodads hanging from the mobile are replaced with slightly different doodads, or if the liner surrounding the crib is changed to one of a slightly different pattern, or if the crib itself is placed in a different room -- the babies will gaze up at the mobile cluelessly, as though they had never seen such a thing in their lives. Evidently babies are equipped with a learning mechanism that comes with a warning label: what you learn in one context will not necessarily work in another.

It is true: what you learn in one context will not necessarily work in another. A child who cries at home gets -- if he's lucky -- attention and sympathy. In nursery school, a child who cries too much is avoided by his peers; in grade school he is jeered at. A child who acts cute and babyish for her daddy evokes a different reaction from her classmates. Children who get laughs for their clever remarks at home wind up in the principal's office if they don't learn to hold their tongue in school. At home the squeaky wheel gets the grease; outside, the nail that sticks up gets hammered down. Or, as in Cinderella's case, vice versa.

Like Cinderella, most children have at least two distinct environments: the home and the world outside the home. Each has its own rules of behavior, its own punishments and payoffs. What made Cinderella's situation unusual was only that her two environments -- and hence her two personalities -- were unusually divergent. But children from ordinary middle-class American families also behave differently inside the home and outside of it. I remember when my children were in elementary school and my husband and I used to go to Back-to-School Night to meet their teachers. Year after year we would see parents talking to their child's teacher and coming away shaking their heads in disbelief. "Was she talking about my kid?" they would say, making it sound like a joke. But sometimes the teacher really seemed to be talking about a child who was a stranger to them. More often than not, this child was better behaved than the one they knew. "But he's so obstinate at home!" "At home she never shuts up for a minute!"

Children -- even preschoolers -- are remarkably good at switching from one personality to another. Perhaps they can do this more easily than older people. Have you ever listened to a couple of four-year-olds playing House?

Stephie (in her normal voice, to Caitlin): I'll be the mommy.

Stephie (in her unctuous mommy voice): All right, Baby, drink your bottle and be a good little baby.

Stephie (whispering): Pretend you don't like it.

Caitlin (in her baby voice): Don't want botta!

Stephie (in her unctuous mommy voice): Drink it, sweetheart. It's good for you!

Stephie plays three parts here: author/producer, stage director, and the starring role of Mommy. As she switches back and forth between them, she gives each one a different voice.

Context and Behavior

The "bottle" that Stephie was pretending to feed to Caitlin was a cylindrical wooden block. Developmentalists are interested in this kind of pretense because it appears to be an advanced, symbolic form of behavior, and yet it appears remarkably early -- before the age of two. Much has been written about the environmental influences that make pretense appear earlier or later; not surprisingly, attention has focused on the role of the child's mother. Researchers have found that a toddler engages in more advanced types of fantasy when the mother joins in the fantasy with the child.

But there is a catch. Greta Fein and Mary Fryer, specialists in children's play, reviewed the research and concluded that, although young children do play at a more advanced level when they are playing with their mothers, "the hypothesis that mothers contribute to subsequent play sophistication receives no support." When the mother encourages the child to engage in elaborate fantasies, the child can do it; but later, when the child is playing alone or with a playmate, it makes no difference what kinds of games she played with her mother.

Other developmentalists attacked this conclusion. Fein and Fryer responded by saying that they "did not intend to disparage the importance of adult caregivers in the lives of young children" and that they hadn't previously realized "how deep is the belief" in the omnipotence of parents. But they stuck to their guns. The evidence indicates that mothers influence children's play only while the children are playing with the mothers. "When theories don't work," Fein and Fryer counseled, "chuck 'em or change 'em." My view precisely.

Learning to do things with Mommy is all well and good, but the infant does not automatically transfer this learning to other contexts. This is a wise policy, because what is learned with Mommy might turn out to be useless in other contexts -- or worse than useless. Consider, for example, a baby I will call Andrew. Andrew's mother was suffering from postpartum depression, an affliction that is not uncommon in the first few months after childbirth. She was able to feed Andrew and change his diapers, but she didn't play with him or smile at him very much. By the time he was three months old, Andrew too was showing signs of depression. When he was with his mother he smiled infrequently and was less active than babies of that age usually are -- his face was serious, his movements muted. Fortunately, Andrew didn't spend all his time with his mother: he spent part of it at a day nursery, and the caregiver at the nursery was not depressed. Watch Andrew with his nursery caregiver and you will see a different baby, smiley and active. The somber faces and muted movements common in the babies of depressed mothers are "specific to their interactions with their depressed mothers," according to researchers who studied babies like Andrew.

Different behaviors in different social contexts have also been noted in older infants, infants of walking age. Researchers have studied how toddlers behave at home (by asking their mothers to fill out questionnaires) and at day-care centers (by observing them there or by asking the caregivers at the center) and found that the two descriptions of the children's behavior do not agree. "There exists the possibility that the toddler's actual behavior differs systematically in the home and day-care settings," admitted one researcher.

Sisters and Brothers

Granted that what children learn from interacting with their mothers might not help them get along with their peers in nursery school, but surely what they learn from interacting with their siblings should be transferable? You would think so -- I would have thought so too. But on second thought, children are probably better off starting from scratch with their peers. The child who dominates her younger brother at home may be the smallest one in her nursery school class; the dominated younger brother may turn out to he the largest and strongest in his. Here is what one team of researchers has to say on this topic:

There was no evidence of individual differences in sibling interactions carrying over into peer interactions....Even the second-born child, who has experienced years in a subordinate role with an older sibling, can step into a dominant role [with a peer].

And this from another:

Few significant associations were found between measures of children's sibling relationships and characteristics of their peer relationships....Children who were observed to be competitive and controlling to their siblings were reported by their mothers to have positive friendships. Children whose mothers reported that they had hostile sibling relationships received higher scores on friendship closeness....Indeed, we should not expect competitive and controlling behaviour toward a younger sibling to be necessarily associated with negative and problematic behaviour with friends.

Unless they happen to have a twin, children's relationships with their siblings are unequal. In most cases the elder is the leader, the younger is the follower. The elder attempts to dominate, the younger to avoid domination. Peer relationships are different. Peers are more equal, and often more compatible, than siblings. Among American children, conflict and hostility erupt far more frequently among siblings than among peers.

Conflict between siblings is the theme of Frank Sulloway's book Born to Rebel, which I mentioned in the previous chapter. In Sulloway's view, siblings are born to be rivals, fighting to get their fair share -- or, in the case of firstborns, more than their fair share -- of family resources and parental love. Children do this, he says, by specializing in different things: if one niche in the family is filled, the next child must find some other way of winning parental attention and approval.

I do not disagree with that. Nor do I doubt that people often drag their sibling rivalries along with them to adulthood and sometimes to the grave. My Aunt Gladys and my Uncle Ben hated each other all their lives. What I doubt is that people drag the emotions and behaviors they acquire in their sibling relationships to their other relationships. With anyone other than her brother Ben, my Aunt Gladys was as sweet and kind as the Cinderella in my childhood storybook.

The patterns of behavior that are acquired in sibling relationships neither help us nor hinder us in our dealings with other people. They leave no permanent marks on our character. If they did, researchers would be able to see their effects on personality tests given to adults: firstborns and laterborns would have somewhat different personalities in adulthood. As I reported in the previous chapter (also see Appendix 1), birth order effects do not turn up in the majority of studies of adult personality. They do, however, turn up in the majority of studies of one particular kind: the kind in which subjects' personalities are judged by their parents or siblings. When parents are asked to describe their children, they are likely to say that their firstborn is more serious, methodical, responsible, and anxious than their laterborns. When a younger brother or sister is asked to describe the firstborn, a word that turns up is "bossy." What we're getting is a picture of the way the subject behaves at home.

At home there are birth order effects, no question about it, and I believe that is why it's so hard to shake people's faith in them. If you see people with their parents or their siblings, you do see the differences you expect to see. The oldest does seem more serious, responsible, and bossy. The youngest does behave in a more carefree fashion. But that's how they act when they're together. These patterns of behavior are not like albatrosses that we have to drag along with us wherever we go, all through our lives. We don't even drag them to nursery school.

Never Leave Home Without It

My favorite example of a failure to transfer behavior from one context to another involves picky eating -- a common complaint among the parents of young children. You would think a picky eater in one setting would be a picky eater in another, wouldn't you? Yes, it has been studied, and no, that's not what the researchers found. One third of the children in a Swedish sample were picky eaters either at home or in school, but only 8 percent were picky in both places.

Ah, but what about that 8 percent? It is time to admit that I have been misleading you: the correlation between behavior at home and behavior outside the home may be low, but it's not zero. I mentioned another example in Chapter 2: the children who behaved obnoxiously with their parents but not with their peers, or vice versa. The correlation between obnoxious behavior in the two settings was only 19, which means that if you saw how a child behaved with her parents you would be unlikely to predict correctly how she would behave with her peers. Still, the correlation was not zero; in fact, ir was statistically significant.

Significant, but surprisingly low. Surprising because, after all, it was the same child behaving in both contexts -- the same child with the same genes. We know from behavioral genetic research that personality traits such as disagreeableness and aggressiveness have heritabilities of around 50 percent. That means a sizable portion of a child's personality (the exact percentage isn't important) is built in, innate, not acquired through experience.Children who have a built-in tendency to be disagreeable take this tendency with them wherever they go, from one social context to another. What they've learned may be tied to the context it was acquired in, but what they were born with they cannot leave behind. The child who is a picky eater both at home and at school may have food allergies or a delicate digestive system. Thus, the fact that some children are picky both at home and at school, and some children are obnoxious both with their parents and with their peers, could be due to direct genetic effects.

Indirect genetic effects -- the effects of the effects of the genes -- can also lead to a carryover of behavior from one context to another. Cinderella's case was unusual: her prettiness put her in danger whenever she was within striking distance of her stepmother. Only in the world outside the cottage was her prettiness an asset. Most pretty children find their prettiness an asset wherever they go. Most homely children learn that homeliness is a disadvantage in every social context. Perhaps some of the children who were obnoxious both with their parents and with their peers were physically unattractive children who had given up trying to get their way by being pleasant, because it didn't work with anyone. Or perhaps they were born with unpleasant dispositions, which made their dealings with all sorts of people problematic. A disagreeable temperament can lead to trouble both directly and indirectly: directly because it makes the child respond unfavorably to other people, indirectly because it makes other people respond unfavorably to the child.

Code-Switching

The carryover of behavior from one context to another due to genetic effects is a nuisance for me -- it gets in the way of the point I am trying to make. I am trying to convince you that children learn separately, in each social context, how to behave in that context. But social behavior is complicated. It is determined partly by characteristics people are born with, partly by what they experience after they are born. The inborn part goes with them wherever they go and thus tends to blur the distinctions between social contexts. To solve this problem I will turn to a social behavior that's acquired entirely through experience: language.

Perhaps I'd better qualify that statement. Language is acquired through experience; yet it is also innate. It is one of the things that we inherit from our ancestors but that does not vary among normal members of our species, like lungs and eyes and the ability to walk erect. Every human baby born with a normal brain is equipped with the ability and desire to learn a language. The environment merely determines which language will be learned.

In North America and Europe, we take it for granted that we must teach our babies how to communicate with language; indeed, we consider that to be one of a parent's important jobs. We start the language-learning lessons early, talking to our babies the minute they're out of the womb, if not before, We encourage their coos and babbles and make a big deal out of their "mamas" and "dadas." We ask them questions and await their replies; if they don't reply we answer the questions ourselves. If they make a grammatical error we rephrase their poorly formed phrase into proper English (or proper whatever). We speak to them in short, clear phrases about things they're interested in.

Thus encouraged, not to say prodded, our babies start talking when they're barely a year old and are speaking in sentences when they're barely two. By the age of four they're competent speakers of English (or whatever).

Now I ask you to imagine a child who goes outside her home for the first time at the age of four and discovers -- as Cinderella did -- that out there everything is different. Only in this case, what's different is that everyone is speaking a language she can't understand, and no one can understand her language. Will she be surprised? Probably not, judging from the reaction of the babies who learned to jiggle the mobile by kicking one foot. Change the liner surrounding the crib and they're in a different world. They assume that the new world has new rules, yet to be learned.

Children of immigrant parents, like the kids of the Russian couple who ran the rooming house in Cambridge (described in Chapter 1), are in exactly that situation. They learn things at home -- most conspicuously a language but other things as well -- that prove to be useless outside the home. Unfazed, they learn the rules of their other world. They learn, if necessary, a new language.

Children have a great desire to communicate with other children, and this desire serves as a powerful incentive to learn the new language. A psycholinguist tells the story of a four-year-old boy from the United States, hospitalized in Montreal, trying to talk to the little girl in the next bed. When his repeated attempts to talk to her in English proved futile, he tried the only French words he knew, fleshed out with a few nonsense syllables: "Aga doodoo bubu petit garçon?" An Italian father living in Finland with his Swedish-speaking wife and son tells of the time he took his three-year-old son to a park and the boy wanted to play with some Finnish-speaking children. He ran up to them shouting the only words of Finnish he had learned: "Yksi, kaksi, kolme...yksi, kaksi, kolme" -- Finnish for "One, two, three."

This fools-rush-in approach is practiced mainly by younger children; older ones are more likely to start off with a least-said-soonest-mended strategy. Researchers studied a seven-and-a-half-year-old boy -- I'll call him Joseph -- who moved with his parents from Poland to rural Missouri. In school, Joseph listened quietly for several months, watching the other children for clues to what the teacher was saying. With neighborhood friends he was more willing to risk making mistakes and he started practicing his English with them almost immediately. At first Joseph's speech sounded like that of a toddler -- "I today school" -- but within a few months he was speaking serviceable English and after two years he was using it like a native, with hardly a trace of an accent. The accent eventually went away entirely, even though he continued to speak Polish at home.

It is common for immigrant children to use their first language at home and their second language outside the home. Give them a year in the new country and they are switching back and forth between their two languages as easily as I switch back and forth between programs on my computer. Step out of the house -- click on English. Go back in the house -- click on Polish. Psycholinguists call it code-switching.

Cinderella's alternate personas are an example of another kind of code-switching. Step out of the cottage -- look pretty, act charming. Go back in -- look homely, act humble. If she had also spoken one language in her home and another language outside it, as Joseph did, that would have been just another difference between life inside the cottage and life outside it. Mastering bilingualism is probably easier for a child than switching back and forth from looking pretty to looking homely.

Code-switching is sort of like having two separate storage tanks in the mind, each containing what was learned in a particular social context. According to Paul Kolers, a psycholinguist who studied bilingual adults, access to a given tank may require switching to the language used in that context. As an example, he mentioned a colleague of his who had moved from France to the United States at the age of twelve. This man does his arithmetic in French, his calculus in English. "Mental activities and information learned in one context are not necessarily available for use in another," Kolers explained. "They often have to be learned anew in the second context, although perhaps with less time and effort."

It is not only book-learning that is stored in separate tanks. "Many bilingual people," reported Kolers, "say that they think differently and respond with different emotions to the same experience in their two languages." If they use one language exclusively at home, the other exclusively outside the home, the home language becomes linked to the thoughts and emotions experienced at home, the other to the thoughts and emotions experienced outside the home. At home Cinderella thought of herself as worthless, outside her home she found that she could win friends and influence people. A bilingual Cinderella might still be scrubbing floors if the prince had addressed her in the language used in her cottage.

Personality theorists don't pay much attention to language. And yet, language, accent, and vocabulary are aspects of social behavior, just as "personality traits" such as agreeableness and aggressiveness are. Like other aspects of social behavior, the language a person uses is sensitive to context, and this is as true for monolingual speakers as it is for bilingual ones. William James said that a person "shows a different side of himself" in different social contexts and gave as his first example the youth who swears like a pirate when he's with his friends but is "demure enough before his parents and teachers." A high school student tells this anecdote about one of his classmates:

A girl at my school was walking down the hall and remembered she forgot something.

"Oh shoot!" she exclaimed.

As she looked around and saw her friends she said, "I mean oh shit."

The girl's parents and teachers make similar adjustments in their verbal behavior. They do not use the same vocabulary or sentence structure when they're talking to a teenager as when they're talking to a two-year-old. They do not use the Same vocabulary or sentence structure when they're talking to their automobile mechanic as when they're talking to their doctor.

Though it is a social behavior, language has the advantage of being free of the genetic complications that plague other kinds of social behaviors. The tendency to be agreeable or aggressive is partly genetic, but the tendency to speak Polish rather than English, or to use swear words with some people and not with others, is entirely environmental.

Language and Social Context

Code-switching is an extreme example; most children's mental tanks do quite a bit of leaking. After all, they carry their memories With them wherever they go, from one context to another. A child who comes out of his house at the age of four and finds that people out there do speak the language he learned at home doesn't have to learn it all over again, although he may be cautious at first about using it outside his home. For most children, the home environment and the outside-the-home environment do not have steel walls between them. The parents come to school to watch their children perform in plays and for conferences with the teacher. The children reveal bits of their home lives in Show-and-Tell and "What I Did on My Summer Vacation." They invite their school friends to their homes for birthday parties.

When William James spoke of the "division of the man into several selves," he said there were two kinds of divisions: harmonious, as exemplified by the man who is tender with his children but stern with his prisoners, and discordant, "as where one is afraid to let one set of his acquaintances know him as he is elsewhere." Cinderella's division was discordant: she was afraid to let her stepmother see her as she was elsewhere. Some psychologists and psychiatrists believe that severe abuse in childhood can lead to multiple personality disorder, the three-faces-of-Eve phenomenon. The connections between the mental tanks are broken, or never get formed, and each personality accumulates its own memories and fails to share them with the others.

Most children do not risk a beating if they reveal bits of their outside-the-home behavior to their parents. But it is common for children to act as though some terrible punishment would ensue if they reveal bits of their home behavior outside the home. Philip Roth, in his novel Portnoy's Complaint, tells an anecdote that is almost certainly autobiographical. Here's Alexander Portnoy -- the son of first-generation Jewish Americans who speak English liberally sprinkled with Yiddish words -- describing an incident from his childhood:

I was already the darling of the first grade, and in every schoolroom competition, expected to win hands down, when I was asked by the teacher one day to identify a picture of what I knew perfectly well my mother referred to as a "spatula." But for the life of me I could not think of the word in English. Stammering and flushing, I sank defeated into my seat, not nearly so stunned as my teacher but badly shaken up just the same...in a state resembling torment -- in this particular instance over something as monumental as a kitchen utensil.

Alexander thought spatula was a Yiddish word -- a home word, a family word -- and he would rather be struck dead than use it in public. I had a similar experience in third or fourth grade when I used the word pinky to refer to my little finger. The girl I was talking to (not a close friend) asked, "What did you say?" and I was struck with panic. I had made a fatal error: pinky must be a home word! The girl asked again, "What did you say?" and I mumbled "Nothing." She became insistent and I became more and more embarrassed but refused to tell her what I had said. Years later I realized that she, too, must have been unsure of the status of the word pinky and was trying to find out if it was a legitimate outside-the-home word.

Joseph spoke Polish to his parents and English to his teachers, schoolmates, and friends. But sometimes his friends came over to his house to play with him and he spoke to them in English, and thus English crept into his home. Or perhaps, like Alexander Portnoy, he was embarrassed to use his home language outside the home, so when he went shopping with his parents he spoke to them in English. However it starts, the children of people who immigrate to English-speaking countries usually end up bringing English home with them, speaking English to their parents. Here's the son of Korean immigrants, describing how he communicated with his mother: "She would mostly speak to me in Korean, and I would answer her in English." Here's an anthropologist explaining why Jewish immigrants from Eastern Europe failed to transmit their language to their children: "They talked Yiddish to their children and the children answered in English." The same sort of thing happens, in a smaller way, in homes in which everyone speaks English: I have heard a great many native-born Americans complain that their children come home talking in the uncouth accents of their peers.

If immigrant parents insist that their children continue to address them in their native language -- that is, the parents' native language -- the children may do so, but their ability to communicate in that language will remain childish, while their ability to communicate in the outside-the-home language continues to grow. Here's a young Chinese-American woman, the child of immigrants, who went to Harvard:

I had never discussed literature or philosophy with my parents. We talked about our health, the weather, that night's dinner -- all in Cantonese, since they do not speak English. While at Harvard, I ran out of words to communicate with my parents. I literally did not have the Cantonese vocabulary to explain the classes I was taking or my field of concentration.

Many immigrant parents see their children losing the language and culture of their homeland and try very hard to prevent it. My local newspaper ran a story about a woman from West Bengal, India, who started a Bengali language school for her children and the children of other Bengali-speaking immigrants.

Like many immigrants, Bagchi wants her children to understand their cultural background. To do that, she believes, they first must be fluent in Bengali, their parents' native tongue and one of 15 languages spoken in India....But learning a language isn't easy if you study it for only a few hours each week. School, television and peer groups immerse children in English, and despite the best of efforts by both parents and children, it often is a challenge to become fluent in the parents' language. "They dream in English. They do not dream in Bengali," Bagchi said, describing Bengali children born in the United States.

They dream in English. It makes no difference whether the first language they learned from their parents was English or Bengali, English has become their "native language." Joseph spoke nothing but Polish for the first seven and a half years of his life, but if he remains in the United States his "native language" will not be Polish. As an adult he will think in English, dream in English, do his arithmetic and his calculus in English. He may forget his Polish entirely.

Parents do not have to teach their children the language of their community; in fact -- hard as it may be for you to accept this -- they do not have to teach their children any language at all. The language lessons we give our infants and toddlers are a peculiarity of our culture. In parts of the world where people still live in traditional ways, no lessons are given and parents generally do very little conversing with their babies and toddlers -- they consider learning the language the child's job, not the parents'. According to psycholinguist Steven Pinker, mothers in many societies "do not speak to their prelinguistic children at all, except for occasional demands and rebukes. This is not unreasonable. After all, young children plainly can't understand a word you say. So why waste your breath in soliloquies?" Compared to American toddlers, the two-year-olds in these societies seem retarded in their language development, but the end result is the same: all the children eventually become competent speakers of their language.

You are thinking, Yes, but even though the mother doesn't speak to the baby, the baby hears her speaking to other people. True. But even this is unnecessary. There is an old story, told by the Greek historian Herodotus, of a king who wanted to find out what language children would speak if left to their own devices. He had a couple of babies reared in a lonely hut by a shepherd and gave instructions that no one was to talk to them or speak a word in their hearing. Two years later he visited the children and, the story goes, they ran up to him saying something that sounded like "bekos," which is the word for bread in an ancient language called Phrygian. The king concluded that Phrygian must have been the world's first language.

Would it shock you to learn that in the United States there are thousands of babies being reared like that? No, it is not an experiment. These are babies born to profoundly deaf couples. Most deaf people marry other deaf people, but more than 90 percent of the babies born to these couples have normal hearing. These babies miss out on some of the experiences we consider crucial to normal development. No one comes running when they scream in terror or in pain. No one encourages their coos and babbles or makes a big deal out of their "mamas" and "dadas." Nowadays most deaf parents use sign language to communicate with their hearing children, but there was a period when the use of sign language was frowned upon, and during that period some deaf parents didn't communicate with their young children at all, except in the most rudimentary ways. And yet, these children suffered no harm. Despite the fact that they didn't learn any language at all from their parents, they became fluent speakers of English. Don't ask them how they learned it; they can't remember and many of them consider the question offensive. I assume they learned the same way Joseph did.

Socialization researchers are unlikely to study families in which the parents speak Polish or Bengali, much less families in which the parents communicate only in sign. They don't worry about how and where children acquire their language because it is a constant: all the parents in their studies speak English and so do all the children, and the researchers assume that children must learn their language from their parents. They make the same assumption about other aspects of socialization. They are wrong about language and I believe they are wrong about other aspects of socialization. Bilingualism is simply the most conspicuous marker of context-specific socialization -- socialization that is tied to a particular social context.

A Place for Everything and Everything in Its Place

As the spatula story suggests, children appear to be motivated to keep their two lives separate. Child abuse often goes undetected because children don't like to talk about it when they're outside their home. They don't want anyone to know that their home is different -- that their step-mother beats them and makes them scrub the floor. Conversely, schoolage children often fail to tell their parents if they've been victims of bullying on the playground. I was a social outcast for four years of my childhood -- none of my classmates would talk to me -- and my parents never knew about it.

But the motivation to keep the home life from leaking out is stronger than the motivation to keep the outside world from leaking in, and it is especially strong in those who have an inkling that their homes might be abnormal in some way. If their mother drinks, their parents throw things at each other, or their father is an invalid, kids don't want anyone to know about it. The child of immigrants might avoid inviting friends over to play. The kid whose parents are wealthier than their neighbors may be as anxious to keep it a secret as the kid whose parents are poorer: what they hate is being different from their peers.

In order to know what has to be concealed, children need some way of learning whether their homes are normal or abnormal. One way they can do this is by watching television; however, this works only if the families they see on television are not too obviously different from the families they see in their neighborhood. If the differences are too great, then children must base their concepts of normal family life on what they learn from their friends and classmates.

Getting information from friends and classmates can be tricky. Mutual efforts by a pair of children to find out about the other's family often fail because both children fear they have something to hide, which is what happened when I used the word pinky to my classmate. But children have a clever way of getting around this problem: they play House. In the game of House, children can cooperatively develop an idea of what a "normal" family is like and at the same time limit their risks, because, after all, it's just a game.

Have you ever listened to children play House or similar games of pretense? The families they depict are straight out of Ozzie and Harriet. Talk about stereotypes! A developmentalist recorded this announcement, made by a little boy playing the part of Daddy: "Okay, I'm all through with work, honey. I brought home a thousand dollars." The girl playing Mommy was quite pleased. But a little boy who wanted to cook dinner was told firmly by his playmate that "daddies don't cook." Another child, a girl, was heard to insist that girls had to be nurses -- only boys could be doctors -- even though her own mother was a physician.

Aside from being sexist, the parents depicted in the game of House are curiously benign. They may argue with each other and scold their "baby," but they seldom go further than that. It isn't that children eschew depictions of violence -- on the contrary, as researchers Iona and Peter Opie observed, "In these playlets children are stolen to be eaten [and] mutilation is accepted almost as commonplace." But in the games of pretend violence, the villains are witches or monsters or robbers and the children themselves often pretend to be orphans, thus explaining why good ol' Mom and Dad aren't around to protect them. If their real parents neglect them or abuse them, that's the last thing they want their friends to know.

Kids want desperately to be normal, and part of being normal is having normal parents. If their parents are different in some way -- and they're bound to be different in some way -- they want to conceal these embarrassing differences from their peers. The humor writer Dave Barry has captured the feeling:

After canteen we'd stand outside the school, surrounded by our peers, waiting for our parents to pick us up; when my dad pulled up, wearing his poodle hat and driving his Nash Metropolitan -- a comically tiny vehicle resembling those cars outside supermarkets that go up and down when you put in a quarter, except the Metropolitan looked sillier and had a smaller motor -- I was mortified. I might as well have been getting picked up by a flying saucer piloted by some bizarre, multitentacled, stalk-eyed, slobber-mouthed alien being that had somehow got hold of a Russian hat. I was horrified at what my peers might think of my dad; it never occurred to me that my peers didn't even notice my dad, because they were too busy being mortified by THEIR parents.

Parents belong in the home; when they come out of the home it makes their children nervous. Aside from the embarrassment, it makes it harder for children to know which context they're in, which rules they're supposed to follow. They are not aware of this, of course; context almost always affects behavior at a level that is not normally accessible to the conscious mind. It isn't until adolescence or adulthood that people occasionally become aware of the way their behavior changes in various social contexts. Perhaps there are people you don't like to be with because you don't like the way you act when you're with them.

The youth described by William James was "demure enough before his parents and teachers" but behaved differently when he was with his friends. He acted the way his parents and teachers had taught him to act, but only in social contexts that included his parents or teachers. It's difficult to teach your dog not to sleep on the sofa when you're not around, because what you are actually teaching him is to stay off the sofa when you are around. When you're not at home, he never gets whacked for jumping up on the sofa.

Seventy years ago a pair of ahead-of-their-time developmentalists tested children's ability to resist temptation. They gave the children opportunities to cheat or steal in a variety of settings: at home, in the classroom, in athletic contests; alone or in the presence of peers. They discovered that children who were honest in one context were not necessarily honest in others. The child who was honest at home might lie or cheat in the classroom or on the athletic field.

When children or adolescents misbehave outside their homes, they are sometimes referred to as "unsocialized" and their parents are blamed. According to the nurture assumption, it is the parents' job to socialize the child. But if children fail to transfer things their parents taught them to other social contexts, it is not their parents' fault.

Will the Real Personality Please Stand Up?

Babies are born with certain characteristics, certain tendencies to behave one way or another. They may have a greater than average tendency to be physically active, or to seek the company of others, or to get angry. These inborn tendencies are built upon and modified by the environment -- that is, by each of the child's environments, separately.

Personality has two components: an inborn component and an environmental component. The inborn part goes with you wherever you go; it influences, to some extent, your behavior in every context. The environmental component is specific to the context in which you acquired it. It includes not only the way you learned to behave in those contexts, but also the feelings you associate with those contexts. If your parents made you feel worthless, those feelings of worthlessness are associated with the social contexts in which your parents did that to you. The feelings of worthlessness will be associated with outside-the-home contexts only if the people you encountered outside your home also made you feel like that.

Stability of personality across social contexts depends in part on how different or similar a person's various contexts have been. Cinderella's two social contexts were unusually divergent, so there was more than the usual amount of variation in her personality. But someone who met her after the prince Carried her off to the castle wouldn't know that. They would see only her outside-the-cottage personality.

The psychologists who study adult personality typically assess it by giving their subjects a self-report personality test -- a standardized list of self-descriptive statements, each of which the subject must agree or disagree with. In most cases the subjects are college students and the test is administered in a classroom or laboratory at the college. Thus, what the test is measuring is the subject's college personality, along with any thoughts or emotions associated with that particular classroom or laboratory. If the test is given again months later, to measure consistency across time, it is again given in a classroom or laboratory -- usually the same one. The subject may be in a better or worse mood this time, but basically it's the same personality, with the same associated thoughts and emotions, so the results are reasonably consistent.

Personality psychologist James Council gave college students a self-report test that was designed to measure their ability to become absorbed in imaginative activities. Then he tried hypnotizing them. The subjects who scored highest on absorption were the easiest to hypnotize, but only if he tried hypnotizing them in the same room where they took the absorption test. When the test was given in one room and the hypnotizing was done in another, there was no significant correlation between the two. In a second experiment, Council asked subjects to fill out a questionnaire that asked them about traumatic childhood experiences such as physical or sexual abuse. Then, immediately afterward, they took a personality test designed to look for signs of emotional problems. There was a significant correlation between reports of childhood trauma and signs of emotional problems. But when Council tried the same thing on a different batch of subjects, this time giving them the personality test first, the correlation disappeared. Taking the trauma test evoked unpleasant thoughts and emotions and associated them with the test-taking setting. The effects of those unpleasant thoughts and emotions could be detected on a personality test if it was given after the trauma test and in the same setting. Council believes that these "context effects" call into question "the validity of a great deal of personality research."

Let's say you wanted to demonstrate that childhood trauma leads to emotional problems in adulthood. One way you could do it is with the method Council used: remind your subjects of their trauma and then, immediately after and in the same room, have them take the personality test. But an even better way would be to bring them back to the place where they experienced the trauma and have them take the personality test there. What you will be demonstrating, however, is not the power of childhood trauma to mess up people's minds. You will be demonstrating the power of context.

When behavioral geneticists study adult personality, they give their subjects personality tests in classrooms or laboratories. They find that the homes in which these subjects grew up had little or no effect on their adult personalities. If behavioral geneticists want to find effects of the home environment, they should take their subjects back to the homes in which they grew up and give them the test there. But what they wilt be demonstrating is not the power of the childhood home to influence adult personality. They will be demonstrating the power of context.

If you never go home again, the personality you acquired there may be lost forever. After Cinderella married the prince she never returned to her stepmother's cottage. Her self-effacing cottage persona was left behind forever, along with her broom and her raggedy clothes.

Most people do go home again. And the moment they walk in the door and hear their mother's voice from the kitchen -- "Is that you, dear?" -- the old personality they thought they had outgrown comes back to haunt them. In the world outside they are dignified, successful women and men, but put them back at the family dinner table and pretty soon they are bickering and whining again; just like they did in the good old days. No wonder so many people hate going home for holidays.

Made of the Myth

One of the reasons you didn't believe me when I told you that the nurture assumption is a myth is that there's so much evidence to support it. You can see that parents have effects on their children! And socialization researchers have collected mountains of data to prove it!

Yes, but where did you see it, and where did they collect it? You are right that parents have effects on their children, but what evidence do you have that these effects persist when the parents are not around? The child who acts obnoxious in the presence of her parents may be demure enough before her classmates and teachers,

Much of the evidence used by socialization researchers to support their belief in the nurture assumption consists of observations of the child's behavior in the presence of the parents, or questionnaires about the child's behavior filled out by the mother. Researchers want to demonstrate effects of the home environment -- for example, after a divorce -- so they observe the children in the home, a home where a lot of unpleasant things have happened recently. Worse yet, they ask the parents -- not exactly what you'd call neutral observers, especially after the turmoil of a divorce -- to fill out a questionnaire about the child's behavior. Predictably, these methods often show that the children of divorce are in significantly worse shape than those whose parents remained married. If the observations are made outside the home, away from the parents, the differences between the offspring of divorced and nondivorced parents get much smaller or go away entirely. (However, some of the differences do persist -- they can be detected even in adulthood. I'll come back to the children of divorce in Chapter 13.)

Context effects are a serious problem in developmental psychology. They produce correlations that don't mean what the researchers think they mean or what they want them to mean. The correlations may turn up in the laboratory as well as in the home. Older children and adolescents are often interviewed or asked to fill out questionnaires in a school classroom or laboratory. This is a method the style-of-parenting researchers often use: they give the kids a self-report personality test or a questionnaire about what kind of trouble they've gotten into lately, and another questionnaire asking them how their parents treat them. Now we have not only a context effect (because the kid fills out both questionnaires in the same setting) but also what might be called a "person effect" -- the same person who's telling you that he smoked four joints this week and flunked a math test is also telling you what jerks his parents are. One team of researchers checked up on their subjects. They gave teenagers a questionnaire asking them about their parents' child-rearing methods and also had their parents fill out the same questionnaire. The correlation between the parents' reports and the kids' reports was only .07 -- in other words, there was no agreement at all. And yet socialization researchers accept at face value kids' descriptions (and parents' descriptions) of what goes on in their homes and use data of this sort to support their theories.

Socialization research has demonstrated one thing clearly and irrefutably: a parent's behavior toward a child affects how the child behaves in the presence of the parent or in contexts that are associated with the parent. I have no problem with that -- I agree with it. The parent's behavior also affects the way the child feels about the parent. When a parent favors one child over another, not only does it cause hard feelings between the children -- it also causes the unfavored child to harbor hard feelings against the parent. These feelings can last a lifetime.

There are hundreds of books that give advice to parents -- books that tell you what you're doing wrong and how to do a better job of raising your kids. Find a good one and it may help to explain why your children behave the way they do when they're at home. My goal is to explain what makes them behave the way they do in the world outside the home -- the world where they will spend the rest of their lives.

Copyright © 1998 by Judith Rich Harris