1.

We all want to know how and why we got to be who we are. Parents have a special interest in answering the “how” and “why” questions with respect to their own children. In addressing the mysteries of human growth, traditional societies have invoked God, the gods, the fates, with luck sometimes thrown in. Shakespeare called our attention to the struggle between “nature and nurture.”1

In our own time the natural sciences and the social sciences have been supplying a bewildering variety of answers. Those with biological leanings look to heredity—the gene complexes of each parent and the ways in which their melded sets of genes express themselves in the offspring. The traits and capacities of the biological parents are seen as in large part determining the characteristics of offspring. Those with a psychological or sociological perspective point to the factors beyond the child’s physiology. Psychoanalysts emphasize the pivotal role of parents, and especially the young child’s relationship to his or her mother. Behaviorists look at the contingencies of reward and punishment in the child’s experience; the character of the child depends on the qualities that are “reinforced,” with those in control of reinforcement in early life having an especially significant influence.

Recently, three new candidates have been proposed to explain “socialization”—i.e., how children grow up within a society and absorb its norms. Impressed and alarmed by the powers of new means of communication, particularly television, students of culture like Marie Winn and Neil Postman have described a generation raised by the electronic media. The historian of science Frank Sulloway has brought new attention to the once discounted factor of “birth order”: on his account, first-borns embrace the status quo, while later-borns are far more likely to support scientific, political, or religious revolutions. And now, in a much publicized new work, Judith Rich Harris suggests that all of these authorities have got it wrong. On her account, the most potent “socializers” are the child’s peers, with parents having little or no effect.

Harris’s work has many things going for it. For a start, she has an arresting hypothesis, one that should strike especially responsive chords in adults who feel they are inadequately involved in the formation of the post-baby boom Generation X and the generations to come. She has an appealing personal story. Kicked out of graduate school in psychology in the early 1960s and a victim of a lupus-like disease, she has hitherto led the life of a semi-invalid, making her living coauthoring textbooks in psychology. One day in 1994, after reading a scholarly article about juvenile delinquency, she was struck by the idea that the role of peers in socialization had largely been ignored while the influence of parents had been much overestimated. She succeeded in publishing a theoretical statement of her view in Psychological Review, the most prestigious journal of psychological theory. She soon gained recognition among scholars and, in a delicious irony, won a prestigious award named after George Miller, the very professor who had signed her letter of expulsion from Harvard almost four decades ago. Harris’s book is well-written, toughly argued, filled with telling anecdotes and biting wit. It has endorsements from some of the most prestigious names in the field. Already it has been widely—and mostly favorably—reported on and reviewed in the popular press.

However, in my view, Harris’s thesis is overstated, misleading, and potentially harmful. Overstated in the sense that she highlights evidence consistent with her thesis and understates evidence that undermines it. Misleading because she treats as “natural” and “universal” what, in my view, is really a characterization of contemporary American culture (and those societies influenced by America). Potentially harmful in that it may, if inadvertently, discourage parents from promoting their own beliefs and values, and from becoming models of behavior, at a time when such values and models should be clearly and continually conveyed to children.

2.

Harris begins by outlining familiar positions in psychology. On her account, Freud’s view of the Oedipal period is quaint and unsupported, while the behaviorists have been widely discredited, both by the cognitivists (who put the mind back into psychology) and the biologists (who reminded us that we are as much a product of our genes as of our experiences). She then turns her keen critical skills to an attack on the branch of empirical psychology that attempts to document important contributions of parents to their children’s personality and character. (Harris uses both terms.)

For over half a century, psychologists and anthropologists have observed parents and children in different settings; they have filled out checklists in which they record predominant kinds of behavior and action, and they have administered questionnaires to the parents and children themselves. These researchers, according to Harris, began with the “nurture assumption”; they presupposed that the most important force in the child’s environment is the child’s parents and then collected evidence to support that assumption. Moreover, while scholars themselves are often guarded in their conclusions, some “pop” psychologists have no inhibitions whatever. They stress the role of parents over all other forces, thus making parents feel guilty if they fail (according to their own criteria), and full of pride when they succeed.

Advertisement

As Harris shrewdly points out, there are two problems with the nurture assumption. First, when viewed with a critical eye, the empirical evidence about parental influences on their children is weak, and often equivocal. After hundreds of studies, many with individually suggestive findings, it is still difficult to pinpoint the strong effects that parents have on their children. Even the effects of the most extreme experiences—divorce, adoption, and abuse—prove elusive to capture. Harris cites Eleanor Maccoby, one of the leading researchers in the field, who concluded that “in a study of nearly four hundred families, few connections were found between parental child-rearing practices (as reported by parents in detailed interviews) and independent assessments of children’s personality characteristics—so few, indeed, that virtually nothing was published relating the two sets of data.”

The second problem with the nurture assumption is potentially more devastating. Harris draws heavily on recent results from behavioral genetics to argue that, even in those cases where children resemble their parents, the presence and actions of parents have little to do with that resemblance. The argument she makes from behavioral genetics runs as follows. Studies of siblings, fraternal twins, identical twins reared together, and identical twins reared apart all point to the same conclusion: about half of one’s intellect and personality results from one’s genes. That is, in any group of people drawn from a particular “population” (e.g., middle-class white youngsters living in the United States), about one half of the variations in an observed trait (for instance, IQ or aggressiveness) is owing to one’s parents’ genetic contribution. The other half is, of course, the result of one’s environment.

For those who assume that the behavior of parents and the models they offer make up a major part of the child’s environment, the results of studies in behavioral genetics are surprising. According to those studies, when we examine any population of children and try to account for the nongenetic variations among them, we find that remarkably few variations can be attributed to their “shared environment”—i.e., when parents treat all of their children the same way, for example, being equally punitive to each child.

In fact, according to the behavioral geneticists, nearly all of the variation is due to what is called the “nonshared environment”—i.e., the variety of other influences, including instances where children are treated differently by the parents (e.g., a brother is punished more than his sister, or differently). In the case of any particular child, we simply do not know with any accuracy what makes up the nonshared environment. We can guess that it consists of siblings, printed matter, radio and television, other adults, school, luck, accident, the different (as opposed to the common or “shared”) ways in which each parent responds to each child, and—if Judith Rich Harris is correct—most especially, a child’s peers.

So much for Harris’s demolition of the importance of parents—except genetically—to the behavior and psyche of the child. Harris adduces evidence from a wide variety of sources, moreover, to stress the important contribution of peers. She goes back to the studies of nonhuman primates to indicate the importance of peer groups in child-rearing—pointing out that monkeys can be successfully reared by peers alone but not by their mothers alone. (It’s not known whether this would be true in “higher” primates.) She cites observations of children in different cultures who play together as much and as early as possible, and routinely gang up on the adults (teachers, parents, masters). She searches in the experimental literature for cases where peers exert an appreciable influence upon one another—for example, adolescents who have the same friends turn out to resemble one another. And she places great emphasis on the human tendency to form groups—and particularly “in-groups” with which one strongly identifies.

Harris also provides many telling anecdotes from her own experiences, and from the press and television, about how adults are ignored and peers admired. British boys who rarely see their parents successfully absorb social values at boarding school. Secretary of Labor Robert Reich quit the Cabinet to be with his sons in Cambridge and found that they would rather hang out “in the Square.” Touchingly she indicates how she and her husband tried to deal with their wayward adopted daughter but finally realized that the peers had more influence. No such problems existed with their biological daughter, who simply followed her biological destiny; the model provided by her parents was no more than an unnecessary bonus.

Advertisement

Harris describes recurrent situations where youngsters overlook the evident models of their parents in favor of those provided by peers. Deaf children of speaking parents ignore their parents’ attempts to teach them to read lips and instead begin to invent gestural signs to communicate with other deaf children and seek opportunities to learn formal signing. The hearing children of deaf parents, Harris points out, learn to speak normally in the absence of a parental model. Analogously, children raised by parents with foreign accents soon begin to speak like their peers, without an accent; like the deaf children, they ignore the models at home and turn, as if magnetized, to the most available set of peers. Arguments like these convince Harris, and apparently many readers (both lay and professional), that young human beings are wired to attend to people of similar age, rather than to those large and obvious authority figures who give them birth and early shelter.

3.

Harris has collected an impressive set of examples and findings to fortify a position that is indeed novel in empirical investigations of “human socialization.” I have sought to do justice to her arguments, though I cannot convey her passion, her missionary sense of having seen the light. Yet I do not find her “peer hypothesis” convincing, partly because I read the literature on the subject differently. My deeper reservations come from my belief that Harris has misconstrued the problem of socialization and, in doing so, has put forth a position that harbors its own dangers.

When we consider the empirical part of Harris’s argument, we find it is indeed true that the research on parent-child socialization is not what we would hope for. However, this says less about parents and children and more about the state of psychological research, particularly with reference to “softer variables” such as affection and ambition. While psychologists have made genuine progress in the study of visual perception and measurable progress in the study of cognition, we do not really know what to look for or how to measure human personality traits, individual emotions, and motivations, let alone character.

Consider, as an example, the categories that the respondents must use when they describe themselves or others on the Personal Attributes Questionnaire, a test used to obtain data about a person’s self-esteem and gender-linked traits.2 Drawing on a list reminiscent of the Boy Scout oath, those who answer the questionnaire are asked whether they would describe themselves as Gentle, Helpful, Active, Competitive, and Worldly. These terms are not easy to define and people are certainly prone to apply them favorably to their own case. Or consider the list of acts from which observers can choose to characterize children from different cultures—Offers Help, Acts Sociably, Assaults Sociably, Seeks Dominance.3 Even if we could agree on what kinds of physical behavior merit these labels, we don’t know with any confidence what these acts mean to children, adolescents, and adults in diverse cultures—let alone to the observers from a distant university. What does a raised fist or a frown mean to a three-year-old or to the thirty-year-old who observes it? The same question could be asked about a wink or an imitated curtsy. We are not measuring chemical bonds or electrical voltage in such cases. We are seeking to quantify the most subtle human characteristics—the sentiments described so finely by Henry James. And therefore it is not surprising when studies—whether by empirical psychologists or behavioral geneticists—do not yield strong results.

I do not want to elevate psychoanalytic theory or practice over other kinds of inquiry, but at least the Freudians were grappling with the deeper aspects of human character and personality—our urgent longings, our innermost fears and anxieties, our wrenching conflicts. We might perhaps find evidence for these complex feelings—and their putative causes—through long narratives, or projective testing (where the subjects respond to ambiguous photographs or inkblots), or by analyzing a series of sessions on the couch. We won’t reach them through questionnaires or checklists; yet Harris relies on many studies that use them.

As social scientists we have been frustrated by our own clumsy efforts to understand personality and character, and even relatively measurable skills, like intelligence or the capacity for problem-solving. And perhaps that is why so many talented psychologists—including the ones quoted on the jacket of The Nurture Assumption—have become drawn to evolutionary psychology and behavioral genetics. Here, at last, is the chance to put psychology and social science (and even squishy inquiries into personality, temperament, and character) on what seems a “real” scientific footing. Physics envy has been replaced by biological bias.

But things are not as clear-cut in the biobehavioral world as outsiders may imagine. Because of the possibility of controlled experiments, sociobiology has made genuine progress in explaining the social life of insects; but its account of human behavior remains controversial. The speculations of evolutionary psychology are just that; as commentators such as Stephen Jay Gould and Steve Jones have pointed out in these pages, it is difficult to know how to disprove a hypothesis in evolutionary psychology. (For example, what evidence can help us decide whether genes, or humans, are really selfish, or really altruistic, or really both?—in which case we are back where we started.)

And what of behavioral genetics? Certainly the opportunity to study twins who have been separated early in life gives us an additional advantage in understanding the heritability of various traits. And Judith Harris rightly calls attention to two enigmas: the fact that identical twins reared apart are almost as alike as those that are reared together; and the fact that identical twins still turn out to be quite different from one another.

But this subject is also dogged by difficulties. We cannot really do experiments in human behavioral genetics; we have to wait until events happen (as when twins are separated early in life) and then study the effects retrospectively. But this approach leaves too many puzzles unaddressed. First of all, for at least nine crucial months, the twins share the same environment—the womb of the birth mother—and we still know very little about the shared chemical and other effects of gestation on their neurological systems. Then, too, they may or may not have been separated right at birth. (And under what extraordinary circumstances does such separation occur?) They may or may not have been raised for a while by family members. The children are not randomly placed; in nearly all cases, they are raised within the same culture and very often in the same community, with similar social settings. Also, infants who look the same and behave the same are likely to elicit similar responses from adults, while those who are raised in the same house may try all the harder to distinguish themselves from one another. Or they may not.

When you add together the uncertainties (and I have only suggested a few of them here) of human behavioral genetics, and the imprecision of the measures used to describe personality and character, it is no wonder that we find little reliable evidence of parental influence. It would be reassuring if we did—but it is not surprising that we do not.

Which brings me to the alternative picture that Harris attempts to construct. She argues that “peers” are the real instrument of socialization. She may be right; but she does not have the evidence to show this. Her assertions depend almost entirely on what she thinks could one day be shown. Indeed, I find it extremely telling that she relies very heavily on the arguments about language—language-learning among the deaf, and the loss of foreign accents. Neither of these has to do with personality, character, or temperament, her supposed topics. In the case of accents, I assume that we are dealing with an unconscious (and presumably innate) process in which the growing child generalizes from his encounters with many of the adults and children he meets outside the home and through television, the movies, and other media. In the case of deafness, the enormous difference between child and parents forces youngsters to make use of resources outside the home—ranging from adult teachers to television and other visual media.

Indeed, despite some imaginative suggestions by Harris, it is very difficult to envision how one could test her hypothesis. For, after all, who are peers? Do they include siblings? Are they the children in the neighborhood? The children in class? The children in after-school activities or in Sunday school? The children on television? In the movies? At some remote spot on the Internet? Who decides? What happens when peers change because the family moves, or one child switches schools, or leaves (or is kicked out of) one group and then enters another? Most important, who selects peers? At least with parents, we researchers stand on fairly firm ground; and with siblings as well. But for all Ms. Harris’s anecdotes, when it comes to peers, we’re afloat.

Undoubtedly, psychological researchers inspired by Harris’s book will seek evidence bearing on her thesis. We will learn from these studies; and some of us who have taken skeptical positions in this debate may have to acknowledge influences we hadn’t sufficiently recognized. Meanwhile, I want to suggest an entirely different approach to the problem, one that might be called “the culture assumption.”

4.

What is socialization about? It is about becoming a certain kind of person—gaining specific knowledge, skills, manners, attitudes, and habits. Animals have little culture; human beings revel in it. Yet what is striking in Harris’s book is that the words “disciplines,” “civilization,” and “culture” (in the sense of civilization) are largely absent from the text and from her thinking. Socialization is reduced to having, or not having, certain personality traits—traits that are measured by rather coarsely conceived and applied tests.

The work of the much-maligned Freud remains the best point of departure for a treatment of these issues. In his Civilization and Its Discontents, Freud defined culture: “the sum of the achievements and institutions which differentiate our lives from those of our animal forebears, namely that of protecting humanity against nature and of regulating the relations of human beings among themselves.” He concentrates particularly on “the one feature of culture which characterizes it better than any other, and that is the value that it sets upon the higher mental activities—intellectual, scientific, and aesthetic achievement.” And he speculates that culture (or civilization) rests upon the human superego—the sense of guilt—which develops (or fails to develop) during the child’s early interactions with his parents. Guilt keeps us from murdering our fellow citizens; guilt prompts us to delay gratification, to sublimate our primordial passions in favor of loftier pursuits.

Whether one examines the least developed preliterate culture or the most advanced technological society, the question remains the same: What structures and practices will enable children to assume their places in that culture and ultimately aid in transmitting it to the generations to come?

Children will have some say in this process, and it is to Harris’s credit (and that of the authorities whom she cites) that she has called attention to this fact. But children are not born just into a family or into a peer group. They are born into an entire culture, whose assumptions begin when the parents say, happily or with a twinge of regret, “It’s a girl,” and continues to exert its influence in nearly every interaction and experience until the funerary rites, burial, cremation, or ascent to heaven takes place.

Earlier, I referred to Eleanor Maccoby’s pessimistic conclusions about documenting parental influence, and I mentioned some of the studies of it that both Maccoby and Harris seem to have had in mind. But let me reconsider the most ambitious of these studies in a different light. In the 1950s and 1960s, John Whiting, Beatrice Whiting, and their colleagues studied childrearing in six cultures, ranging from a small New England town to agricultural settings in Kenya, India, Mexico, the Philippines, and Okinawa.4 What emerges from that study is that childrearing practices are distinctly different around the globe: different in treatment of infants, in parental sleeping patterns, in how children do chores, in their helping or not helping in rearing younger siblings, in initiation rites, in ways of handling aggression, and in dozens of other variables. So differently are children reared in these cultures that no one would confuse an adult New Englander with an adult Gusii of Kenya or an adult Taira of Okinawa—whether in their knowledge, skills, manners, habits, personality, or temperament.

For the social scientist, the analytic problem is to find the source of these differences. Parents behave differently in these cultures, but so do siblings, peers, other adults, and even visiting anthropologists. And of course the adult roles, natural resources, technology, and means of communication (primitive or modern) differ as well. In all probability, each of these factors makes its contribution to the child’s “personality and character.” But how to tell them apart? Harris chooses to minimize these other factors and zooms in on the peers, but her confident choice is not justified.

5.

Harris takes little note of a crucial fact: all but a few of the studies that she reviews, including several of the most influential behavioral genetic ones, were carried out in the United States. The United States is not a country without culture; it has many subcultures and a more general “national” culture as well. Harris and most of the authorities that she cites are not studying child-rearing in general; indeed, they are studying child-rearing largely in the white, middle-class United States during the last half-century.

From the time of Alexis de Tocqueville’s visit to the United States in the early 1830s, observers have noted the relative importance in this country of peers, friends, or fellow workers of the same age, the members of one’s own community. Tocqueville commented, “In America the family, in the Roman and aristocratic signification of the word, does not exist. All that remains are a few vestiges in the first years of childhood….” As a sociologist might put it, America is a more horizontal, “peer-oriented” society than most others, and particularly more so than most traditional societies.

When empirical social science began in this country, these unusual cultural patterns were noted as well. Studying the America of the 1940s, the sociologist David Riesman and his coauthors called attention to the decline of tradition-centered and “inner-directed” families, where the parental models were powerful; and to the concomitant rise of the “other-directed families” that made up “the Lonely Crowd.” In this increasingly common family constellation, much socialization occurred at the behest of the peer group, whether for adults or for children. Riesman wrote, “The American peer group, too, cannot be matched for power throughout the middle-class world.”

Examining the America of the 1950s and 1960s, the psychologist Urie Bronfrenbrenner noted that children spend more time with peers than with parents and reached the same conclusion: “Whether in comparison to other contemporary cultures, or to itself over time, American society emerges as one that gives decreasing prominence to the family as a socializing agency…. We are coming to live in a society that is exaggerated not only by race and class, but also by age.”5 Thus not only has the peer group had an important part in American society from the first; but in recent decades this trend has accelerated.

But there are many possible peer groups. To which ones are children drawn and why? Here I believe (and Harris concedes this) that parents have a decisive role—by the friendships they encourage or discourage, by the schools they select or avoid, by the after-school activities they encourage and summer camps they approve of, parents contribute substantially to the choice of possible peer groups. I would go one step further. Children themselves select—and are selected for—various peer groups according to parental predilections. The work of the social psychologist Mihaly Csikszentmihalyi on “talented teens”6 strongly suggests that the values exhibited at home—integrity vs. dishonesty, hard work vs. laziness, artistic interests vs. philistinism—imprint themselves on children and in turn serve as major determinants of the peer groups to which children are attracted and, not incidentally, the ones where they are welcomed or spurned.

6.

It seems that in every passing decade—perhaps in every passing selection of fall books—we are told of a new approach to bringing up children or of a new, villainous influence on family life. Certainly, we do not have the feeling of a steady scientific march toward truth. It is more as if we are on a roller-coaster, with each new hypothesis tending to invalidate the previous one.

Still, it would be defeatist simply to embrace the opposite perspective, to declare that each of the various factors—mother, father, grandparents, same- sex siblings, different-sex peers, television, etc.—is important and be done with it. As a scientific community, we can do better than this. To do so, we should be undertaking two activities.

First, even as we welcome the clarifications provided by evolution and genetics, we cannot lose sight of the different cultural settings in which research is carried out and the different meanings attached to seemingly similar traits and actions. Parents and peers have different meanings in Japan, Brazil, and the United States; what we learn from the Whitings, and from much other sociological and anthropological research, is that these “independent variables” cannot simply be equated in designing research or in interpreting findings. In fact, a father may be treated more like a sibling in one society, and an older sibling more like a father in another; parents may encourage children to associate with peers in one culture and to steer clear of them in another and, in yet another, to combat their influence in every way they can.

Second, even as we discover genes or gene clusters that appear to influence important social or psychological variables, we must not assume that we have “solved” the problem of socialization. We still don’t know the physical mechanisms by which genes actually affect the brain and cause people to make one choice or another. What triggers (or fails to trigger) genes will vary across cultural settings; and how their expression is understood will also vary. Young men, for example, may have a proclivity to imitate other young men of similar size and power, but that proclivity can be manipulated, depending upon whom the child is exposed to and which rewards and punishments are contingent upon imitation or non-imitation.

Each of the numerous influences on a child’s personality I have mentioned can surely have an effect, but the effect will vary among different children, families, and cultures. As science progresses, we may someday be able to predict the relative importance of each across these different factors. My reading of the research suggests that, on the average, parents and peers will turn out to have complementary roles: parents are more important when it comes to education, discipline, responsibility, orderliness, charitableness, and ways of interacting with authority figures. Peers are more important for learning cooperation, for finding the road to popularity, for inventing styles of interaction among people of the same age. Youngsters may find their peers more interesting, but they will look to their parents when contemplating their own futures.

Parental attitudes and efforts will determine to a significant extent how a child resolves the conflicting messages of the home and the wider community as well as the kind of parent the child one day becomes. I would give much weight to the hundreds of studies pointing toward parental influence and to the folk wisdom accumulated by hundreds of societies over thousands of years.7 And I would, accordingly, be skeptical of a perspective, such as Ms. Harris’s, that relies too heavily on heritability statistics and manages to reanalyze numerous studies and practices so that they all somehow point to the peer group.

To gain attention, an author often states a finding or hypothesis very strongly. (I’ve been guilty of this myself.) In Harris’s case, this has led to a belittling of the roles of parents in child-rearing and to a stronger endorsement of the role of peers than the current data allow. I do not question Harris’s motives but I do question her judgment, which might have been better guided by the old medical oath “first, do no harm.”

It is all to the good if parents do not become crushed with anxiety when they have problems with their children or when their children turn out differently than they would like. Guilt is not always productive. But to suggest, with little foundation, that parents are not important in socialization borders on the irresponsible. Perhaps, on the average, those of us who are parents are not particularly successful in encouraging the personality traits we would hope to see in our children, whether because we do not know how to get their attention, or because they are “primed” to pay attention to their peers and we are not aware of how long and how hard we must work to counter these proclivities.

But children would not—could not—grow up to be members of a civilized culture if they were simply left to the examples of their peers. Indeed, parents are especially important when children’s peers set strong and destructive examples. In the absence of credible parents and other adults, most children will not be able to deal effectively with life. A social science—or a layman’s guide—that largely left out parents after birth would be absurd. So would a society.

Whether on the scene, or behind the scenes, parents have jointly created the institutions that train and inspire children: apprenticeships, schools, works of art and literature, religious classes, playing fields, and even forms of resistance and rebellion. These institutions, and t6he adults who run them, sustain civilization and provide the disciplines—however fragile they may seem—that keep our societies from reverting to barbarism.

Sad to say, these most important parts of life—which make life satisfying and fascinating—are largely absent from The Nurture Assumption. They are absent as well from most of the work emanating from the biotropic pole of contemporary social science. Until their importance is realized, and the biological and cultural perspectives are somehow deeply integrated with one another, scientific claims about children and family life are bound to remain barren.

This Issue

November 5, 1998