Tuesday, April 01, 2003

Extreme Problems with Essential Differences

The Essential Difference: The Truth about the Male & Female Brain

By: Lesley J. Rogers, D.Phil., D.Sc.


At the beginning of his new book, Simon Baron-Cohen tells us that he is speaking out on a subject that has been taboo: the innate nature of sex differences. I am not sure when or where this topic has been taboo, however, since sex differences have been debated going back to Aristotle. In the mid to late 1800s, Charles Darwin and Gustave Le Bon, who was one of the founders of social psychology, had plenty to say about them—and about the inferiority of women.

The issue of whether there are essential differences between men and women and, if so, what causes them, has never been laid to rest. Baron-Cohen believes that alleged male superiority on spatial tasks and female superiority in language skills are surface differences. At the real core, he argues, are men’s ability to systemize information and women’s to empathize. He suggests that the ability to empathize may underlie better language skills, since the empathizer pays more attention to listening to others and to communicating well. The ability to analyze information and construct systems may underlie effective use of spatial information.

A lecturer in psychopathology at the University of Cambridge, Baron-Cohen has substantial expertise on both autism and Asperger’s syndrome. He presents his theory of sex differences in the context of a hypothesis about these two disorders, which he interprets as manifestations of the “extreme male brain.” But the question for this book is whether his theory of essential sex differences stands up to scientific scrutiny and aids our understanding of autism and Asperger’s syndrome.

By “essential,” Baron-Cohen means that these sex differences are programmed in the genes. He states the theory on page 1: The female brain is “predominantly hard-wired for empathy” and the male brain “hard-wired for understanding and building systems.” I am far from convinced by the evidence he presents for this idea. Nor am I persuaded that his theory is particularly courageous, as he attempts paint it. On the contrary, scientists who echo the beliefs, including the prejudices, of their society are always heard, and the topic of sex differences is no exception. Even during the now legendary 1970s and 1980s, the heyday of second-wave feminism, traditionalists who believed in essential differences between men and women had no apparent difficulty being published. Female scientists who argued along with traditionalists (for example, Corinne Hutt, Katharina Dalton, and Doreen Kimura) attracted a following and continue to do so today, despite the attempts of some feminists conversant with biology to have their criticisms of theories of genetically (and hormonally) determined sex differences heard. In truth, it is ideas other than those based on “essential” differences that have remained in the minority, and today there is a new emphasis on genetic explanations of a wide range behaviors, sex differences among them. 

So, to clarify the baseline of this debate: Baron-Cohen is in the company of the majority, not an oppressed minority. His attempt to win over his audience by claiming to be a lone voice crying some sort of truth against the constraints imposed by misguided feminism of the last three decades comes across as insincere. I have no objection to a debate on any new theory about sex differences, but the point is not that some feminists agree with the author, as he says, but how good is his scientific evidence and how new are his ideas?


In the 1960s, Donald Broverman put forward a theory similar to Baron-Cohen’s. Broverman said that men are better than women at complex restructuring tasks, whereas women excel over men on simple repetitive motor tasks. This division between the sexes seems consistent with Baron-Cohen’s proposal for the male brain, although he does not cite Broverman. The female characteristics of emotional expression, nurturing, and empathizing have an even longer history, so it is difficult to concur with Baron-Cohen’s statement that “two neglected dimensions [of sex differences] are empathizing and systemizing.”

Baron-Cohen is at pains to reassure the reader that he does not wish to “repeat old forms of oppression in a new guise.” No doubt readers will applaud. The problem is that in the area of sex differences emotions tend to run high, and ideology is on hand with alleged facts to support its (usually traditional) views. In my opinion, Baron-Cohen does not achieve his aim to avoid further oppression of women. He attempts to blur the differences between men and women by pointing out that there is substantial overlap between the sexes and that he is discussing average, not absolute, differences. By telling us that the differences are “hard-wired,” however, he reinforces age-old myths about the behavioral differences between men and women. For example, a superior quality of women, he says, is their ability to host a large party, tactfully making everyone feel included. But this is a minor anecdotal piece of information concerning a small minority of women in a specific social situation. Even this skill, by the way, may take years of learning and experience to reach perfection and reflects specific social values. I am not aware of any convincing evidence for the idea that women’s ability to host a party well is an innate quality.

According to Baron-Cohen’s theory, women are more likely to have what he calls a Type E (empathizing) brain, manifested in reading magazines on fashion, romance, intimacy, and beauty and spending their time on coffee hours and caring for people and pets. Men, he says, are more likely to have a type S (systemizing) brain that directs their attention to magazines on computers, cars, boats, science, and consumer guides and, because they like to engage in activities with rules, they are more likely to practice plane or bird spotting and photography. In the wider arena, a type S would be a scientist, politician, or physician, and he would be a leader, whereas a type E would be a teacher, homemaker, or nurse and, of course, a mother. This clearly reinforces traditionally held divisions between the sexes. I fail to see anything new here. The female proclivity to empathize differs little from the long-asserted claim that female nurturing behavior is innate.

I do not think there is anyone today who seriously contests that sex differences in brain function exist. What is open to question is how large the average differences are and how meaningful they are in everyday life. Even more questionable are the assertions that natural selection acted on humans to ensure that those differences became encoded in our genes. The main point, however, is that the existence of any average difference between the sexes tells us nothing about what causes it. That is the essential question for science, and its answer is the only basis for supporting or changing social attitudes and programs.


Baron-Cohen advances three lines of evidence for his claim that human sex differences in empathizing and systemizing are determined genetically: the age at which these differences first appear, the lack of apparent variation across cultures, and the existence of sex differences in animals.

Age at Which Sex Differences First Appear

If boys and girls show significant differences in behavior very early in life, can we use this as evidence that the differences are genetically caused, rather than being learned as a consequence of the different ways in which adults treat boys and girls? The first fact to recognize here is that the differential treatment begins from the day of birth with, for example, subtle differences in the way that nursing staff and parents interact with the newborn. The second relevant point is that most studies of sex differences in early life have tested infants at one to three years of age, which gives ample time for cultural influence.

Although Baron-Cohen does not cite much of it, there is research evidence that adults actively reinforce sex stereotypes in infants in their first years of life. For example, in 1976, J.A. Will, P. A. Self, and N. Datan tested the responses of women to the same year-old child, whose true sex they did not know. When the child was dressed as a girl, they encouraged play with dolls and spoke gently. When dressed as a boy, the child was encouraged to play with trucks and tools and interactions were less verbal. Baron-Cohen points out that parents may treat their sons and daughters differently because they are, in fact, different and the parents are responding to that difference. This would not explain the results obtained in this experiment, however. 

More recent research by Estelle Campenni in 1999 showed similar encouragement by mothers of sex-stereotyped play in boys and girls. A cross-cultural study by Hugh Lytton and David Romney in 1991 showed that parents are stricter with boys than girls, and that fathers tend to differentiate between their sons and daughters more than do mothers. In fact, the typical male behavior in boys tends to be enforced by coercion, compared to encouragement of girls to adopt typical female behavior. No wonder, therefore, that as early as two years old, boys show more aggressive behavior than girls. The same applies to girls’ capacity for greater empathizing at only a year old and to the evidence that by three years old girls have a “theory of mind” (that is, know that another person has his or her own separate way of thinking). There has been sufficient time to learn these sex-typical characteristics.

Notably, one of Baron-Cohen’s own students has found sex differences in a sample of one-day-old infants. The baby girls looked for a longer period at a face than at a mobile, while the boys did the reverse. It is unlikely that this difference could have been learned from sex-differentiated treatment, unless the very first interactions with adults have some immediate effect. In the absence of any supporting evidence for that, though, we must accept Baron-Cohen’s interpretation as likely to be correct: The cause of this particular difference depends on expression of different genes in boys and girls. But it is a vast extrapolation to say that this result is the first evidence that females are type E and males type S. Looking longer at a face does not mean that the infant is empathizing with it or will do so later. It could even mean the reverse: that the child has taken longer to habituate because she has not been able to relate to the face. In addition, since development is characterized by marked shifts in attention and behavior, sometimes changing from week to week, it is possible that the male infants might have found the face more interesting than the mobile when they were a little older, or that the opposite would occur in the female infants. It is an even greater leap to say that the difference found in the one-day-old children explains male-female divisions in employment patterns in adulthood, as Baron-Cohen suggests at one point.

Continual sex-specific conditioning throughout life coupled with social prejudice is by far the most likely reason some occupations are heavily dominated by men. Baron-Cohen’s acceptance of claims by male physicists among his colleagues that they do not discriminate against women in making new appointments speaks of a lack of understanding of social pressures, which are both

subtle and blatant and occur throughout life. By calling on such anecdotal support for his beliefs, Baron-Cohen displays a lack of scientific rigor. It is dangerous to mix empirical facts with unreliable information gleaned from chatting over lunch.

Lack of Variation Across Cultures

As evidence of genetic determinism for patterns of male and female behavior, Baron-Cohen points to the commonality of sex roles or traits across cultures. But it is possible that these consistent sex-role differences are largely the result of culturally determined roles adopted by our common ancestors and maintained across the generations by learning, which can be as resistant to change as any behavior coded in the genes. Also, apparent invariance across cultures is often claimed on the basis of testing only a few groups from cultures that do not differ greatly from one another. For example, the claimed cultural invariance of male superiority in spatial ability is undermined by John Berry’s 1966 study showing that Eskimo women are superior to Eskimo men in spatial ability. This one exception is sufficient to throw the argument for genetic causation into doubt.

Existence of Sex Differences in Animals

The existence in animals of sex differences is often taken as evidence for their genetic causation and against any effects of learning. But this incorrectly assumes that animals acquire their sex-typical behavior entirely without learning or, if learning is involved, only to a degree that is minimal compared to the effect of genes. Elegant studies by Celia Moore in the 1980s, which Baron-Cohen does not cite, have shown just how important early experience is for the development of sex differences in rats. In my own 1999 book, Sexing the Brain, I reported in detail on Moore’s work showing how mother rats lick the anogenital region of their male pups more than of their female pups and how this difference in stimulation leads to development of sex-typical behavior —an example of how a particular environmental stimulus causes specific sex-defined behaviors. This was confirmed by Moore’s subsequent observation that female pups receiving artificial stimulation of their anogenital region in early life develop male-typical behavior. Litters in which the male and female pups are licked equally by their mother display no sex differences in behavior as adults. Moore inserted small tubes into the mothers’ nostrils to ensure that they did not discriminate between their male and female pups, licking them equally because they were unable to smell any difference. This also showed that smell was the main basis on which they made this sex discrimination. 

Male pups have higher levels of testosterone than do female pups, which causes them to produce urine that attracts their mothers to lick their anogenital region. If female pups are injected with testosterone, they are licked more and then develop male-typical behavior. We can conclude from this that behaving as an adult male or female does not depend on the direct action of testosterone on the brain, as many have assumed, but on the interactive influences of genes, hormone secretion, and maternal treatment during early life. More important, we can see that altered early experience (here: anogenital licking) can override any effect of the genes, although genes, of course, play an inseparable role along with experience during the normal course of development.

These detailed, controlled experiments on rats are a model of accurate understanding of the complex interactive processes of development that lead to sex differences. They should caution us not to assume that levels of testosterone or other hormones either before or after birth change behavior by acting directly on the brain. Baron-Cohen mentions a study that found that lower fetal levels of testosterone in humans correlate with higher levels of eye contact and a larger vocabulary at one and two years of age, supposedly a manifestation of the female-typical brain. Although he interprets this to mean that the testosterone acts directly on the brain to masculinize it, the hormone’s influence could have been indirect by changing the parents’ behavior in some way.

Are the sex differences observed in monkeys and apes similar to those in humans? Baron-Cohen suggests that female monkeys, apes, and humans all show more interest in babies than do males, which may be a marker of the heightened emotional sensitivity of the female. General statements like this about primates have little validity, since attention to offspring varies greatly from one species of primate to another—and among individuals of the same species. In marmoset monkeys, for example, the male does much of the caring for offspring. In apes, as zookeepers have learned from long experience, the female has to learn to be a good parent. Even in the wild, individuals vary greatly in acquiring maternal behavior beyond suckling.

Likewise it is risky to consider, as Baron-Cohen suggests, that the spatial ability of male rats, and especially laboratory bred ones, may be a precursor to human male superiority in systemizing. Many species have remarkable spatial abilities—far better than ours—but the assumption that there are consistent sex differences in spatial ability across all species cannot be upheld. Far too few species have been studied in depth in this context, and overall the proposition makes no sense in terms of survival in the natural environment.

Both male and female animals may need excellent spatial abilities to survive in the same environments. Migratory birds, male and female, are equally successful in reaching their destination using their sophisticated spatial ability. The same applies to mammalian species that migrate. Moreover, food gathering may require excellent spatial skills and memory. In many avian species, the female raises her offspring alone and, if anything, in these species one ought to expect superior spatial ability in females. Baron-Cohen’s attempt to demonstrate universality of sex differences in spatial ability across species, as well as across human cultures, fails due to selective use of scanty scientific information.


Spatial processing is interesting because it is mainly a function of the right hemisphere in humans and other species. The left and right hemispheres of the brain are specialized to process different information and to control different functions. In most humans, the left hemisphere is specialized for language and the right for spatial ability. This specialization of the right hemisphere for processing spatial information has been shown in a number of mammalian and avian species, and so has its dominance in the expression of intense emotions. Some researchers have hypothesized that there are sex differences in the degree of asymmetry, and, at least in the case of language, there is some support for this (female brains are less lateralized than male brains). But it is not known whether this difference is a result of the effect of testosterone in causing the right hemisphere to develop earlier and grow faster than the left, as hypothesized by the neurologist Norman Geschwind, or of how and when the child learns language. Baron-Cohen is convinced, however, that Geschwind’s hypothesis is correct and extends it to encompass empathizing, which he links to language and so to the left hemisphere, and systemizing, which he links to spatial ability and so to the right hemisphere. Hence females should be left-hemisphere dominant and males right-hemisphere dominant.

There are two problems here. The first is that empathizing should involve emotions, as Baron-Cohen admits later, and emotions are controlled primarily by the right hemisphere. The second is the evidence in humans and other species (rats, chicks, and pigeons) that it is the left hemisphere that analyzes and categorizes information. This seems to me to be closer to Baron-Cohen’s definition of “systemizing” than is spatial ability. Therefore, if the sex differences in brain lateralization are meaningful, I believe we have a confusing picture: Females are superior for language (left hemisphere) and emotions (right hemisphere), and males are superior in systemizing (left hemisphere) and for spatial ability (right hemisphere). No simple effect of testosterone on the development of the right hemisphere would explain the development of this more complex pattern.

Individuals who score equally on empathizing and systemizing do exist and, according to Baron-Cohen, have a balanced brain. By contrast, there are the unbalanced conditions of being an extreme female brain (scoring high on empathizing and low on systemizing) or an extreme male brain (high on systemizing and low on empathizing). Baron-Cohen suggests that having an extreme male brain underlies both autism and Asperger’s syndrome. The author’s expertise in these two conditions is apparent in the chapters covering this topic, which make fascinating reading. Indeed, I can accept that autism is a disorder of the ability to empathize, and that it has all the other features that Baron-Cohen ably describes. But, for all the reasons discussed herein, I am unconvinced that labeling it “the extreme male brain” aids either its description or our understanding of how it develops.

Also, if autism is caused by high levels of testosterone in the fetus, as Baron-Cohen suggests it could be, we would have to remind ourselves that the same causal mechanism has been suggested (incorrectly, in my opinion) for lesbianism. Would the logical next step, then, be to label lesbians as autistic, with “extreme male” brains?


Although The Essential Difference is written in an engaging style, it presents one side of the complex issue of differences in brain function between the sexes, ignoring most of the scientific information that might muddy or contradict the author’s theory. Unfortunately, such opinions tend to have an enormous impact on social attitudes, so Baron-Cohen’s approach is not just scientifically unsound but, alas, appears to be socially prejudiced. Let me hasten to fend off criticism that I am merely being politically correct. Ultimately, any hypothesis must be scientifically verifiable; if there is telling evidence against it, one cannot simply ignore that evidence. It should be used to alter the hypothesis. There is abundant scientific evidence showing the complex, interactive, and ever-changing influences of experience and genes that take place as an organism develops. I fail to see how we can deepen our understanding of the functions of a host of contributing factors by parceling out the largest portion of control to the genes, which merely reduces the complex processes involved to untenable simplistic explanations.

I do not share the author’s optimism that “we can be confident that genes controlling empathizing and systemizing will be identified.” Time will confirm or refute this prediction, but already many early claims of discovering genes that control complex human behavior patterns, including schizophrenia, homosexuality, and depression, have evaporated under scrutiny and failed to be duplicated (many discussed in Gisela Kaplan and Lesley Rogers’s Gene Worship). Doubtless there are readers who will be heartened by the way Baron-Cohen reinforces their beliefs, but their sense of security will be an illusion.  


From The Essential Difference: The Truth about the Male & Female Brain, by Simon Baron-Cohen, Ph.D. © 2003 by Simon Baron-Cohen. Reprinted with permission of Perseus Publishing.

Little is known about the genetics of systemizing, but there are clues that genes play a role. For example, studies of mathematical ability in twins (one of the clearest examples of systemizing) show that identical twins are more alike in their mathematical ability than are non-identical twins. And children with developmental dyscalculia are “born not to count,” as Brian Butterworth at UCL [University College London] puts it. They are of normal intelligence and are sociable, but cannot systemize for genetic reasons.

With the human genome mapped, and the determinations of the functions of genes now a major industry, we can be confident that genes controlling empathizing and systemizing will be identified. Such genes will not rule out the role of culture and environment. Genetically and/or hormonally based neural systems underlying empathizing and systemizing still require the right environmental input (sensitive parenting, for example, in the case of empathizing) in order to develop normally. But identifying such genes or hormones will help us understand why, despite all the relevant environmental factors, some children are worse at empathizing, or better at systemizing, than others.

As the genetics of both systemizing and empathizing become better understood, this will raise an obvious question: Why should these abilities have become encoded in our genes? The traditional answer to this type of question is the one that Charles Darwin formulated, in terms of evolutionary theory: traits usually only come under genetic control when they confer some survival and reproductive advantage to the organism in two battles: surviving to reach adulthood, and being selected to become a parent. In the next chapter we confront this evolutionary question directly. In what ways might it have been adaptive or advantageous to have either the male or female brain types?...

Some theories suggest that our male and female ancestors occupied quite different niches and had very different roles. If true, the selective pressures are likely to have been very different for each, and could have led to the evolution of different types of cognitive specialization. Naturally, what may have been adaptive for one sex may not have been adaptive for the other.

In this chapter I speculate about why it might have been advantageous for females to have brain type E, and why it might have been advantageous for males to have brain type S. (Recall from Chapter 1 that brain type E is the description given to those who are able to empathize with greater skill than they are to systemize, and brain type S is the mirror image of this.)

Because empathizing is such a fundamentally human ability that must be as old as the Homo sapien brain itself, you might be prepared to accept that a relative talent in empathizing could characterize the female brain. But you might balk at the idea that a relative talent in systemizing could possibly characterize something as ancient as the male brain, since systemizing resembles the hypothesis-testing of a scientist, and scientific thinking is a recent human development. Let me put you right. Although academic science is relatively recent—a mere few hundred years old—“folk” science is as old as humans themselves. Tribal peoples have been developing their own understanding of natural systems, building their own technologies, formulating their own medical systems and establishing systems to govern their social groups, for tens of thousands of years.  

About Cerebrum

Bill Glovin, editor
Carolyn Asbury, Ph.D., consultant

Scientific Advisory Board
Joseph T. Coyle, M.D., Harvard Medical School
Pierre J. Magistretti, M.D., Ph.D., University of Lausanne Medical School and Hospital
Helen Mayberg, M.D., Icahn School of Medicine at Mount Sinai 
Bruce S. McEwen, Ph.D., The Rockefeller University
Donald Price, M.D., The Johns Hopkins University School of Medicine
Charles Zorumski, M.D., Washington University School of Medicine

Do you have a comment or question about something you've read in CerebrumContact Cerebrum Now.