Book: The Agile Gene

Previous: 2. A Plethora of Instincts
Next: 4. The Madness of Causes

Professors are inclined to attribute the intelligence of their children to nature, and the intelligence of their students to nurture.

Roger Masters

Disagreement thrives on uncertainty. In the 1860s, uncertainty over the source of the Nile was the source of a bitter dispute between two English explorers, John Hanning Speke and Richard Burton. Only two men who have shared a camp for many months could disagree so violently. Speke favored Lake Victoria, which he had discovered while Burton lay ill in a tent at Tabora; Burton insisted that the source lay in or near Lake Tanganyika. The feud did not end until 1864, when Speke shot himself (perhaps accidentally) on the day he was to debate with Burton in public. Speke, by the way, was right.

Watching this dispute from an influential position in the Royal Geographical Society, and occasionally fanning the flames on behalf of Burton, was a distinguished geographer named Francis Galton. It was Galton’s fate to ignite an even bigger feud in 1864, one that would run for more than a century: nature versus nurture. The nature–nurture debate is a bit like the argument over the source of the Nile. Both debates also thrived on ignorance; the more that came to be known, the less the argument seemed to matter. Both debates also seemed unnecessarily petty. Surely, what mattered more than which lake was the source of the Nile was that Africa contained two vast lakes new to western science. Likewise, it surely matters less whether human nature is more innate or more learned, but instead the precise way in which it is both. The Nile is the sum of thousands of streams, no one of which can be truly called its source; the same is true of human nature.

Galton’s passion was quantifying. In a long career, he invented, coined, or discovered a wide range of things: northern Namibia, anticyclone weather systems, the study of twins, questionnaires, fingerprints, composite photographs, statistical regression, and eugenics. But perhaps his most lasting legacy is to have inaugurated the nature–nurture debate and coined the very phrase. Born in 1822, he was a grandson of the great scientist, poet, and inventor Erasmus Darwin by Darwin’s second wife. He found his half-cousin Charles Darwin’s theory of natural selection both convincing and inspiring, ascribing this immodestly to “an hereditary bent of mind that both its illustrious author and myself have inherited from our common grandfather, Dr Erasmus Darwin.” Emboldened by his own pedigree, he now found his true calling in the statistics of heredity. In 1865, deserting geography, he published an article on “hereditary talent and character” in Macmillan’s Magazine, in which he revealed that distinguished men had distinguished relatives. He expanded it into a book called Hereditary Genius in 1869.

Galton was simply asserting that talent runs in families. Exhaustively and enthusiastically, he described the pedigrees of famous judges, statesmen, peers, commanders, scientists, poets, musicians, painters, divines, oarsmen, and wrestlers. “The arguments by which I endeavour to prove that genius is hereditary, consist in showing how large is the number of instances in which men who are more or less illustrious have eminent kinsfolk.” It was not very sophisticated reasoning. After all, one might just as well argue the opposite, that the rise of humble men to great eminence would reveal innate talents triumphing over the disadvantages of circumstance; the clustering of talent in families might indicate shared teaching. Most reviewers thought Galton had overstated the role of heredity and had ignored the contribution of upbringing and family. In 1872 a Swiss botanist, Alphonse de Candolle, asserted as much at book length. Candolle pointed out that great scientists in the previous two centuries had come from countries or cities with religious tolerance, widespread trade links, a moderate climate, and democratic government—suggesting that achievement owed more to circumstance and opportunity than to native genius.

Candolle’s attack stung Galton into writing a second book, English Men of Science: Their Nature and Nurture, in 1874, in which he used a questionnaire for the first time and repeated his conclusion that scientific geniuses were born, not made. It was in this book that he coined the famous alliteration:

The phrase “nature and nurture” is a convenient jingle of words, for it separates under two distinct heads the innumerable elements of which personality is composed.

He may have borrowed the phrase from Shakespeare, who in The Tempest has Prospero insult Caliban thus:

A devil, a born devil, on whose nature nurture can never stick.

Shakespeare was not the first to juxtapose the two words. Three decades before The Tempest was first performed, Richard Mulcaster, an Elizabethan schoolmaster who was the first headmaster of the Merchant Taylors’ school, was so fond of the antiphony of nature and nurture that he used it four times in his book Positions Concerning the Training Up of Children (1581):

… [Parents] will have their children nursed as well as they can, without question where, or quarrelling by whom: so as they may have that well brought up by nurture, which they love so well, bequeathed them by nature…. God hath provided that strength in nature, wherby he entendes no exception in nurture, for that which is in nature…. Which naturall abilities, if they be not perceived, by whom they should: do condemne all such, either of ignoraunce, if they could not judge, or of negligence, if they would not seeke, what were in children, by nature emplanted, for nurture to enlarge…. Which being thus, as both the truth tells the ignorant, and reading shewes the learned, we do wel then perceave by naturall men, and Philosophicall reasons, that young maidens deserve the traine: bycause they have that treasure, which belongeth unto it, bestowed on them by nature, to be bettered in them by nurture.

Mulcaster repeated the phrase in his next book, Elementaries, in 1582: “whereto nature makes him toward, but that nurture sets him forward.” Mulcaster was a curious character. Born in Carlisle, he was a distinguished scholar and famous, if strict, educational reformer. He quarreled irascibly with the school governors and was a passionate advocate of the game of football: “The foteball strengtheneth and brawneth the whole body,” he observed. Mulcaster also dabbled in drama, writing several pageants for the royal court and educating the playwrights Thomas Kyd and Thomas Lodge at his school. He is supposed by some to have been the model for the character of Holofernes, the vain schoolmaster in Love’s Labour’s Lost, so there is a good chance that Shakespeare either knew Mulcaster or read his works.

Shakespeare may also have been the inspiration for the next of Galton’s ideas. Two of Shakespeare’s plays turn on the confusion of twins: The Comedy of Errors and Twelfth Night. Shakespeare was himself the father of twins, and he used mistaken twins to make fiendishly ingenious plots. But, as Galton pointed out, in A Midsummer Night’s Dream Shakespeare introduced a pair of “virtual twins”—unrelated individuals who had been reared together. Hermia and Helena, despite being “like to a double cherry, seeming parted, but yet an union in partition,” not only look physically unlike each other but are attracted to different men and end up quarrelling violently.

Galton followed up the hint. The next year he wrote an article entitled “The History of Twins, as a Criterion of the Relative Powers of Nature and Nurture.” At last he had a respectable way to test the hypothesis of heredity, free of the objections raised against his pedigrees. Remarkably, he deduced that there were two sorts of twin: identical twins, born from “two germinal spots in the same ovum”; and nonidentical twins “each from a separate ovum.” This is not bad. For “germinal spot” read “nucleus” and you are close to the truth. Yet in both kinds, the twins shared nurture. So if identical twins resembled each other in behavior more than fraternal twins, then the influence of heredity was supported.

Galton wrote to 35 pairs of identical twins and 23 pairs of non-identical twins, collecting anecdotes about their similarities and differences. Triumphantly he recounted the results. Twins who resembled each other from birth remained similar throughout their lives, not only in appearance but also in ailments, personality, and interests. One pair suffered severe toothache in the same tooth at the same age. Another pair bought identical sets of champagne glasses as presents for each other at the same time at different ends of the country. Twins who were born different, by contrast, grew more different as they grew older. “They were never alike either in body or mind, and their dissimilarity increases daily,” said one of his respondents. “The external influences have been identical; they have never been separated.” Galton sounded almost embarrassed by the strength of his conclusion: “There is no escape from the conclusion that nature prevails enormously over nurture…. My fear is, that my evidence may seem to prove too much, and be discredited on that account, as it appears contrary to all experience that nurture should go for so little.”


With hindsight one can pick all sorts of holes in Galton’s first twin study. It was anecdotal and small, and the argument was circular: twins who appeared identical behaved identically. Galton had not distinguished identical twins from fraternal twins genetically. Yet the study was remarkably persuasive. By the end of his life Galton had seen his hereditarian beliefs move from surmise to orthodoxy. “Nature limits the powers of the mind as definitely as those of the body,” said The Nation in 1892. “On these points, among thinkers everywhere, [Galton’s] opinions have prevailed.” The old empiricism of John Locke, David Hume, and John Stuart Mill, whereby the mind was seen as a blank sheet of paper on which experience would write its script, had been replaced by a sort of neo-Calvinist notion of inherited individual destiny.

There are two ways to look at this development. You can damn Galton for being seduced by his “convenient jingle” into presenting a false dichotomy. You can see him as one of the evil spirits of the twentieth century, cursing the three generations that followed so that they swung like a pendulum between ridiculous extremes of environmental and genetic determinism. You can note with horror that from the beginning, Galton’s motives were eugenic. On the very first page of Hereditary Genius in 1869 he was already extolling the virtues of “judicious marriage,” lamenting the “degradation of human nature” by the propagation of the unfit, and invoking the “duty” of the authorities to exercise power to change human nature by progressive breeding. These suggestions would grow into the pseudoscience of eugenics. With hindsight, therefore, you can blame him for an idea that would cause misery and cruelty to millions in the century to come, not just in Nazi Germany but in some of the most tolerant countries of the world.

All this would be true, though it is a little harsh to expect that none of it would have happened without Galton, let alone that he should have foreseen where his ideas would lead. Even the convenient jingle would have soon occurred to somebody else. A more charitable reading of history would see Galton as a man far ahead of his time who hit upon a remarkable truth: that many aspects of our behavior start within us in some way, that we are not putty in the hands of society or victims of our surroundings. You could even—though this might be stretching it—assert that this notion was vital in keeping alive the flame of liberty in the environmentalist despotisms of the twentieth century: those of Lenin, Mao, and their imitators. Galton’s insights into heredity were remarkable, considering that he knew nothing about genes. He would have had to wait more than a century to see that the study of twins did in the end prove much of what he had suspected. To the extent that they can be teased apart, nature prevails over one kind of (shared) nurture when it comes to defining differences in personality, intelligence, and health between people within the same society. Note the caveats.

This is a recent development. Twenty years ago, the picture was very different. By the 1970s the whole notion of studying twins to learn about heredity was in eclipse. Two of the largest studies of twins since Galton were in disgrace. In Auschwitz, Josef Mengele was notoriously fascinated by twins. He sought them out among new arrivals at the concentration camp and segregated them into special quarters for study. Ironically, this “favoritism” led to a higher survival rate among twins than singletons—most of the small children who survived Auschwitz were twins. In exchange for submitting to procedures that were often brutal and sometimes fatal, they were at least better fed. All the same, few survived.

In Britain, the educational psychologist Cyril Burt was slowly accumulating a set of identical twins reared apart, which enabled him to calculate the heredity of intelligence. In 1966, when he published the full set of results, he claimed to have found 53 pairs of such twins. This was an extraordinarily large sample, and Burt’s conclusion that IQ was highly heritable influenced British educational policy. But it later emerged that at least some of the data were almost certainly faked. The psychologist Leon Kamin noticed that the correlation had remained exactly the same, to the third decimal place, even while the data set had expanded over several decades. The Sunday Times simultaneously asserted that two of Burt’s coauthors probably did not exist (one has since reappeared, however).

With a history like this, it was little wonder that twin research was a tainted subject in the 1970s. Yet today the study of twins has been reborn as the principal method of a scientific discipline known as behavior genetics that has flowered especially in the United States, Holland, Denmark, Sweden, and Australia. It is sophisticated, argumentative, mathematical, and expensive—everything that a thoroughly modern science should be. But at its core lies Galton’s insight: that human twinning provides a beautiful natural experiment for discerning the contributions of nature and nurture.

In this respect, fortune has been generous to human beings. The ability to produce identical twins seems to be rare in the animal kingdom. It is unknown in mice, for example, which produce litters of non-identical littermates. Human beings occasionally produce litters, too. Among white people, about one birth in every 125 consists of two non-identical, fraternal, or “dizygotic” twins—derived from two zygotes or fertilized eggs. The rate is higher among Africans and lower among Asians. But one birth in every 250 consists of identical (or monozygotic) twins, derived from a single fertilized egg. Without a genetic test, identical twins cannot be reliably distinguished from fraternal twins, though there are telltale signs. Their ears tend to be identical.

Behavior genetics is a simple matter of measuring how similar identical twins are, how different fraternals are, and how both identicals and fraternals turn out if separately adopted into different families. The result is an estimate of “heritability” for any trait. Heritability is a slippery concept, much misunderstood. For a start, it is a population average, meaningless for any individual person: you cannot say that Hermia has more heritable intelligence than Helena. When somebody says that the heritability of height is 90 percent, he does not and cannot mean that 90 percent of my inches came from my genes and 10 percent from my food. He means that the variation in height in a particular sample is attributable 90 percent to genes and 10 percent to environment. There is no variability in height for the individual and therefore no heritability.

Moreover, heritability can measure only variation, not absolutes. Most people are born with 10 fingers. Those with fewer have usually lost some through accidents—through the effects of the environment. The heritability for finger number is therefore close to zero. Yet it would be absurd to argue that environment is the cause of our having 10 fingers. We grow 10 fingers because we are genetically programmed to grow 10 fingers. It is the variation in finger number that is environmentally determined; the fact that we have 10 fingers is genetic. Paradoxically, therefore, the least heritable features of human nature may be the most genetically determined.

So, too, with intelligence. It cannot be right to say that Hermia’s intelligence is caused by her genes: it is obvious that you cannot become intelligent without food, parental care, teaching, or books. Yet in a sample of people who have all these advantages, the variation between who does well in exams and who does not could indeed be a matter of genes. In that sense, variation in intelligence can be genetic.

Through accident of geography, class, or money, most schools have pupils from similar backgrounds. By definition, they give these pupils similar teaching. Having therefore minimized the differences in environmental influences, the schools have unconsciously maximized the role of heredity: it is inevitable that the difference between the high-scoring and the low-scoring pupils must be set down to their genes, for that is just about all that is left to vary. Again, heritability is a measure of what is varying, not what is determining.

Likewise, in a true meritocracy, where all have equal opportunity and equal training, the best athletes will be the ones with the best genes. Heritability of athletic ability will approach 100 percent. In the opposite kind of society, where only the privileged few get sufficient food and the chance to train, background and opportunity will determine who wins the races. Heritability will be zero. Paradoxically, therefore, the more equal we make society, the higher heritability will be, and the more genes will matter.


I’ve labored the caveats deliberately before even mentioning the results of modern twin studies. The story of those studies begins in 1979, when there appeared in a Minneapolis newspaper an account of a pair of identical twin men from western Ohio reunited at the age of 40. Jim Springer and Jim Lewis had been reared apart in adopted families since they were a few weeks old. Intrigued, the psychologist Thomas Bouchard asked to meet them to record their similarities and differences. Within a month of their reencounter, Bouchard and his colleagues examined the Jim twins for a day and were astonished by the similarities. Though they had different hairstyles, their faces and voices were almost indistinguishable. Their medical histories were very similar: high blood pressure, hemorrhoids, migraines, “lazy eye,” chain-smoked Salem cigarettes, bitten nails, weight gain at the same age. As expected, their bodies showed remarkable similarity. But so did their minds. Both followed stock-car racing and disliked baseball. Both had carpentry workshops. Both had built a white seat around a tree trunk in the garden. They went to the same Florida beach on vacation. Some of the coincidences were, well, coincidences. Both had dogs named Toy. Both had wives named Betty. Both had divorced women named Linda. Both had named their first children James Alan (though one spelled it James Allen).

It occurred to Bouchard that maybe twins reared apart would turn out to be not just as similar but more similar than twins reared together. In the same family, differences might become exaggerated: one twin would start to do a little more of the talking and the other less, or something like that. This is now known to contain a germ of truth. In some respects, twins who were separated early in life have more similarities than twins separated at a later age.

The news reporter who had first written about the Jim twins interviewed Bouchard after his meeting with them, and the resulting article brought a flood of interest from the media. The twins appeared on the Tonight show, with Johnny Carson, and that was when things began to snowball. Twins started calling. Bouchard invited them to Minnesota and put them through a battery of physical and psychological tests, administered eventually by a team of 18 people. By the end of 1979, 12 pairs of reunited twins had contacted Bouchard. By the end of 1980, 21; a year after that, he had 39 pairs.

That was the year Susan Farber published a book disparaging all studies of identical twins reared apart as unreliable. The studies exaggerated similarities, ignored differences, and skated over the fact that many twins had spent many months together as infants before their adoption or had been reunited many months before being seen by scientists. Some of the studies, such as Cyril Burt’s, were perhaps even fabricated altogether. Farber’s book was seen as the last word on the matter, but Bouchard merely saw it as a challenge to do a flawless study. He was determined not to leave himself open to such accusations, and he carefully recorded everything about his pairs of twins. Anecdotes aside, he was gathering real, quantitative information on similarity. By the time he published, his data were all but impregnable to Farber’s criticisms. Not that this impressed the establishment. His critics still charged that he was proving nothing but his own assumptions. Of course these people resembled each other—they lived in similar middle-class suburbs of similar cities; they swam in the same cultural sea; they were taught the same western values.

All right, then, said Bouchard, and he set out to find fraternal (dizygotic) twins reared apart. These were people who had shared a womb as well as a western upbringing. If his critics were right, then they too should show remarkable similarities of mind. Do they?

Take religious fundamentalism. In a recent study Bouchard measured how fundamentalist individuals are by giving them questionnaires about their beliefs. The correlation between the resulting scores for identical twins reared apart is 62 percent; for fraternal twins reared apart it is just 2 percent. Bouchard repeats the exercise with a different questionnaire designed to elicit a broader measure of religiosity and still gets a strong result: 58 percent versus 27 percent. He then shows a similar contrast between sets of identical twins reared together and fraternal twins reared together. His colleague repeats the exercise with a different questionnaire designed to discover “right-wing attitudes.” Again there is a high correlation in identical twins reared apart (69 percent) and no correlation at all in fraternal twins reared apart. He gives the twins a different questionnaire that simply lists single phrases and asks for approval or disapproval: immigrants, death penalty, X-rated movies, etc. Those who reply no to immigrants, yes to the death penalty, and so on are judged more “right-wing.” The identical-apart correlation is 62 percent, the fraternal-apart correlation only 21 percent. Similar huge differences emerge from similar large studies in Australia.

Bouchard is not trying to prove that there is a “god” gene or an antiabortion gene. Nor is he trying to claim that the environment plays no part in determining details of religious observance. It is absurd to argue, for instance, Italians are Catholic and Libyans are Muslim because they possess different genes. He is simply claiming that, astonishingly, even in such a prototypically “cultural” thing as religion, the impact of genes cannot be ignored and can be measured. There is a partly heritable aspect of human nature, which might be called religiosity, and it is distinct from other attributes of personality (it correlates poorly with other measures of personality such as extroversion). This can be detected using simple questionnaires, and it predicts fairly well who will end up becoming a fundamentalist believer within any particular society.

Notice how even this one simple study refutes many of the objections raised by critics of behavior genetics. Many people argue that questionnaires are unreliable, crude measures of people’s real thoughts; but that simply makes these results conservative. The effects would probably be bigger if measurement error could be ruled out. Many argue that identical twins reared apart have not really lived such separate lives as is claimed. The twins have often been reunited for many years before the experiment is done. But if this is true, it will be just as true for the fraternal twins reared apart. The same response demolishes the frequent objection that Bouchard, by attracting self-selected twins to his studies, preferentially attracts those who are more similar to each other. But it is the differences between identical and fraternal twins that are revealing, not the absolute similarity. Others say you cannot separate nature from nurture, because they interact. True, but the fact that twins reared apart do not differ greatly from twins reared together suggests that such an interaction is less powerful than many believe.

In researching this book, I encountered a vitriolic opinion of Bouchard’s research among many people. Not content with making the long-since-answered arguments in the last paragraph, they would pointedly remind me to check where Bouchard got the funds for his research: the Pioneer Fund. This fund, founded in 1937 by a textile billionaire, is unashamedly in favor of eugenics. Its charter reads: “To conduct or aid in conducting study and research into the problems of heredity and eugenics in the human race generally and such study and such research in respect to animals and plants as may throw light upon heredity in man, and research and study into the problems of human race betterment with special reference to the people of the United States.” Based in New York, it is run by a board consisting mainly of aging war heroes and lawyers.

Their motive in supporting Bouchard’s research is presumably that they want to believe genes influence behavior, so they give money to a researcher who seems to be getting results, which support such a conclusion. Does this mean that Bouchard and all his many colleagues (not to mention the similar twin-studiers in Virginia, Australia, Holland, Sweden, and Britain) have faked their data to please their funders? Seems pretty far-fetched. Besides, you only have to meet Bouchard for a few minutes to realize that he is nobody’s patsy and nobody’s fool, let alone a raving determinist itching to unleash a new eugenics movement on the world. He takes money from the Pioneer Fund because it has no strings attached. “My rule is that if they don’t make any restrictions on me—what I think, what I write, what I do—I’ll accept their money.”

There is, of course, a problem with how such studies are reported. The headline “the gene for x” does much mischief, not least because of the reputation genes have garnered for being invincible bulldozers of all that stands in their path. However, the champions of nurture must bear some responsibility for creating this reputation in the first place, by equating genes with inevitability in the process of arguing that since behavior is not inevitable, genes cannot be involved. Champions of nurture repeatedly state that “the gene for x” means a gene that always and only causes behavior x; the champions of nature reply that they merely mean the gene increases the probability of behavior x, compared with other versions of the same gene. When the British twin-researcher Thalia Eley announced in 1999 that evidence from 1,500 identical versus fraternal twin pairs in Britain and Sweden suggested a strong genetic influence on whether an individual child would become a school bully, should she have complained or apologized when a reporter described her conclusion in the usual shorthand: “Bullying behaviour may be genetic”? The truer statement would be “Variations in bullying behavior may be genetic in typical western societies,” but few reporters can expect news editors to insert such caveats.

It is worth recalling how much of a shock the carefully controlled twin studies of the 1980s were when they first came out. Until then it was genuinely thought that differences in experience even among middle-class westerners would produce differences in personality with no help from the genes. The hypothesis on trial was not “all in the genes” but “not in the genes at all.” Here is a quotation from a leading textbook of personality psychology, published in 1981, the year Bouchard first had good data: “Imagine the enormous differences that would be found in personalities of twins with identical genetic endowments if they were raised in two different families.” That is what everybody thought, even Bouchard. “Look,” Bouchard says openly; “when I started, I did not believe these kinds of things could be influenced by genes. I was persuaded by the evidence.” The twin studies have caused a genuine revolution in the understanding of personality.

However, the very success of behavior genetics has been its undoing. Its results are boringly predictable: everything turns out to be heritable. Far from being able to parcel the world into genetic and environmental causes, as Galton wanted, twin studies have found almost everything to be equally strongly heritable. When Bouchard began, he expected to find that some measures of personality were more heritable than others. But at the end of two decades of such studies of separated twins in many countries, with larger and larger samples of twins, there is an unambiguous conclusion. For nearly all measures of personality, heritability is high in western society: identical twins raised apart are much more similar than fraternal twins raised apart. The difference between one individual and another owes more to differences in their genes than to factors in their family background.

Psychologists nowadays define personality in five dimensions—the so-called “big five” factors: openness, conscientiousness, extroversion, agreeableness, and neuroticism (OCEAN for short). Questionnaires can elicit personal scores for each of these dimensions, and they seem to vary independently. You can be open-minded (O), fussy (C), extrovert (E), jealous (A), and calm (N). In each case a little over 40 percent of the variation in personality is due to direct genetic factors, less than 10 percent due to shared environmental influences (i.e., mostly the family), and about 25 percent due to unique environmental influences experienced by the individual (everything from illness and accidents to the company he or she keeps at school). The remaining 25 percent or so is simply measurement error.

In a sense what these twin studies have proved is that the word “personality” means something. When you describe somebody as having a certain personality, you are intending to refer to some intrinsic part of their nature that is beyond the influence of other people—the content of their character, to borrow a famous phrase. By definition, you mean something unique to them. It is, however, counterintuitive after a century of Freudian certainties to find how little that intrinsic character is influenced by the families they grew up in.

In this respect, personality is about as heritable as body weight. The correlation between two siblings in weight, according to one study, is 34 percent. The similarity between parents and children is a little lower, 26 percent. How much of this similarity is due to the fact that they live together and eat similar food, and how much to the fact that they share many of the same genes? Well, identical twins reared in the same family have a correlation of 80 percent while fraternal twins reared together have only 43 percent similarity, which suggests that genes matter rather more than shared eating habits. What about adoptees? The correlation between adoptees and their adoptive parents is only 4 percent, and that between unrelated siblings in the same family is just 1 percent. By contrast, identical twins reared apart in different families are still 72 percent similar in weight.

Conclusion: weight is largely due to genes, not eating habits, so throw away the diet advice and let rip with the ice cream? Of course not. The study says nothing about the causes of weight; it only reveals something about the causes of differences in weight within a particular family. Given the same access to food, some people will put on more weight than others. People are getting fatter in western societies, not because their genes are changing but because they are eating more and taking less exercise. But when everybody has similar access to food, the ones who put on weight fastest will be the ones with certain genes. So variation in weight can be inherited, even while changes in the average can be environmental.

What kind of gene could cause personality to vary? A gene is a set of instructions for making a protein molecule. To leap from this epitome of digital simplicity to the complexity of personality sounds impossible. Yet it can now, for the first time, be done. The changes in genetic sequence that lead to changes in character are being found: the haystack is revealing its first few needles. Take the gene for a protein called brain-derived neurotrophic factor, or BDNF, on chromosome 11. It is a short gene, a chunk of DNA text just 1,335 letters long—exactly the same length as this paragraph, by good fortune. The gene spells out in four-letter code the complete recipe for a protein that acts as a sort of fertilizer in the brain encouraging the growth of neurons, and probably does much else besides. In most animals, the 192nd letter in the gene is G, but in some people it is A. About three-quarters of human genes carry the G version, the rest the A version. This minuscule difference, just one letter in a long paragraph, causes a slightly different protein to be built—with methionine instead of valine at the 66th position in the protein. Since everybody has two copies of each gene, that means there are three kinds of people in the world: those with two methionines in their BDNFs, those with two valines, and those with one of each. If you give people a questionnaire about their personality and simultaneously find out which kind of BDNF they have, you will find a striking effect. The met–mets are noticeably less neurotic than the val–mets, who are noticeably less neurotic than the val–vals.

The val–vals are the most, and the met–mets the least, depressed, self-conscious, anxious, and vulnerable—four of the six facets that make up the psychologists’ dimension of neuroticism. Of the other 12 facets of personality, only one (openness of feelings) shows any association. This gene, in other words, specifically affects neuroticism.

Do not get carried away. This finding accounts for only a small portion of the variation between people, perhaps 4 percent. It may prove to be a peculiarity of 257 families in Tecumseh, Michigan, where the study was done. It is most definitely not “the” neuroticism gene. But at least in Tecumseh it is a gene whose variation explains some of the personality differences between any two individuals and in a way that is consistent with the standard way of describing personality. It is also the first gene to be associated so strongly with depression; this fact gives a faint glimmer of medical hope for one of the least treatable and commonest disorders of modern life. The lesson I wish to draw from it is not that this particular gene will prove especially significant, but that it proves just how easy is the leap from a spelling change in a DNA code to a real difference in personality. Neither I nor anybody else can yet begin to tell you how or why such a tiny change results in a different personality, but that it does so seems almost certain. The appeal to incredulity beloved by some of the critics of behavior genetics—“genes are just recipes for proteins, not determinants of personality”—just will not wash. A change in a protein recipe can indeed result in a change in personality. There are other candidate genes emerging, too.

So it is not crazy to conclude that people differ in personality more if they have different genes than if they are reared in different families. Hermia is less like Helena, despite being raised with her, than Sebastian is like Viola even though they were raised apart. This might seem obvious to the point of banality. Any parent who has more than one child notices dramatic differences in personality and knows for sure that he or she did not put them there. But then parents are almost bound to notice innate differences because parents are holding the environment fairly constant by raising each child in the same family. The surprise of the studies of twins raised apart is that they seem to show, even when the environments are varied somewhat, the differences in personality are still mostly innate. Even when the family environment does vary, it leaves no mark on personality. This conclusion emerges most starkly from the study of twins, but it is fully supported by other studies of adoption and of the relations of twins and adoptees.

The effect of being reared in the same home is negligible for many psychological traits.


The shared environment plays only a small and non-significant role in the creation of personality differences in adults.

Quickly but imperceptibly, statements like this seem to evolve into the assertion that families do not matter. Go ahead, neglect your kids, the logic seems to follow; their personality will not be affected. Some blame the researchers themselves for leaving this impression. Read the small print, however, and you will always find careful denials of such a fallacy. A happy family gives you other things than personality—things like happiness. Families do matter for personality; a child desperately needs to be reared in a family in order to develop her personality. So long as she does have a family to grow up within, it does not terribly matter whether the family is big or small, rich or poor, gregarious or solitary, old or young. A family is a bit like vitamin C: you need it or you will become ill, but once you have it, consuming extra does not make you healthier.

For those attached to the idea of the meritocracy, this is an encouraging discovery. It means there is no excuse for discriminating against people from underprivileged backgrounds, or to be wary of people brought up in unusual families. A disadvantaged childhood does not condemn a person to a certain personality. Environmental determinism is at least as heartless a creed as genetic determinism, a theme I shall have cause to revisit throughout this book. So it is lucky we do not have to believe in either.

There is a criticism to be made of twin studies of personality, one that I shall weave into my argument that genes are the agents of nurture at least as much as they are the agents of nature. The criticism rests on the fact that heritability depends entirely on context. The heritability of personality may be high in a group of middle-class Americans who have experienced equivalent, even identical, patterns of nurture. But throw a few orphans from Sudan or the offspring of headhunters from New Guinea into the sample and the heritability of personality would probably drop rather fast: now environment would matter. Hold the environment constant and it’s the genes that vary: what a surprise! “I can prove in a court of law,” says Tim Tully, who studies the genes of memory but has no time for twin studies, “that heritability has nothing to do with biology.” To the extent, therefore, that researchers studying twins try to suggest that the measurement of heritability is an end in itself, they are deluding themselves. And having once produced surprisingly strong evidence that genes do affect personality, it is not clear what they go on to do. Twin studies alone are notoriously unhelpful at revealing which actual genes are involved.

Here’s why. Heritability is usually highest for those features of human nature caused by many genes rather than by the action of single genes. And the more genes are involved, the more the heritability is actually caused by the side effects of genes rather than the direct effect. Criminality, for instance, is quite highly heritable: adopted children end up with a criminal record which looks a lot more like that of their biological parents than like that of their foster parents. Why? Not because there are specific genes for criminality, but because there are specific personalities that get into trouble with the law and those personalities are heritable. As Eric Turkheimer, a researcher who studies twins, puts it, “Does anyone really suppose that unintelligent, unattractive, greedy, impulsive, emotionally unstable, or alcoholic people are no more likely than anyone else to become criminals or that any of these characteristics could be completely independent of genetic endowment?”


Despite the sweeping successes of twin studies, a few features of human behavior prove to be less heritable. The sense of humor shows low heritability: adopted siblings seem to have quite similar senses of humor, while separated twins have rather different ones. People’s food preferences seem to be barely heritable—you get your food preferences from your early experience, not your genes (so do rats). Social and political attitudes show a strong influence from the shared environment—liberal or conservative parents seem to be able to pass on their preferences to their children. Religious affiliation, too, is passed on culturally, rather than genetically, though not religious fervor.

What about intelligence? The debate about the heritability of IQ has been scarred by controversy since its inception. The first IQ tests were crude and culturally biased. In the 1920s, convinced that intelligence was largely hereditary and alarmed at the thought of excessive breeding by stupid people, governments in the United States and many European countries began to sterilize mental defectives to prevent them from passing on their genes. But in the 1960s came a sudden revolution, as in so many other debates. From then on, even the assertion of heritable IQ led to vitriolic campaigns of denunciation, assaults on your reputation and demands for your dismissal. The first to suffer this treatment was Arthur Jensen in 1969, following his article in the Harvard Educational Review. By the 1990s, the argument that society was segregating itself by assortative mating along intellectual and therefore racial lines—asserted in The Bell Curve by Richard Herrnstein and Charles Murray—provoked another wave of rage among academics and journalists.

Yet I suspect that if you took a poll of ordinary people, they would hardly have changed their views over a century. Most people believe in “intelligence”—a natural aptitude or lack of it for intellectual pursuits. The more children they have, the more they believe in it. This does not stop them from also believing in coaxing it out of the gifted and coaching it into the ungifted through education. But they think that there is something innate.

The studies of twins reared apart or together unambiguously support the idea that although some people are good at some things and others are good at other things, there is such a thing as unitary intelligence. That is to say, most measures of intelligence correlate with each other. People who are good at general knowledge tests or vocabulary tests are usually good at abstract reasoning or at tasks that involve completing number series. This was first noticed a century ago by a follower of Galton’s, the statistician Charles Spearman, who dubbed the common factor g for general intelligence. Today, a measure of g derived from correlating different IQ tests remains a powerful predictor of how well a child will do at school. There has been more research on g than on any other subject in psychology. Theories of multiple intelligence come and go, but the notion of correlated intelligence just will not go away.

What is g? Something that appears so real in statistical tests must surely have a physical manifestation in the brain. Is it something to do with speed of thought or size of brain, or is it something subtler? The first thing to be said is that the search for the genes of g has been a huge disappointment. None of the genes that are capable of causing mental retardation when broken prove to have any effect on intelligence when altered more subtly. Searching at random through the genes of intelligent people to find ways in which they consistently differ from genes of normal people has so far turned up just one decent statistical correlation (for the IGF2R gene on chromosome 6) and more than 2,000 no-shows. This may just mean that the haystack is too big and the needles too small. Candidate genes, such as the PLP gene that seems to affect speed of neuronal transmission, have proved capable of explaining only a small degree of reaction time and do not correlate well with g: the speedy-brain theory of intelligence does not look promising.

The one physical feature that does clearly predict intelligence is brain size. The correlation between brain volume and IQ is about 40 percent, a number that leaves much room for the small-brained genius and the big-brained dullard but is still a strong correlation. Brains are composed of white matter and gray matter. When, in 2001, brain scanners reached the stage that people could be compared for the amount of gray matter in their brains, two separate studies in Holland and Finland found a high correlation between g and volume of gray matter, especially in certain parts of the brain. Both also found a huge correlation between identical twins in volume of gray matter: 95 percent. Fraternal twins had only a 50 percent correlation. These figures indicate something that is under almost pure genetic control, leaving very little room for environmental influence. Gray matter volume must be “due completely to genetic factors and not to environmental factors” in the words of Danielle Posthuma, the Dutch researcher. These studies bring us no closer to the actual genes of intelligence, but they leave little doubt that the genes are there. Gray matter consists of the bodies of neurons, and the new correlation implies that clever people may literally have more neurons, or more connections between neurons, than normal people do. After the discovery of the role of the ASPM gene in determining brain size through neuron number (), it is beginning to look as if some of the genes of g will soon be found.

However, g is not everything. Twin studies of intelligence also reveal a role for the environment. Unlike personality, intelligence does seem to receive a strong influence from the family. Studies of the heritability of IQ in twins, adoptees, and combinations of the two have all gradually converged on the same conclusion. IQ is approximately 50 percent “additively genetic”; 25 percent is influenced by the shared environment; and 25 percent influenced by environmental factors unique to the individual. Intelligence therefore stands out from personality in being much more susceptible to family influence. Living in an intellectual home does make you more likely to become an intellectual.

However, these average figures conceal two very much more interesting features. First, you can find samples of people in which variation in IQ is much more environmental and much less genetic than the average. Eric Turkheimer found that the heritability of IQ depends strongly on socioeconomic status. In a sample of 350 pairs of twins, many of whom had been raised in extreme poverty, there emerged a clear difference between the richest and the poorest. Among the poorest children practically all the variability between individual IQ scores was accounted for by shared environment and none by genetic type; in the richer families, the opposite was true. In other words, living on a few thousand dollars a year can severely affect your intelligence for the worse. But living on $40,000 a year or $400,000 a year makes little difference.

This is a finding with obvious significance for policy. It implies that raising the safety net of the poorest does more to equalize opportunity than reducing inequality in the middle classes. It is dramatic confirmation of the truth I alluded to earlier: that even when variation in achievement is explained entirely by genes, this does not mean the environment does not matter. The reason you find such strong genetic effects in most samples is that most of the people in the samples live in adequately happy, supportive, affluent families. If they did not, they would suffer enormously. It is a point that is almost certainly true of personality, too. Your parents may not have been able to alter your adult personality by being a little bit strict. But you can be sure that they would have done so if they had locked you in your room 10 hours a day for weeks on end.

Recall the heritability of weight. In a western society, with ample access to food, those who put on weight faster will be the ones with the genes that nudge them into eating more. But in a desolate part of the Sudan, say, or Burma, where extreme poverty is rife and famine just around the corner for many people, everybody is hungry and the fat people are probably the rich ones. Here variation in weight is caused by the environment, not the genes. In the jargon of the scientist, the effect of the environment is nonlinear: at the extremes, it has drastic effects. But in the moderate middle, a small change in the environment has a negligible effect.

The second surprise hidden in the average figures is that the influence of genes increases and the influence of shared environment gradually disappears with age. The older you grow, the less your family background predicts your IQ and the better your genes predict it. An orphan of brilliant parents adopted into a family of dullards might do poorly at school but by middle age could end up a brilliant professor of quantum mechanics. An orphan of dullard parents, reared in a family of Nobel Prize-winners, might do well at school but by middle age may be working in a job that requires little reading or little deep thought.

Numerically, the contribution of “shared environment” to variation in IQ in a western society is roughly 40 percent in people younger than 20. It then falls rapidly to zero in older age groups. Conversely, the contribution of genes to explaining variation in IQ rises from 20 percent in infancy to 40 percent in childhood to 60 percent in adults and maybe even 80 percent in people past middle age. In other words, the effect of being reared in the same environment as somebody else is influential while you are still in that environment but does not endure beyond the period of shared rearing. Adoptive siblings do have partly similar IQs while living together. But as adults their IQs are wholly uncorrelated. By adulthood, intelligence is like personality: mostly inherited, partly influenced by factors unique to the individual, and very little affected by the family you grew up in. This is a counterintuitive discovery exploding the old idea that genes come early and nurture late.

What this seems to reflect is that the intellectual experience of a child is generated by others. An adult, by contrast, generates his or her own intellectual challenges. The “environment” is not some real, inflexible thing: it is a unique set of influences actively chosen by the actor himself or herself. Having a certain set of genes predisposes a person to experience a certain environment. Having “athletic” genes makes you want to practice a sport; having “intellectual” genes makes you seek out intellectual activities. The genes are agents of nurture.

As a parallel, how do genes affect weight? Presumably through controlling appetite. In an affluent society, those who gain most weight are hungrier and so eat more. The difference between a genetically fat and a genetically thin westerner lies in the fact that the first is more likely to buy ice cream. Is it the gene or the ice cream that causes fatness? Well, it is obviously both. The genes are causing the individual to go out and expose himself to an environmental factor, in this case ice cream. Surely it is bound to be the same in the case of intelligence. The genes are likely to be affecting appetite more than aptitude. They do not make you intelligent; they make you more likely to enjoy learning. Because you enjoy it, you spend more time doing it and you grow more clever. Nature can act only via nurture. It can act only by nudging people to seek out the environmental influences that will satisfy their appetites. The environment acts as a multiplier of small genetic differences, pushing athletic children toward the sports that reward them and pushing bright children toward the books that reward them.

The main conclusion in behavior genetics is counterintuitive in the extreme. It tells you that nature plays a role in determining personality, intelligence, and health—that genes matter. But it does not tell you that this role is at the expense of nurture. If anything, it proves rather dramatically that nurture matters just as much, though it is inevitably less good at discerning how (there is no environmental equivalent to the natural experiment created by identical and fraternal twins). Galton was utterly wrong in one important respect. Nature does not prevail over nurture; they do not compete; they are not rivals; it is not nature versus nurture at all.

Paradoxically, if western society has reached the point where the heritability of intelligence is so high, then it means we have achieved something approaching a meritocracy, where your background does not matter. But this also reveals something truly surprising about genes. They do vary within the normal range of human behavior. You might expect that genes would be like vitamin C or families—they become limiting only when they are malfunctional. So broken genes might cause rare broken minds, just as they cause rare diseases. Severe depression, mental illness, or mental disability might be caused by rare variations in genes, just as all these things could be caused by a rare and bizarre upbringing. This would then be the perfect utopia in which, so long as all had normal genes and a normal family, everybody would have the same potential personality and intelligence. The details would then come down to accident or circumstance.

But it is not like that. Behavior genetics reveals very starkly that there are genetic differences which are common and which affect our personalities within the range of normal human experience. There are val–vals and met–mets among us, not just for the BDNF gene but for many other genes affecting personality, intelligence, and other aspects of the mind. Just as some people are genetically better at gaining muscle strength than others, according to which version of the ACE gene they possess on chromosome 17, so some people are genetically more able to absorb education according to which versions they possess of some unknown genes. These mutations are not rare; they are common.

From the point of view of the evolutionary biologist this is a scandal. Why is there so much “normal” genetic variation, or, to give it its proper name, polymorphism? Surely, the “clever” variants on genes would gradually drive the “dull” ones to extinction, and the phlegmatic ones would drive out the excitable ones. One kind must inevitably be superior to the other in providing survival or mating advantages. One kind must therefore endow its owner with greater ability to become a fecund ancestor. Yet there is no evidence of genes going extinct in this way. There seems to be a sort of happy coexistence of different versions of genes within the human population.

Enigmatically, there is more genetic variation in the human population than science has a right to expect. Behavior genetics, remember, does not discover what determines behavior; it discovers what varies. And the answer is that genes vary. Contrary to popular opinion, most scientists love enigmas. They are in the business of finding new mysteries, not cataloging facts. The white-coated ones in the labs live in the dim hope of finding a really fine conundrum or paradox. And here is a fine one.

There are plenty of theories to explain the enigma, though none that is entirely satisfactory. Perhaps we human beings have simply relaxed natural selection so much by keeping ourselves alive with technology that our mutations have proliferated. But then why is the same variation present in other animals? Perhaps there is a delicate form of balancing selection that always favors the rare variants, thus keeping rare genes from going extinct. This idea certainly seems to explain variability in the immune system because disease favors rare versions of genes by attacking the common ones, but it is not immediately obvious why this should preserve polymorphism in personality. Perhaps mate choice encourages diversity. Or perhaps some new idea, as yet unheard of, will explain the phenomenon. Rival explanations for polymorphism were already causing bitter divisions among evolutionists in the 1930s, and they are not settled yet.


Normally at this point, a book about behavior genetics would lurch into vitriolic criticism of one side or the other in the nature–nurture argument. Either I would argue that twin studies are dubious in motive, flawed in design, idiotic in interpretation, and likely to encourage fascism and fatalism, or I would argue that they are a moderate and sensible corrective to the crazy dogma of the blank slate, which has forced us to try to believe there is no such thing as innate personality or mental talent and everything is the fault of society.

I have some sympathy with both views. But I am resolutely resisting the temptation to go in for this kind of comment, which has bedeviled the nature–nurture debate. The philosopher Janet Radcliffe-Richards catches the gist nicely: “If you follow up in detail any of the claims about what opponents are supposed to have said in this debate, you may be quite startled by the extent of misquoting, quoting out of context, looking for the worst interpretation of what is said, and flagrant misrepresentation that goes on.” In my experience, scientists are most often wrong when they are being critical of each other. When they assert that their preferred idea is true and another idea is therefore false, they can be right about the first and wrong about the second: both ideas can be partly true. Like explorers arguing over which tributary is the source of the Nile, they are missing the point that the Nile needs both tributaries or it would be a creek. Any geneticist who says he has found an influence for genes and therefore there is no role for the environment is talking bunk. And any nurturist who says he has found an environmental factor and therefore there is no role for genes is equally talking bunk.

The story of IQ contains a very clear example of this phenomenon. Called the Flynn effect, after its discoverer, James Flynn, it is the remarkable fact that average IQ scores are rising steadily at the rate of at least five points per decade. This shows that the environment does influence IQ; it implies that compared with our grandparents we are all teetering on the brink of genius, which seems unlikely. Nonetheless, something about modern life, whether it is nutrition, education, or mental stimulation, is making each generation better at IQ tests than its parents. Therefore, one or two nurturists (but not Flynn) argued triumphantly, the role of genes must be smaller than had been thought. But the analogy of height shows that this is a non sequitur. Thanks to better nutrition, each generation is taller than its parents, but nobody would argue that therefore height is less genetic than was thought. In fact, because more people now reach their full potential stature, the heritability of variation in height is probably increasing.

Flynn himself now thinks he understands his own effect by referring to the way appetite reinforces aptitude. During the twentieth century society progressively made it more rewarding for children to seek intellectual, school-based achievement. Thus rewarded, they responded by exercising the relevant parts of the brain more. By analogy, the invention of basketball has encouraged more children to practice their basketball skills. As a result each generation is better at basketball. Two identical twins resemble each other in their basketball ability because they started out with a similar aptitude, which gave them the same appetite for the game, which brought them the same opportunities for practice. It is aptitude and appetite, not one or the other. An identical twin, having the same genes as his co-twin, therefore goes out and gets himself the same experience.


Toward the end of his long life, Francis Galton succumbed to a temptation that befalls many prominent men. He wrote a utopia. Like all descriptions of the ideal society, from Plato’s and Thomas More’s onward, it depicts the sort of totalitarian state that nobody in his right mind would want to inhabit. It is a useful reminder of a theme that will recur throughout this book: pluralism in the causes of human nature is vital. Galton was right about the strength of heritable factors in human nature, but wrong to think that nurture therefore does not matter.

Galton wrote his book in 1910, when he was in his eighties. It was called Kantsaywhere, and it purports to be the diary of a man named Donoghue, a professor of vital statistics. Donoghue arrives in Kantsaywhere, a colony governed by a council along entirely eugenic lines. He meets Miss Augusta Allfancy, who is about to take an honors examination at Eugenics College.

Kantsaywhere’s eugenics policies were invented by a Mr. Neverwas, who left his money to be used for the improvement of the human stock. Those who do well in the eugenic exams by having heritable gifts are rewarded in various ways; those who merely pass are allowed to breed in only a small way; those who fail are sent to labor colonies, where their duties are not especially onerous but they must remain celibate. Propagation by the unfit is a crime against the state. Donoghue accompanies Augusta to various parties where she meets potential mates, for she will marry at 22.

Fortunately for Galton, Methuen rejected the novel for publication and his great-niece Eva managed to keep it from wide circulation. She at least realized how embarrassing it was. She could never have realized that Galton’s controlled society would also be horribly prophetic for the twentieth century.

Previous: 2. A Plethora of Instincts
Next: 4. The Madness of Causes