• Email
  • Single Page
  • Print

What’s Not in Your Genes

The same problem arises in Ridley’s chapter on mental illness. Again his facts are right: schizophrenia is about equally common in all races and there is some evidence that it’s partly genetic. But Ridley leaps to big conclusions. First he suggests that the mutations causing schizophrenia are ancient, “having occurred before the ancestors of all non-Africans left Africa.” But the fact that a disease is as common in Eskimos as in bushmen tells us nothing about the age of the mutations involved. Worse, Ridley can’t resist reciting an adaptive tale to explain the persistence of these mutations. Maybe, he says, natural selection keeps them around because carrying a few mutations is good for you (making you a genius), while carrying many is bad (making you schizophrenic). While this idea might seem to explain the apparent association between genius and mental illness within families, it is demonstrably false. A well-known result in population genetics shows that the kind of natural selection Ridley invokes here cannot maintain genetic variation in populations.

Similar problems appear as Ridley turns to the debate over nature and nurture. To understand these, we need to define more clearly what the debate is about, since the controversy has meant rather different things to different people. To some, the question has been whether genes have a part in determining human capacities, such as language or the ability to act morally. To others, it has been whether differences between groups—most notoriously, human races—are in the genes. But the traditional question at the heart of the debate—the one that has consumed everyone from Galton to Gould and that has inspired a century of heated debate over twin studies, genetic determinism, IQ, and race—concerns differences between individuals. It is this: Given the differences that we see among individuals in a trait such as intelligence, how much is due to genes and how much to environment? Biologists and psychologists express the answer to this question in terms of “heritability,” the percentage of variation in a trait caused by genes. If none of the differences that we see in IQ were genetic, its heritability would be zero percent; if all of the differences were genetic, its heritability would be 100 percent; and if half of the differences were genetic, its heritability would be 50 percent, and so on. Heritability is measured by assessing the similarity of individuals who have a known degree of genetic relatedness, e.g., by asking how similar twins are or how closely children resemble their parents. Measurement of heritability does not, therefore, require that we know anything about which genes affect a trait. (Indeed it doesn’t even require that we know that genes are made of DNA.)

Though Ridley devotes much attention to heritability, he doesn’t seem completely clear about it. He claims, for instance, that “heritability is usually highest for those features of human nature caused by many genes rather than by the action of single genes.” But since we have no idea how many genes underlie virtually any aspect of human nature, it’s hard to see how we could possibly know this. Also, while a trait that’s affected by many genes might sound “more genetic,” there is no known relationship between the number of genes affecting a trait and heritability in any species.

Despite all this, it could still be true that recent discoveries about the genome have blurred the distinction between nature and nurture. If so, we might happily overlook the above sorts of slips. Who cares, after all, that Darwin got his physical theory of inheritance wrong (basing it on a nonexistent entity called the “gemmule”)? He got the big theory—natural selection—right. And in science getting the big question right more than makes up for getting the small ones wrong. It seems to me, though, that Ridley’s attempt to recast the nature–nurture debate doesn’t succeed. Much as I wish that he’d laid the controversy to rest, I don’t think he has. There are at least two problems. The first is that he seems to have the right answer to the wrong question.

It should be clear that the traditional question “How heritable is a trait?” is different from Ridley’s question “Do genes respond to experience?” The first question is statistical. It asks about the percentage of variation in, say, IQ, that arises from inherited differences among individuals (do some parents pass on smart genes to their kids?) versus the percentage that arises from environmental differences (do some parents pass on books to their children?). The second question is mechanistic. It asks about how genes behave within individuals. But knowing that my DNA switches on and off during my lifetime tells me nothing about whether you and I were born with different DNA, much less how much these differences affect our IQs. Nor does it render this question meaningless. The fact that genes respond to experience is certainly interesting and important. It’s the sort of fact we pay neurobiologists to find out. But it’s the wrong kind of fact to settle the nature–nurture debate.

This is an important point and we need to look at it more carefully. To do so, consider a trait that’s less controversial than IQ and a species that’s less politically sensitive than Homo sapiens—the height of a plant. Imagine a population of sunflowers in which all plants are genetically identical: all carry exactly the same DNA. Imagine further that none of the genes in these plants responds to experience. Though different sunflowers will inevitably grow to different heights (some may get more water than others), the heritability of height in this population is zero. None of the differences in height is caused by differences in the genes inherited among individuals (since there were no genetic differences).

Now imagine that all individuals are again genetically identical but that all genes respond to experience, switching on and off in response to subtle differences in water and light, among other factors. Ridley would claim that nurture acts via nature here. But the heritability of height hasn’t changed one bit. It is still precisely zero. It’s still true that none of the differences in height arise from differences in the genes inherited among plants. (Since it’s still true that there were no differences in the genes inherited among plants.) In the more complicated case where genetic differences among individuals do exist, the lesson stays the same: the switching on and off of genes has no necessary effect on heritability. It might go up, down, or remain unchanged.2

What’s going on here? The problem is that the question of nature versus nurture asks about the effects of inherited differences among individuals, while Ridley’s answer talks about the behavior of genes within individuals. Because the question of nature versus nurture concerns the quantitative effects of inherited differences among individuals, it is statistical; because Ridley’s answer concerns the molecular behavior of genes within individuals, it is mechanistic. There is, then, a mismatch between the levels at which the traditional question and Ridley’s answer are posed. And, unfortunately, a statistical question cannot have a molecular answer.

This is not to say that molecules don’t physically underlie traits, genes, and even environments. Of course they do. In any given person, IQ or height or anything else you can point to reflects the physical arrangement of molecules. But it is to say that molecules don’t provide the right level of abstraction at which to talk about the statistical contributions of nature and nurture. If I ask about the gross national product and you answer in terms of the molecular makeup of money, something has gone wrong. The findings that Ridley surveys might, I suppose, force us to abandon some particularly simple ways of talking about nature and nurture (heritability isn’t the only way of talking), but I don’t think they touch the hard problem at the historical heart of the controversy.

Ridley’s apparent failure to see this leads him to partly mislocate the source of our troubles with nature and nurture. The IQ debate hasn’t proved interminable because of any philosophical problem with heritability3 ; and it certainly hasn’t proved interminable because we lacked a key piece of data about how genes act in brains. It has dragged on forever because we cannot perform the required critical experiment with humans. For ethical and practical reasons, we cannot force a large set of parents and children to live in a common environment, allowing us to compare the IQs of parent and child, thereby computing a percentage of heritability.4 We can, however, do the analogous kind of experiment in other species. This is why there’s a nature–nurture debate for IQ in human beings and not for milk yield in dairy cows.

The second problem with Ridley’s view is even more important and has nothing to do with heritability. If all you care about is that organisms respond to experience, it doesn’t matter if this response involves the switching on and off of DNA or not. Any molecular mechanism will do. Organisms might, for instance, respond to experience by folding their proteins into different three-dimensional shapes in different environments—a process that has nothing to do with DNA but that could still change the way an organism looks or acts. The fact that DNA is sometimes involved in response to experience is interesting but it doesn’t accomplish anything that a non-DNA mechanism doesn’t accomplish. Either way, organisms still “absorb experience,” reshaping themselves in response to the environment. Indeed so far as I can tell, the only real difference between the DNA and non-DNA mechanisms is that one—because it happens to involve the genetic material—tempts us into talk about recasting the nature–nurture debate while the other doesn’t.

Despite these problems, you might argue that Ridley has at least shown that we are not slaves to our genes. An understanding of how genes respond to experience might even open the door to novel medical and other interventions that could, say, boost a child’s intelligence. (Knowing about Johnny’s genes, I might also know that reading to him at age six months turns on gene G in neuron N.) This may well be true. But there is nothing fundamentally new here. We have long been able to intervene between genes and traits. (Do you have genes for skinny arms? Lift weights.) As Ridley himself emphasizes, genetic endowment has rarely meant a fixed fate and this was never the locus of the nature–nurture debate.


One of the most interesting questions about Nature via Nurture is: Why does Ridley reach for the wrong level of analysis, confounding statistics and mechanisms? Why does he exaggerate the significance of certain discoveries about how genes work? Though this requires some speculation, the answer seems plain. Ridley, a self-styled champion of “techno-optimism,”5 seems to have succumbed to genome hype. Nature via Nurture is riddled with breathless claims about the wonders of the genome. We are assured that “the genome has indeed changed everything,” are promised “bizarre stories from the deepest recesses of the genome,” and are alerted to a “startling new truth” that has just emerged “from the human genome.” Ridley’s enthusiasm even leads him into subtly revisionist history. Despite the hype, the fact is that almost none of the findings he discusses has anything to do with the much-publicized Human Genome Project. I doubt the lay reader of Ridley’s book would figure this out.6

The point is that Ridley, like the rest of us, is immersed in a scientific culture, both popular and professional, that encourages the facile assumption that all questions about organisms will emerge from Big Science working on small molecules. But there is no principled reason why the answers to all questions about biology, much less human nature, will be explained by the genome or molecular mechanism. This should be neither surprising nor controversial. There are many questions in all fields of science that have little to do with underlying physical detail. The entire discipline of cognitive science, including much of linguistics, is built on the premise that, though minds are made of brains, we can learn a good deal about the mind—e.g., how we go about making new sentences—without worrying over the details of neurobiology. Software, not hardware, provides the right level of abstraction. It follows, of course, that the answers to such higher-level questions do not change when the underlying mechanisms are discovered. The answer to “How hot is Mars?” did not change when statistical mechanics showed that heat is molecular motion. And the answers to “How heritable are plant height, milk yield, Huntington’s chorea, and IQ?” do not change once we know that genes switch on and off inside organisms.

It is probably true, however, that it is harder to resist this kind of reflexive reductionism in biology than in any other science today. Much of the reason is momentum. Biology has made extraordinary progress over the last century by following a relentlessly reductionist road. Furthermore biology, unlike physics, possesses little mathematical theory of the sort that floats defiantly above physical detail and that inevitably inspires respect for autonomous principles or laws. While physics has enjoyed a long tradition that reveres the abstract and disdains the mechanistic,7 this attitude is almost unimaginable in biology. Ridley’s book, with its alluring mix of undeniable good intention and deft description, will surely add to the fascination with genome and mechanism. And to some extent this is fine. But it’s important not to lose sight of a subtler message—facts can sometimes be both fascinating and irrelevant.

  1. 2

    Biologists might think Ridley is concerned with “genotype-by-environment interaction,” where genes have different effects in different environments. But in one of the oddest passages in his book, Ridley announces: “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 any case, genotype-by-environment interaction can occur whether genes turn on and off or not; and heritability remains measurable regardless.

  2. 3

    This isn’t to say that there aren’t any, only that they haven’t been the chief barriers to progress.

  3. 4

    This experiment would likely yield an upper bound on heritability. Twin studies are harder to interpret than this experiment. Note also that choosing a common environment for humans would involve many arbitrary decisions and would not eliminate important nongenetic differences among the families in which children grow up.

  4. 6

    Ridley’s previous book, Genome (HarperCollins, 1999), lays it on even thicker: “Some time in the year 2000, we shall probably have a rough first draft of the complete human genome. In just a few short years we will have moved from knowing almost nothing about our genes to knowing everything. I genuinely believe that we are living through the greatest intellectual moment in history. Bar none.”

  5. 7

    This tradition is best represented by the physicist Pierre Duhem. See The Aim and Structure of Physical Theory (Princeton University Press, 1954), especially Part I.

  • Email
  • Single Page
  • Print