A few commentators have welcomed genetic enhancement as just the latest step in the struggle to improve human life. Many more are appalled. They warn that it is a Faustian grab at divine powers that will never be used wisely by us mortals. They worry that it will spawn the ultimate inequality, a genetic caste system. In his book Our Posthuman Future (just released in paperback), the conservative thinker Francis Fukuyama warns that genetic enhancement will change human nature itself and corrode the notion of a common humanity that undergirds the social order. Bill McKibben, writing from the political left, raises similar concerns in his new jeremiad Enough: Staying Human in an Engineered Age.
Whether they welcome or decry it, almost everyone agrees that genetic enhancement is inevitable if research proceeds on its current course. In America, genetic enhancement is a major concern of the president's Council on Bioethics; its chairman, Leon Kass, and several of its members, including Fukuyama, are outspoken worriers.
As it happens, some kinds of genetic enhancement are already here. Anyone who has been turned down for a date has been a victim of the human drive to exert control over half the genes of one's future children. And it is already possible to test embryos conceived in vitro and select ones that are free of genetic defects such as cystic fibrosis.
But when it comes to direct genetic enhancement - engineering babies with genes for desirable traits - there are many reasons to be sceptical. Not only is genetic enhancement not inevitable, but it is not particularly likely in our lifetimes. This skepticism comes from three sources: the limits of futurology, the science of behavioural genetics, and human nature itself.
The history of the future should make us raise an eyebrow whenever the experts tell us how we will live 10, 20, or 50 years from now. Not long ago we were assured that by the turn of the century we would live in domed cities, commute by jet-pack, and clean our homes with nuclear-powered vacuum cleaners wielded by robot maids. More recently we were promised the paperless office, interactive television, the internet refrigerator, and the end of bricks-and-mortar retail. It's not just that these developments have not yet happened, many of them, like domed cities, never will happen. Even in mundane cases, technological progress is far from inexorable. Air travel, for example, is barely faster or more comfortable today than it was when commercial jets were introduced 50 years ago.
Why are technological predictions usually wrong? Many futurologists write as if current progress can be extrapolated indefinitely - the fallacy of climbing trees to get to the moon. They routinely underestimate the number of things that have to go right for a development to change our lives. It takes more than a single eureka!; it takes a large number of more boring discoveries, together with the psychological and sociological imponderables that make people adopt some invention en masse. Who could have predicted the videophones of the 1960s would sink like a stone while the text messaging of the 1990s would become a teenage craze?
Finally, futurologists tend to focus their fantasies on the benefits of a new technology, whereas actual users weigh both the benefits and the costs. Do you really want to install software upgrades on your refrigerator or reboot it when it crashes?
Many prognosticators assume that we are in the midst of discovering genes for talents such as mathematical giftedness, musical talent and athletic prowess. The reality is very different. The Achilles heel of genetic enhancement will be the rarity of single genes with consistent beneficial effects.
Behavioural genetics has uncovered a paradox. We know that tens of thousands of genes working together have a large effect on the mind. Twin studies show that identical twins (who share all their genes) are more similar than fraternal twins (who share half their genes, among those that vary from person to person), who in turn are more similar than adopted siblings (who share even fewer of the varying genes). Adoption studies show that children tend to resemble their biological relatives in personality and intelligence more than they resemble their adopted relatives.
But these are effects of sharing an entire genome, or half of one. The effects of a single gene are much harder to show. Geneticists have failed to find single genes that consistently cause schizophrenia, autism or manic-depressive disorder, even though there is overwhelming evidence that these conditions are substantially heritable. And if we can't find a gene for schizophrenia, we're even less likely to find one for humour, musical talent, or likeability, because it's easier to disrupt a complex system with a single defective part than to improve it by adding a single beneficial one. The 1998 report of a gene that was correlated with a four-point advantage in IQ ( intelligence quotient) was recently withdrawn because it did not replicate in a larger sample - a common fate for putative single gene discoveries.
So don't hold your breath for the literary creativity gene or the musical talent gene. The human brain is not a bag of traits with one gene for each trait. Neural development is a staggeringly complex process guided by many genes interacting in feedback loops. The effect of one gene and the effect of a second gene don't produce the sum of their effects when they're simultaneously present. The pattern of expression of genes (when they are turned on or off by proteins and other signals) is as important as which genes are present.
Even when genes should be at their most predictable - in identical twins, who share all their genes, and hence all the interactions among their genes -we don't have foregone conclusions. Identical twins reared together (who share not only their genes but most of their environments) are imperfectly correlated in personality measures such as extroversion and neuroticism. The correlations, to be sure, are much larger than those for fraternal twins or unrelated people, but they are seldom greater than 50 per cent. This tells us there is an enormous role for chance in the development of a human being.
It gets worse. Most genes have multiple effects, and evolution selects the ones that achieve the best compromise among the positive and the negative ones. Take the most famous candidate for genetic enhancement: the mice that were given extra copies of the NMDA receptor, which is critical to learning and memory. These poster mice did learn mazes more quickly, but they also turned out to be hypersensitive to pain. Closer to home, there is a candidate gene in humans that appears to be correlated with a 10-point boost in IQ. But it is also associated with a 10 per cent chance of developing torsion dystonia, which can confine the sufferer to a wheelchair with uncontrollable muscle spasms.
This places steep ethical impediments to research on human enhancement. Even if some day it might be possible, could you get there from here? How can scientists try out different genes to enhance the minds of babies given that many of them could have terrible side effects?
Genetic enhancement faces another problem: most traits are desirable at intermediate values. Wallis Simpson said that you can't be too rich or too thin, but other traits don't work that way. Take aggressiveness. Parents don't want their children to be punching bags or doormats, but they also don't want Attila the Hun either. Most want their children to face life with confidence rather than sitting at home cowering in fear, but they don't want a reckless daredevil out of Jackass. So even if a gene had some consistent effect, whether the effect was desirable would depend on what the other tens of thousands of genes in that child were doing.
The third obstacle to re-engineering human nature comes from human nature itself. We are often told that it's only human for parents to give their children every possible advantage. Stereotypical yuppies who play Mozart to their pregnant bellies and bombard their newborns with flash cards would stop at nothing, it is said, to give their children the ultimate head start in life.But while parents may have a strong desire to help their children, they have an even stronger desire 'not to hurt' their children. Playing Mozart may not make a foetus smarter, but it probably won't make it stupider or harm it in other ways. Not so for genetic enhancement. It is not obvious that even the most overinvested parent would accept a small risk of retardation in exchange for a moderate chance of improvement.
Another speed bump from human nature consists of people's intuitions about naturalness and contamination. People believe that living things have an essence that gives them their powers and which can be contaminated by pollutants. These intuitions have been powerful impediments to the acceptance of other technologies. Many people are repelled by genetically modified foods even though they have never been shown to be unsafe or harmful to the environment. If people are repulsed by genetically modified soybeans, would they really welcome genetically modified children?
Finally, anyone who has undergone in-vitro fertilisation knows that it is a decidedly unpleasant procedure, especially in comparison to sex. Infertile couples may choose the procedure as a last resort, and some kooks may choose it to have a child born under a certain astrological sign or for other frivolous reasons. But people who have the choice generally prefer to conceive their children the old-fashioned way.
It is misleading, then, to assume that parents will soon face the question, 'Would you opt for a procedure that would give you a happier and more talented child?' When you put it like that, who would say no? The real question will be, 'Would you opt for a traumatic and expensive procedure that might give you a slightly happier and more talented child, might give you a less happy, less talented child, might give you a deformed child, and probably would make no difference?' For genetic enhancement to 'change human nature' not just a few but billions of people would have to answer yes.
My point is not that genetic enhancement is impossible, just that it is far from inevitable. And that has implications. Some bioethicists have called for impeding, or even criminalising, certain kinds of research in genetics and reproductive medicine, despite their promise of improvement in health and happiness. That is because the research, they say, will inevitably lead to designer babies. If genetic enhancement really were just around the corner, these proposals would have to be taken seriously. But if the prospect is very much in doubt, we can deal with the ethical conundrums if and when they arise. Rather than decrying our posthuman future, thinkers should acknowledge the frailty of technological predictions and should base policy recommendations on likelihoods rather than fantasies.
Steven Pinker is Peter de Florez Professor in the Department of Brain and Cognitive Sciences at MIT, United States