It's hard to predict where genetic engineering is heading, but it's easy to predict what different scientists will say about where it's heading. The key is the fax test. Biotechnology enthusiasts fax to you. Ring them up, and they send corporate reports, congressional testimony, and biographical information on their CEOs. They set up conference calls so you can bask in the visions of senior executives. "Augmenting intelligence may prove to be feasible within the next 30 to 40 years," says Chuck Tyler, vice president for new ventures at Applied Biosystems, And, as for mortality: "Perhaps Ponce de León was just three or four centuries too early."

The less visionary scientists also communicate by fax, but you do the sending. Want to talk to a Nobel Prize-winning molecular biologist? Send your interview request by fax. Then maybe you'll get permission to fax the questions. Not that it's worth the trouble; the forecast will, at its most venturesome, amount to "time will tell."

Those cautious typos tend to be found at the National Institutes of Health's Human Genome Project, a massive effort to decipher the three-billion-digit genetic sequence that encodes the recipe for making a human being. For biotech companies, bullishness is good business, but for the scientists whoare spending $3 billion in taxpayer money on the Genome Project, speculation about messing around with human nature and society can only bring sorrow. "It's critical that people appreciate the limits of what can be done, so that their fears don't intrude on the benefits that could come out of genetic research," says Dr. Nancy Wexler, a psychologist and hereditary disease specialist who sits on the Genome Project's advisory board.

There's a third contingent in the debate over biotechnology's future: the Rifkinites. When people like Wexler downplay the issues that lie ahead, people like Andrew Kimbrell, policy director of Jeremy Rifkin's biotechnology watchdog group, the Foundation on Emerging Technologies, say, "That's disingenuous." Kimbrell says, "We're beginning to identify genetic material that gives us a predisposition to any number of mental and emotional disorders. You could see genetic engineering of human beings from the fetal level on up." For example, "If they found that your child who's in kindergarten was predisposed to shyness, they would alter that child not be shy…. These technologies have an enormously eugenic potential."

It is an odd spectacle, From the political left—which historically has vilified anyone who suggested that genes had much to do with behavior—comes the insistence that genes have lots to do with behavior. Meanwhile, molecular biologists and geneticists are insisting that molecular biology and genes aren't too consequential; can-do scientists who just convinced the nation to fork over $3 billion to decode the recipe for humankind are stressing all the things that can't be done with the recipe.

The economic and scientific forces that may be propelling us toward a bio-talent industry, and a host of associated large policy questions, are already on display in more conventional markets. In agriculture, biotech firms and universities are working on everything from tougher tomatoes to leaner pigs to plants that can make their own pesticides and heal their own wounds. Companies say that within the next decade they'll learn how to engineer the nutritional content of food. In ten years, "All the major commodities, both animal and plant, will be engineered," promises Dr. Arnold Foudin, who tracks private ventures for the Department of Agriculture's biotech division. You hate oats but want the health boost of oat bran? "There's no reason on earth we can't put it into other crops," he says.

Private researchers also boast of genetically altered microorganisms that can devour organic wastes in groundwater and landfills—and that come complete with a self-destruct mechanism to prevent the Frankenstein problem. "The only carbon they can eat is the carbon you want destroyed," says Dr. Leonard Guarraia, Monsanto's director of science and technology policy. "When it's gone, they're gone." Biotech wizards also promise to displace the known hazards of chemical pesticides with organic plant "vaccines."

The Rifkinites have gotten most of their publicity so far by lobbying against such interventions—the engineering of single-cell forms of life. But this cause may be waning. Though mainstream scientists initially shared some of the concern, they seem little troubled these days about mutant organic pesticides running amok. Unless some such catastrophe strikes during the next few years, as genetic engineering heads out of the lab and into everyday life, Rifkin will get little reward for playing this theme. Attention will turn to the next set of big questions: ones directly involving human beings.

The initial round of human genetic therapies is expected in medicine. First come cures for single-gene disorders, such as cystic fibrosis, muscular dystrophy, sickle cell anemia, kidney disease, and Huntington's disease. In five years: insertion of therapeutic genes into human cancer patients. In ten years: genetic therapies for a wide variety of cancers. In 20 years: cures for autoimmune disorders and rheumatoid arthritis. In 40 to 50 years: cures for diabetes, Alzheimer's, and multiple sclerosis. Then, at some point, we're expected to confront a realm of cancers, heart disease, arthritis, and mental disorders that, while genetically treatable, involve so many genetic or environmental variables that a sure-fire genetic cure is impossible.

It's hard to decide how seriously to take these sunny forecasts, since scientists both in and out of the business sector find it in their interests to make them. In any event, the really big issues will come in the next generation of DNA technology. If the frontiers of genetic engineering can advance from one-gene to three-gene traits within a few years, what stands in the way of discovering the larger gene combinations that influence height, weight, and skin color? For that matter, won't we eventually be able to identify the genetic roots of intelligence, ambition, courage, and altruism? And if so, whose children will be permitted to benefit?

Ask someone affiliated with the Human Genome Project about these spooky future technologies, and they'll probably dismiss the question by saying that environmental factors overshadow the influence of genes. "If you deprive a child of all sensory input, it can’t do anything when it grows up," notes David Botstein, a vice president at Genentech and a board member of the Genome Project. Or, to take a slightly less extreme ease: if you have the same genes as Einstein but you grow up in Okarche, Oklahoma, you may not do too well on your SATs. This point is so firmly embedded in the repertoire of cautious types that they have developed pet metaphors to express it. Here's Botstein's: "The brain is the hardware, and education—in fact, arguably all experience—is the software. Now, the question is, can I increase the ability of an IBM-PC to do word processing by giving it the brains of a Cray?" Here's Wexler's: "All kinds of people take the same scale and combine the notes in incredibly different ways," she says. "You put a C next to an A or a B, and you get a Bach chord or a Rolling Stones chord,"

This all sounds nice—and, for the most part, it is true; the expression of genes (even genes as simple as the ones for skin pigmentation) is exquisitely sensitive to the environment. But this doesn't mean that it is in principle impossible to meaningfully alter behavioral patterns by manipulating the genes. Indeed, very simple reasoning leads to the conclusion that, in principle, it almost certainly is possible to affect intelligence, aggression, and other traits through genetic engineering.

Just about all psychometricians now agree that genetic differences underlie some substantial fraction of the differences in intelligence—as crudely measured by, say, IQ, scores—among people. Granted, this consensus says nothing for certain about individual cases. The difference in intelligence between Sylvester Stallone and Stephen Hawking is, for all we know, entirely due to the environment. And, granted, this consensus provides no guarantee about the ultimate manifestation of genetic differences between two people. Even if part of the difference is genetic, that does not mean that anyone with Hawking's genes would have wound up with his mind: the role of the environment in shaping the mind is too critical. What is safe to say is this: some babies have constellations of genes that are more likely to lead to genius than other babies' constellations of genes. And growing evidence—some from studies of identical twins reared apart—suggests that the same may go for courage, altruism, moodiness, etc. If these constellations of genes, along with their statistical implications for behavior, can ever be isolated, that will ensure decades' worth of articles for bioethics journals. For these statistical correlations will no doubt entice more than a few expectant parents into an intense interest in genetic engineering.

Actually, even if observed differences in intellectual and emotional traits had no genetic underpinning, there would be reason to believe genetic engineering is in principle possible. After all, no scientist denies that evolution, by recombining genes, has raised the average intelligence of the population (or, more accurately, the average potential for intelligence). And there is no reason to assume that this process of recombination has reached any fundamental limit. It is strange to hear government scientists, many of them steeped in the study of how natural selection has manipulated the genes to create the human brain, acting as if there were no reason to believe that manipulating the genes could have any effect on the human brain.

Once these theoretical questions are settled, the questions of practice arise: Will science ever be able to specify genes that make a person more likely to be smart, ambitious, emotionally stable, etc? Here complexity rears its ugly head, There are three billion genes in the human genome, and they can interact in astoundingly diverse ways. To take the simplest case: children of dwarves who inherit one normal gene and one dwarf gene turn out more normal than those who inherit two dwarf genes. So a child who inherits only one gene for Huntington's disease will be better off than one who inherits two bad genes, right? Wrong. Each case is different. This kind of variation may seriously limit our ability to extrapolate from one gene-trait relationship to another, and vastly complicate attempts to alter whole complexes of genes to predictable effect.

The three billion genes in the genome—many of which affect the expression of others—are nothing compared to the multiplicity of environmental factors that affect genetic expression. Keeping track of the resulting chains of causality is probably hopeless. Granted, the same exact collision of genetic and environmental factors, if replayed, would produce the same result. But even if you know all of the factors involved, you couldn't predict the outcome unless you had seen the same situation play out before. That's unlikely, and arguably impossible. Wait, it gets worse. The central factor in every collision, the person, is a function not just of three billion genes but also of a history of gene-environment collisions.

Again: none of this means that some sets of genes aren't more likely to get you into Harvard than other sets. But it does say that no set can guarantee you a slot there. And it says, further, that uncovering any statistical correlations between genes and the chances of getting into Harvard is going to be a mammoth task. Here's a likely scenario. Researchers looking for genes that predispose people toward, say, anti-social behavior will find a fuzzy pattern involving a dozen genes and suggest there's a causal link. Other scientists will challenge their hypothesis, test it with other data, declare it falsified, and accuse the original researchers of failing to consider the influence of such factors as education, family situation, and religious upbringing. If you liked watching scientists argue about what happens inside a jar, you'll love watching them argue about what happens outside.

But adroit entrepreneurs and anxious parents won't wait for the last word on smartness genes. Once word gets out about a study linking certain genes to, say, a 160-point average difference in SAT scores, the same folks who send their kids to Princeton Review to give them an "equal chance" will start to think about how unfair it is that their offspring are genetically disadvantaged.

The ensuing problems are obvious. Such gene therapy, if it became available, would at first be so expensive that only the rich could afford it. Even the purest free marketeer would have to think twice about that. And if society chooses not to let the market allocate this resource, should we pay to make gene therapy available to everyone, or should we ban it? (And will Congress ban it if Japan doesn't?) Before you decide, answer these questions: Certainly parents should have access to genetic therapies for alleviating serious diseases, right? Including mental retardation? And if that's fair game, then how about below-average intelligence? Similarly: Certainly parents should be allowed to head off severe manic-depressive illness (the genetic roots of some variants of this are now coming to light). So what's wrong with making your child less prone to more subtle mood swings? To playground violence? Totemper tantrums?

Another frontier in eugenics is the question of programming immortality—or, at least, holding mortality at bay. Roughly speaking, geneticists lend to lean toward either the mechanical model or the genetic model of aging. The mechanical view: "Aging is simply exposure to living for years. There's a certain wear and tear." The genetic view: "Aging is encoded in the genes, it's part of the basic biology of the organism. It's not a question of the parts wearing out. It's programmed." Then there is the middle ground, occupied by Dr. George Martin, a genetic specialist on aging. Ultimately, Martin believes aging is largely genetic, but very complex, with many genes involved. If so, the good news is that genetic engineering can probably prevent deaths due to pathologically early aging, perhaps fairly soon. The bad news is that it will be far more difficult, perhaps impossible, to postpone significantly aging and death in general.

The ethical questions that may grow out of all this 30 or 40 years down the road will be colored in more ways than one by the question of the complexity of genetic causation. Rifkin has been talking for years about the ethics of parents choosing one trait over its alternatives; tall over short, for example, or white over black. But choices may also involve deciding whether some whole categories of traits are more important than others. The reason is that a single gene or group of genes can exert simultaneous effects on more than one trait. The same gene that contributes to your hooked nose may reduce your chances of tooth decay, for all we know, or even limit your fluency in foreign languages. So you can't order everything from the trait menu, and you sometimes can’t order even one item without worrying about a bunch of others. Now the question is, do you care more about your kid's nose or her linguistic facility? And if there's any downside at all, should society let you choose?

The other problem is, the more nuanced your definition of a mental strength or virtue, the less likely you are to find consistent genetic correlates for it. So the tendency will be to design a standard that weeds out nuance. In the case of intelligence, you ask multiple-choice questions instead of essay questions. You eliminate from the test those subjects in which scientists can't agree on an objective standard. You're left with IQ, a decent predictor of professional success. You compare genomes and IQ scores and discover a strong link. In your scholarly journal article, you stipulate that IQ measures only one aspect of intelligence. But alert entrepreneurs are already hyping the story on TV. Expectant parents are lining up to get into pilot IQ gene "therapy" programs. The danger, it turns out, is not that these IQ genes will explain the subtlety of intelligence, but that they will replace it.

William Saletan is editor of Hotline, a political news service.

This article originally ran in the July 17, 1989, issue of the magazine.