The point here is that a modest difference in the mean of some trait can have a tremendous effect on the frequency with which members of a group exceed a high threshold. If some important cultural task can only be accomplished by individuals who are unusually good at solving certain kinds of puzzles, then the course of cultural evolution may change radically with modest changes in the group's average puzzle-solving ability. There are many other factors that might influence such events, but a difference in mean ability due to genetic differences is one of them. And both of these factors—social phase transitions andincreases in the frequency of people with specific talents—may have played a part in the birth of modern science.
Science as we know it got its official start in Europe in the sixteenth century with the publication of Copernicus's work De revolutionibus in 1543. The closest thing to modern science seen before that would have been the protoscience practiced by the Greek and, later, Arab civilizations—but they're not that close. The productivity and intensity of modern science far outshines earlier efforts. Some of the most important European scientists, such as Isaac Newton, James Clerk Maxwell, and Charles Darwin, made larger intellectual contributions as individuals than other entire civilizations did over a period of centuries.
We believe that science requires communication and cooperation between people who are unusually good at (and interested in) puzzle-solving. Science is a social enterprise, and scientists never truly work alone: They always build on the work of others. It was Newton who said, "If I have seen further it is by standing on the shoulders of Giants," and he ought to have known. So the number of such people, and their social connections, is crucial to the progress of science. We also know that modest differences in mean ability can have a big effect on how common such people are.
You see, there can also be phase transitions in connectivity. Imagine that the average budding scientist in Europe in 1450 knew a few other people like himself. Those acquaintances knew others, but since such people were rare, the potential scientists of Europe fell into small, isolated groups rather than a single connected community. There was no efficient way for new ideas and discoveries to spread. We are positing that as the frequency of such people increased, there was a sharp transition at a certaincritical value. Suddenly all groups connected, and there was a path between any two members. Something similar happens in epidemiology: If the number and density of vulnerable individuals exceeds a certain threshold, the infectious disease is certain to spread to the entire community. Below that threshold, the disease is confined to a small cluster of people and dies out.
Thus, the scientific "revolution" may well have resulted from modest changes in gene frequencies affecting key psychological traits. What traits would have favored the birth of science? Increases in abstract reasoning or numerical abilities might have helped, and it's possible that those traits were favored by selection in complex, hierarchical societies. Generally, though, we think there was no direct selection favoring creativity itself, and that creative individuals are accidental by-products of selection for other traits, traits that really did pay off in everyday life, such as low time preference and the ability to make complex mental models.
Our view is in sharp contrast to those who have argued that creativity conferred fitness benefits. It has been shown that poets are unusually likely to be manic-depressive.28 Building on this, others have argued that alleles underlying manic-depression should have increased in frequency because of the social rewards received by poets and other creative artists.29 Of course, few people carrying those alleles had a chance to be poets: Most (in recent millennia) must have been hardscrabble farmers, and it's hard to see how manic-depression could have been an advantage in that situation.
In fact, poets have seldom received large rewards, and their fitness has often been low—particularly among those with manic-depression, as a result of its high suicide risk. More generally, creativity seldom confers large fitness advantages, because good new ideas can be rapidly copied by others. The copiers receive the fitness benefits without paying the associated costs. In fact, it's been obvious for a long time that innovators seldom harvest much of the benefit generated by their innovations. Public policy has aimed at increasing those rewards—for example, through patent systems and public support of scientific research. Such support is limited and fairly recent, however, and over the long run of human history and prehistory, direct selection for creativity seems unlikely.
Technical and social factors must have been important in increasing social connectivity: Better transportation, regular mail services, and the printing press, for example, played essential roles. Although inventions such as the printing press were undoubtedly important, they seem to have been necessary rather than sufficient, since science either does not exist or is appallingly feeble in the majority of the world's populations, even among those that have access to those favorable technological factors. If a region or population produces major advances in knowledge, science there is real and alive, otherwise not. By that standard, science does not exist in sub-Saharan Africa or in the Islamic world today. As Pervez Hoodbhoy (head of the physics department in Islamabad) has written, "No major invention or discovery has emerged from the Muslim world for well over seven centuries now."30
Although we do not as yet fully understand the true causes of the scientific and industrial revolution, we must now consider the possibility that continuing human evolution contributed to that process. It could explain some of the odd historical patterns that we see. For example, if people hadn't yet changedenough, the failure of Hellenistic science to take off may have been inevitable. In addition, such ideas may help explain why some populations with an early start on agriculture and state formation have found it easy to participate in these revolutions, while those with late starts have not. In particular, we think that the story of the Ashkenazi Jews, many of whom have played important parts in the later phases of those two revolutions, was shaped by this kind of evolution—evolution over historical time.
5
GENE FLOW
GENETIC HISTORY
Geneticists have traditionally traced the flow of genes in order to study the movements and origins of peoples. They've studied particular variants of the Y chromosome in an attempt to determine which ethnic group in Asia is most closely related to the Amerindians.1 They've tried to determine the extent to which modern Europeans are descended from ancient Europeans of the Upper Paleolithic period who adopted farming, or from Neolithic immigrants from the Middle East.2 Researchers have used mitochondrial DNA (mtDNA) and Y-chromosome data to determine which groups contributed maternal and paternal ancestry to a mixed population—they have determined, for example, that most Mexican Y chromosomes are of Spanishorigin, whereas most Mexican mtDNA is Amerindian.3 They have tried to analyze other ancient population movements in this way as well, most notably the original human expansion out of Africa.
There are two ways of looking at these types of informative gene variants. In the kind of analysis conducted to determine paternal and maternal ancestral lines, on the one hand, researchers are only interested in these gene variants as markers of past population movement and admixture rather than in the functions of the variants themselves. The assumption is that one Y chromosome functions just like any other and that all mtDNA variants have the same properties: That is to say, they're neutral. But if their properties varied—if the bearers of some Y-chromosome variants had noticeably higher fitness than those bearing other variants—this whole brand of analysis would be thrown into question, particularly when used to look far back into prehistory.
We, on the other hand, are interested in alleles because of their effects, precisely because they do make a difference. Generally, we're interested in how population movements and admixture have helped to spread new adaptive variants rather than in how using the variants can help us to track the movements.
Every new mutation, including any rare but important beneficial mutation, starts out as a single co
py in one individual. It's local. If it's going to be important, if it's ever going to influence a significant fraction of the human species, it must first spread. Looking at the bigger picture, we can see that the flood of favorable mutations involved in the recent acceleration of human evolution will have major impacts only if they spread widely. Presumably, if our theories are correct, many of them are still inthe midst of sweeping through large segments of our world population. This means that the average person today bears many of these new favorable mutations. You are, most likely, significantly different genetically from your ancestors of a few thousand years ago. And although natural selection hardly operates in a way guaranteed to maximize our convenience, it is still the case that many adaptive changes have welcome results. It's hard to argue against something that keeps you alive.
BREEDING LIKE RABBITS
The settler Thomas Austin released 24 wild rabbits on his Australian farm, called Barwon Park, in 1859, and some other Australian farmers later followed his example. Rabbits are sexually mature at about six months, and they have a 31-day gestation period. Given a favorable environment, rabbits can easily increase their population fourfold in a year. Try to imagine the growth of the rabbit population in Australia: first 24 rabbits, then 100 rabbits after a year, 20,000 in five years, and 25 million after ten years. That's roughly what happened: At the end of ten years, shooting or trapping 2 million a year had no noticeable effect on their population.
At first growth was slow—an increase of 75 in a year doesn't sound that impressive, not in a country the size of the lower 48. But growth speeded up as the rabbit population increased. Another way of putting it is that the percentage of growth per year stayed the same, but that the percentage was multiplied by a larger and larger population as time passed. A process that at first seemed unimpressive left a whole continent literally swarming with rabbits in a single decade. It took only two or threetimes longer to fill a continent with rabbits than it had to fill a single farm.
A favorable allele, such as the one that confers lactose tolerance, spreads in much the same way, although the process takes thousands of years, largely because human generations are much longer than rabbit generations. But for a sweep to happen rapidly, the population must be "well mixed." This is not always the case, because mixing genes over long distances—over rivers, mountains, deserts, and oceans, or through hostile tribes—is far from automatic. Sometimes it happened, sometimes it didn't. These sweeps were strongly influenced by history—and they influenced history right back.
HOW A SWEEP BEGINS
Every selective sweep starts out as a change in the DNA of a sperm or egg. Such changes can be caused by chemicals, radiation, or just random jostling of molecules—but what matters to us is that such changes do occur. Mutations favorable enough to initiate a sweep are extremely rare. One set of human DNA has about 3 billion nucleotides, and an average person has about 100 new mutations. Most of those changes are in DNA that apparently does nothing at all—only 2 percent of our DNA does anything (as far as we know)—but on average, two or three of those mutations affect functional DNA. Still, they do not usually make a significant difference, either in a positive or a negative way.
When a mutation does make a significant difference, the effect is almost always negative: Random changes in an incredibly complex piece of machinery are likely to screw things up. Sometimes a change in a single nucleotide can kill or cause serious disability. For example, achondroplasia, the most common kind of dwarfism, is caused by a change in a single nucleotide on chromosome 6—almost always the same exact change. Such negative alleles never become common: Their bearers have fewer children than average, so there will be fewer copies in the next generation. Very rarely, a mutation happens that has a positive effect—a good difference. These rare but supremely important events are the raw material of evolution.
LIMONE SUL GARDA
In 1980, Italian researchers found that a man from Limone sul Garda (a small lakeside village in northern Italy) had very low levels of HDL ("good" cholesterol) and high levels of triglycerides, yet showed no sign of heart disease. Both of his parents
Limone sul Garda
had lived to advanced ages. Their curiosity whetted, the researchers performed blood tests on all 1,000 inhabitants of Limone and found a total of 43 people with this same unusual blood-lipid profile. The local church had birth records going back centuries, and the researchers were able to determine that all those individuals could trace their ancestry back to the same couple (Giovanni Pomaroli and Rosa Giovaneli), who had married in 1780.4 This genealogical pattern suggested that these villagers shared a mutation, which turned out to be a change in the protein called ApoA-I (Apolipoprotein A-I), a major component of high-density lipoprotein (HDL). ApoA-I helps to clear cholesterol from arteries, but this variant, ApoA-^ (M for Milano), apparently does a considerably better job of it. A change in a single nucleotide modified an amino acid in the protein, completely changing its chemical action.
ApoA-IM is much more effective at scouring out arteries than the standard version of the protein is, and carriers have substantial protection against atherosclerosis. They have a much-reduced risk of heart attacks and strokes, and they often reached an advanced age.5 Not only that, these effects of the ApoA-IM mutation have been duplicated in mice, and it protects them against artery plaque as well.6 Preliminary tests show that intravenously administered synthetic ApoA-IM actually shrinks preexisting artery plaque in humans: Nothing else we know of does that.
Judging from the records we have, this mutation seems to have increased in number, from 1 copy to 43 in ten generations. Chance and general population growth must have played a role, but let's suppose that freedom from heart attacks and strokes is driving a gradual increase. What would happen if it were given several thousand years to expand?
Let's say that its true advantage—the long-term average— is 7 percent, so that carriers raise 7 percent more children than average. In that case, you'd expect most Europeans to have a copy in 6,000 years or so. This assumes, of course, that Europe will still exist thousands of years from now, that we won't have developed a universal cure for atherosclerosis in the meantime, and that the robots won't have taken over first. We know the future is uncertain—bear with us.
Success in 6,000 years is nothing to hold your breath about, but the point is that mutations with a similar advantage that started in a single village at the dawn of recorded history have had time to become common in just this way. That estimate assumes that genes (and people) are well mixed, but that's clearly not the case in Limone sul Garda. The village is quite isolated: The mountains and the lake hem it in, and there wasn't even a road until the 1930s. Such isolation doesn't make the occurrence of a favorable mutation more or less likely, but the concentration of carriers in one village may have made them easier to notice. However, it certainly interferes with the spread of the gene.
So how did a favorable mutation spread thousands of years ag°?
THE GIRL NEXT DOOR
Few villages today are as geographically isolated as Limone: Most have other villages fairly close to them, and generally there's traffic between them. People make frequent visits to these places close to home. The simplest and oldest mechanism of gene flow—marrying someone from the next village over— therefore still prevails. More often than not, this has meant women leaving their homes to join their husbands' communities,an ancient pattern that we still see in chimpanzees. This village- to-village contact has been a factor in gene flow ever since people settled in villages, and it has been one of the most important ways in which favorable alleles have spread. Given time, neighborhood marriages can carry an allele thousands of miles. A beneficial allele originating in one band or village could, through such intermarriage, gradually spread to neighboring populations, and then to neighbors' neighbors, and so on. Al- leles with a big advantage would spread more rapidly than alle- les with a small advantage.
In any case, with some simplifying assumption
s, it's possible to model the spread of an adaptive allele with a mathematical formula. In that model, the frequency of the favored allele spreads in the form of a wave with a constant speed. The speed depends on the selective advantage and the root-mean-square distance separating the parents' and the child's birthplaces. If we call that marital distance a and the selective advantage of the allele s, the speed of advance is approximately a X (2s)% miles per generation.
Hunter-gatherers can be amazingly mobile, and since most recent hunter-gatherers were spread very thinly, there often were no girls next door. So hunter-gatherers, especially in sparsely settled areas, had to find mates at a considerable distance. A generation ago, when many Bushmen were still wandering freely, their average marital distance was over 40 miles. This may not have been typical in prehistory. In the days before agriculture, when everybody and his brother was a hunter-gatherer, most lived in choice territories, not in the marginal habitats like the Kalahari Desert where that way of life has persisted. Population density would have been higher in those conditions thanamong Bushmen today, and people may not have had to search so far for a mate. However, it is clear that agriculture eventually led to crowding. Peasant farmers usually marry people living nearby, not least because there are plenty of people living nearby to choose from. In an example discussed by Alan Fix, based on census records from a densely settled part of rural England about 150 years ago, the average marital distance was only 6 or 7 miles.7
The 10,000 Year Explosion Page 12