The Mating Mind: How Sexual Choice Shaped the Evolution of Human Nature (12 page)

After mating, the choosy females start producing offspring. Their sons have longer-than-average tails that they inherited from their fathers. (Their daughters may also inherit longer tails—a phenomenon we shall consider later.) The non-choosy females produce sons whose tails are about the same length as those of their fathers—but these mediocre tails are no longer average. They are now below average, because the average tail length has increased in this generation, due to sexual selection through mate choice. The genes for long tails have spread.

The question is, will they keep spreading? Fisher's key insight

was that the offspring of choosy females will inherit not just longer tails, but also the genes for the sexual preference—the taste for

long tails. Thus, the genes for the sexual preference tend to end up in the same offspring as the genes for the sexually selected trait. When genes for different traits consistently end up in the same

bodies, biologists say the traits have become "genetically correlated." Fisher's runaway process is driven by this genetic correlation between sexual traits and sexual preferences in off spring, which arises through the sexual choices their parents made. This genetic correlation effect is subtle and counterintuitive, which is one reason why biologists took fifty years to prove that Fisher's idea could work.

Of course, when the sons of choosy females inherit the genes underlying their mother's sexual attraction to long tails, they may not express this preference in their own mating decisions. But they can pass their mother's sexual preferences on to their own daughters. Since their long tails make them sexually attractive, they tend to produce not only more sons than average, but more daughters as well. In this way, the sexual preference for long tails can genetically piggyback on the very trait that it prefers. This gives the runaway process its positive-feedback power, its evolutionary momentum.

The Runaway Brain

Did the runaway sexual selection process play a role in the evolution of the human brain? To see how this would work, take the previous example, and in the place of "bird," substitute "hominid"—meaning one of our ape-like ancestors that walked erect. For "long tail," substitute "creative intelligence." If hominid males varied in their creative intelligence, and if that creative intelligence was genetically heritable, two out of three prerequisites for sexual selection would be present.

The only other requirement would be for hominid females to develop a sexual preference for creative intelligence, for whatever reason. If they did, then males with higher creative intelligence would attract more sexual partners and produce more offspring, assuming our ancestors were not completely monogamous. Those offspring would inherit higher-than-average creative intelligence, and would also inherit the sexual preference for creative intelligence. Intelligence would become genetically correlated with the sexual taste for intelligence. The sexual taste would piggyback on

the evolutionary success of the sexual trait that it favors. The sexual trait and the sexual preference would both spread through the population. The hominids would become more creatively intelligent, and demand more creative intelligence of their sexual partners. The key here is that creative intelligence need not have given the hominids any survival advantages whatsoever, but through runaway it could evolve as a pure sexual ornament.

In the early 1990s, the runaway process seemed to me ideally suited to explaining why the human brain evolved so quickly, and to such an extreme size, during a period when it seemed to make our ancestors no better at making tools or competing against other species of African hominids. It became the focus of my research and the subject of my 1993 Ph.D. thesis at Stanford, which was titled "Evolution of the Human Brain through Runaway Sexual Selection." The human brain's evolution clearly looked as if it was driven by some sort of positive-feedback process. Other theorists proposed other candidates for the positive feedback. In 1981, E. O. Wilson suggested that larger brains permitted more complex cultures, which in turn selected for larger brains. This could initiate an evolutionary feedback loop between brain size and cultural complexity. Richard Dawkins has supported this view, seeing the human brain as a repository of learned cultural units called "memes." Larger brains permit more memes, which in turn favor bigger brains.

Two other positive-feedback ideas have proven influential in evolutionary psychology. In 1976 Nicholas Humphrey proposed that pressures for social intelligence could have turned into a positive-feedback process that drove human brain evolution. In 1988 Andy Whiten and Richard Byrne extended this idea by focusing on the survival advantages of social deception and manipulation. Their "Machiavellian intelligence" hypothesis has been accepted by many primate researchers and psychologists interested in human social intelligence. Apart from social competition within groups, another positive-feedback possibility was competition between groups. In 1989 Richard Alexander proposed that perhaps tribal warfare turned into an arms race for

ever greater technological and strategic intelligence. This military competition could drive brain size and intelligence upwards.

These theories all have some validity. Cultural, social, and military selection pressures were probably significant. But these positive-feedback loops seemed too speculative. They had not been admitted into the pantheon of evolutionary forces by biologists, and were not routinely used to explain interesting traits in other species. They were slightly ad hoc hypotheses restricted to primate and human evolution. The runaway process was different: it was part of mainstream evolutionary theory, one of the leading contenders for explaining complex, costly, ornamental traits in other species. Yet it had never been proposed as the driving force behind the evolution of the human brain.

This seemed a peculiar oversight in need of vigorous correction, and for several years I gave dozens of talks about the idea of human mental evolution through runaway sexual selection. Matt Ridley kindly gave the idea some attention in the final chapter of his book
The Red Queen.
However, I now think that the runaway brain idea is only partly successful. It has some strengths that can help account for some of the sex differences in human behavior and some of the differences between our species and other primates. However, it also has some serious problems, so it will constitute only a small part of my overall theory.

The Requirements of the Runaway Process

One possible problem is that runaway sexual selection demands polygyny—a mating pattern in which some males mate with two or more females. For runaway to work, some males must prove so attractive that they can copulate with several females to produce several sets of offspring. The least attractive males, as a rule, must be left single, heartbroken, and childless. Sexual competition must be almost a winner-take-all contest. In elephant seals, for example, one dominant male may account for over 80 percent of all copulations with females on a particular beach, and almost as high a proportion of all offspring. (Polygyny does not mean that every male gets to father the offspring of many females-

that would be a mathematical impossibility, given an equal sex ratio. It means rather that a few males mate often and produce many offspring, and most males mate rarely, producing very few offspring.)

If our ancestors were perfectly monogamous, runaway sexual selection could not have favored large brains, or creative intelligence, or anything else. Runaway would never have started. A crucial question is how polygynous our ancestors were. The more polygynous they were, the more potent runaway sexual selection could have been. The modern understanding of human evolution suggests that our ancestors were moderately polygynous—neither as polygynous as elephant seals, gorillas, or peacocks, nor as perfectly monogamous as albatrosses. The evidence comes from many sources, but I shall mention just two: body size differences and anthropological records. Across primates, species where males are much larger than females tend to be highly polygynous. This is because males compete more intensely and violently in more polygynous species where the stakes are higher, and this competition drives up their relative size and strength. Generally, the larger the sex difference in body size, the more polygynous the species. In humans, the average male is about 10 percent taller, 20 percent heavier, 50 percent stronger in the upper body muscles, and 100 percent stronger in the hand's grip strength than the average female. By primate standards, that is a moderate sex difference in body size, implying a moderate degree of polygyny.

Other evidence of polygyny comes from anthropological studies of human cultures and human history. Most human cultures have been overtly polygynous. In hunter-gatherer

cultures the men who are the most charming, the most respected, the most intelligent, and the best hunters tend to attract more than

their fair share of female sexual attention. They may have two or

three times as many offspring as their less attractive competitors. In pastoral cultures the men who have the largest herds of animals

attract the most women. In agricultural societies the men who have the most land, wealth, and military power attract the most women. Before the middle ages, in urban civilizations with high

population densities, the men at the top of the hierarchy almost always had harems of hundreds of women producing hundreds of babies. The first emperor of China reputedly had a harem of five thousand. King Moulay Ismail of Morocco reputedly produced over six hundred sons by his harem. In European Christian societies from the medieval era onwards, monogamous marriage became the religious and legal norm, though powerful men still tended to attract many mistresses and to re-marry more quickly if their first wife died. For example, anthropologist Laura Betzig showed that throughout American history, presidents tended to mate more polygynously than men of lower political status. (This may be little consolation to politicians of mediocre musical ability, since popular male musicians such as Bob Marley and Mick Jagger allegedly behaved even more polygynously than presidents.)

Those of us brought up in European-derived cultures tend to think of humans as monogamous, but in fact mating in our species has almost always been moderately polygynous. For millions of years, there was enough variation in male reproductive success to potentially drive runaway sexual selection during human evolution.

Runaway Is Unpredictable

The runaway process is very sensitive to initial conditions and random events. Runaway's initial direction depends on the female preferences and male traits that happen to exist in a population. Runaway's progress depends on several kinds of random genetic events such as sexual recombination, which mixes genes randomly every time two parents produce offspring, and the evolutionary process called genetic drift, which eliminates some genes by chance in small populations, as a result of an effect called "sampling error." Because runaway is a positive-feedback process, its sensitivity to initial conditions and random events gets amplified over evolutionary time. These effects make runaway's outcome quite unpredictable. It never happens the same way twice. Runaway's unpredictability is apparent if you look at the

diversity of sexual ornamentation in closely related species. Of a dozen species of bowerbirds, no two construct the same style of courtship nest. Of three hundred species of primate, no two have the same facial hair color and style. These differences cannot be explained as adaptations to different environments—they are the capricious outcomes of sexual selection.

Computer simulations confirm runaway's unpredictability. In the early 1990s when we were psychology graduate students at Stanford, Peter Todd and I spent months running simulations of runaway sexual selection. We would run the same program, repeatedly, while just changing the initial conditions slightly, or changing the random numbers used by the computer to simulate random events like mutation. The results were quite capricious. Two populations can start out very similar to each other, and evolve slightly different sexual preferences, which lead their sexual ornaments to evolve in slightly different directions, which reinforce their sexual preferences, and so forth. The populations end up in opposite corners of the range of possibilities, sprouting different sexual ornaments, with different sexual preferences. And if you run the same simulation again, with just slightly different random numbers influencing mutations, the populations will evolve in yet another set of directions. A population will often split apart spontaneously into two clusters that are reproductively isolated, creating two distinct species. If you went out for a coffee while running a simulation and came back ten minutes later, the population would usually have moved where you least expected it—not through the physical space of its simulated habitat, but through the abstract space of possible ornament designs.

Suppose you take a dozen species of ape that lived in social

groups in Africa about ten million years ago. Think of these species as nearby clusters in the space of all possible sexual

ornaments and courtship behaviors. Now turn runaway sexual

selection loose in each species. One species might develop a runaway preference for large muscles, and turn into gorillas.

Another species might develop a runaway preference for constant sex, and turn into bonobos (previously known as "pygmy