Who We Are and How We Got Here
Page 26
Star Clusters are not limited to Asia. The geneticist Daniel Bradley and his colleagues identified a Y-chromosome type that is present in two to three million people today and derives from an ancestor who lived around fifteen hundred years ago.16 It is especially common in people with the last name O’Donnell, who descend from one of the most powerful royal families of medieval Ireland, the “Descendants of Niall”—referring to Niall of the Nine Hostages, a legendary warlord from the earliest period of medieval Irish history. If Niall actually existed, he would have lived at about the right time to match the Y-chromosome ancestor.
Star Clusters capture the imagination because they can be tied, albeit speculatively, to historical figures. But the more important point is that Star Cluster analysis provides insights about shifts in social structure that occurred in the deep past that are difficult to get information about in other ways. This is therefore one area in which Y-chromosome and mitochondrial DNA analysis can be instructive, even without whole-genome data. For example, a perennial debate among historians is the extent to which the human past is shaped by single individuals whose actions leave a disproportionate impact on subsequent generations. Star Cluster analysis provides objective information about the importance of extreme inequalities in power at different points in the past.
Two studies, one led by Toomas Kivisild and the other led by Mark Stoneking, have compared the results of Star Cluster analysis on Y-chromosome sequences and on mitochondrial DNA sequences and arrived at an extraordinary result.17 By counting the number of differences per DNA letter between pairs of sequences, which reflects mutations that accumulated in a clocklike way over time, these studies estimated the time since different pairs of individuals shared common ancestors on the entirely male (Y-chromosome) and entirely female (mitochondrial DNA) lineages.
In mitochondrial DNA data, all the studies found that most couples living in a population today have a very low probability of sharing a common ancestor along their entirely female line in the last ten thousand years, a period postdating the transition to agriculture in many parts of the world. This is exactly as expected if population sizes were large throughout this period. But on the Y chromosome, the studies found a pattern that was strikingly different. In East Asians, Europeans, Near Easterners, and North Africans, the authors found many Star Clusters with common male ancestors living roughly around five thousand years ago.18
The time around five thousand years ago coincides with the period in Eurasia that the archaeologist Andrew Sherratt called the “Secondary Products Revolution,” in which people began to find many uses for domesticated animals beyond meat production, including employing them to pull carts and plows and to produce dairy products and clothing such as wool.19 This was also around the time of the onset of the Bronze Age, a period of greatly increased human mobility and wealth accumulation, facilitated by the domestication of the horse, the invention of the wheel and wheeled vehicles, and the accumulation of rare metals like copper and tin, which are the ingredients of bronze and had to be imported from hundreds or even thousands of kilometers away. The Y-chromosome patterns reveal that this was also a time of greatly increased inequality, a genetic reflection of the unprecedented concentration of power in tiny fractions of the population that began to be possible during this time due to the new economy. Powerful males in this period left an extraordinary impact on the populations that followed them—more than in any previous period—with some bequeathing DNA to more descendants today than Genghis Khan.
From ancient DNA combined with archaeology, we are beginning to build a picture of what this inequality might have meant. The period around five thousand years ago north of the Black and Caspian seas corresponds to the rise of the Yamnaya, who, as discussed in part II, took advantage of horses and wheels to exploit the resources of the open steppe for the first time.20 The genetic data show that the Yamnaya and their descendants were extraordinarily successful, largely displacing the farmers of northern Europe in the west and the hunter-gatherers of central Asia in the east.21
The archaeologist Marija Gimbutas has argued that Yamnaya society was unprecedentedly sex-biased and stratified. The Yamnaya left behind great mounds, about 80 percent of which had male skeletons at the center, often with evidence of violent injuries and buried amidst fearsome metal daggers and axes.22 Gimbutas argued that the arrival of the Yamnaya in Europe heralded a shift in the power relationships between the sexes. It coincided with the decline of “Old Europe,” which according to Gimbutas was a society with little evidence of violence, and in which females played a central social role as is apparent in the ubiquitous Venus figurines. In her reconstruction, “Old Europe” was replaced by a male-centered society, evident not only in the archaeology but also in the male-centered Greek, Norse, and Hindu mythologies of the Indo-European cultures plausibly spread by the Yamnaya.23
Figure 28a. Human populations have expanded dramatically in the last fifty thousand years. We can see this in trees of relationships constructed based on mitochondrial DNA, where the rarity of recent shared ancestors in this period reflects the large size.
Any attempt to paint a vivid picture of what a human culture was like before the period of written texts needs to be viewed with caution. Nevertheless, ancient DNA data have provided evidence that the Yamnaya were indeed a society in which power was concentrated among a small number of elite males. The Y chromosomes that the Yamnaya carried were nearly all of a few types, which shows that a limited number of males must have been extraordinarily successful in spreading their genes. In contrast, in their mitochondrial DNA, the Yamnaya had more diverse sequences.24 The descendants of the Yamnaya or their close relatives spread their Y chromosomes into Europe and India, and the demographic impact of this expansion was profound, as the Y-chromosome types they carried were absent in Europe and India before the Bronze Age but are predominant in both places today.25
Figure 28b. On the Y chromosome, many people shared ancestors around five thousand years ago. This corresponds to the dawn of the Bronze Age—a period of the first highly socially stratified societies—when some males succeeded in accumulating wealth and making an extraordinary contribution to the next generation.
This Yamnaya expansion also cannot have been entirely friendly, as is clear from the fact that the proportion of Y chromosomes of steppe origin in both western Europe26 and in India27 today is much larger than the proportion of steppe ancestry in the rest of the genome. This preponderance of male ancestry coming from the steppe implies that male descendants of the Yamnaya with political or social power were more successful at competing for local mates than men from the local groups. The most striking example I know of is from Iberia in far southwestern Europe, where Yamnaya-derived ancestry arrived at the onset of the Bronze Age between forty-five hundred and four thousand years ago. Daniel Bradley’s laboratory and my laboratory independently produced ancient DNA from individuals of this period.28 We found that approximately 30 percent of the Iberian population was replaced along with the arrival of steppe ancestry. However, the replacement of Y chromosomes was much more dramatic: in our data around 90 percent of males who carry Yamnaya ancestry have a Y-chromosome type of steppe origin that was absent in Iberia prior to that time. It is clear that there were extraordinary hierarchies and imbalances in power at work in the expansions from the steppe.
The Star Cluster work rests on Y chromosomes and mitochondrial DNA. What can whole-genome analysis add? When whole-genome data are used to reconstruct the size of the ancestral population of most agricultural groups in the last ten thousand years, they document population growth throughout this period, with no evidence of the Bronze Age population bottlenecks detected from Y chromosomes.29 This is not what one would expect from averaging the mitochondrial DNA and Y chromosomes. Instead, it is clear that the Y chromosome was a nonrepresentative part of the genome where certain genetic types were more successful at being passed down to later generations than others. In principle, one possible explanation for this is n
atural selection, whereby some Y chromosomes gave a biological advantage to those who carried them, such as increased fertility. But the fact that this genetic pattern manifested itself around the same time in multiple places around the world—in a period coinciding with the rise of socially stratified societies—is too striking a pattern to be explained by natural selection at multiple independently occurring advantageous mutations. I think a more plausible explanation is that in this period, it began to be possible for single males to accumulate so much power that they could not only gain access to large numbers of females, but they could also pass on their social prestige to subsequent generations and ensure that their male descendants were similarly successful. This process caused the Y chromosomes these males carried to increase in frequency generation after generation, leaving a genetic scar that speaks volumes about past societies.
It is also possible that in this period, individual women began to accumulate more power than they ever had before. Yet because it is biologically impossible for a woman, even a very powerful one, to have an extremely large number of children, the genetic effects of social inequality are much easier to detect on the male line.
Sex Bias in Population Mixture
There are many ways that populations come together—for example through invasions, migrations into each other’s homelands, demographic expansion into the same territory, and trade and cultural exchange. Potentially, populations could mix as equals—for example through the overlapping of two equally resourced populations moving peaceably into the same area. But much more often there is asymmetry in the relationship, as reflected in mixture involving males from one group and females from the other, as occurred in the history of African Americans and in the history of the Yamnaya. The different histories of men and women recorded in different parts of the genome make it possible to study this mixture, and thereby to obtain clues about cultural interactions that occurred long ago.
Some of the examples of sex bias evident from genetic data are truly ancient. Take for example the founding of the ancestral population of non-Africans. Any genetic analysis of non-Africans reveals evidence of a population bottleneck dating to some time before fifty thousand years ago—that is, a small number of individuals giving rise to many descendants today. In 2009, I worked with Alon Keinan, a postdoctoral scientist in my laboratory, to compare genetic variation on the X chromosome, the larger of our two sex chromosomes, to the rest of the genome. To our surprise, we found much less genetic variation in non-Africans on chromosome X than would be expected from the level of variation in the rest of the genome, assuming that males and females participated equally in the founding of the ancestral population of non-Africans. The pattern was too extreme to be explained by a simple scenario of more men than women participating in the founding of the ancestral population of non-Africans. But we discovered that one scenario that could explain the pattern is that after the initial budding off of the ancestral population of non-Africans, there was genetic input into this population from males of other groups. Since males carry one copy of chromosome X for every two copies of other chromosomes, a process of repeated waves of male immigration would decrease X-chromosome diversity (meaning that there would be less genetic variation in the population) compared to the rest of the genome, producing the pattern we observed.30
This hypothesis gains some plausibility from what we know of the interaction of central African Pygmy hunter-gatherer populations with the Bantu-speaking agriculturalist populations that surround them. When the Bantu first expanded out of west-central Africa several thousand years ago, they had a profound influence on the indigenous rainforest hunter-gatherer populations they encountered, as is evident from the fact that today no Pygmies speak a non-Bantu language and all harbor substantial Bantu-related ancestry. Even today, the overwhelming pattern is that Bantu men mix with Pygmy women and the children are raised in Pygmy communities.31 The waves of Bantu-related gene flow into the Pygmy population are similar to the scenario that Keinan and I had suggested for the ancestral population of non-Africans. The genetic consequence of this anthropological pattern is reflected in Pygmies having a substantially reduced degree of genetic diversity on chromosome X compared to the expectation from the rest of the genome.32 Perhaps similar processes were at work in the shared history of non-Africans, explaining the reduced X-chromosome diversity relative to the rest of the genome in that case too.
Evidence of sex bias in the mixture of human populations is becoming commonplace. The male-biased European contribution to admixed populations in the Americas is stark in African Americans, but it is truly extraordinary in populations in South and Central America, reflecting stories like that of Hernán Cortés and La Malinche. Andrés Ruiz-Linares and colleagues have documented how in the Antioquia region of Colombia, which was relatively isolated between the sixteenth and nineteenth centuries, about 94 percent of the Y chromosomes are European in origin, whereas about 90 percent of the mitochondrial DNA sequences are of Native American origin.33 This reflects social selection against Native American men. Because nearly all the male ancestry comes from Europeans and nearly all the female ancestry comes from Native Americans, one might naively expect that the people of Antioquia today would derive about half their genome-wide ancestry from Europeans and half from Native Americans, but this is not the case. Instead, about 80 percent of Antioquian ancestry comes from Europeans.34 The explanation is that Antioquia was flooded by male migrants over many generations. The first European men to arrive mixed with Native American women. Additional European male migrants came later. Through repeated waves of male European migration, the proportion of European ancestry kept increasing everywhere in the genome except for mitochondrial DNA, because mitochondrial DNA is passed to the next generation entirely by females.
Massive sex bias in population mixture also occurred between four thousand and two thousand years ago during the formation of the present-day populations of India.35 As discussed in part II, endogamous groups in India with traditionally higher social status tend to have more West Eurasian–related ancestry than groups with traditionally lower social status,36 and the effect is highly sex-biased, as mitochondrial DNA tends to be largely of local origin, whereas a much higher proportion of Y-chromosome types have affinity to West Eurasians.37 This pattern plausibly reflects a history in which males of West Eurasian–related ancestry were more highly placed in the caste system and sometimes married lower-ranking females. It speaks to a dramatic coming together of socially unequal populations to form the present genetic structure of India.
DNA has the power to overturn expectations from other fields, though, and in this case it has also revealed a surprise about sex-biased mixture. Today, almost every Pacific island population harbors some of its ancestry from people of mainland East Asian origin. As described in part II, this ancestry derives from people whose ancestors originated on Taiwan island and who invented long-distance seafaring and used it to disperse their people, language, and genes. But almost every Pacific island population also harbors Papuan ancestry related to the indigenous hunter-gatherers of the island of New Guinea. Surprisingly, in light of the theme that males from an expanding population tend to mix with local females, initial studies of mitochondrial DNA and Y chromosomes showed that the mixed populations of the Pacific today derive most of their East Asian origin DNA not from male but from female ancestors.38
One explanation that has been suggested for this pattern is that in early Pacific island societies, property usually passed down the female line and males were the primary people who moved across islands.39 But there is another process that may also have contributed. As described in part II, my laboratory showed that the first people of the open Pacific had little Papuan-related ancestry.40 We showed that later west-to-east waves of migration of people of mixed Papuan and mainland East Asian ancestry explain the ubiquity of Papuan ancestry in the remote Pacific today. If males from this later-arriving population had social advantages relative to the previously resident
population, this could have resulted in newly arriving males of primarily Papuan ancestry mixing with previously established females of primarily East Asian–related ancestry.
The Pacific islander example highlights the importance of not simply assuming that genetic analyses of sex-biased events will fulfill expectations from anthropology. Now that the genome revolution has arrived, with its power to reject long-standing theories, we need to abandon the practice of approaching questions about the human past with strong expectations. To understand who we are, we need to approach the past with humility and with an open mind, and to be ready to change our minds out of respect for the power of hard data.
The Future of Genetic Studies of Inequality
At the present time, our methods for using genetic data to study sex bias in human history are frustratingly primitive. Many of the most interesting findings about sex bias so far have been based on just two locations in the genome, the Y chromosome and mitochondrial DNA, which reflect only tiny fractions of our family trees. Studies of sex-biased population dynamics using these sections of the genome become nearly useless for understanding events that occurred more than around ten thousand years ago, because at that time depth, everyone in the world descends from only a small handful of male and female ancestors who are too few in number to support a statistically precise measurement of sex bias.