by David Reich
The agriculturalist expansion that had the greatest impact on Africa is the one associated with people who speak languages of the Bantu family.14 Archaeological studies have documented how beginning around four thousand years ago, a new culture spread out of the region at the border of Nigeria and Cameroon in west-central Africa. People from this culture lived at the boundary of the forest and expanding savanna and developed a highly productive set of crops that was capable of supporting dense populations.15 By about twenty-five hundred years ago they had spread as far as Lake Victoria in eastern Africa and mastered iron toolmaking technology,16 and by around seventeen hundred years ago they had reached southern Africa.17 The consequence of this expansion is that the great majority of people in eastern, central, and southern Africa speak Bantu languages, which are most diverse today in present-day Cameroon, consistent with the theory that proto-Bantu languages originated there and were spread by the culture that also expanded from there around four thousand years ago.18 Bantu languages are a subset of the larger Niger-Kordofanian family spanning most of the languages of West Africa,19 which likely explains why today the frequencies of mutations in groups in Nigeria and in Zambia are more similar than the frequencies of mutations in Germany and Italy despite the former two countries being separated by a far greater geographic distance.
Ultra-sensitive genetic methods, which can detect shared relatives of pairs of individuals in the past few thousand years, have now made it possible to learn something about the geographic path of the Bantu expansion. The genetic variation in Bantu speakers in East Africa is more closely related to the genetic variation in Malawi to the south of the Central African rainforests than it is to genetic variation in Cameroon.20 This suggests that the initial Bantu expansion was largely to the south and that the movement to East Africa was a later expansion from a southern staging ground. This contrasts with the theory of a direct eastward movement from Cameroon, a theory that had been plausible prior to the genetic data.
Figure 26. Today, West African–related ancestry is predominant in eastern and southern Africa due to the Bantu expansion of the last four thousand years.
Another agricultural expansion that had a profound impact is the one that spread Nilo-Saharan languages, spoken by groups from Mali to Tanzania. Many Nilo-Saharan speakers are cattle herders, and a common view is that the Nilo-Saharan expansion was driven by the spread of farming and herding in Africa’s dry Sahel region during the expansion of the Sahara Desert over the last five thousand years. One important branch of Nilo-Saharan is the Nilotic languages, which are mostly spoken by cattle herders along the Nile River and in East Africa, including the Maasai and Dinka. The genetic data make it clear that Nilotic-speaking herders were not always socially disadvantaged relative to farmers in the frontier regions where they encountered each other. For example, the Luo group of western Kenya (to which former U.S. president Barack Obama’s father belonged) are a primarily farming people who speak a Nilotic language. But George Ayodo, a Luo scientist from Kenya who spent time in my laboratory, found that the mutation frequencies in the Luo are much more similar to those of the majority of Bantu speakers, likely reflecting a history in which a Bantu-speaking group in East Africa adopted a Nilo-Saharan language from its high status neighbors.21
The African language expansion whose origin is most unclear is the one associated with Afroasiatic languages. They are most diverse in present-day Ethiopia, which throws weight behind the theory that northeastern Africa was the homeland of the original speakers of these languages.22 But the Afroasiatic language family also contains a branch localized to the Near East that includes Arabic, Hebrew, and ancient Akkadian. It has been hypothesized on this basis that the spread of Afroasiatic languages, or at least some branches of them, could have been related to the spread of Near Eastern agriculture,23 which introduced barley and wheat and other Near Eastern crops into northeast Africa up to seven thousand years ago.24 New insights are already emerging from ancient DNA, which makes it possible to document ancient migrations between the Near East and North Africa that could have spread languages, culture, and crops. In 2016 and 2017, my laboratory published two papers showing that a shared feature of many East African groups, including ones that do not speak Afroasiatic languages, is that they harbor substantial ancestry from people related to farmers who lived in the Near East around ten thousand years ago.25 Our work also found strong evidence for a second wave of West Eurasian–related admixture—this time with a contribution from Iranian-related farmers as might be expected from a spread from the Near East in the Bronze Age—and showed that this ancestry is widespread in present-day people from Somalia and Ethiopia who speak Afroasiatic languages in the Cushitic sub-family. So the genetic data provide evidence for at least two major waves of north-to-south population movement in the period when Afroasiatic languages were spreading and diversifying, and no evidence of south-to-north migration (there is little if any sub-Saharan African related ancestry in ancient Near Easterners or Egyptians prior to medieval times).26 Genes do not determine what language a person speaks and so genetic data cannot by themselves determine how languages spread, and thus cannot provide definitive evidence in favor of one theory or another about whether the ultimate homeland of Afroasiatic languages was sub-Saharan Africa, North Africa, Arabia, or the Near East. But there is no question that the genetic data increase the plausibility of a Near Eastern agriculturalist source for at least some Afroasiatic languages, and the genetic findings raise the question of what languages were spoken by these north-to-south migrants.
The fourth great agriculturalist expansion in Africa is the one associated with the Khoe-Kwadi languages of southern Africa. Like the two language groups spoken by hunter-gatherer groups in the south of Africa—Kx’a and Tuu—Khoe-Kwadi languages are characterized by click sounds. Based on shared words for herding, it has been hypothesized that Khoe-Kwadi languages were brought from East Africa by cattle herders who came to southern Africa after eighteen hundred years ago and who may also have picked up click sounds from local populations.27 The genetic data support the hypothesis of a major genetic contribution by East Africans to Khoe-Kwadi-speaking populations today. In 2012, Joseph Pickrell in my laboratory showed that Khoe-Kwadi speakers share a disproportionate amount of their ancestry with Ethiopians compared to the Kx’a and the Tuu, as might be expected from a migration from the north.28 The size of East African–derived DNA segments in some of the Khoe-Kwadi–speaking populations is what would be expected from mixture eighteen to nine hundred years ago with a ghost herder population, consistent with the arrival of herders around this time and a delay before the mixture with local populations was complete. Within the segments matching East Africans, Pickrell found even smaller segments that matched Near Easterners more than they did any other populations, and that had lengths expected for an average mixture date of around three thousand years ago. That is the average date of mixture between people of West Eurasian–related ancestry and sub-Saharan ancestry in many groups in Ethiopia,29 so this finding provides further support for the hypothesis of an East African source.
Ancient DNA has now verified this hypothesis. In 2017, Pontus Skoglund analyzed ancient DNA from the approximately thirty-one-hundred-year-old remains of an infant girl from Tanzania in equatorial East Africa, and an approximately twelve-hundred-year-old sample from the western Cape region of South Africa, both buried among artifacts and animal bones that identified them as being from herder populations.30 The Tanzanian girl was a member of the ghost herding population that Pickrell and I had predicted: a group that derived most of its ancestry from ancient East African hunter-gatherers, and the remaining part from an ancient West Eurasian–related population. This population almost certainly played a major role in spreading cattle herding from the Near East and North Africa across sub-Saharan Africa. Our ancient DNA evidence from the southern African herder also strongly supported this idea, showing that this individual derived about one-third of her ancestry from the pastoralist popu
lation of which the Tanzanian girl was a part, and her remaining ancestry from local groups related to present-day San hunter-gatherers. The mixture of ancestries in the twelve-hundred-year-old southern African herder was very similar to that in present-day Khoe-Kwadi speakers, many of whom are herders, supporting the theory that early Khoe-Kwadi languages, herding, and this type of East African ancestry all spread to southern Africa through a movement of people.
The landscape of human biological and cultural diversity in Africa today, dominated as it is by the effects of the agricultural expansions of the last few thousand years, is extraordinary, but it is also distracting if one’s interest is in understanding the big picture of what happened. A trap that researchers of African genetics, archaeology, and linguistics repeatedly fall into is celebrating Africa’s present-day diversity, epitomized by a slide showing the faces of people from across the continent who look very different from each other that many of us use when presenting on Africa. It is tempting to think that in order to comprehend deep time in Africa we need to be able to hold all of that diversity in our heads and explain all of it at once. But most of the present-day population structure of Africa is shaped by the agricultural expansions of the past few thousand years, and so focusing on describing Africa’s mesmerizing diversity paradoxically does the project of understanding the big picture of humans in Africa a disservice just as much as focusing on the common origins of all modern humans in Africa does Africa a disservice. We need to stop focusing on describing the veil and instead rip it away, and for this we need ancient DNA.
Reconstructing Africa’s Forager Past
Who lived in Africa before the expansion of food producers, the people who so profoundly transformed the human landscape of the continent? Answering this question is extraordinarily difficult based on patterns of present-day variation. In the introduction to this book, I described how Luca Cavalli-Sforza made a bet in 1960 that it would be possible to reconstruct the deep history of human populations based entirely on patterns of genetic variation in present-day groups.31 However, he lost his bet, as ancient DNA has revealed that there has been so much migration and population extinction that in most instances it is very difficult even with sophisticated statistical methods to recover the details of ancient demographic events from the traces left behind in the DNA of present-day people.
The breakthrough that is making it possible to get beyond this impasse will not be surprising to the reader. It is genome-wide ancient DNA, which can be coanalyzed with data from groups that have been genetically and culturally isolated compared to their neighbors, among them the Pygmies of Central Africa, the San hunter-gatherers of the southern tip of Africa, and the Hadza of Tanzania, whose languages with clicks are very different from the languages of the Bantu who surround them and whose genetic ancestry is highly distinctive as well. Some of these populations harbor genetic lineages that are highly divergent from their neighbors. We can compare data from these ancient samples to probe events that occurred deeper in time than those that can be accessed only by analyzing the DNA of present-day populations.
Well-preserved ancient DNA has until recently been hard to find in most parts of Africa because of the hot climate, which accelerates chemical reactions that degrade DNA. But in 2015, the ancient DNA revolution finally arrived in Africa because of improvements in the efficiency of DNA extraction techniques and a better understanding of which bones yielded the most DNA.
The first genome-wide ancient DNA data from Africa came from a forty-five-hundred-year-old skeleton found in a highland cave in Ethiopia.32 This ancient individual was much more closely related to one group living in Ethiopia today, the Ari, than to many others. Today there is an intricate caste system that shapes the lives of many people within Ethiopia, with elaborate rules preventing marriage between groups with different traditional roles.33 The Ari include three subgroups—the Cultivators, Blacksmiths, and Potters—who are socially and genetically differentiated from one another and from non-Ari groups.34 Since the Ari have a distinctive genetic affinity to the forty-five-hundred-year-old ancient highland individual compared to other Ethiopian groups, it is clear that there were strong local barriers to gene exchange and homogenization within the region of present-day Ethiopia that persisted for at least forty-five hundred years. This is the best example of strong endogamy that I know of—even more ancient than the evidence of endogamy in India that so far is only documented as going back a couple of thousand years.35
Ancient DNA keeps surprising us. In 2017, Pontus Skoglund in my laboratory analyzed sixteen individuals from Africa: foragers and herders from South Africa who lived between about twenty-one hundred and twelve hundred years ago, foragers from Malawi in southern Africa who lived between about eighty-one hundred and twenty-five hundred years ago, and foragers, farmers, and herders from Tanzania and Kenya who lived between about thirty-one hundred and four hundred years ago.36 While these individuals are very recent compared to some of the oldest Eurasian ancient DNA, they nevertheless provide insights into African population structure before the arrival of the food producers who transformed much of Africa’s human geography.
A great surprise that emerged from our ancient DNA analysis was that there was evidence of a ghost population dominating the eastern seaboard of sub-Saharan Africa that appears to have been largely displaced by the expansion of agriculturalists.37 This population, which we called the “East African Foragers,” contributed all of the ancestry of two ancient hunter-gatherer genomes in our dataset from Ethiopia and Kenya, as well as essentially all of the ancestry of the present-day Hadza of Tanzania, who today number fewer than one thousand. We also found that the East African Foragers were more closely related to non-Africans today than they were to any other groups in sub-Saharan Africa. The close relationship to non-Africans suggests that the ancestors of the East African Foragers may have been the population in which the Middle to Later Stone Age transition occurred, propelling expansions outside of Africa and possibly within Africa too after around fifty thousand years ago. So the population that became the East African Foragers had a pivotal role in our history.
The East African Foragers were not a homogeneous population. This is evident from the fact that our data include at least three distinct East African Forager groups within Africa—one spanning the ancient Ethiopian and ancient Kenyan, a second contributing large fractions of the ancestry of the ancient foragers from the Zanzibar Archipelago and Malawi, and a third represented in the present-day Hadza.38 Based on the sparse data we had, we were not able to determine the date when these groups separated from one another. But given the extended geographic span and the antiquity of human occupation in this region, it would not be surprising if some of the differences among these groups dated back tens of thousands of years. There is precedent for such separations within forager populations in Africa. In 2012, my laboratory and another showed that a group that I think of as the “South African Foragers”—a lineage that is as divergent from the East African Foragers as any present-day human population—contained within it two highly divergent lineages that separated from each other at least twenty thousand years ago.39 East Africa is at least as rich a human habitat as southern Africa, and it would not be surprising if separations among foragers in East Africa were at least as old.
The second surprise was our discovery that some of our ancient African forager population samples shared ancestry from both South African Forager lineages and East African Forager lineages. Today, South African Forager lineages are essentially entirely restricted to southernmost Africa, where they form an important part of the ancestry of nearly all of the populations that use languages with clicks in them, and where they contribute almost all the ancestry of present-day San foragers as well as the ancient forager genomes we generated from southern Africa. But our ancient samples show that the term “South African Forager” may be misleading about where the ancestral population of this group arose. Two approximately fourteen-hundred-year-old individuals from Zanzibar
and Pemba islands off the coast of Tanzania—an island chain that separated from the mainland approximately ten thousand years ago as sea levels rose and thus plausibly harbors isolated descendants of a forager population that lived in East Africa around that time40—were a mixture of approximately one-third South African Forager–related ancestry and the remainder East African Forager ancestry.41 A series of seven samples from three different archaeological sites in Malawi in south-central Africa, which we dated to between about eighty-one hundred and twenty-five hundred years ago, were part of a homogeneous population that harbored about two-thirds South African Forager–related ancestry and the remainder East African Forager ancestry. So South African Forager ancestry was in the past distributed over a much broader swath of the continent, making it hard to know where this ancient population originated.
Figure 27. Ancestry currently restricted to San hunter-gatherers of southern Africa (F) was once spread across eastern Africa at least to Tanzania. Ancestry currently restricted to the isolated Hadza hunter-gatherers of Tanzania (C) was once widespread too.
Ancient DNA is teaching us that the history of modern Africa has its roots in ancient population separations and mixtures even before the arrival of agriculture. Thus the human story in Africa is complex at all levels and at all time depths, as might be expected from the continent’s huge size, its varied landscape, and the antiquity of the presence of our species there. The ancient DNA revolution is only just getting a toehold in Africa. In the coming years, Africa will be fully included in the ancient DNA revolution, and data will arrive from remains from more locations and from deeper times. These data will surely transform and clarify our view of what happened in the deep African past.