And truth shall come out of the mists of legend

By Razib Khan | October 11, 2013 11:08 am

And he will be a wild man; his hand will be against every man, and every man’s hand against him; and he shall dwell in the presence of all his brethren.

– Genesis 16:12

By now you may have seen or read two important papers which just came out in Science, 2000 Years of Parallel Societies in Stone Age Central Europe, and Ancient DNA Reveals Key Stages in the Formation of Central European Mitochondrial Genetic Diversity. The details have been extensively explored elsewhere. If you don’t have academic access I highly recommend the supplement of the second paper. It’s also very illuminating if you don’t have a good grasp of the nuts and bolts of archaeology (I do not). I can’t, for example, confirm whether the merging strategies of different archaeological cultures were appropriate or not, because I’m not totally clear in my own head about the nature of these distinct archaeological ‘cultures’ (quotations due to the fact that archaeologists infer culture from material remains, and so they may not be cultures in the sense we understand culture). But the overall finding is clear, in ancient Europe thousands of years ago there were multiple demographic replacements and amalgamations. The post-World War II thesis in archaeology that one could not infer changes in the demographic character from material remains (because the latter can diffuse purely through memetic means) seems to be false. The correspondence is surprisingly tight.

In its broad outlines this was clear before these papers emerged. There is very little I would change from my post The last days of Grendel. This confusing welter of societies in prehistoric Europe is hard for us to conceive of (or reconstruct with any plausibility) today, and as one of the authors of the broader mtDNA paper observes you could not infer this pattern of replacements based on modern patterns of variation. Phylogeography inferring the past from present distributions of variation clearly has limitations, because it is constrained by the necessity to adhere to parsimony in the absence of a dense enough data set. In the early aughts the argument was between scholars who adhered to a more dominant role for demographics in transferring the farming lifestyle (L. L. Cavalli-Sforza et al.)  from the Neolithic societies of the Middle East to Europe, and those who pushed forward the thesis of cultural diffusion (Sykes et al.). These are obviously stylized extreme positions, but they capture the essence of the dispute in regards to how cultures transform and expand. Scholars looked at present day European and Middle Eastern populations, and compared their genetic relatedness, usually with male and female lineages (Y and mtDNA).

There was a major problem with this model: the ancient DNA we have is telling us that present population genetic distributions are poorly correlated with past population genetic distributions. And, not only are the ancient populations of Europe rather well mixed and overturned, like a well tilled field, but it seems entirely likely that those of the Middle East are too. Therefore the methodology was bound to mislead from the get-go; the premise of a few major population movements was false. But there was I believe another major lacunae in our understanding: prehistoric people were not entirely atomized. Whether one believed in the central role of demographic movements or cultural transmission, both theses seemed to posit that prehistoric human populations were mobilizing and interacting mostly on a small scale. Diffusing. This seems likely to be wrong. Or at least it misses enough of the picture that it turns out to give a false impression.

To understand what I’m getting on, consider the American migration west in the 19th century. There were multiple forces at work. First, there was a real demographic pressure in many parts of the United States. New England for example was literally at capacity. It simply had no more land for subsistence agriculture which could support a larger population beyond the Malthusian limit. There were three primary responses. A transition up the “value chain” toward industry, made possible by the natural endowments of water power available in the region. Decreased total fertility rate (related to the first). And finally, a mass migration west, first to upstate New York, but then across the Great Lakes and out even to the Pacific. To a great extent these shifts can be modeled as individual (for family/firm) dynamics. People are responding rationally to changing incentives. But this misses “higher level” structural shifts.

As we all are now well aware the United State government entered into a massive program of ethnic cleansing and pacification of the native populations of the western territories, making migration a viable option. It acquired the western seaboard states through victory in war (California) or diplomatic bluster and coordinated demographic assault (Oregon and Washington). These events are linked to macroscale cultural dynamics, encapsulated in a slogan such as Manifest Destiny. Increasing the geographic scale of the model cultural and demographic changes in Europe itself also made itself felt in the United States (i.e., European migration to places such as the Midwest were important contributors to settlement of the nation, and this migration was often due to social and political dynamics in source nations). The reality of these macroscale dynamics means that demographic shifts often occurred in pulses, in a discontinuous fashion.

Credit: dbachman

Because prehistory is defined by the lack of writing from which we can draw detailed narratives, we will always be in the dark as to the specific macroscale dynamics which resulted in the cultural and genetic shifts we infer (barring the development of time machines). But, we can at least construct a correct framework get a true flavor of the context of how humans interacted in the past. As I have stated elsewhere, I believe that once autosomal and Y chromosomal results come on line (mtDNA is more copious and so easier to extract) we will see that many of the discontinuities and shifts are actually attenuated in the female lineage. What I mean here is that the picture from these papers may actually be less radical than the real shifts truly were. In India the source populations for admixture were distinct enough that it seems clear that admixture was male-mediated. West Eurasian Y chromosomal lineages are more well represented fractionally than West Eurasian autosomal ancestry, which is more well represented than West Eurasian mtDNA. The whole zone from West Asia out toward Atlantic Europe was more of a continuum, so solid inferences will have to wait on the ancient DNA.

Finally, one last big picture aspect which I think is important to note is that the genetic distances between ancient populations across small spatial scales was very large. I suspect that with the rise of agriculture, and imperial states, we have seen a massive process of genetic requilibration across vast swaths of Eurasia in particular. Though I think we mislead ourselves if we view prehistory purely as an affair of small scale bands with vague higher order structure, it is still the fact that the scale was smaller than what came later. That leads me to conclude that population genetic diversity as a function of distance in the far past was likely greater than it has been across most of recorded history. So inferences about the character of human genetic diversity derived from contemporary variation is misleading.* The large differences between Bushmen populations may be highly representative of what was the norm in the past.

* To be clear, Fst between continental populations may be the same. But Fst over small scales may have been larger.

CATEGORIZED UNDER: Anthroplogy, Archaeology
  • Karl Zimmerman

    I find myself wondering if the best model for the persistence of hunter-gatherers in Neolithic Europe for 2,000 years following the introduction of agriculture is Pygmies in Sub-Saharan Africa.

    The Bantu migration into the core pygmy area happened approximately 2,500 years ago. Nonetheless, the pygmies were not wiped out or dispersed to the fringes of the Bantu expansion, as was the case for the Khoisan (and whatever other extinct peoples used to live in Southern Africa). Instead they fell into a semi-symbiotic relationship with Bantu and other agriculturalists, trading hunted/foraged items for agricultural products, and ultimately losing their own languages.

    • razibkhan


  • andrew oh-willeke

    It is also appropriate to note that the highly distinct hunter-gatherer and farmer populations would not just differ in food production methods.

    They would have been very obviously racially distinct (at least as much if not more than the an Australian Aborigine v. a modern Western European in genetic difference as measured by FST implying very obvious visual differences as well), as well as having profoundly unrelated languages (perhaps as different as Mandarin v. English). Surely, they would also have differed profoundly in religious beliefs and moral/ethical values as well.

    A movie like Avatar (the blue alien one) are probably closer to the spirit of the differences between farmers and hunters in this time period than some of the more PC quasi-historical dramas trying to capture the experience.

    • Onur

      They would have been very obviously racially distinct (at least as much if not more than the an Australian Aborigine v. a modern Western European in genetic difference as measured by FST implying very obvious visual differences as well)

      Physical anthropology and genetics both show that both the hunter-gatherers and agriculturists of West Eurasia were racially Caucasoid. The racial differences between the two groups were within the limits of the variation of the Caucasoid race. That is why it is much harder to differentiate between hunter-gatherers and agriculturists in West Eurasia than between those of South Asia, where the two groups seem to have been racially quite distinct (the ANI-ASI divergence).

  • ryanwc

    One caveat that only affects your theory on the edges – the basis for describing some of these populations in the ‘2000 years of coexistence’ paper as hunter-gatherer was isotope analysis that suggested a “forager and freshwater fish diet.” Depending on how much freshwater fish, these may be settled populations with similar levels of technology to the agricultural people.

  • Supercessionist

    Mr. Khan, your thesis is safe. Many of your fellow geneticists even, for example, think of the Irish and Scottish as ‘Celts’ when they are no more so than the Basques (granting I have a better grasp on the archaeology and history and I’m sure you knew this). Language is no way to go in terms of plotting genetic change per se, and one would think everyone would appreciate this with regard to cultural artifacts. Hence “The Celts” may not have been particularly closely related, but shared a culture and eventually, mostly, a language, and when they conquered more settled pastoral people, it was like drops into full buckets and may have been more than a slow admixture than a violent invasion or invasions.
    You might find ancient Irish myth pretty interesting as it might relate to Western Europe.
    Apropos of little, I have some interest on the Khazarian hypothesis and appreciated your article. I do think that some of your criticisms, while technically valid, don’t actually disprove the broader point – genetic contributions, significant ones, that are not Levantine. It would be interesting to see cultural/linguistic analysis. My own sense is that the Khazarian contribution was not large, but not insignificant – that was a major power, and the evidence for conversions is indubitable, the question is to what extent it trickled down to the plebs.
    Lastly, the recent paper on the largely European maternal ancestry of Ashkenazi Jews seems very difficult to argue with {less shaky than the Khazarian paper}. I have not seen very much news coverage at all – which tells me that *politics* may be holding up or frustrating science on this.
    I’m not sure why any of this matters other than notions that rights to land or title or the kingdom to heaven depend on genetics… but a walk through Williamsburg, Brooklyn will vouchsafe that the Ultra Orthodox, who largely came from Poland and the Baltics, are not Middle Eastern in appearance.
    This, of course, isn’t a scientific conclusion, but the notion of direct lineage stems, in part, from the idea that conversion didn’t happen. It did – all the time, and both ways, but much less since, I’d argue, the Reformation…
    Anyway, lastly I’ll note that Tay Sachs occurs in not only Jews, but Gaels (Irish and Scots and in a practical sense, Welsh). Curious, isn’t it. Might this be some pre-Indo-European marker picked up in central Europe 10k or more years ago? Not marker but.. trait…

    • razibkhan

      1) you’re too loquacious for the thin data density of what you’re saying. t he basque/irish connection was based on Y chromosomes was pretty much superseded by autosomal work years ago. keep up on the literature.

      2) have not seen very much news coverage at all –

      you don’t count the new york times?

      i agree there’s politics at work. but hasn’t stopped major media coverage.

    • Paul Conroy

      The problem is that a lot of people claim to be “Irish” and few are really “Native Irish” – as in all known ancestry from before the Normans. My father is one such person – and the only one that I know of, He consistently shows more Scandinavian-like and more Basque-like ancestry than every other “Irish” person, whose results are a vailable.

      Here are his recent Admixture results from Eurogenes K16

      Admix Results (sorted):

      # Population Percent
      1 Atlantic 37.20
      2 North_Sea 32.89
      3 Baltic 10.74
      4 West_Med 8.50
      5 Eastern_Euro 8.13
      6 West_Asian 2.54

      # Primary Population (source)Secondary Population (source)Distance
      1 80.9%Irish+19.1%French_Basque@3.76
      2 78.9%West_Scottish+21.1%French_Basque@3.84
      3 77.6%Danish+22.4%French_Basque@4.53
      4 83.6%Southeast_English+16.4%French_Basque@5.26
      5 83.4%Irish+16.6%Southwest_French@5.3
      6 84.7%Irish+15.3%Spanish_Aragon@5.34
      7 80.5%West_Scottish+19.5%Southwest_French@5.46
      8 82.1%West_Scottish+17.9%Spanish_Aragon@5.55
      9 86%Irish+14%Spanish_Castilla_La_Mancha@5.57
      10 86.2%Irish+13.8%Spanish_Valencia@5.61

      Using 4 populations approximation:
      1 French_Basque + West_Scottish + West_Scottish + West_Scottish @ 4.924
      2 French_Basque + Irish + West_Scottish + West_Scottish @ 4.944
      3 French_Basque + Irish + Irish + West_Scottish @ 4.995
      4 French_Basque + Irish + Irish + Irish @ 5.074
      5 Danish + French_Basque + West_Scottish + West_Scottish @ 5.179
      6 Danish + French_Basque + Irish + West_Scottish @ 5.193
      7 Danish + French_Basque + Irish + Irish @ 5.237
      8 Danish + Danish + French_Basque + West_Scottish @ 5.473
      9 Danish + Danish + French_Basque + Irish @ 5.483
      10 French_Basque + Irish + Irish + Orcadian @ 5.494

    • Paul Conroy

      BTW, my father and I, and my oldest daughter are Tay-Sachs carriers

  • Generalista


    I agree with some of what you say, but not with all. Most importantly, European autosomal DNA is extremely close to each other – which to me means that the Gravettian had a unifying consequence, and everything after (agriculture, Bronze age, iron age) had firstly only a mild impact and secondly, as such, a broad one.

    Uniparental markers don’t tell you much, at all. Where are those that distinguish the Finns and Hungarians, in Europe? Autosomally, Hungarians are extremely close to Germans, today, and the latter also to anyone else in that vicinity. How can this be?

    It just is. You can take DNA samples in a small village in Central Germany, today, and get three people that share as much autosomal DNA as you would expect from almost-relatives – but one has y-DNA haplogroup R1b, one has I, the third has R1a. And from ancient DNA we know it has been so for probably over 4,000 years, and likely even longer.

    Uniparental DNA differences clearly exaggerate autosomal differences.

    As to the “HGs,” yes, we know that many of them were actually settled, or at least seasonally settled.

    • razibkhan

      Most importantly, European autosomal DNA is extremely close to each other – which to me means that the Gravettian had a unifying consequence,

      probably not, though perhaps. the gravettian is old, the drift parameter probably would have diversified over 20-30 K years. more likely the homogenization is due to recent admixture. this seems evident in the few autosomal ancient sequences we have today from HGs. some of it is published. some of it will be by pontus skoglund et al. in the near future. if you want to make a bet, i’m game :-)


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Gene Expression

This blog is about evolution, genetics, genomics and their interstices. Please beware that comments are aggressively moderated. Uncivil or churlish comments will likely get you banned immediately, so make any contribution count!

About Razib Khan

I have degrees in biology and biochemistry, a passion for genetics, history, and philosophy, and shrimp is my favorite food. In relation to nationality I'm a American Northwesterner, in politics I'm a reactionary, and as for religion I have none (I'm an atheist). If you want to know more, see the links at


See More


RSS Razib’s Pinboard

Edifying books

Collapse bottom bar