Roman cultural history has almost no demographic imprint


Several friends have asked that I weigh in the recent dust-up between Nassim Nicholas Taleb and Mary Beard. I haven’t for a few reasons. First, I can’t really be bothered to go incognito and see every detail of Taleb’s argument, as he has me blocked on Twitter (he called me a fucking idiot or something at some point). Second, the passion around this topic has little to do with genetics or ancient history from what I can see, two topics which I am actually interested in. Rather, it’s more about contemporary geopolitics. This is interesting too, and I have opinions on that, but I try really hard to keep history and politics in separate silos unless I am explicit about the connections and relevance. That’s because I don’t see classical history as simply something instrumentally important for modern times, but interesting in and of itself (the same goes for population genetics).

And for what it’s worth, Mary Beard says the same in her conclusion to SPQR. The ancients were ancients. Let them be what they were.

That being said, as someone with knowledge sets in ancient history and historical population genetics, I will make a few statements and let others interpret them however they wish (to be frank, I’m not going to cede ground to any of the experts I’ve seen who have spoken on the intersection of these two knowledge sets, so I figured it was time to put something somewhere beside Twitter).

* The prior probability that a Roman officer of the period in Britain would have visible black African ancestry (as seems clear by the cartoon, though no one has asked the cartoonists what their intent was) probability seems rather low. But it is non-zero, because a small minority of Roman subjects and citizens would have been defined as black by their physical appearance if they were alive today (they are mentioned passingly in the literature and texts from the period). Including in Britain.

* The probability conditional that he was based on an officer in Britain who was a native of Tidis is low, but higher. Several historians have pointed out in defense of the cartoon’s plausibility that there were many North Africans in Roman service, as well as prominent North Africans in Roman history (to name three of note, Septimius Severus, Tertullian, and St. Augustine). Whole tribes of what we’d today term Berbers enrolled in the Roman military a federates.

There are several separate issues to note. First, of the many North African genotypes I’ve seen detectable Sub-Saharan ancestry is found in almost all of them. But, many (most?) North Africans do not look visibly of Sub-Saharan African ancestry (see list of heads of states of Algeria). Second, both historical and genetic evidence indicates that this admixture from Sub-Saharan Africa is overwhelmingly (though not exclusively) from the period after Islam and the rise of a much bigger trans-Saharan trade (see Genomic Ancestry of North Africans Supports Back-to-Africa Migrations). Modern North Africa does have a large population today of people who are black or of obvious part-black ancestry, but this is due to the slave trade under Islam, and not antiquity.

* As evidence of the lack of non-European ancestry the paper The fine-scale genetic structure of the British population has been submitted. This is a great paper with best-of-breed methods and a massive data set of native English, with regional data. How do we resolve textual and archaeological evidence of people born outside of Britain during the Roman period in Britain with their lack of long-term genetic footprint among native modern Britains?

These sorts of questions need to be integrated in a broader context of the demography and genetics of antiquity that we have. On the whole looking at papers on modern and ancient DNA I am surprised by the lack of perturbation on the genetic structure attributable to the Roman period across Western Eurasia. I will offer two likely reasons that are related.

First, Classical civilization was an urban one, and the textual evidence we have is going to be highly skewed culturally in terms of our perception. The Roman world was predominantly written in cultured Latin and Greek (from what I have read the early translations of the Bible are indicative of a more pedestrian background of Christians due to the class markers of their lexical choices and idioms). But it was not necessarily spoken in cultured Latin and Greek across vast swaths of its territory. Even in St. Augustine’s time Punic was still spoken in the North African countryside, while the persistence and resurgence of Basque and Berber, and perhaps Brythonic Celtic in Britain, attest to vast reservoirs of people who were under the Roman peace, but not of it (also, the persistence of Albanian from a native Illyrian substrate). Because of the resources historians have on hand, text, there is going to be a major lacunae in our understanding and perception of the past. We hear the urban elites speaking to us. Not the rural majority.

Second, Classical civilization was an urban one, and this might have a major impact on the demographic consequences of migration. At any given size the effective breeding population is smaller than the census population, and the breeding population may not be representative of the overall population in terms of their genetic character. More specifically, it seems highly possible that the cosmopolitan urban Roman cities were massive demographic sinks. Rome before the Gothic Wars was a very populous city, not too far on the path of decline from its early imperial peak. But by the year 600 it had decreased its population to the point that vast swaths of the city were abandoned. Where did these people go? No doubt some of the elites scattered. Cassiodorus simply moved when barbarism came to his front step. But this was less possible for the urban proletariat. There is strong evidence that slaves in the ancient world were not replacing themselves reproductively due to brutality under which they lived. Some of the same was likely true of the urban proletariat.

* There is a difference between the inheritance pattern of culture and genes. In The Geography of Recent Genetic Ancestry across Europe this passage has always stuck out for me: “There is relatively little common ancestry shared between the Italian peninsula and other locations, and what there is seems to derive mostly from longer ago than 2,500 ya…The rate of genetic common ancestry between pairs of Italian individuals seems to have been fairly constant for the past 2,500 years, which combined with significant structure within Italy suggests a constant exchange of migrants between coherent subpopulations.”

The straightforward conclusion from this is that the Latinization of the Italian tribes and Magna Graecia occurred with no great demographic transformation. Modern Italy has within it the ghost of tribes long gone. This is notable because if you read the historical records of the Roman period you see evidence of trade, transport, and migration. But the genetic data would not lead you to this conclusion outside of Sicily and a few parts of Southern Italy.*

Above I have presented my reasoning for why this might be. But I think what it tells us that genetic data can informs us when there is a demographic turnover, and therefore a cultural turnover, but it will miss cultural turnovers which don’t have demographic impacts. These are many. To give a few examples, the rise of Islam in South Asia and Southeast Asia, the Latinization of the Western Mediterranean, the de-Latinization of Britain after the withdrawal of Roman legion and before the mass arrival of Saxons, and arrival of Buddhism in East Asia. All these are massive historical and cultural events, but they would not be visible in the genetic record.

If you want to learn about Roman history there are many books you could read. But I do recommend you try Bryan Ward-Perkins’ The Fall of Rome: And the End of Civilization. It’s a nice materialist take, and I think it gets to the underlying dynamics of institutional fragility of ancient civilization which was so easily wiped away by barbarism.

Addendum: The migration of the Slavs, Anglo-Saxons, and the Islamic Empires, all seem to differ from antiquity in having a major demographic impact. Why? In the case of massive institutional collapse, as in the first two cases, very old dynamics of inter-group competition arise, and famine probably does the rest of the trick. For Islam, it was a genuinely cosmopolitan civilization, with a more complex gradation between free and slave than in antiquity. Though it was quite brutal, African and Turkish slaves became free, and their genetic impact can be seen throughout the Islamic world.

* Like Spain, a substantial proportion of the Sicilian gene flow exchange is almost certainly due to the Islamic period. There are segments of North African and Sub-Saharan ancestry in Sicilians, albeit to a smaller extent than in Spain (in keeping with the shorter time period as part of the Islamic world).

29 thoughts on “Roman cultural history has almost no demographic imprint

  1. What’s your take on the date of Calabrians’ and Sicilians’ East Mediterranean affinities? They seem to be at least as Levantine-shifted as North African-shifted—some cluster with Sephardim and Cypriots.

  2. What’s your take on the date of Calabrians’ and Sicilians’ East Mediterranean affinities? They seem to be at least as Levantine-shifted as North African-shifted—some cluster with Sephardim and Cypriots.

    i think greek, roman era, and emirate all combined.

  3. Anyone know if these results have held up?:

    “Leach et al. (2009) provide evidence for intense foreign settlement. At one burial ground near Roman York, craniometric analysis revealed that 66% of the individuals clustered most closely with Europeans, 23% with sub-Saharan Africans, and 11% with Egyptians. At another, the proportions were 53% European, 32% sub-Saharan, and 15% Egyptian (Leach et al., 2009).”

    Leach, S., M. Lewis, C. Chenery, G. Müldner, & H. Eckardt. (2009). Migration and diversity in Roman Britain: A multidisciplinary approach to the identification of immigrants in Roman York, England, American Journal of Physical Anthropology, 140, 546-561

    http://evoandproud.blogspot.com/2010/07/african-community-in-roman-britain.html

  4. “Rather, it’s more about contemporary geopolitics”

    Geopolitics? Or do you mean identity politics?

    @Syonredux:

    Read your link and the following post…a lot of nonsense by Peter Frost there, citing Geoffrey of Monmouth as a historical source, lol. Still, I’d be interested in the current state of archaeological research as well.

  5. @ German Reader:”Read your link and the following post…a lot of nonsense by Peter Frost there, citing Geoffrey of Monmouth as a historical source, lol. Still, I’d be interested in the current state of archaeological research as well.”

    Yeah, there’s a lot of silly stuff there. As you said, though, I would be interested in informed commentary on the skeletal remains. Here’s a link to a Twitter discussion. Not exactly the best forum for this kind of thing, but better than nothing:

    https://twitter.com/caitlinrgreen/status/895350086267547648

  6. craniometric stuff seems to have low power to pick up differences. so i don’t trust it too much.

    re: geo/identity politics. struggled with how to say it. note that the stuff about how ancient romans and greeks were replaced by levantines was predicated on ideas that they were nordic whites. so in various forms this stuff has old roots, not just modern identity politics.

    in the medieval period you have tapestries depicting asian and african peoples as if they were european. hard tick for ppl to avoid.

  7. I read a bunch of those blog post summaries by professors Frost and Green about the discovery of craniums and teeth that imply the presence of sub-Saharan Africans as one of the founding populations of Great Britain but it all seems pretty thin especially since most of the mere handful of finds (4% of skulls at 79 locations) are from the time of Roman Britain. No DNA has been discovered and both professors do carefully point out that the dental oxygen isotopes from the Nile Valley and Delta are also found in Iberia. No doubt there were some blacks that came to Britain over the centuries but it all seems highly speculative wishful thinking as to their being large numbers of them that were around after the legions pulled out.

    “Further research is needed” is their conclusion and given today’s ideological climate I’m sure funding will be made available for a politically correct snipe hunt if the proposal is well written.

    https://youtu.be/5LNsASlXzkU

  8. Estimates of shared ancestry averaged from 4% in some places to 10-12% in the general population, the populations of the Canary Islands outputted up to 20% of shared ancestry with north Africans. The data are no totally consistent with the Islam domination of parts of the Iberian Peninsula as cause of the North African gene flow in South Western Europe. The North African gene flow to the Canary Island –geographically part of Africa- is from Neolithic times and very likeably the population in the moment of the Spanish conquest (between 1402 and 1496) were 100% from Maghreb origin.
    In the Iberian Peninsula, the pattern of the Maghreb ancestry seems to present an inconsistently predominance in the west, decreasing to east, with parts of western Andalusia (in the south) and Galicia (in the northwest) in the high percentage (20%). If western Andalusia had a very long period of Islam domination (from 711 to 1290 -in eastern Andalusia the Islamic presence extended to 1492), in Galicia the presence of Islam was very short (718 to 740). In a counterintuitive output, the Mediterranean Coast has the lowest percentage of shared ancestry with the Maghreb.
    Further, the “Moriscos” (descendants of Islamic believers) were expelled from Spain in 1609: estimates of the number expelled range between 275,000 and 300,000 Moriscos (or 4% of the total Spanish population at the time). This counter emigration not seems had have genetic consequences.
    Maybe movements of population in the Roman and also the Carthaginian Empires –which conquered Iberia and recruited and trained there the army of Hannibal- contributed also to the shared ancestry between the Iberian Peninsula and the Maghreb.

  9. @Razib Khan:
    “note that the stuff about how ancient romans and greeks were replaced by levantines was predicated on ideas that they were nordic whites. so in various forms this stuff has old roots, not just modern identity politics.”

    Yes, I know…though I’m doutbful this was a widespread view prior to the 18th and especially 19th centuries, it seems closely connected to modern nationalism/racialism.
    Re geopolitics, that just brings up very different connotations in my mind (its definition seems to be fuzzy, but think about people like Mackinder, Haushofer or Brzezinski…so more about the connection between geography and great power politics). Don’t know though if there’s a good term for what you describe.

  10. I don’t get this talk of Levantine and North African shifts of South Italians. Only Sicilians are 5% Arab/Berber and that’s about it. All Italians are shifted toward CHG and EHG (Indo-Europeans) relative to Neolithic Anatolians, not toward the Levant or Africa. Cypriots are the same (but more CHG) and Jews cluster close-by because they’re mostly Southern European.

    The recent Lazaridis paper showed that Greece, Cyprus, Albania, and Italy are almost identical to Mycenaeans and Minoans (even a little more “northern” in some cases), so obviously not much demographic change since the Bronze Age.

  11. What about the Maltese? This is a population that is culturally “Italianate” in many ways, but speaks a language that is, from my understanding, a vestige of what was spoken in Arab Sicily in the Middle Ages. I’m curious about their genetic heritage.

  12. The recent Lazaridis paper showed that Greece, Cyprus, Albania, and Italy are almost identical to Mycenaeans and Minoans (even a little more “northern” in some cases), so obviously not much demographic change since the Bronze Age.

    hm. most of the genetic character is due to the bronze age. not sure ‘not much’ though is totally accurate though this is semantic. PCA

    http://gnxp.nofe.me/wp-content/uploads/2017/08/nature23310-f1.jpg

    i’ve looked at a larger greek data set (4 grandparents born in greece) than lazaridis (private dtc company data). there is clear admixture cline to the balkans (i think the ‘italian’ cluster are greeks from the islands)

    https://gnxp.nofe.me/2015/08/14/greeks-with-slavic-ancestry-and-without/

  13. Re geopolitics, that just brings up very different connotations in my mind (its definition seems to be fuzzy, but think about people like Mackinder, Haushofer or Brzezinski…so more about the connection between geography and great power politics). Don’t know though if there’s a good term for what you describe.

    started with paul joseph watson. he’s concerned with migration.

  14. In the Iberian Peninsula, the pattern of the Maghreb ancestry seems to present an inconsistently predominance in the west, decreasing to east, with parts of western Andalusia (in the south) and Galicia (in the northwest) in the high percentage (20%).

    don’t assume it’s a simple cline due to time under muslim rule. eg in the early period the raw berbers were settled on the northern frontier.

  15. re: north african in spain. double checked http://www.pnas.org/content/110/29/11791.full

    1) west african correlates with north african ancestry. this means more likely it’s islamic period, since in north africa the LD decay dates to an islamic period

    2) the low bound time of admixture using variance btwn individuals is very low. probably too low. but suggests recent

    #1 is the stronger point…not sure about robustness of method used in #2.

  16. “started with paul joseph watson. he’s concerned with migration.”

    That’s just some guy on Youtube though, he can’t change the meaning of “geopolitics”.
    I think what you’re alluding to re medieval Europeans depicting non-Europeans with European features is the concept of the “somatic norm”, that is the physical characteristics a group regards as its own standard appearance, which can then be naively considered to be a universal norm (especially so in pre-modern societies).
    But maybe there isn’t a really satisfying terminology for those issues in all their complexity.

  17. @EvolutionistX

    I thought that although they almost certainly migrated from Italy in the dark ages, that Ashkenazi Jews show a significant proportion of both Germanic, Slavic, and in the case of the Hungarian Ashkenazim some Magyar DNA, given that the Ashkenazi Jews have become an indigenous European population over the 1000+ years that they have inhabited the continent.

  18. James,

    It’s true that Central and Eastern Euro ancestry in Ashkenazim has been buried to some degree under the Italian-heavy narrative of the last few years. But can you share some evidence for the Magyar claim? Never heard that. Plus, Jews arrived in Hungary in large numbers relatively late.

  19. That’s just some guy on Youtube though, he can’t change the meaning of “geopolitics”.

    what are you fucking talking about? i know what geopolitics is. i said geopolitics cuz watson’s tweet is what started the current debate (taleb was later). i’m alluding to the fact that all i know about watson is he talks in front of a map and is concerned with migration and was critiquing article in context of british being native.

    the original cartoon was influenced by identity politics. but that’s not the crux of the initial response.

  20. I thought that although they almost certainly migrated from Italy in the dark ages, that Ashkenazi Jews show a significant proportion of both Germanic, Slavic, and in the case of the Hungarian Ashkenazim some Magyar DNA, given that the Ashkenazi Jews have become an indigenous European population over the 1000+ years that they have inhabited the continent.

    i can double check, but i’ve never seen evidence of very much NW/NE european DNA in them. i’ve looked at a lot of jewish autosomes.

  21. @Razib Khan

    Upon further inquiry it appears that the matters of the size and the timing of European DNA into the Ashkenazi genome is currently a matter of recent research some current scientific debate but even the most conservative estimates being in the 12% range. At any rate every time I see a closeup pic of the Israeli political leadership there are a hell of a lot of blue and green eyes.

  22. bro, it’s around 50%. but it’s SW european. not northern european. that was my point.

    the conservative estimates there are bullshit. no way.

  23. Judging by the mtDNA results of the Gaunches, the aboriginal population of the Canary Islands, at least half if not more of the Sub-Saharan African ancestry in modern North Africans, Berber and Arabized Berbers in particular, predates the trans-Saharan slave trade. E1a, L3 and L2 were all present in those remains. That’s not to say that the slave trade didn’t have an impact… Egyptian Muslims are generally 3-6% more African than their Christian counterparts, who are still about ~10-15% African.

    Also, while groups like the Haratin are definitely relics of the slave trade, many “black” groups in the Sahara undoubtedly date to antiquity. The Bafour (the original inhabitants of the Western Sahara and Mauritania), Toubou of the Tibesti range, Nubians of the Upper Nile Valley, and Beja of the Eastern Desert are all notable examples.

  24. Razib, how does the islamic presence in Spain explain the fact that places that were occupied for much shorter periods of time (notably Galicia, Léon and Asturias) have higher NA and SSA ancestry than for instance Valencia? Because the former were under muslim control for else than 100 years and the latter for over 500

  25. Is there really much controversy about the genetics of the Ashkenazim? About 40-50% southern European (former East Roman Empire) and 50-60% Near Eastern. Not much, if any, mixture in from Germanic or Slavic populations.

  26. @Joe Q

    Apparently there is some mixture. While I stand corrected in my false assumption that the majority of Ashkenazi genome that is of European derivation was Slavic or Germanic, there is still debate over the details:

    “Another event post-dated the founder event and likely occurred in Eastern Europe.”

    http://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1006644

    and a more back of the envelope/scratch pad analysis:

    http://eurogenes.blogspot.com/2016/09/estonian-biocentre-human-genome.html

Comments are closed.