Substack cometh, and lo it is good. (Pricing)

The Orantes has not mixed much with the Tiber

In a moment of weakness I decided to read some of Mary Beard’s SPQR: A History of Ancient Rome. I say weakness because I want to wean myself off of excessive reading of Roman history, as in terms of inferential utility I’ve long reached diminishing returns. But I quite enjoy the topic, and so here I am.

The author is an excellent writer as well as a scholar, and I quite enjoyed Roman Triumph, so I am entirely not surprised that SPQR has me hooked. Some of my correspondents have exhibited some disdain toward it because of Beard’s attempts to draw some connections to present day mores and values from that of Rome, presumably with a progressive bent.

Myself, this does not bother me. I don’t come into reading about Rome as an ignorant, so I can sort that from the nuggets of fact and positivistic interpretation. In any case, I think of it rather like how Islamic philosophers viewed Aristotle through their own religio-cultural lens. Obviously this was an issue that caused resistance to the transmission of Aristotle to the Christian West, but ultimately it did not stop what was inevitable. At the end of the day it was more about Aristotle than the glosses.

Though I highly recommend SPQR (I’m halfway through), that’s not the point of this post. Going along I kept thinking about the section on the Etruscans. The Rasena. Their origins have a genetic connection that is clouded and uncertain right now. I would like to dig deeper into this issue in the future; no doubt some day it will be cleared up. But that day is not this day.

Modern Italians have more “Indo-European” admixture than they do “Middle Eastern”

Rather, I want to address the idea that modern Italians are genetically a distinct people from ancient Roman Italians. Because on that score we have the answers. Ultimately the idea that this is even a debate goes back to Juvenal:

It is that the city is become Greek, Quirites, that I cannot tolerate; and yet how small the proportion even of the dregs of Greece! Syrian Orontes has long since flowed into the Tiber, and brought with it its language, morals, and the crooked harps with the flute-player, and its national tambourines, and girls made to stand for hire at the Circus. Go thither, you who fancy a barbarian harlot with embroidered turban….

These comments are rooted in the reality that Rome during Juvenal’s period was quite a cosmopolitan city, with large numbers of Greeks and people from the Eastern Mediterranean who were Hellenized to various degrees (in the early 3rd century Rome was ruled by a family of Hellenized Syrians). We know this because we have plenty of observations and complaints, and there are a plethora of inscriptions and graffiti in the new languages.

In the 19th and early 20th century the ascendency of Nordic racial theories about the origins of white supremacy across the world presented a problem. The Mediterranean peoples had been in decline for centuries, and were perceived to be Orientalized and inferior. Yet in the past they had achieved greatness which Northern Europeans were attempting to emulate. How could a racially inferior people have created such excellence?

A simple explanation for this condition for Victorians and their Continental fellow travelers was one of racial degradation. The ancient Romans were in this telling fundamentally a different people than modern Romans, with the latter being derived from migrants from the eastern Mediterranean who had arrived during the period of the Empire.

Though most of the racially derogatory elements are gone form this narrative, it is still strongly persistent in public consciousness. Being a Cavalli-Sforza nerd (there is such a thing), I have a copy of Consanguinity, Inbreeding, and Genetic Drift in Italy, and there was data in it which made me skeptical of wholesale replacement in the middle 2000s. Then there was Peter Ralph and Graham Coop’s 2013 paper, The Geography of Recent Genetic Ancestry across Europe, which reported lots of deep regional structure across Italy.

This is important because it suggests a local stability to the demographic character of the regions for a long time. Probably earlier than the period of the Roman Empire. Though one can imagine scenarios of demographic replacement which would produce this result, they’re generally less parsimonious than the model whereby modern Italian population structure maintains the general outline it had at the beginning of the Iron age.

Finally, over the past seven years I have done a lot of analysis and manipulation of tens of thousands of Europeans and Middle Easterners in relation to their genetic data for personal and professional reasons. Some patterns jump out at you, and some subtle tendencies come into the foreground. It is pretty clear that Italians are not a transplanted Middle Eastern population (though there is some recent non-Italian ancestry; Sicilians often have minor components of clear North African ancestry as well as small percentages of Sub-Saharan heritage, which I think is almost certainly due not to Greek and Roman cosmopolitanism, but the legacy of the Arab emirate which existed on the island for a few centuries).

But now I have realized probably the best illustration of this. The Reich lab has been generating a massive genotype dataset over the past five years on the Human Origins Array. And not only do you have modern populations, but you have ancient ones (from ancient DNA). The PCA plots in their papers make what I’m saying above pretty clear.

I’ve modified the PCA plot from Genomic insights into the origin of farming in the ancient Near East. Notice where various Italian groups and Greeks are. I’ve also labeled the Druze; they are almost certainly an excellent representation of Near Eastern Syrians from 2,000 years ago. They have been endogamous for nearly 1,000 years in the Lebanese highlands, and don’t have admixture that is more common in Syrian Muslims from the lowlands.

Notice that the most of the Greeks are shifted further toward Northwestern Europeans than Southern Italians. I say most, because I’ve had access to a larger data set of Greeks, and it becomes clear that a minority of Greeks cluster more with Southern Italians, and the majority have a minority admixture element from a Northern European population. This is Slavic ancestry that arrived after the middle of the 6th century, when the East Roman state basically abandoned most of the Balkans to focus on maintaining control over Constantinople, Salonika, and the Peloponnese.

Northern Italians are shifted toward Sardinians and Spaniards. The Sardinians are important, because we now know that they are the closest modern Europeans to the agriculturalists who arrived from the eastern Mediterranean during the early Neolithic. This population, “Early European Farmers” (EEF), once dominated most of the continent. But ~5,000 years ago migrations from the steppe brought a new element which replaced and assimilated them in Northern Europe.

But in Southern Europe their genetic legacy remains strong and to a great extent dominant. Iberia and the Italian peninsula have been impacted by the migrations out of the steppe, with Sardinia the least so. In the smaller plot above you can see that the early Neolithic individuals are close to the Sardinians, with mainland Italians being shifted toward other populations.

The Northern Italians in particular show some influence from Northern European populations. Some of this may be gene flow through diffusion due to proximity, but the Alps are a rather formidable barrier. Rather, I suspect it reflects episodic migration. I generally do not weight the Lombards too highly as a major influence. Rather, I suspect that it is a combination of Gaulish settlement in the Po river valley, and early impacts from the Indo-Europeans who arrived in the Italian peninsula.

The Southern Italian shift toward the Middle East probably does indicate some gene flow, but it is important to remember that this was also Magna Graecia, so there is probably a Greek element here similar to what occurs among those Greeks without Slavic admixture (please note that Byzantine Greek rule also persisted in Southern Italy up until the Norman conquest ). And if you look at how they relate to the Neolithic samples, they exhibit a lot of shift on the plane toward the steppe populations, parallel to the Levantines. In other words, a lot of the change since the Neolithic in Southern Italy is attributable to the influence of the steppe migration, not Roman era gene flow from Syrians.

I will probably do some formal analysis at some point so that the numbers can get out there now that there are so many ancient genotypes available too. But really this shouldn’t be a discussion anymore.

Addendum: You may be asking, if there are so many literary comments about non-Italians during the Roman Empire in Italy, where did they go? I think the big thing to remember is that there is an ascertainment bias toward what we know in urban areas. There is a high likelihood that urban areas were population sinks, which could not maintain themselves without constant migration.


21 thoughts on “The Orantes has not mixed much with the Tiber

  1. While urban areas may have been population sinks even in the best of times, there is the further consideration that the population of Rome (and other Italian cities I would assume?) totally collapsed after the imperial period. I suspect most of the foreign elements did not migrate to the countryside and take up peasant farming.

  2. I find those PCA maps fascinating, but based on what I have previously read about the genetic origins of the Jews, I would have thought the mainstream Ashkenazi and Sephardi groups would have been shifted more over towards the Druze / Cypriots / Lebanese rather than being basically “right on top” of Sicilians / South Italians.

  3. Is that really a thing? Reminds me of the theories of Theo Vennemann ( ), who claims to have discerned Punic influences in the Germanic languages and thinks the Carthaginians exercised some sort of trading hegemony plus colonization over the North sea region. Sounds outlandish at first, but who knows?

    Btw regarding the Etruscans, I thought there was some genetic evidence that they did indeed come from Asia Minor, as had been claimed by the ancients…still not that clear after all?

  4. re: etruscans, the stuff in 2007 suggested there was an association. later work muddied it up. we need genome-wide data.

    spencer knows the Y inside out. so i think it’s credible if he asserts it. roman period gene flow to britain was low, but it obv. occurred.

  5. Do Levantine groups show any evidence of the Steppe admixture? Either from original IE expansions or from later Roman/Greek population movement?

  6. Do Levantine groups show any evidence of the Steppe admixture? Either from original IE expansions or from later Roman/Greek population movement?

    not really. the major ‘tell’ that it is steppe is ANE. this is really high % in the caucasus, but low as you go toward the levant.

  7. Thanks. Is the reason that Sicilians and Greeks cluster nearer Levantines then 100% due to the shared EEF ancestry, or is there any other component they also have in common, such as HG ancestry from the Mediterranean area?

  8. Razib, the Italians North in that PCA are 13 individuals from Bergamo HGDP and 8 individuals from Tuscan HGDP, all labelled as Northern Italians, but only the former from Bergamo are North Italian, those who are closer to Greeks in that PCA are the Tuscans (Central Italian).

  9. On that, bear in mind these West Eurasia wide PCA are only offering the most general view of trends that cut across the whole of West Eurasia. There can be a few more divergences between populations that aren’t covered and would be more apparent in a PCA which was more restricted.

    To kind of tl:dr it a bit more, Fst divergence between Askhenazi Jews and Sicilians is about 0.005. That’s not actually so trivial among modern European people, as that’s a little larger than between Norwegians and Spanish (0.004), Irish and Polish (0.004), Spanish and Greek (0.003).

    AJ are also quite a bit more further in Fst distance from European populations compared to Sicilians: AJ-English distance is about 0.01 while Sicilian-English about 0.006 / AJ-Polish about 0.011 compared to Siclian-Polish 0.008. (As true when comparing to ancient Europeans).

    This isn’t just an effect of more overall drift among Ashkenazis, but is probably also true net of distance to an outgroup, as distance to African (Luo) or East Asians (Han) between AJ and Sicilian only differs by Fst 0.001 (AJ less close).

    The Ashkenazis have had a unique, interesting population history (combining from Roman and Levant populations with quite large population size, then a sharp bottleneck and a pretty huge population expansion), and there’s a more differentiation there from is apparent in the PCA.

  10. Thanks. Is the reason that Sicilians and Greeks cluster nearer Levantines then 100% due to the shared EEF ancestry, or is there any other component they also have in common, such as HG ancestry from the Mediterranean area?

    i think it’s mostly EEF. but there has been some gene flow into these groups from the levant surely…though sicily as i said is unique because like spain it was ruled by muslims and some of the genetic profiles clearly have ‘moorish’ type ancestral mixes in minor components (affinities with afro-asiatic groups, but some element of sub-saharan).

    note with the greeks that i assume ‘pontic greeks’ who were resettled in the 1920s are probably going to be somewhat different, since they didn’t get impacted by the migrations of the slavs, and, may have been hellenized for anatolian substrate anyway (apparently a substantial fraction of the ‘greeks’ who left turkey were native speakers of turkish who still retained their greek orthodox religion, just as many of the ‘turks’ from greece were the inverse).

  11. What of Anatolians and ANE ancestry? There were the speakers of the Anatolian branch of IE languages, and the Greeks for millenia, do they have much ANE/Steppe ancestry?

  12. What of Anatolians and ANE ancestry? There were the speakers of the Anatolian branch of IE languages, and the Greeks for millenia, do they have much ANE/Steppe ancestry?

    doesn’t look like it. latest matheison paper has some samples. nothing. perhaps higher number would show it? not much in modern turks last i checked, but i can check again.

    remember, the hitties, the nesa, were a small elite. this may have been much like how indo-europeans spread in the mediterranean.

  13. Shocking, really, how Greeks could dominate regions of Anatolia for 2 millennia or more and there’s not ANE/Steppe ancestry. Presumably the Greeks who left the Ottoman Empire for Greece do show ANE/Steppe ancestry?

  14. first, we know how greek colonies emerged: it was male migration. so *genetically* the ones on the asia minor coast were more half-greek. second, it looks like a lot of the ANE ancestry in greece may come through slavs. i think indo-europeanization in greece was as a more elite emulation affair.

    finally, hellenization was to a great extent cultural for much of history. many of the byzantine dynasties were greek speaking and orthodox christian, but their ethnic origins were diverse (often armenian).

  15. Has there been any effort to try to differentiate skeletons from Ancient/Classical Greece by class, in order to look for genetic differences between elite and non-elite remains?

    Or has this been done anywhere in regards to Indo European expansion? You would assume here the elites were more ANE than the commoners for a time, I suppose.

  16. Those Greek samples are the same used in Lazaridis et al 2014: according to the supplementary information, 14 of them are from Thessaloniki and 6 are from Athens.

  17. [comment deleted.

    you are disrespectful. if you leave a comment like this you will be banned.

    don’t allude to ‘several genetic studies.’ i know many of the people who study these sorts of topics personally. i can tell you often i can a later deeper look because i have the tie and inclination. the moorish ancestry in w. med. ppl is easy to spot: there is west african ancestry in long(ish) segments in many spanish and sicilians.

    second, i don’t talk much about slaves because they often don’t reproduce at high level. if you are a ‘history buff’ you would know that.

    i apologize your long comment was totally removed, but you sir as a patronizing asshole.


Comments are closed.