Tuesday, November 12, 2019

Ancient Roman DNA

At the forest level, the only slightly surprising quirk in a new (paywalled) survey of ancient Roman DNA in the journal Science released last Friday is that its second demographic transition is unclear and on the late side, but still in the Bronze Age. There are no samples from 1700 BCE to 700 BCE, so there simply isn't the data to resolve this transition in detail with direct evidence (although deeper analysis of, for example, linkage disequilibrium in early Iron Age steppe ancestry components might help resolve this question in later studies even without new data). This is something that may be due as much as anything to Rome being a backwater until the early Iron Age when it was really founded as the city that it is today. 

The Near Eastern component largely drops out of the the Roman gene pool after the Imperial Roman period, although this could be simply due to the peculiarities of which samples with particular biases data to which eras.

The new paper and its abstract are as follows (emphasis added):
Ancient Rome was the capital of an empire of ~70 million inhabitants, but little is known about the genetics of ancient Romans. Here we present 127 genomes from 29 archaeological sites in and around Rome, spanning the past 12,000 years. We observe two major prehistoric ancestry transitions: one with the introduction of farming and another prior to the Iron Age. By the founding of Rome, the genetic composition of the region approximated that of modern Mediterranean populations. During the Imperial period, Rome’s population received net immigration from the Near East, followed by an increase in genetic contributions from Europe. These ancestry shifts mirrored the geopolitical affiliations of Rome and were accompanied by marked inter-individual diversity, reflecting gene flow from across the Mediterranean, Europe, and North Africa.
Margaret L. Antonio, et al. "Ancient Rome: A genetic crossroads of Europe and the Mediterranean" 366 (6466) Science 708 (November 8, 2019). DOI: 10.1126/science.aay6826

Bernard's Blog has a post that gets past some of the paywall limitations.

Razib notes that Rome was something of a population sink which absorbed cosmopolitan influences but didn't necessarily spread them. He has more analysis in an earlier post that observes that "Modern Romans descend from Italian peasants, who were less impacted by the predations of the Goths and Byzantines, and had higher fertility than urban dwellers even in peaceful times" and also notes that Christianity may have helpfully reduced inbreeding and clan based social organization in Europe.

Eurogenes opens up two threads for discussions of the fine details. 

* His first post asks "What's the difference between ancient Romans and present-day Italians?" A question the new paper answers with "not much." 

* His second post is entitled: "Open analysis and discussion thread: Etruscans, Latins, Romans and others", the biggest upshot of which is that Etruscan ancient DNA differs surprisingly subtly from that of nearly contemporaneous Indo-European language speakers. This is arguably the most notably inference that can be drawn from the data including this new paper. Some notable comments from that thread (my emphasis in bold):

Romulus said...
It's really weird that the Iron Age Italians prefer Yamnaya over all these Beaker or CWC groups as a 2 way fit with Copper Age Italians.

It seems to imply that the Steppe group which brought M269 to Italy branched off before the group that led to the Lech Valley Beakers. Giving Italians a less diluted form of Steppe.

There were Copper Age Italian like groups (Hungarian Beaker) mixing with Yamnaya in the Carpathian Basin.

But what it probably is a reflection of is a migration of this Bronze Age Croatian like group represented by R437 bringing additional Steppe from the Balkans after the initial Beaker groups. The J2b Croatian had a lot of Steppe and R437 shows that on the PCA.
ǵenh said...
@ Davidski

That's right, David.

The Proto-Villanovan culture is a Late Bronze culture present throughout Italy from the Alps to eastern Sicily, which is considered Proto-Italic but not exclusively Italic. In the sense that almost all subsequent cultures of the Iron Age of Italy derive from Proto-Villanovan culture, both Iron Age Italic (Latin, Osco-Umbrians) and non-Italic cultures (Veneti, Etruscans). Especially in the past, even the Culture of Golasecca (which is typical of north-west Italy from which derive the Celts who speak a lepontic language) was believed to be derived from the Proto-Villanovan culture.

While the Villanovan culture is an Iron Age culture that is the first phase of the Etruscans, the beginning of the Etruscan civilization, according to what is now the most accepted Etruscan chronology by scholars.

The misunderstanding between the two names was born because the Villanovan culture associated to the Etruscans was the first to be discovered by archaeologists around the middle of the 1800s. When in the 1930s archeologists also discovered settlements of Protovillanovan culture, at first they thought that it was only an earlier phase of Villanovan culture. Only later they understood that the Protovillanovan culture was instead the previous phase of many other cultures present in Italy, including the Italic ones, but the name was never changed, and this has contributed to creating confusion.
Samuel Andrews said...
"It seems to imply that the Steppe group which brought M269 to Italy branched off before the group that led to the Lech Valley Beakers. Giving Italians a less diluted form of Steppe."

What I'm seeing, is Latins perfer Bell Beaker from Germany & Czech.

"There were Copper Age Italian like groups (Hungarian Beaker) mixing with Yamnaya in the Carpathian Basin."

R1b P312+ isn't from Yamnaya. It's from Corded Ware. A R1b L51+ has been found in Corded Ware. It's unlikely there were groups with mostly Yamnaya/Kurgan ancestry in Central Europe in the Bronze age. The groups who went to Italy were already probably under 50% Yamnaya.
Gaska said...

1-They are practically identical to Latins (24-28% Yamnaya)
2-Descend from BBs (45-50%)
3-We have cases of Bbs in Parma with Iberian signal in its Autosomal DNA
4-Italian Chalcolithic has a strong Iberian signal-
5-Etruscans are a mixture of local Eneolithic + Balkans + BBS-
6-Very similar to the Iron Age Iberians and Northern Italians
7-They have more Balkan mix than Latins
8-Obviously they were NOT Africans or Anatolians, nor Levantines-
9-Its mitochondrial markers are typically western (WHG-EEF and Iberian)
10-Heirs of the Villanovan culture that comes from UrnField Culture
11-We only have a male marker and I think it comes from Illyria
12-They spoke a non-Indo-European language-

Also, keep in mind that Sam is an expert in declaring the end of the story without any reason to do so.

It is essential to know more information about the Bronze Age in Italy (we do not have a single sample yet) and more data on the uniparental markers of the Etruscans, because they are also direct descendants of the BBs and therefore the doubts about the language spoken by the BB culture still exist
Archi said...
[Quoting Gaska above] . . .

1. R475 is almost identical, yeah. When there are only three of them. R850 is also a sample.
2. thus it is possible to name any other percent and anybody else.
5. You can name any other combination. If you forget about R475.
6. And we can say that Very unsimilar to the Iron Age Iberians and Northern Italians
7. Can we say that they have not Balkan mix than Latins
8. We will forgot further, the main thing is not to turn to R475 (and R850)
9. They have typical mito for Europe and Anatolia
10. Villanova are not their culture.
11 This male marker is much closer to Nuragic Sardinia and without any steppe components, it was not there before the nuragas.
12. Nobody knows the language of which group they spoke, but definitely not Basque. He had the strongest connections with Lydia.
Andrzejewski said...
Can we claim then that Etruscans are basically almost identical to Italics, save for an extra layer of admixture stemming from migrants from the Balkan who were rich in EEF and must’ve spoken a farmer language? Is it from areas close to Lemnian? If so, would Etruscan be considered a Lengyel sourced language?

I would interject that Etruscan was indeed not Indo-European and also not Basque, although its associations with other attested non-Indo-European language of Europe like Lemnian is cryptic because non-Indo-European languages of the Mediterranean other than Etruscan, Basque and the Afro-Asiatic languages of the region (e.g. Phoenician, Maltese and Hebrew) are very poorly attested. Etruscan was still in living language in classical Rome at a stage when history and linguistics had finally started to cross over from myth and legend to serious scholarly pursuits with recognized experts. Classical Roman scholars were wrong about many things, and their work in this area can't easily be tested because Etruscan hasn't been fully deciphered, but they wouldn't have missed anything too obvious to someone who had access to the information available in the capital of the Roman Empire.

FrankN said...
Do I get this correctly? Some models, such as the original ones from the paper, and also those run by Matt, speak in favour of an Anatolian_BA (CHG-enhanced) element in the genesis of Etruscans and IA_Italics, while others seem to favour a BB descent instead. Could someone try to clarify this, possibly by treating each sample separately? E.g., "Italic" R1021, 7th cBC, is according to the SuppMats coming from a site half-way between Rome and Naples characterised by "Italic archaeological findings and Pelasgic walls, characteristic of the people of the Bronze Age Aegean" - IOW: We may be dealing with an outpost of Magna Graecia rather than a typical "Italic" assemblage.

Linguists have long been puzzled by Anatolian - ItaloCeltic isoglosses such as the passive marked by infixed -tu-. A parsimonious explanation could be that Italo-Celtic represents an originally Anatolian language (probably Luvic rather than Hittite-like) which entered during the MBA and was later overformed by some kind of Illyro-Germanic that came in with the Urnfield expansion.
As to Etruscan-HurroUrartean connections: A genetic relation isn't yet universally accepted, but seems to enjoy increasing support. From https://en.wikipedia.org/wiki/Alarodian_languages:

"The term "Alarodian languages" was revived by I.M. Diakonoff for the proposed language family that unites the Hurro-Urartian and Northeast Caucasian languages.(..)

The inclusion of Etruscan and the related Tyrsenian languages has also been proposed, first by Orel and Starostin in 1990, on the basis of sound correspondences.[13] Facchetti has argued that there is a "curious" set of isoglosses between Etruscan and Hurrian[14], while Pliev proposed instead that Etruscan had a Nakh substrate.[15] In 2006, Robertson developed the hypothesis for including Tyrsenian further by presenting reconstructions of common ancestral forms of the numerals, and proposed cases of apparent sound correspondences between Etruscan and Nakh, with discussion also of Hurro-Urartian, Lemnian and the various Dagestanian branches.[16]

Apparently, Gamkrelidze/Ivanov 1990 have stressed particular closeness between NE Caucasian and Etruscan. A Fournet, while sceptical about a HU-NEC genetic relation, speaks out in favour of Tyrrhenian-HU relatedness (see my comments to the previous post). Kozyrski e.a. 2015 (http://www.aiscience.org/journal/j3l) came up with a number of fresh Etruscan-NEC (&HU) isoglosses.

Last but not least, V.V. Ivanow "Comparative Notes on Hurro-Urartian, Northern Caucasian and Indo-European" glosses over various HU/NEC - Etruscan connections, e.g. as concerns the plural on -(a)r shared by HU, NEC and Etruscan, Etruscan eis-er/ais-e/ar vs. Hurr. e-en-za-a-ri "gods", HU *pur(r)a "slave, servant" also present in Etruscan (plus Latin puer "boy", w/o satisfying IE etymology), and the etymological relation between the Etruscan toponym Mantua and the Urartian Mantupa.


Last but not least, a couple of Austrian archeologists have related the Taurus (S. Anatolia) to the Tauern (E. Alps) massives (c.f. the Taurisci in IA N. Italy). Note in this context Etruscan tul, Chechen t'o "stone" [albeit the t<->r sound shift would require explaination].
FrankN said...
Correction to my previous post: I mistook the Taurisci, who actually settled around the Tauern massive in Austria, for the NE Italian Taurini, who lent there name to the city of Turin.
FrankN said...
Ric: "So what you are basically saying is that Italic and Celtic originated in Anatolia to try and make the Hurro-Urartian thing work...?"

Not quite. What I am saying that linguists have for long known about a specific closeness between Anatolian and Italo-Celtic, w.o. so far being able to explain the reason for this closeness. An EBA migration out of Anatolia into Italy would provide a sensible explanation.

Otherwise, ItaloCeltic also has a lot in common with Germanic (albeit some phylogenies, e.g. the one proposed by D. Ringe, group Germanic with Albanian, others have Germanic clustering with Balto-Slavic). In this sense, the terminology "originated in [Anatolia]" is probably misplaced. We should rather think in terms of admixture and overforming, as also happened with English (Latin overforming Insular Celtic, to be overformed by first W., than N. Germanic, and ultimately heavily absorbing Franco-Norman, which in itself was a Gaulish-Latin-Germanic hybrid).

IA Italy by ca. 500-400 BC, displayed a huge linguistic diversity, including
- several fairly differentiated Italic languages falling into three different families (Osco-Umbrian, Latin-Faliscan, Venetic [phonetically Italic, grammatically probably a distinct language/family in its own right]);

- Other IA, especially Celtic (Lepontic, possibly further Celtic language), Ancient Greek in S.Italy, plus possibly also Illyrian languages along the Adriatic coast;
- Non-IE languages such as Etruscan, Rhaetic, Semitic (Punic), possibly Ligurian, Paleosardinian, and maybe a couple more.

The Punic and Old Greek cases seem relatively clear, but otherwise this linguistic diversity is so far poorly understood. Even if we just focus on non-Greek IE languages: Their diversity is unlikely to have evolved from just a single introgression, e.g. Urnfield-related ca. 1.200 BC. One possibility would be that IE was already spoken in Italy since a long time, maybe EEF, or BB. Both scenarios are IMO unlikely: There is substantial linguistic evidence against EEF speaking IE, and the BB impact on Italy (aside from the questions on BB language thrown up in Iberia) was extremely limited, both archeologically and genetically.

As such, I think the explaination for Italy's linguistic diversity during the middle IA lies in a series of IE introgressions from the MBA onwards, which brought several already fairly differentiated IE languages there (interactction with non-IE substrate, and Etruscan/ Semitic superstrate, provided for further differentiation). One obvious source is Urnfied via the E. Alps, a second one would be the (Pre-Proto-)Illyrian-speaking Balkans, but in addition to those two, Anatolia also looks linguistically promising.
Samuel Andrews said...
I looked at the Imperial Roman samples. 48 samples are in G25 PCA These are my opinons......

41 of 48 are over 50% Middle Eastern. On average, they are 66% Middle Eastern.

Where in the Middle East did they come from????
Asia Minor=22
Samuel Andrews said...
These was real Middle Eastern admixture in Roman-era Italy not the kind the ancient Greeks had. Ancient Greeks had Western Anatolia, EEF-rich, IranN-low, Levant-low kind of Mid East admixture.

This new Mid East admix in Roman era Italy came from the "interior" Middle East not the Aegean/Western Anatolia.

Meaning, they weren't all majority AnatoliaBA-like. They had complex ancestry which had significant doses from all over the ancient Middle East: Anatolia, Levant, Caucasus, Iran.

Overall, Cyrpiots are the best modern references. Because Cypriots are mostly ANatolia-BA but also have significant recent Levant-BA and recent Mesoptamia like ancestry.
FrankN said...
I would for the time being suggest some caution as concerns the "Etruscan" samples. The locations in question may rather have been assigned to Etruscans on political than on linguistic grounds. After all, Rome was also an "Etruscan" city until they kicked out "Etruscan" Tarquinius Superbus in 509 BC. Let's wait for aDNA from Tuscany, where there (hopefully) is some more certainty that not only the political elite, but also commoners actually spoke Etrurian.

Otherwise, that HRV-IA I1331 sample has been found not too far away from the area of the IA Liburnians.
"Following studies of the onomastics of the Roman province of Dalmatia, Géza Alföldy has suggested that the Liburni and Histri belonged to the Venetic language area.[4][5] In particular, some Liburnian anthroponyms show strong Venetic affinities, a few similar names and common roots, such as Vols-, Volt-, and Host- (< PIE *ghos-ti-, "stranger, guest, host"). Liburnian and Venetic names sometimes also share suffixes in common, such as -icus and -ocus.
Other toponymical and onomastic similarities have been found between Liburnia and other regions of both Illyria and Asia Minor, especially Lycia, Lydia, Caria, Pisidia, Isauria, Pamphylia, Lycaonia and Cilicia, as well as similarities in elements of social organization, such as matriarchy/ginecocracy (gynaikokratia) and the numerical organization of territory. These are also features of the wider Adriatic region, especially Etruria, Messapia and southern Italy.[10] (..)
The old toponym Liburnum in Liguria may also link the Liburnian name to the Etruscans,[12][13] as well as the proposed Tyrsenian language family.

Right- that's the linguists' story, that to some extent matches, and to some extent not, what aDNA now starts to reveal. To grab the full story, we'll need lots of MLBA aDNA.

Rob: "Etruria was able to rise amidst all this because of it contacts with the Nuraghi, which were one of the few powers which were able to maintain their trans-Mediterranean contacts amidst the said collapses, via Cyprus. "
I don't think that is the full story. Etruria's rise also has to do with control of Italy's richest iron ore deposits, especially those on Elba (and in addition also sophisticated iron smithing and Welding craftsmanship). See for details

1 comment:

Ryan said...

Do Andrzejewski and Gaska's points about a Balkan origin of the Etruscans, I think this pretty much rules that out actually. The Etruscans are pulled in the PCA away from the Balkans, not towards it. If anything the Etruscans seem similar to Sardinians.