Consequences of O&M 2018 (I): The latest West Yamna “outlier”

This is the first of a series of posts analyzing the findings of the recent Nature papers Olalde et al.(2018) and Mathieson et al.(2018) (abbreviated O&M 2018).

As expected, the first Y-DNA haplogroup of a sample from the North Pontic region (apart from an indigenous European I2 subclade) during its domination by the Yamna culture is of haplogroup R1b-L23, and it is dated ca. 2890-2696 BC. More specifically, it is of Z2103 subclade, the main lineage found to date in Yamna samples. The site in question is Dereivka, “in the southern part of the middle Dnieper, at the boundary between the forest-steppe and the steppe zones”.

NOTE: A bit of history for those lost here, which appear to be many: the classical Yamna culture – from previous late Khvalynsk, and (probably) Repin groupsspread west of the Don ca. 3300 BC creating a cultural-historical community – and also an early offshoot into Asia – , with mass migrations following some centuries later along the Danube to the Carpathian Basin, but also south into the Balkans, and north along the Prut. There is thus a very short time frame to find Yamna peoples shaping these massive migrations – the most likely speakers of Late Proto-Indo-European dialects – in Ukraine, compared to their most stable historical settlements east of the Don River.

There is no data on this individual in the supplementary material – since Eneolithic Dereivka samples come from stored dental remains – , but the radiocarbon date (if correct) is unequivocal: the Yamna cultural-historical community dominated over that region at that precise time. Why would the authors name it just “Ukraine_Eneolithic”? They surely took the assessment of archaeologists, and there is no data on it, so I agree this is the safest name to use for a serious paper. This would not be the first sample apparently too early for a certain culture (e.g. Catacomb in this case) which ends up being nevertheless classified as such. And it is also not impossible that it represents another close Ukraine Eneolithic culture, since ancestral cultural groups did not have borders…

NOTE. Why, on the other hand, was the sample from Zvejnieki – classified as of Latvia_LN – assumed to correspond to “Corded Ware” (like the recent samples from Plinkaigalis242 or Gyvakarai1), when we don’t have data on their cultures either? No conspiracy here, just taking assessments from different archaeologists in charge of these samples: those attributed to “Corded Ware” have been equally judged solely by radiocarbon date, but, combining the known archaeological signs of herding in the region arriving around this time with the old belief (similar to the “Iberia is the origin of Bell Beaker peoples” meme) that “only the Corded Ware culture signals the arrival of herding in the Baltic”. This assumption has been contested recently by Furholt, in an anthropological model that is now mainstream, upheld also by Anthony.

We already know that, out of three previous West Yamna samples, one shows Anatolian Neolithic ancestry, the so-called “Yamna outlier”. We also know that one sample from Yamna in Bulgaria also shows Anatolian Neolithic ancestry, with a distinct ‘southern’ drift, clustering closely to East Bell Beaker samples, as we can still see in Mathieson et al. (2018), see below. So, two “outliers” (relative to East Yamna samples) out of four samples… Now a new, fifth sample from Ukraine is another “outlier”, coinciding with (and possibly somehow late to be a part of) the massive migration waves into Central Europe and the Balkans predicted long ago by academics and now confirmed with Genomics.

I think there are two good explanations right now for its ancestral components and position in PCA:

Modified image from Mathieson et al. (2018), including also approximate location of groups from Mittnik et al. (2018), and group (transparent shape outlined by dots) formed by new Bell Beaker samples from Olalde et al. (2018). “Principal components analysis of ancient individuals. Points for 486 ancient individuals are projected onto principal components defined by 777 present-day west Eurasian individuals (grey points). Present-day individuals are shown.”

a) The most obvious one, that the Dnieper-Dniester territory must have been a melting pot, as I suggested, a region which historically connected steppe, forest steppe, and forest zone with the Baltic, as we have seen with early Baltic Neolithic samples (showing likely earlier admixture in the opposite direction). The Yamna population, a rapidly expanding “elite group of patrilineally-related families” (words from the famous 2015 genetic papers, not mine), whose only common genetic trait is therefore Y-DNA haplogroup R1b-L23, must have necessarily acquired other ancestral components of Eneolithic Ukraine during the migrations and settlements west of the Don River.

How many generations are needed for ancestral components and PCA clusters to change to that extent, in regions where only some patrilocal chiefs but indigenous populations remain, and the population probably admixed due to exogamy, back-migrations, and “resurge” events? Not many, obviously, as we see from the differences among the many Bell Beaker samples of R1b-L23 subclades from Olalde et al. (2018)

b) That this sample shows the first genetic sign of the precise population that contributed to the formation of the Catacomb culture. Since it is a hotly debated topic where and how this culture actually formed to gradually replace the Yamna culture in the central region of the Pontic-Caspian steppe, this sample would be a good hint of how its population came to be.

See e.g. for free articles on the Catacomb culture its article on the Encyclopedia of Indo-European Culture, Catacomb culture wagons of the Eurasian steppes, or The Warfare of the Northern Pontic Steppe – Forest-Steppe Pastoral Societies: 2750 – 2000 B.C. There are also many freely available Russian and Ukrainian papers on anthropometry (a discipline I don’t especially like) which clearly show early radiocarbon dates for different remains.

This could then be not ‘just another West Yamna outlier’, but would actually show meaningful ‘resurge’ of Neolithic Ukraine ancestry in the Catacomb culture.

It could be meaningul to derive hypotheses, in the same way that the late Central European CWC sample from Esperstedt (of R1a-M417 subclade) shows recent exogamy directly from the (now more probably eastern part of the) steppe or steppe-forest, and thus implies great mobility among distant CWC groups. Although, given the BB samples with elevated steppe ancestry and close PCA cluster from Olalde et al. (2018), it could also just mean exogamy from a near-by region, around the Carpathian Basin where Yamna migrants settled…

If this was the case, it would then potentially mean a “continuity” break in the steppe, in the region that some looked for as a Balto-Slavic homeland, and which would have been only later replaced by Srubna peoples with steppe ancestry (and probably R1a-Z93 subclades). We would then be more obviously left with only two options: a hypothetic ‘Indo-Slavonic’ North Caspian group to the east (supported by Kortlandt), or a Central-East European homeland near Únětice, as one of the offshoots from the North-West Indo-European group (supported by mainstream Indo-Europeanists).

How to know which is the case? We have to wait for more samples in the region. For the moment, the date seems too early for the known radiocarbon dating of most archaeological remains of the Catacomb Culture.

Diachronic map of Late Copper Age migrations including steppe groups ca. 2600-2250 BC

An important consequence of the addition of these “Yamna outliers” for the future of research on Indo-European migrations is that, especially if confirmed as just another West Yamna sample – with more, similar samples – , early Palaeo-Balkan peoples migrating south of the Danube and later through Anatolia may need to be judged not only in terms of ancestral components or PCA (as in the paper on Minoans and Mycenaeans), but also and more decisively using phylogeography, especially with the earliest samples potentially connected with such migrations.

NOTE. Regarding the controversy (that some R1b European autochthonous continuists want to create) over the origin of the R1b-L151 lineages, we cannot state its presence for sure in Yamna territory right now, but we already have R1b-M269 in the eastern Pontic-Caspian steppe during the Neolithic-Chalcolithic transition, then R1b-L23 and subclades (mainly R1b-Z2013, but also one xZ2103, xL51 which suggests its expansion) in the region before and during the Yamna expansion, and now we have L51 subclades with elevated steppe ancestry in early East Bell Beakers, which most likely descended from Yamna settlers in the Carpathian Basin (yet to be sampled).

Even without express confirmation of its presence in the steppe, the alternative model of a Balkan origin seems unlikely, given the almost certain continuity of expanding Yamna clans as East Bell Beaker ones, in this clearly massive and relatively quick expansion that did not leave much time for founder effects. But, of course, it is not impossible to think about a previously hidden R1b-L151 community in the Carpathian Basin yet to be discovered, adopting North-West Indo-European (by some sort of founder effect) brought there by Yamna peoples of exclusively R1b-Z2103 lineages. As it is not impossible to think about a hidden and ‘magically’ isolated community of haplogroup R1a-M417 in Yamna waiting to be discovered…Just not very likely, either option.

As to why this sample or the other Bell Beaker samples “solve” the question of R1a-Z645 subclades (typical of Corded Ware migrants) not expanding with Yamna, it’s very simple: it doesn’t. What should have settled that question – in previous papers, at least since 2015 – is the absence of this subclade in elite chiefs of clans expanded from Khvalynsk, Yamna, or their only known offshoots Afanasevo and Bell Beaker. Now we only have still more proof, and no single ‘outlier’ in that respect.

No haplogroup R1a among hundreds of samples from a regionally extensive sampling of the only cultures mainstream archaeologists had thoroughly described as potentially representing Indo-European-speakers should mean, for any reasonable person (i.e. without a personal or professional involvement in an alternative hypothesis), that Corded Ware migrants (as expected) did not stem from Yamna, and thus did not spread Late Indo-European dialects.

This haplogroup’s hegemonic presence in North-Eastern Europe – and the lack of N1c lineages until after the Bronze Age – coinciding with dates when Uralicists have guesstimated Uralic dialectal expansion accross this wide region makes the question of the language spread with CWC still clearer. The only surprise would have been to find a hidden and isolated community of R1a-Z645 lineages clearly associated with the Yamna culture.

NOTE. A funny (however predictable) consequence for R1a autochthonous continuists of Northern or Eastern European ancestry: forum commentators are judging if this sample was of the Yamna culture or spoke Indo-European based on steppe component and PCA cluster of the few eastern Yamna samples which define now (you know, with the infallible ‘Yamnaya ancestral component’) the “steppe people” who spoke the “steppe language”™ – including, of course, North-Eastern European Late Neolithic

Not that radiocarbon dates or the actual origin of this sample cannot be wrong, mind you, it just strikes me how twisted such biased reasonings may be, depending on the specific sample at hand… Denial, anger, and bargaining, including shameless circular reasoning – we know the drill: we have seen it a hundred times already, with all kinds of supremacists autochthonous continuists who still today manage to place an oudated mythical symbolism on expanding Proto-Indo-Europeans, or on regional ethnolinguistic continuity…

More detailed posts on the new samples from O&M 2018 and their consequences for the Indo-European demic diffusion to come, indeed…

See also:

13 thoughts on “Consequences of O&M 2018 (I): The latest West Yamna “outlier”

  1. About supremacists, I have never ever encountered anyone who denied a Steppe homeland, and the people living in the area today (Russians and Ukranians) all tend to deny such attribution to them.

    Now, how to explain Central amd South Asia, Andronovo, Srubnaya, Shintashta, the Indo-European languages and the great frequency of R1a? The Anatolian component, along with R1a, came to these places with Corded Ware – why Sanskrit isn’t an Uralic language then?
    The same way people were jumping to conclusions about Yamnaya>CW>Beakers, you also are with CW=Uralic. There’s no evidence for that, and if anything, what you yourself said above:
    Yamnaya = Northwest IE (plus Tocharian)
    CW = Indo-Slavonic
    Makes so much more sense.
    Absense of evidence is not Evidence of absense – it’s a fact that the CW spread IE, now it’s just a matter of how either:
    1. A R1a community got overwhelming Steppe ancestry if they weren’t of the R1b patrilineage.
    2. A Steppe community (probably from the Baltics) got their R1b replaced by R1a without changing the language.

    My best bet is in Scenario 2: in the Baltics, a Tribe Chief lineage must have changed by accident from a R1b to R1a due to, I don’t know, the then Chief not having sons and his Sister having one with a R1a man.

    1. My experience with Russians and Ukrainians is with academics, and their trend is to date languages further back in time than Western European schools of thought – due, in my opinion, to their Nostraticist tradition. So at best they bet for a Mesolithic origin, and that obviously makes a common admixed R1a-R1b (and WHG:EHG) community in the Baltic quite ‘comfortable’ to discuss. That this theory is obviously easier to accept by anyone regardless of their haplogroup does not mean it is correct, though, and is not without its ‘continuist’ trends either.

      And about common Russians or Ukrainians, I cannot say, nor about any other people. My experience is that normal people anywhere in the world don’t care about Y-DNA haplogroups at all, whatever you tell them, until they know theirs, and begin to fantasize about the origins of their paternal line: then things change, and some kind of genealogical quest for ethnicity, or purity, or continuity, or supremacy begins.

      About the possibilities you propose, there are a million explanations as to how (certain) Slavic populations became mostly of R1a lineages. But only two reasonable models remain, judging by the data we have.

      The main question now, as always, is the linguistic classification, and that determines if Balto-Slavic comes from the east, or from the west. Not that the linguistic community is really divided, since ‘Indo-Slavonic’ is a term only used by two or three known linguists, but Kortlandt is one of them, and he is a renown expert Slavicist and Balticist, so we need genomic data to finally reject his Temematic substrate hypothesis (adopted from Holzer), which he uses to explain the obvious North-West Indo-European influence in Balto-Slavic.

      Data on the Tollense valley is already showing, in my opinion, hints to the definitive answer: . But in any case we will need a lot of data to connect this early Indo-European-speaking Yamna population with the (quite late) Proto-Slavic and Proto-Baltic homelands, and there can always be surprises.

  2. So you basically consider the R1a that also seems very associated with proto/early Slavs as a Uralic marker that pre-dates IE, so Slavic too, in northeast Europe?

    Like say Mycenaeans with the pre-Greek-associated J2a spreading an Indo-European language to Cyprus at the time of palatial destruction or Scandinavians with the pre-Germanic-associated I1 spreading East Germanic to Poland.

    I can consider it as a general idea even if I don’t necessarily agree with it but the question is: how can you show you’re right about this specifically with genetics, especially when most people on the internet geneto-anthropological sphere are likely to interpret Corded Ware as strictly IE in the first place and without strong breaks in Y-DNA? Similarly, what’s the cultural succession you have in mind that takes Balto-Slavic from the West to the East? The Slavic expansion from East to West later on is easier to argue for.

    As a funny aside tangentially related to this sort of topic, here’s another scenario that has been often proposed that I don’t find to be necessarily the most likely but I always found interesting and should be controversial enough: proto-IE (or post-“Indo-Hittite”) came to the steppe with the southern CHG-EEF mixed population via the Caucasus which also brought the productive economy there. The relevant R1b clades, as opposed to various steppic dead-ends, might have come from there too but that one might not be necessary either way. A potential parallel in Britain? A mostly EEF population with only some WHG admixture but dominated by WHG-associated Y-DNA and, for all one knows, speaking EEF languages.

    As for the outliers, Outlier 2 seems like a UkraineN-BalkanN mix, maybe one of those few representatives we see of earlier cultures that still remain around like some of those HG-heavy individuals you see in Neolithic Romania. Outlier1 looks like a ‘regular’ Yamnaya individual with more of the ArmeniaBA-IranCA ancestry that represents the southern component of the ‘steppe’ ancestry.

    1. As you say, it is strong breaks in Y-DNA what we look for.

      Since the 2015 papers, people have not had any difficulties accepting that Corded Ware represented all Indo-European branches, later diffused to
      West Europe.

      Then they gradually came to accept – because of the obvious R1b-R1a divide – that maybe CW represented only “Indo-Slavonic”, or “Germanoc-Indo-Slavonic”.
      What I have argued is that the only branches not (exactly) represented by Yamna and its expansions are the phonetically more divergent ones, i.e. Balto-Slavic and Indo-Iranian. Hence there were not multiple founder or diffusion events since some general, long-lasting steppe migrations as proposed by Kristiansen; just two main ones (in Balto-Slavic and Indo-Iranian territory), and quite old, just after the Yamna expansions. Hence e.g. their old shared ‘Satem’ innovation.

      As to how language succession within genomic change happens, the process is being found everywhere, so it is not a rarity anymore ( Also, now that CWC does not represent a Yamna offshoot, you cannot find anyone anywhere that doesn’t try to fit some sort of diffusion or founder effect event into their narratives, so continuity is (at last) out of the question…

      About Balto-Slavic from west to east, there are already multiple theories of continuity from one culture to another until they reached the steppes, from where they obviously migrated east to west in proto-historic times (see e.g. for Slavic ).

      My main argument is a linguistic one, obviously: That Balto-Slavic is, in essence, a North-West Indo-European dialect. I looked for archaeological alternatives, and the model I proposed was the most reasonable one to me ( ). It is obvious, however, that tracing Slavs and Balts to some culture B.C. is as difficult as proposing a homeland for Late PIE, so I just proposed a possibility as to what R1a subgroup might represent both migrations, and this is pure speculation right now.

      The main question for me, now that we have some confirmation of Yamna in Polish and Czech lands, and then a resurge of R1a (and we already had that for the North Caspian steppe) is not if this event happened, but if that resurge actually means A) that Balto-Slavic came from that region – my preferred option – , or B) if it actually represents Holzer’s Temematic, and Balto-Slavic is thus part of an older Indo-Slavonic group, and is thus related probably to steppe cultures and possibly R1a-Z93 (and admixed and suffered a founder effect in the steppe), which I believe is not the case.

      As I said in another comment, that option should be easily dismissable in Linguistics, but Kortlandt is a renown Balticist and Slavicist and supports this, so we need genomic data to reject his proposal, at least from the point of view of population movements.

      Even if we follow Slavic peoples to the steppe and back, it will be difficult to convince certain sectors that nothing happened there, so – like the “Latin is related to Anatolian” meme, we will probably be stuck with the “Slavic is related to (Indo-)Iranian” for a looong time. But at least we will also have genomic proof against old or continued contacts.

      As for theories of ancient origins of steppe peoples (to the formation of Early PIE), we certainly need more samples from around the steppe, including south, but also north and east. And even then, I don’t think we might have the needed anthropological data to make assumptions. Linguists are barely scratching the surface of Indo-Uralic – and not much more is probably possible right now – , and archaeological cultures can only give you that much information…

    2. @RandomName

      I’ve entertained similar scenarios for early Indo-European. It’s interesting what you say about R1b – there seems to be a weird gap in Europe when it comes to M269 despite quite extensive sampling. IIRC there’s an iron age R1b-Z2103 from Azerbaijan that plots rather south of modern Iranians found in a Hurro-Uratrian archaeological context. It’s quite difficult to say more than that with so few samples from that region though.

  3. It is difficult to discern any kind of consensus in a field so small, but I think that modern linguists tend to be in favor of an eastern origin of Uralic. A leak of an upcoming population genetics already foreshadowed that North-Eastern Europe received an influx of eastern genes at the time when Ante Aiko and the likes predicted Uralic languages to have arrived in Fenno-Scandia and surroundings – namely in the iron age. A western origin would also upend the quite straightforward tree of Uralic, with the eastern languages usually serparating earlier and being associated with more easterly autosomal genetics and a higher percentages of Y-DNA N1c.

    IMHO, if you want to propose an non-IE origin for CWC, something like Aiko’s Paleo-European languages, which according to him survive as substrates in Sami, would make more sense. But there’s so little that can be gathered with any kind of certainty from that region that we’ll probably never really know.

Leave a Reply

Your email address will not be published.