The PIE homeland controversy: August 2019 status report

Archeologist David Anthony has a new paper on the Indo-European homeland debate titled Archaeology, Genetics, and Language in the Steppes: A Comment on Bomhard. It's part of a series of articles dealing with Allan R. Bomhard's "Caucasian substrate hypothesis" in the latest edition of The Journal of Indo-European Studies. It's also available, without any restrictions, here.

Any thoughts? Feel free to share them in the comments below. Admittedly, I found this part somewhat puzzling (emphasis is mine):

It was the faint trace of WHG, perhaps 3% of whole Yamnaya genomes, that identified this admixture as coming from Europe, not the Caucasus, according to Wang et al. (2018). Colleagues in David Reich’s lab commented that this small fraction of WHG ancestry could have come from many different geographic places and populations.

I think that's highly optimistic. It really should be obvious by now thanks to archeological and ancient genomic data, including both uniparental and genome-wide variants, that the Yamnaya people were practically entirely derived from Eneolithic populations native to the Pontic-Caspian (PC) steppe. So, in all likelihood, this was also the source of their minor WHG ancestry.

Indeed, they clearly weren't some mishmash of geographically, culturally and genetically disparate groups that had just arrived in Eastern Europe, but the direct descendants of closely related and already significantly Yamnaya-like peoples associated with long-standing PC steppe archeological cultures such as Khvalynsk and Sredny Stog. I discussed this earlier this year, soon after the Wang et al. paper was published:

On Maykop ancestry in Yamnaya

I hope I'm wrong, but I get the feeling that the scientists at the Reich Lab are finding this difficult to accept, because it doesn't gel with their theory that archaic Proto-Indo-European (PIE) wasn't spoken on the PC steppe, but rather south of the Caucasus, and that late or rather nuclear PIE was introduced into the PC steppe by migrants from the Maykop culture who were somehow involved in the formation of the Yamnaya horizon.

Inexplicably, after citing Wang et al. on multiple occasions and arguing against any significant gene flow between Maykop and Yamnaya groups, Anthony fails to mention Steppe Maykop. But the Steppe Maykop people are an awesome argument against the idea that there was anything more than occasional mating between the Maykop and Yamnaya populations, because they were wedged between them, and yet clearly distinct from both, with a surprisingly high ratio of West Siberian forager-related ancestry (see here and here).

Despite all the talk lately about the potential cultural, linguistic and genetic ties between Maykop and Yamnaya, including claims that the latter possibly acquired its wagons from the former, my view is that the Steppe Maykop and Yamnaya wagon drivers may have competed with each other and eventually clashed in a big way. Indeed, take a look at what happens after Yamnaya burials rather suddenly replace those of Steppe Maykop just north of the Caucasus around 3,000 BCE.

Yamnaya_RUS_Caucasus
RUS_Progress_En_PG2001 0.808±0.058
RUS_Steppe_Maykop 0.000
UKR_Sredny_Stog_II_En_I6561 0.192±0.058
chisq 13.859
tail prob 0.383882
Full output

Yep, total population replacement with no significant gene flow between the two groups. Apparently, as far as I can tell, there's not even a hint that a few Steppe Maykop stragglers were incorporated into the ranks of the newcomers. Where did they go? Hard to say for now. Maybe they ran for the hills nearby?

Intriguingly, Anthony reveals a few details about new samples from three different Eneolithic steppe burial sites associated with the Khvalynsk culture:

The Reich lab now has whole-genome aDNA data from more than 30 individuals from three Eneolithic cemeteries in the Volga steppes between the cities of Saratov and Samara (Khlopkov Bugor, Khvalynsk, and Ekaterinovka), all dated around the middle of the fifth millennium BC.

...

Most of the males belonged to Y-chromosome haplogroup R1b1a, like almost all Yamnaya males, but Khvalynsk also had some minority Y-chromosome haplogroups (R1a, Q1a, J, I2a2) that do not appear or appear only rarely (I2a2) in Yamnaya graves.

As far as I can tell, he suggests that they'll be published in the forthcoming Narasimhan et al. paper. If so, it sounds like the paper will have many more ancient samples than its early preprint that was posted at bioRxiv last year.

For me the really fascinating thing in regards to these new samples is how scarce Y-haplogroup R1a appears to have been everywhere before the expansion by the putative Indo-European-speaking steppe ancestors of the Corded Ware culture (CWC) people. It's basically always outnumbered by other haplogroups wherever it's found prior to about 3,000 BCE, even on the PC steppe. But then, suddenly, its R1a-M417 subclade goes BOOM! And that's why I call it...

The beast among Y-haplogroups

At this stage, I'm not sure how to interpret the presence of Y-haplogroup J in the Khvalynsk population. It may or may not be important to the PIE homeland debate. Keep in mind that J is present in two foragers from Karelia and Popovo, northern Russia, dated to the Mesolithic period and with no obvious foreign ancestry. So it need not have arrived north of the Caspian as late as the Eneolithic with migrants rich in southern ancestry from the Caucasus or what is now Iran. In other words, for the time being, the steppe PIE homeland theory appears safe.

Update 20/12/2019: A note on Steppe Maykop

See also...

Is Yamnaya overrated?

The PIE homeland controversy: January 2019 status report

Late PIE ground zero now obvious; location of PIE homeland still uncertain, but...

The PIE homeland controversy: January 2019 status report

Last year, the preprint that claimed to have presented archaeogenetic data that opened up the possibility of the Proto-Indo-European (PIE) homeland being located south of the Caucasus was, ironically, also the preprint that considerably strengthened my confidence that the said homeland was actually located north of the Caucasus.

Of course, I'm talking about the Wang et al. manuscript at bioRxiv, which is apparently soon to be published as a peer-reviewed paper in Nature Communications (see here).

It'll be fascinating to observe if and how the peer-review process has impacted on the preprint, and especially its conclusion. My impression was that the authors seemed pretty sure that the Maykop people gave rise to the Yamnaya culture, or at least Indo-Europeanized it. But, as far as I saw, the archaeogenetic data didn't bear this out at all, and instead showed a lack of any direct, recent and meaningful genetic relationship between Maykop and Yamnaya (see here). Was this also picked up by the peer reviewers? We shall see.

Moreover, there was some exceedingly interesting fine print in the manuscript's supplementary information:

Complementary to the southern [Darkveti-Meshoko] Eneolithic component, a northern component started to expand between 4300 and 4100 calBCE manifested in low burial mounds with inhumations densely packed in bright red ochre. Burial sites of this type, like the investigated sites of Progress and Vonyuchka, are found in the Don-Caspian steppe [10], but they are related to a much larger supra-regional network linking elites of the steppe zone between the Balkans and the Caspian Sea [16]. These groups introduced the so-called kurgan, a specific type of burial monument, which soon spread across the entire steppe zone.

Always read the fine print, they say. And they're right. Imagine if I only read the preprint's conclusion and missed this little gem; I'd probably think that the PIE homeland was located south of the Caucasus rather than on the Don-Caspian steppe.

Wow, proto-kurgans with inhumations densely packed in bright red ochre? A supra-regional network linking the elites of the steppe all way from the Balkans to the Caspian Sea? An expansionist culture? And, as evidenced by the ancient DNA from the Progress and Vonyuchka sites, a people who may well have been in large part ancestral to the Yamnaya, Corded Ware and Andronovo populations, that have been identified based on archeological and historical linguistics data as the main vectors for the spread of Indo-European languages as far as Iberia in the west and the Indian subcontinent in the east.

I wonder if the authors actually asked themselves who these people may have been, before so haphazardly turning to Maykop and, ultimately, the Near East, as the likely sources of the Yamnaya culture? To me they look like the Proto-Indo-Europeans and true antecedents of Yamnaya.

So as things stand, my pick for the PIE homeland is firmly the Don-Caspian steppe. And I genuinely thank Wang et al., and indeed the Max-Planck-Institut für Menschheitsgeschichte (aka MPI-SHH), for their assistance.

But, you might ask, what about the Hittites? Yes, I realize that no one apart from me and a few of my readers here can find any steppe ancestry in the so called Hittite genomes published to date. However, consider this: if the PIE homeland really was on the steppe, and a dense sampling strategy of Hittite era Anatolia fails to turn up any unambiguous steppe ancestry in at least a few individuals, then there has to be an explanation for it. But let's wait and see what a dense sampling strategy of Hittite era Anatolia actually reveals before we go that far.

See also...

The PIE homeland controversy: August 2019 status report

Yamnaya: home-grown

Late PIE ground zero now obvious; location of PIE homeland still uncertain, but...

Big deal of 2018: Yamnaya not related to Maykop

I was going to write this post after the genotype data from the Wang et al. preprint on the genetic prehistory of the Greater Caucasus became available, because I wanted to demonstrate a few key points with analyses of my own. But I've got a hunch that the formal publication of the manuscript, and thus also the release of the data, has been indefinitely delayed for one reason or another. So here goes anyway, the big deal of 2018...

This year, ancient DNA has revealed that the populations associated with the Maykop and Yamnaya archeological cultures were genetically distinct from each other, and, in all likelihood, didn't mix to any significant degree. Case in point: an ADMIXTURE analysis from Wang et al. 2018.

No doubt, this is quite a shock for many people, especially those of you who consider Maykop to have been a Proto-Indo-European-speaking culture that either gave rise to Yamnaya or at least Indo-Europeanized it. So now, if you still want to see Maykop as the Indo-Europeanizing agent in the Pontic-Caspian steppe, you'll have to rely solely on archeological and linguistics data, and also keep in mind that ancient DNA has slapped you in the face.

In just a few years, ancient DNA has provided us with plenty of shocks, but this is arguably among the biggest.

However, I honestly can't say that it was a huge surprise for me, because I tentatively predicted this outcome more than two years ago based on a handful of mitochondrial (mtDNA) haplotypes (see here). Certainly, analyzing genome-wide genetic data is what I thrive on, but if that's off limits, then eyeballing even a few mtDNA markers can also be very useful.

Wang et al. easily demonstrate the lack of any meaningful genetic relationship between Maykop (including Steppe Maykop, which shows an unusual eastern influence) and Yamnaya using a range of methods. But, judging by their conclusion, in which they still seem to want to see Maykop as the said Indo-Europeanizing agent in the Pontic-Caspian steppe, they're not exactly enthused by their own results. And they also make the following claim (emphasis is mine):

Based on PCA and ADMIXTURE plots we observe two distinct genetic clusters: one cluster falls with previously published ancient individuals from the West Eurasian steppe (hence termed ‘Steppe’), and the second clusters with present-day southern Caucasian populations and ancient Bronze Age individuals from today’s Armenia (henceforth called ‘Caucasus’), while a few individuals take on intermediate positions between the two. The stark distinction seen in our temporal transect is also visible in the Y-chromosome haplogroup distribution, with R1/R1b1 and Q1a2 types in the Steppe and L, J, and G2 types in the Caucasus cluster (Fig. 3A, Supplementary Data 1). In contrast, the mitochondrial haplogroup distribution is more diverse and almost identical in both groups (Fig. 3B, Supplementary Data 1).

I'd say that what they're almost suggesting there is that the Caucasus and Steppe clusters, hence also the Maykop and Yamnaya populations, shared significant maternal ancestry. If this were true, then perhaps it might mean that the Pontic-Caspian steppe was Indo-Europeanized via female-biased migrations from Maykop? Yes, perhaps, if this were true. However, it's not.

To be sure, Yamnaya does show a close genome-wide genetic relationship with an earlier group from the North Caucasus region: the so called Eneolithic steppe people. But they can't be linked to Maykop or even the roughly contemporaneous nearby Eneolithic Caucasus population, and seem to have vanished, at least as a coherent genetic unit, just as Maykop got going. Wang et al. managed to sequence three Eneolithic steppe samples with the following mtDNA haplogroups: H2, I3a and T2a1b.

H2 is too broad a haplogroup to bother with, but here are the results for I3a and T2a1b from the recently launched AmtDB, the first database of ancient human mitochondrial genomes (see here).

In a database of 1,131 ancient samples, I3a shows up in just five individuals, all of them associated with Yamnaya-related archeological cultures and populations: Poltavka (BARu), Unetice (UNC), Corded Ware (CWC), and Bell Beaker (BBC). Similarly, T2a1b shows up in just four individuals, all of them associated with Corded Ware (CWC) and Bell Beaker-derived Bronze Age Britons (BABI). And if I go back a step to T2a1, then the list reveals two Yamnaya individuals from what is now Kalmykia, Russia.

Thus, using just two mtDNA haplotypes I'm able to corroborate the results from genome-wide genetic data showing a close relationship between Eneolithic steppe and Yamnaya. So like I said, useful stuff.

This obviously begs the question: what does the AmtDB reveal about Maykop mtDNA haplotypes, especially in the context of the genetic relationship, or rather lack of, between Yamnaya and Maykop? Yep, again, the AmtDB basically corroborates the results from genome-wide genetic data.

But don't take my word for it. Stick the currently available Maykop mtDNA haplogroups into the AmtDB and see what happens (for your convenience I've made a list available here). Considering the close geographic and temporal proximity of Maykop to Yamnaya, you won't see an overly high sharing rate with Yamnaya and closely related populations. Moreover, Maykop shows several haplogroups that appear highly unusual in the context of the Eneolithic and Bronze Age steppe mtDNA gene pool, and, instead, link its maternal ancestry to those of the early European farmers, West Asians or even Central Asians, such as HV, M52, U1b, U7b and X2f.

See also...

Yamnaya: home-grown

Yamnaya isn't from Iran just like R1a isn't from India

Big deal of 2016: the territory of present-day Iran cannot be the Indo-European homeland

Late PIE ground zero now obvious; location of PIE homeland still uncertain, but...

All of the post-Middle Neolithic samples from the recent Mittnik et al. and Saag et al. preprints on the ancient population history of the Baltic region belonged to Y-chromosome haplogroup R1a. And most of them belonged to the R1a-M417 (R1a1a) subclade that makes up almost 100% of the R1a lineages in the world today. This is what the results look like in a table (the sample IDs are of my own design):

Earlier samples from the same region belonged to Y-haplogroups I2a and R1a, but this was a subclade of R1a defined by the YP1272 mutation that is extremely rare today even in Northeastern Europe.

And now shifting our focus west of Scandinavia: all but two of the post-Middle Neolithic samples from around the North Sea from the recent Olalde et al. preprint on the Bell Beaker phenomenon and ancient population history of Northwest Europe belonged to Y-chromosome R1b, and more specifically to the R1b-M269 (R1b1a1a2) subclade, which makes up almost 100% of the R1b lineages in the world today. Here's a table:

Earlier samples from the same region belonged to Y-haplogroups I2a, I, G2a and CF, and most of the instances of I and the CF would probably be classified as I2a if not for missing data.

Interestingly, despite the R1a vs R1b dichotomy between these post-Middle Neolithic obvious newcomers to the Baltic and North Sea regions, respectively, they were very similar in terms of overall genetic structure, obviously closely related, starkly different from Middle Neolithic Northern Europeans, and in all likelihood mainly derived from the same homeland that was not located in Northern Europe.

So can we locate this homeland with any degree of certainty, you might wonder? In fact, you might ask, isn't this a futile search for the time being, as we await ancient DNA from many prehistoric Eurasian populations?

Not at all, because when attempting to answer this question we're bounded by two key constraints: the exceptionally high frequencies of R1a and R1b in the post-Middle Neolithic Baltic and North Sea samples, and their close genetic affinity to earlier and contemporaneous populations from the Pontic-Caspian steppe, part of which is due to significant Caucasus Hunter-Gatherer (CHG) admixture that was lacking in Middle Neolithic Northern Europeans.

Indeed, to date, the Pontic-Caspian steppe is the only region where both R1a and R1b have been found in ancient remains from the same sites dating to the Mesolithic, Neolithic and Eneolithic. Here's a table based on results from Mathieson et al. 2015 and 2017. The R and R1 might really be R1a or R1b if not for missing data.

The Pontic-Caspian steppe also abuts the Caucasus foothills, and we know that CHG admixture was a major feature of its inhabitants from at least the Eneolithic. So odds are, and make no mistake, these are indeed excellent odds, that the homeland we're looking for was on the Pontic-Caspian steppe.

But of course I2a has also been recorded in prehistoric samples from the Pontic-Caspian steppe. So, you might ask, why did the populations migrating out of the steppe belong to R1a and R1b, and why did some of them seemingly carry only R1a while others only R1b? This can be explained by local founder effects on the steppe due to patrilocality. Moreover, it's possible that some groups moving out of the steppe did carry high frequencies of I2a, but they're yet to enter the ancient DNA record. [Edit: Maybe they already have? See here]

Now, the aforementioned post-Middle Neolithic newcomers to the Baltic and North Sea regions are most certainly in large part the direct ancestors of modern-day Northern Europeans, speaking languages belonging to the three daughter branches of late Proto-Indo-European (PIE): Balto-Slavic, Celtic and Germanic. It's highly unlikely that languages ancestral to these present-day languages were spoken by Middle Neolithic farmers, nor introduced into Northern Europe after it was colonized by the migrants from the Pontic-Caspian steppe.

What this strongly suggests is that the Pontic-Caspian steppe was also the late PIE homeland.

But, you might argue, the Pontic-Caspian steppe may have just been the expansion point for some of the late PIE language branches. No, that won't work. For one, modern-day populations speaking languages belonging to all other late PIE branches, such as Armenian, Greek, Indo-Iranian and Italic, show signals of the same population expansion from the Pontic-Caspian steppe that gave rise to modern-day Northern Europeans, in the form of Yamnaya-related genome-wide genetic admixture and appreciable frequencies of Y-chromosome haplogroups R1a-M417 and/or R1b-M269.

Some of these signals are certainly due to fairly recent admixture from Northern Europeans, like in much of Greece as a result of the Slavic expansions during the Early Middle Ages, but most cannot be explained in this way.

Secondly, Balto-Slavic, Celtic and Germanic are not more closely related to each other than to some of the other late PIE branches. For instance, Balto-Slavic is considered far more closely related to Indo-Iranian than to Celtic, which is generally seen as a sister branch to Italic. Therefore, if Balto-Slavic and Celtic derive from a homeland on the Pontic-Caspian steppe, then logically this is also where we should look for the origins of Indo-Iranian and Italic.

So as far as the late PIE homeland is concerned, thanks to ancient DNA, the debate is now practically over. But the PIE homeland debate is still wide open, or so we're told.

Apparently, Mathieson et al. 2017 aren't comfortable with putting the PIE homeland on the Pontic-Caspian Steppe because they can't find any evidence in their ancient DNA dataset of a significant migration through the Balkans that would potentially bring Anatolian languages from the Pontic-Caspian steppe to Anatolia. From the paper:

One version of the Steppe Hypothesis of Indo-European language origins suggests that Proto-Indo European languages developed in the steppe north of the Black and Caspian seas, and that the earliest known diverging branch – Anatolian – was spread into Asia Minor by movements of steppe peoples through the Balkan peninsula during the Copper Age around 4000 BCE, as part of the same incursions from the steppe that coincided with the decline of the tell settlements. [51] If this were correct, then one way to detect evidence of it would be the appearance of large amounts of characteristic steppe ancestry first in the Balkan Peninsula, and then in Anatolia. However, our genetic data do not support this scenario. While we find steppe ancestry in Balkan Copper Age and Bronze Age individuals, this ancestry is sporadic across individuals in the Copper Age, and at low levels in the Bronze Age. Moreover, while Bronze Age Anatolian individuals have CHG/Iran Neolithic related ancestry, they have neither the EHG ancestry characteristic of all steppe populations sampled to date [20] , nor the WHG ancestry that is ubiquitous in southeastern Europe in the Neolithic (Figure 1A, Supplementary Data Table 2, Supplementary Information section 1). This pattern is consistent with that seen in northwestern Anatolia [11] and later in Copper Age Anatolia [23], suggesting continuing migration into Anatolia from the East rather than from Europe.

And this...

On the other hand, our data could still be consistent with the Steppe-Balkans-Anatolia route hypothesis model, albeit with constraints. It remains possible that populations dating to around 1600 BCE in the regions where the Indo-European Luwian, Hittite and Palaic languages were spoken did have European hunter-gatherer ancestry. However, our results would require that such ancestry was not ubiquitous in Bronze Age Anatolia, and was perhaps tightly linked to Indo-European speaking groups. We predict that additional insight about the genetic origins of the potential speakers of early Indo-European languages will be obtained when ancient DNA data become available from additional sites in this key period in Anatolia and the Caucasus.

But I'd say the authors are taking that one particular version of the Steppe Hypothesis way too seriously. They might even be implying things that the creator(s) of the said hypothesis never posited.

Why do they seemingly expect a massive surge of steppe admixture into the Balkans during the Copper Age? If the steppe people are just shooting through the Balkans on their way to Anatolia, why would they leave a lot of admixture along the way? And if the locals are abandoning their tell settlements and running for the hills as far away from the oncoming steppe invaders as they can, how exactly would they acquire steppe admixture? Osmosis or what?

The Balkans is not Northern Europe, and the hypothesized migration of the proto-Anatolians from the Pontic-Caspian Steppe to Anatolia through the Balkans was never, as far as I know, meant to parallel the massive Corded Ware expansion across Northern Europe. In other words, why should all of the early Indo-European expansions have been of the same character, especially considering that they moved into such starkly different areas of Eurasia?

Indeed, as Mathieson et al. 2017 point out in the quote above, the evidence for the fleeting presence of steppe peoples in the Copper Age Balkans is in their dataset. For instance, in their Varna 1 sample set from Bulgaria, three out of the five individuals show significant steppe admixture. One of these individuals is almost 50% Yamnaya-like. Surely, there's really no need to expect anything more than that when looking for signals of a proto-Anatolian migration from the Pontic-Caspian Steppe to Anatolia.

In fact, even though I do appreciate the incredible work these guys are doing and the data they're making available to myself and everyone else, I suspect that there's a little bit of, shall we say, schadenfreude going on here.

They sequenced all of three Early Bronze Age Anatolians of obscure origin (are they actually suspected Anatolian speakers, like Luwians?), and apparently it's a big deal that they can't find any steppe admixture in Early Bronze Age Anatolia. Come on.

And then we're offered just three Yamnaya samples from the Pontic Steppe in Ukraine. One happens to be a massive outlier towards the Caucasus. Wow, what are the chances of that? And guess what, all three of these Yamnayans are females, so of course we're left wondering about the Y-haplogroups of the Yamnaya males on the Pontic Steppe. What happened to the males? Next paper, that's what.

Update 19//05/2017: Please note that the authors are not holding back any Yamnaya males from Ukraine for a future paper, as per my claim in the last paragraph above. They used what they had for the time being.

Update 21/05/2017: Actually, I suspect that we already have a population from the Bronze Age steppe in the ancient DNA record with a high frequency of Y-haplogroup I2a. See here.

See also...

R1a-M417 from Eneolithic Ukraine!!!11

Ancient herders from the Pontic-Caspian steppe crashed into India: no ifs or buts

Eastern Europe as a bifurcation hotspot for Y-hg R1

Globular Amphora people starkly different from Yamnaya people