search this blog

Friday, January 30, 2015

Half of our ancestry comes from the Pontic-Caspian steppe

Here's the latest teaser for the new David Reich et al. paper on the ethnogenesis of present-day Europeans. It's part of an abstract for a seminar to be held by Professor Reich at Jesus College, Oxford, on February 9. Interestingly, it argues that migrations from the steppe resulted in a ~50% population turnover across northern Europe from the late Neolithic onwards, which is very much in agreement with recent discussions on the topic at Eurogenes (for instance, see here).

By ~6,000-5,000 years ago, a resurgence of hunter-gatherer ancestry had occurred throughout much of Europe, but in Russia, the Yamnaya steppe herders of this time were descended not only from the preceding eastern European hunter-gatherers, but also from a population of Near Eastern ancestry. Western and Eastern Europe came into contact ~4,500 years ago, as the Late Neolithic Corded Ware people from Germany traced ~3/4 of their ancestry to the Yamnaya, documenting a massive migration into the heartland of Europe from its eastern periphery. This steppe ancestry persisted in all sampled central Europeans until at least ~3,000 years ago, and comprises about half the ancestry of today’s northern Europeans. These results support the theory of a steppe origin of at least some of the Indo-European languages of Europe, and show the power of genome-wide ancient DNA studies to document human migrations.

Source: Ancient DNA documents three ancestral populations for present-­day Europeans

Update 11/02/2015: Massive migration from the steppe is a source for Indo-European languages in Europe (Haak et al. 2015 preprint).

Wolfgang Haak et al., Massive migration from the steppe is a source for Indo-European languages in Europe, bioRxiv, Posted February 10, 2015, doi:


«Oldest   ‹Older   601 – 646 of 646
ryukendo kendow said...

@ Simon W

I think it is quite unlikely that the PIE language was the language of the Armenian-like population, as long as it seems that R1a arrived with EHG and not the pseudo-Armenians. Languages correlate strongly with fatherhood, but very little with motherhood. What correlates with motherhood linguistically is accent and phonetics from the mother's tongue. These have been proven empirically in a variety of sociolinguistic studies of mixed couples in diverse contexts.

Also, there is a debate on the linguistic affinities of PIE going on at Dienekes', but it is basically accepted that similarities btw PIE and Uralic are in some core vocab and structure, while similarities on the Kartvelian and Southern side generally are the high diversity of phonemes and some twice-loaned loanwords--aka the loans did not arrive directly.

ryukendo kendow said...

A few musing I'm gonna dump here from the perspective of social sciences.

It has been known for a long time that there are two major models for explaining response to perceived provocation among males (mostly) in different societies; one sees the free, proportional demonstration of irritation at slights, which coexists with a social scene that tolerates rudeness and brusqueness and satire at a high level, but generally favours peaceful negotiation and material, utilitarian measures of benefit to people. The other model sees highly exxagerated demonstrations of hospitality and ritualised politeness, and tendency to suppress irritation at slights, unless some breach of protocol is seen to give great disrespect, after which no-holds barred conflict is allowed to rise to the open; this tends to coexist with a social scene that is intolerant of brusqueness and carelessness in social situations and values a male's pursuit of honour often at ruinous cost to himself and his associates.

It has also been known that the second complex of values and behaviour tends to correlate with areas of pastoral economy, because herds are almost impossible to defend from stealing and pastoralism allows for extremely decentralised distribution of coercive capability amongst the male population. This causes 'one man's law' situations to arise extremely often and makes a person's reputation, a.k.a honour, the only social institution capable to curtailing violence directed at one's kin, family or property. It also favours the development of highly clannish, patriarchal and aggressive societies.

These can exist in basically the same population subject to different pressures, e.g. the scots-irish highlanders compared to their anglo counterparts (think hatfields and mccoys vs whigs and benjamin franklin), Dinka and Nuer vs. Ethiopian Amhara, north nigerians vs. Igbos, Pashtuns vs. punjabis, and appalachian whites vs. boston brahmins, or alps french vs parisians as late as the 17th century. Those who read sons of albion would know that these cultural differences have been inherited in the American N vs. S division for literally hundreds of years after the pastoral economy of the scots-irish ancestors of the southerners had been left behind, and were a great reason behind the appallingly large fraction of young southern men who literally threw themselves to their deaths in the civil war. Sherman burned atlanta to the ground because he understood that the entire southern self-conception as hard honour-bound warrior nobles vs the northern mercantile softies had to be utterly shattered if this kamikaze-like social situation was to be broken and thus for the war to finish early.

It seems to me quite likely that the introduction of pastoralism via domesticated animals into the Ural region just a few hundred years prior to PIE resulted in the sort of cultural changes one has seen repeatedly in pops like the Dinka and Nuer, and especially the Arabs, after pastoralism became the dominant economic model, aka a kind of 'barbarisation'--the cities of hadrahmaut, the queen of sheba and the gold and perfumes of the empire of aksum and etc. all being left behind for endless intercine tribal violence in the case of Arabs. These might then have been passed on to the daughter socieities as a cultural inheritance and resulted in long-term tendency to warlikeness and patriarchal norms in the societies they generated that were only submerged with the incporporation of Buddhism and vegetarianism into post-Vedic Brahmanism in India, and the Catholic Church taming the germanic tribes via feudalism in Europe.

ryukendo kendow said...

@ Shaikorth

Thank you for your comments.

That question has always been bugging me as ADMIXTURE components discovered in modern pops define important dimensions of their own in the variance, but it is not clear how much relevance they have for ancient genomes. Personally, I think the fact that e.g. later HGs don't score high in East Euro until the IEs appear, and e.g. Danubian neolithic shares components with Caucasus while Cardial aDNA only with Mediterranean, seem to suggest that they do have some relevance. Furthermore the way ADMIXTURE is structured is to find those dimensions in which the putative admixing pops differ in ways most conducive to explain present variation, so I won,t be surprised if, e.g. Atlantic maximised in BB, etc.

Nirjhar007 said...

''PIE and Uralic are in some core vocab and structure''
Uralic has Zero donation to PIE etc and such structures also exist between PIE-semitic.
''I think it is quite unlikely that the PIE language was the language of the Armenian-like population, as long as it seems that R1a arrived with EHG and not the pseudo-Armenians.''
R1a is not Exclusively= Indo-Europeans!
There was the 6000 BC migration from Zagros to Urals which could have brought R1a to the E Europe as i agree with Maju.
However as we see in case of Maykop Invasion of Yamnaya ~3000 BC it was the Prominent Indo-European intrusion to Europe.

Mike Thomas said...

@ RK
"Languages correlate strongly with fatherhood, but very little with motherhood"

Again with the non-specific; unsubstantiated generalisations . U guys can model till the cows come home; but your perspectives will meet with more success when your social theory improve

After all, why is it called a 'mother tongue "?

Nirjhar007 said...

Good reasoning! lets see what the ''Delete Master'' suggests!;)).

ryukendo kendow said...

" After all, why is it called a 'mother tongue "? "

So this is your 'social theory', and what passes for 'good reasoning'.

The conclusion that no more than 10% of a migration of males was needed to invoke language change in the ancient world was reached by... Colin Renfrew.

If I cite everyone who I read that would take far too long.

You don't seem particularly interested in finding the truth?

Mike Thomas said...

@ RK

"So this is your 'social theory', and what passes for 'good reasoning'."

Clearly, this was a tongue in cheek comment, and not extensive analysis. But in jest there's truth.
"The conclusion that no more than 10% of a migration of males was needed to invoke language change in the ancient world was reached by... Colin Renfrew."

Perhaps, But so what? Language change can begin at the level of one (!) household; an incoming woman. So 1% , and a female !

"You don't seem particularly interested in finding the truth? "

I am. But i don;t think it'll be from those who brandish clown-ish, broad-brush, non-contextual statements just to make sense of their 'numbers'.

postneo said...


on assymetric language inheritance. The main constraint on language is mutual intelligibility not father and mother stuff. In a multilingual scenario the fathers language will prevail when the mother is a minority speaker and has married in. This would be the norm in archaic societies.

on agression; In males aggression is very sophisticated. There is many layers of bluffs threats, role play, games sometimes ritualized. In females its more binary. This is seen in all mammals e.g. even seemingly simple animals like Bulls with whom I have interacted some...! vs cows.

Davidski said...

Can anyone access this supp info pdf to see if there are any interesting plots of the ancient Danish and Hungarian genomes?

Roy King said...

It is a methodological paper on aDNA--no results about specific samples.

Chad Rohlfsen said...

I've contacted Eske, to see if the genomes are available and if he has any plots.

Krefter said...

"It is a methodological paper on aDNA--no results about specific samples."

Thanks for the info. That saves an hour of useless searching.

Davidski said...

It'd probably take you about 5 minutes to look over the supp info pdf to see if there were any PCA plots there. I couldn't do it because the download keeps failing.

Arch Hades said...

If the British have half their ancestry coming from the Steppe from the Bronze age why is there so little R1a in Britain? Despite this being the major lineage of those Steppe groups. Some sort of founder effect from a non R1a carrying male/group of males who still had a lot of steppe ancestry?

What type of genomes are they using as a model for pre steppe ancestry in Britain?

It might take years for this paper to pass peer review and come out into print. All we have right now is an abstract from a conference that hasn't even taken place yet.

Stoneville Inc said...

I'm sorry if I'm off topic here, but does anyone know if they are still working on this Summerian DNA paper or was this a bogus story?

Davidski said...


I think it's already clear that Bell Beakers and Celts brought most of the ANE to the UK and Ireland. They probably mostly carried R1b.

What isn't clear just yet is how the Bell Beakers and Celts acquired their ANE and R1b.

And I doubt the study will take several years to peer review. The longest I've seen a manuscript on population genetics being reviewed was six months. So I'd say the paper should be out within a couple of months.

Davidski said...

Stoneville Inc,

Yes, the story is genuine. The rediscovery of that Ur skeleton was big news late last year. I'd say the DNA analysis will probably take a year or more.

Also, the Rise project is testing ancient samples from Armenia. From memory, I think they date back to the Iron Age, but maybe there are older ones being analyzed as well?

ryukendo kendow said...

@ Davidski
Holy shit thats amazing!

The Sumerians--now thats a mystery.

Pre-arab, pre-african slave trade and pre-semitic. I wouldn't be surprised if they turned out even more ENF than Bedouin. If they turned out almost pure Basal, that would open up so many possibilities for analysis here.

@ postneo

Actually agree, peers>parents for linguistic dominance except under exceptional circumstances.

@ Mike Thomas

I'm sorry, your logic just boggles the mind sometimes. You seem to have this idea that any valid 'social science' deconstructs phenomena to reveal a mass of human interest stories one can watch on the news?

How does a single person with a language change the language of even one other person in a foreign group, when language is a social and not an individual tool for comm? Why would anyone feel any need to learn her language?

Maybe that happens in, like, one in 10,000 cases, but social scientists are under no obligation to explain that any more than animal ethologists have to explain videos of water-skiing budgerigars.

Why do trendy teens in Nagaland start speaking hindi, despite their ethnic insurgency? Why do speakers of basolects shift lect in the presence of a acrolect-speaker, but acrolect-speakers never change lect? Why are Nagaland teens and basolect speakers nonplussed/shocked when you point out the apparent discrepancies? Obviously this is not a conscious thing, much less 'choice'.

The fact that PIE were once a single linguistic community or closely related communities is something that is accepted, because e.g. expansion of romance or arabic lects from once again a single linguistic community--a very small group in both cases--is documented to produce a series of regular sound correspondences in the historical period, and the methods linguists use just extrapolates this methodology into older and older linguistic communities. As for latinised or arabised culture, you tell me.

If a solar storm destroys technology and humans develop pre-industrially for another 5000 years, what would happen to the soc. and cultural practices of Arab-speaking MENA by that time, to say nothing of their language? Something like.... IEs today? Copernican principle dude. Furthermore look at Papua New Guinea. Language=their version of politicised identity, dividing them more or less exactly. So this phenomenon is old.

'human behavior cannot be reduced to a law'. That is so sixties. Try telling that to any economists, structural historians, cliodynamicists, social scientists, linguists and sociolinguists, political scientists,, economic hists, ecological hists, etc.etc. The only place that still holds is in SOME parts of anthropology, SOME parts of archaeology and in marxist/racial/gender cultural criticism, which says more about those fields than about the way the world works.

I suggest you read up more social scientists than those in anthropology or archaeology.

Krefter said...


I really doubt Sumerians were pure basal or even ENF. Does anything suggest Semetics brought more ANE? I think Sumerians will be 100% ENF or ENF+WHG, or had some ANE.

I'm interested in seeing Sumerian genomes because....

West Asia has a very complex genetic history people seem to ignore and think can be understood by looking at distant cousins in Europe(EEF). EEF and modern west Asian mtDNA mostly split from each other in west Asia long before the Neolithic.

I think west Asian genetic history mostly involves local diversity(People within that region mixing with each other), unlike in Europe(foreign ANE and ENF).

I think "ENF" existed in west Asia long before the Neolithic, because Greek HGs had similar mtDNA to Neolithic near easterns(and all modern west Eurasians).

Also, something important to remember is ANE came to west Asia separate from EHG and south Asia. This and Yamna genomes suggest very old admixture between ENF and ANE.

Krefter said...

@Sumerian article,
"The ancient Sumerians, builders of the world’s first known civilization, are a mystery to us. Settling in what we would now call southern Iraq from about 5400 BCE on, they produced a written language, a complex system of mythology, impressive architecture, and a lost world that held regional hegemony for thousands of years."

This kind of ticks me off. How do we even define civilization? There was no sudden moment when civilization began.

This reminds me of this...

I think there's a Eurocentric approach to "ancient civilizations" influence on modern people. Old Euro historians treat themselves as *the* descendants of the ancient civs, even though before the middle ages they weren't at all. The middle east is as much descended of those civs as Europe is, but the history books barely mention that.

"Civilization" wasn't passed down from the Sumerians, to the Egyptians, to the Greeks, to the Romans, and to Europeans. That's such a simplistic and false interpretation of ancient history, that schools teach.

My history books in school treat modern Europe as a continuation and the same entity("western civilization")as ancient Greece, which is total bull shit.

If Christian Europe is a continuation of Greece, then the Muslim Middle east is a continuation of the Persians. See how crazy that is?

IMO, The only reason Europe is associated with those people, is the *late* Roman empire, limited ancient writings, Christianity(Bible writings), and historians.

Western civilization(eg, Rome, Christianity) was a foreign thing to most Euro's ancestors, and were adopted in the last 1500 years. Islam is an even more recent phenomenon in the middle east.

Teaching that ancient Rome and Greece as the sole ancient Euros, is the same as treating ancient Arabs as the sole ancient Middle easterns.

Davidski said...

There's a good article about that Ur skeleton at the History Blog.

Tesmos said...

I am convinced that the northern Bell Beakers still had alot of CWC like ancestry.

ryukendo kendow said...

@ Krefter

" because Greek HGs had similar mtDNA to Neolithic near easterns(and all modern west Eurasians)."

Where did you read that? That's incredibly interesting.

I agree that the probably paraphyletic basal groups in the ME would be incredibly difficult to untangle.

ADMIXTURE's caucasus-centered components track semitic speakers as opposed to cushitic speakers in Ethiopia. Cushitics share red-sea centered comps with Bedouin--as do all Afroasiatic speakers except sometimes berbers-- but all semitics share comps with caucasus in addition.

This, coupled with YHap J and the fact that Semitic appears to have come into mesopotamia from the Northern levant/mesopotamia area, makes me think that the proto-semites might have carried ANE, thus explaining the high level of ANE there even if the proto-afroasiatics did not and in all probability were from Africa. We know that ANE in ME has to be later than neol because neol migrants into from both anatolia-->Danube or Levant-->Med carried none.

For those wondering about semitic in Africa or S Levant, I don't see how one can get around p-semitic loans in PIE and words for horse, camel, and petroleum minerals e.g. bitumen in p-semitic.

PersonaMan said...

From Wikipedia, referenced are Cunliffe and Needham:

The Lower Rhine region had, by 3000 BC, adopted a burial rite characterized by single inhumation accompanied by a beaker decorated with cord zone impressions, and frequently by a perforated stone battle-axe.[3] This cultural package was characteristic of belief systems which extended across the North European Plain and into Russia.[3] The arrival of the Maritime Bell Beaker from the west a century or two later initiated a period of borrowing and experimentation in what has been called the Primary Bell Beaker/Corded Ware contact zone and cultural traits developed here, such as single burial and the shaft-hole axe, were transmitted westwards along the exchange networks from the Rhine to the Loire.[20][21] It was from this fusion zone that the modified Beaker package spread northwards across the Channel to Britain.

Just thought i'd throw it in there, i'm pretty new to this sort of period, all interesting to me.

It's the perfect opportunity for northern Bell Beakers to 'have a lot of CWC like ancestry' as Tesmos said.

Just a point of discussion, i'm not really adding anything, just want to watch from the sidelines. :)

G. Dekaen said...

Warning, I'm not sure how many posts this will be, maybe three or four. Sorry for that, not trying to spam or anything!

Hello everyone, I'm a new poster and have been trying (unsuccessfully) to keep up with the torrid pace of discussions the past few weeks. I had a hypothesis to propose before Reich's hopeful revelation of something big in that presentation, hopefully they're not too ignorant.

I ran the numbers for CO1, BR1 and BR2 with the K7 ANE Calculator to get an idea of the genetic change in Central Europe with roughly 1000 year intervals (2800BC-2100BC-1200BC). CO1 numbers I took from Vesuvian Sky's post on The Apricity since his numbers seem to match overall quite well with Eurogenes' official figures. Otherwise, the numbers come from Chad and Davidski. Here are the following numbers:

CO1: 60.4% WHG-UHG, 39.6% ENF
BR1: 84.3% WHG-UHG, 7.3% ENF, 7.74% ANE
BR2: 63.8% WHG-UHG, 21.7% ENF, 11.8% ANE
Yamna: 35-40% WHG-UHG (closer to 40%), 35-40% ENF (closer to 40%), 20-25% ANE (closer to 20%)
Catacomb: 50-55% WHG-UHG (closer to 55%), 15-20% ENF (closer to 15%), 30% ANE

My estimates for Yamna derive from Pattern/Reich's statements that Yamna was 50-50 EHG/Armenian using various possibilities for EHG from being identical to SHG (to give me a max WHG range), all the way to hyper ANE (up to 68%, to give me a max. ANE. range), with the assistance of mtDNA trying to narrow down what was more likely/reasonable. Catacomb estimates are based on extrapolations from mtDNA comparing it to Yamna/LBK mtDNA and patterns in autosomes. Anyways, the results lead me to believe that Yamna/Catacomb. didn't really have much of an impact on the European gene pool or dispersion of IE languages, basically a dead end. It looks like things are going to be a lot more complicated and interesting than a simple Yamna dispersion model. Hopefully there will be some clues to this effect in the paper/presentation.

The change from CO1 to BR1 required an 80%-100% population influx/replacement according to my numbers, essentially, indicating near total replacement or that the incoming Mako culture didn't incorporate descendants of the Baden culture. The cumulative population change (or the ancestral population if it was just one wave) that led to BR1 had 87-90.3% WHG-UHG, 0-3.7% ENF, 8.6-9.6% ANE. It seems extremely difficult to fit Yamna/Catacomb as being in any way responsible for any aspect of this change (even with a model of multiple waves from multiple sources, Yamna/Catacomb statistically probably couldn't contribute more than a token few percent if that), it had too much ENF, ANE and too little WHG-UHG. Rather, this indicates that there was either a major population wave from NE Europe/S. Baltic/Eastern Corded where the people were rather "pure"/direct descendants of S. Baltic HGs (Poland-Lithuanian HGs prior to 3000BC had 6 U5b and a single U4 mtDNA, they likely were 7-14% ANE, 86-93% WHG-UHG). I doubt this came from Corded Ware in Germany or even Poland since ENF was likely too high there (around 70% of mtDNA is farmer derived), but more likely Belarus or even further East Corded. The problem is the further East you go, the more ANE is encountered and this incoming population had a relatively low ratio of ANE:WHG 1:9, so an alternative scenario (and I think slightly more probable) would involve two influxes, one from the West of purely HG types to increase the WHG-UHG and a second influx from the East (Corded Ware) which provided the modest ANE and insignificant ENF. If we assume Eastern Corded Ware had about 15-20% (30% at most) ANE, then this eastern migration could have been responsible for 30-60% of BR1's ancestry, enough to transmit their "Geminate" language, more on that below.

G. Dekaen said...

Also, since BR1's ANE is only about half of modern Hungarians, this indicates that the increase in ANE in Europe was not a single-stage event, but involved many phases and since Yamna/Catacomb was not responsible for BR1, ANE cannot exclusively be connected with (P)IE expansion. Instead, I think this northern population expansion (BR1) reflected the hypothesized "Language of Geminates" family which seems to be a major substrate in northern Europe, influencing via substrates the Celtic, Germanic, Balto-Slavic and possibly even Finnic languages (!). More on this later. In effect, Yamna/Catacomb seems to have had no connection with BR1 and BR1 likely did not speak an IE language.

I'm not saying that Yamna/Corded didn't expand in the copper age, maybe they did (which is what Reich will likely show us), what I'm saying is that that expansion was not IE or that BR1 replaced/wiped out the Yamna/Corded expansion, which were then in-turn largely replaced and assimilated by EBA/MBE IEs from the steppe/forest-steppe(BR2 ancestors).

The change from BR1 to BR2 was more modest with a minimum influx of 20% new population. Unfortunately, the change in the proportion of the components wasn't radical enough to be able to triangulate the source/eliminate probabilities. The range of possible replacement was from a minimum of 20% to 90% with the incoming population having a range of (going from 20% to 90% replacement): 0-61.5% WHG-UHG, 79.3-23.3% ENF, 28.2-12.3% ANE. The most important clue comes from the increase in ANE, with a cumulative increase of at least 12.3% and a max. of 28.2%. This indicates continued influx of ANE, most likely from the Steppe. It's also coming from a population which came with double or more ANE than BR1's ancestral population, and it is in this time period (MBE), that I think IE languages began their major expansion into C. Europe. More on this in just a bit.

The second partial clue we might have is in CO1 having 60.4% WHG-UHG (and since "Thracian" V2 from 1100-1500BC south of CO1 has roughly the same EEF-WHG-ANE Calc score as CO1, 88/7.6/4.4 CO1, 82.9/7.8/9.3 V2, we can logically assume V2/SE Euros had a similar K7 score to CO1) meaning that scenarios where the incoming population(s) is less than 50% WHG-UHG (this would basically mean an incoming population outside of Europe, Anatolia/Caucasus/Central Asia) can be considered as unlikely (especially since BR1 lacks West Asian, Central Asia is still theoretically possible though). This would mean the likely influx was between 60-90% incoming new population/change with the new population containing a cumulative average of: 50.1-61.5% WHG-UHG, 31.3-23.3% ENF, 14.5-12.3% ANE. The ENF is too high here for a single migration from E. European steppe/forest-steppe, so there may have been two migrations. If we use IR1 (57% WHG-UHG, 18% ANE, 22% ENF) and V2 (CO1=V2 in EEF Calc, they are virtually identical, so likely around 60% WHG-UHG, 40% ENF, 0% ANE, but possibly in low single digits) as rough steppe and SE. Euro bronze-age proxies, a migration of 70% IR1/steppe and 30% V2/SE. Europe gets us some pretty sharp averages right on the money, for example: 70%*22 IR1 ENF=15.4 + 30%*40% V2 ENF=12-->27.4% ENF (right in the middle of the cumulative replacement numbers), 70%*18 IR1 ANE=12.6% ANE (again, right in between the cumulative range of 14.5-12.3%), 70%*57 IR1 WHG-UHG=39.9 + 30%*60% V2 WHG-UHG=18-->57.9% WHG-UHG.

G. Dekaen said...

Basically, it boils down to an IE expansion (of Germanic and Balto-Slavic, and probably Italo-Celtic; Indo-Iranian expanded a bit before this as did Greek, and Hittite well before this from the Caucasus southwards) deriving out of Srubna/Unetice Cultures-->Tumulus-->Urnfield and boom, huge spread. Something I haven't heard anyone mention so far in tracing the PIE homeland is that the word for the "yew tree", taxus baccata (according to the "Oxford Intro to PIE and the PIE World") which is solidly reconstructed in Hittite, Greek, IR, Balto-Slavic and Celtic languages combined with Dybo's research showing PIE had an exceeding familiarity with mountainous terrain makes the Caucasus and Carpathian mountains the most likely spots for PIE because yew trees do not exist on the steppe at all (nor in Ukraine, only in Crimea, so outside of the range of Yamna), nor in the Urals, but DO exist in the N. Caucasus mountains/along the coasts of Anatolia and in the Carpathians (their easternmost border in Europe). Given the gradual increase of ANE coinciding with the appearance of IE languages, the Caucasus is more likely than the Carpathians. Beavers also seem to have existed in the Caucasus in the Neolithic and possibly later.

So my hypothesis is that PIE was spoken in the N. Caucasus (Maykop Culture) until it started breaking up no earlier than 3500/3000BC (PIE knew plough, cart, copper), Hittite went south, linguists don't seem certain when Tocharian split, if right after Hittite or a while after, but basically, IEs stayed in the N. Caucasus slowly expanding into southern Very Late Yamna/Early Catacomb (or just Catacomb, not 100% sure, the finding of mtDNA R1 in Catacomb attests to likely Caucasian input) which came into contact with non-IE Poltavka/Abashevo/Fatyanovo-Balanovo (higher WHG-UHG, lower ANE/ENF, higher blue eyes, the "western" part of this probably spoke "Language of Geminates", eastern part/Abashevo spoke perhaps Uralic) cultures, and mixed with them. The eastern version went on to create Indo-Iranian Sintashta/Petrovka-->Andronovo, while the western variant which mixed with Geminate Language family created Srubna/Unetice representing Italo (?)-Celtic (Celtic has minor Geminate substrate, not sure of Italic), Germanic, Balto-Slavic and later Finnic and spread via the following Tumulus and Urnfield cultures. The major expansion of IE languages (except Hittite, possibly Tocharian, Greek/Armenian/Albanian) started probably around 1800BC, just a bit after BR1.

This might also make sense of some otherwise rather fantastical claims by certain Celtic and Germanic peoples of their descent from "Scythians," maybe they weren't too far off after all; at least they seem to have gotten the location more or less right, and just mistook the Scythians for being their ancestors when it was in-fact the people immediately preceding the Scythians.

I'm really sorry about the long post, I tried being reasonably comprehensive yet succinct, but I'm sure there are a lot of holes that still need to be plugged, in time I'm sure, but I'm confident there is potential here for those specialists who can help develop it further (especially with archaeology and physical anthropology, I'm a bit fuzzy with those); Maykop is after all the second most popular PIE homeland after Yamna, the difference between my idea and most is that I see IE expansion as an EBA or MBA phenomenon rather than Copper Age/Yamna/Corded/Bell Beaker. I can provide all the full numbers if anyone is curious.

G. Dekaen said...

The above autosomal data and hypothesis seems to be corroborated by the general paucity of blue eyes (a WHG diagnostic trait; core/homeland ANE were likely brown-eyed, only in mixed WHG-ANE fringe areas did ANE have blue eyes) in Yamna. Isn't it a little odd for a culture like Yamna, which only had a frequency of 7% blue eyes (2/28 G/G with an additional 4/28 with A/G in rs12913832, overall frequency of G allele at 14%) to contribute 50% of the ancestry of northern Europeans who have on average around 60-80% blue eyes AFAIK? While I'm well aware that Gok2 had blue eyes (small sample size, 1/1), in order for this 50-50 mixture to reach 60-80% blue eyes through simple mixture, the native 50% Gokhem/TRB people would need a frequency of 113-153% blue eyes which is obviously impossible. I also know that selection has been posited as a possible reason for the increase in blue eyes over the course of 4000 years, but frankly, that doesn't seem convincing considering that only a couple hundred years after Catacomb2, 2000-2500BC (2/15 G/G, 13% blue eyes, an additional 1/15 A/G, total frequency of G allele is 17%) we see blue eyes at a frequency of 80% (4/5) among samples with non-East Eurasian uniparental markers/admixture in Andronovo, 1800-1400BC. That's a huge jump in blue eyes without much time or space passing by between Andronovo and Catacomb, the most logical explanation for this change would be that Andronovo descends from a migration from a population group very different from Yamna/Catacomb (presumably with elevated WHG). According to the archaeology, Andronovo descends from Sintashta-Petrovka which in-turn descends from a combination of Abashevo and Poltavka. And while I know that Poltavka is supposedly an archaeological descendant of Yamna (perhaps just acculturated HGs?), maybe the increase in blue eyes, as well as WHG/ANE came from this nexus of cultures: Abashevo, Poltavka and possibly Fatyanovo-Balanovo/Eastern Corded Ware in the Oka-Volga-Kama area rather than Yamna/Catacomb.

An interesting piece of potentially supportive evidence would be from a recent paper on milk consumption covered by Davidski which found 1/1 sample from the Balanovo Culture had evidence of milk consumption, meaning Balanovo was possibly partly responsible for the spread of lactose tolerance throughout Europe along with WHG, since the two correlate rather well. Also notice that the two Corded-aged samples from Denmark, Oster Hurup and Gjerrild both tested negative for milk consumption which could indicate a difference in people/culture between Eastern (Balanovo) and Western (SGC) Corded. Increasingly, I think we should start looking at a post-Corded, EBA-MBA time-frame to capture the main-rearrangement in genes that brought about the IE languages and increase in blue eyes for Europe. Unetice was almost certainly involved in this as was Tumulus Culture and its successor, Urnfield. I wonder if we're paying far too little attention to the impact these later cultures from the MBA had on reshaping the genes and languages of Europe (IE expansion).

Anyways, just an idea. Can't wait for the paper/presentation.

G. Dekaen said...

Also, a Romanian study that analysed mtDNA from the Bronze to Iron Ages just came out if anyone can access it:

Someone on eupedia forum posted the mtDNA recovered from the study, but I don't have an account and can't see the image, the link is below. Can anyone check that out, the study had something like 50 samples tested, so it may reveal some interesting and important bronze age changes in Romania that help uncover the expansion of IE. On that note, I'd like to hypothesize that the Thracian V2 sample from 1500-1100BC was a pre-Thracian, and possibly pre-IE speaker, while P192-1 from 800-500BC was almost certainly Thracian speaking, as the EEF-WHG-ANE Calculator shows a significant increase in WHG (+15%), a significant decrease in EEF (-20%), and a minor increase in ANE (+5%). We'll see if the results from this study support these assertions.

As for Sumerians, they almost certainly had ANE present in their population since we have two U4 samples and a single R sample found in Sumeria dated 2700-2900BC and 2550BC. U4 is strongly correlated with ANE and R was also likely present in strongly ANE populations. These three lineages seem to point Northeast.

Also, Figure 2 in the following study of dental traits might also affirm the existence of ANE in Turkmenistan as early as 4000BC since it shows Painted Ware Culture of Turkmenistan clustering in the left-most branch (#9) with European Mesolithic HGs from Latvia and Karelia (who are known to have ANE) in addition to European Neolithic cultures from the steppe, like Dnieper-Donets instead of clustering with the right-most branch which has Neolithic cultures from 9000-7000BC in Anatolia. Intriguingly, population #8 are Ukrainian Tripolye and form the tightest cluster with Painted Ware Turkmenistan; could that imply that Tripolye absorbed some ANE from the previous Dnieper-Donets or Yamna?

G. Dekaen said...

Oops, forgot the link:

Davidski said...

Er, the Yamnaya guesstimates we've been discussing here are based on the K8, not the GEDmatch K7.

The two tests give very different WHG vs. WHG-UHG estimates, with the K8 being much more accurate.

Simon_W said...

Damn, I was being crazy! :-D I don't need GEDmatch, I had used DIYDodecad before. And BR1 seems to have enough SNPs in common with the illumina chip for a useful analysis with DIYDodecad:

Alright, BR1 is

in K7b =
88.52% Atlantic_Baltic
11.47% Southern
0.01% West_Asian

in K12b =
54.65% North_European
43.81% Atlantic_Med
1.53% Caucasus
0.00% Gedrosia

in Globe13 =
63.81% North_European
35.95% Mediterranean
0.15% Southwest_Asian
0.00% West_Asian

So definitely no late Neolithic/EBA highland West Asian admixture. Gedrosia neither.
I'm not sure atm how to interpret that. If the PIE homeland was in West Asia, it would follow that BR1 wasn't an IE. Or it could mean that she was IE and that the homeland wasn't in West Asia. But if the homeland was in West Asia, then IE language arrived in the Carpathian Basin at some time in the 2nd millennium BC, in the developped Bronze Age, and I wonder if that wasn't too late for Italo-Celtic to arrive.

Nirjhar007 said...

BR1 wasn't Indo-European.

Simon_W said...

Call me crazy, but the fact that BR1 had 0% Gedrosia makes it unlikely to me that there was R1b in this population, in the Carpathian Basin, around 2000 BC. Apparently it was the result of an early wave of EHG admixture. And if the Globular Amphora people to the north had some early EHG admixture too (which seems likely to me, because of the craniometrics), then it was probably similar to the early wave into the Carpathian Basin, that is, also lacking R1b and Gedrosia.

Yet there was R1b in Bell Beaker people in Kromsdorf, and the area of elevated levels of Gedrosia in Europe overlaps strongly with the extent of the Bell Beaker culture. So there is only one reasonable conclusion: R1b, Gedrosia and Bell Beaker spread together from somewhere in western Europe. And R1b and Gedrosia had arrived there somewhat earlier from West Asia. The question is: Did they spread Italo-Celtic? Or an Iberoid language?

Similarly in the east, in the Yamnaya culture, there was 50% Armenian-like admixture. I'm quite sure that this was accompanied mostly by R1b. J2 is too rare in the whole east of Europe. R1a might be an alternative, but to me its distribution looks more EHG derived. And R1b is rather common in Armenians, unlike R1a. Similarly to the Bell Beaker case, the question is: Did these pseudo-Armenians spread Indo-European? Or a Kartvelian language?

So we've got two waves of West Asian and probably R1b related populations into late Neolithic Europe. BUT of course this isn't the only population expansion we've seen. At roughly the same time there was an expansion of EHG admixture into the Carpathian Basin and to the Rhine. The North European autosomal component also spread in central Asia and into northern India (see the MDLP World-22 Northeast European component!). And there is European R1a + Balkanic I2a in southeastern Turkey, evidence for an EHG admixed population movement from the Balkans. Estimates for the North European admixture in Iberia date it to about 1500 BC, the Middle Bronze Age. So really, from the genetic point of view either expansion may have been the IE one.

Furthermore we cannot safely deduce the spread of languages from admixure proportions. Does the ruling class always enforce their language? Of course not, see the change of the Bulgarian elite from Turkic to Slavic and the change of the Frankish elite from Germanic to Gallo-Romance. Does the language of the majority always prevail? Certainly not, see the change of the Irish populace from Gaelic to English, etc. Also I'm not sure if the Bell Beaker expansion spread any language at all. If these were primarily traders they may have been keen to learn the languages of the locals.

But back to the PIE homeland. If I had to guess I would right now say it was the steppe. Because – why the heck should Uralic speakers let their word for water be influenced by West Asian newcomers? Seems unlikely to me.

Nirjhar007 said...

''But back to the PIE homeland. If I had to guess I would right now say it was the steppe. Because – why the heck should Uralic speakers let their word for water be influenced by West Asian newcomers? Seems unlikely to me.''
That word is also similar with Dravidian and goes back to Nostratic periods common heritage nothing special and basic structure wise PIE also share similarity with Semitic!.

Davidski said...

BR1 belonged to the Mako culture, which emerged in the Carpathian Basin after Yamnaya migrations into the area. Map...

Nirjhar007 said...

Any News from the Conference?.

Davidski said...

I should have something in a couple of hours.

Nirjhar007 said...

Thank You!

Simon_W said...

Yes, initially I also thought that only the local Yamnaya can explain the change from CO1 to BR1. Chad pointed out that it may also stem from pre-Yamnaya steppe admixed eastern Balkan populations expanding with the early Bronze Age. That might explain the lack of West Asian admixture, if Yamnaya did have such an admixture, which we don't know for sure, but which the rumors about 50% Armenian-like ancestry kind of suggest.

Simon_W said...

BTW it occured to me today that the high incidence of R1b in Basques, combined with their strong Gedrosia component seems to suggest that there is a lot of Bell Beaker ancestry there. But what differentiates the non-IE Basques from their IE neighbours, according to Dodecad K12b? Spaniards and Portuguese have more Northeuropean admixture. And also more ANE, according to Laz. et al. 2014. If correct, this would mean that the incoming IEs brought in more Northeuropean admixture and ANE, but Gedrosia and R1b were already in place before them, combined with a bit of ANE.

Seriously, the correlation between R1b and Gedrosia looks so well that it would be a great surprise to me if there was any R1b around in the Carpathian Basin around 2000 BC, in a population with 0% Gedrosia. BR2 in contrast does have some slight Gedrosia admixture, 1.14%. He also had appreciable West Asian admixture and most of all, he had haplogroup J2. So it looks very much as though R1b moved to the Carpathian Basin late, and together with J2, probably both from a West Asian source.

Simon_W said...

With the Carpathian Basin (as per above) having no R1b until after 2000 BC, with the Corded Ware being R1a dominated, and with potential pre-Corded Ware EHG admixed migrants (Globular Amphora) presumably being similar to BR1, it's hard to derive the Kromsdorf R1b from an EHG source to the east. A spread with Iberian Bell Beakers to central Europe looks more likely. This does in no way rule out a presence of R1b in Yamnaya.

G. Dekaen said...

@ Davidski
Yes, I'm aware that the K8 is more accurate in measuring ANE and WHG. I had to use the K7 because BR1 doesn't have a K8 score and my intended purpose was to see what changes happened between 3000-2000BC and from 2000-1000BC if any.
Looking back at my argument(s), I must admit two main problems:
1) I speculate that BR1 came from a non-Yamna/Catacomb (non-IE, instead Geminate language) population far to the NE of the steppe, possibly from E. Corded Ware/Abashevo/Poltavka who were overwhelmingly WHG-UHG and blue-eyed who would go on to later mix with NW. IE speakers (Germanic, Balto-Slavic and Italo-Celtic likely in Catacomb), transmitting blue eyes to NW IEs who would later spread it into Europe circa 2000-1000BC via Srubna/Unetice/Tumulus/Urnfield cultures. Funnily enough, despite BR1 being so radically WHG with limited ANE and ENF, he actually had brown eyes, not blue. That could be an indication that selection for blue eyes did indeed occur. We'll have to see.
2) I've been further toying around with the WHG-UHG/ANE/ENF estimates based on mtDNA proportions, tentatively identifying U2e as 100% ANE with U4, U5a, T1, W, I and R lineages being assigned as 50% ANE and 50% WHG (U4, U5a) or 50% ENF (T1, W, I, R). mtDNA C is also a strong candidate for being a primarily or significantly ANE marker considering its early presence in E. Europe and in Neolithic Syria (PPNB) (although I can't seem to find the reference anymore on (ancestraljourneys). Anyways, I don't have too much confidence in the proportions I listed in the above posts because they depend too much on methodologies with too many unknowns, namely Y-DNA.
The only one we can have a half-decent idea for is Yamna given Patterson/Reich's admission of being 50-50 Armenian-EHG. That puts Yamna at around 40% ENF as we have to remember that EHG's were found to have carried mtDNA H, probably taken from interactions with Dnieper-Donets further south who were 40% H. It means that it's possibly EHG's were ever so slightly ENF, 5-10%. Otherwise, using a wide variety of models/possibilities, Yamna was anywhere between 6-44% WHG-UHG and 20-86% ANE. At the very least, this gives us an upper limit of 44% WHG-UHG in Yamna (based on assuming EHGs were identical to SHGs to maximize the possibly WHG score) which may be useful. The Catacomb mtDNA would seem to show that the huge influx of U4 (26%) represented an increase in ANE from the E. or NE., so their WHG-UHG was likely even lower than 40%. That's about all we can say with a smidge of confidence.

G. Dekaen said...

@Simon W
According to the latest information from Jaska, IIRC, proto-Uralic's earliest contacts with PIE are actually not with PIE itself, but rather with what he calls: "late archaic IE/NW IE" and "early proto-Aryan" dated to 2800BC. Proto-Uralic does not seem to have had direct contacts with PIE, but rather with later, derived forms of IE - basically, it had already split. If we accept his date of 2800BC, then that would mean PIEs fragmentation roughly coincides with Catacomb culture's expansion/development. Catacomb culture has mtDNA R1 which indicates influence from the Caucasus which, IMO is the more likely homeland of PIE itself. I would've also thought that R1a and R1b may have come from the Caucasus into Catacomb and then spread into Europe at about this time, but that's cutting it a little close since we see R1a in German Corded already in 2600BC. Then again, AFAIK, the R1a found in Corded is some sort of outlier with few modern descendants, so maybe it was wiped out/pushed by a later, more successful EBA R1a/R1b migration. Here's a map of Catacomb with surrounding cultures, as you can see it contacts Abashevo which could possibly be the proto-Uralic homeland:
Personally, I think Basques could potentially be a copper-age migratory people, possibly represented by BR1 or Bell Beaker types. IIRC, Basques have some folk-tale/myth about giants erecting stone structures (dolmens/megaliths?) which could mean either they forgot it was they themselves who did that long ago :P, or it was truly a different people. Also, Spaniards and Basques all seem to have roughly 7-8% ANE and since we know IE languages never had a strong grasp on Iberia until after Latin took hold, it's logical to assume that most of the ANE that exists in Iberia was brought by non-IE peoples.
As for R1b, I think the most likely possibility is that it came from the Steppe via the Caucasus. That said, some R1b may have come from Anatolia because in my preceding posts, I showed that BR2 can be modeled as a 70%-30% mixture of IR1 and V2, used as proxies for Bronze-Age Steppe and SE. Euro populations.

Loki Sigurdsson said...

R1b is not from Asia.

«Oldest ‹Older   601 – 646 of 646   Newer› Newest»