search this blog

Saturday, March 2, 2019

Maykop: a multi-ethnic layer cake?


Let's speculate about the linguistic affinities of the currently available ancient populations from the Caucasus and surrounds. I put together a series of outgroup f3-stats to help things along. They're available for download here.

Maykop
Georgian 0.258224
Abkhasian 0.257899
Latvian 0.257376
Swedish 0.257301
Turkish_Trabzon 0.256996
Basque_Spanish 0.256589
Chechen 0.256514
Icelandic 0.256418
Norwegian 0.256325
Lezgin 0.256272
Irish 0.256227
Tabasaran 0.256092
Italian_Bergamo 0.25605
English_Cornwall 0.256032
Polish_East 0.255991
Scottish 0.255955
Adygei 0.255913

Steppe_Maykop
Latvian 0.261845
Russian_North 0.26145
Estonian 0.260355
Finnish 0.260211
Lithuanian 0.260072
Udmurd 0.259804
Ingrian 0.259663
Surui 0.259637
Vepsa 0.259608
Karelian 0.259532
Karitiana 0.259482
Russian_West 0.259397
Russian_Central 0.259274
Wichi 0.259106
Saami 0.258982
Komi 0.258945
Icelandic 0.258854
Swedish 0.258814
Mordovian 0.258604
Irish 0.25859

Eyeballing the stats might be enough to get a general impression about what they mean, but to understand them properly it's necessary to get technical with something like PAST3 (see here). That's because f3-stats pick up shared genetic drift from all drift paths, and don't especially focus on more recently shared ancestry. This can often lead to confusing outcomes.

Below are a few examples of linear models based on my f3-stats. Note that many Indo-European speakers, especially from Northern Europe, are foremost attracted to ancient samples from the Pontic-Caspian steppe. On the other hand, non-Indo-European speakers, from such far flung locations as the Caucasus and Iberia, show relatively stronger affinity to ancient samples from Anatolia and the Caucasus. Moreover, Uralic speakers show elevated affinity to ancient hunter-gatherer samples from Eastern Europe and Siberia. Makes sense, right?
Based on these and other data, I'd say that Maykop and the culturally related Steppe Maykop were something of a multi-ethnic polity, with many near and far related languages spoken by its people, including perhaps Kartvelian, Northwest Caucasian, Yeniseian and Indo-European. But it seems to me that Proto-Indo-European was spoken by steppe foragers turned pastoralists just outside of the Maykop zone. And I'm quite sure that after the Maykop collapse various early Indo-European groups pushed across the Caucasus and deep into the Near East. Just take a look at the f3-stats and linear model for Hajji_Firuz_BA to see what I mean.

See also...

An exceptional burial indeed, but not that of an Indo-European

The Steppe Maykop enigma

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

98 comments:

Davidski said...

Look for these files in the zip folder if you want to plug something quickly into PAST.

Caucasus_ancient-modern_f3-stats.txt

Caucasus_ancient-modern_f3-stats_West_Eurasia.txt


What other stats can I add to this line up?

Cpk said...

"And I'm quite sure that after the Maykop collapse various early Indo-European groups pushed across the Caucasus and deep into the Near East."

No change in Arslantepe which was very close to Armi.

Davidski said...

Well, there was a change at Hajji Firuz, which is fairly deep in the Near East.

I'll make my mind up about the changes at Arslantepe when I get my hands on the data.

Matt said...

In case anyone would like to look at some of these stats plotted against distances from Global25: https://imgur.com/a/42BjyWe

Grey said...

"it seems to me that Proto-Indo-European was spoken by steppe foragers turned pastoralists just outside of the Maykop zone"

that always seemed the most likely to me - some adjacent population (possibly intrusive) catalyzed steppe foragers into pastoralists leading to the invention of steppe cavalry,

oops.

Andrzejewski said...

@Davidski @Grey so it would be the CHG rich population rather than the EHG rich ones who invented the labguage, right?

Andrzejewski said...


"it seems to me that Proto-Indo-European was spoken by steppe foragers turned pastoralists just outside of the Maykop zone"

DL P said...

Whatever the case, it's clear Maykop played an integral role in the development of the Proto-Indo-Europeans, whether it be directly or indirectly through the introduction of north-Middle-Eastern pastoral animals and possibly even metallurgy.

Them meee said...


I’m oddly pleased by the fact these samples are now called “Hajji_Firuz_BA”.

@Andrzejewski

There was no massive invasion of CHG people of EHG people. It’s unlikely that EHG and CHG were separate populations so late in history, given that ancient steppe foragers were already a mix of EHG and CHG or even EHG with raw basal, ergo not much CHG, by the Mesolithic or Neolithic.

Them meee said...

But this means Maykop survived and became modern Caucasians to a certain extent, right?

Grey said...

Andrzejewski said...
"@Grey so it would be the CHG rich population rather than the EHG rich ones who invented the labguage, right?"

i think the people who won the eventual conflict were probably the ones who invented the language.

my analogy would be if the Sumerians started to colonize America catalyzing the Comanche but the Comanche eventually won would they more likely end up speaking Sumerian influenced by Comanche or Comanche influenced by Sumerian?

it could be either but i think the latter is more likely.

Andrzejewski said...

But we don’t know whether EHG or CHG won out. My bet is on CHG being the aristocracy

Ric Hern said...

@ Andrzejewski

So you think the Women transferred their language...

Grey said...

Andrzejewski said...
"But we don’t know whether EHG or CHG won out. My bet is on CHG being the aristocracy"

i think the horse dudes won in their eventual conflict with the people(s) who catalyzed them and therefore the language was probably the language of the horse dudes at the time (imo) but otherwise yeah - i guess if the horse dudes had an intrusive aristo layer then the "horse dude language" might originally have been the language of that aristo layer.

Andrzejewski said...

No. I think it’s males in a patrilineal society. But it seems like the CHG rich population might’ve been a founder of PIE culture somehow

Andrzejewski said...

5 years ago I wasn’t sure if it was the R1a1 or the R1b people who invented the language

Slumbery said...

@Andrzejewski

"But we don’t know whether EHG or CHG won out. My bet is on CHG being the aristocracy."

It does not seem that way. The YDNA found it elite burials on the Steppe as well as YDNA associated with IE expansion elsewhere so far came out as WHG/EHG. Not only there is no R1(a or b) in CHG (Progress Eneolithic has R1b, but is already EHG mixed), but also typical CHG J lineages are rather lacking in Steppe and early IE context.

So it looks like that the Neolithic and Bronze Age age steppe (and generally eraly Indo-European) elite males are mostly do not have male lineages originating in CHG.

The source of the language is a different story. However I'd rather don't use the verb "invent" in this context. It sounds like you think that PIE was developed as an organized, planned effort and that is definitely not the case.

Andrzejewski said...

@Slumbery “The source of the language is a different story.“

What do you mean by that? That it was female mediated?

Ric Hern said...

If all this mixing happened during the Mesolithic then surely we have to look at cultural practices of Hunter Gatherers....

AWood said...

Should be able to look at the core lexicon and determine if the words for flora and fauna in PIE are native to the steppe. If so, that should be a solid argument in favour of the steppes. Pastoralism could have been taught from adjacent groups such as early Neolithic Iranians, or via trade from these groups. The same thing goes for metallurgy which could have ultimately been borrowed from Maykop. I don't see how these things would refute a steppe origin of PIE. Horses are native to the steppes, not the mountains or the Middle East. Wheels also favour a steppe advantage, but possibly not the origin. Also to consider are the animal bone necklaces and sceptres which I believe to be steppe in origin and possibly linked to PIE hierarchy. I'm going off memory, certainly no archaeology or linguistics expert, but these hints are how you build the hypothesis.

Andrzejewski said...

@Ric Hern If modern Northwest Caucasus people are mostly CHG, the maybe the supposed ancient phylogenetic connections attested to by Nichols and Colarusso (Pontic hypothesis) hold water more than the Indo-Uralic. Especially that Uralic people don’t have the ANE component that EHG and CHG both possess.

AWood said...

Definitely the "CHG" rich, elite burial lineages of Maykop were J2 and the like, but I don't see any argument that they spoke PIE. The steppe "elites" have not been CHG related male lineages.

Andrzejewski said...

@AWood I know that PIE came from the Steppes, just wondering if it was a CHG or an EHG language

Ric Hern said...

@ Andrzejewski

If CHG and ANE/EHG mixed near the Lower Don Kammenaya Balka area during the Mesolithic then we have to look at Hunter Gatherer Practices. For example: Did Hunter Gatherers practice Bride Kidnapping ? Was there a more violent process going on ? Or did they do the bride price thing ?

JuanRivera said...

It's rather moot to debate whether EHG or CHG originated PIE. Whereas the Caucasus shieded northern CHG from Anatolian and Iranian admixture, there was nothing to shield it from ANE and EHG. The presence of ANE outside of EHG in Piedmont shows that it is of Paleolithic origin. And Sidelkino already shows some CHG, plus J is present in one of the karelian EHG samples.

Matt said...

Sticking all the f3 stats through correspondence analysis - https://imgur.com/a/3GhT0hO

Suggests that most of the columns roughly correspond to linear combinations that we would expect - Piedmont as EHG+CHG, Yamnaya_Samara on the same cline displaced towards Sredny_Stog/GAC Poland, Steppe_Maykop_o as between Steppe Maykop and post-Eneolithic Caucasus.

Also though, compared to expectation from that, very, very, very slight excess of Yamnaya+Piedmont+Darkveti_Meshoko relatedness to Northern Europe, Northern South-Central Asia and North Caucasus, which makes sense since they're directly ancestral?

JuanRivera said...

So the steppe was likely a hodgepodge of R1a, R1b, Q1a, I2a, J and C1a2.

Slumbery said...

@Andrzejewski

"What do you mean by that? That it was female mediated?"

No I mean that I do not know what ancient group's language was the main precursor for PIE, but it cannot be decisively determined just from the YDNA of the PIE&IE elites. So even if the PIE/IE YDNA was mainly from WHG/EHG sources, the language's main root still _could_ be an earlier CHG language.

JuanRivera said...

CHG didn't exist in pure form since the Neolithic. For the steppe, pure CHG disappeared since the latest Paleolithic.

Mouthful said...

@Andrzejewski

"Especially that Uralic people don’t have the ANE component that EHG and CHG both possess."


I still don't understand how you read and post in this blog for quite some time, and still get such simple basics so wrong.

Them meee said...

@Andrzejewski

There was no massive invasion of CHG people of EHG people. It’s unlikely that EHG and CHG were separate populations so late in history, given that ancient steppe foragers were already a mix of EHG and CHG or even EHG with raw basal, ergo not much CHG, by the Mesolithic or Neolithic.

Do you understand what I said? EHG and CHG did not mix around 4000 BC to create PIE. Steppe foragers already were a mix of that prior to 5000 or 6000 BC, and there is the possibility that this CHG signal is actually just Basal Eurasian for the most or whole part.

So no, there wasn’t a CHG conquest of EHG. This is now as fantastical as PIE being Iranian or the idea that a later CHG wave lightened up all of Europe and raised the rate of haplogroup H across the continent.

Them meee said...

Also it’s not like Uralics are ANE-deficient. In fact, they were derived from ANE-rich peoples of Siberian and steppe (but very Siberian, Native American or WSHG-like or derived) origin, not from a East Asian or WHG-like population.

When will Andrzejewski understand...

Huck Finn said...

@Them mee and re: "When will Andrzejewski understand..."

I wouldn't hold my breath if I was you.

Open Genomes said...

Global25 restricted nMonte for I4243, the Hajji Firuz Bronze Age woman, dated 2400-2100 BCE:

Bronze Age and earlier ancestry composition of sample: I4243 Population: Hajji_Firuz_BA Bronze Age Iran

32.6% Poltavka-Afanasievo, 23.2% Andronovo-Sintashta (this may be too late for this individual, but if the dating is 2100 BCE, it would be right for Andronovo), 23.8% Darkveti-Meshoki / Maykop / Kura-Araxes, 10.2% Armenia Chalcolithic, 4.0% extra BMAC Iran-like ancestry, and only 1.8% I2327 Hajji Firuz Chalcolithic outlier. Almost no Levantine-related ancestry.

Chalcolithic and earlier ancestry composition of sample: I6561 Population: Ukraine_Eneolithic Chalcolithic Steppe

55.4% Anatolian Neolithic-related ancestry, 10.2% Varna Chalcolithic outlier-related ancestry, 18.2% Progress Eneolithic, 14.8% Khvalynsk Eneolithic, and only 1.4% Levantine Neolithic-related ancestry.

There's no doubt that this individual has substantial Steppe ancestry. The question is, did this ancestry come straight down south through the Caucasus, or around the east side of the Caspian?

The Steppe ancestry seems to be evenly divided between Yamnaya-like Poltavka-Afanasievo ancestry and Proto-Indo-Iranian Andronovo-like ancestry. This seems to point to a slightly later date in the range, around the time of the 4.2 Kiloyear Event.

The Chalcolithic and earlier results seem to indicate that perhaps the Iranian-like BMAC ancestry was mediated via the Progress Eneolithic, and so there was a direct migration southward through the Caucasus, when Srubnaya-Andronovo type ancestry was already present on the Pontic-Caspian Steppe in the former Yamnaya-like Poltavka region.

The presence of Proto-Indo-Iranian Srubnaya-Andronovo-like ancestry, unlike that found in Bronze Age Anatolia, would indicate something more like the later Bronze Age Mitanni than proto-Anatolians.

Andrzejewski said...

@Them meee "Do you understand what I said? EHG and CHG did not mix around 4000 BC to create PIE. Steppe foragers already were a mix of that prior to 5000 or 6000 BC, and there is the possibility that this CHG signal is actually just Basal Eurasian for the most or whole part."

I understand that both EHG and CHG populations making up "Steppe ancestry" were original to the Steppe going back at least until the Epipaleolithic. I was just not sure if PIE or its predecessors originated with CHG or EHG. There is the Indo-Uralic theory that only Bomgaard or Kunstadt support today, and most of the lexomes quoted by Kristiina were loan words from PIE to Uralic trajectory. On the other hand, there is the "laryngeal theory" in which PIE only retained its laryngeals in its archaic Anatolian branch. Joanna Nichols and James Colarusso tout the "Pontic theory" by which NW Caucasian and PIE had a common ancestor 12,000 ybp. In most other branches laryngeals transformed into vowels, like in the word "Father".

Open Genomes said...

I4243, the Hajji Firuz Bronze Age woman, on the Global25 Ward's distance-squared clustering tree clusters with Late Kubano-Tersk and very closely with the I1917 Yamnaya Ukraine outlier:

I4243 Hajji Firuz Bronze Age female on the Global25 Ward's distance-squared clustering tree

Andrzejewski said...

@Them meee "Also it’s not like Uralics are ANE-deficient. In fact, they were derived from ANE-rich peoples of Siberian and steppe (but very Siberian, Native American or WSHG-like or derived) origin, not from a East Asian or WHG-like population.

When will Andrzejewski understand..." @Mouthful

OK, it's easy. First off, @Davidski himself posted a couple of months ago, that modern Uralic speakers have 2/3 of their aDNA as "European" and only a very small trace amount of WSHG.

Second, Uralics are paternally (uniparentally) N1c1 (mainly) and it stems from an NO population in Northern China 5,000 years ago. They do have some "Baikal HG" admixture but although ANE originated near Lake Baikal, it's not exactly "ANE".

Thirdly, Uralic languages do NOT sound or have anything resembling Nivkh, Dene-Yeniseian, Chukto-Katamachdal etc. All these populations are Ancient North Eurasian-rich and have Haplogroup Q. In fact, Botai, Okunevo and the Xungnu/Huns, and not few Scythians and Saka, Sarmatians and Cimmerians have ANE (and East Asian) DNA markers. Not Finno-Ugric.

Also, all the possible cognates quoted by Kriistina as corroborating a putative Indo-Uralic origins turned out to be loan words from PIE or Indo-Iranian into Proto-Uralic, including words like "water" and "name".

Slumbery said...

@Andrzejewski

One of the root causes of your confusion that you apparently do not know what ANE is and can't perceive the different time depth of ancestries referred in the discussions.

For example ..., that modern Uralic speakers have 2/3 of their aDNA as "European" and only a very small trace amount of WSHG."
+"Especially that Uralic people don’t have the ANE component that EHG and CHG both possess."

That trace amount is quite substantial in Siberia, but more importantly, "European" means they have as much or more ANE ancestry as IE speakers, because North Europeans have a lot of EHG ancestry and EHG had a lot of ANE ancestry. In fact, some modern Uralic populations have extra EHG ancestry above their steppe mediated EHG ancestry. However that is also European, since EHG was European. That is, since EHG and CHG itself had ANE ancestry, ANE ancestry van be a part of "European" in the context where Davidsky said that. Different time depths at play...

Also you mix up genetics and language in a way you should not. I know this blog contains a lot of linguistic speculations on genetic base, but you should never forget that language is not genetically coded, especially not coded by YDNA. Seriously, I am R1b-Z2103 and I am a native Uralic speaker. Does this mean that Yamnaya was Uralic?:)

Davidski said...

@Andrzejewski

Uralic people don’t have the ANE component that EHG and CHG both possess.

For your own sake, please stop rambling.

Mouthful said...

@Andrzejewski

Nothing you have written is relevant, for starters try to look up what makes up the bulk of EHG ancestry and how it's modeled. David posted in this blog quite a few times, use the search function.

Davidski said...

@Matt

Sticking all the f3 stats through correspondence analysis - https://imgur.com/a/3GhT0hO

Suggests that most of the columns roughly correspond to linear combinations that we would expect - Piedmont as EHG+CHG, Yamnaya_Samara on the same cline displaced towards Sredny_Stog/GAC Poland, Steppe_Maykop_o as between Steppe Maykop and post-Eneolithic Caucasus.


Thanks, very interesting.

Leron said...

The question of chronology is crucial. Could the Hajji Firuz woman actually be from the Iron Age? Not that she time travelled but she might have been dumped at an older layer. She seems almost like a Western Scythian.

Them meee said...

@Andrzejewski

I didn’t say steppe of the Khvalynsk type, but Siberian-like or ANE-like. So not PIE, I’m not Ryuk.

Thirdly, Uralic languages do NOT sound or have anything resembling Nivkh, Dene-Yeniseian, Chukto-Katamachdal etc. All these populations are Ancient North Eurasian-rich and have Haplogroup Q. In fact, Botai, Okunevo and the Xungnu/Huns, and not few Scythians and Saka, Sarmatians and Cimmerians have ANE (and East Asian) DNA markers. Not Finno-Ugric

Uralic speakers had no ANE and were rather Han-like, Etruscan is a CHG language, PIE is derived from a CHG conquest of EHG...

How the... do you arrive to these conclusions?

Davidski said...

@Leron

Could the Hajji Firuz woman actually be from the Iron Age?

Not likely, considering a C14 dating of 2465-2286 calBCE.

That's the Catacomb period just north of the Caucasus.

Catacomb > Armenia_MLBA

Andrzejewski said...

@Them meee "How the... do you arrive to these conclusions?"

I extrapolate linguistics from genetic data. Etruscans have CHG as well as Mycenaean Greeks and post-4000 BC Anatolians. So the mainstay of ancestry may still be Anatolia_N but the new component of CHG is another layer.

Also, from 4000 BC to 2000 BC even before the Steppe invasion the proportion of mtDNA H of all clades rose sharply from 19% of early farmers to over 40%.

I didn't say anything about a "conquest of EHG by CHG", merely saying that it seems to me that the ruling elite of Khvalynsk and other cultures had more CHG than other samples.

Andrzejewski said...

@Them meee "Uralic speakers had no ANE and were rather Han-like"

https://en.wikipedia.org/wiki/Haplogroup_N-M231

Quote: "Haplogroup NO-M214 – its most recent common ancestor with its sibling, haplogroup O-M175 – is estimated to have existed about 36,800–44,700 years ago.

It is generally considered that N-M231 arose in East Asia approximately 19,400 (±4,800) years ago and re-populated northern Eurasia after the Last Glacial Maximum. Males carrying the marker apparently moved northwards as the climate warmed in the Holocene, migrating in a counter-clockwise path (through modern China and Mongolia), to eventually become concentrated in areas as far away as Fennoscandia and the Baltic.(Rootsi 2006). The apparent dearth of haplogroup N-M231 amongst Native American peoples indicates that it spread after Beringia was submerged (Chiaroni 2009), about 11,000 years ago."

And,
Y-chromosomes belonging to N1b-F2930/M1881/V3743, or N1*-CTS11499/L735/M2291(xN1a-F1206/M2013/S11466), have been found at relatively high levels in South China and adjoining areas of southeastern Asia."

Finally: "Ancient peoples
Most samples from the Liao civilization in northeastern China and northern Korea belonged to y-DNA N. N is found in many ancient samples of northeastern China and Siberia. It is thus suggested that the ancestors of the Uralic-peoples and of the Turkic-Yakut peoples originated in this region about 8000-6000 years ago."

I rest my case...

Davidski said...

@Andrzejewski

What's stopping you from checking facts before you formulate your theories and present them here?

The Khvalynsk sample with by far the most CHG is the one with no grave inventory that appears to have been murdered.

See here...

The Khvalynsk men

Mouthful said...

@Andrzejewski

I'll try to make this as short as possible since I don't want to go down the path into long arguments, because you missed out on a lot of data. Y-DNA markers =/= autosomal admixture, you base your assumption that Uralics don't have ANE admixture which couldn't be further from the truth, based solely on haplogroup N origins.

Them meee said...

@Andrzejewsk

So you still believe there was a massive CHG wave into Central and Western Europe?

Sorry, but that’s just comical.

Andrzejewski said...

@Them meee No. There was a wave of WHG/Villabruna cluster circa 15,000 years ago. Then 8000 years ago a Dzuduzana like Neolithic Farmers from Anatolia swept through (closer to Gravettians) into Europe (somewhat of a "resurgence" of Gravettian-like ancestry). Then about 6,000 years ago a CHG-rich wave of mostly Anatolian Farmers swept into Southern Europe, partially replacing the earlier Anatolian wave. Then 4,900 years ago a Yamnaya group swept the GAC culture, forming Corded Ware, whose western Rhenish outlier became the Bell Beakers.

But Yamnaya was a mix of EHG : CHG + some `15% EEF which was itself rich in WHG.

Them meee said...

@Andrzejewski

Then about 6,000 years ago a CHG-rich wave of mostly Anatolian Farmers swept into Southern Europe, partially replacing the earlier Anatolian wave.

Much of it happened roughly during or after the steppe migrations, at least in the northern Balkans.

But this:

Also, from 4000 BC to 2000 BC even before the Steppe invasion the proportion of mtDNA H of all clades rose sharply from 19% of early farmers to over 40%.

sounds like you still believe there was a CHG wave further north too.

So I don’t know what you believe.

Open Genomes said...

@David BTW, a mistake:
BenzigerodeHeimburg_LN is really Bell Beaker from the Bronze Age. Classifying this as Late Neolithic throws off the Global25 nMontes, which is why it kept showing up for the various Steppe influenced individuals when the comparison samples were restricted to the Chalcolithic or Neolithic!

I'm fixing it in my runs, but you should relabel it as Beaker_BenzigerodeHeimburg or something similar.

Map showing the Beinzegerode Heimberg samples with the source link

Open Genomes said...

@David, given now that Beinzegerode Heimburg is not European Neolithic, it means that these Chalcolithic Caucasus-region samples won't be presumed to have Neolithic ancestry from Europe, but rather may prefer Anatolia. Now I'll try to rerun these again, and see what comes up.

MomOfZoha said...

Checking too late again without much insight though just cannot help but express what struck me:

It is very strange that all those groups are sandwiched BETWEEN Abkhasian and Adygei when comparing to Maykop. Extremely strange... Given that Abkhasian and Adygei are amongst the top closest pop-averages to Maykop in David's PCA25 spreadsheet -- not to mention the fact that they are sister languages/peoples distant from most other languages -- it is like "squishing" all those groups across a weird axis/projection...

And one more thing glancing through the comments:
@Andrzejewski guy (any relation to Pat Benatar? independently discovered her maiden name the other week): Geesh buddy... I don't know much about pop-gen, but even *I* know that Uralics are not exactly lacking in ANE. Come on, dude...

Andrzejewski said...

@MomofZoha Andrzejewski is a very common Polish surname. Means "son of Andrew".

Slumbery said...

@Davidski

I have a maybe weird idea. I cannot entirely reconcile it with archaeology, but it should be OK as one wild speculation.
So the basic idea that Yamnaya is LPIE and Anatolian is a mystery branch that separated earlier. What if it is the other way around? Yamnaya is the predecessor of Anatolian (and then actual Anatolian speakers came from Catacomb)? In this scenario PIE is Ukraine 6000 years ago and the ancestors of the extant IE branches remain in the West until the Abasevo-Sintashta / Srubnaya expansion.
One of the problems of course that Yamnaya at some point took over the region where the PIE homeland should be in this scenario, so this would either force us to think CWC as the LPIE node or to consider less obvious archaeological cultures as important to the spread of IE. It would also pretty much assume that Repin was itself started in its western parts and intrusive in the Volga region or some kind of lingua franca effect.

Andrzejewski said...

@Slumbery "One of the problems of course that Yamnaya at some point took over the region where the PIE homeland should be in this scenario, so this would either force us to think CWC as the LPIE node"

Exactly what I've been thinking all along! All non-Anatolian non-Tocharian branches trace back to Corded Ware, including Indo-Iranic. That includes let's assume substrate words and agricultural vocabulary transfer from the local assimilated EEF farmers.

So what if the reconstructed PIE was nothing but the Corded Ware lingua franca, meaning that Repin or Yamnaya spoke a much older mother/para-language?

Them meee said...

Just out of curiosity, what do the Benzigerode samples look like?

Diego Arroyo de Lagasca Encinas said...

@"Then 4,900 years ago a Yamnaya group swept the GAC culture, forming Corded Ware, whose western Rhenish outlier became the Bell Beakers"


Is this what the Kurganists think?

That the CWC Western Rhenish outliers became BBs?

Davidski said...

@All

For what it's worth...

Kumtepe4 and Kumtepe6 are now in the Global25 datasheets...

https://drive.google.com/open?id=1FSzKKknFGcOgfyA76q9PD7B-n-MJs7L8

epoch said...

@AWood

Anthony has a thing or two to say about reconstructing a homeland based on shared roots of flora and fauna:

https://erenow.net/ancient/the-horse-the-wheel-and-language/5.php

Ric Hern said...

@ Open Genomes

Map very interesting. A U5b2a2 in Poland....So in Germany, Poland and Ireland...

Ric Hern said...

@ Open Genomes

And a U5a1b in Poland in the same area...

ambron said...

Because it was about I6561...

Individual I6561 has identical correlation distribution to contemporary populations as I6531. Both were R1a. One came from the steppe, the other from Poland. They shared a difference of 2,000 years. They both gave autosomaly more to Western Europe. This supports the hypothesis that PIE distributed R1a.

Slumbery said...

Of course the PIE Ukraine (and Yamnaya is already a subbranch) theory also has to be somehow reconciled with the fact that early CWC Baltic is clearly very Yamnaya-like in terms of genome-wide (autosomal) ancestry. In later CWC Baltic this ancestry is much less (Ukraine Eneolothic becomes dominant + some Globular Amphora), but a very real looking direct Yamnaya affinity is still present and significant.

"sample": "CWC_Baltic_early:Average",
"fit": 2.1501,
"Yamnaya_Samara": 70,
"Ukraine_Eneolithic": 29.17,
"Globular_Amphora_Ukraine": 0.83,

"sample": "CWC_Baltic:Average",
"fit": 2.2562,
"Ukraine_Eneolithic": 60,
"Yamnaya_Samara": 29.17,
"Globular_Amphora_Ukraine": 10.83,

"sample": "Sintashta_MLBA:Average",
"fit": 1.372,
"Ukraine_Eneolithic": 48.33,
"Yamnaya_Samara": 36.67,
"Globular_Amphora_Ukraine": 15,

sample": "Srubnaya_MLBA:Average",
"fit": 1.3015,
"Ukraine_Eneolithic": 50.83,
"Yamnaya_Samara": 38.33,
"Globular_Amphora_Ukraine": 10.83,

Ric Hern said...

A migration route of R1b L51 seems to take form. Up the Dnieper, Down the Bug, Down the Vistula and West along the Notec and eventually reaching the Oder....?

Matt said...

@Slumbery: What if it is the other way around? Yamnaya is the predecessor of Anatolian (and then actual Anatolian speakers came from Catacomb)? In this scenario PIE is Ukraine 6000 years ago and the ancestors of the extant IE branches remain in the West until the Abasevo-Sintashta / Srubnaya expansion.

Thoughts, not particularly well organized, off the top of my head on that:

1) Anthony's arguments that lack of reconstructed shared wheel / wheeled vehicle vocab (shared with rest of IE) in Anatolian indicate earlier separation at a pre-wheel stage would seem to become unsound if they are derived from Yamnaya (who are in shorthand described as wagon people), and Yamnaya separate from CWC at a wheel using stage. Hence (in part) argument for Suvorovo-Novodanilovka Complex move to Danube Valley at 4200BCE -> Anatolian.


2) The sequence of Repin offshoot->Afanasievo, leading to Tocharian language, shaky as it is, seems to become more difficult to reconcile? Maintaining a sequence order where split is Anatolian->Tocharian->rest, and Anatolian->rest split is a split off from Yamnaya horizon to CWC. (If you have the split off of CWC earlier, pre-Yamnaya, then it seems like a lot of the elaborate argument for identifying features of Yamnaya culture that were innovations of theirs over previous cultures, guest-host relationship etc and linking these to "rest", surviving PIE today, as evidence supporting steppe hypothesis, that all goes away.)

3) The post-Tocharian branches seem like they have to be more compressed in time. Anthony's suggestion (Horse, Wheel 2007, possibly outdated) for CW I believe was that early branching dialects of IE that would become Germanic split off from the common PIE homeland at 3300 BCE, as the earliest post-Tocharian split, splitting off from Yamnaya horizon, and were the languages that late TRB culture spoke ("If I had to hazard a guess I would say that this was how the Proto- Indo-European dialects that would ultimately form the root of Pre- Germanic first became established in central Europe: they spread up the Dniester from the Usatovo culture through a nested series of patrons and clients, and eventually were spoken in some of the late TRB communities between the Dniester and the Vistula. These late TRB communities later evolved into early Corded Ware communities" Of course this is not the same model as we now know appears right for the CWC). Then Celtic-Italic splits off under a transfer of languages through Danube Valley at 3100 BCE.

So this probably all would have to be compressed under branches off from CW horizon a fair bit later in time.

(Corded Ware ancestor for all surviving PIE today would be later I think so tougher to reconcile with time depths implied by the core lexicon trees, but I think those are already difficult to reconcile with the Yamnaya hypothesis ("All surviving IE languages derived from offshoots of Yamnaya horizon") anyway because the branching order under their tree phylogeny doesn't really fit well with the timeline of migrations established under that hypothesis, as articulated by Anthony, and in fact are virtually the opposite. So that's not really particularly anything new.).

Slumbery said...

@Matt

That you for the detailed explanation.

Still, on the genetic side, in the time depth of 6000 years BP the dominant ancestry of CWC, Sintashta and Srubnaya seems to be Ukraine Eneolithic as opposed to anything east/south-east of the Don. De only group that score less than 50% is Yamnaya proper and its direct descendants. Given that Yamnaya YDNA (at least what is actually found by now) is not actually ancestral to any later groups dominant YDNA (although closely related to the Beaker R1b), while YDNA ancestral to CWC, Sintashta, Srubnaya was found in Ukraine Eneolithic context, Yamanya has some weaknesses and questions to answer as the main LPIE node. At least from the genetic side.

It also follows that if we accept Suvorovo-Novodanilovka as Anatolian because of Archeological and linguistic arguments, then we might very well just crossed out Yamnaya entirely from the story. Well, except Tocharians, because if Anatolian is earlier Balkan and everybody else is coming from Ukraine (give or take), then Tocharian could be a Catacomb descendant group from the rather messy Andronovo horizon.

I cannot comment on the cultural-linguistic argument for Yamnaya, because I am not familiar with it. Could it be terminology picked from Yamnaya by contact or during the westward expansion of Yamnaya?

Grey said...

"Also, from 4000 BC to 2000 BC even before the Steppe invasion the proportion of mtDNA H of all clades rose sharply from 19% of early farmers to over 40%."

wild speculation but mtdna is involved in energy production in the cell

https://en.wikipedia.org/wiki/Mitochondrial_DNA

so i wonder if different types of mtdna can have selective advantages in certain conditions e.g. diet?

DKj said...

Did Northern and Eastern Europeans inherit ABCC11 gene from the steppes admixture? or did hunter gatherers carry it as well? It peaks in Asians but also seems to be present in N & E Europe.

ABCC11 gene frequency:

Deutsche 16%
Westeuropäer 17%
Osteuropäer 20%
Chinesen 94%
Koreaner 100%

I'm guessing it has something to do with Steppe/ANE in case of Europe?

JuanRivera said...

It's also present in western and southern europeans, even in Iberia. It clearly has to do with steppe.

JuanRivera said...

Did models of both Kum6 and Kum4: Kumtepe_LN_6_low_res: (Tepecik_Ciflik_N+Seh_Gabi_ChL, 4.7981, Tepecik_Ciflik_N 99.17% Seh_Gabi_ChL 0.83%) (Tepecik_Ciflik_N+Seh_Gabi_ChL+Barcin_N, 3.5753, Tepecik_Ciflik_N 17.5% Seh_Gabi_ChL 15.83% Barcin_N 66.67%) (Tepecik_Ciflik_N+Maykop_Novosvobodnaya+Barcin_N, 3.5266, Tepecik_Ciflik_N 21.67% Maykop_Novosvobodnaya 15% Barcin_N 63.33%) (Tepecik_Ciflik_N+Maykop_Novosvobodnaya+Barcin_N+Peloponnense_N, 3.4995, Tepecik_Ciflik_N 15.83% Maykop_Novosvobodnaya 13.33% Barcin_N 40% Peloponnense_N 30.83%)

Kumtepe_LN_4_low_res: (Kumtepe_LN_6_low_res, 17.0489) (Kumtepe_LN_6_low_res+Khvalynsk_Eneolithic, 14.2498, Kumtepe_LN_6_low_res 75% Khvalynsk_Eneolithic 25%) (Kumtepe_LN_6_low_res+Ukraine_Eneolithic, 14.6833, Kumtepe_LN_6_low_res 61.67% Ukraine_Eneolithic 38.33%) (Kumtepe_LN_6_low_res+Progress_Eneolithic, 13.9845, Kumtepe_LN_6_low_res 69.17% Progress_Eneolithic 30.83%) (Kumtepe_LN_6_low_res+Vonyuchka_Eneolithic, 13.9934, Kumtepe_LN_6_low_res 68.33% Vonyuchka_Eneolithic 31.67%) (Kumtepe_LN_6_low_res+Varna_o, 15.1443, Kumtepe_LN_6_low_res 55.83% Varna_o 44.17%) (Kumtepe_LN_6_low_res+Armenia_ChL, 14.0856, Kumtepe_LN_6_low_res 25.83% Armenia_ChL 74.17%) (Kumtepe_LN_6_low_res+Armenia_ChL+Progress_Eneolithic, 13.9304, Kumtepe_LN_6_low_res 34.67% Armenia_ChL 54.17% Progress_Eneolithic 9.17%) (Kumtepe_LN_6_low_res+Armenia_ChL+Varna_o, 13.9342, Kumtepe_LN_6_low_res 22.5% Armenia_ChL 61.67% Varna_o 15.83%) (Kumtepe_LN_6_low_res+Armenia_ChL+Progress_Eneolithic+Varna_o, 13.8931, Kumtepe_LN_6_low_res 30.83% Armenia_ChL 50% Progress_Eneolithic 6.67% Varna_o 12.5%). Curiously, the best model I did of Kum4 was one of Kum6+Armenia_ChL+Peloponnense_N+Progress_Eneolithic, which isn't shown. Kum4 clearly shows steppe ancestry.

Davidski said...

Kum4 does have steppe ancestry. A lot of it too.

This sample is from western Anatolia, with a dating of 5,500–4,800 BP.

Pity that it's such a shitty sequence, because otherwise it would nail the steppe PIE hypothesis.

Them meee said...

@Davidski

Exactly.

And we need more high-quality samples from the region, because otherwise the Max Planck Institute and other wanton supporters of the Out-of-Iran or Armenia hypothesis will continue with their hijinks.

Samuel Andrews said...

Modern Greek Cretans & Aegeans (from near coast of Turkey) are probably mostly Anatolian not Greek. Despite this, they have about 15% Steppe ancestry. There are only two possible explanations...
-Greek/Balkan ancestry had lots of Steppe ancestry (30-40%)
-Both their Greek/Balkan & Anatolian ancestors had Steppe admixture.

The second possibility is more likely. This, means it is likely there were Anatolians with Steppe admixture.

JuanRivera said...

I'm eagerly waiting for the samples of the Dzudzuana and Yana papers.

Them meee said...

I’m waiting for Fatyanovo samples, maybe they’re the ancestors of the Indo-Iranians?

JuanRivera said...

That too.

Them meee said...

I also wish we had proto-Corded Ware samples, either from post-Stog groups or the immediate predecessors of CWC (Dnieper zone?)

Kristiina said...

At the moment, the yfull tree and the ancient evidence point to K2 having arisen in South Central Asia with O thriving in China, N in Western/Central Siberia, Q in Northeast Asia and R in West Asia and Caspian steppe.

Ric Hern said...

P1 in Northeast Siberia and R in Southern Siberia...How did you get to R originating in West Asia ?

Kristiina said...


@ Ric Hern I did not say what you claim me to have said. I said R THRIVED in West Asia and Caspian steppe.

It is stupid to think that during a period of 30 000 years a haplogroup would stay put in one tiny area. People were moving around and haplogroups arose and went extinct.

epoch said...

@David

Is Kum4 good enough to eke out some significant D-stats? Such as this?

Mbuti EHG Kumtepe_LN4 Kumtepe_LN6

JuanRivera said...

With such low quality genomes, Kum4 best models as Kum6+Steppe+Something else. If we had higher quality genomes, we could see the spread of steppe ancestry in Anatolia.

Matt said...

Davidski: What other stats can I add to this line up?

Considering this question, I had a look at some f3 stats you'd run from me before. These were f3(Mbuti,X;Ancient), rather than the f3(Mbuti;Ancient,X) which I think you've run here.

They give a closely correlated measure that is more like a distance: https://imgur.com/a/GLJzbnQ (whereas the other more like a similarity index)

Anyways, looking at those, besides the obvious Levant_N, Levant_CA and Levant_BA, the only ones in the set that I had which seemed like they could give some more resolution on differences between modern are Iberia_CHL (which behaves like GAC_Poland but splits Spain away) and the Baltic_BA (which behaves like Sredny_StogII but splits NE Europe away): https://imgur.com/a/fz1Fh5k. Dutch Beaker might also be useful for NW Europe. Other than that, really not very much turned up in regard to splitting Western Eurasian populations that's not already represented here.

Davidski said...

@epoch

Nope.

Ric Hern said...

@ Kristiina

Okay, Sorry missed the THRIVED part. However I did not say that nobody migrated since 30 000 years ago.

But I also can not see how lineages who evolved in an extremely cold environment just suddenly prefer a Desert environment(Evolution of Bactrian Camels, Onager, Gerbils and Karakul Sheep etc.) to settle in and prosper without acquiring some advantages over the local inhabitants of that Desert Regions.

For me personally the Caucasus Piedmont area does not fall within the definition of what West Asia is mostly all about.

Just my personal opinion.

As for K2. The connection of Vindija Neanderthals to Modern Eurasian populations and finding K2 close descendants already way up North both in Europe and Western Siberia maybe rather points to an origin in West Asia and Europe (Caucasus, Anatolia, Balkans) rather than South Central Asia ?

Davidski said...

Based on ancient DNA, Kristiina's proposal that Y-hg R "thrived" in West Asia before the Bronze Age looks like a fantasy.

I mean, let's use some common sense and logic here.

Ric Hern said...

@ Davidski

Yes, and that is why I said "...settle in and prosper without acquiring some advantages over the local inhabitants of that Desert Regions."

@ Kristiina

What David said. ;)

Kristiina said...

@ David

You omit yDNA R2. We do have pre-Bronze Age samples of R2 from west Asia and one sample that was defined R1.

These are Iran Neolithic samples from the Harward paper
Ganj Dareh NE Zagros I1945 R2a
Ganj Dareh NE Zagros I1949 R1
Iran NE Ganj Dareh Zagros I1947 GD22 R2a
Iran NE Ganj Dareh Zagros I1954 GD41 R2a
Iran NE Ganj Dareh Zagros I1946 GD20 R2a
Iran NE Ganj Dareh Zagros I1952 GD40 R2a

The samples are dated c. 8000-7500 BCE!

Davidski said...

@Kristiina

That sample was wrongly defined as R1. It's actually R2, just like the other one.

So all you have is a low frequency of R2 in what is now Iran.

Davidski said...

@Matt

https://drive.google.com/file/d/1g-oG1GaYJZ7oeEuNyWdAC4xl5DXwnI4p/view?usp=sharing

Matt said...

Thanks, look much like as expected: https://imgur.com/a/jlCHDaG

Matt said...

@Davidski, this is probably asking a bit too much but any possibility of doing Indus_Periphery, Swat_IA, Ust_Ida_LN, Iron_Gates_HG, WHG for (Eurasian) completeness sake?

Davidski said...

@Matt

No Ust_Ida_LN, but I put in Hovsgol_BA and Tianyuan.

https://drive.google.com/file/d/1g-oG1GaYJZ7oeEuNyWdAC4xl5DXwnI4p/view?usp=sharing

Matt said...

Cheers.

Minor note: Something may be up with row values for Swedish, Russian_North, Russian_Central and Russian_West? They seem oddly high and separated from close relatives when I put those through PCA?

Bunch of random graphics using these stats: https://imgur.com/a/gTPaXCa

Davidski said...

@Matt

This might be an artifact of only using two people per population (each of those populations are based on just two people).

I'll try and bolster those sets with more people and hopefully the problem disappears.

Davidski said...

@Matt

Check out the latest datasheets. At least some of those artifacts should now be gone, but if they persist, then I can't do much about it, and you just have to remove the outlier samples from the relevant analyses.

https://drive.google.com/file/d/1g-oG1GaYJZ7oeEuNyWdAC4xl5DXwnI4p/view?usp=sharing