search this blog

Saturday, December 15, 2018

Some German guy once said...


If you repeat a lie often enough, people will believe it, and you will even come to believe it yourself.

On a totally unrelated note, the Max-Planck-Institut für Menschheitsgeschichte (aka MPI-SHH) is apparently still claiming that its southern Proto-Indo-European (PIE) homeland theory has been corroborated by archaeogenetic data. For instance, check out the Youtube clip here.

Below is a screen cap from the clip showing a map that summarizes what the folks at the MPI-SHH are thinking in regards to the PIE question and the early spread of Indo-European languages.


Unfortunately, this map doesn't make any sense. Why? Here it is, in point form, as simply as I can put it:

1) There's no evidence in any archaeogenetic data of migrations during the Neolithic from what is now Armenia and surrounds to Western Europe, the Pontic-Caspian steppe, or, indeed, South Asia, that may have brought Indo-European languages to these regions. In fact, the currently available ancient DNA data outright contradict this scenario, because:

A) the Corded Ware and Yamnaya archeological cultures, which are generally considered to have been the main vectors for the spread of Indo-European languages from the Pontic-Caspian steppe into Northern and Central Europe, weren't founded by migrants from south of the Caucasus (see here)

B) the Neolithic farmer populations that migrated deep into Europe and eventually colonized the western third of the continent were especially poor in Caucasus-related ancestry, and, realistically, could only have come from well to the west of the Caucasus

C) conversely, the Neolithic farmer populations that moved deep into South Asia are inferred to have been especially poor in Anatolian-related ancestry, and, realistically, could only have come from well to the east of the Caucasus (see here)

D) Caucasus-related ancestry, of basically the same type that is being associated by the MPI-SHH with the PIE expansion, did move into Western Europe across the Mediterranean, but this happened during the Bronze Age and it impacted the island of Sardinia, which is generally regarded to have been inhabited by non-Indo-European speakers until the Romans got there (see here). Oops.

2) There's now overwhelming evidence both in ancient and modern DNA data that Eastern Europeans and Indians, especially Indo-European-speaking Indians, share significant ancestry, in particular paternal ancestry, from essentially the same Bronze Age populations living on the Pontic-Caspian steppe (not south of it!), and this is the only obvious, important genetic link between these two linguistically closely related but geographically far flung groups within the last...tens of thousands of years?

3) Ancient samples from Mycenaean, and thus Indo-European-speaking, Greece and parts of Iron Age Iberia where Indo-European languages were attested at the time also show steppe-derived ancestry, and, in fact, of a very similar character to that shared by Eastern Europeans and Indo-European-speaking Indians (see here and here, respectively).

4) However, Pre-Mycenaean and likely non-Indo-European Minoan samples, also from the Aegean region, don't show any steppe ancestry, but they do show Caucasus-related ancestry, of basically the same type that is being associated by the MPI-SHH with the PIE expansion. Oops again.

Thus, at the very least, these undeniable and, surely, easy to grasp facts that I've just set out should give pause to anyone who still claims that the Near East, rather than the Pontic-Caspian steppe, was the main staging point for the expansions of the early Indo-Europeans. Indeed, methinks it's now time to admit by all those concerned that the most likely homeland of all surviving branches of the Indo-European language family, and thus of late PIE, was the Pontic-Caspian steppe.

Honestly, I'm shocked, and even disturbed, that none of this seems to have filtered down to the linguists at the MPI-SHH, especially since the MPI-SHH is also heavily populated by scientists who apparently know a thing or two about archaeogenetics.

Now, it's true that archaeogenetic data are yet to reveal an unambiguous signal of steppe ancestry in samples from Hittite era Anatolia (five have been published to date), which may perhaps suggest that the people who brought Hittite and the other Anatolian languages to Anatolia didn't come from the steppe. Of course, Anatolian languages represent the earliest, most basal split in the Indo-European phylogeny, and thus aren't part of the late PIE node. So if the Indo-European-speaking ancestors of the Hittites didn't come from the steppe, then it stands to reason that early PIE didn't either.

But this isn't relevant to my criticism of the MPI-SHH, because even if early PIE didn't come from the steppe, then like I said, there's very solid evidence now that late PIE did, and the problem is that the linguists and geneticists at the MPI-SHH appear to be missing this point, or they're unwilling to accept it.

Moreover, please note that I'm not arguing that the linguists at the MPI-SHH are getting things wrong when it comes to actual linguistics. For all I know, their approach in this area might well be perfect, and perhaps it has indeed revealed insights that have been missed by others using more traditional methods?

For instance, it's possible that the phylogeny of Indo-European languages as shown in the video linked to above reflects the truth better than anything else offered to date. I don't know, so I'm keeping an open mind about that. But admittedly, I'm skeptical, considering how lousy the MPI-SHH's interpretation of the archaeogenetic data has been to date in this context, even at the most basic level.

See also...

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

392 comments:

1 – 200 of 392   Newer›   Newest»
An Idiot said...

Even though I believe in a West Asian homeland of sorts, it's just laughable to think that it spread to Europe and India from the Middle East and not the Steppe. Unless it was a VERY silly mistake, there has to be some other questionable motives involved here - they probably just don't like the idea of blonde Steppe colonists.

Davidski said...

I seriously doubt that the steppe-derived groups that moved into South Asia were very blond.

Palacista said...

That map seems to imply the primary split is betweeen Balto Slavic and Indo Iranian.

Davidski said...

Yeah, it looks like total BS.

M. Myllylä said...

In general scientists in population genetics are prone to repeat what older branches of science and the big readership wants to see. Not all, but too many of them. Science should be a trailblazer instead of being a fearful mousekin. Does this reveal something about the scientific community? IDK.

Al Bundy said...

Ok it spread from what is now Armenia in 3 directions.I don't get the map either, it spread to the PC steppe and then stopped?

Davidski said...

I suppose it represents the peopling of Eastern Europe by the ancestors Balto-Slavs, since the map is focusing on the spread of the main, early IE nodes, rather than everything.

Al Bundy said...

What they actually say in the video, Gray dubbed in German, is very standard.

Davidski said...

Doesn't matter, the map is from the Twilight Zone and Heggarty was in the New Scientist and Guardian recently basically saying that there were migrations from Iran that gave rise to Yamnaya.

Al Bundy said...

It makes them look bad because I think an ultimate homeland south of the Caucasus is a possibility pending more Hittite, or actual Hittite, samples.

Davidski said...

That's not the only thing that makes them look bad.

Al Bundy said...

Especially considering the fact that Haak is at Max Planck and that 2015 paper made a strong, and pretty much accepted case, that steppe migrations brought IE to N Europe.

Al Bundy said...

Another issue is that timeframe 8000 ybp.Did I read that right? Yea I guess that's not the only thing.

Cpk said...

I'm sure that Late PIE is on the steppes. But although we don't have a lot of ancient Hittite genomes we have a lot of skulls and they don't show any change until 2000 BC. Plus Anatolian names from 2500 BC.
So IMO early Pie is either (more likely) West Asian or it came down from steppes through Caucasus very early.

An Idiot said...

I wonder why they claimed that Western IE spread to Europe with farming? That's what puzzles me the most, as it clearly isn't true at all.

Presumably this whole thing is due to them finding R1b L23 in the Caucasus at a surprisingly early date, and perhaps finding L51 in Western Europe at a surprisingly early date, as well as maybe Z2103 in the BMAC. That, or they really don't know what they're talking about.

@Davidski If Corded Ware is anything to go by it's likely the Steppe incursion into India was by people who were blonde - that doesn't mean they all had "yellow" Scandinavian hair though.

Also, unlike with CW, we have remains from the Eastern IE expansion with blondism, and of course you can always just google "blonde Kalash"

An Idiot said...

As for the possibility of R1a originating in Armenia and spreading out as that video described - just, no.

Slumbery said...

It looks like there were "steppe-CHG" populations very early and CHG in later steppe comes from them as Davidsky suggested earlier. In this case PIE could have evolved mainly from their language even if whatever demographic/cultural processes made EHG/WHG paternal lineages dominant at some point.
At one hand this would still not mean a particularly close linguistic connection between the Steppe and the Transcaucasia or Iran, because the split between steppe-CHG and the populations of those regions is very early. At the other hand this could explain the existence of a high CHG but low EHG early IE group that does not register as a genetically steppe input in Anatolia.

Of course I am just wildly speculating.

EastPole said...

Link to Youtube clip is not right.

“For instance, it's possible that the phylogenetic relationships of Indo-European languages as shown in the screen cap below (from the video linked to above) reflects the truth better than anything else offered to date.”

I am very skeptical about their method. Glottochronology is pseudoscience. Reconstructed PIE is not a real language and has never been spoken by anybody etc.

But assuming there is some truth in their method, i.e. it works although we don’t know why, and the split between Indo-Iranian languages (red arrow) and Balto-Slavic languages (blue arrow) occurred before 4000 BC:

https://i.postimg.cc/NjZSfPQb/screenshot-462.png

then Sredny-Stog seems to be the archeological culture from which these two language families emerged, not Armenia. We have R1a from Dereivka culture.

Davidski said...

OK, fixed the link to the clip.

Matt said...

Voiced my specific doubts about their ideas in the other thread where this was discussed (comments of "Europe's ancient proto-cities may have been ravaged by the plague"), so won't repeat them, and there's nothing in Davidski's post that strikes me as unreasonable or obviously wrong, so nothing for me to comment there.

Only thing I'd add is that if the consensus Steppe theory is correct, the way to square with Jena work on cognates in basic lexicon seems to imply that contact / substrate effects were pretty large, reconfiguring the basic lexicon of expanding IE to a huge degree. E.g. if Avestan and Vedic_Sanskrit actually diverging for <1000 years when first attested(?), but look as if diverging for >3000 years in basic lexicon*, and that's due to contact effects, bit of a challenge to even thinking of them as descending in a tree like way from proto-Indo-European (and really to proto-Indo European reconstruction as well).

*Heggarty uses the example that Avestan and Vedic_Sanskrit only seem to share about 57.5% cognacy of basic lexicon (I guess from their 200 word lists).

Example of a tree from what I think is a recent-ish presentation by Heggarty linked in the above mentioned comments thread - https://i.imgur.com/bOtbgQE.jpg (each grindline 1000 years, thinner branches indicate slower lexicon evolution and more solid branches more certainty over the tree position of inferred ancestor, I think). Version of that tree from Gray's documentary seems to allow more sharing and admixture between IE branches - https://i.imgur.com/hocKRoA.jpg (e.g. transfer between Western European branches) and timescales for divergence are compressed.

Ebizur said...

Al Bundy wrote,

"Another issue is that timeframe 8000 ybp.Did I read that right? Yea I guess that's not the only thing."

I'm afraid you have not read it right. "8.000 v. Chr." (i.e. 8.000 vor Christus) means "8,000 B.C.," not "8,000 ybp."

Bob Floy said...

This is turning into great comedy, especially since Raveane et al. didn't get the memo. I wonder if they'll get scolded?

Them meee said...

That map shows steppe PIE as a dead end. What are they even thinking? Are they ignoring the EHG/steppe signal in Europe or what?

Bob Floy said...

I don't see how anyone can deny what's going on here now, honestly. This might come to a head soon.

Dragos said...

Wait - are they offering this as their theory, or just one possibile theorem which exists. If it's the latter, then much ado about nothing.
If it's the other, then is it based on aDNA find yet to be published, or just their own left -field theory ?

Davidski said...

@Dragos

Yeah, they've got all sorts of other theories, but they haven't had a chance to roll them out yet, because they've been flogging their Armenian "hybrid" theory for the last couple of years.

Dead cat bounce

And yeah, they probably have unpublished ancient data showing massive migrations from Armenia to Western and Eastern Europe and contradicting all of the published DNA to date, but they're keeping it all under wraps because it's fun and crazy.

Them meee said...

@Bob Floy

Turning?

http://eurogenes.blogspot.com/2015/06/paul-heggarty-desperate-or-clueless.html?m=1

Vara said...

This doesn't make any sense. Looking at the "dead cat bounce" model it has PIE around 8kya and the 3 wave migrations happening at different stages. The 8000 BCE explosive migration is a bit too much, IMO.

This is most likely an extremely simplified model. I'd take this with a pinch of salt.

Grey said...

reading about this stuff as a kid years ago the recurring theme was everything started in the fertile crescent and spread out from there so i guess it's difficult to let go of that idea.

i think it's probably still half-true i.e. people from the fertile crescent spread along trade routes looking for valuable stuff and in some way acted as a catalyst for the steppe horse culture (explaining the language connections) but my default assumption would be if the Comanche had won they'd probably be speaking Comanche with some influence from English, not English with some influence from Comanche so unless there was compelling evidence to the contrary i'd have thought that would be the default assumption about PIE as well (unless proved otherwise).

andrew said...

@Matt "seems to imply that contact / substrate effects were pretty large,"

I think that this is precisely what is going on. And, there is research out there (to recount in detail for another day), to suggest that the relative importance of contact/substrate effects to linguistic drift is much more strongly in favor of the former and much more weakly in favor of the latter than is commonly recognized.

I think that this is also what is going on with the seemingly great time depth of the Anatolian languages. All of the European branches of the Indo-European languages share a substrate from a common Neolithic Anatolian farmer substrate, while the Anatolia itself, by the time that Anatolian language family languages are first attested there had Caucasian-ish substrate language. So, some of what is believed to be Proto-Indo-European sourced lexical and other influences are probably actually shared influences from a common substrate that is not shared in the Anatolian languages because they did not share this substrate.

The archaeological data and the earliest contemporaneous historical accounts point to a much later arrival of Indo-European languages in Anatolia, probably only two or three centuries at most before 2000 BCE, not that far apart in time from IE expansions into the Balkans, Greece, South Asia and the Tarim Basin.

But, if contact and substrate influence are more important than random drift in the rate of which languages evolve, then many seeming contradictions in reconciling archaeo-linguistics to other evidence is neatly and harmoniously resolved.

Leron said...

They are noticing southern genetic contributions into a population that would later expand with IE languages in the north. Their mistake is correlating genetic evidence directly to languages. Unlike genes, languages are a lot more fluid. Either way, it's still data and worth considering even though you can't accept their conclusions.

Al Bundy said...

@Ebizur Thanks.They must view things differently than the Reich lab, who don't seem to dispute a steppe migration to N.Europe.

Bob Floy said...

@An Idiot

"Unless it was a VERY silly mistake, there has to be some other questionable motives involved here"

It's the same motives that we were discussing in that other thread, this is more evidence of that. I'd love to hear someone suggesting that there's not some kind of agenda behind this now.

Al Bundy said...

Some of the people at Max Planck still support the Anatolian theory so this might be a way to salvage part of it, but they're doing it pretty poorly.

Bob Floy said...

@Al Bundy

Any attempt to salvage that theory is stupid and, if you're a professional scientist, unethical, since the data plainly says something very different. The only way anyone is propping up the Anatolian theory is by being dumb or dishonest, and I'm speaking as a former supporter of that theory. These guys are a disgrace.

Matt said...

@andrew, yes, I don't know about the Hittite case but in general I think it seems like something that is possible.

I actually think that the Jena group do take it seriously and seem to be trying to try and account for it.

For example - https://www.findaphd.com/phds/project/contact-and-change-in-the-diversification-of-the-indo-iranic-languages/?p104664 (seems to be trying to get a handle on contact driven change) and their modelling on contact and change in Austronesian - https://phys.org/news/2017-10-myth-language-history-languages.html (where grammar interestingly seemed more affected by contact than lexicon, for the given lexical and grammar features they were looking at).

Uralic and Turkic also offer some potential considerations. If there are rapid accelerations in change in basic lexical change generally due to substrates and contact.

It should be very obvious in Turkic particularly, and they should look a lot older and more diverse under these sorts of methods than the more or less historical attestations of dates that we have (pretty much certain about 2,500 YBP) should suggest.

Latin American Spanish might also be another example to look at, in an even more recent timescale.

It's intuitively a likely / plausible thing, seems like potentially tricky thing to put on a firm footing where you're not moving back to the question of relying on expert impressions and intuitions about what is possible and likely and which features are affected and which are not and so on, where substrate effects can be recruited in support of an argument or not in a way that the person making an argument has any choice about.

Al Bundy said...

@Floy I agree

Al Bundy said...

They're trying to prop up a part of it which as you say is stupid, but a PIE homeland Armenia Iran is possible.

Dragos said...

@ Andrew
that's an interesting perspective, although very much a minority view these days.

With regard to ''all of the European branches of the Indo-European languages share a substrate from a common Neolithic Anatolian farmer substrate, while the Anatolia itself, by the time that Anatolian language family languages are first attested there had Caucasian-ish substrat''

Should not this make Indo-Iranian, and esp. Indic, the most diverged (because they encountered Iran _N & AASI) ?

Al Bundy said...

Actually scratch that, a PIE homeland in Armenia is not possible.

Peter Klevius said...

Perhaps MPI confused gene flow with money flow?

Bob Floy said...

@Al

I'll be generous and say that the Armenian theory is not a very good one.

@Peter Klevius

I think you may have hit the nail on the head there.

Al Bundy said...

Not a linguist so if what Andrew says is correct the steppe theory could still account for Anatolian , but I guess the Kroonen crowd favors a much earlier arrival.I'd also like to see more elite Myceneans, hopefully they're in the pipeline.

mickeydodds1 said...

It's all political, of course.

Because of fairly obvious reasons relating to events in Germany from 1933 to 1945.

epoch said...

If someone manages to crack Linear A we don't even need "more samples". If Linear A is non-IE than we basically know that Anatolian languages were derived from the steppe.

Also, Hittites cremated their dead. If they derived from early Balkans Bronze Age, who had a small (~25%) steppe amount and the spread in Anatolia was elite spread, than it would mean that the elites already were highly diluted. And they were cremated.

epoch said...

I keep reading that Hittites had many Hattic loanwords for local flora. Another interesting thing is that Hittites seemed to have had a special liking for saffron, possibly related to spring festivities. The Hittite word for saffron, Azupiru, is derived from the same Semitic root from which the word "saffron" is derived.

But saffron is indigenous in N.W. Iran as well. Why, if the plant was so important, wouldn't they use a native IE word?

https://academic.oup.com/jxb/article/60/1/6/573757

Dragos said...

Lads you need to remember there was more than Hittites. Virtually, the entire central & western Anatolia had IE speakers. So there essence will be discoverable, never fear. And Hittites did not exclusively cremate.

epoch said...

@Dragos

Yes, the Luwians. They're attested from 2000 BC onwards, its language has a Hurrian substrate. Cemeteries roughly like Hittite cemeteries. Anatolia knew different burial practices: Cremations, jar burials and other burials. There are even mixed cemeteries with jar-burials, cist burials and cremations.

The point is there is a continuous jar burial tradition from before Hittites and Luwians, and since we know that at least Hittites were intrusive, which is even documented by Assyrians, we can safely assume that burials in jars are non-IE Anatolians.

Simon_W said...

Just on a side-note, apparently Goebbels didn't say that thing about lying, it's an urban legend. I was just googling to find out which German guy is supposed to have said that, because I didn't know, and stumbled upon this blog:

http://falschzitate.blogspot.com/2017/12/eine-luge-muss-nur-oft-genug-wiederholt.html

Them meee said...

It’s funny because back in the day it was believed the Proto-Indo-European homeland was Scandinavia, originating in Funnelbeaker and Corded Ware being derived from Funnelbeaker rather the steppe.

And yeah, let’s not forget other Anatolians.

Matt said...

@Dragos: "Should not this make Indo-Iranian, and esp. Indic, the most diverged (because they encountered Iran _N & AASI) ?"

Not really sure whether that implication is correct (I would have had the same intuition but andrew may have an idea around that), though I think under this theory, there should also be some shared features between Indo-Aryan and European branches which were acquired in the "Neolithic European substrate" because of a pulse back from Central Europe to the steppe (may offset some increasing difference from a more different substrate?).

JuanRivera said...

PIE has a lot of vocabulary which disqualifies an early origin. Among those are metals (copper/bronze, gold, silver), milk products (milk, cheese, butter, wax), plants (oats, whose actual domestication wasn't in the neolithic), animals (horse, who weren't also domesticated until the late neolithic at earliest), equipment (wheels) and defense (forts). A steppe origin is supported by the animal product vocabulary, horses (who have at least two terms), tortoises (whose members mostly live in semiarid environments), sand, beaver, birch, disproportionate amount of terms for domestic animals (also observed for other pastoralists), and as bonus, water-related vocabulary (such as river, lake, water, wave, fishes, sea, ship, etc), caused by the climate forcing people to live near bodies of water, where they would develop fishing and navigation skills. The Pontic-Caspian steppe happens to have plenty of bodies of water, especially the Dnieper, Don, Volga and Ural rivers. Another point in favor is that every recorded IE languages, including those in Anatolia, Armenia and Iran, have a substrate, the unsubstrated steppe IE languages going extinct without trace.

JuanRivera said...

As for milk products, a more accurate way (counting wax, and wool, that I forgot) would be animal products.

JuanRivera said...

And also I forgot shell, who belongs in the water-related vocabulary.

Simon_W said...

@epoch

"If someone manages to crack Linear A we don't even need "more samples". If Linear A is non-IE than we basically know that Anatolian languages were derived from the steppe."

Linear A is basically "cracked". The difficulty that remains is to understand the language. Which implies that it isn't an IE language, otherwise it wouldn't be so hard to understand.
But your argument is deeply flawed. I mean, there are various, unrelated language families with heavy CHG ancestry now, why should this have been different in the Bronze Age?

Simon_W said...

@Cpk

"But although we don't have a lot of ancient Hittite genomes we have a lot of skulls and they don't show any change until 2000 BC. Plus Anatolian names from 2500 BC.
So IMO early Pie is either (more likely) West Asian or it came down from steppes through Caucasus very early."

Apparently you didn't quite get the point of my reply in the other thread. That date you keep citing, 2000 BC, is wrong. The change in cranial and facial shape that you have in mind, dates to the MLBA. So according to modern dating, I think, 2500 BC or something. 2000 BC is old School dating (likely predates c14 calibration).
Also I have to specify that it's not quite right either saying that there was absolutely "no" change in Anatolia prior to that MLBA/Hittite change. The devil may be in the details. For instance the Early European farmers who had marched out of western Anatolia had quite a bit shorter upper faces and broader noses than those central Anatolians of the EBA, while they were similar in the cranial shape. So the genomic shift that affected central Anatolia during the Neolithic and up to the EBA, which brought along stronger CHG affinity, or in other words an eastern shift, was carried out by longheaded leptomorphic people of a Mediterranoid type, maybe akin to what modern Azeris tend to look like. In the MLBA, when the shift to a higher cranial index occured, and when the Anatolian IEs are supposed to have appeared, we don't notice an additional eastward shift in the genomes IIRC. The genomic eastward shift predates the roundening of the head and face.

Simon_W said...

So in any case it's dead wrong to think that the genomic eastward shift in central Anatolia was brought along by roundheaded Alpinoid or Armenoid folks, even though that might seem plausible from the modern distribution of morphological and metrical features.

Simon_W said...

Actually, considering the rather low upper faces of the EEF, the Bronze Age spread of non-IE CHG admixture throughout the Mediterranean appears to have contributed a lot to the forming of the classic leptomorphic Mediterranean type. Another possible factor in this process may be non-obvious Nordoid admixture from Corded Ware-related people.

epoch said...

@Simon_W

"But your argument is deeply flawed. I mean, there are various, unrelated language families with heavy CHG ancestry now, why should this have been different in the Bronze Age?"

If IE languages really are related to the Anatolian BA samples which show no steppe admixture IE languages must have been spoken as well by any of the related samples such as Minoan Greeks.

If the Minoan samples are *not* associated with IE, then why would then absence of steppe admixture in Anatolian BA samples mean anything with respect to the languages they spoke? If the Anatolian influx of Iranian in Anatolia would be related to IE but the Minoan isn't, we can safely assume the Iranian admixture is not related to any language shift. In that case the Iranian admixture would indeed be ancestral to many language families.

However, if Linear A is a non-IE language and has no steppe, whereas Linear B is Greek, and the samples show steppe ancestry, we have a confirmed language shift from non-IE to IE associated with an uptick of steppe admixture.

That has consequences for every interpretation of Anatolian BA sampes as well.

Them meee said...

@Simon_W

Because to them it seems more convenient to focus on the CHG in Yamnaya than the CHG-like ancestry in that Neolithic Peloponnese dude.

JuanRivera said...

Eh, mycenaean samples have been demonstrated to have steppe admixture. In contrast, minoan ones don't have it. While linear A has been transliterated, it hasn't been translated with IE languages, which makes the minoan language a non-IE one.

epoch said...

@ JuanRivera

Is that a certainty? Or likely? Or the majority of opinion?

old europe said...


Here's a study on the possible IE nature of minoan. Of course it is an open debate. But I remember that for many decades hittite was not considered IE............

https://www.google.it/url?sa=t&rct=j&q=&esrc=s&source=web&cd=6&cad=rja&uact=8&ved=2ahUKEwitwMmQqKXfAhXhUBUIHWQCCc0QFjAFegQICRAC&url=https%3A%2F%2Fwww.teicrete.gr%2Fdaidalika%2Fdocuments%2Fminoan_language%2Fjies27.pdf&usg=AOvVaw3ur3u-QTwyRnZbiFpWSxKh

Andrzejewski said...

There are linguistic connections between Proto-Indo-European and a language family near Baikal Lake, which probably date back to MA1:

https://en.wikipedia.org/wiki/Chukotko-Kamchatkan_languages

JuanRivera said...

Recently ran models of minoans and mycenaeans. Minoans, represented by Minoan_Lasithi, have no steppe nor its proxies WSHG, EHG and Ukraine_N, and those made fits worse. Mycenaeans, in contrast, has detectable steppe proxies WSHG, EHG, Ukraine_N, Ukraine_Eneolithic and Khvalynsk_Eneolithic, and 13%+ contribution from steppe and steppe-containing populations. Best fits are with steppe itself. As for the minoan language, it's likely to represent the pre-greek substrate.

Matt said...

epoch, I would tend to agree with Simon. I suspect you're being a bit optimistic, in that if you said that to Jena (or at least Heggarty) that you had proof Linear A was not IE, they'd go "A fascinating example, most likely, of a group with some migrating CHG ancestry either adopting or continuing to speak a paleo-European Aegean Neolithic language".

I mean, it seems like it's not like they can't postulate an expansion of CHG-IranN related people, that just happened to largely avoid any place where people had early writing, other than Hittites (as writing is rare enough), and they don't either say that PIE was the only language of CHG heavy folk (they hardly could). There's probably enough territory where writing wasn't attested for a long time that this is just about doable. This will be more tricky for them, I'd guess as sharper genetic distinctions made between CHG/IranN and various Neolithic Caucasian groups (more likely) - or their theory will be proved right (less likely).

JuanRivera said...

Chukotko-Kamchatkan are in Kamchatka, the Aldan coast and Chukotka in the present, though they were over a larger area in the past. That connection has been found in a 2014 paper, in which removing CK-like words in Celtic still kept CK as IE's closest relative. Also, I compared some PIE words with PCK ones, and a number of words with similar structure and meaning, which may be either cognates or false cognates. Though, IE seems more likely related to Uralic.

JuanRivera said...

As for the date of CK-IE separation (ignoring possible groups in which IE would be included), it may be some thousands of years after Afontova Gora 3.

JuanRivera said...

Correction: 2015 paper.

EastPole said...

The fact is that languages change really fast and a lot when people mix, when their religions and cultures mix, and change very slowly and very little when people don’t mix and have stable cultures and religions.

Here is an interesting article on linguistic changes in Armenian language:

https://www.quora.com/What-is-the-most-peculiar-sound-change-ever-recorded

thanks to Lazaridis:

https://twitter.com/iosif_lazaridis/status/1074390616728064000

They give an example of Armenian numerals and how they changed.
It may be interesting to compare Armenian numerals with Indo-Iranian, Slavic, Baltic and PIE.
Slavic and Indo-Iranian numerals are almost identical, Baltic are slightly different from Slavic and Indo-Iranian but apparently closer to PIE.

For example number five:
Ved. pañca, Av. panca
Pol. Pięć (=pienci) OCS pętĭ (=penti)
OPruss. pēnkjāi, Lith. penkì
PIE: *pénkʷe
Armenian: hing/hing/hink

Number four:
Ved. catvāras, Av. caθuuārō (=čatuaro)
OCS četyre, Pol. cztery (cz=č), Russ. četyre/ četvertij(=fourth)
OPruss. keturjāi, Lith. keturì
PIE: *kʷetwor-
Armenian: čork῾/čors/čors


https://en.wikipedia.org/wiki/Proto-Indo-European_numerals

Armenian PIE homeland is not a linguistic argument. There is nothing close to PIE in Armenian language. It must be something else. What? I don’t know.

JuanRivera said...

And the change of *dw to *(e)rk in Armenian.

Dragos said...

@ Matt

As I understood it, Andrew suggests a common origin for all IE, incl Anatolian, thus a level / rake kind of tree instead of a deeper dbifurcation between Anatolian and NPIE. As such they have would all a common formative substrate, according to the scenario. He links Anatolian's rapid divergence to the distinct substrate to which it subsequently enters (CHG/ Caucasus). Therefore, the same principle should apply to I-A, or Tocharian , even moreseo. Of course, behaviour can be erratic..

I am also confused by the suggestion of historical attestion mentions, uinless a new document has been discovered which details an arrival narrative, we should not necessarily conflate time of first attestation with actual time of arrival. Indeed, there is no attestation of anything much prior to 2000 BC. So I’m not sure about that theory at present.

On other topic '' I'd guess as sharper genetic distinctions made between CHG/IranN and various Neolithic Caucasian groups (more likely''

Absolutely. Same goes even for steppe groups, potentially.

andrew said...

@epoch @cpk I agree on the archaeology mentioned which was important in shaping my perspective. The majority view based on linguistics alone and not giving a great deal of weight to contact/substrate is that Anatolian IE languages need to date from something like 3500 BCE, to which the quibble over 2500 BCE v. 2000 BCE is nothing. I don't think that the archaeology or historically attestations or any ancient DNA to date supports this very early date of Anatolian IE and abandoning a steppe hypothesis for IE is increasingly disfavored.

JuanRivera said...

I would say that PIE people were unique. They were agropastoralists, who raised crops in highlands, rivers and moister zones, and pastoralist everywhere else. They supplemented with significant quantities of wild animals and fish, some wild plants, and a little of seals (monk and caspian ones) and other marine mammals. They had both permanent houses and ox/horse-driven carts. They were good observers who studied the sky and the weather. They also used ships in the rivers (to fish and navigate) and the seas (to fish, navigate, and going to the small islands in them). They had some of the most advanced technologies back then. They also were traders. Overall, they were closer to modern-day steppe peoples in lifestyle. Doubt that we'll see an exact parallel.

JuanRivera said...

And their homeland is most likely the Pontic-Caspian steppe.

Davidski said...

@Dragos

Same goes even for steppe groups, potentially.

The divergence times for CHG and Iran_N groups might be as much as tens of thousands of years.

It's even possible that CHG and Iran_N formed independently via the admixture of ANE-related groups from Siberia and Central Asia into Anatolian-related foragers living in and around the Caucasus. In other words, we could be looking here at parallel processes that gave rise to various CHG/Iran_N groups, rather than divergence from a common Iran_N-like root, with limited or no gene flow between them.

On the other hand, Bronze Age steppe groups look like they're all derived from the same Eneolithic base and, in terms of genome-wide ancestry, only differ in terms of the ratios of the same components (CHG/EHG/EEF). Obviously, they also sometimes differ sharply in regards to Y-chromosome haplogroups, but this shouldn't be overemphasized for reasons that I won't go into, but should be obvious to most reading this.

In other words, it's likely that most CHG and Iran_N groups spoke languages from highly differentiated language families, while it's possible that Bronze Age steppe groups spoke closely related languages form the same language family.

Unknown said...

Pelasgians, and not just Southern Italians, are rich in CHG-like ancestry. To wit, all the richness of J1 and J2 male Haplogroups prevalence within the current Hellenic population.

Davidski said...

To wit, all the richness of J1 and J2 male Haplogroups prevalence within the current Hellenic population.

Yeah, from Minoan and related non-Indo-European substrate groups.

By the way, I don't allow nicks like Unknown here, so for future reference, you must post with a unique name.

Andrzejewski said...

Btw, in regards to the Q1a sample found within Yamnaya Samara ~6.5kya, is it possible that it stems from merely a genetic drift P —> Q in lieu of P —> R(1), rather than being considered “foreign”? After all, they all derive from the ANE/Afontova Gora3/EHG cluster.

AWood said...

@Andrzejewski

Probably not because most, if not all of the Yamnayan samples are derived from L23+ which is well downstream of R1b, let alone when R and Q split from P. That must be at least 25,000 years ago, so it's not all that relevant. I would imagine the Q guys are related to contemporary Okunevans who have a completely different autosomal structure, yet were in near proximity and overlapped with Yamnaya and Afanasievo.

JuanRivera said...

Eh, that Q1a was actually Khvalynsk. Khvalynsk has a WSHG (West_Siberia_N) shift in relation to EHG (plus a CHG shift already known for years), and that the Q1a Khvalynsk guy had the most CHG out of all northern steppe samples (and as much as the Steppe Eneolithic samples in the southern steppe, specifically right next to the North Caucasus). All of that suggest that the Khvalynsk Q1a guy had his Q1a derived from local EHG groups.

Andrzejewski said...

@AWoods I’m conflicted about that as well: some researchers assume that Okunovo is related to Botai and that both have some Yenisseyan affiliation. I think that both KC/Yenisseyan and Proto-Proto-IE stem from MA1 ANE—> EHG, and that blondism in Yamnaya came with the AG3 mutation. What I am not sure about is the Botai affiliation as I understand R1b samples have been found so maybe some Yamnaya/Khvalynsk link did exist somehow?

JuanRivera said...

But the differentiation of R and Q from P1 were deep in Siberia. And still, Q1a was likely just above I2a in frequency.

Andrzejewski said...

@JuanRivera I wonder if the CHG in Khvalynsk were the aristocracy which may’ve brought with them the chariots, horse domestication and the kurgan burial customs while the local “commoners” were ANE/EHG or whether the incoming EHG “Indo-Uralic” or “Indo-Yenisseyan” we’re moving westbound from Central Asia and assimilating the local autochthonous CHG population speakers of Hatti/Hurrian/Kartvelian etc.?

Ric Hern said...

As I see it some diseases are more adapted to certain climates and environments and so also the people who live in that environments. Before modern medicine people used to stick to their latitude and those that didn't suffered severe culling before adaptation kicked in. The only way for peoples offspring to adapt quickly was by mating with the locals. When climate change hits people usually migrate. The stubborn ones usually perish. That is why we will not see a lot of Steppe overall ancestry in the South for example Hittites or a lot of Neolithic Farmer ancestry in the North. The Hunter Gatherer ancestry most probably bounced back because of their adaptation to a certain type of climate and environment. The unadmixed Farmers mostly went the way of the Dodo and more mobile Steppe people migrated to an environment more similar to where they originally came from with diseases they were more adapted to.

This is why the most probable spread of Steppe related people to the South were via Armies, Elites and Traders rather than massive tribal movements....

Andrzejewski said...

Also, the assumption that the ANG/EEF migrated with a sex ratio of 1:1 v. Yamnaya whose sex ratio was 5:1 M:F. I doubt that it happened. As in, what happened to all the PIE females left behind in the Steppes...

PF said...

I don't think the contact/substrate theory makes much sense to explain Anatolian. Whichever Steppe group spread IE (at least to Europe) already acquired farmer ancestry early on. Also, it's a bit hard to believe that there was a uniform "farmer language" and that the social dynamics that led to language influence were the same everywhere Steppe people migrated. In some places there were no EEFs at all -- e.g., all the pre-Steppe Baltic samples (Latvia_MN, Narva, etc.) are either pure EHG or EHG/WHG hybrids. Similarly, as Dragos mentioned, it's hard to square with the Indo-Aryan branches.

So for the theory to work and offer that the early branching isn't actually early but due to CHG-related contact/substrate, there'd need to be some isolated Steppe-related IE-speaking group that never meaningfully interacted with EEFs, somehow made it to Anatolia, and popped up on the scene shortly before 2000 BC. Not to mention that if there was indeed a CHG-related language shift in Anatolia, shouldn't it too have a strong EEF substrate?

A late date for Anatolian makes no sense to begin with; apparently already by ~2000 BC Luwian and Hittite were different languages. My opinion is that instead of trying to make Anatolian fit a late date, we should do the opposite, and try to make it fit a very early date.

Indeed, to connect the Steppes with Anatolia and IE, it does seem there needs to be some early branched group that basically goes unnoticed in the archaeological record for 1-2K years... and from the evidence we have now, didn't have anywhere near the same level of genetic impact that Steppe groups had everywhere else.

"Forget it Jake, it's Anatolia." :-D

JuanRivera said...

There appears to be a cline of EHG and CHG, in which there was an abrupt shift somewhat north of the Caucasus. While there's some CHG in EHG, it's in the order of single digits, and the bulk of steppe's CHG (which was different from south caucasian CHG) arrived in the Neolithic and later, bringing with them the Neolithic package and the North Caucasian-like influence observed in PIE (although CHG-heavy people like the Steppe Eneolithic trio and the Khvalynsk Q1a guy spoke languages related to PIE, which went extinct, but not before having an impact upon it). There was also a WHG-ANE cline, in which there appears to be a fast decline in ANE as one goes west, as exemplified by the transition between 65%+ ANE-derived EHG and 30-45% ANE-derived Ukraine_Mesolithic/Ukraine_N. So, while ANE bought R and Q, and pre-PIE language, WHG bought I2a and CHG bought Basal Eurasian (as early as Mesolithic EHG samples), the North Caucasian-like influence in PIE and the neolithic package.

Ric Hern said...

@ Andrzejewski

The Kamennobalkovsky culture of the Lower Don River (Kamennaya Balka) apparently had some kind of link to the Paleolithic Imeretian Culture of the Caucasus...

Andrzejewski said...

The ethnolinguistic phylogenetic picture of NE Europe is far from simple: how else can we square the fact that the Saami language with its strong Uralic N1c3 marker has roughly 1/3 of its vocabulary deriving from some obscure substrate?

And, in case this substrate relates to an assimilated WHG/SHG language(s), is it by any chance related to Erteboelle?

JuanRivera said...

Steppe-like groups left impact in MLBA Anatolians, whereas Anatolian Neolithic groups made no direct impact upon Eastern Iranians and Indians, nor in the Xiaohe mummies. Also, there's the issue of missing farming vocabulary, extensive domestic animal and products vocabulary, metal vocabulary, extensive water vocabulary, evidence of contact of proto-Germanic and proto-Indo-Iranian with Proto-Uralic, evidence of substrates in every IE group (including Anatolians), the words for sand, beaver, tortoise and birch, and more. Also, they were recorded in the archeological record as Suvorovo, Coțofeni, Ezero and Kumtepe, where steppe-like influence is visible.

JuanRivera said...

I heard that the Saami substrate was Basque-like, which if true, would make Basque and the Saami substrate WHG and SHG languages. As for Ertebølle, it was likely related to it because of the proximity to Scandinavia.

Andrzejewski said...

Now, speaking of Erteboelle Culture, it seems that the Hunter Gatherer I2a Erteboelle + EEF LBK merged to form the Funnelbeaker with its mixed forager/Farmer economy. When IE (R1a1 Rich Corded Ware + R1b Rich Bell Beaker) took over Funnelbeaker/Globolar Amphora and form the basis of the Proto-Germanic Branch, does it indicate that the uniqueness of this branch is because of the Satem (Corded) + Beaker (Centum) merger, or does it alternately indicate a pre-IE substrate?

I also ponder if the historic distinction between Northern Germans (Prussia, Saxony) v. Southern Germans (Bavaria, Austria) and between Protestants v. Catholics, respectively are the outcome of the different origins: WHG/SHG in the north, EEF/Anatolian Neolithic in the South?

JuanRivera said...

There appears to be a farming substrate in Proto-Germanic. As for the distinction, Northern Europe used more hunting, gathering, fishing and pastoralism, whereas Southern Europe depended more on agriculture (although some Italic-speaking groups were heavily pastoralist).

Cpk said...

@Simon_W
"That date you keep citing, 2000 BC, is wrong. The change in cranial and facial shape that you have in mind, dates to the MLBA. So according to modern dating, I think, 2500 BC or something"

Can you show me a source/research on this?

Andrzejewski said...

@JuanRivera - Not sure if it is a "farming" substrate or "Hunting" one: lots of words related to fish like "carp" are regarded pre-IE...who knows?

Second, the so-called non-IE substrate in the German language seem to be obsolete. It is actually strange that during Hitler's time ("pure Aryanism") the currency of having up to 1/3 of the German language being supposedly non-IE had a momentum whereas nowadays no serious linguist mentions it anymore. https://en.wikipedia.org/wiki/Germanic_substrate_hypothesis

Andrzejewski said...

Here's a list of alleged words which are or were considered "non-IE" but they overwhelmingly cover "hunting" vocabulary, not "farming" one:

https://eupedia.com/linguistics/non-indo-european_germanic_words.shtml

Ric Hern said...

As far as I understand Germanic evolved more to the North mostly within Scandinavia. So my guess is a more Hunter Gatherer like Substrate. The fjords and Mountains certainly were more formidable terrain and less suitable for Farming...

Andrzejewski said...

We see further shift from Anatalian_N into CHG-related ancestry in Greeks, same pattern that happened concerning Southern Italians and Sicilians:

https://www.biorxiv.org/content/biorxiv/early/2017/05/09/135616.full.pdf

"This can partly be explained by the above differences along with major Slavic influence during the middle ages, and also the Hellenic colonization of Magna Graecia. However, I think there may be a deeper ancestral similarities between these regions. “By contrast, data from five southern Greek Neolithic individuals (labelled ‘Peloponnese_Neolithic’)—three (plus one that has previously been published26) from Diros Cave and one from Franchthi Cave—are not consistent with descent from the same source population as other European farmers. D statistics (Supplementary Table 2) show that these ‘Peloponnese Neolithic’ individuals, dated to around 4000 bc, are shifted away from WHG, and towards CHG, relative to northwestern Anatolian Neolithic and Balkan Neolithic individuals. We detect the same pattern in a single Neolithic individual from Krepost in present day Bulgaria (I0679_d, 5718–5626 bc). An even more marked shift towards CHG has previously been observed in individuals associated with the Bronze Age Minoan and Mycenaean cultures 26, suggesting gene flow into the region from populations with CHG-rich ancestry throughout the Neolithic, Chalcolithic and Bronze Age. Possible sources are from people related to the Neolithic population of the central Anatolian site of Tepecik Çiftlik21, or the Aegean site of Kumtepe 11, who are also shifted towards CHG relative to northwestern-Anatolian Neolithic samples, as are later Copper and Bronze Age Anatolians”

Ric Hern said...

My other guess is West Germanic = Corded Ware + SHG + Bell Beaker. And East Germanic = Corded Ware + SHG + Proto-Finnic ?

Andrzejewski said...

CHG shift during Mid-Neolithic may have something to do with an increase of mtDNA?

I believe that the increase of mtDNA H from 19% to 40% during the Middle Neolithic times has anything to do with a partial replacement of original LBK-like Anatalian_N/Levant_N with a more CHG-shifted population. Note also that the rate increased from a Middle Eastern frequency of 19% to over 40% in Europe. (It is also speculated that the CHG had BOTH alleles for fair skin SPL instead of just one, as was the case with Anatolian farmers:


Here's an article published by Nature Communications about a study that suggests Europe's modern mtDNA signature was largely established about 6000 years ago, in the mid Neolithic, by people of an unknown origin who largely replaced the early Neolithic farmers, for reasons that aren't yet clear. Although it does also indicate that Bell Beaker folk expanding out of Iberia did have a significant impact during the late Neolithic.

www.nature.com/ncomms/journal/v4/n4/full/ncomms2656.html

Here's the abstract.

Haplogroup H dominates present-day Western European mitochondrial DNA variability (>40%), yet was less common (~19%) among Early Neolithic farmers (~5450 BC) and virtually absent in Mesolithic hunter-gatherers. Here we investigate this major component of the maternal population history of modern Europeans and sequence 39 complete haplogroup H mitochondrial genomes from ancient human remains. We then compare this ‘real-time’ genetic data with cultural changes taking place between the Early Neolithic (~5450 BC) and Bronze Age (~2200 BC) in Central Europe. Our results reveal that the current diversity and distribution of haplogroup H were largely established by the Mid Neolithic (~4000 BC), but with substantial genetic contributions from subsequent pan-European cultures such as the Bell Beakers expanding out of Iberia in the Late Neolithic (~2800 BC). Dated haplogroup H genomes allow us to reconstruct the recent evolutionary history of haplogroup H and reveal a mutation rate 45% higher than current estimates for human mitochondria.

Ric Hern said...

@ Andrzejewski

As far as I understand the latest study about the Bell Beaker people there were very little Ancestry from South of the Pyrenees Mountains which reached Central Europe....

Andrzejewski said...

One additional interesting point of reference here is that two Non-Indo-European speaking populations have *very* Europoid-looking phenotypes: Mordvins/Mari, who speak Uralic languages and are either EHG/WSHG:

https://images.search.yahoo.com/yhs/search?p=mordvins&fr=yhs-itm-001&hspart=itm&hsimp=yhs-001&imgurl=http%3A%2F%2Fwww.elupuu.org%2Ffilebank%2F31-group_1_600px.jpg#id=2&iurl=http%3A%2F%2Fwww.elupuu.org%2Ffilebank%2F31-group_1_600px.jpg&action=click

The second population may be a mixture of CHG + Anatolian Farmers, which is the Adyghe/Cherkes who speak a Northwest Caucasus language (postulated by Johanna Nichols to have a phylum with PIE dating back to 12,000YBP (10,000BC):

https://pbs.twimg.com/media/BsdvkmNIIAEIlHc.jpg:large

It is surprising to find out that the putative proximate sources for Yamnaya or other Steppe population speaking Proto-Indo-European languages both look very modern European, despite their non-IE origins.

Andrzejewski said...

Mari/Komi (Uralic speaking in European Russia) - https://tse3.mm.bing.net/th?id=OIP.KHiGMkdIoyVMph9cURbPwQHaFO&pid=15.1&P=0&w=261&h=185

Davidski said...

@Andrzejewski

Komi and Mari people are phenotypically very heterogeneous, with some looking totally North-Central European and others very Siberian. This is because they have 15-25% Siberian admixture, depending on how you calculate and define it, and that's a lot compared to the vast majority of Indo-European-speaking Europeans. See here...

Corded Ware people =/= Proto-Uralics (Tambets et al. 2018)

And nope, latest ancient DNA data refute that idea of any significant migrations of Iberian Bell Beakers out of Iberia. See here...

Migration of the Bell Beakers—but not from Iberia (Olalde et al. 2018)

Them meee said...

@Andrzejewski

Also Mordvins are of mostly Eastern European descent, obviously.

Also isn't proto-proto-IE called pre-proto-IE?

Andrzejewski said...

@Them meee I do not know what came before "Proto-Indo-European"; it is merely a construct of scientists rather than a living language. Additionally, no records of it survived because it was not written. My estimate is perhaps that during the shift from Samara -> Khvalynsk the language started to form and my guess is that it mostly came with the Yeniseian- Kamchatka- like lexicon. Derivation from Afontova-Gora3 is likely so the Indo-Europeans DID have blond hair, at least some alleles for it. What boggles my mind are some unresolved issues:

1. Did the ANE/EHG AG3 migrate west into the Steppe from Lake Baikal post-LGM, then assimilate the original Caucasus HG to incorporate the substrate of the CHG into the mostly Steppe ANE-derived language (close to Kamchatka language family), or did the CHG incursion from South of the Caspian transform Samara into Khvalynsk, with the foreigners being the "Aryan" or the ruling elite, with kurgan burial style, chariots, metallurgy and so on coming in from the "South". (or via later Maykop).

2. Were the original IE R1b speakers? It seems as though both R1a1 and R1b stem from the Mal'ta Boy or a close lineage so perhaps the languages were of the same language family?

3. Is it probable (if it was not the case above), that the R1a1 were speakers of some "Old Europe" HG language that was "Indo-Europeanized" by the Khvalynsk folks?

4. Maybe the alleged "farmer substrate" in IE languages did NOT come with a Corded Ware substrate but with Yamnaya long term (2 millenia) of contacts with farmer societies to the West of it, such as Cucuteni Tripolye?

5. If indeed the overwhelming majority of Yamnaya (and of Bell Beaker) were R1b-L23 subclade, then who were the (predominantly R1a1) Corded Ware and what's their relationship to Yamnaya? Is it likely or at least possible they were a sister-Steppe Culture?

I don't have the answers. I hope Mathieson, Haas, Lazaridis, Reich et al. do, or will soon have :)

Dragos said...

@ Davidski

''In other words, it's likely that most CHG and Iran_N groups spoke languages from highly differentiated language families, while it's possible that Bronze Age steppe groups spoke closely related languages form the same language family.''

Yes I agree in part.
However, before the steppe groups began to converge in the 5th century, they had been significantly diverged. So it might be that an R1a-M17 in Dereivka spoke a very different language to an R1b in the pre-Volga steppe. Unless one takes Andrzejewski's view, which IMO is moving into the realm of pseudo-science.

Davidski said...

@Dragos

However, before the steppe groups began to converge in the 5th century, they had been significantly diverged. So it might be that an R1a-M17 in Dereivka spoke a very different language to an R1b in the pre-Volga steppe.

I doubt that because Yamnaya-like populations were already expanding during the Eneolithic from the same relatively compact area, probably from between the Black and Caspian Seas.

We know this because the R1a-M417 sample from Dereivka has a lot of ancestry of this type, unlike the preceding samples from Dereivka. And obviously so do the Eneolithic samples from the North Caucasus Piedmont steppe and Samara.

Davidski said...

Actually, that Sredny Stog II R1a-M417 sample is from Alexandria not Dereivka. But it's from eastern Ukraine anyway.

EastPole said...

@Davidski
“Actually, that Sredny Stog II R1a-M417 sample is from Alexandria not Dereivka. But it's from eastern Ukraine anyway”

Alexandria site with R1a-M417 belongs to Sredny Stog II Dereivka culture.

Matt said...

@Dragos: I am also confused by the suggestion of historical attestion mentions, uinless a new document has been discovered which details an arrival narrative, we should not necessarily conflate time of first attestation with actual time of arrival. Indeed, there is no attestation of anything much prior to 2000 BC. So I’m not sure about that theory at present.

It seems tougher to believe in a long presence of IE in a region where there is long history writing, and other languages attested in writing, without an IE presence in writing. As you point out though, writing is much confined within some of these regions for much of this time (Mesopotamia, small band of SW Iran, Egypt, I think?).

Absolutely. Same goes even for steppe groups, potentially.

Potentially... maybe. I'll go into more detail - so, as I interpret that, this is kind of like an idea where the Bronze Age to late Iron Age steppe was probably more linguistically diverse than when it first enters history, and we kind of just don't "catch" these early steppe groups "in the act" of speaking lots of non-IE languages because the attestation of what they did speak is pretty late after first writing, away from the centers of writing and in a narrow window of leveling by Sarmatian / Scythian etc. before lots of replacement by Turkic / Hunnic and other peoples.

Working under Reich's model (as I understand it) where the Steppe_EMBA group (Yamnaya and Afanasievo and Poltavka) are all expansion of a single pulse of people probably originating from a contact region where they learn to use the wheel, which demically replaces all these smaller groups and cultures that came before based out of herding / hunting camps on the Pontic-Caspian steppe, that kind of seems less possible.

If we were using the model where Steppe_EMBA sort of emerges from a kind of multi-regional leveling and admixture between different steppe complexes, though, none of which were actually replaced and all merged in, then that's more likely, that they'd be speaking and retain multiple languages.

I tend to believe that the "Steppe_EMBA is a small group that replaces everyone else with more EHG ancestry" is more likely and is probably proto-Indo-European speaking, perhaps acquiring Uralic speakers as well then or a bit later in history is more likely though. Under a "multi-regional emergence of Steppe EMBA" though, multiple language families seems intuitively reasonably possible and perhaps more likely.

Davidski said...

@Matt

All of the Eneolithic samples (plus the later Yamnaya) in my map here are likely to be closely related within a recent time frame because they share significant ancestry of the Yamnaya type.

We don't yet know where this ancestry was expanding from, but it was probably a relatively small area because of how specific this signal is, and considering the dates of the samples, it's likely to have been during the Eneolithic, not the Early Bronze Age.

Yamnaya: home-grown Map

So it's hard for me to believe that this was a process involving the spread of different language families.

I don't think the possibility that, as per archeology and also now ancient DNA data, they later merged again to form Yamnaya contradicts this.

Andrzejewski said...

@Matt what makes you think that these early Steppe groups were speaking “lots of non-IE languages” instead of IE ones?

Davidski said...

@Matt what makes you think that these early Steppe groups were speaking “lots of non-IE languages” instead of IE ones?

That's the point I made as well.

As I see it, there was a PIE expansion on the steppe during the Eneolithic already.

Then there was a homogenization on the steppe during the Early Bronze Age that gave rise to Yamnaya, and then to the expansions of late PIE via Corded Ware and Yamnaya.

Slumbery said...

@Awood
"I would imagine the Q guys are related to contemporary Okunevans who have a completely different autosomal structure, yet were in near proximity and overlapped with Yamnaya and Afanasievo."

I do not think so. For two reasons.

First, because Okunev is not really contemporary with anything on the table here. It is second half of third millennium BC and early second millennium BC, so contemporary with late Poltavka and Sintashta, but a way too late even for Yamnaya proper, let alone Steppe Maykop or especially Khvalynsk.

Secondly because the genetics do not add up.
Global25 nMonte Okunev (fit 1.4163)
Afanasievo: 18.33%
Botai: 49.17%
Shamanka N: 32.5%
Weest Siberia N: 0%

Okunev is a heavy with BHG-like ancestry and up-to-date the earliest known Volga-Ural-Steppe region population with this kind of ancestry is Mezhovskaya. The earliest that is west from Middle Siberia is Bolshoy Oleni Ostrov.

Slumbery said...

@Davidsky
"The divergence times for CHG and Iran_N groups might be as much as tens of thousands of years.

It's even possible that CHG and Iran_N formed independently via the admixture of ANE-related groups from Siberia and Central Asia into Anatolian-related foragers living in and around the Caucasus. In other words, we could be looking here at parallel processes that gave rise to various CHG/Iran_N groups, rather than divergence from a common Iran_N-like root, with limited or no gene flow between them."

It is possible, but it seems to be unlikely for me. I think they are too similar to each other for tens of thousands of years divergence. As for the other option, what is the probability that two populations that live among different neighbors become this similar just because of accidentally similar deep ancestry ratios?

A model where CHG-Iran become a thing (one could say crystallized) during LGM isolation and then expanded right after LGM (so 17-18 kya?) can explain the situation, because the changes during the time passed, combined with different regional mixtures are sufficient to explain the divergence.

Davidski said...

@Slumbery

It is possible, but it seems to be unlikely for me. I think they are too similar to each other for tens of thousands of years divergence. As for the other option, what is the probability that two populations that live among different neighbors become this similar just because of accidentally similar deep ancestry ratios?

CHG and Iran_N aren't very similar. Have a look at the linear models here...

On the enigmatic early Neolithic farmers from Iran

And also the PCA here...

Ust'-Ishim man x2

Honestly, I don't think there ever was any significant gene flow between CHG and Iran_N.

They both look to be the parallel outcomes of Dzudzuana-related foragers separately absorbing ANE-related and ENA-related admixtures, the latter of which seems more pronounced in Iran_N.

Dragos said...

@ Matt

''It seems tougher to believe in a long presence of IE in a region where there is long history writing, and other languages attested in writing, without an IE presence in writing. As you point out though, writing is much confined within some of these regions for much of this time (Mesopotamia, small band of SW Iran, Egypt, I think?).''

I had to do a quick search. If i gathered correctly, excluding scripts like Vinca & Indus, the earliest transliteratable scripts attest
- Sumerian c. 3000 BC
- Akkadian c. 2500 BC (the earliest Semitic)
- Hittite names attested in Akkadian (Old Assyrian) c. 2000 BC
At this time other foreigh idioms are also attested in Akkadian - Elamite, Mari (Syria)..
So there is not much of a lag to really speak of.
More difficult to evaluate are the proposed contacts of Sumerian with IE (Pokorny, Whittaker, Sahala)

Slumbery said...

@Davidsky

Both on the D stat based models and on the PCA from your other article modern Iranians and even Iran_Clh are often placed much closer to CHG than to Iran_(E)N. That cannot be right in therms of their ancestry proportions, so I have to assume that their position on these plots is heavily influenced by relatively recent mixing and that is exactly my point.

I do not say that your alternative ideas are impossible or fundamentally wrong, but I can't see a decisive argument against a more recent (~10 000 years) common root either.

Davidski said...

@Slumbery

That cannot be right in terms of their ancestry proportions, so I have to assume that their position on these plots is heavily influenced by relatively recent mixing and that is exactly my point.

I'd say that it's due to Anatolian admixture in Iran_ChL and a heavier Dzudzuana-related base in CHG, as well as more eastern admixture in Iran_N.

Genetic drift also has a very heavy bearing on the outcomes in the PCA, which suggests that Iran_N and CHG don't share "recent" genetic drift.

By the way, in terms of a straight D-stats comparison, Iran_N doesn't form a clade with CHG to the exclusion of Iran_ChL, so what this means is that your interpretation of what is right in terms of their ancestry proportions isn't correct. :)

Chimp CHG Ganj_Dareh_N Seh_Gabi_ChL 0.0087 2.824 931497
Chimp Seh_Gabi_ChL Ganj_Dareh_N CHG -0.0018 -0.566 931497
Chimp Ganj_Dareh_N CHG Seh_Gabi_ChL 0.0105 3.234 931497

Them meee said...

I have seen people argue that Yamnaya-like or with Yamnaya/Khvalynsk-like ancestry could have included para-IEs. How likely is it that? Seems a bit implausible to be because the ancestry is too similar across all groups.

Davidski said...

@Them meee

I think it's possible that the Eneolithic steppe groups from the North Caucasus Piedmont steppe were para-IE, depending on how that's defined.

But if Yamnaya and Corded Ware were really late PIE, then I don't see why or how Sredny Stog II and Khvalynsk weren't PIE.

Slumbery said...

@Davidski
"I'd say that it's due to Anatolian admixture in Iran_ChL and a heavier Dzudzuana-related base in CHG, as well as more eastern admixture in Iran_N."

I tend to agree, but we are saying the same thing differently here. On those linked plots populations are being tossed around by mixing events that are much more recent than tens of thousands of years, so the fact that they do not cluster together does not disprove a more recent common root (more recent than tens of thousands of years).

"Genetic drift also has a very heavy bearing on the outcomes in the PCA, which suggests that Iran_N and CHG don't share "recent" genetic drift."

Would you say a 8-10 ky old shared drift is "recent" as far as it matters for the PCA? Something that would still pull them closely together after 8-10 ky additional drift + mixing from different sources?
I calculate that 8-10 ky back from Iran N, it is less on the CHG side, because it is much older, but this still a very long time.

"By the way, in terms of a straight D-stats comparison, Iran_N doesn't form a clade with CHG to the exclusion of Iran_ChL, so what this means is that your interpretation of what is right in terms of their ancestry proportions isn't correct. :)"

I simply meant that modern Iranians and Iran Clh cannot have more CHG ancestry than Iran N ancestry, so the cause of this clustering on the linked plots must be relatively recent (post Iran N) mixing. If this recent mixing can cause the mixed populations to plot closer to a population that is not a main ancestor of them than to a population that is a main ancestor of them, then the distance between CHG and Iran N on those plots is not very heavy argument against their possible common roots (in the 8-10 ky time depth).

Davidski said...

@Slumbery

I simply meant that modern Iranians and Iran Clh cannot have more CHG ancestry than Iran N ancestry, so the cause of this clustering on the linked plots must be relatively recent (post Iran N) mixing.

Yes, this is due to Anatolian and also steppe input into post-Iran_N populations in Iran.

But this is irrelevant to our discussion, so there's no point focusing on it, because CHG harbors elevated Anatolian/Villabruna affinity not as a result of such recent admixture but from its Dzudzuana-related forager base ancestry.

Hence, my interpretation is that CHG comes from a different, more Anatolian/Villabruna-like root population than Iran_N, and thus is better described as a parallel outcome of a very similar ancient mixture process as Iran_N, rather than its sister clade per se.

Chimp Barcin_N Ganj_Dareh_N CHG 0.0179 5.893

Chimp Villabruna Ganj_Dareh_N CHG 0.0253 5.722

I suppose that another explanation could be that a sister clade to the population ancestral to Iran_N moved into the Caucasus and mixed with the Dzudzuana-related foragers there to form CHG. But I'm not a big fan of this scenario at the moment, due to the somewhat different types of eastern ancestry that I'm picking up in CHG and Iran_N.

Slumbery said...

@Davidsky

We are both exhausted our arguments (well, I admit that I mostly just trying to shoot down yours), so this discussion is nearing its end. So just two things.
1. "a sister clade to the population ancestral to Iran_N moved into the Caucasus and mixed with the Dzudzuana-related foragers there to form CHG." is exactly what I think. This is the theory I am pitting against the theory of accidentally similar ancestry proportions without any "recent" (less than tens of thousands of years) common ancestry. I find this much more likely.
And the different types of eastern ancestries could have arrived later for both of them.
2. One of the reasons I prefer this version is the J2 paternal lineage. It links CHG to Iran N and its TMRCA is pretty much right for an expansion at the end of the LGM, let's say from the South/SW Caspian coastal region into the surrounding higher regions.

Andrzejewski said...


Further proof that Bell Beaker did *not* come from Spain:

4,500 years ago there was a Steppe-based invasion which wiped out practically every single male in the Iberian peninsula. Does it explain why Basque people are predominantly R1b-L23? Could be!

https://elpais.com/elpais/2018/10/03/inenglish/1538568010_930565.html

Taymas said...

@Matt and @andrew, I really appreciate you guys bringing up the contact issue with linguistic evolution. My intuition tells me it's a huge factor, look forward to digging into the links. Regarding the Turkic datapoint, wouldn't Anatolian/CentralAsian states/priesthoods/literacy be a serious confound? I would guess that while neglecting contact overestimates age, neglecting {phonetic} literacy would underestimate (where applicable). Thinking also of Arabic.

I am not a linguist, not in the slightest, just curious.

JuanRivera said...

The migration of ANE to the Pontic-Caspian steppe was somewhere between the end of the LMG and the Younger Dryas. Before ANE came (besides that in CHG), there was a WHG-CHG cline, in which CHG-heavy groups suddenly gave way to CHG-poor groups somewhere between the Black and Caspian sea. Archeology seems to indicate an epigravettian influenced by ideas from the Near East. When ANE came, the WHG-CHG cline was turned into an EHG-CHG cline, given that EHG is ANE+WHG and a little of CHG. When the Neolithic came, CHG-heavy groups contributed more CHG to populations in the north, and with it the Neolithic package. Maykop has nothing to do genetically with the steppe as it has additional Anatolian and Iranian ancestry that's absent in steppe groups, plus its CHG is different from the one in steppe groups. As for R1a and R1b, both were part of the same population, as R1a was detected in Neolithic and Eneolithic ukrainians and also a sample in Khvalynsk. Likely, there was a cloud of PIE and para-PIE languages stretching from the forest-steppe/forest transition to the North Caucasus, and from the Dniester to somewhat east of the Ural river. As time progressed, also WSHG (West_Siberia_N, which was 80%+ ANE) entered to Khvalynsk and descendant steppe groups from the east. As for blondism, not only AG3 had it, but also it was detected in Samara_EHG, in Khvalynsk and certainly present in Sredny Stog, Steppe Eneolithic and Yamnaya, although their remains didn't survive. Homogeneization took place in the steppe in such a way that EEF was introduced to the eastern part, WSHG was introduced in the western part, CHG increased in the north, and EHG increased in the south. R1a was sorted more to the north, while R1b was sorted more to the south, resulting in predominantly R1a Corded Ware and predominantly R1b Bell Beakers. While CHG south of the Caucasus didn't contribute to steppe, not even to EHG, it transmitted the Neolithic package and metallurgy to different "steppe CHG" in the foothills of the North Caucasus, who actually contributed to steppe groups. Meanwhile, the horse seems to be domesticated in the Late Neolithic or Eneolithic by Khvalysnk and Sredny Stog groups, who were poorer in CHG, except for the Q1a sample. Para-PIE languages in the forest-steppe/forest transition and between the Black and Caspian seas went extinct, being replaced by PIE, but not before contributing some terms.

Slumbery said...

@Davidsky

Just to see what comes from it I tried to model CHG as Hotu HG + others and Hotu HG as CHG + others in G25 nMonte. I can't get very far with this, because there are no data from the right older populations, but the results are still useful if we keep their limits in our mind.

So CHG came out as something like 80+% Hotu HG + something that best represented by Anatolian Neolithic (among the samples of the database). No other populations I tried (including any European HG or Central Asian - Siberian groups) improves the fit, or grabs any ancestry share in the presence of those two.
Hotu HG came out as 2/3 CHG + very significant ANE (around 30% represented by MA1) + some minor but significant southern ancestry (I could not find anything better than Natufians for that). Ganj_Dareh_N is the same composition, but more CHG and Natufian and less ANE.

This is pretty consistent with what you said about their differences and also explains why Anatolian admixture into Iran Clh pulls it closer to CHG.
And also these are the kind of admixtures you would expect based on geography. NW expansion meets circum-Pontic admixture (most similar to later Anatolians among the available proxies), while the SE group gets a lot of contact from Central Asian ANE and then something from the South.

PF said...

@Slumberry

You should use AG3 for ANE, as MA1 already has CHG-related admixture (from the non-Basal portion of CHG). Just to see how this result held up on G25, AG3 came out as 100% MA1, while MA1 came out as 84% AG3 + 16% CHG.

Using your model I got for Hotu_HG

CHG,68.2
AfontovaGora3,20.8
Natufian,11

distance%=14.5636

So ANE goes down from ~30% to ~20%. Whatever was happening way back in the day in central Asia was super interesting and likely quite complex. Anyways, probably better to hold off trying to figure out the subtleties between CHG/Iran until we get the Dzudzuana genome. When is that being published already?

JuanRivera said...

I think that AG3 was also CHG-admixed. If we want an ANE proxy without CHG admixture, we will have to wait for the Yana RHS genomes. It would be interesting to see the interactions between Dzudzuana-derived peoples and Yana-derived ones.

EastPole said...

@JuanRivera
“Para-PIE languages in the forest-steppe/forest transition and between the Black and Caspian seas went extinct, being replaced by PIE, but not before contributing some terms.”

Could you tell us in more detail about how you see this process taking place. Where, in which cultures, para-PIE languages were spoken and how, when and where PIE was formed and from where and when it expanded.
I think that West Yamnaya where Post-Stog tribes mixed with incoming from the east Yamnaya, forming pre-Corded Ware which then migrated north, is a very interesting period. It is a pity we don’t have aDNA from this time.

Dragos said...

Davidski
If the Caucasus was depopulated during LGM, as it is written, then the Dzudzuana ancestry found in later CHGs might be imported from elsewhere, perhaps near Iran.
The fact that Anatolian neolithics form a clade with Dzdz., despite bearing extra basaloid admixture, suggests that Anatolia was subject to less turnover during the LGM, and harboured some refuges near Antalya for example

Slumbery said...

@Dragos
"If the Caucasus was depopulated during LGM, as it is written, then the Dzudzuana ancestry found in later CHGs might be imported from elsewhere, perhaps near Iran."

Not likely from near Iran, because CHG have more of that ancestry than Iran HG.

@PF
Thank you, that is interesting.

Ric Hern said...

If someone can read Russian please read about the Late Upper Paleolithic Imeretian Culture apparently with ties to the Baradostian and influence upon the Kammenobalkovsky Culture at the Lower Don. Apparently there were three cultural layers found at Kammenaya Balka...Maybe this is where ANE met the Caucasus population to form CHG...ANE from the Southern Urals down the Volga and Don and Proto-CHG from the Caucasus up the East Black Sea Coast...

Andrzejewski said...

@Ric Hern So ANE formed the basis of BOTH EHG & CHG populations?

JuanRivera said...

It was the basis of EHG and also was a really significant component in CHG (30-40%). The difference is that EHG has WHG and more ANE, whereas CHG has a lot of Dzudzuana and some ANA. As an aside, EHG's sibling is WSHG (West_Siberia_N), the latter having barely any WHG, having more ANE (80%+), and 10-15% Baikal_HG which wasn't present in EHG. As for how PIE displaced its relatives in Eastern Europe, it was through prestige and some amount of dominance, but as said before, they left some traces. Western para-PIE transmitted some farming terms and the term for bull (*(s)táwros), southern para-PIE had an even heavier North Caucasian influence, northern para-PIE had some heavy influence from and to Proto-Uralic, and eastern para-PIE had some influence from Botai, and transmitted the word for sand (*sámh₂dʰos, derived from a root meaning 'to pour') and, if truly a PIE root, *kebʰ-.

JuanRivera said...

Northern para-PIE contributed beaver and birch, while southern para-PIE contributed mountain, mist and some trade terms.

Ric Hern said...

Something like that yes. More ANE in EHG and less in CHG...Let's rather say one of the basis...

Andrzejewski said...

I wonder where what they call “Indo-Hittite” and “Indo-Uralic” fit in the mix.

Andrzejewski said...

Or maybe it is likely that they CHG admixture in the Khvalynsk —> Stredny Stog —> Yamnaya is just a genetic drift from ANE —> EHG?

JuanRivera said...

Indo-Hittite would fit in the early Eneolithic, with Anatolian migrating out through the Balkans, leaving archeological and genetic traces, whereas core IE remained more time. Indo-Uralic seems EHG. As for CHG, it isn't just genetic drift as it has Basal Eurasian and ANA ancestry, plus it matches Kotias (a 14 kya sample from the Caucasus) very well, also, the Basal Eurasian and ANA in EHG seems to come from CHG. The CHG in steppe groups seems to match with surface Northwest and Northeast Caucasian (hence North Caucasian) influences in PIE (at the core, PIE is another North Eurasian language family deriving from ANE).

JuanRivera said...

Correction: IE instead of PIE in the last paretheses.

Them meee said...

CHG is not a drifted EHG. The CHG in the steppe is... different, but it’s not drifted EHG since it clearly has different origins than EHG.

JuanRivera said...

CHG seems to originate in the Near East, from a mixture of ANE and Dzudzuana, whereas EHG is E European in origin, arising out of WHG+ANE+CHG. The two only shared ANE and little Dzudzuana, and then, CHG's ANE is Mal'ta-like, whereas EHG's ANE is AG3-like.

Andrzejewski said...

@ JuanRivera I bet the finding that AG3 was the first population to possess the blondism allele greatly inconvenienced many scientists and researchers, lest it would vindicate the Third Reich's/Thule Society's link between the Yamnaya Proto-Indo-European "Aryans" with blond hair ;)

Davidski said...

It's rather unlikely that anyone lost any sleep over a single allele in a heavily damaged ancient sample, especially also since Yamnaya samples don't show a high frequency of alleles associated with blond hair or light eyes.

The earlier finding that Andronovo from South Siberia was predominantly a blond/light-eyed population probably did cause a bit of panic in some academic circles, but as it turns out now, the reason for this wasn't Andronovo's high level of Yamnaya ancestry, but probably admixture from (non-Indo-European) Middle Neolithic farmers living west of the steppe.

Andrzejewski said...

When it comes to various subclades of R1b I am often at a loss: apparently,a male buried at Lebyazhinka approximately 7,000 years BP and often referred to by scholars of archaeogenetics as the "Samara hunter-gatherer" appears to have carried the rare Y-DNA haplogroup R1b1* (R-L278*). https://en.wikipedia.org/wiki/Samara_culture

Now, not even a few hundreds years later, and another subclade of R1b seems to be part of a different horizon:

https://en.wikipedia.org/wiki/Ancient_North_Eurasian#EHG

"Eastern European Hunter-Gatherer (EHG) is a lineage derived predominantly from ANE. It is represented by two individuals from Karelia, one of Y-haplogroup R1a-M417, dated c. 8.4 kya, the other of Y-haplogroup J, date 7.2 kya, and one individual from Samara, of Y-haplogroup R1b-P297, dated 7.6 kya. This lineage is closely related to the ANE sample from Afontova Gora, dated c. 18 kya".

_______

So, how come the Samara R1b-P297 individual belongs to an "Eastern HG" cluster, whereas an R-L278* (R1b1) from the SAME AREA/Culture is regarded as "Samara Hunter Gatherer" instead of EHG? And what is "Samara HG" anyway?

Also, if all Y-Haplogroups are derived from ANE/AG3, then why would the Y-Haplogroup J individual be part of "EHG" himself?


Davidski said...

Samara_HG is EHG from Samara.

But some of the info on that page is wrong anyway. Karelia_HG doesn't belong to R1a-M417. So I wouldn't rely on Wikipedia for this sort of stuff.

JuanRivera said...

EHG's CHG ancestry explains the J1 EHG sample. CHG inherited it from Dzudzuana ancestors, and passed it to EHG when it, WHG and large amounts of ANE met in the steppe to form EHG. It was even rarer than Q1a and I2a. About that other thing, is because of focusing on a specific area. Also, it appears that Samara EHG is the same individual with discrepant dating and Y phyllogeny.

Andrzejewski said...

@DAvidski "The earlier finding that Andronovo from South Siberia was predominantly a blond/light-eyed population probably did cause a bit of panic in some academic circles, but as it turns out now, the reason for this wasn't Andronovo's high level of Yamnaya ancestry, but probably admixture from (non-Indo-European) Middle Neolithic farmers living west of the steppe."

_____________

It seems that maybe the reason that Northwestern Caucasus speaking populations such as the Adyghe (Cherkes) https://images.search.yahoo.com/yhs/search;_ylt=A0geK957ORhc2yYAn5oPxQt.;_ylu=X3oDMTByMjB0aG5zBGNvbG8DYmYxBHBvcwMxBHZ0aWQDBHNlYwNzYw--?p=adyghe&fr=yhs-itm-001&hspart=itm&hsimp=yhs-001 look like modern Europeans has perhaps something to do with the almost even mix of Neolithic Anatolian Farmers with Caucasus HGs? Maybe the Nordicists were wrong all along and it actually happened to be Middle Eastern-derived lineages that are responsible for the so-called "Nordic" phenotype instead of the Yamnaya/Steppe ancestors?

But it seems as if the Siberian-descendant lineages like the Mari (Finno Ugric) possess overwhelmingly very Europoid phenotype, albeit with kind of a strong "Ladogan" or "Turanic" WSHG admixture to it.

https://images.search.yahoo.com/yhs/search;_ylt=AwrDQq5aOhhcFFIAZUMPxQt.;_ylu=X3oDMTByMjB0aG5zBGNvbG8DYmYxBHBvcwMxBHZ0aWQDBHNlYwNzYw--?p=mari+russia&fr=yhs-itm-001&hspart=itm&hsimp=yhs-001

So, if Steppe populations were largely ANE/EHG-CHG cline, it's likely that they look like modern Europeans, mainly Eastern Baltics?

It puzzles me why so many non-IE populations from the Caucasus to Lake Baikal look almost identical to modern Europeans. Does it have to do with the mdDNA actually? Hard to know. I DO know that the Steppe populations' mtDNA vary much: V, H, T, N, K, Z, R, HV, U, etc.

JuanRivera said...

Steppe folk certainly resembled modern europeans, although with some shift towards the south and the east. As for the looks, ANE was a descendant of ANS (represented by the Yana RHS samples), which in turn was 29.2% Tianyuan-like and the rest being similar to K14 and Sunghir. Also, there is steppe admixture in Central Asia and Southwest Siberia up to lake Baikal and Mongolia, being present as early as the Early Bronze Agge.

JuanRivera said...

EHG and Steppe shared U2, U4, C1, C4a3, C4a6 and C5 with Siberians. T, H, V, K, J, I and W, among others, were introduced by steppe populations to southern Siberia.

Andrzejewski said...

The are lots of things that fall between the cracks. One of them is the "Semitic question" i.e. where do the Semites come from? The original Natufians as we understand them were one of the first groups to transition from HGs into farming, circa 14,500ybp (around the same time of AG3). Somehow there was a mutual admixture between Natufian and Anatolian Farmers, the latter as I understand it were mostly autochthonous HGs transitioning to agriculture ~12,000 years ago with infusions of Natufians, Iran-related population and WHG from the Balkan?

Now, the story in the Book of Genesis about some mythical warlord figure named "Abraham" originating from Harran, which was inhabited by Hurrians, or even perhaps belonged to the Mittani kingdom 3,700 years ago (Hurrians ruled by a Yamnaya-derived "marianu" elite). Lots of bible researchers analysing the biblical stories of Genesis came to the conclusion that "Abraham" or whoever he represented was NOT a Semite but a Hurrian, and that many of the customs (e.g. passing a wife off as a sister, etc) derived from Hurrian folklore.

Now, modern geneticists conclude that the current Middle Eastern population is an admixture of the original "Semites" (=Natufians) with a strong demic diffusion of Iran_N/CHG. The Hurrians may as well be descendants of the Kura-Araxes culture. Modern Lebanese DNA is 92% Natufian + Iran_N.

Now, to complicate things even more, perhaps what happened after the Uruk migration is that those "Hurrians" carried J1 and J2 into the Levant, and that the aboriginal Y-DNA Haplogroups of E1b1b were replaced by Caucasus Hunter Gatherer/Iran_N DNA?

Some researchers theorize that the Amorites were actually Hurrians/Urartu/CHG who switched their speech into Semitic languages.

IF all these data point are correct and interconnected, it may turn out that a Yamnaya-style male migration of Hurrians/CHG altered the original Y-DNA of the ME. The only problem that underlies it all is a simple one: HOW DID THE SEMITIC LANGUAGE SURVIVE LIKE THE BASQUE INSTEAD OF THE INDIGENOUS "SEMITES" LANGUAGE-SHIFTING TO HURRIAN/Northwest Caucasian?

Andrzejewski said...

Now, in the Southern-Central European context: Is it plausible that following the Middle Neolithic Age some EEF were being replaced by CHG in Greece and Italy? Greeks's and Etruscan's most common y-dna is J1 or J2, which were likely carried from the Caucasus or East of Anatolia. (Minoans and Mycenaeans are both shifted to CHG). However, it turns out that there is a widespread discontinuity between mdDNA H subclades in Central Europe AFTER the LBK, and maybe it signals either a CHG or a Yamnaya/Steppe-based infusion and partial replacement?

JuanRivera said...

Replacement of H lineages was from both steppe and CHG, the latter localized to the south. The CHG input in southern Europe certainly began in the Neolithic, as evidenced by Pelopponense_N. As for Semitic, they survived because they were also desert pastoralists, giving them a base from which expand.

Andrzejewski said...

This guy, from some TV show on NatGeo looks like a CHG rep sample, doesn’t he?

https://scontent-lga3-1.xx.fbcdn.net/v/t1.0-9/fr/cp0/e15/q65/48394614_194033464885135_7968672546205728768_o.jpg?_nc_cat=102&efg=eyJpIjoidCJ9&_nc_ht=scontent-lga3-1.xx&oh=72cd9af8aed74777ba66dacc54a78af7&oe=5CA9911A

Ric Hern said...

Remember Central Europe was influenced by GAC after LBK...

Andrzejewski said...

The GAC (Globular Amphora Culture) was the founding upon the CWC was installed. I am starting to think that the Corded Ware was a natural growth of the preceding Stredny Stog and perhaps even stretching back as far as the Bog-Dniester and Dniester-Donets. R1a1 was indigenous to its area and spanning all the way to contemporary Poland with its own Satemized version of the Para-PIE language(s) while Yamnaya was R1b (what happened to all the Samara Culture abundant R1a1, I wonder), Centum, and it later gave rise to Bell Beaker.

Ric Hern said...

"Samara Culture abundant R1a1" ?

Andrzejewski said...

@Ric Hern There were found specimens with R1a1 although the overwhelming majority of y Haplogroups (due to founding effect) were R1b (like Yamnaya).

Ric Hern said...

My personal view is that R1a was more abundant in the Forest Steppe to the North. R1b was more the Open Plains guys. Contact zone of R1a and R1b I was in the area where Forest gave way to Grassland.

Andrzejewski said...

@Ric Hern do you agree that both R1a1 and R1b were originally speakers of some variations of a PIE language, or do you think R1b was “first” and R1a1 was “Indo-Europeanized”? What’s your opinion about the transformation of Samara —> Khvalynsk Culture and the role CHG elements might’ve played in it?

Ric Hern said...

Maybe the split between R1a and R1b happened near the Lower Don with R1a moving up the Don and R1b migrating to the Balkans along the Black Sea Coast.

Andrzejewski said...

@JuanRivera Which might indicate the practice of exogamy was very widespread and prevalent among Steppe populations. Do you believe that this is the way that the CHG element became so central over time within Yamnaya?

Ric Hern said...

I think neither R1a or R1b spoke PIE before they started to mingle during the Eneolithic. When you bake a cake you have to have all the ingredients otherwise it will not be a cake. Heheheeh...

Davidski said...

R1a, and especially its R1a-M417 subclade, doesn't look like it was common anywhere before the Indo-European expansions. It certainly was nowhere near Poland until the Corded Ware people got there.

R1b was much more common in the steppe and forest steppe throughout the Mesolithic, Neolithic and Eneolithic, all the way from the Baltic to Central Asia.

Ric Hern said...

My personal feeling is that Women were mostly the binding factor within the Steppe Cultures. Women like to keep in touch with their Mothers and Sisters...

George said...

Greek language have Anatolian substratum, in central Greece placenames, so possibly Minoan was related to Anatolian.
While western Corded Ware in Holand is ancestral to Bell Beaker culture, and Basque are almost identical to western Bell Beaker.
According to Grigoriev theory the Indoeuropean languages came to Northern and Central Europe not earlier than late Bronze age.

George said...

Indoeuropean Migration in northern/central Europe 1400 BC according to Archeologist Grigoriev
s001.radikal.ru/i196/1304/06/cee1935a25f8.jpg

Also pre-Celtic Bronze age Britain is almost identical to eastern Corded Ware, they are possibly the ancestors of Picts that are possibly related to Basque , and it is considered that German languages have substratum that is related to North Caucasian or to Basque

George said...

While in Southern Europe, for example Tyrhenians are related to other Nostratic people
the words from Swadesh list of Tyrhenian language are close to Kartvelian and Indoeuropean

Davidski said...

@George

Eastern Corded Ware was closely related to Trzcieniec culture and to modern Balto-Slavs. Everything matches, including the subclades of R1a and mtDNA. See here...

Modern-day Poles vs Bronze Age peoples of the East Baltic

It shows no relation to Basques at all.

Also, Picts were Indo-Europeans. So Grigoriev was obviously wrong about a lot of things.

George said...

I didn't wrote that, I wrote what Basque are identical to Western Bell Beaker, and ancestral to Bell Beaker is western Corded Ware from Holland.

According to Grigoriev the monuments and structures in late Sintashta(not in early Sintashta) and in Irtysh Siberia late Bronze age have more ancient analogs in northern mesopotamia and in northwestern Iran and southern Azerbaijan (Uzerlik-tepe), later this type of structures are apearing in Central Europe

Davidski said...

@George

It's like this...

Corded Ware > Trzciniec culture > R1a-M417(Z282+) Balto-Slavs

Corded Ware > Sintashta culture > R1a-M417(Z93+) Indo-Iranians

There's no need to reinvent the wheel based on outdated archeological data.

George said...

Also in west Eurasia Tyrhenians Kartvelians and Indoeuropeans have the "m-T" pronoon isogloss,
The Basque and North Caucasians don't have this isogloss

George said...

Grigoriev is modern archeologist

Dragos said...

Davidski
The Balto-Slavs are said to form in the Sosnica-Trziniec-Komarov complex.

Davidski said...

@George

Grigoriev is modern archeologist

But his views are outdated, because they're totally at odds with the latest ancient DNA data.

And you can't tell me that Corded Ware wasn't Indo-European and was more closely related to Basques than to Balto-Slavs and Indo-Iranians. That's like arguing that up is down and down is up.

George said...

According to ancient dna Iron age and Late Bronze age Northen Pakistan is mostly E1b1 with some J1 J2a G2, in cultures that are early Indo-Aryan
The E1b1 in this cultures prooving the Grigoriev theory that Indo-Iranians are from northen Messopotamia.
Also the Indoeuropean languages have Semitic adstratum, so it is another plus for Northern Mesopotamian theory.
E1b1 is also in early Iron age in northern Kazakhstan "Central Sacae" another plus to that theory

George said...

It is like Hungarian is not Slavic, but Hungarians are almost Slavic people genetically
and Germanic and Balto-Slavic are not CWC, but they are close genetically

Davidski said...

@George

That sounds like a fringe theory.

The fact is that Balto-Slavs and Indo-Aryans are related linguistically and genetically, and E1b1 has nothing to do it.

So those ancient E1b1 samples weren't Indo-Aryans. Indo-Aryans are derived from R1a-rich Sintashta.

See here...

The mystery of the Sintashta people

George said...

We should compare those Indo-Iranians that geneticaly didn.t or less affected from White Huns Hephtalites, South-Central Asian Turks , Moghols and others from southcentral Asia,
People like the Makrani Balochi and Zoroastrians from Pakistan,
If you compare them, with Slavic people you will see that only genes(autosomal) in common they have are west asian genes

And also the separation of Baltic and Slavic is 3100 years ago, at that time there is western Asian non European Neolithic branch J2a-CTS900 in Kyjatice Culture 1,110–1,270 B.C.E.
And E1b1 J2a J1 G2 among early Iranians and Indo-Aryans are also western Asian

Folker said...

Just a reminder to some: we don’t have Anatolian names from 2500 BC. Some names found in tablets wrote in cuneiform in a semitic language could be read in a way that they look like constructed with radicals found in names found in Anatolia centuries later, some among Hittites. The corpus of these names is extremely small (less than 20 with perhaps only 9 different). Even if (which remains to be proved) those names where from an IE language, it has diverged from the main Anatolian branch centuries before (read the paper). The Anatolian branch is only attested after 2000 BC, with Hittite, and later with Luwian and Palaic. Geographic extent of IE languages in Anatolia before around 1400 BC is unknown. We could say that around 2000BC Hittite was present in Central Anatolia around Kadesh (with Luwian and Hattic), Luwian likely in the Konya plain, and Palaic in the NW, near the Black Sea. The Luwian kingdoms are only attested later, so can be the result of later expansion (linked or not to Hittites older kingdom).
By the way, cremation is considered as intrusive and linked with Balkans.
On another point: at some point, Anatolian languages could have evolved in close vicinity of some IE languages from a more recent split of PIE. Some common features are difficult to explain in another way.
Last, but an important point: since the EBA, Anatolia was a crossroad between Balkans and Middle East. Trade routes were not empty.

Slumbery said...

@George
"And also the separation of Baltic and Slavic is 3100 years ago, at that time there is western Asian non European Neolithic branch J2a-CTS900 in Kyjatice Culture 1,110–1,270 B.C.E."

How is that lonely J2a on the Hungarian plain connected to early Balto-Slavs in any way? You are bringing up facts that are true, but irrelevant to the question at hand.

George said...

It looks like there is another George on this blog.

Andrzejewski said...

@George “According to ancient dna Iron age and Late Bronze age Northen Pakistan is mostly E1b1 with some J1 J2a G2, in cultures that are early Indo-Aryan
The E1b1 in this cultures prooving the Grigoriev theory that Indo-Iranians are from northen Messopotamia.”

Pakistan is not a European country. The Indo-Aryan elite is greatly outnumbered by natives (Dravidians, Elamite-related, BMAC), all of them came from Iran_N Zagros mountains. E1b1 can readily indicate a Muslim invasion.

George said...


@Slumbery
Autosomally and by IBD segments the J2a BR2 from Hungary 3200BP is closer to Polish and Lithuanians than the samples from Corded Ware
http://www.tropie.tarnow.opoka.org.pl/images/br2_luz.png

Davidski said...

@George

The Balto-Slavic paternal marker is R1a-M417.

It comes from Corded Ware. So Corded Ware was Proto-Balto-Slavic.

There were no other later migrations that had any real impact on the Balto-Slavic gene pool.

Slumbery said...

@George
"Autosomally and by IBD segments the J2a BR2 from Hungary 3200BP is closer to Polish and Lithuanians than the samples from Corded Ware."

1. The heat map you linked does not show that, because it does not show any comparison with CWC. (Not to mention that Wales and some parts of Italy have the same color as Lithuania and that suggest a rather mixed genome-wide ancestry.)
2. So what?
BTW, YHg J2a is rare among modern Polish and pretty much nonexistent among Lithuanians.

Cpk said...

@folker I have not seen any linguist rejecting Indo-European origin for those names. Please direct me to a source if there is one.

Folker said...

A small correction to my post: Kanesh not Kadesh.
By the way, to the ones ignorant of the discussion, the tablets were found in Ebla, and the names were found in some tablets connected to Armi (another city, of unknown location). The paper is from Kroonen.

George said...

@Andrzejewski which muslims 3000 years ago? And from muslims are mostly J1c2d.
And yes there is almost no E1b1b1b2 in Northern Pakistan today
Here below early Iron age early Indo-Aryan cultures
Mostly E1b1b1b2(is it from early Semitic influence to Indoeuropepeans in Messopotamia?)
and also some L1a J2a1 G2a E1a Q1b2 and some local R2a and H1a



I8220 I8220 970-550 BCE Pakistan_IA_Aligrama_all Pakistan_IA_Aligrama_all Aligrama Pakistan .. L1a
I8245 I8245 970-550 BCE Aligrama_IA Aligrama_IA Aligrama Pakistan .. R2a3a2b2b1
I8246 I8246 970-550 BCE Aligrama_IA Aligrama_IA Aligrama Pakistan .. G2a2a
I6548 I6548 1000-800 BCE SPGT Barikot_IA Barikot Pakistan H20a H1a1
I5396 I5396 904-817 calBCE (2715±20 BP, PSUAMS-2790) SPGT Katelai_IA Katelai Pakistan U4d J2a1
I5399 I5399 1000-800 BCE SPGT Katelai_IA Katelai Pakistan J1d R2a3a
S8194.E1.L1 I8194 1100-900 BCE SPGT Udegram_IA Khyber-Pakhtunkwa, Swat, Babozai tahsil, Gogdara Pakistan R30a1b H1a1a
S7725.E1.L1 I7725 1200-800 BCE SPGT Udegram_IA Khyber-Pakhtunkwa, Swat, Babozai tahsil, Udegram Pakistan M65a1 E1b1b1b2
S8191.E1.L1 I1799 1200-800 BCE SPGT Udegram_IA Khyber-Pakhtunkwa, Swat, Babozai tahsil, Udegram Pakistan T2a1b E1b1b1b2
S8195.E1.L1 I8195 1200-800 BCE SPGT Udegram_IA Khyber-Pakhtunkwa, Swat, Babozai tahsil, Udegram Pakistan U8b1a2b E1b1b1b2
I5400 I5400 927-831 calBCE (2745±20 BP, PSUAMS-2793) SPGT Loebanr_IA Loebanr 1 Pakistan M5a Q1b2
I6554 I6554 831-796 calBCE (2645±20 BP, PSUAMS-2796) Loebanr_IA_father.I6292 Loebanr_IA_father.I6292 Loebanr 1 Pakistan T2g1 L1a
I6555 I6555 906-820 calBCE (2720±20 BP, PSUAMS-2797) SPGT Loebanr_IA Loebanr 1 Pakistan M4 L1a
S10001.E1.L1 I10001 1300-1000 BCE SPGT Loebanr_IA Loebanr 1 Pakistan R30b1 L1a
S10974.Y1.E1.L1 I10974 900-800 BCE SPGT Loebanr_IA Loebanr 1 Pakistan .. L1a
S8997.E1.L1 I8997 900-800 BCE SPGT Loebanr_IA Loebanr 1 Pakistan W3a1b R2a
I1799 I1799 1044-922 calBCE (2830±20 BP, PSUAMS-2632) Udegram_IA_son.I13262 Udegram_IA_son.I13262 Udegram, Babozai tahsil, Swat District , Khyber-Pakhtunkwa Province Pakistan T2a1b E1b1b1b2
I1985 I1985 1192-939 calBCE (2880±30 BP, Beta-428667) SPGT Udegram_IA Udegram, Babozai tahsil, Swat District , Khyber-Pakhtunkwa Province Pakistan M E1b1b1b2a
I1992 I1992 1195-978 calBCE (2890±30 BP, Beta-428665) SPGT Udegram_IA Udegram, Babozai tahsil, Swat District , Khyber-Pakhtunkwa Province Pakistan H2a2 E1a
I3262 I3262 976-832 calBCE (2760±25 BP, PSUAMS-2157) Udegram_IA_father.or.son.I1799 Udegram_IA_father.or.son.I1799 Udegram, Babozai tahsil, Swat District , Khyber-Pakhtunkwa Province Pakistan H14a E1b1b1b2a
I6899 I6899 1044-830 cal BCE (2785±45 BP, CEDAD LTL13328A) SPGT Udegram_IA Udegram, Babozai tahsil, Swat District , Khyber-Pakhtunkwa Province Pakistan .. E1b1b1b2
I6900 I6900 1400-1126 cal BCE (3018± 45 BP, CEDAD LTL13327A) SPGT Udegram_IA Udegram, Babozai tahsil, Swat District , Khyber-Pakhtunkwa Province Pakistan .. E1b1b1b2

Ric Hern said...

@ George

In the British Isles 90% of the earlier population were replaced by Bell Beaker people from Central Europe via the Netherlands. The MtDNA shows influence from the GAC and LBK Cultures if I remember correctly, while Basque are closer to Sardinians than to Northwestern Europeans. There were two different Neolithic migrations towards the West. One following the Mediteranean Coastline from which the Basque descends and one to the Northwest from where Central and Northwestern Europeans descends...

So nope very little Basque/Sardinian related ancestry survived North of the Pyrenees Mountains.

George said...

@Slumbery
For example one Russian(?) man with hg N1c1 his comparision with IBD segments
http://forum.molgen.org/index.php/topic,7120.msg267375.html#msg267375


BR2 82,25 ---//Confidence: very low
Balt 70,29 ---//Confidence: high
Ukrainian-West-and-Center 68,64 ---//Confidence: high
Ukrainian-East-and-Center 65,92 ---//Confidence: high
Russian-South 65,69 ---//Confidence: very high
Russian-West 64,96 ---//Confidence: very high
Belarusian 64,31 ---//Confidence: very high
Veps 62,81 ---//Confidence: medium
Finnish 61,41 ---//Confidence: low
Polish 59,81 ---//Confidence: high
Moksha 59,58 ---//Confidence: low
German 59,25 ---//Confidence: low
Russian-North-East 57,58 ---//Confidence: very high
Russian-North-Kargopol 57,49 ---//Confidence: very high
Estonian 56,17 ---//Confidence: high
Karelian 53,07 ---//Confidence: high
Swedish 51,06 ---//Confidence: high
Hungarian 51,01 ---//Confidence: very high
Croatian 50,16 ---//Confidence: very high
Komi 49,85 ---//Confidence: high
British 48,97 ---//Confidence: very high
Loschbour 48,79 ---//Confidence: very low
Erzya 48,26 ---//Confidence: medium
Bulgarian 47,79 ---//Confidence: high
Norwegian 47,7 ---//Confidence: high
Ust-Ishim 44,96 ---//Confidence: very low
Romanian 44,81 ---//Confidence: medium
Saami 44,58 ---//Confidence: high
Chuvash 44,14 ---//Confidence: very high
French 43,57 ---//Confidence: very high
NE1 42,02 ---//Confidence: very low
Mari 40,44 ---//Confidence: high
Spanish 40,38 ---//Confidence: very high
Tatar-Kazan 39,84 ---//Confidence: high
Tatar-Crimean 39,1 ---//Confidence: low
Udmurt 38,81 ---//Confidence: high
Ashkenazi 38,29 ---//Confidence: high
Italian 38,23 ---//Confidence: very high
Tatar_Lithuanian 37,47 ---//Confidence: very low
Greek 36,62 ---//Confidence: very high
Adygei 36,48 ---//Confidence: high
Mansi 36,38 ---//Confidence: medium
Basque 36,02 ---//Confidence: very high
Chechen 35,82 ---//Confidence: very high
Bashkir 35,59 ---//Confidence: low
Kumyk 35,24 ---//Confidence: high
Turkmen 34,68 ---//Confidence: high
Ossetian 34,47 ---//Confidence: high
Sardinian 34,33 ---//Confidence: very high
Sicilian 34,3 ---//Confidence: very high
Balkarian 33,92 ---//Confidence: very high
Nogay 33,92 ---//Confidence: high
Sephard 33,11 ---//Confidence: high
Lezgin 32,74 ---//Confidence: very high
Italian-South 31,6 ---//Confidence: high
Abkhazian 31,02 ---//Confidence: very high
Turkish 30,62 ---//Confidence: very high
Armenian 30,56 ---//Confidence: high
Georgian 29,38 ---//Confidence: very high
Cypriot 28,9 ---//Confidence: medium
Azerbaijani 27,71 ---//Confidence: high
Uygur 27,38 ---//Confidence: medium

Ric Hern said...

@ George

I wonder if you noticed that those E1bs although many were only from a few specific sites. It is almost like sampling an Armish community and expect huge differences in lineage....

George said...

@Ric Hern
Basque autosomaly are a mix of Neolithic and plus CWC/Yamnaya
Sardinians are a mix of Neolithic and plus West Asian

Pre-Sardinian language is unknown, but some consider that it was close to Tyrhenian

And I want to remind again,
In west Eurasia Tyrhenian Kartvelian and Indoeuropean have the "m-T" pronoon isogloss, but Basque and NorthCaucasians don.t have

Slumbery said...

@George
OK, but what do you hope to prove with this? It has significant North-Eastern genome-wide ancestry apparently, but it is still a single J2a from site that is very peripheral for the early Balto-Slavs and J2a is still rare among Slavs and super rare among Balts.

George said...

@Slumbery
N1c1 is also very rare among Hungarians, almost absent

Folker said...

@Cpk
As said previously, those names are identified as IE by Kroonen in the paper found as a supplement of Damgaard. This is not the main subject of the paper. There is likely a publication coming of the subject. The problem is: no paper, no possibility to object. There are many points which are needing some discussion:
- the fiability of transcription. Kroonen is using the works of several specialists. But as said, the clerk spoke a semitic language and the tablet is wrote in his language, not in the language of the people bearing the possible IE name.
- the corpus is very small. Are a dozen different names of unknown signification enough to identify a language?
- other names can be found in tablets, with a real phonetic proximity with those names. They are not cited as IE by Kroonen.
And so on.

«Oldest ‹Older   1 – 200 of 392   Newer› Newest»