search this blog

Thursday, October 26, 2017

Tollense Valley Bronze Age warriors were very close relatives of modern-day Slavs


This is strongly suggested by the Principal Component Analysis (PCA) below, which shows that many of the Tollense Valley warriors (Welzin_BA) cluster in the Slavic-specific part of the plot. The relevant datasheet is available here.

I designed this PCA with the sole purpose of using Balto-Slavic-specific genetic drift to differentiate Slavs from Germans, except of course those Germans with a lot of Slavic ancestry, who are usually from eastern Germany and Austria. I can assure you, people who don't harbor significant Slavic ancestry never cluster in this part of the plot.

The only other ancient samples that cluster in the Slavic zone are, as expected, an early Slav from Bohemia and, interestingly, a Bronze Age individual from what is now Hungary. But we've already seen strong genetic, and indeed genealogical, links between another Hungarian Bronze Age genome and present-day Slavs (see figure 3 here).

So what's going on? Did the proto-Slavs come into existence during the Bronze Age, as opposed to the more generally accepted early Medieval Period? And did they expand from what is now Hungary? Or did they migrate there from the Baltic region? Thanks to Matt in the comments for the table below.


See also...

Tollense Valley Bronze Age battle: preliminary ancient DNA analysis

Genetic and linguistic structure across space and time in Northern Europe

111 comments:

Lukasz M said...

Can you add all numbers to WEZ samples?

Davidski said...

I'm having a beer. Maybe someone else can. The datasheet is here.

https://drive.google.com/file/d/0B8XSV9HEoqpFdURNZFFRbExLZHM/view?usp=sharing

Lukasz M said...

Maybe I will add:)You use only samples from WEZ_Modern_Ancient plink files from the authors?

Davidski said...

Yeah, from those PLINK files.

zardos said...

An open question is where and how Proto-Germanic came from. Because my guess was even before that a more Northern-Western area.
Did the studies say something about the cultural affiliations of the Tollense ware parties?

Lukasz M said...

Really WEZ54 and WEZ57 which were considered southern mercenaries are in or near Germanic cluster??

Davidski said...

WEZ54 is a total outlier and shouldn't be on this plot. He doesn't cluster with anyone here and might well be from Southwestern Europe.

WEZ57 might be a warrior from France, no? He's quite Celtic-like anyway.

zardos said...

Your plot could be interpreted as if there was a battle between a Slavic-like and a Germanic-like group, no? Most of the individuals cluster clearly in one of the groups with virtually nothing in between.

Davidski said...

We'd need to see the archaeological, including isotopic, reports for each individual before considering such things.

zardos said...

Sure, more data is needed. But if you plot is reliable, the variation seems to be to wide for one single group. So either we deal with two ethnicities along those lines (Germanic-like vs. Slavic-like) or with mixed alliances on both sides, which I think is much less likely. Single mercenaries are one thing, but that would be another.

Michał said...

@Davidski

Have you tried to place those Welzin samples on your Polish aDNA PCA plot?

The key distinguishing feature for WEZ is an increased shared drift with WHG, so any PCA that does not include any WHG/SHG/EHG samples is unlikely to differentiate between WEZ and modern Central European or Western European populations.

RISE569 is an Early Slavic sample from Bohemia that (in contrast to RISE568, not used in your analysis) seems to have a very strong Germanic admixture, as shown on Łukasz’s maps (see links below), so this should not be used as a good example of an unadmixed Early Slav.
https://fusiontables.googleusercontent.com/embedviz?q=select+col39%3E%3E1+from+1VfCtRKDni3U3FzwRmuQ-4tv2Zhw2v3kYh1suJI7a&viz=MAP&h=false&lat=50.39422413594516&lng=40.427433437499985&t=1&z=4&l=col39%3E%3E1&y=2&tmplt=2&hml=KML
https://fusiontables.googleusercontent.com/embedviz?q=select+col39%3E%3E1+from+1yEM1thWij5HFRgQ0OBHzfTEuq_bM4WEURLxNAASu&viz=MAP&h=false&lat=49.943859680927034&lng=34.758488124999985&t=1&z=5&l=col39%3E%3E1&y=2&tmplt=2&hml=KML

On your plot, most of the Welzin samples are placed around the “border region” between the Germanic and Slavic samples, and this is also where all Hungarian samples (marked as “Finno-Ugric”) have landed, so this actually shows most similarity between WEZ and the Hungarians (rather than Slavic or Germanic people). Importantly, the vast majority of those Welzin samples are located “west” of the mixed (Slavic-Germanic) RISE569 sample. Also, I am pretty sure that most of those Slavic (stateblue) samples located “west” of RISE569 (and thus close to the majority of WEZ samples) are from Slovenia and Slovakia (ie. from Hungarian-like populations), so if not counting a single Baltic-like sample (WEZ58), this leaves us with only three Welzin samples that landed “east” of RIS569 (and “east of the Germanic-Slavic border”) and thus are likely to show some similarity to modern Polish people (in this particular type of analysis), although all this may simply be an artefact of using an inappropriate set of reference samples.

Taymas said...

I've long felt unsatisfied with my understanding of what was going on with those Bronze Age Hungary samples on the PCA. (A) For such a small geography they vary so much. (B) They form a pretty straight line. (C) Yet this cline doesn't end at Hungary_CA. (D) Nor does the other end of the cline end quite at Beakers. Maybe this is all unsurprising to someone more informed. The samples are from a while ago, I haven't seen anyone re-analyze them in light of all the new ancient gene evidence.

Arza said...

@ Michał
The key distinguishing feature for WEZ is an increased shared drift with WHG

"I designed this PCA with the sole purpose of using Balto-Slavic-specific genetic drift to differentiate Slavs from Germans"


RISE569 is an Early Slavic sample from Bohemia that seems to have a very strong Germanic admixture

Totally unsupported claim. In David's PCA spreadsheet for Western Eurasia RISE569 clusters with Croatians and Slovenians.

Additionally now we know that such autosomal composition existed 2000 years earlier. What is even more important it existed in the same place as birthplace of RISE569, as another study showed that Bohemia and Moravia are very strong candidates for a source of some of the warriors:

"Multi-isotope proveniencing of human remains from a Bronze Age battlefield in the Tollense Valley in northeast Germany" https://doi.org/10.1007/s12520-017-0529-y


as shown on Łukasz’s maps

On which type of analysis those maps are based?


most of the Welzin samples are placed around the “border region” between the Germanic and Slavic samples

What now is a border between Germanic and Slavic samples once was a West Slavic territory.


the vast majority of those Welzin samples are located “west” of the mixed (Slavic-Germanic) RISE569 sample

Once again, she's not mixed, and modern Poles are shifted North-East due to historical reasons.


this leaves us with only three Welzin samples

This leaves us with a whole bunch of possible West Slavs proper and proto-South Slavs in case of the ones with a Balkan vibe.

Synome said...

I think it would be a mistake to label these people "Proto-Slavs". The Slavic expansion involved genetic input, but it also certainly involved language shifts. So modern day Slavic speakers are not an accurate genetic representation of the earliest Slavic speakers. Even the Early Bohemian Slav samples from the 7th century or later may already be admixed or even cultural adopters of the Slavic ethnicity.

What we really need are samples from the Kiev culture. It is the best agreed upon candidate for the location of the pre expansion Slavs. Even then, genes aren't language. Many people could have adopted Slavic language and culture without a large input from Kiev culture migrants.

In the case of the Kurgan hypothesis for PIE, all of the genetic, linguistic, archaeological, and historical evidence generally points the same way and confirms the basic claims of a Bronze Age expansion from the Pontic-Caspian steppe. If you try to put the Slavic expansion further back than the 6th century CE based only on broad genetic continuity, you run into huge contradictions from history, linguistics, and archaeology.

Michał said...

@ Arza

"I designed this PCA with the sole purpose of using Balto-Slavic-specific genetic drift to differentiate Slavs from Germans"

Just take a look at Figures 4.24 and A4.8 in the original Welzin paper, and you will see that when using an appropriate set of ancient reference samples (including WHG, SHG and EHG) one will get a clear separation of Welzin from all modern Central-European and Western European populations (including both Slavic and Germanic people). Also, I am pretty sure that when Davidski adds these Welzin samples to his Polish aDNA PCA plot, they won’t cluster with any Slavs or Germans anymore.


Totally unsupported claim. In David's PCA spreadsheet for Western Eurasia RISE569 clusters with Croatians and Slovenians.

Are you suggesting that modern Croatians and Slovenians are the best model for the Early Slavs (or Proto-Slavs)?
BTW, I know from Łukasz that Davidski has already admitted that RISE569 might indeed be mixed.


On which type of analysis those maps are based?

The author is posting here (as Lukasz M), so he may explain it to you. Here is a link to the relevant discussion on the Anthrogenica forum:
http://www.anthrogenica.com/showthread.php?10347-(Unofficial)-oracle-for-Eurogenes-K36-256-reference-populations


What now is a border between Germanic and Slavic samples once was a West Slavic territory.

Firstly, I was talking about a “virtual” border on the PCA plot, not about a real border. Secondly, the Slavs have only arrived at that modern Slavic-German border in the 6th-7th century AD.


modern Poles are shifted North-East due to historical reasons.

Agreed, and this is exactly the result of a very strong (and relatively recent) “Slavic admixture”.


This leaves us with a whole bunch of possible West Slavs proper and proto-South Slavs in case of the ones with a Balkan vibe.

You are apparently not aware that there is only one WEZ sample (WZ40) that clusters with modern Poles on that PCA plot by David, and this is actually a sample for which the f3 outgroup analysis shows more genetic drift with modern people from Ireland, Austria, Basque country, Germany and Finland than with those from modern Poland (although no statistically significant differences are seen, of course). See this Anthrogenica thread for more details:
http://www.anthrogenica.com/showthread.php?12405-Addressing-challenges-of-ancient-DNA-sequence-data-obtained-with-next-generation-meth&p=301683&viewfull=1#post301683

Slumbery said...

Davidski

We have 20 samples.
- 8 of them are outside of the modern Slavic range.
- 8 of them outside the modern Germanic range.
- 17 of them are in the section of Slavic-Germanic and the parts of both that are close to the other.
These are just as Germanic as Slavic. (And strictly speaking possibly neither.)

Synome said...

I should really have put the earliest Slavic expansion in the 5th century. This predates the widespread adoption of the Slav ethnonym, but it lines up with the archeological descendants of the Kiev culture.

In my opinion, the most probable archaeological sequence all the way from late PIE to Slavic is thus:

Yamnaya--Late PIE
Corded Ware--Indo-Slavic?
Trzciniec--Proto Balto-Slavic
Chernoles--Early-Proto Slavic
Zarubinets--Middle-Proto Slavic
Kiev-- Late Proto-Slavic.Expansion.
Prague-Korchak, Penkovka, Kolochin--early Slavic cultures

The Slavic expansion may have involved quite a lot of genetic input, but it would have happened during the migration period and not earlier.

supernord said...

aah, there are not numbers at all WEZ red points, eeh

Michał said...

@Synome

While I agree with your general concept, I doubt your scenario assuming Trzciniec origin for Chernoles is sufficiently supported by archaeology. Therefore, I would rather replace Chernoles with Milograd/Pidhirtsi, a Trzcieniec-derived culture/population that apparently contributed to both Zarubintsy and Post-Zarubintsy horizon, so this would be the major link between the "Proto-Balto-Slavic" Trzciniec-Sosnica-Komarov horizon and the Proto-Slavic Kiev culture.

Chad Rohlfsen said...

Highly unlikely these are Slavs. That's not until the 6th Century here.

Matt said...

Nice PCA Davidski.

It seems like it could, tentatively, be more sensitive at picking up the relationships were ancients are actually ancestral to different populations in Europe.

Even though at the same time PCA based on very deep ancients and a big West Eurasian / world cross section, are more sensitive for picking up the differences in the big WHG:EHG:IranN:LevantN proportions between ancients and their modern descendants. Those proportions are what drive the big, general patterns in how Europeans are related to non-European West Eurasians, or even world populations. (But sometimes you want to understand how people relate to and descend from the proximate, recent ancestors of the Bronze Age and later...)

Few graphics: https://imgur.com/a/PD8bc

(another version of PC1 vs PC2 figure in Davidski's main post, with different labeling (the pops in his .dat file), some neighbour joining trees and finally, some distance calculations using all 10 dimensions.)

Matt said...

It seems like in terms of ancients affinities to moderns there are some broad groupings of the ancient populations:

Central European affinity: https://i.imgur.com/2Ba2IUj.png - Closest to West Slavs (Czech, Slovak, Slovene, West Ukrainian), Hungarian, Austrian and East German = WelzinBA, HungaryBA, Slav_Bohemia.

North Germanic affinity: https://i.imgur.com/Vb7KONq.png - Closest to North Germanics, then West Germanics = Nordic_IA, Nordic_LN, Unetice

British and Irish Isles affinity: https://i.imgur.com/gGObAIF.png -
Closest to BI, Irish and North Sea populations = Bell Beaker Germany, Ireland EBA, England IA, England Roman, England Anglosaxon (though England Anglosaxon could arguably also be put in North Germanic grouping and Bell Beaker Germany is more generally West European).

Steppe_EMBA: https://i.imgur.com/Mv11emK.png - No clear longitude affinity and closest to a mix of North European groups from East-West = Yamnaya Samara, Poltavka, and, marginally, Sarmartian Pokrovka.

Steppe_MLBA and Corded Ware Germany: https://i.imgur.com/uVW6fMU.png - Similar patterns to North Germanic group, but slightly more Central and less Western = Corded Ware Germany, Srubnaya, Potapovka (and Hungary_IA only for a lack of a better place to put it).

Scythian_Samara: https://i.imgur.com/sA9SQfB.png - Russian and Finnic affinities, possibly relating to presence of Siberian ancestry.
No ancient population, as a whole, in this analysis seems to have a Baltic character, though as an individual Welzin_BA_outlier_2 has some of this character.

Some concern there may potentially be overfitting in the dataset, but certainly the general patterns seem pretty sensible...

Synome said...

@Michal

That's fair. I prefer to see Milograd as Proto-Baltic, but I think your scenario is plausible too. I think Chernoles-Milograd is the time and place of most ambiguity.

Lukasz M said...

First of all. The SNP count is so miserable in those samples (few thousands snps in many, max 22000, 10 000 on average more or less), that any definite conclusion is impossible, it's only loose prediction.

I wonder if Davidski knows something about next samples. Will be better?

supernord said...

WEZ56 is not close to Baltic, it is close to CWC (more EHG?). It is archaic subject.
It is evident that the Slavs are direct descendants of the Lusatian culture.

Arza said...

Just take a look at Figures 4.24 and A4.8 (...)

And? We are talking here about the drift shared with Balto-Slavs and not about general EHG/CHG/WHG/EFF proportions.


Are you suggesting that modern Croatians and Slovenians are the best model for the Early Slavs (or Proto-Slavs)?

Huh?!? Maybe you should reread what I wrote.


Davidski has already admitted that RISE569 might indeed be mixed

In my ranking of trust David is on the second place. His spreadsheets and stats are on the fist one. David always can change his mind, aDNA won't change.

Oh, and there is a little detail - since yesterday we know for sure that medieval Germanic admixture is not the only option, nor even probable.


so he may explain it to you.

I wish that you're unable to do so.


I was talking about a “virtual” border on the PCA plot

So did I.


the Slavs have only arrived at that modern Slavic-German border in the 6th-7th century AD

This is not even an argument, but another unsupported statement.

Agreed, and this is exactly the result of a very strong (and relatively recent) “Slavic admixture”.

YEAH, Polish-Lithuanian Commonwealth was osom.


You are apparently not aware (...)

I'm looking now at never-ending animation of 10 dimensions projected at once on a 2D plane, so...


that there is only one WEZ sample

We are talking here about things that happened on a western and southern fringes of a Slavic world over 3200 years ago and you expect that they will align perfectly with modern Poles?

And why they should in the first place if isotopic analysis shows that apparently they were from around Moravia?

*cough* Kyjatice *cough*

Matt said...

Few more of these graphics with the dataset Davidski posted here -
https://drive.google.com/file/d/0B8XSV9HEoqpFbFBQc2xGRWdleFU/view?usp=sharing

Here: https://imgur.com/a/gnYYJ

This could be useful for some nMonte models, though I guess it will be less accurate for deep ancestry proportions.

Steven said...

So Slavic-like ancestry in East Germany could be ancient rather than from recent migrations?

Davidski said...

@Michal

Have you tried to place those Welzin samples on your Polish aDNA PCA plot?

Not much point in doing that, as that PCA is mostly driven by broad ancient ancestry components, like ANE, CHG, WHG and basal. So the Welzin samples will simply be attracted to other samples with similar ratios of these components.

The PCA above is mostly based on Balto-Slavic specific drift, so even samples that have different ratios of ANE, CHG, WHG etc. cluster in the same Slavic zone if they have Slavic ancestry.

Samuel Andrews said...

@Rob,

But I thought everyone in eastern Europe before Slavic expansions was like Hungary BA.

Davidski said...

Also, Michal, formal stats tend to focus on deep phylogeny as well, rather than recent drift, even f3 stats do that, so they're not very good at picking up recent ethnic links.

Think of the PCA above as a type of ethnic rare allele analysis, and my usual West Eurasian PCA as a type of formal stats analysis.

Rob said...


@ Sam

"But I thought everyone in eastern Europe before Slavic expansions was like Hungary BA."

And why did you think that ?

Romulus said...

It makes sense. If you go by the record of Germanic expansions in this wikipedia article Germanics only started migrating to the area in 750-500 B.C., although it is very close.

https://en.m.wikipedia.org/wiki/West_Germanic_tribes

Davidski said...

@Steven

So Slavic-like ancestry in East Germany could be ancient rather than from recent migrations?

But this wouldn't explain the high level of very recent shared IBD between many Germans and Slavs, especially Poles, nor the young Slavic-specific clades of R1a in Germany.

So Germans do have a lot of Slavic ancestry, both from eastern Germany and Poland, but some of their eastern affinity might be due to Bronze Age ancestry from people like the more eastern Tollense warriors.

Davidski said...

This could be useful for some nMonte models, though I guess it will be less accurate for deep ancestry proportions.

Exactly.

Samuel Andrews said...

@Rob,
"And why did you think that ?"

I'm being sarcastic. Remeber last year you strongly pushed that before Slavic expansions everyone in eastern Europe had hardly any R1a and was something like Hungary BA in terms of EEF-WHG-Steppe ratio? But I argued the Slavs were one of many related peoples derived from Corded Ware who continuously inhabited Corded Ware territory after its end.

Michał said...

@Davidski

Also, Michal, formal stats tend to focus on deep phylogeny as well, rather than recent drift, even f3 stats do that, so they're not very good at picking up recent ethnic links.

Firstly, those Welzin samples are more than 3000 years old, so this is not about any recent drift with modern Poles/Slavs. Secondly, by making the above statement you seem to admit that the deep phylogeny of the Welzin samples was quite different from that of the modern Poles, and since I guess you assume that modern Poles constitute the least admixed Slavic population and resemble the Proto-Slavs quite well, this also means that those Welzin samples were neither (Pre-)Proto-Slavs nor Proto-Balto-Slavs nor any population closely related to them. Thirdly, since we know that the ancient Balts from the Late Bronze Age were already nearly identical to the modern Balts (at least autosomally), this makes it practically impossible to derive the Balts from any such EEF-rich Welzin-like population, so all this indicates that the Proto-Balto-Slavs were very similar to both ancient and modern Balts, and thus the Slavs are likely to have arisen as a result of a significant western (but non-Welzin, ie. not shifted towards WHG) admixture in a Baltic-like population. Thirdly, once we admit that the f3 stats do not suggest that Welzin shared any specific drift with modern Poles/Slavs (when compared to other modern European populations), there is practically nothing in those Welzin data that would suggest any specific association with the Slavs.

Davidski said...

@Michal

Many of the Welzin samples share very specific drift with modern-day Slavs and an ancient Slav that very few other samples do (except Slavic-admixed Germans and Bronze Age Hungarians).

There must be a very good reason for this, like, for instance, the Welzin samples being ancestral to modern Slavs.

And as you can see in this thesis I linked to, Poles top the f3 stats list with the Welzin samples, so there's no conflict with my results, although I'll reiterate again that f3 stats don't just focus on ethnic links, but all shared drift, so they're often difficult to interpret.

http://eurogenes.blogspot.com.au/2017/10/tollense-valley-bronze-age-battle.html

You really need to stop focusing on overall deep ancestry proportions here. Please re-read what I said, and also look at Matt's comments, to get a better idea of what my PCA is designed for.

As things stand, you basically have things backwards.

Michał said...

@Arza

*cough* Kyjatice *cough*

Kyjatice is actually yet another proof for the Slavs having arrived to Poland from outside. I have already discussed it elsewhere, so let me just quote what I have written on this subject:

This is additionally supported by the autosomal results for the Bronze Age Lusatian culture (HAL36 from East Germany) and for its southern (Slovakian/Hungarian) variant called Kyjatice culture (sample BR2 from Hungary). They both are strongly west-shifted, with HAL36 resembling modern Dutch and Norwegian populations and BR2 looking like a modern French sample.

Importantly, BR2 turned out to be a very good quality sample, for which it was possible to perform an additional analysis using the haplotype-based approach. It showed that although that BR2 male did not resemble the modern Poles, he shared the largest number of haplotypes with those Poles rather than with any other modern population. This in turn suggested that although BR2 seems to represent an extinct population (resembling Western and Southern Europeans rather than modern Central Europeans), his haplotypes (or rather the haplotypes of his relatives from the Kyjatice/Lusatian culture) have been best preserved in modern Poles. The final conclusion is that modern Poles descent from a BR2-like (Lusatian?) population and are (in a sense) his closest modern relatives, but this BR2-like (or Lusatian-like) ancestry constitutes only a relatively small fraction of their overall ancestry, with the other part being very different (so this is why modern Poles do not look like BR2 at all). Apparently, that "other part" of Polish ancestry must have been of Eastern origin (to counter-balance the westward shift) and likely corresponds to shared Slavic ancestry. Most importantly, those haplotypes inherited from BR2-like people (from the Lusatian culture?) are not common neither among Eastern nor Southern Slavs, as all those remaining Slavic populations show less affinity to BR2 than let's say people from Wales or Sicily. This additionally supports the conclusion that it could not have been Poland from where the Slavs expanded east and south, as in such case we should see more of those BR2-like haplotpes in other Slavic populations (rather than in Western or Southern Europe).

Arza said...

@ Michał
Thirdly, since we know that the ancient Balts from the Late Bronze Age were already nearly identical to the modern Balts (at least autosomally), (...) , so all this indicates that the Proto-Balto-Slavs were very similar to both ancient and modern Balts

No, this indicates that Proto-Balto-Slavs lived on the steppe and were equal to (late-)PIE. Balts and Baltic languages are derived from Baltic_BA, thus possibly whole Trzciniec horizon was already Proto-Slavic.

It's now pretty clear that there was no Balto-Slavic outside of the steppe and all later similarities arose due to convergence of both neighbouring groups that continued for millennia, practically until last century.

E.g. Endzelins model:
https://upload.wikimedia.org/wikipedia/commons/thumb/b/bd/Balto-Slavic_theories.svg/400px-Balto-Slavic_theories.svg.png

The Genetic History of Northern Europe
The statistic D(Lithuanian, Baltic_BA; X,
Mbuti) reveals significant positive results for many modern Near Eastern and
Southern European populations which can be caused by Lithuanians having received
more genetic input from populations with higher farmer ancestry after the Bronze Age
(Supplementary Information Table S8). As this applies to nearly all modern
populations besides Estonians, especially for Central and Western Europe, limited
gene-flow from more south-western neighbouring regions
is sufficient to explain this
pattern.


Do anyone talks about deep genetic links between Balts and Germanics? Is a Balto-Germanic language a thing? If not, then who lived to the south-west of Lithuanians?

Thirdly, once we admit that the f3 stats do not suggest that Welzin shared any specific drift with modern Poles/Slavs (when compared to other modern European populations)

Cool that you've finally admitted that Poles genetically do not differ in a significant way from other Central or even West European populations.

supernord said...

@ Michał

All your texts are incorrect in all items. BR2 is not the Lusatian culture. It is in Hungarian place, in Ludas-Varju-dulo the Lusatian culture there has never been. Etc.

Davidski said...

@Michal

Take a close look how populations with highest levels of WHG cluster on that PCA. They don't cluster together if they're from different language groups.

Poles don't cluster with the Irish, but rather with Slovenians, despite the fact that Poles are more similar to Irish in terms of ancient ancestry proportions.

So WHG has nothing to do with the positions of the eastern shifted Welzin samples.

Arza said...

@ Michał

BR2-like (or Lusatian-like) ancestry constitutes only a relatively small fraction of their overall ancestry

It's a marker. Unique and very precisely pinpointed in both, time and space.

Apparently, that "other part" of Polish ancestry must have been of Eastern origin

Ever heard about Yamnaya? Corded Ware? PL_N17? RISE598?

those haplotypes inherited from BR2-like people (from the Lusatian culture?) are not common neither among Eastern nor Southern Slavs

They don't have to be common.
On a side-note. Germans, Poland was full of Germans. But they've packed their stuff and moved away. If this BR2 affinity comes from "Germanic substrate", as you claim on other forums, then where are all that BR2 haplotypes which moved away from Poland with the Germanic masses?

Oh, now we talk about Germans, so they can magically disappear.

all those remaining Slavic populations show less affinity to BR2 than let's say people from Wales or Sicily

And do Welsh and Sicilians show substantial affinity to Poles? Or maybe this affinity to BR2 involves different part of BR2 genome, e.g. his other parent?

it could not have been Poland from where the Slavs expanded

And where is the problem in your opinion?

Twasztar said...

@Synome

I should really have put the earliest Slavic expansion in the 5th century. This predates the widespread adoption of the Slav ethnonym, but it lines up with the archeological descendants of the Kiev culture.

What you call "Slavic expansion" was in fact West Slavs/Veneti from the Lusatian culture being overrun by Germanic bands from Scandinavia during the Migration Period, and then restored at the beggining of the Middle Ages. Widespread adoption of the Slav ethnonym (that is, Byzantines extending ethnonym of one of the attacking tribes onto all other "Slavic" peoples) is of little importance. Neither Slavs exist only since the Slav ethnonym exists/is widely used, nor was that the only ethnonym of peoples nowadays known as "Slavs". Also, what archeological descendants of the Kiev culture are you talking about? If you mean Sukow-Dziedzice and Prague cultures, they are not descended from Kiev, but from Przeworsk.


In my opinion, the most probable archaeological sequence all the way from late PIE to Slavic is thus:

I see it a tad bit differently:

Sredny Stog - Proto-Indo-Slavic ((Late) PIE)
Corded Ware - Proto-Balto-Slavic
Trzciniec - Balto-Slavic
Lusatian - Proto-West Slavic | Chernoles - Proto-East Slavic
Przeworsk - West Slavic | Zarubinets, Kiev - East Slavic
--- Migration Period - temporary Germanic presence ---
Sukow-Dziedzice, Prague - West Slavic | Penkovka, Kolochin - East Slavic

Twasztar said...

@Michał

Thirdly, since we know that the ancient Balts from the Late Bronze Age were already nearly identical to the modern Balts (at least autosomally), this makes it practically impossible to derive the Balts from any such EEF-rich Welzin-like population

This is supposed to be an argument against what exactly? Why on earth would anyone need to derive Balts from an EEF-rich Weltzin-like population? I have no idea why you bring Balts and ancient Balts to this discussion, as if that was any relevant.


@Arza

Y-DNA is needed definitely. It will shut mouths of those people once and for all.

Michał said...

@Davidski

Many of the Welzin samples share very specific drift with modern-day Slavs

Where is your evidence for this? The fact that a very small fraction of those Welzin samples clusters with the Slavs in some (not all!) PCAs is definitely not enough to suggest any "very specific drift with modern Slavs", unless you also claim that a very similar "specific" drift is also seen for people of Germanic, Finno-Ugric or South European ancestry (as there are also quite many Welzin samples that cluster with those populations on your PCA).


There must be a very good reason for this, like, for instance, the Welzin samples being ancestral to modern Slavs.

In case you mean ancestral to all modern Slavs (ie. to Proto-Slavs), this seems extremely unlikely (or practically impossible). Shouldn't we expect to see the Welzin samples cluster mostly with those Slavic samples that resemble the Proto-Slavs in such case (and I hope you don't consider the Hungarians and Slovenians to resemble the Proto-Slavs much more strongly than it is the case for modern Poles, Ukrainians or Belarussians). Please remember that only one Welzin sample cluster with the Poles in your PCA. What proportion between Welzin and non-Welzin would you expect for the Proto-Slavic population? What would be the autosomal (and Y-DNA) profile of that putative non-Welzin component?

In case you mean ancestral to some selected Slavic populations, this cannot be ruled out, but it would have nothing to do with the Slavdom as a whole. Instead this would rather represent some local contribution of Welzin-admixed populations to some West Slavic and South Slavic (West Balkan) populations.


Poles top the f3 stats list with the Welzin samples, so there's no conflict with my results, although I'll reiterate again that f3 stats don't just focus on ethnic links, but all shared drift, so they're often difficult to interpret.

Your explanations are inconsistent. Let's state it very clearly: The f3 stats list for the only Welzin sample that clusters with modern Poles in your PCA (which is WEZ40) starts with the Irishmen, Austrians, Basques, Germans and Finns, while the Poles come only sixth. On the other hand, some Welzin samples that supposedly show most drift with the Poles (according to f3), cluster with very different (often non-Slavic) populations in your PCA, so this demonstrates that your PCA results are not consistent with the f3 stats results, hence you shouldn't refer to those results to support your claims.

Michał said...

@Davidski

Take a close look how populations with highest levels of WHG cluster on that PCA. They don't cluster together if they're from different language groups.

Why should they cluster together? They won't cluster together even if WHG is included in your PCA, and this is because they show other important differences. By contrast, the Welzin samples will be shifted to a separate Welzin-specific cluster that is almost equally distantly related to all remaining populations from Western and Central Europe, so after removing WHG they are nearly randomly assigned to different modern populations (and this assignment won't be very stable, as shown below).

Let's use your approach to determine whether Welzin shares any "recent drift" with some selected Romance-speaking populations. PCA shown in Fig. 6 (in the original Welzin paper) uses only modern samples (no WHG nor any other "distracting stuff") and indeed nicely separates all Romance populations from the Slavic and Germanic ones. "Importantly", most Welzin samples cluster with Southern Europeans now (mostly with the Italians and Spaniards). The very same sample WEZ40 that on your PCA plot clustered with the Poles now clusters with the Bulgarians (a former Roman province!), but two other samples (WEZ24 and WEZ77) that were previously shifted towards the Eastern Slavs now cluster with the Italians and Spaniards. Will you consider this to be a definitive proof that Welzin was ancestral specifically to people speaking the Romance languages?

Michał said...

@supernord

BR2 is not the Lusatian culture. It is in Hungarian place, in Ludas-Varju-dulo the Lusatian culture there has never been.

BR2 is from the Kyjatice group/culture which is quite commonly considered a local Southern (Slovak-Hungarian) variant of the Lusatian culture.

Anyway, in case BR2 is not related to the Lusatian culture (which is certainly possible) then this particular example cannot be used to support the Lusatian ancestry of the Poles/Slavs (which is what Arza apparently tried to communicate in his post).

Michał said...

PCA shown in Fig. 6 (in the original Welzin paper)

I meant Fugure 3.20.

supernord said...

People, you realize that the word "Proto-" means only that the community (thing) isn't recorded live, but only reconstructed. Between "Proto-Balto-Slavic" and "Balto-Slavic" there is no difference in description ancestral community, except that first you report that the community is not recorded in written sources, and reconstructed to you. And in the second case you do not notifying.


@ Twasztar
I see it a tad bit differently:

Your scheme is impossible. It contravenes everything.

Rob said...

@ Sam

No you've misunderstood , but these matters are highly complex so it's okay
I argued that modern R1a throughout Eastern Europe is largely (apart from the east Baltic enclave) due to Slavic movements in late antiquity/ EMA
And they are.
They did not continuously inhabit CWC territory, they are new arrivals from somehwere, probably west Ukraine
And of course this is correct. Just take a look at the wielbark genomes from the Iron Age; they do appear more "west European" in comparison; and modern Slavs do have BA admixture (just read Dave's lede)
These Bronze Age genomes might "look" Slavic, and they might be termed para-Baltic-Slavic, but they're probably extinct or rare subclades of R1a-Z20 nowadays . Let's wait see what they show

Samuel Andrews said...

But Rob you also claimed Hungary BA was the norm in Eastern Europe before Slavic expansions. And those Wielbark/Goth guys may very well have been recent arrivals from Scandinavia.

Twasztar said...

@supernord

Correction. "Trzciniec - Balto-Slavic" should have been "Proto-Slavic".

By Proto-X i mean the earliest possible stage identifiable as X.

Michał said...

@Twasztar

Y-DNA is needed definitely. It will shut mouths of those people once and for all.

Ok. let's then see your predictions for Welzin Y-DNA. My prediction is that no Slavic (or Balto-Slavic) specific clades of R1a are going to be found there. This includes M458, CTS1211 and Z92, all predating Welzin by nearly 2000 years and constituting together nearly 100% of modern Balto-Slavic R1a.

Instead, I expect to find some rare subclades of I2 or I1, plus some G2a lineages, and maybe even some rare subclades under R1b and R1a (except the typical Balto-Slavic ones, as mentioned above).

Davidski said...

@Michal

Here's the same PCA with SHG and WHG added. As you can see, Welzin_BA cannot be described as Bronze Age North/Central Europeans with enriched SHG/WHG, otherwise they would cluster in the hole well above the Slavic cluster.

https://2.bp.blogspot.com/-lNfM_P7sVJ8/WfJ7ApH5vWI/AAAAAAAAGLU/OgxkPyLN_nw0g0pe1o0ibs7F6iZptJ66ACLcBGAs/s1600/Tollense_Valley_warrirors_PCA2.png

There's something very specific about many Welzin_BA that makes them push left and down, towards Slavs, so they overlap with Western Slavs and even Slovenians, even though they have somewhat different ancient ancestry proportions.

I can assure you that I've tested this on many different plots, and it's a phenomenon that cannot be explain by any casual similarity between Welzin_BA and Slavs.

I'm not sure how else to explain to you that you're hopelessly wrong. If someone else wants to continue this discussion with you, then by all means, but I can't be bothered.

Rob said...

@ Sam

"But Rob you also claimed Hungary BA was the norm in Eastern Europe before Slavic expansions. "

As if. You do realise the Baltic and Russia are in Eastern Europe ? If so , who would make such claims that all EE looked like BA Hungary ?
Not me. Feel free to dig up anything to contrary

Rob said...

"And those Wielbark/Goth guys may very well have been recent arrivals from Scandinavia."

I said they might *plot with BA Hungary", or Western Europeans. We all know the goths didn't come from Hungary, right ?

Davidski said...

@Michal

Let's use your approach to determine whether Welzin shares any "recent drift" with some selected Romance-speaking populations. PCA shown in Fig. 6 (in the original Welzin paper) uses only modern samples (no WHG nor any other "distracting stuff") and indeed nicely separates all Romance populations from the Slavic and Germanic ones. "Importantly", most Welzin samples cluster with Southern Europeans now (mostly with the Italians and Spaniards). The very same sample WEZ40 that on your PCA plot clustered with the Poles now clusters with the Bulgarians (a former Roman province!), but two other samples (WEZ24 and WEZ77) that were previously shifted towards the Eastern Slavs now cluster with the Italians and Spaniards. Will you consider this to be a definitive proof that Welzin was ancestral specifically to people speaking the Romance languages?

LOL

No, I won't consider it as proof that Welzin_BA was ancestral to Romance populations. I'll consider it as definitive proof that the PCA is total garbage.

I can run a PCA like this. How much do you want to bet that only two of the Welzin_BA will come out looking Romance?

Michał said...

@Davidski

Here's the same PCA with SHG and WHG added.

There is something definitely wrong with your PCA. I have never seen WHG clustering with the Finno-Ugric samples. How many of those samples are WHG? It looks like nearly all are SHG.


There's something very specific about many Welzin_BA that makes them push left and down, towards Slavs, so they overlap with Western Slavs and even Slovenians, even though they have somewhat different ancient ancestry proportions.

Well, if this is real then it might be some local non-Slavic substratum, both for the Western Slavs and for the Slovenians (or generally for the West Balkan Slavs), so this would point to some ancient Germanic populations that had some Welzin ancestry, but that's all.

Davidski said...

@Michal

There's nothing wrong with my PCA.

You're just failing to grasp that it's a PCA designed specifically to find Slavic-related ancestry.

Michał said...

@Davidski

I can run a PCA like this. How much do you want to bet that only two of the Welzin_BA will come out looking Romance?

This is still more than just one sample looking Polish in your PCA. ;)

Anyway, this is exactly my point. By selecting a very specific set of samples (and changing some settings) you can of course try to get the results that fit your expectations as much as possible, but this is hardly a fair attitude unless your results are very straightforward and unequivocal, which is definitely not the case here.

Twasztar said...

@Michał

For people blinded by ideology like you, results will always be not enough "straightforward and unequivocal", no matter what.

Davidski said...

@Michal

By selecting a very specific set of samples (and changing some settings) you can of course try to get the results that fit your expectations as much as possible, but this is hardly a fair attitude unless your results are very straightforward and unequivocal, which is definitely not the case here.

OK, you're totally full of shit.

All I did was to design a PCA that splits the main linguistic groups within Northern Europe. I didn't tell the algorithm who spoke what. It was an unsupervised run. Do you know why the algorithm came up with a coherent result? Because people who speak similar languages in Northern Europe are also usually closely related.

I then projected the ancient samples onto this genetic/linguistic canvass. Again this was an unsupervised run.

Many of the Welzin_BA cluster very strongly with Western Slavs because they share a specific relationship with them. This is not a casual similarity and it's a result that is easily repeated on very different plots.

SHG/WHG cluster sort of in the middle between the Finnics and Balts here, because they're more closely, albeit still distantly, related to these groups.

Michał said...

@Twasztar

For people blinded by ideology like you, results will always be not enough "straightforward and unequivocal", no matter what.

What about your predictions regarding the putative Slavic Y-DNA lineages in Welzin? I hope you are not afraid of verifying your hypothesis about the Slavic identity of those ancient warriors.

Michał said...

@Davidski

OK, you're totally full of shit.

Don't loose your temper please (otherwise I will need to report your improper behavior to the blog owner) ;)

Sorry in case you felt offended by my remarks. I really want this discussion to stay civil.

Davidski said...

It makes no difference what Y-DNA lineages they had since it's already obvious that many of them were very closely related to modern Western Slavs.

There's no explaining away these results, certainly not by citing some of the woeful PCA in that thesis.

It's actually easy to run the PCA I'm running and to reproduce these results in a variety of ways. But I'm sure that when better genome-wide sequences of the Welzin_BA samples are available, then IBD and rare allele tests will show the same thing.

Michał said...

@Davidski

It makes no difference what Y-DNA lineages they had since it's already obvious that many of them were very closely related to modern Western Slavs.

Since this cannot be verified by any Y-DNA results, I guess this "close relationship" is supposed to be of a very specific nature. Could you please briefly describe this specific kind of relationship?

Twasztar said...

@Michał

What about your predictions regarding the putative Slavic Y-DNA lineages in Welzin?

I reckon we'll see mainly R1a (especially M458) and I2 on the Slavic side, and mainly I1 on the Germanic side.

No matter the results, they'll inform us only about Y-DNA structure of the Bronze Age Carpathian Basin, because that's where the warriors came from, unless they release more samples including from warriors who came from Poland.

Davidski said...

@Michal

Since this cannot be verified by any Y-DNA results, I guess this "close relationship" is supposed to be of a very specific nature. Could you please briefly describe this specific kind of relationship?

It looks to me that about half of the Welzin_BA set came from a population that was directly ancestral to Western Slavs.

I don't expect this to be confirmed just with more genome-wide DNA, but also with mtDNA lineages and, if we get enough samples, Y-DNA.

But if no typically Western Slavic Y-DNA lineages are found in the Welzin_BA set, even with, say, 100 samples, then this of course won't mean that there's no paternal relationship between Western Slavs and Welzin_BA. It could mean that the founder effects that characterize modern Western Slavic Y-DNA happened later and elsewhere than Western Pomerania.

Chetan Vit said...

Parpola (2012) provides a good model of the origins of the various Indo-European groups. One of the likeliest models in my opinion

http://www.academia.edu/23437942/Parpola_Asko_2012_2013_._Formation_of_the_Indo-European_and_Uralic_Finno-Ugric_language_families_in_the_light_of_archaeology._Pp._119-184_in_R._Gr%C3%BCnthal_and_P._Kallio_eds_A_Linguistic_Map_of_Prehistoric_Northern_Europe._MSFOu_266._Helsinki

Davidski said...

@Matt

I used one of your tables in my blog post. Cheers

supernord said...

@Michał

Kyjatice group/culture which is quite commonly considered a local Southern (Slovak-Hungarian) variant of the Lusatian culture.

- Nо, Kyjatice culture is not part or variant of the Lusatian culture. This culture is part the Urnfield system culture derived from the Pilini culture and Gavа culture under the influence of the Lusatian culture.

then this particular example cannot be used to support the Lusatian ancestry of the Poles/Slavs

It is not matter because Lusatian not come from Kyjatice culture or from BR2. Lusatian could be as many like BR2 and equally vary because different populations have a common origin and the difference. Slavs is not ancestry of BR2, but BR2 shifts to Slavs.

Synome said...

I think if we are going to be honest, we can't date Proto-Slavic to the Bronze age.

It's historically documented that Slavic didn't differentiate into separate languages until around the second millennium CE. If you assume it occupied roughly the range it did in 1000 CE all the way back in 1250 BCE, you're going to need a very good explanation for why Proto Slavic remained almost totally homogeneous in such a large area for more than 1,000 years before rapidly diverging into separate dialects and then separate languages from ~500 to ~1000 CE, which aligns with what we know from both written historical accounts of their expansion and the archaeological record.

Look at Matt's chart. The Bohemian Slav is shifted significantly towards Ukraine and away from Germany and Austria compared to both Welzin and Hungary_BA. Welzin is closer to German, Austrian, Swedish and Orcadian than it is to Polish.

Davidski said...

@Synome

Welzin is closer to German, Austrian, Swedish and Orcadian than it is to Polish.

Did it slip your mind that there are at least two clusters within Welzin_BA?

So who are the typically Slavic Welzin_BA samples? Germanics?

supernord said...

@ Synome


1. Slavic can differentiate into separate languages that disappeared.
2. If the language of a certain community lives in a certain limited area, it forms a dialect continuum which does not turn into separate languages. Lusatian culture is not divided into a plurality of cultures.
3. You are exaggerating the rate of division of languages, languages not sharing can exist for many centuries, for example the Hungarian language, the English language since the early middle ages and so on. The separation of Germanic language and the Celtic language originated ~3000 BC, however, these people began to spread themselves only in the 1 Millennium BC. A huge time. In this sense, the Slavic language is no different from any other IE language, the same everywhere, languages are not divided into other languages every century.
And the Lusatian culture existed for ~600 years.

Arza said...

@ Synome
It's historically documented that Slavic didn't differentiate into separate languages until around the second millennium CE.

Prove it, if you can.
(you can't)

you're going to need a very good explanation

Let me cite EastPole's comment from under previous post:
"The fact is that languages change really fast when people mix, when their religions and cultures mix, and change very slowly when people don’t mix and have stable cultures and religions."

Slavs have not only expanded, but were cut off from each other by other migrating people. Different religions were enforced. Different writing systems. State borders appeared. Nationalisms, linguistic reforms, finally modern (linguistic) scholarship.

And you know what? After 1300 years you can broadcast tv that will be easily understood by all Slavs, see here:

http://steen.free.fr/interslavic/

Here goes your proof for lack of differentiation (reported intelligibility of prayers written in OCS, I presume).

Synome said...

@Davidski
Actually, that did slip my mind. Here's my issue. The Slavic expansion involved extensive assimilation of local people. I'm not surprised that these ancient east central Europeans share specific drift with Western Slavs. Just as I would not be surprised if a similar shared drift can be shown between Bronze age Balkans people and modern southern Slavs. The shared ancestry dates to before the Slavic expansion.

Simon_W said...

That shift in the affinities of Welzin and Hungary_BA vs the Bohemian Slavs in Matt's table is indeed telling and can't be a coincidence. Reminds me of Jürgen Udolph's theory of the Slavic Urheimat and expansions:
https://jpst.it/16IW4

He based this on the distribution of Slavic place names, field names and hydronyms. The area of the proposed homeland is where names of East, West and South Slavic type overlap.

Matt said...

If we were to say that all the Slavic languages were as widespread and differentiated as they are today by the Bronze Age, then this seems like it is equivalent to us saying:

- South Slavic languages never expanded in LA/EM over the territories associated to the Dacian, Thracian and Illyrian languages, or that these languages actually were just South Slavic, misunderstood by ancient writers.

- East Slavic languages never gained ground over territories in the east of Europe, in Russia and Belarus, over areas where the languages spoken were Finno-Ugric or from Scythian or Baltic branches of Indo-European. E.g. some (Gimbutas for'ex) think Baltic was dominant across a wider swathe of land well into Belarus and even Russia. Based on hydronyms, found into eastern Poland and as far as the Dniepr and Kiev. If Slavic never expanded, and this region was Slavic speaking to begin with long before the late LA/EM, then this didn't happen?

Is this a fair or a misleading paraphrase?

(I'm guessing these are the points of controversy when we discuss if the Slavic language expanded in LA/EM, because the historical expansion of East Slavic across the Russian Empire has to be uncontroversial.)

Synome said...

@ Arza

"Prove it, if you can."

What I'm stating is the academic consensus. I'm sure you can Google or go to Wikipedia and find plenty of proof in the references. I'm not a linguist, and I'm not going to post annotated comparisons of Old Slavonic manuscripts and Novgorod birch bark letters. The professionals have already done that for us.

I would like to see the proof of your claims. Can you cite academic papers or books that support your theory? I would be happy to look at them.

Written linguistic evidence from the time periods in question points to Slavic diverging from around 500 CE but still being a single mutually intelligible language until around 1000 CE. Let's see the academic sources that say otherwise.

Matt said...

@Simon_W: That shift in the affinities of Welzin and Hungary_BA vs the Bohemian Slavs in Matt's table is indeed telling and can't be a coincidence.

Quick distance visualisation of distance from these (Welzin BA average, Hungary BA average, Slav Bohemia) against each other -

Based on all the dimensions of the 10 dimension Northern PCA from this post: https://i.imgur.com/Ji6ih8I.png

Based on all the dimensions of the 6 dimension Europe wide PCA: https://i.imgur.com/k7shMWv.png

The 10 dimension should isolate the substructures specific to the Northern European groups, while the European wide has some of that (more than a West Eurasia wide PCA) but is mixing in the WHG:EEF:EHG:CHG:Siberian contrasts as well. Essentially because the variance attributable to WHG:EEF:EHG:CHG should be lower in the pan-Northern only, so the PCA procedure, provided sufficient samples and SNPs, should capture more variance orthogonal to this.

(For more visualisation, some other 6 dimensions distance plots: https://imgur.com/a/ZZ4h6
10 d plots: https://imgur.com/a/8Hh2f).

Svento Suava said...

@Synome:

First, you can't hide behind an "academic consensus". In the 1500s the academic consensus was that the sun revolved around the earth... If you can't be bothered to understand what the academics are saying that don't bring them up in a discussion.

Second, is there really an academic consensus? Who is doing the consenting? What are the %s? Have you researched that? I am assuming, the answer is "no" so let's not bring that up.

Third, note that you claimed it is "historically documented" that Slavic differentiated around the turn of the millennium. So your argument was not about "academic positions" but about documentation. So the question is where is the documentation?

Svento Suava said...

@Lukasz M:

How are you defining "Slavic" for purposes of the base case for your modeling? I think you said at historycy that Eastern Germans are Germans by birth and language. I am not sure what that means then for your modeling. If someone is born in Germany and speaks German as his birth language, then that person is German? If we go with that definition then any kind of genetic testing does not really make any sense.

And to your point, about Poles from Suwalki region, I guess, I would have thought that they are in fact "Balts." At least for purposes of modeling.

But I may be getting wrong what you meant.

EastPole said...

@Chetan Vit
“Parpola (2012) provides a good model of the origins of the various Indo-European groups. One of the likeliest models in my opinion”

Not quite. Parpola writes: “I have suggested correlating Late PIE with the Late Tripolye culture”. I don’t think that PIE or Late PIE originated in Tripolye culture, but I have no doubt that the language and religion of Tripolye influenced very strongly the development of Slavic language and religion, We can see Tripolye cultural influence in Late Sredny Stog Dereivka culture where R1a-M417 individual was found with some Tripolye autosomal admixture. And the area of Dereivka culture overlaps with Slavic homeland.

@Arza
“Let me cite EastPole's comment from under previous post:
"The fact is that languages change really fast when people mix, when their religions and cultures mix, and change very slowly when people don’t mix and have stable cultures and religions."”

Agreeing with the good linguists like Oleg Trubachyov who stresses uninterrupted IE origin of Slavic since 3000 BC which correlates very well with the origin of Corded Ware culture as a result of mixing of Sredny Stog pastoralists with Tripolye farmers, we can now ask why Slavic languages have been changing so slowly. My answer is Slavs didn’t mix much and didn’t change religion.

Notice fertilizing fire symbol on the belly of Tripolye figurine:

https://i1.wp.com/upload.wikimedia.org/wikipedia/commons/0/01/MotherGoddessFertility.JPG

http://s28.postimg.org/tv6lds999/Cucutene_Trypillias01ok3_A.jpg1ok3_A.jpg

This symbol is still popular in Slavic cultures. Here is example from Boris Rybakov book “Jazyczestwo driewnich sławian” page 29:

https://s28.postimg.org/i0u92ellp/screenshot_1017.png

Here is Polish money from around 960 AD -Mieszko I denar:

https://s28.postimg.org/4mlrk44bh/mieszko-i.png

The cross with four dots later became a sacred Hindu symbol:

http://i1076.photobucket.com/albums/w443/priwas/indian_swastika.png

They call it ‘svástika’ which means "lucky charm" from ‘sú’ - “good, well” and ‘ásti’- "to be"

We call it 'swarzyca’, ‘świaszczyca’ ‘swarożyca’ or ‘swarga’. This name is clearly linked to Slavic fertilizing fire god ‘Swarożyc’ the son of the heaven ‘Swar’, worshiped first by Tripolye farmers.
This is a symbol of love and happiness and has nothing to do with German symbol of hate and racism.

The fact that elements of Slavic religion, customs or Slavic religious vocabulary which have survived till modern times can be used to explain the religions of ancient India, Greece and other IE which they have long lost and forgotten is a remarkable sign of religious continuity in Slavic lands between Elbe river and Ural since the Bronze Age.

Synome said...

@Svento Suava

Here's a citation for you. I don't want to take up too much space with this so, if you want more, just ask.

"Until the middle of the 1st millennium CE, the Slavs were known to other peoples as the inhabitants of the vast territories between the Dnieper and Vistula rivers. In the 6th century CE the Slavs expanded to the Elbe River and the Adriatic Sea and across the Danube River to the Peloponnese (southern Greece). In that period, according to the oldest Greek and Latin writings about the Slavs, they were already divided into several groups. The Slavic language, however, was uniform in its phonological and grammatical structure, with important dialectal variations occurring only in the vocabulary."
...

"The next period in Slavic linguistic history began with the loss of the “reduced” vowels ŭ and ĭ, called yers, that resulted from Indo-European short u and i; that loss caused a wide-ranging change in many words and forms. Although that process was common to all the Slavic dialects, which were still connected with each other at that period, it took place slowly and at different rates in different dialects, beginning in the 10th to the 12th century and expanding from the southwest to the northeast. With the loss of the yers, which gave different results in different dialectal groups, the uniformity of the Slavic language area finally disappeared, and separate branches and languages emerged."

'Slavic Languages' Encyclopedia Britannica, 2017.

Primary Contributors:

Vyacheslav Vsevolodovich Ivanov

Director, Research Institute of World Culture, Moscow Lomonosov State University. Professor, Department of Slavic Languages and Literatures, and Professor, Program of Indo-European Studies, University of California, Los Angeles

Wayles Browne

Professor of Linguistics, Cornell University, Ithaca, New York. Author of 'Relative Clauses in Serbo-Croatian in Comparison with English'. Co-author of 'A Handbook of Bosnian, Serbian, and Croatian'.
Co-editor of several volumes of 'Formal Approaches to Slavic Linguistics'.

Open Genomes said...

PCA of the Welzin Tollense LBA Battlefield victims along with ancient and modern samples projected as a map of Europe

Based on this projection, I would say that the two sides were the Proto-Celtic Hallstatt culture and the Proto-Germanic Nordic Bronze Age culture.

Welzin Tollense LBA Battlefield aDNA K6 3D Interactive PCA:
http://www.open-genomes.org/analysis/PCA/Eurogenes_Welzin_Tollense_LBA_Battlefield_aDNA_PC_plot_1-2-3.html

Chetan Vit said...

@EastPole Parpola's suggestion is based on the fact that evidences of the first wheeled wagons appear in the Trypolye culture region and early PIE probably had no words for wheel terminology since they are missing from Anatolian languages. So he theorizes that a group of early PIE speakers invaded the Trypolye culture and imposed their language on them, which then developed into late PIE after the Anatolian speakers had left for Anatolia. Subsequent expansion occured from the Balkan region back into the steppe and northern Europe.

Sounds like a credible theory to me. I don't know if there are any better theories which derive the Corded Ware directly from the Sredny-Stog or some steppe culture.

Davidski said...

@Simon_W

That shift in the affinities of Welzin and Hungary_BA vs the Bohemian Slavs in Matt's table is indeed telling and can't be a coincidence.

But some of the Welzin_BA are much more eastern than Slav_Bohemia, so what appears to have happened is that there was gene flow from the west into that part of Europe since the Tollense battle.

Davidski said...

@Open Genomes

I have to say that your screen cap looks unusual. It doesn't appear to match my PCA based on PCs 1 & 2 (how did you get so many Welzin_BA clustering with Scandinavians??), nor distances based on all 10 PCs.

Svento Suava said...

@ Synome

And how do Ivanov and Browne know this exactly? Were they there?

I could give you a cite to a bunch of other people who thought differently.

(Kostrzewski, Boguslawski, Ketrzynski, Manczak, Alioneli, Poesche etc)

Arza said...

@ Davidski

It matches a 3D view of PC1/2/3 at a certain angle. What Open Genomes forgot to mention is that Hungary_BA:I1502 is in the middle of that crowd.

Here is a 1/2/4 (6-dim spreadsheet) rotated to be most readable and familiar for everyone ("standard" West Eurasian PCA viewed from SHG perspective with Yamnaya - farmer cline going from left to right):

https://s6.postimg.org/eicpy3p5t/Welzin_6_D.png

2D PC1/PC4 will be nearly identical to this view (rotated mostly). It's the same structure that Matt spotted in DoHA, and which is visible there on 1/2/6.

Arza said...

This is how the plot from above correlates with geography:

https://s6.postimg.org/tdai3m8bl/Welzin_Geography.png

3 not-a-clines meet somewhere around Moravia, Slovakia and Hungary.
One goes to the North-West, second to Latvia, third to Greece. As WEZ56 shows, Balto-Slavic part existed already in the Bronze Age, since he is shifted from Baltic_BA towards Hungary_BA.

In other words Slavs are a part of a complicated Central European genetic structure that developed in the Bronze Age.

Matt said...

Few more graphics trying to compare, contrast and combine the K6 PCA from Davidski's comments in the last thread against Global 10 PCA, using the overlapping pairs: https://imgur.com/a/TcSa0

Note, no Welzin here as I have no Global 10 for them.

There's a contrast in where the Steppe_EMBA and MLBA, and the ancient Bronze Age Europeans join in the tree between the neighbour joining trees for the K6 vs Global 10 vs the merge of the two of them.

In the Global tree, the Ireland_EBA, Unetice and Nordic_LN populations seem high in steppe (look higher than present day people) so join with Northeastern Europeans who have higher HG ancestry, and also join with Steppe_EMBA, MLBA while the lower Steppe ancestry Hungary_BA joins with Western Europe. In the European K6, in an almost opposite pattern, these groups simply join with later people from the same region. Essentially in this plot Bronze Agers are 1) High Western affinity, high Steppe_EMBA affinity, 2) High Eastern affinity, lower Steppe_EMBA affinity.

In the merged tree this creates a bridge between present day NW Europeans and Steppe_EMBA.

Should work out differently once samples are in from the Baltic_BA, or maybe even Welzin, which be high Eastern affinity, high Steppe_EMBA affinity and create a bridge there (and also possibly when more Bell Beaker groups that have been are sampled are on).

In combining both PCAs together through one PCA on them, the dimensions combine so there's a clear split in PC1 and PC2, which pool similar elements from Global10 and also from the K6 to fit what look like mostly the most ancient proportions, then PC3 and PC4 which takes mainly from the K6 to form simple dimensions of West-East present day European and then to differentiate Finno-Ugric and Scythian, etc. populations from others.

Kind of interesting that at the moment, at least on these, there are these individuals who look ancestral and similar to present day Europeans in the K6 view looking at recent drifts, but they aren't the same ones that look to have the right deep ancestry proportions. So more sampling of either more populations, or more individuals within populations needed to capture individuals and populations who fit on both levels...

epoch2013 said...

This is very odd.

If this affinity really can be tied to Slavs than why are East-Slavs so different? Furthermore, the area was Germanic in the Iron Age, there is enough proof for that (Hydronyms + Tacitus + Runes). So either this is a very old population structure and the coming Germanic culture did not have much genetic impact on it. But then why wouldn't the same be valid for the coming of Slavic culture that came after the Germanics. Or after the Germanics all left the area was resettled from a similar genetic pool.

The Veneti come to mind. The name could then be tied to "Wend". These could already have been a mixture of Slavs and Germanics.

Synome said...

@epoch2013

I think migrant Germanics and Slavs both had some impact on the ancestry in the area of east central Europe, but the fundamentals were already in place by the Bronze age, before either of these groups arrived. So the first option. This genetic cluster is older than both Slavs and Germanics. Ethnicity was very fluid in those times. Groups were assimilated by elite dominance, tribes confederated with other tribes who didn't necessarily speak related languages etc. But I think the basic ancestral make up in this area was set by the Bronze age. Probably Trzciniec related.

epoch2013 said...

@Synome

But that is refuted by archaeological evidence if I understand it well, which points to massive depopulation during migrational period in Eastern-Germany.

Synome said...

@epoch2013
Could you cite something that argues for the depopulation? I'm willing to believe that could have happened.

In either case, both the ancient Bronze age Welzin people from this area that would have been depopulated, and the later Slavs would ultimately trace their their ancestry to the Trzciniec culture in my hypothesis here. So even if the original ancients all left or died, this accounts for the Slavic affinity of a subset of the Welzin samples. They may have been Proto-Baltic related.

epoch2013 said...

@synome

https://www.thefreelibrary.com/The+beginnings+of+Slavic+settlement+east+of+the+river+Elbe.-a0118956836

Between the early medieval and the late migration period settlement (i.e. in the decades about 600 AD) there is a hiatus, and so far archaeology is not able to bridge this gap. The often cited "Germanic-Slavic contacts" can not be verified archaeologically, even in view of the positive arguments of toponomy (Leube 1995; Brather 1996c: 180-186; Donat & Fischer 1994). Palynological observations indicate a fast and drastic reduction in settlement during the sixth and seventh centuries, as it is shown by a minimum of cerealia and an increasing proportion of trees; nevertheless a total lack of settlement can not be deduced (Wiethold 1998:269-281). We are confronted with unsolved problems of archaeological chronology. Whatever region was occupied by the Sorbs in 631/632, Slavs should have been recognised somewhere east of the Frankish kingdom in the early seventh century.

epoch2013 said...

@Synome

I think both assimilation of locals and total migration happened. There are stories which describe both. Paul the Deacon describes how the Longobards included all the people they conquered. Ceasar describes how Dumnorix burned down all the houses of the Celts that entered Gallia under his command so they wouldn't gave a way back.

There was a paper mentioned in some thread on Anthrogenica which mentioned partial continuity in the areas as well as massive depopulation but I can't seem to find it back

Arza said...

@ epoch2013

When it comes to this part of Europe ancient sources are very imprecise. On top of that they are often misinterpreted on purpose to fit some agendas. It's better to leave them aside.

Pre-Medieval Germanic hydronymy is virtually non-existent in this area (if we talk about Poland).

Yes, there are few findings with runes, just like medieval Slavic pottery in Sweden or a ring with "Allah" engraved on a jewel in one of the Viking graves. And no one says that never ever any Germanic tribe didn't dwell in Poland.

epoch2013 said...

@Arza

"When it comes to this part of Europe ancient sources are very imprecise. On top of that they are often misinterpreted on purpose to fit some agendas. It's better to leave them aside."

Tacitus is overall trustworthy on this.

"Pre-Medieval Germanic hydronymy is virtually non-existent in this area (if we talk about Poland)."

Tollense is in Germany, area around Neubrandenburg. Hafel, Spree.

"Yes, there are few findings with runes, just like medieval Slavic pottery in Sweden or a ring with "Allah" engraved on a jewel in one of the Viking graves. And no one says that never ever any Germanic tribe didn't dwell in Poland."

Again, the Tollense site is in Germany.

Arza said...

Ah, Dołęża you mean. Therefore, I don't know where the problem is. Nevermind.

Twasztar said...

@epoch2013

You are wrong of course.

Furthermore, the area was Germanic in the Iron Age, there is enough proof for that

There aren't any proofs for that, unless you mean that some Gothic armies were roving this area during the Migration Period.

Hydronyms

Hydronyms at least as far West as Oder are of Venetic i.e. Slavic origin, certainly not Germanic. Some people claim they are "old European", but that's still not Germanic.

Tacitus

There's no evidence that "Germanic" form the times of Tacitus is the same as today's "Germanic".

Runes

What runes? And why do you assume that runes must have something to do with the so called "Germanics" (that is, Nordics)? Venets used runes too. Actually, many runic inscriptions form Central and Eastern Europe can be read in Slavic.

epoch2013 said...

@Tasztar

Read the reply to Arza.

Gaspar said...

I cannot believe some people think that because you speak slavic you must be slav or that if you speak german you must be german ...........clearly then we are ALL English from ancestors form England on this forum, if this is the method some are adopting.

The battle was between Germanic people pushing east from denmark and north german areas into west baltic tribes ...........if these remnant west balts became germanic or slavic later on does not count ......Gimbatus gave you a clue who was living in these ancient lands

Sofia Aurora said...

There seems to be a tremendous conflict that devasted the Trypillian mega-towns and which is another proof for all the anti-kurganist folks that YES!!!
The Indoeuropeanization was not a pathway of sunshine and rainbows!!
Below is the link:

http://onlinelibrary.wiley.com/wol1/doi/10.1002/oa.2633/abstract

Synome said...

@Gaspar

I think it's too early to be jumping to conclusions about exactly which groups of people were involved in this battle and what languages they spoke. Proto-Germanic speaking people may have been involved. Proto-Baltic or related speakers may have been involved. I think the presence of Proto-Slavic speakers is unlikely, but considering the fact that some of these people seem to have traveled considerable distances to the battle site, I certainly can't say for sure. One possibility I have seen barely any attention given is the possible presence of Proto-Italo-Celtic related groups. The associated Urnfield culture expanded to this area around this time period.

But we really don't know. This is just the beginning, we need more intensive sampling over a variety of time periods in this area if we are going to try and pin languages on these ancient warriors who left no written records.

Svento Suava said...

Language is irrelevant here. The question is are these people the genetic ancestors of some Slavs, Germanics, Celts, all or none of the above.

Kuba Krchak said...

I think this shows existing Celtic admixture in modern Slavs.

twój stary said...

I refuse to believe that something like Proto-Slavic or maybe even Proto-Balto-Slavic(at this point) wasn't a thing. After all, Corded Ware were Satem speakers with close affinity to later Indo-Iranian languages. There is also a Proto-Balto-Slavic substrat in Germanic languages, meaning it simply had to be brought by Corded Ware. So I don't what's so shocking about it.

Fanty said...

Tollense... Tollense.

Isnt that that battle where, some month ago, scientists claimed there had been at least 3 different types of people amoung the dead? "Scandinavian like", "Polish like" and "Southern European like" people?

And didnt they say, no one of the corpses was local to that area but arrived from quiet far away?

Or was that a different battle? Or are their conclusions of who these people match, "outdated" or "wrong"?