search this blog

Sunday, October 22, 2017

Tollense Valley Bronze Age battle: preliminary ancient DNA analysis


This dissertation, I'm guessing, is a prelude to a paper on the genetic origins of the victims of what was probably a large scale Bronze Age battle in the Tollense Valley, northern Germany:

Addressing challenges of ancient DNA sequence data obtained with next generation methods.

I blogged about the Tollense Valley project last year, following a Science feature which posited that the battle fallen may have come from very different parts of Europe (see here). But judging by the results in this thesis, that might not be the case after all. Emphasis is mine:

The 21 samples available to this study stem from skeletal remains found in the Tollense valley in north eastern Germany and date to the bronze age (ca. 3200 BP), except for sample WEZ16, which dates to the neolithic (ca. 5000 BP) and was found in a burial context. Although several samples from the Welzin site have been dated using the C 14 method, from the samples used for this study only the neolithic WEZ16 (2960BC ±66) and the Bronze Age sample WEZ15 (1007BC ±102) were radiocarbon dated. All individuals except WEZ16 were found in a non burial context, widely dispersed and dis-articulated [48] along the river bank of the Tollense river.

...

The PCA in Figure 4.24 shows modern Eurasian individuals in grey and ancient individuals in colour according to their assigned population (for details on the modern populations see Figure A.48). The majority of Welzin individuals fall within the variation of modern populations from the northern central part of Europe (compare Figure A.48), with hunter gatherers, the Yamnaya and the LBK populations appearing on the outer range of PC1 and PC2.

...

Outliers from the Welzin cluster are: WEZ16, which falls closer to the Sardinians and neolithic LBK along PC2, WEZ54, which clusters with the Basques and also fall closer to LBK individuals along PC2, WEZ57, which falls in between the former individual and the Welzin cluster, and WEZ56, which separates from the main cluster of Welzin individuals along PC2 in the opposite direction as the former three, towards the Corded Ware or Yamnaya.

...

The ancient population that share the most drift with the Welzin group are WHG and the SHG population followed by the Unetice, the Bell Beaker and the Corded Wear. Starting with the Unetice the following f3 values fall in the range of the standard error of each other. The average difference between two consecutive f3 values is 0.0021 ± 0.0024 and the average standard error in each f3 value is 0.0037 ± 0.0007. The most similar modern populations are the Polish, Austrians and the Scottish.

...

Any interpretation regarding possible parties that might have been involved in the conflict in the Tollense valley ∼ 3200 ago can only be speculative with regards to the here shown data. With the resolution given here, an educated guess for different involved parties could be, that both parties were relatively local and more closely related than any ancient DNA study was able to separate so far. Maybe similar to people from Hessen versus people from Rhineland-Palatinate in modern Germany.



Sell, Christian, Addressing challenges of ancient DNA sequence data obtained with next generation methods, Mainz : Univ. iii, 109 Seiten, 2017, Urn:nbn:de:hebis:77-diss-1000012793

See also...

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

98 comments:

Samuel Andrews said...

Yeah, the media leaks gave the impression the armies had lots of mercenaries from southern Europe. Instead I guess one of the soldiers came from a Neolithic/EEF isolate tribe? And the confusing "most similar to Poles and Scandinavians" leak came from analysis not able to split apart European ethnic groups.

On pg. 104 is some newly sequenced mtDNA. It includes one new result from Mesolithic Greece.....mHG X. Recall, the previous result from Mesolithic Greece was K1c.

Also new results from ancient Iran....
Qaleh Rostam, Iran 6400 BC. rCRS (assume HV)-16234

Haftavan Tepe, Iran, 1600 BC. U5a1b1, J2b1a.

Samuel Andrews said...

Also some really old results from "Minino." Anyone know where that is.

Here are the results....
MIN4, 8740-8420 bc, U5a1.
MIN1, 8850-8550 bc, U4

Samuel Andrews said...

Table A.42. Pg. 103.

They got mtDNA and genome-wide coverage of each sample. SO, that means genome wide data from Mesolithic Greece!

Davidski said...

Sounds good. We'll probably see those samples in papers next year.

Davidski said...

By the way, the Minino site is apparently in northern Russia, so those mtDNA haplogroups make sense.

Grey said...

off-topic : interesting article suggesting sabre toothed tigers may have gone extinct in Eurasia but then replaced later from America.

http://www.nature.com/news/sabre-toothed-cats-prowled-europe-200-000-years-after-supposedly-going-extinct-1.22861

Davidski said...

Yeah, the full paper is open access here...

http://www.cell.com/current-biology/abstract/S0960-9822(17)31198-3

Rob said...

See what Y lineages turn up ?
Maybe some nowadays rarer R1a, some I1, L23, etc

EastPole said...

Sell projected his Weltzin samples on the reference space used in Hoffmanova et al. 2017

I tried to estimate ‘Polish’ area of Hoffmanova’s PCA:

https://s1.postimg.org/4wa8exiycf/Hofmanova_PCA.png

and then compared it with Sell’s PCA:

https://s1.postimg.org/1ur3rvh9i7/Sell_PCA.png

It looks like Weltzin samples are SHG/WHG shifted in respect to ‘Polish’:

https://s1.postimg.org/6c55pfxujz/WHG_shift.png

Maybe HG component was higher in warrior class?

Matt said...

Does look like there's an offset of these individuals generally having a slight enrichment of HG ancestry compared to the German Corded->German Beaker line. Might make sense with location.(I used Google maps for my own understanding to place this site compared to the Corded and Bell Beaker ones, screencaps to save anyone else the bother if they want to look: https://imgur.com/a/sWxns).

Though, that said, the ADMIXTURE analysis seems to suggest that's perhaps not the case? See p40.

They have a kind of PCA where they use European samples and the unrefined whole genome data and then run them all together, and then PCA where they've projected it on. I don't know if they get much of a comprehensible result out of running all the samples together, seems like a shame the author did not also run a separate Europe only PCA and then do projection on that.

https://imgur.com/a/sz5N5

@EastPole, you may want to look at p95 to save time on estimating areas and positions, as it has labeled recent samples.

Kristiina said...

Minino is in Pskovskaya Oblast', Russia, close to the Belarusian order:
https://geographic.org/geographic_names/name.php?uni=-4144035&fid=5308&c=russia

rozenfag said...

@Davidski @Kristiina

There are a lot of villages that name, Minino: https://de.wikipedia.org/wiki/Minino

However, if I google "Minino mesolithic" in Russian, the only I could find is Minino in Moscow region: https://www.google.com/maps/place/56%C2%B040'30.0%22N+37%C2%B059'00.0%22E/

Actually, there seems to be a paper about it: https://journals.uair.arizona.edu/index.php/radiocarbon/article/view/16448/pdf

Samuel Andrews said...

Another Mesolithic GReek mtDNA result I didn't notice...
Theo5, 7050 bc, T2.

That's a big deal because it really strongly suggests Mesolithic Greece had Basal Eurasian/Near Eastern stuff.

Then another ancient Greek result. I don't know if this is Mesolithic or Neolithic.
Xir2, 6200 bc, T1a.

Theo5 has low genome wide coverage. Theo1 (K1c) though is 16.68% "On Target." Is that good?

Lenny Dykstra said...

"Does look like there's an offset of these individuals generally having a slight enrichment of HG ancestry compared to the German Corded->German Beaker line. Might make sense with location."

Yeah, was gonna say the same thing. They appear to have more WHG remnant vs. moderns. We know that the Baltic around Estonia was a neolithic refugia for relatively unadmixed WHGs and Tollense isn't far from there.

Will this paper have y-dna and any idea when its due to come out? It might take years and multiple papers trickling out if they decide to excavate the whole site...

Rob said...

Good spot, Sam

Anthro Survey said...

Seems to confirm the idea yet again that the North Sea and Atlantic seaboards had a greater share of HG ancestry by the late Neolithic than other areas of Europe prior to the infusion of steppe ancestry there.

SwedenSkoglund_MN(Haak 2015) and WEZ16 were very much the norm.

Very much surprised to see the Basque-like WEZ64 that late in the game, though. Perhaps a merc from France or some isolated hill/forest tribesman.

Sofia Aurora said...

Guys sorry for interupting the continuum of the thematology of the posts but have you seen the news?

Extremely UNPRECENTED palaeontological-palaeoanthropological news!!!

Please see below:

http://www.independent.co.uk/news/science/teeth-fossil-human-history-evolution-development-germany-rhine-mainz-archaeology-a8010506.html

Also here:

http://m.dw.com/en/archaeology-fossil-teeth-discovery-in-germany-could-re-write-human-history/a-41028029

There is a free download paper here:

https://www.researchgate.net/publication/320518472_A_new_great_ape_with_startling_resemblances_to_African_members_of_the_hominin_tribe_excavated_from_the_Mid-Vallesian_Dinotheriensande_of_Eppelsheim_First_report_Hominoidea_Miocene_MN_9_Proto-Rhine_Riv

And a video:

https://m.youtube.com/watch?v=67axuBYGhzY

And all this just after a month from these findings:

http://www.sciencedirect.com/science/article/pii/S001678781730113X

And after only 3 months since the Graecopithecus smoking gun!!!

Lukasz M said...

Davidsky, it can be possible they use "Estonian Poles" as Polish proxy again in this study?

Davidski said...

Nope, I think the Poles they're using are from Poland. I have that sample set, and it's different from the Estonian Polish one, with no outliers with Siberian admixture.

Matt said...

Lenny: They appear to have more WHG remnant vs. moderns. We know that the Baltic around Estonia was a neolithic refugia for relatively unadmixed WHGs and Tollense isn't far from there.

Yep, it does seem to fit with that result as well?

(Very subtle Changes in f3 drift rank ordering compared to the Bell Beaker Germany and Corded Ware Germany might actually relate to a little bit of filtering in of Narva WHG related ancestry specifically... Compare - https://imgur.com/a/rcAtz. Though it's hard to say if that has any meaning without any systematic comparison.)

Btw, I would say scratch worrying about their ADMIXTURE analysis at K=2. Based on the patterns in their Europe_EN vs LBK_EN subsamples (which seem to indicate patterns of WHG difference which wouldn't be consistent with what we know).

I was slightly surprised by the f3 shared drift stats for the four outliers: https://imgur.com/a/rcAtz

MN European like female WEZ16, Iberia_BA / Basque like male WEZ54, particularly. WEZ16 shows highest shared drifts with BI populations, rather than Sardinians, which is unusual for a MN like sample - perhaps due to slightly higher WHG ancestry than usual for MNs combined with relatively high Anatolian pushing her into an unusual position....

Anthro Survey said...

@Davidski

Since we mentioned Slavs, I wanted to point out that Siberian admixture tends to be exaggerated in Russians.

When I modeled Russian_Smolensk in your set, it didn't ask for any Scythian_Pazryk, Buryat or other populations with a Siberian signal. Models rather will with just Slav_Czech and Latvian_HG. Proper Western Russians largely derive ancestry from a massive Slavic influx from a region encompassing parts of modern-day Poland/Belarus/Slovakia/West Ukraine, right?

The Kargopol Russians, on the other hand, do clearly require some additional Siberian input. As Slavic entities expanded eastward and north-ward, it is natural that locals with Siberian admixture were Slavicized(Russified). Just like Gascons are Latinized Vasconians.

Davidski said...

Russians from Smolensk are very similar to Belorussians and also Poles from Masovia, and there are probably different reasons for that.

mickeydodds1 said...

Perhaps this is the original ancient homeland of the people who later migrated to north Britain and became known to be the Scots.

Kristiina said...

@rozenfag
Thanks for correction. I was too quick and Russia is big. I would be very happy to get more data from Minino.

Rob said...

@ Anthro
A 9th century Czech wouldn't be ideal for Porto- Slavic due to absorbed Lombard etc
It's also interesting, if this PCA true, early Baltic-Slavs are more HG shifted than today, due to previously discussed BA Hungary signal introgressing sibsequently (?)

EastPole said...

Who was fighting with whom in Tollense Valley?
We will need Y-DNA to try to answer this question.

If R1a and R1b are found then the possible hypothesis may be: HG-shifted Celto-Germanics and HG-Shifted Slavs.
This would indicate that Bell Beaker derived Celto-Germanics around North Sea and Baltic Sea were HG-shifted.
And so Corded Ware derived Slavs around Baltic Sea were also HG-shifted.

What if N is found? Then we can speculate about Balts or Finns taking part in it.
I read somwhere about Celtic-Finnic linguistic connections.

AWood said...

@mickeydodds1

To be clear and true, people didn't high tail it for Scotland. Many of the tribes were pushed out of southern Britain by the Romans.

Synome said...

@EastPole

Slavs didn't expand from their original range until the medieval period. In my mind, that range is much further south and east of this area.

Also, the earliest cultures that we could possibly identify with "Proto Slavs" to the exclusion of "Balto-Slavs" would just barely be taking form during this time.

This battle site looks to be situated in an area that borders the future "Germanic" "Celtic" and "Baltic" zones of the Iron age, so maybe it's not surprising that there was already conflict in this possibly contested area. The Lusatian culture existed around this time and place, and it shows influences from all three of the precursors to these cultural areas (Nordic Bronze Age, Urnfield, Trzciniec, respectively).

EastPole said...

@Synome
“Slavs didn't expand from their original range until the medieval period. In my mind, that range is much further south and east of this area.”

Nobody knows from where and when Slavs expanded. We are hoping that genetics will help us to solve this puzzle as we are hoping that it will help us to solve PIE question.
We have R1a-Z280 in Lusatian culture and in Trzciniec culture so far. Present paper shows that some Welzin individuals had a lot of shared drift with modern Poles. So everything is possible.
Maybe the puzzle of Slavic origin is inherently linked with the puzzle of PIE origins. We cannot exclude that original Corded Ware spoke language which was closer to Slavic than to any other language. I think it is very likely taking into considerations similarity between Indo-Iranian and Slavic languages, similarities in culture and religion etc.
Celts and Germanics are also a great puzzle but I think they are more linked with Western Europe and Bell Beaker culture.
Let’s wait for more aDNA from Tollense Valley especially Y-DNA. Then we can discuss it.

Twasztar said...

Finally something interesting. Do we know whether they tested/will test also yDNA of the fallen? I expect lots of R1a, especially M458 and Z280.

Twasztar said...

@Synome

Slavs didn't expand from their original range until the medieval period. In my mind, that range is much further south and east of this area.

That's not what data suggests, including the data from this very study. Thanks to genetics, every year we know more and more about our ancestors, and it becomes more and more obvious that what has been claimed by certain fraction of archeologists is wrong and must be verified.

EastPole said...

@Twasztar
“Finally something interesting. Do we know whether they tested/will test also yDNA of the fallen? I expect lots of R1a, especially M458 and Z280.”

I am not sure about it. If Slavs won and the battle ended in the massacre of Celto-Germanic forces (Austrians and the Scottish) then majority of the killed in this battle should be R1b.

Samuel Andrews said...

@EastPole,

The word Slav refers to a single people group that expanded in the Middle Ages. There were no Slavs in 1000 BC. Referring to a Bronze age people as Slavs is like referring to the proto-Slavs as Poles.

Samuel Andrews said...

All I need is wikipedia to understand you can't refer to a Bronze age people from 1000 BC as Slavs. I guess if they were proto-SLavs, then yes, like how you can call Anglo Saxons English.

EastPole said...

@Samuel Andrews
“The word Slav refers to a single people group that expanded in the Middle Ages.”

This is a false definition of Slavs. It is clear from genetics and linguistics.
Don’t waste your time on things you don’t understand and don’t compare small and very young Germanic tribes to huge and ancient ethnos which populated half of Europe and is clearly linked with Sredny Stog and Corded Ware cultures.
Try better to explain the origin of Celtic and Germanic ethnicity because this is a real puzzle. I have no idea where they came from and when.

Matt said...

There are currently genetic results, in terms of the Ralph and Coop IBD expansions, and also Busby's CHROMOPAINTER results, which are hard to explain *without* a recent founder effect and expansion within Eastern Europe, during the late first millennium AD. (Ralph and Coop 2013 is the "old" paper, but they have also published this last year - http://www.genetics.org/content/205/3/1335).

This seems to align awfully well in Garrett and Chang's results showing a split of Slavic languages at around that time (http://cdn.sci-news.com/images/enlarge/image_2516_2e-Indo-European-Languages.jpg), which also aligns very well with the consensus of a late Antiquity Slavic expansion. (There are kinds of special pleading someone could do to get around this "Balto-Slavic is unusually conservative, so must have diverged millennia earlier". But I think those probably don't work - taking the above tree and even setting an extreme arbitary estimate at Balto-Slavic evolving at half the rate, then Slavic would still be an Iron Age diverging group).

All this said, the split of Balto-Slavic from Celto-Germanic groups looks to occur at roughly 2700BC, and their split from each other in the 700-500BC range. So possibly there were other languages related to these languages and on the Balto-Slavic side of the split about in Eastern Europe, and also Balto-Slavic-like people in some sense, before the recent expansion. Which then got replaced by the expansion of the Slavic languages as we know them...

supernord said...

@Matt - this picture doesn't matter. Dates it is not authoritative.

See https://upload.wikimedia.org/wikipedia/ru/8/89/Starostin_IE.png


Therefore, the argument that the Slavs did not exist cannot be taken.

Twasztar said...

@EastPole

You are right, but judging by the autosomal results I'm afraid ours might not win.


@Samuel Andrews

The word Slav refers to a single people group that expanded in the Middle Ages. There were no Slavs in 1000 BC.

Firstly, there is no evidence of any Middle Age expansion. Secondly, you are confusing two things. The birth of a people is not tantamount to the first recording of their ethnonym in a written source. That the name "Slav" doesn't appear in the sources before a certain date (which is not at all sure as someone may argue that for example the name Suevi/Suavi is in fact the same as later Sławi-anie (pron. Sŭavi-anie) - see e.g. Grimm), obviously doesn't mean Slavs couldn't exist prior to that date (under different name or unrecorded). It has already been explained on this forum that the word "slav" is attested in the very archaic Vedic language with the exact same meaning, ethymology and use.

Davidski said...

I've got these Tollense samples. Some basically look Slavic in the modern sense in terms of drift shared with Slavs. Didn't expect this.

Chad Rohlfsen said...

@Twasztar

You aren't referring to the Suebi/Suevi are you? They are not Slavs.

Twasztar said...

@Chad Rohlfsen

You aren't referring to the Suebi/Suevi are you?

I'm referring to the Suevi/Suavi as it's these forms that exist in sources almost exclusively, while the form "Suebi" is a German distortion.

They are not Slavs.

Prove it. Preferably by showing the language they spoke.

Synome said...

@Davidski

Are you planning on updating this post with some of your analysis of these individuals? I'd be interested to see how your conclusions compare with those of the authors.

Chad Rohlfsen said...

LOL! Are you serious? Romans even said they were an alliance of several German tribes. They're considered Elbe German speakers, ancestral to High German. Then of course there are things like the Kingdom of Suebia in Iberia, and then there is modern Swabia. Are you really suggesting Slavs covered most of Germany and moved to France and Iberia?

Ric Hern said...

I get the feeling that languages and people are more complicated....They behave like a Consertina, splitting up and recombining many times over but keeping some basics throughout.

The later Slavic expansion could have been a levelling effect of already existing dialects that resembled Slavic in some forms or the other thus making it seem as if Slavic is very young.

This I suspect of Proto-Indo-European to. Late Proto-Indo-European spread over Older Indo-European languages and leveled out the different dialects that is why some languages are left with Archaisms not found within the Majority.....

kony1_1 said...

@Davidski
In that sense, is there any distinction between Balts and Slavs, or in their relation to the Tollense?

@Chad
The ethnonym may have survived a language switch. The Germanic Suebi area neighbors or overlaps some proposed Slavic homelands.

@Matt
Mycanean offers a good date callibration point in my opinion.
It must have left Sintashta between 2100BC (Sintashta style artefacts) and 2000BC (no chariots).
It contains trace amounts of ruki, so this sound change had just started at that time, and no traces of satemization, so this change and the satem branches emerged after that time.

@slavocentrists
There is genetic evidence of the expansion 1400 years ago.
The expansion itself is historically recorded.
The ethnic landscape prior to it is historically recorded.
Language vise, there is only one possible candidate for the origin of the expansion, the (proto) West Baltic.
If the West Baltic had any substructure, it was too subtle for the historians to notice and would qualify as dialects at most.
Neither East Baltic nor Slavic did exist prior to the expansion, and both came to existence due to the expansion.
Hydronymy suggests that East Baltic expanded eastwards up to Moscow, but eventualy Slavic got to expand over it again.
Genetically, the conquered layers contributed to further expansion.

I believe originally Baltic and Thracian were pushed into Europe from Srubna or even more eastwards by Iranics, either as a single entinty and diverged later, or independently.
1500 and 1200BC are possible dates.

Helgenes50 said...

@ Davidski

Is the coverage high enough to have some of these new samples in Global10 ?

Davidski said...

@kony

I believe originally Baltic and Thracian were pushed into Europe from Srubna or even more eastwards by Iranics, either as a single entinty and diverged later, or independently.
1500 and 1200BC are possible dates.


Srubna gave rise to Scythians (shared R1a-Z93 and mtDNA). Balts and Slavs were offshoots from Trzciniec (shared R1a-Z280 and strong genome-wide drift). Ancient DNA has already shown this.

Davidski said...

@Helgenes50

Is the coverage high enough to have some of these new samples in Global10?

Don't think so.

Samuel Andrews said...

@Davidski,

Are the other ancient genomes like from Mesolithic Greece also available?

Davidski said...

@Samuel

Are the other ancient genomes like from Mesolithic Greece also available?

Don't think so.

Twasztar said...

@Chad Rohlfsen

Romans even said they were an alliance of several German tribes.

Yes, except that Germans from the Roman times probably have as much to do with what we today understand as "Germanic", as 10th century Prussians with 19th century Prussians.

They're considered

Different things can be considered.

Then of course there are things like the Kingdom of Suebia in Iberia

Of course, and a lot of weird Slavic-like toponyms there, and lots of strangely Slavic sounding names of their rulers too.

and then there is modern Swabia.

Modern Swabians are rather name-shifted Alleman imigrants than the ancient Suevi.

Are you really suggesting

I'm not suggesting anything as I haven't adopted any definitive position on this topic. I merely hinted that "one could argue" about the dating of the "Slav" ethnonym. I'm not going to discuss this any further.

André de Vasconcelos said...

@Twasztar

"Of course, and a lot of weird Slavic-like toponyms there, and lots of strangely Slavic sounding names of their rulers too."

Mind giving some very clear examples?
Because my whole family is from that area and I don't recall anything slavic-like in the toponyms

Cossue said...

With respect to the Suevi -as most local documents and chronicles called the ones who settled in Galicia and northern Portugal in 411, and this name is preserved in several villages and parishes called Suevos and Suegos- their names were Germanic, many even East Germanic: Hermericus, Reckila, Reckiarius, Heremigarius, Maldras, Massila, Framta, Frumarius, Reckimundus, Remismundus, Ariamirus, Theodemirus, Miro, Eboricus, Audeca, Malaricus, Wittimer, Nitigis, Adoric, Pantardus, Anila, Siseguntia, Ildericus, Neufila, Hildemirus...

Also, the chronicler Gregory of Tours, who was well informed on what was happening in Galicia in the 6th century, in a passage of his History of the Franks when referring to the Suevi of Galicia he equated them with the Alamanni.

Other arguments on the ethnicity of these Sueves can be given;
- Galician language have hundreds of Germanic loanwords, most of them received either directly from the language of the Sueves or Visigoths, or indirectly through French; but probably just a few loanwords taken from Slavic languages in recent times.

- There are thousands of place names in Galician (and Portugal) derived from a Germanic noun or from a Germanic personal name… But again near none derived from Slavic.

- In Galician and N Portugal R1a is just a 1-2%, but for example R1b-u106 (pre-Germanic, but probably expanded into southern Europe through Germanic peoples) is just under 5%...

Matt said...

@supernord, well, the main point of what I was saying was about the splits *within* the Slavic language family. The family looks as if (is reconstructed statistically as a tree as if) it they were languages which split in the mid-late first millennium CE.

If it had spread across a large swathe of Europe in the Bronze Age, then the divergences between the branches would be much greater than they are (e.g. family level divergences). Even if you assume that its within family rate of divergence is twice as conservative as, say, Romance or West Germanic, then the internal divergences still cannot really reconcile with a Bronze Age spread.

Trees do vary for the family level relationships (simply because there is still admixture and contact between language branches and a simple tree history will not hold*, and slightly differing weighting on different characteristics), and I won't say that Garrett and Chang's tree is slam right (though to me it makes the most sense**). But I think all that deal with internal splits within language families, and the one you refer to doesn't, will agree on the point of the internal splits within West Germanic, or Romance, or Slavic not looking that old.

I did bring in the macrofamily relationships to suggest that there may have been in Eastern Europe other IE language families which are most similar to Balto-Slavic out of the survivors today.

*Nets may be a more realistic assumption - http://rspb.royalsocietypublishing.org/content/early/2010/11/23/rspb.2010.1917 - or trees with admixture edges.

**One interesting point in favour of Garrett+Chang's tree is that it is a match to the structure of folktales across Europe - http://rsos.royalsocietypublishing.org/content/3/1/150645 - "even after accounting for spatial relationships among linguistically related Indo-European groups".

Matt said...

@Ric Hern: The later Slavic expansion could have been a levelling effect of already existing dialects that resembled Slavic in some forms or the other thus making it seem as if Slavic is very young.

Yes, it seems like that is possible. The thing is that it is a bit of butchering on the lingustics to call those existing IE dialects that may have resembled Slavic, Slavic. Slavic is a specific group of present day languages, and those present day languages have specific reconstruction time of internal divergences.

I don't think we can contest that proto-Slavic, the language, is "older" or "younger" than any other IE dialect. The question is whether the divergence (and expansion of Slavic languages across Europe) is "more ancient" or "more recent".

@kony : "Mycanean offers a good date callibration point in my opinion. It must have left Sintashta between 2100BC (Sintashta style artefacts)"

I wouldn't say this is crazy, but because artefacts are borrowed and traded, seems difficult to pin a particular kind of artefacts to a language, unless its something like the linguistic reconstruction that proto-Indo-European had to know the wheel, yoke, etc.

@Davidski:I've got these Tollense samples. Some basically look Slavic in the modern sense in terms of drift shared with Slavs. Didn't expect this.

Is this stuff like projecting onto a modern Europeans only PCA, where when we do it for e.g. Bell Beaker and Corded Ware, IRC, they tend to sit on a position to slightly further to the "north" and "west"?

Rob said...

I think the optimal usage for "paleo-lexicon" is chronology- such as the wheel for non-Anatolian IE and Iron/ chariots for proto-Celtic

With that in mind, what archaeologists might call Porto-Slavic / early Slavic culture is actually the late/ final linguistic stage of proto-Slavic . In fact, there was no clean split with Baltic, and Slavic is merely a southern Baltic langauge which expanded north into Belarus and Russia
So while it's worthile to investigate further the identity of przeworsk culture bearers (and their difference cf "Germanic" Wielbark) Matt and Sam are correct in pouting to an obvious expansion phase in 500 AD. It is simply disingenuous to claim for "no evidence" for such an event when there clearly is historic archaeological linguistic and genetic support for it

Simon_W said...

The Weltzin site is in the area of the Nordic Bronze Age, which in all likelihood was Germanic:

https://jpst.it/16Dct

It's well known that Germanic tribes in Antiquity often fought each other, so there's no need to assume an LBA "Battle of the Nations" of Celto-Germanics vs Balto-Slavs, lol. Especially given the close autosomal similarity of the samples, mentioned in the dissertation.

Simon_W said...

If these samples are close to Poles and share a lot of drift with them, this might explain why according to O'Sullivan et al., 5 out of 8 Alemannic genomes of 7th century Southern Germany were East European like.
http://eurogenes.blogspot.ch/2016/09/isba7-palaeobarn-abstracts.html

EastPole said...

@Matt
“There are kinds of special pleading someone could do to get around this "Balto-Slavic is unusually conservative, so must have diverged millennia earlier".”

“If it had spread across a large swathe of Europe in the Bronze Age, then the divergences between the branches would be much greater than they are (e.g. family level divergences). Even if you assume that its within family rate of divergence is twice as conservative as, say, Romance or West Germanic, then the internal divergences still cannot really reconcile with a Bronze Age spread.”

Your argument is false or pseudoscientific because you are assuming something you don’t know i.e. that the rate of divergence of Slavic is twice as conservative as, say, Romance or West Germanic.

Look at some data: Slavic change 1-2% /millennium :

https://s1.postimg.org/1wc1l7bgfj/screenshot_297.png

“The new arboretum of Indo-European “trees”. Can new algorithms reveal the phylogeny and even prehistory of Indo-European?” Hans J. Holm

http://www.tandfonline.com/doi/full/10.1080/09296170701378916

Why then only twice and not five or ten times as conservative as Romance or West Germanic? Why do we compare Romance which originated from Vulgar Latin overlaid over various languages like Gaelic etc or Germanic with its not-IE substratum with Slavic which was spoken by very homogeneous population for millennia.
This is a false argument used by glottochronology: if all languages change approximately at the same rate then Slavic couldn’t expand 5000 years ago but had to expand1500 years ago.
But do all languages change at approximately the same rate? Is glottochronology a science? I don’t think so.
It is true that Slavic languages change at a very slow rate. OCS recorded in Macedonia in IX century is still intelligible to me in Poland in XXI century and probably Slavic spoken in Poland in V century or earlier would also be intelligible to us now.
How do we know it was not so 2000 BC. If the rate of change for Slavic is 1-2% per thousand of years then even 3000 BC a large portion of basic words would be in intelligible for modern Slavic speakers.
We now have genetics and can verify opinions of various scholars about the history of Slavic languages:
Kortland:
“Starting from the linguistic evidence and trying to fit the pieces into a coherentwhole, we arrive at the following picture. The best candidate for the original IndoEuropean homeland is the territory of the Sredny Stog culture in the eastern Ukraine.

The Indo-Europeans who remained after the migrations became speakers of Balto-Slavic.”

http://www.kortlandt.nl/publications/art111e.pdf

Burrow “The Sanskrit Language” p. 14:

https://postimg.org/image/cvz03e7qn/

3000 BC is the time period when most likely proto-Slavic originated according to Oleg Trubachyov who stresses uninterrupted origin of Slavic (which excludes mixing and development outside PIE homeland):

“Currently, there is an objective tendency to deepen the dating of ancient Indo-European dialects. This also applies to Slavonic as one of the Indo-European dialects. However, the question now is not that the history of Slavonic may be measured by the scale of the II to III millenniums B.C. but that we can hardly date the ‘emergence’ or ‘separation’ of proto-Slavonic or proto-Slavonic dialects from Indo-European dialects because of the proper uninterrupted Indo-European origin of Slavonic.
The latter belief is in line with the Meillet’s indication that Slavic is an Indo-European language of archaic type, vocabulary and grammar of which has not experienced shocks in contrast to, for example, the Greek (vocabulary)”

Trubačёv, O. N. 2003. Ėtnogenez i kul’tura drevnejščix slavjan: Lingvističeskie issledovanija.
Moskva: “Nauka”.

Now such opinions seem to be confirmed by discovering R1a-M417 in Late Sredny Stog Dereivka culture which overlaps with Slavic homeland:

http://eurogenes.blogspot.com/2017/09/the-genomic-history-of-southeastern.html?showComment=1506060604577#c3839922991527525496

EastPole said...

Sorry, Gaulish not Gaelic. Romance languages are very mixed and therefore change fast.

Davidski said...

@Matt

Try this. The first 3-5 dimensions are enough to see very clear substructures, even between one Baltic-like outlier and a whole bunch of Slavic-like, or rather West Slavic-like, individuals.

https://drive.google.com/file/d/0B9o3EYTdM8lQcEhyVmxscktCLWc/view?usp=sharing

Romulus said...

Whats the speculation on why these 1000 BC samples uniformly show such high affinitu to WHG? Other Urnfield samples from nearby are I2a2, r1b, and I think 1 R1a? Talking about the Lichtenstein cave.

Jaap said...

What an interesting thread! Language families are not like trees, are they? They're like river-deltas, or nets. And for a split to be significant five millennia later would be truly amazing. Germanic(like) and Slavic(like) split on the steppe, continued as neighbours for all this time ... Ah, my brain hurts! Good sign! Likewise for the R1a-R1b dichotomy. Of course there are explanations, but explanations don't always 'explain', or turn into a narrative. The plot thickens ...

Lenny Dykstra said...

"If these samples are close to Poles and share a lot of drift with them, this might explain why according to O'Sullivan et al., 5 out of 8 Alemannic genomes of 7th century Southern Germany were East European like.
http://eurogenes.blogspot.ch/2016/09/isba7-palaeobarn-abstracts.html"

"Try this. The first 3-5 dimensions are enough to see very clear substructures, even between one Baltic-like outlier and a whole bunch of Slavic-like, or rather West Slavic-like, individuals."

------

That's very interesting... could the position of modern Germans be explained by Scandinavian and/or French admixture beginning around the early Middle Ages?

I also know Frederick William I encouraged migration of educated Hugenots and Belgians into Prussian territory but hadn't really considered that to be genetically signficant... maybe Germanic tribes really have become more "Westernized" (vis a vis their Slavic cousins) over the past 1200 years.

Chad Rohlfsen said...

You're full of it rookie

Grey said...

Romulus said...
"Whats the speculation on why these 1000 BC samples uniformly show such high affinitu to WHG?"

When I first read about Tollense I looked around that region on google maps and IIRC it's very flat with big lakes and little villages squashed on little hills so my guess is those villages used to be little islands in a big swamp.

It would be interesting to know how recently it was drained.

Samuel Andrews said...

@Lenny,

Instead, those ancient Germans are all probably Slavic and Slavic-related and not representative of actual Germanic people from the period. DNA from actual Germanic tribes people should turn out related to Germans and Scandinavians.

First Y DNA results from Goths have come out mostly I1 and leaked information from Polish media says they're closely related to Scandinavians.

Really, the affinities Iron age/Medieval peoples is still a mystery but I can't imagine German tribes had any special affinity to Slavs.

zardos said...

Rather they were not Germanic to begin with and the Germanic tribes expanded from somewhere else into these territories later if those results will be confirmed.

I think there were a lot of population movements of which we have no proof so far in the archaeogenetic data. Those were not as big as the Neolitic or LN/EBA ones, but still close to population replacement more than once. The differences between the competing populations were just not as big any more.

Matt said...

@EastPole, I don't want to hugely continue down that argument, but I think you may be confusing intelligibility and overall change in the language as a whole, with changes and replacement in stable, frequent used subsets of the language, which is what lexicostatistic trees are built around. English has changed, and so is not intelligble with old English because of, grammatical, phonological changes and borrowing of vocabulary - but this is not what is being used to built these trees.

Still, you are right about non-equivalence of rates of change even in these languages, which is constrained in models like G+C by attested ancestors, which is not possible for the whole Slavic clade. This said, analysis of English, Russian, Spanish, Greek written corpora did not seem to find much variance at rate of change relative to frequency of use of words http://sci-hub.io/http://www.nature.com/nature/journal/v449/n7163/abs/nature06176.html.

Why then only twice and not five or ten times as conservative as Romance or West Germanic?

At 1x as conservative, then it would seem to indicate that Slavic languages diverged roughly 15000 YBP, or at 5x, at 7500 YBP...?

Matt said...

@Davidski: Try this. The first 3-5 dimensions are enough to see very clear substructures, even between one Baltic-like outlier and a whole bunch of Slavic-like, or rather West Slavic-like, individuals.

Thanks. I've completed some PAST3 with these - plots, neighbour joining and distance measurements - https://imgur.com/a/nfSWa

The whole average of the group (excluding the outliers) seems closest to German (which I guess is mainly East German), then Slovenian, then Slovakian, then Ukrainian, Swedish, Polish.

Do you believe it would be possible to repeat with other ancient samples on, and perhaps Norwegian to fill out the NW space and Hungarian, Croat, Bulgarian to fill out the Southeast? I wonder particularly if some of the French / Spanish like samples could become less so if there were more modern SE samples in

apostateimpressions said...

"The majority of Welzin individuals fall within the variation of modern populations from the northern central part of Europe"

So, the population of Europe was established during the Bronze age with the steppe IE invasions, which is exactly what anthropologists have been saying for over a century and which "right on" lefty-liberal "anthropologists" have been trying to deny for the last 30 years. Look who turned out to be right. Of course that population is now being wiped out through mass immigration of workers to keep the capitalist system going. A lot else was said over the last 150 years, lets wait and see who was right about that too...

Chad Rohlfsen said...

Modern Germans are likely more Mediterranean-like due to Protestant migration to Northern Europe and the IR. We will see this to be the case, I'm pretty sure, when we get late Iron Age / Migration Age Germans. Remember, Saxons looked like Norwegians and Swedes.

Rob said...

Modern Germans South shift could be due post-Roman movement from west Germany and France ?
As Carolingians and Ottonians settled the Slavic areas. Have much of the ancient Germani east of the Elbe have effectively dissipated at a population level ?

Rob said...

Charlemagne's Saxon wars impacted north- central Germany also.

Romulus said...

@Samuel Andrews

"First Y DNA results from Goths have come out mostly I1 and leaked information from Polish media says they're closely related to Scandinavians. "

You have a source/reference on that? Please elaborate.

Chad Rohlfsen said...

Rob,

I don't think Franks will be all that different, really. They were northern neighbors of Alemanni, who also appear very north eastern. I think even Roman genetic impact will be microscopic compared to the Protestant and IR movements. Post and Pre-Roman Brits really don't look any different from each other either.

Chad Rohlfsen said...

I also think we're going to see that migration era movements are not as impressive as people think. I am betting the Saxon input of 20-30% is a preview of things to come. Y changes are 2-3x higher than autosomal changes. I wouldn't be surprised if the average say, Polish individual is 20-30% of the Slavic group, on top of the 80% Gothic and Elbe group, which is itself only a 20-30% Germani imprint on another late PIE group we no longer have or maybe Balts. I could be wrong, but some whole pop movements into abandoned places just seems a unlikely.

Rob said...

@ Chad

"I don't think Franks will be all that different, really"

Franks won't be (they mostly cremated), but later Carolingians coming from France will be. That's what I mean by post-Roman. I'm talking on top of head, haven't looked into the figures to see where exactly colonists came from and how many there were.

Chad Rohlfsen said...

I'm not so sure about that either, on a whole. You see some intermarriage with other groups around the Carpathians or Baltic in the elites, but I don't think there will be much difference there in the general populations. The hold on territory in modern Germany was pretty short. It's probably more professional armies by this time and not a large folk movement. Probably no more of an input than Crusaders on Middle Eastern genetics. But, we'll see.

Rob said...

"Probably no more of an input than Crusaders on Middle Eastern genetics."
Do you reckon ?
There was active settlement to parts east of the Elbe, where there had been virtually no German speakers east of the Elbe for few hundred years

Chad Rohlfsen said...

That window was the 6th through 9th century. It wasn't everything east of the Elbe. More like Mecklenburg-Vorpommern to the south/ southeast. Like you say too, probably never really completely voided of Germans, but a minority for that window.

But, kind off topic for this stuff or what I'm referring to.

AWood said...

This window seems to line up with the late Bronze Age, maybe during the Amber trade and an entry of R1b-DF27/U152 in the late Nordic Bronze Age? I1, and R1a, probably R1b-U106 were already native to the northern region. It seems there is some input from the central-south of Europe if we take the "Basque" signal as real. Let's wait for the full analysis, but it seems there is at least some southern, (but not too southern) contribution. Switzerland, France, and northern Italy might fit just as nicely.

Davidski said...

@Matt

Try this...

https://drive.google.com/file/d/0B8XSV9HEoqpFbFBQc2xGRWdleFU/view?usp=sharing

Simon_W said...

Quite frankly, you're not seeing the wood for the trees.

Check LBA Europe: https://jpst.it/16FBQ

See the core of the Hallstatt culture (early Iron Age): https://jpst.it/16FCk

Watch its maximum extent: https://jpst.it/16FCq

Enjoy a map of the La Tène culture (late Iron Age): https://jpst.it/16FCO

Then you'll see that the area of modern-day southern and central Germany was all the time covered by cultures belonging to central Europe. In all likelihood these were not Germanic speakers, but probably in the Italo-Celtic range and autosomally inbetween modern French and BR2 from LBA Hungary. Only northern Germany belonged to a different sphere, and this was the early Germanic sphere, I suppose. With the spread of the Jastorf culture in the Iron Age these early West Germanics spread southwards and assimilated the more southern, French-like locals and mixed with them. Simple as that. No need to assume mass migrations of Huguenots and Romance speaking Franks in the Carolingian era. Sure there was migration of this kind, but that should be relatively negligible.

Rob said...

Simon

The Germans who colonised east Germany in the Middle Ages were different to Iron Age inhabitants
That's a completely new/ differerent population and a break in previous continua
What's the bet that Iron Age mecklenberg Pomorania etc are more northern than today ?

Simon_W said...

I think it's possible that Germanic arose as a mix of an R1b-U106 dominated Italo-Celtic-related population and an R1a-dominated population that shared more with Balto-Slavic. In fact, Germanic shares some morphological features with the latter at the exclusion of the former. For instance:
For the genitive singular of nouns, we have Slavic (OCS) and Baltic (Lithuanian) using -ó, Baltic and Germanic (Gothic) using -eso, and Celtic (OIR) and Italic (Latin) using -í.
For indirect and dative cases, we find Slavic, Baltic, Germanic, and Tokharian using -m, and Celtic and Italic using -bhos (dative singular). (source: http://webspace.ship.edu/cgboer/indoeuropean.html )
And if the mixture hypothesis is true, then there may have been a continuum in the early Germanic area, with more eastern groups having more eastern affinity. We've already seen what the Anglo-Saxons were like, and I suppose they were not quite as eastern as the Welzin samples. But they didn't originate in eastern Germany either.

Simon_W said...

@ Rob

Yes, absolutely. Another map: https://jpst.it/16FJE
There we see the German/Slavic border at around 1000 AD.

Simon_W said...

A map showing the temporal progress of the Medieval German "Ostsiedlung": https://jpst.it/16FK2

And these were the main vectors of this population movement according to dialects:
https://jpst.it/16FKl

It's clear that the Germans assimilated and absorbed some of the Slavs, and it's likely that the Slavs in turn had absorbed some Germanic remainders when they moved in. But how much? Probably not large quantities.

Rob said...
This comment has been removed by the author.
Rob said...
This comment has been removed by the author.
epoch2013 said...

@David

How well do they compare to the North-Germans (Niedersachsen) and (North-)Dutch?

Open Genomes said...
This comment has been removed by the author.
Open Genomes said...

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


What does everyone make of this?

My take is that whoever was fighting, the Germanic tribesmen lost the battle.

Tesmos said...

@Simon_W

''I think it's possible that Germanic arose as a mix of an R1b-U106 dominated Italo-Celtic-related population and an R1a-dominated population that shared more with Balto-Slavic. ''

R1b-U106 probably dominated an earlier western IE group which was responsible for the formation of Proto-Germanic, rather than an Italo-Celtic related population. We should not forget that R1b-U106 has been found in Battle Axe Sweden and Bell Beaker Netherlands.(Rise98 and I4070)

EastPole said...

@Matt
Thank you for the link to Pagel et al. 2007 paper, but I really don’t think that mathematicians, computer scientists and biologists will be able to tell us the real history of our languages.
I think the real linguists, those able to explain words, discover their etymology and history, those who are able to understand ancient religions and cultures together with the geneticists who are able to tell us when and how people mixed will tell us the true story.

“At 1x as conservative, then it would seem to indicate that Slavic languages diverged roughly 15000 YBP, or at 5x, at 7500 YBP...?”

You see how stupid conclusions can follow from pseudoscientific assumptions about the steady rate of change used in computational linguistics.
Slavic languages were changing slowly but surely Slavs didn’t exist 15000 YBP.

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.


Some time ago I wrote about elements of archetypal Indo-European poetic language present in the most archaic phases of Rig-Vedic and Greek composition, which were also found in Slavic poetry:

http://s22.postimg.org/jp60mfvg1/screenshot_150.png

http://chs.harvard.edu/CHS/article/display/6429

But Slavonic metre which represents a heritage of considerable antiquity was recorded in much more recent times :

http://s18.postimg.org/k64z31z3d/screenshot_112.png

https://books.google.pl/books?id=ZXrJA_5LKlYC&pg=PA161&lpg=PA161&dq=%E2%80%9Cghee%E2%80%9D+is+a+secret+sacred+word&source=bl&ots=WcAwklg4mn&sig=25udoVHnWgJn-99t3L2Htg4iADQ&hl=pl&sa=X&ved=0ahUKEwjqnLrnx-bPAhUhJJoKHYXEBXsQ6AEIRjAG#v=onepage&q=%22When%20it%20comes%20to%20slavonic%20metre%22&f=false]https://books.google.pl/books?id=ZXrJA_5LKl...etre%22&f=false

What does it prove. Well, it indicates that Slavic culture has been quiet stable for some time. Up to modern times Slavic poets have been using archaic poetic meter that their proto-Slavic ancestors used when living in PIE homeland together with the ancestors of proto-Greeks and proto-Indo-Iranians.
But Slavs are still living in the same area and are similar to their ancestors living there in PIE times so maybe this is why all is changing so slowly here.

We are still using religious terminology in our translated Bibles which can also be found in Rigveda, but we can explain many of these words, i.e. they have Slavic etymology, while Indo-Iranians can’t.

Slavs probably didn’t exist before 3000 BC. I think they may have originated when R1a dominated Late Sredny Stog Dereivka culture pastoralists mixed with Tripolye farmers and settled in Vistula-Dnieper area, which took place around 3000 BC. At that time languages, religions, cultures were changing very fast but after that changes have been really slow.
Computational linguistics will not have much to say about it.

Svento Suava said...

@Cossue:

First, no one disputes that there are Germanic toponyms in Spain/Portugal. The question, however, is how many of those are from the Visigoths and how many from the Suevi.

Second, Why do you think that names like Miro or Ariamirus are Germanic? I could point to some others too. What would you consider a "Slavic" name then?

Third, you are assuming that those people who called themselves Suevi in Portugal in the 5th century spoke the same language as the people who called themselves Suevi in Caesar's time.

olaf larsson said...

Are there any results regarding which y-haplogroups these individuals have?

MomOfZoha said...

@EastPole:
"Thank you for the link to Pagel et al. 2007 paper, but I really don’t think that mathematicians, computer scientists and biologists will be able to tell us the real history of our languages.

I think the real linguists, those able to explain words, discover their etymology and history, those who are able to understand ancient religions and cultures together with the geneticists who are able to tell us when and how people mixed will tell us the true story."

Computational assumptions on the rate of change for languages and genomes have far more to do with attempting to tractably model distance-preserving phylogenetic trees than they do with pure mathematics. The most "mathematical thing" that one may attempt to say about them might be that the relevant effect observable from some average constant rate of change versus a variable rate of change is provably indistinguishable under such and such other assumptions over such and such time spans in some "law of large numbers" kind of way. But, that requires encoding a lot of other biological principles into mathematics in the first place, which may be more difficult for biology than for statistical physics...

For a "layperson" the bottleneck into mathematics is often the level of formalization required. But linguists eat formality for breakfast, far more than mathematicians themselves. It would be a useful exercise for every linguist to attempt to formalize and prove the "linguistic theorems" they wish to express. The beauty of such an attempt, even if it proves too ambitious, is that you realize aspects of your own assumptions, and your own principles driving your thoughts -- you feel the topology of your own mind. If there is a tautology -- a short-circuit so to speak -- you might be able to catch it that way.

Otherwise, if we are to consider a "yes" or "no" answer to "Is the PIE 'homeland' in location X?" to be a "theory" in and of itself, then a random coin would be just as profound as the majority of linguists and anthropologists (assuming that "PIE Urheimat" is a fundamentally meaningful concept -- ok, two more random coins).

Then again, as long as the underlying fiber of PIE talk that attempts to connect South Central Asian poetry and religion to Slavic genetics does not ever disturb the graves of Attar and Mevlana, I will be ok. It should be a great relief to give the Europeans credit for the Hindu caste system anyway, though that of course is not mine to give.

If I were to make a poetic bargain with you, I’d ask you to trade Russian Pushkin for Iranian Ferdowsi. This should be mutually desirable: Pushkin must be far too Sub-Saharan for your taste on account of his great-grandfather. And, the Iranians who pay homage to the extravagant shrine of Ferdowsi are those who place one foot forward to pose for the camera while naming their children “Aryan” so many years after World War II. Ferdowsi was really more linguist than poet.

You can tell a lot about a dead poet by those who visit his grave.

The birds themselves pay homage to Attar in the garden surrounding his most humble tomb. My nineteen-month-old daughter found and chased them happily from one end of the garden to the next. I swear they were thirty, all flying in the same breath.