Tracing the Spread of Celtic Languages using Ancient Genomics

[Andour]

Not sure male haplos will prove very helpful in mapping migrations in this particular instance.

By the end of Bell Beaker times, there are already heaps of L2 not only along the Danube, but also around present-day Prague. Most likely those clans (some of them at least) survived in situ through the Bronze Age and were still there when Knoviz formed. If people moved north into Czechia from, say, the banks of the Danube in early Urnfield times, the autosomal mix was impacted (as per the paper under discussion), but the male lineages probably remained predominantly L2. The same goes for other places in a vast area stretching from Hungary to the Rhine - and perhaps some distance west of it. The only thing that might help would be to get in-depth subhaplo analyses from both the different places and the successive periods under scrutiny. And even then we probably wouldn't get a sufficient corpus to properly track the paths of each individual clade.

Edit : exactly what Riverman posted above while I was typing... (Always been a bit of the slow type !)

[Mitchell-Atkins]

Not sure male haplos will prove very helpful in mapping migrations in this particular instance.

By the end of Bell Beaker times, there are already heaps of L2 not only along the Danube, but also around present-day Prague. Most likely those clans (some of them at least) survived in situ through the Bronze Age and were still there when Knoviz formed. If people moved north into Czechia from, say, the banks of the Danube in early Urnfield times, the autosomal mix was impacted (as per the paper under discussion), but the male lineages probably remained predominantly L2. The same goes for other places in a vast area stretching from Hungary to the Rhine - and perhaps some distance west of it. The only thing that might help would be to get in-depth subhaplo analyses from both the different places and the successive periods under scrutiny. And even then we probably wouldn't get a sufficient corpus to properly track the paths of each individual clade.

Edit : exactly what Riverman posted above while I was typing... (Always been a bit of the slow type !)

At present we have a Hallstatt in eastern France and a La Tene sample in Slovakia on the U152>L2>Z49>Z142>Z12222, Z150 branch, but nothing further down on my line in the 1st millennium BC.  But these 2 samples kinda bracket the area in question.

CGG023693 530 BC Hallstatt Z150
"Les_Moidons_Tumulus_8_1920_51  Bourgogne_Franche_Compt Les_Moidons Barrow 46.878971 5.842749 Bone Petrous 750BCE-450BCE"

I11712, 190-1 BC, 95 BC, La Tene Z150>BY33535 in Bratislava, Slovakia.

[Strider99]

This part appears quite speculative on their end, but it's still an interesting interpretation of the Tollense battle:

The Late Bronze Age was a period of cultural transitions that were formative to the successive Iron Age, including the rise of intensified agriculture, metal craftsmanship and new forms of warfare. Archaeologically, this era saw the emergence and expansion of the Urnfield Culture (3300–2750 BP) across much of Europe. Several factors contributed to its expansion, including the aforementioned advancements in metalworking and agricultural techniques, which provided economic stability and surplus.

The introduction of new crops, such as millet, may have further aided population growth. This fostered expansion and facilitated the long-distance exchange of important commodities, including bronze, salt, and other raw materials.

In contrast, the northward expansion from Central Europe into northern Germany and southern Scandinavia appears to have culminated in a significant defeat, as suggested by the Tollense battle around 3200 BP and evident also from the lack of Knovíz related ancestry in Scandinavia prior to the Migration Period.

It seems to me that they mostly base this conclusion on the lack of Urnfield-related ancestry in Scandinavia until ~500 AD. Their interpretation could fit with the Weltzin samples carrying R1b-P312, R1b-CTS6889 which the Tollense sample shares with a Knoviz culture individual, and I2 lineages like Z2069, the latter which has also been found in a Tumulus culture context at Konobrže. 

This taken together with the rather southern-shifted profiles of the Tollense samples makes for a fairly convincing argument, in my opinion. Especially since we know that many of them (but not all) had non-local strontium isotope ratios.

[Riverman]

Not sure male haplos will prove very helpful in mapping migrations in this particular instance.

By the end of Bell Beaker times, there are already heaps of L2 not only along the Danube, but also around present-day Prague. Most likely those clans (some of them at least) survived in situ through the Bronze Age and were still there when Knoviz formed. If people moved north into Czechia from, say, the banks of the Danube in early Urnfield times, the autosomal mix was impacted (as per the paper under discussion), but the male lineages probably remained predominantly L2. The same goes for other places in a vast area stretching from Hungary to the Rhine - and perhaps some distance west of it. The only thing that might help would be to get in-depth subhaplo analyses from both the different places and the successive periods under scrutiny. And even then we probably wouldn't get a sufficient corpus to properly track the paths of each individual clade.

Edit : exactly what Riverman posted above while I was typing... (Always been a bit of the slow type !)

Well, we currently had the paper on Unetice in Austria, opposite to the Bell Beaker derived Unterwölbling group, and the Uneticians had different yDNA, most notably branches of Epi-Corded R-Z283. One might argue that the Uneticians were not as exclusive as to fully exterminate earlier BB people in the areas they expanded into, BUT they still did expand and colonise into the areas their core cultures spread too. This became very apparent in the recent Austrian paper, since the Danube served as the borderline (similar to the Tumulus culture vs. Suciu de Sus-Igrita etc. border formed by the Tisza in the Carpathian basin), and the population North of the Danube, in Unetician colonised territories, was quite different from the people to the South in the Unterwölbling group.

Looking at this generalised map:



You see the borderzone between Unetice (4) and Unterwölblinger group (8) which follows a clear cut genetic separation.

Czechia being by and large occupied by Uneticians means that if we talk about regional continuity, these locals should have had at least al of of R-Z283 AS WELL.

Knoviz is a mixed group, but not as mixed, and the mixed elements seem to have come more recently, from the East, with the Lusatian-Urnfield influence.

Therefore Knoviz R-L2 is definitely not the main surviving-important part of the puzzle. Because the Tumuluc culture clans which expanded into the Carpathian basin, replacing locals which had other haplogroups (like Encrusted Pottery was I2/G2 by and large), must have been coming from a "pure"/exclusive source group of people, and this is indeed the Southern German core Tumulus culture zone, from which the TC warbands/pastoralist clans expanded from, replacing and largely annihilating Western Uneticians and related groups as well as Encrusted Pottery, Vatya etc., all very different cultures and people.

You also have Nitra culture (11) and Otomani (15) and Mierzanowice (18) on the map. All these groups were dominated by Epi-Corded R1a/R-Z283. And all of them got largely destroyed by the Tumulus culture first and the things they set in motion second (like Carpathian people moving North fleeing TC, attacking them, which helped to form Eastern Urnfield/Lusatians cultures there).

We can therefore say with confidence that most of the R-L2 which expanded Eastward came from a Tumulus culture core zone in Southern Germany and there was little to no continuity in its Eastern zone of distribution.
Similarly, but probably not to the same extent, when La Tene Celts moved in, they too brought a fresh badge of R-L2 to the East, largely replacing or mixing with locals which were dominated by different patrilineages.

[Riverman]

This taken together with the rather southern-shifted profiles of the Tollense samples makes for a fairly convincing argument, in my opinion. Especially since we know that many of them (but not all) had non-local strontium isotope ratios.

The Tollense warriors are a wild bunch of allies called to arms from a wide range of Urnfield regions. We have Tumulus culture derived Southern Germans beside various Lusatians and from the far East of their range Balto-Slavic kind of auxiliaries or mercenaries too. This shows first and foremost, to me, how Urnfield was successful and managed to overcome opposition: They gathered forces from other areas of their network, other believers, from all corners of the Urnfield sphere, best compared with say the Christian crusades in my opinion.
It also shows that different ethnicities worked together under the Urnfield banner in such instances, even if many of the warriors might have been paid mercenaries.

[Andour]

One might argue that the Uneticians were not as exclusive as to fully exterminate earlier BB people in the areas they expanded into...

Yes. The 2024 Penske paper on Leubingen Unetice has 26 male samples : 1 I2a, 10 R1a, 15 R1b, 7 of which are L2 (plus 1 U152). Though Germany may have received less of a north-eastern influence (the Latvia_BA signal in Unetice) than Czechia, it goes to show that a substantial proportion of the BB-era population did survive among uneticians.

[dancingfragments]

This part appears quite speculative on their end, but it's still an interesting interpretation of the Tollense battle:

The Late Bronze Age was a period of cultural transitions that were formative to the successive Iron Age, including the rise of intensified agriculture, metal craftsmanship and new forms of warfare. Archaeologically, this era saw the emergence and expansion of the Urnfield Culture (3300–2750 BP) across much of Europe. Several factors contributed to its expansion, including the aforementioned advancements in metalworking and agricultural techniques, which provided economic stability and surplus.

The introduction of new crops, such as millet, may have further aided population growth. This fostered expansion and facilitated the long-distance exchange of important commodities, including bronze, salt, and other raw materials.

In contrast, the northward expansion from Central Europe into northern Germany and southern Scandinavia appears to have culminated in a significant defeat, as suggested by the Tollense battle around 3200 BP and evident also from the lack of Knovíz related ancestry in Scandinavia prior to the Migration Period.

It seems to me that they mostly base this conclusion on the lack of Urnfield-related ancestry in Scandinavia until ~500 AD. Their interpretation could fit with the Weltzin samples carrying R1b-P312, R1b-CTS6889 which the Tollense sample shares with a Knoviz culture individual, and I2 lineages like Z2069, the latter which has also been found in a Tumulus culture context at Konobrže. 

This taken together with the rather southern-shifted profiles of the Tollense samples makes for a fairly convincing argument, in my opinion. Especially since we know that many of them (but not all) had non-local strontium isotope ratios.

Most of the Tollense samples available for analysis have high HG levels, which is generally unusual for Knoviz. In the graph, the Tollense warriors with high HG are located between Encrusted_Pottery and Trzciniec. It looks like they were part of a continuum of Middle Bronze Age cultures in Eastern Europe. This cannot be connected with the Celts. But perhaps the sample of Tollense warriors is skewed due to its insignificance and is insufficient to draw adequate conclusions.

[Riverman]

The high HG ones are Lusatians and Balto-Slavic people mentioned above. Lusatians are the majority, Balto-Slavic a small minority. But beside those two are R1b dominated ans TC derived warriors which the authors connected to Southern Germany and Czechia.

[dancingfragments]

The high HG ones are Lusatians and Balto-Slavic people mentioned above. Lusatians are the majority, Balto-Slavic a small minority. But beside those two are R1b dominated ans TC derived warriors which the authors connected to Southern Germany and Czechia.

I'm not sure I got it all right. It seems to me that Kisapostag, Trzciniec and Baltic_BA are all part of a larger picture that resulted from the mixing of different farming and hunting groups in Eastern Europe in the late Neolithic. I think there was another group like that, intermediate between Kisapostag and Baltic_BA, from which the main Tollense warrior cluster came. Then they mixed with the Corded Ware people, of course.

[nomad01]

The high HG ones are Lusatians and Balto-Slavic people mentioned above. Lusatians are the majority, Balto-Slavic a small minority. But beside those two are R1b dominated ans TC derived warriors which the authors connected to Southern Germany and Czechia.

I'm not sure I got it all right. It seems to me that Kisapostag, Trzciniec and Baltic_BA are all part of a larger picture that resulted from the mixing of different farming and hunting groups in Eastern Europe in the late Neolithic. I think there was another group like that, intermediate between Kisapostag and Baltic_BA, from which the main Tollense warrior cluster came. Then they mixed with the Corded Ware people, of course.

Tollense warriors are believed to come from Moravia (eastern Czechia). They look mostly like a mix of Kisapostag (I-L1229) and a BBC-derived group (R1b). But their R1b is mostly not the Italo-Celtic U152, but other, rare branches.

[dancingfragments]

I'm not sure I got it all right. It seems to me that Kisapostag, Trzciniec and Baltic_BA are all part of a larger picture that resulted from the mixing of different farming and hunting groups in Eastern Europe in the late Neolithic. I think there was another group like that, intermediate between Kisapostag and Baltic_BA, from which the main Tollense warrior cluster came. Then they mixed with the Corded Ware people, of course.

Tollense warriors are believed to come from Moravia (eastern Czechia). They look mostly like a mix of Kisapostag (I-L1229) and a BBC-derived group (R1b). But their R1b is mostly not the Italo-Celtic U152, but other, rare branches.

I thought so too, but a source like Kisapostag has a higher Balto-Slavic drift than WEZ58 and WEZ59, and their HG level is comparable. So when modeling WEZ58 and WEZ59 with Kisapostag, we would need additional HG. I would assume there was someone else besides Kisapostag.

[nomad01]

I'm not sure I got it all right. It seems to me that Kisapostag, Trzciniec and Baltic_BA are all part of a larger picture that resulted from the mixing of different farming and hunting groups in Eastern Europe in the late Neolithic. I think there was another group like that, intermediate between Kisapostag and Baltic_BA, from which the main Tollense warrior cluster came. Then they mixed with the Corded Ware people, of course.

Tollense warriors are believed to come from Moravia (eastern Czechia). They look mostly like a mix of Kisapostag (I-L1229) and a BBC-derived group (R1b). But their R1b is mostly not the Italo-Celtic U152, but other, rare branches.

I thought so too, but a source like Kisapostag has a higher Balto-Slavic drift than WEZ58 and WEZ59, and their HG level is comparable. So when modeling WEZ58 and WEZ59 with Kisapostag, we would need additional HG. I would assume there was someone else besides Kisapostag.

WEZ56 is R1a and he looks kind of similar to the Lusatians or Lithuania_BA. So maybe 58 and 59 have a bit of that admixture as well.

[Telemachus]

Hi guys! Do you think haplogroups such as R1b-U152 and G-L497 in the balkans were from the celtic or germanic migrations? considering the fact there were many celtic groups in illyria and rivers in montenegro and serbia such as Tara ans Lim trace their names to celtic languages. Even specific tribes such as the Mataruga tribe that trace their name to a spear brought by the celts.

Celts, including Romanized Celts, and descendants of Germanic tribes are plausible candidates for G-L497's presence in the Balkans. Sample I15516 (127-233 CE; Viminacium, Pirivoj Necropolis, Serbia: G-Z727) plots relatively closely to Iron Age Celtic groups and modern Cornish and Belgian samples in G25, while G-L497 was present among a decent number of known Gothic samples (e.g., 11 percent of the Wielbark samples were G2a, most of which were confirmed to be downstream from G-L497). A noteworthy lineage for Balkan G-L497s, containing about half of the Balkan G-L497s in FTDNA, is G-L42 (see FTDNA and YFull).

[Kyu]

Proto-Italic is thought to be quite young, only having split a bit earlier than 1000 BC, so that fits a more recent origin

If the Urnfield era is considered late or disintegrating Proto Celtic then most likely Proto Italic already existed in this period (1200-850) or even before it. The key lies in the dating and duration of the Italo-Celtic. How long did it last? From what I gather there are 2 competing theories; a brief early node that probably ended by or during the EBA, or a much longer period of existence for Italo-Celtic that possibly lasted until the Urnfield era  (which I feel is less likely).

Another clue could be dating the split of (proto) Celtiberian from Proto Celtic. Did it happen before or during the appearance of Urnfields in Iberia or was it the result of a later migration? Did proto Celtiberian arrive in Iberia fully formed or was it a variety of Proto Celtic that the migrants brought with them, that then evolved in situ in Iberia over many centuries into Celtiberian? (And afaik some Iberian linguists contemplate that there were other Hispano-Celtic languages and not just Celtiberian).

If the 2nd Celtic language to break off was Lepontic and the earliest attestation of Lepontic was in the 6th or 5th century BC then proto Celtiberian had to of split before the earliest proposed dates for the inscriptions bearing the Lepontic language. As Jaska would probably say one cannot 'see' language from DNA or Archaeology, but an idea of continuity in material culture from Canegrate to Golasecca would make it harder to dismiss an even earlier date for the split of Lepontic from Common Celtic. But if there were interruptions in the settlement pattern in the long long timeline from Canegrate to the end of Golasecca culture, then in theory that could possibly imply a much later intrusion of Celts into Northern Italy but still being before the supposed Celtic invasions of the La Tene era

It would also help to roughly date the formation of the (central!?) Italic genepool which afaik is directly ancestral to the IA Italics and Etruscans. As Jaska would say you cant 'see' language from DNA, but if the (central!?) Italic genepool stayed roughly the same from the BA to the IA, it could be indicative of at least a BA date for Proto Italic.

I think Italic and Celtic could have split during the EBA-MBA transition, but both would have generally remained within Tumulus and Urnfield which could explain some shared features which can't be from Proto-Italo-Celtic, but Italic probably did start to expand to Italy during this time considering the links Terramare had with Pannonia

Celtiberian is probably from Urnfield, and would have expanded more into Iberia with the end of Cogotas I

[Jaska]

As Jaska would probably say one cannot 'see' language from DNA or Archaeology, but an idea of continuity in material culture from Canegrate to Golasecca would make it harder to dismiss an even earlier date for the split of Lepontic from Common Celtic. But if there were interruptions in the settlement pattern in the long long timeline from Canegrate to the end of Golasecca culture, then in theory that could possibly imply a much later intrusion of Celts into Northern Italy but still being before the supposed Celtic invasions of the La Tene era

It would also help to roughly date the formation of the (central!?) Italic genepool which afaik is directly ancestral to the IA Italics and Etruscans. As Jaska would say you cant 'see' language from DNA, but if the (central!?) Italic genepool stayed roughly the same from the BA to the IA, it could be indicative of at least a BA date for Proto Italic.

Someone called me?
Genetic or archaeological continuity is an indicator but not proof. Such continuity is at least partially evident practically everywhere, and everywhere someone has claimed that this continuity proves linguistic continuity. Such claims are unfortunately mutually contradictory (as the same continuity can be linked to several different language families in the region), which shows us that the method is totally unreliable and scientifically worthless.

Moreover, we know that the linguistic landscape is everywhere much younger than the genetic or cultural continuity would let us assume. As the result, genetic or archaeological continuity often corresponds to linguistic discontinuity – not linguistic continuity.

What continuity can tell us is that there was no total population replacement in the region. But continuity cannot tell the result of language contacts during the admixture process. There are several possible outcomes:
1. Older language prevailed, and the newcomers adopted it.
2. Older language influenced the newcomer language (loanwords and other substrate features).
3. Older language left no trace in the newcomer language.

Although we can say that the first option is more likely, if the older population and culture also clearly prevailed against the newcomers (“strong continuity”), there is no way to ascertain the result from the genetic or archaeological data alone. Only linguistic results can tell us since when certain language was spoken in certain region. Even if there is a superficial contradiction (strong genetic and cultural continuity seem to correspond to linguistic discontinuity), the results just must be accepted. There is nothing impossible in such a result, as we know from the documented cases of language shifts.