Yamnaya ancestry: mapping the Proto-Indo-European expansions

steppe-ancestry-expansion-europe

The latest papers from Ning et al. Cell (2019) and Anthony JIES (2019) have offered some interesting new data, supporting once more what could be inferred since 2015, and what was evident in population genomics since 2017: that Proto-Indo-Europeans expanded under R1b bottlenecks, and that the so-called “Steppe ancestry” referred to two different components, one – Yamnaya or Steppe_EMBA ancestry – expanding with Proto-Indo-Europeans, and the other one – Corded Ware or Steppe_MLBA ancestry – expanding with Uralic speakers.

The following maps are based on formal stats published in the papers and supplementary materials from 2015 until today, mainly on Wang et al. (2018 & 2019), Mathieson et al. (2018) and Olalde et al. (2018), and others like Lazaridis et al. (2016), Lazaridis et al. (2017), Mittnik et al. (2018), Lamnidis et al. (2018), Fernandes et al. (2018), Jeong et al. (2019), Olalde et al. (2019), etc.

NOTE. As in the Corded Ware ancestry maps, the selected reports in this case are centered on the prototypical Yamnaya ancestry vs. other simplified components, so everything else refers to simplistic ancestral components widespread across populations that do not necessarily share any recent connection, much less a language. In fact, most of the time they clearly didn’t. They can be interpreted as “EHG that is not part of the Yamnaya component”, or “CHG that is not part of the Yamnaya component”. They can’t be read as “expanding EHG people/language” or “expanding CHG people/language”, at least no more than maps of “Steppe ancestry” can be read as “expanding Steppe people/language”. Also, remember that I have left the default behaviour for color classification, so that the highest value (i.e. 1, or white colour) could mean anything from 10% to 100% depending on the specific ancestry and period; that’s what the legend is for… But, fere libenter homines id quod volunt credunt.

Sections:

  1. Neolithic or the formation of Early Indo-European
  2. Eneolithic or the expansion of Middle Proto-Indo-European
  3. Chalcolithic / Early Bronze Age or the expansion of Late Proto-Indo-European
  4. European Early Bronze Age and MLBA or the expansion of Late PIE dialects

1. Neolithic

Anthony (2019) agrees with the most likely explanation of the CHG component found in Yamnaya, as derived from steppe hunter-fishers close to the lower Volga basin. The ultimate origin of this specific CHG-like component that eventually formed part of the Pre-Yamnaya ancestry is not clear, though:

The hunter-fisher camps that first appeared on the lower Volga around 6200 BC could represent the migration northward of un-admixed CHG hunter-fishers from the steppe parts of the southeastern Caucasus, a speculation that awaits confirmation from aDNA.

neolithic-chg-ancestry
Natural neighbor interpolation of CHG ancestry among Neolithic populations. See full map.

The typical EHG component that formed part eventually of Pre-Yamnaya ancestry came from the Middle Volga Basin, most likely close to the Samara region, as shown by the sampled Samara hunter-gatherer (ca. 5600-5500 BC):

After 5000 BC domesticated animals appeared in these same sites in the lower Volga, and in new ones, and in grave sacrifices at Khvalynsk and Ekaterinovka. CHG genes and domesticated animals flowed north up the Volga, and EHG genes flowed south into the North Caucasus steppes, and the two components became admixed.

neolithic-ehg-ancestry
Natural neighbor interpolation of EHG ancestry among Neolithic populations. See full map.

To the west, in the Dnieper-Dniester area, WHG became the dominant ancestry after the Mesolithic, at the expense of EHG, revealing a likely mating network reaching to the north into the Baltic:

Like the Mesolithic and Neolithic populations here, the Eneolithic populations of Dnieper-Donets II type seem to have limited their mating network to the rich, strategic region they occupied, centered on the Rapids. The absence of CHG shows that they did not mate frequently if at all with the people of the Volga steppes (…)

neolithic-whg-ancestry
Natural neighbor interpolation of WHG ancestry among Neolithic populations. See full map.

North-West Anatolia Neolithic ancestry, proper of expanding Early European farmers, is found up to border of the Dniester, as Anthony (2007) had predicted.

neolithic-anatolia-farmer-ancestry
Natural neighbor interpolation of Anatolia Neolithic ancestry among Neolithic populations. See full map.

2. Eneolithic

From Anthony (2019):

After approximately 4500 BC the Khvalynsk archaeological culture united the lower and middle Volga archaeological sites into one variable archaeological culture that kept domesticated sheep, goats, and cattle (and possibly horses). In my estimation, Khvalynsk might represent the oldest phase of PIE.

(…) this middle Volga mating network extended down to the North Caucasian steppes, where at cemeteries such as Progress-2 and Vonyuchka, dated 4300 BC, the same Khvalynsk-type ancestry appeared, an admixture of CHG and EHG with no Anatolian Farmer ancestry, with steppe-derived Y-chromosome haplogroup R1b. These three individuals in the North Caucasus steppes had higher proportions of CHG, overlapping Yamnaya. Without any doubt, a CHG population that was not admixed with Anatolian Farmers mated with EHG populations in the Volga steppes and in the North Caucasus steppes before 4500 BC. We can refer to this admixture as pre-Yamnaya, because it makes the best currently known genetic ancestor for EHG/CHG R1b Yamnaya genomes.

From Wang et al (2019):

Three individuals from the sites of Progress 2 and Vonyuchka 1 in the North Caucasus piedmont steppe (‘Eneolithic steppe’), which harbour EHG and CHG related ancestry, are genetically very similar to Eneolithic individuals from Khvalynsk II and the Samara region. This extends the cline of dilution of EHG ancestry via CHG-related ancestry to sites immediately north of the Caucasus foothills

eneolithic-pre-yamnaya-ancestry
Natural neighbor interpolation of Pre-Yamnaya ancestry among Neolithic populations. See full map. This map corresponds roughly to the map of Khvalynsk-Novodanilovka expansion, and in particular to the expansion of horse-head pommel-scepters (read more about Khvalynsk, and specifically about horse symbolism)

NOTE. Unpublished samples from Ekaterinovka have been previously reported as within the R1b-L23 tree. Interestingly, although the Varna outlier is a female, the Balkan outlier from Smyadovo shows two positive SNP calls for hg. R1b-M269. However, its poor coverage makes its most conservative haplogroup prediction R-M343.

The formation of this Pre-Yamnaya ancestry sets this Volga-Caucasus Khvalynsk community apart from the rest of the EHG-like population of eastern Europe.

eneolithic-ehg-ancestry
Natural neighbor interpolation of non-Pre-Yamnaya EHG ancestry among Eneolithic populations. See full map.

Anthony (2019) seems to rely on ADMIXTURE graphics when he writes that the late Sredni Stog sample from Alexandria shows “80% Khvalynsk-type steppe ancestry (CHG&EHG)”. While this seems the most logical conclusion of what might have happened after the Suvorovo-Novodanilovka expansion through the North Pontic steppes (see my post on “Steppe ancestry” step by step), formal stats have not confirmed that.

In fact, analyses published in Wang et al. (2019) rejected that Corded Ware groups are derived from this Pre-Yamnaya ancestry, a reality that had been already hinted in Narasimhan et al. (2018), when Steppe_EMBA showed a poor fit for expanding Srubna-Andronovo populations. Hence the need to consider the whole CHG component of the North Pontic area separately:

eneolithic-chg-ancestry
Natural neighbor interpolation of non-Pre-Yamnaya CHG ancestry among Eneolithic populations. See full map. You can read more about population movements in the late Sredni Stog and closer to the Proto-Corded Ware period.

NOTE. Fits for WHG + CHG + EHG in Neolithic and Eneolithic populations are taken in part from Mathieson et al. (2019) supplementary materials (download Excel here). Unfortunately, while data on the Ukraine_Eneolithic outlier from Alexandria abounds, I don’t have specific data on the so-called ‘outlier’ from Dereivka compared to the other two analyzed together, so these maps of CHG and EHG expansion are possibly showing a lesser distribution to the west than the real one ca. 4000-3500 BC.

eneolithic-whg-ancestry
Natural neighbor interpolation of WHG ancestry among Eneolithic populations. See full map.

Anatolia Neolithic ancestry clearly spread to the east into the north Pontic area through a Middle Eneolithic mating network, most likely opened after the Khvalynsk expansion:

eneolithic-anatolia-farmer-ancestry
Natural neighbor interpolation of Anatolia Neolithic ancestry among Eneolithic populations. See full map.
eneolithic-iran-chl-ancestry
Natural neighbor interpolation of Iran Chl. ancestry among Eneolithic populations. See full map.

Regarding Y-chromosome haplogroups, Anthony (2019) insists on the evident association of Khvalynsk, Yamnaya, and the spread of Pre-Yamnaya and Yamnaya ancestry with the expansion of elite R1b-L754 (and some I2a2) individuals:

eneolithic-early-y-dna
Y-DNA haplogroups in West Eurasia during the Early Eneolithic in the Pontic-Caspian steppes. See full map, and see culture, ADMIXTURE, Y-DNA, and mtDNA maps of the Early Eneolithic and Late Eneolithic.

3. Early Bronze Age

Data from Wang et al. (2019) show that Corded Ware-derived populations do not have good fits for Eneolithic_Steppe-like ancestry, no matter the model. In other words: Corded Ware populations show not only a higher contribution of Anatolia Neolithic ancestry (ca. 20-30% compared to the ca. 2-10% of Yamnaya); they show a different EHG + CHG combination compared to the Pre-Yamnaya one.

eneolithic-steppe-best-fits
Supplementary Table 13. P values of rank=2 and admixture proportions in modelling Steppe ancestry populations as a three-way admixture of Eneolithic steppe Anatolian_Neolithic and WHG using 14 outgroups.
Left populations: Test, Eneolithic_steppe, Anatolian_Neolithic, WHG.
Right populations: Mbuti.DG, Ust_Ishim.DG, Kostenki14, MA1, Han.DG, Papuan.DG, Onge.DG, Villabruna, Vestonice16, ElMiron, Ethiopia_4500BP.SG, Karitiana.DG, Natufian, Iran_Ganj_Dareh_Neolithic.

Yamnaya Kalmykia and Afanasievo show the closest fits to the Eneolithic population of the North Caucasian steppes, rejecting thus sizeable contributions from Anatolia Neolithic and/or WHG, as shown by the SD values. Both probably show then a Pre-Yamnaya ancestry closest to the late Repin population.

wang-eneolithic-steppe-caucasus-yamnaya
Modelling results for the Steppe and Caucasus cluster. Admixture proportions based on (temporally and geographically) distal and proximal models, showing additional AF ancestry in Steppe groups and additional gene flow from the south in some of the Steppe groups as well as the Caucasus groups. See tables above. Modified from Wang et al. (2019). Within a blue square, Yamnaya-related groups; within a cyan square, Corded Ware-related groups. Green background behind best p-values. In red circle, SD of AF/WHG ancestry contribution in Afanasevo and Yamnaya Kalmykia, with ranges that almost include 0%.

EBA maps include data from Wang et al. (2018) supplementary materials, specifically unpublished Yamnaya samples from Hungary that appeared in analysis of the preprint, but which were taken out of the definitive paper. Their location among Yamnaya settlers from Hungary is speculative, although most uncovered kurgans in Hungary are concentrated in the Tisza-Danube interfluve.

eba-yamnaya-ancestry
Natural neighbor interpolation of Pre-Yamnaya ancestry among Early Bronze Age populations. See full map. This map corresponds roughly with the known expansion of late Repin/Yamnaya settlers.

The Y-chromosome bottleneck of elite males from Proto-Indo-European clans under R1b-L754 and some I2a2 subclades, already visible in the Khvalynsk sampling, became even more noticeable in the subsequent expansion of late Repin/early Yamnaya elites under R1b-L23 and I2a-L699:

chalcolithic-early-y-dna
Y-DNA haplogroups in West Eurasia during the Yamnaya expansion. See full map and maps of cultures, ADMIXTURE, Y-DNA, and mtDNA of the Early Chalcolithic and Yamnaya Hungary.

Maps of CHG, EHG, Anatolia Neolithic, and probably WHG show the expansion of these components among Corded Ware-related groups in North Eurasia, apart from other cultures close to the Caucasus:

NOTE. For maps with actual formal stats of Corded Ware ancestry from the Early Bronze Age to the modern times, you can read the post Corded Ware ancestry in North Eurasia and the Uralic expansion.

eba-chg-ancestry
Natural neighbor interpolation of non-Pre-Yamnaya CHG ancestry among Early Bronze Age populations. See full map.
eba-ehg-ancestry
Natural neighbor interpolation of non-Pre-Yamnaya EHG ancestry among Early Bronze Age populations. See full map.
eba-whg-ancestry
Natural neighbor interpolation of WHG ancestry among Early Bronze Age populations. See full map.
eba-anatolia-farmer-ancestry
Natural neighbor interpolation of Anatolia Neolithic ancestry among Early Bronze Age populations. See full map.
eba-iran-chl-ancestry
Natural neighbor interpolation of Iran Chl. ancestry among Early Bronze Age populations. See full map.

4. Middle to Late Bronze Age

The following maps show the most likely distribution of Yamnaya ancestry during the Bell Beaker-, Balkan-, and Sintashta-Potapovka-related expansions.

4.1. Bell Beakers

The amount of Yamnaya ancestry is probably overestimated among populations where Bell Beakers replaced Corded Ware. A map of Yamnaya ancestry among Bell Beakers gets trickier for the following reasons:

  • Expanding Repin peoples of Pre-Yamnaya ancestry must have had admixture through exogamy with late Sredni Stog/Proto-Corded Ware peoples during their expansion into the North Pontic area, and Sredni Stog in turn had probably some Pre-Yamnaya admixture, too (although they don’t appear in the simplistic formal stats above). This is supported by the increase of Anatolia farmer ancestry in more western Yamna samples.
  • Later, Yamnaya admixed through exogamy with Corded Ware-like populations in Central Europe during their expansion. Even samples from the Middle to Upper Danube and around the Lower Rhine will probably show increasing contributions of Steppe_MLBA, at the same time as they show an increasing proportion of EEF-related ancestry.
  • To complicate things further, the late Corded Ware Espersted family (from ca. 2500 BC or later) shows, in turn, what seems like a recent admixture with Yamnaya vanguard groups, with the sample of highest Yamnaya ancestry being the paternal uncle of other individuals (all of hg. R1a-M417), suggesting that there might have been many similar Central European mating networks from the mid-3rd millennium BC on, of (mainly) Yamnaya-like R1b elites displaying a small proportion of CW-like ancestry admixing through exogamy with Corded Ware-like peoples who already had some Yamnaya ancestry.
mlba-yamnaya-ancestry
Natural neighbor interpolation of Yamnaya ancestry among Middle to Late Bronze Age populations (Esperstedt CWC site close to BK_DE, label is hidden by BK_DE_SAN). See full map. You can see how this map correlated with the map of Late Copper Age migrations and Yamanaya into Bell Beaker expansion.

NOTE. Terms like “exogamy”, “male-driven migration”, and “sex bias”, are not only based on the Y-chromosome bottlenecks visible in the different cultural expansions since the Palaeolithic. Despite the scarce sampling available in 2017 for analysis of “Steppe ancestry”-related populations, it appeared to show already a male sex bias in Goldberg et al. (2017), and it has been confirmed for Neolithic and Copper Age population movements in Mathieson et al. (2018) – see Supplementary Table 5. The analysis of male-biased expansion of “Steppe ancestry” in CWC Esperstedt and Bell Beaker Germany is, for the reasons stated above, not very useful to distinguish their mutual influence, though.

Based on data from Olalde et al. (2019), Bell Beakers from Germany are the closest sampled ones to expanding East Bell Beakers, and those close to the Rhine – i.e. French, Dutch, and British Beakers in particular – show a clear excess “Steppe ancestry” due to their exogamy with local Corded Ware groups:

Only one 2-way model fits the ancestry in Iberia_CA_Stp with P-value>0.05: Germany_Beaker + Iberia_CA. Finding a Bell Beaker-related group as a plausible source for the introduction of steppe ancestry into Iberia is consistent with the fact that some of the individuals in the Iberia_CA_Stp group were excavated in Bell Beaker associated contexts. Models with Iberia_CA and other Bell Beaker groups such as France_Beaker (P-value=7.31E-06), Netherlands_Beaker (P-value=1.03E-03) and England_Beaker (P-value=4.86E-02) failed, probably because they have slightly higher proportions of steppe ancestry than the true source population.

olalde-iberia-chalcolithic

The exogamy with Corded Ware-like groups in the Lower Rhine Basin seems at this point undeniable, as is the origin of Bell Beakers around the Middle-Upper Danube Basin from Yamnaya Hungary.

To avoid this excess “Steppe ancestry” showing up in the maps, since Bell Beakers from Germany pack the most Yamnaya ancestry among East Bell Beakers outside Hungary (ca. 51.1% “Steppe ancestry”), I equated this maximum with BK_Scotland_Ach (which shows ca. 61.1% “Steppe ancestry”, highest among western Beakers), and applied a simple rule of three for “Steppe ancestry” in Dutch and British Beakers.

NOTE. Formal stats for “Steppe ancestry” in Bell Beaker groups are available in Olalde et al. (2018) supplementary materials (PDF). I didn’t apply this adjustment to Bk_FR groups because of the R1b Bell Beaker sample from the Champagne/Alsace region reported by Samantha Brunel that will pack more Yamnaya ancestry than any other sampled Beaker to date, hence probably driving the Yamnaya ancestry up in French samples.

The most likely outcome in the following years, when Yamnaya and Corded Ware ancestry are investigated separately, is that Yamnaya ancestry will be much lower the farther away from the Middle and Lower Danube region, similar to the case in Iberia, so the map above probably overestimates this component in most Beakers to the north of the Danube. Even the late Hungarian Beaker samples, who pack the highest Yamnaya ancestry (up to 75%) among Beakers, represent likely a back-migration of Moravian Beakers, and will probably show a contribution of Corded Ware ancestry due to the exogamy with local Moravian groups.

Despite this decreasing admixture as Bell Beakers spread westward, the explosive expansion of Yamnaya R1b male lineages (in words of David Reich) and the radical replacement of local ones – whether derived from Corded Ware or Neolithic groups – shows the true extent of the North-West Indo-European expansion in Europe:

chalcolithic-late-y-dna
Y-DNA haplogroups in West Eurasia during the Bell Beaker expansion. See full map and see maps of cultures, ADMIXTURE, Y-DNA, and mtDNA of the Late Copper Age and of the Yamnaya-Bell Beaker transition.

4.2. Palaeo-Balkan

There is scarce data on Palaeo-Balkan movements yet, although it is known that:

  1. Yamnaya ancestry appears among Mycenaeans, with the Yamnaya Bulgaria sample being its best current ancestral fit;
  2. the emergence of steppe ancestry and R1b-M269 in the eastern Mediterranean was associated with Ancient Greeks;
  3. Thracians, Albanians, and Armenians also show R1b-M269 subclades and “Steppe ancestry”.

4.3. Sintashta-Potapovka-Filatovka

Interestingly, Potapovka is the only Corded Ware derived culture that shows good fits for Yamnaya ancestry, despite having replaced Poltavka in the region under the same Corded Ware-like (Abashevo) influence as Sintashta.

This proves that there was a period of admixture in the Pre-Proto-Indo-Iranian community between CWC-like Abashevo and Yamnaya-like Catacomb-Poltavka herders in the Sintashta-Potapovka-Filatovka community, probably more easily detectable in this group because of the specific temporal and geographic sampling available.

srubnaya-yamnaya-ehg-chg-ancestry
Supplementary Table 14. P values of rank=3 and admixture proportions in modelling Steppe ancestry populations as a four-way admixture of distal sources EHG, CHG, Anatolian_Neolithic and WHG using 14 outgroups.
Left populations: Steppe cluster, EHG, CHG, WHG, Anatolian_Neolithic
Right populations: Mbuti.DG, Ust_Ishim.DG, Kostenki14, MA1, Han.DG, Papuan.DG, Onge.DG, Villabruna, Vestonice16, ElMiron, Ethiopia_4500BP.SG, Karitiana.DG, Natufian, Iran_Ganj_Dareh_Neolithic.

Srubnaya ancestry shows a best fit with non-Pre-Yamnaya ancestry, i.e. with different CHG + EHG components – possibly because the more western Potapovka (ancestral to Proto-Srubnaya Pokrovka) also showed good fits for it. Srubnaya shows poor fits for Pre-Yamnaya ancestry probably because Corded Ware-like (Abashevo) genetic influence increased during its formation.

On the other hand, more eastern Corded Ware-derived groups like Sintashta and its more direct offshoot Andronovo show poor fits with this model, too, but their fits are still better than those including Pre-Yamnaya ancestry.

mlba-ehg-ancestry
Natural neighbor interpolation of non-Pre-Yamnaya EHG ancestry among Middle to Late Bronze Age populations. See full map.
mlba-chg-ancestry
Natural neighbor interpolation of non-Pre-Yamnaya CHG ancestry among Middle to Late Bronze Age populations. See full map.
mlba-anatolia-farmer-ancestry
Natural neighbor interpolation of Anatolia Neolithic ancestry among Middle to Late Bronze Age populations. See full map.
mlba-iran-chl-ancestry
Natural neighbor interpolation of Iran Chl. ancestry among Middle to Late Bronze Age populations. See full map.

NOTE For maps with actual formal stats of Corded Ware ancestry from the Early Bronze Age to the modern times, you should read the post Corded Ware ancestry in North Eurasia and the Uralic expansion instead.

The bottleneck of Proto-Indo-Iranians under R1a-Z93 was not yet complete by the time when the Sintashta-Potapovka-Filatovka community expanded with the Srubna-Andronovo horizon:

early-bronze-age-y-dna
Y-DNA haplogroups in West Eurasia during the European Early Bronze Age. See full map and see maps of cultures, ADMIXTURE, Y-DNA, and mtDNA of the Early Bronze Age.

4.4. Afanasevo

At the end of the Afanasevo culture, at least three samples show hg. Q1b (ca. 2900-2500 BC), which seemed to point to a resurgence of local lineages, despite continuity of the prototypical Pre-Yamnaya ancestry. On the other hand, Anthony (2019) makes this cryptic statement:

Yamnaya men were almost exclusively R1b, and pre-Yamnaya Eneolithic Volga-Caspian-Caucasus steppe men were principally R1b, with a significant Q1a minority.

Since the only available samples from the Khvalynsk community are R1b (x3), Q1a(x1), and R1a(x1), it seems strange that Anthony would talk about a “significant minority”, unless Q1a (potentially Q1b in the newer nomenclature) will pop up in some more individuals of those ca. 30 new to be published. Because he also mentions I2a2 as appearing in one elite burial, it seems Q1a (like R1a-M459) will not appear under elite kurgans, although it is still possible that hg. Q1a was involved in the expansion of Afanasevo to the east.

middle-bronze-age-y-dna
Y-DNA haplogroups in West Eurasia during the Middle Bronze Age. See full map and see maps of cultures, ADMIXTURE, Y-DNA, and mtDNA of the Middle Bronze Age and the Late Bronze Age.

Okunevo, which replaced Afanasevo in the Altai region, shows a majority of hg. Q1b, but also some R1b-M269 samples proper of Afanasevo, suggesting partial genetic continuity.

NOTE. Other sampled Siberian populations clearly show a variety of Q subclades that likely expanded during the Palaeolithic, such as Baikal EBA samples from Ust’Ida and Shamanka with a majority of Q1b, and hg. Q reported from Elunino, Sagsai, Khövsgöl, and also among peoples of the Srubna-Andronovo horizon (the Krasnoyarsk MLBA outlier), and in Karasuk.

From Damgaard et al. Science (2018):

(…) in contrast to the lack of identifiable admixture from Yamnaya and Afanasievo in the CentralSteppe_EMBA, there is an admixture signal of 10 to 20% Yamnaya and Afanasievo in the Okunevo_EMBA samples, consistent with evidence of western steppe influence. This signal is not seen on the X chromosome (qpAdm P value for admixture on X 0.33 compared to 0.02 for autosomes), suggesting a male-derived admixture, also consistent with the fact that 1 of 10 Okunevo_EMBA males carries a R1b1a2a2 Y chromosome related to those found in western pastoralists. In contrast, there is no evidence of western steppe admixture among the more eastern Baikal region region Bronze Age (~2200 to 1800 BCE) samples.

This Yamnaya ancestry has been also recently found to be the best fit for the Iron Age population of Shirenzigou in Xinjiang – where Tocharian languages were attested centuries later – despite the haplogroup diversity acquired during their evolution, likely through an intermediate Chemurchek culture (see a recent discussion on the elusive Proto-Tocharians).

Haplogroup diversity seems to be common in Iron Age populations all over Eurasia, most likely due to the spread of different types of sociopolitical structures where alliances played a more relevant role in the expansion of peoples. A well-known example of this is the spread of Akozino warrior-traders in the whole Baltic region under a partial N1a-VL29-bottleneck associated with the emerging chiefdom-based systems under the influence of expanding steppe nomads.

early-iron-age-y-dna
Y-DNA haplogroups in West Eurasia during the Early Iron Age. See full map and see maps of cultures, ADMIXTURE, Y-DNA, and mtDNA of the Early Iron Age and Late Iron Age.

Surprisingly, then, Proto-Tocharians from Shirenzigou pack up to 74% Yamnaya ancestry, in spite of the 2,000 years that separate them from the demise of the Afanasevo culture. They show more Yamnaya ancestry than any other population by that time, being thus a sort of Late PIE fossils not only in their archaic dialect, but also in their genetic profile:

shirenzigou-afanasievo-yamnaya-andronovo-srubna-ulchi-han

The recent intrusion of Corded Ware-like ancestry, as well as the variable admixture with Siberian and East Asian populations, both point to the known intense Old Iranian and Old/Middle Chinese contacts. The scarce Proto-Samoyedic and Proto-Turkic loans in Tocharian suggest a rather loose, probably more distant connection with East Uralic and Altaic peoples from the forest-steppe and steppe areas to the north (read more about external influences on Tocharian).

Interestingly, both R1b samples, MO12 and M15-2 – likely of Asian R1b-PH155 branch – show a best fit for Andronovo/Srubna + Hezhen/Ulchi ancestry, suggesting a likely connection with Iranians to the east of Xinjiang, who later expanded as the Wusun and Kangju. How they might have been related to Huns and Xiongnu individuals, who also show this haplogroup, is yet unknown, although Huns also show hg. R1a-Z93 (probably most R1a-Z2124) and Steppe_MLBA ancestry, earlier associated with expanding Iranian peoples of the Srubna-Andronovo horizon.

All in all, it seems that prehistoric movements explained through the lens of genetic research fit perfectly well the linguistic reconstruction of Proto-Indo-European and Proto-Uralic.

Related

Genetic continuity among Uralic-speaking cultures in north-eastern Europe

east-europe-bronze-age

The recent study of Estonian Late Bronze Age/Iron Age samples has shown, as expected, large genetic continuity of Corded Ware populations in the East Baltic area, where West Uralic is known to have been spoken since at least the Early Bronze Age.

The most interesting news was that, unexpectedly for many, the impact of “Siberian ancestry” (whatever that actually means) was small, slow, and gradual, with slight increases found up to the Middle Ages, compatible with multiple contact events in north-eastern Europe. Haplogroup N became prevalent among Finnic populations only through late bottlenecks, as research of modern populations have long suggested, and as ancient DNA research hinted since at least 2015.

I risked to correlate the arrival of chiefs from the south-west with the infiltration of N1c-VL29 subclades during the transition to the Iron Age, coupled with that minimal “Siberian” ancestry (see e.g. here and here). Now we know that the penetration of this non-CW ancestry started, as predicted, in the Iron Age; that it was highly variable in the few samples where it appeared, with ca. 1-4%, while most Iron Age individuals show 0%; and that it was not especially linked to individuals of N1c-Vl29 lineages.

It is also basically confirmed, based on the (ancient and Modern Swedish) N1c-L550 subclades found among Iron Age Estonians, that N1c-VL29 lineages and the so-called “Siberian” ancestry will be found simultaneously around the Baltic coastal areas, and that different lineages must have suffered later founder effects among Finns, which suggests that these alliances through exogamy brought exactly as much language change in Sweden, Lithuania, or Poland, as they did in the East Baltic region…

On the other hand, the paper has also shown a potential movement of Corded Ware-derived peoples, if the change from LBA to IA samples is meaningful; in fact, even more Corded Ware-like than Baltic and Estonian BA populations. The exact origin of that movement is difficult to pinpoint, and it may not be related to the arrival of Akozino warrior-traders from the south-east, since theirs seems to be a minor impact proper of elites in a chiefdom system around the Baltic.

fortified-settlements-lba-ia
Distribution of fortified settlements (filled circles) and other hilltop sites (empty circles) of the Late Bronze Age and Pre-Roman Iron Ages in the East Baltic region. Tentative area of most intensive contacts between Baltic and Balto-Finnic communities marked with a dashed line. Image modified from (Lang 2016).

Also suggesting a potential movement is the ‘southern’ shift observed in the West and East Baltic areas, likely showing the arrival of Proto-East Baltic speakers (such as the Trzciniec outlier), as we have already discussed in this blog. The unexpected increase in Corded Ware-like ancestry in the Eastern Baltic, coupled with the expected large continuity of hg. R1a-Z283 in the homeland of Balto-Finnic expansions, gives even more support to the known complex system of exogamy along the Baltic coasts, and offers another potential reason for the rise of Baltic-speaking territories in the West Baltic: elite domination.

It is nevertheless important to understand that, even among the most “genetic continuous” regions like Estonia, not a single population in Europe is heir of some ancestral, immutable people. Not in terms of haplogroups, and not in terms of admixture. Balto-Finnic speakers, however continuous they might seem (e.g. in Southern Estonians) aren’t an exception.

After all, this blog was (re)born to fight the currently prevalent sheer stupidity surrounding the simplistic “R1a/steppe ancestry=Indo-European” association, so I wouldn’t like to see it replaced with some other stupid continuity or purity ideas within 10 to 20 years…

Late Uralic stems from East Corded Ware groups

With the currently available tools – linguistics, archaeology, and now genetics -, I don’t think there is any argument to date to question the direct connection of the Late Proto-Uralic expansion with all Eastern Corded Ware groups (i.e. Battle Axe, Fatyanovo-Balanovo, and Abashevo), and thus at least with the unifying A-horizon of Corded Ware and the bottlenecks under R1a-Z645.

NOTE. The only out-group among Corded Ware cultures is the Single Grave culture. It appears to be an early Corded Ware offshoot, reflected in their non-unitary cultural traits (distinct from later unifying waves), in their varied patrilineal clans, and in the short-lasting cultural effect in northern Europe before their complete demise under pressure of expanding Yamna/Bell Beaker peoples from the Danube. The culture’s minimal (if any) effects on succeeding peoples might be seen mostly in the (mainly phonetic) Uralic substrate found in Balto-Slavic – although this may also stem from a more eastern influence, close to the Baltic – and in the contacts of Celtic with Uralic. The huge time depth between this early hypothetic Uralic layer in northern Europe and the emergence of peoples inhabiting these territories in recorded history have no doubt been erroneously interpreted as a lack of Uralic presence in the area.

1) That connection was evident in the Yamna – CWC differences in archaeology, and especially later, with at least Fatyanovo-Balanovo and Abashevo representing the obvious replacement of the Volosovo culture before further expansions of CWC-related groups west and east of the Urals.

The mythical millennia-long continuity of Volosovo hunter-gatherers, including centuries among Corded Ware peoples, as expected lately by the Copenhagen group (and anyone who doesn’t want to question the 1960s association of Indo-European with CWC) must be rejected today in population genomics, as the recent studies of ancient and modern populations show, and as ancient DNA from the region will confirm.

2) In linguistics, the survival of Volosovo as The Uralic-speaking culture was also hardly believable. From Kallio (2015):

While we can say at least something about Uralic substrates in Northeastern Europe, non-Uralic substrates cannot at all easily be identified, because of multiple language shifts, viz. first from non-Uralic to Uralic and then from Uralic to Russian. Yet the Soviet Uralicist Boris Serebrennikov (1956, 1959) argued that there are some non-Uralic substrate toponyms in the Volga-Oka region, but his idea was never taken seriously in the west (cf. Sauvageot 1958), and it pretty soon also sank into oblivion in Russia, even though it can still occasionally pop up there in non-onomastic circles (cf. Napolskikh 1995: 18–19). However, not all the hypotheses on non-Uralic substrates in Northeastern Europe should be rejected (see e.g. Helimski 2001b).

bronze-age-early-languages-east-europe
Tentative map of the distribution of known languages in Eastern Europe during the Early Bronze Age. See full map.

Helimski (2001) argues for a non-Uralic topo-hydronomy in Northern Russia, whose population may have kept their languages up to the Common Era despite the Corded Ware expansion, which is in line with the survival of some non-Indo-European languages everywhere in Europe after the expansion of Yamna and its offshoots:

It should be borne in mind that these [Uralic] hydronyms reached us mainly through Northern Russian and, accordingly, with a tendency to phonetic-morphological adaptation and unification (for river names it is “natural” to be, like the word ‘river’ itself, feminine and to end in -a). Taking into account this circumstance, it may turn out to be non-useless for etymological identification of at least some of the hydronyms on the Finno-Ugric basis.

On the other hand, I wouldn’t exclude the possibility that some parts of this large geographical area were never (completely) Finno-Ugric. The population that created the most important part of the hydronymy of the Russian North could be finally pushed aside or assimilated only at the end of the 1st – beginning of the 2nd millennium AD, during the Russian colonization, retaining the memory of the White-Eyed Chude in its own memory.

NOTE. For more on this non-IE substrate in (especially West) Uralic, see e.g. Zhivlov (2015),

The same non-Uralic substrate is most likely behind most of the shared traits by Mordvinic and Balto-Finnic (see below).

3) In genetics, I don’t think the picture could get any clearer. I don’t know what “Steppe ancestry = Indo-European” proponents expected from 2019, if they expected anything at all (I haven’t seen any coherent model, proposal, or prediction for a long time now), but I doubt the recent results are compatible with any of their implied expectations.

corded-ware-pca-sub-neolithic-europe
Detail of the PCA of the Corded Ware expansion. See full PCA and more related files.

Notice, from the PCA above, how this Baltic Late Neolithic group shows actually a shift from Sredni Stog (see PCA with Sredni Stog) towards typical Khvalynsk-Urals-related ancestry, i.e. populations from eastern European forested regions, derived from hunter-gatherer pottery groups, as I have proposed for a very long time, since the first time a Baltic LN “outlier” appeared. It’s amazing how some amateurs can find 0.1% of any Siberian outlier’s ancestry among Uralians 4,000 years later, but fail to see the direct connection here. The esoteric uses of qpAdm, I guess…

Especially noticeable is the extra WHG-like ancestry and corresponding shift, seen especially marked in late Polish CWC samples, but also in Baltic CWC and especially in one Sweden Battle Axe sample, all of them shifting apparently closer to Pitted Ware and SHG. While that may have been interpreted as an in situ admixture in Scandinavia before, the late Polish CWC samples show likely a resurgence of local populations, so we can assume that both shifts (to SHG- and EHG-like populations) of available CWC samples around the Baltic are clearly part of the WHG:EHG continuum that will be found in the eastern European sub-Neolithic cultures, from Narva to Volosovo.

This WHG-related ancestry is clearly predominant in groups with which Battle Axe peoples admixed, based on the shift towards Pitted Ware, which – I can only guess based on modern Volga Finns – is different from the shift we will see in Netted Ware, more towards the Khvalynsk-Urals cluster. This is in line with the expansion of Battle Axe eastward through coastal areas (West to East Baltic and Finland into Sweden), while Fatyanovo peoples probably emerged from a slightly different route, but also a northern one, if one is to follow archaological similarities and their chronology.

bronze-age-europe-baltic
Detail of the PCA of European Bronze Age populations. See full PCA and more related files.

During the Iron Age, the only peoples that probably shifted strongly (based on modern populations) are West Baltic ones, getting closer to the available Late Trzciniec samples, and even closer to the Trzciniec outlier, i.e. away from the earlier Eastern Corded Ware cluster, and towards Central European groups like Czech EBA or Poland EBA, both of them clearly derived from Bell Beakers, but also admixed with (and thus shifted toward) CW-like populations.

If one looks carefully at the previous PCA on Bronze Age populations, and the next one on Iron Age clusters, it is evident that adding the Swedish LN outlier to East Baltic BA (both strongly related to Battle Axe populations) essentially gives us the continuity of East Baltic BA into the Iron Age. This cluster is continued also in two outliers from Sigtuna, a Viking town close to the Gulf of Finland, known to be an important trading site, 1,500 years later. Not much of a change around the Gulf of Finland, then:

iron-age-eastern-europe
Detail of the PCA of East and North European Iron Age populations. See full PCA and more related files.

Based on the two simplistic Uralic clines one might see described (among the many that certainly existed, from Corded Ware to different Eurasian populations), and just like BOO was for some months fashionable as “Samic”, some may be tempted to say that certain Sintashta or Srubna outliers close to the Urals mark the True Uralic™ peoples. Because, of course they do. Ghost haplogroup N and stuff. And Corded Ware never ever Uralic. Because Gimbutas, and my IE R1a grandfather.

NOTE. Funny thing here: there might be Corded Ware, Iranian, Slavic, Germanic, etc… outliers or out-groups, and they might form the widest genetic clusters ever seen, but they are all of one language, because archaeology and linguistics; however, one “outlier” (also, put your own definition of “outlier” here, let’s say 1% of whatever, and strontium isotope potentially from 100 km away) ca. 600 BC in the Baltic who (surprise!) happens to show hg. N, and he signals the first incoming True Uralic™ speaker from wherever… It won’t be the first or the last time some people resort to “the complexity of Uralic-speaking peoples” in ancestry, just to look for “hg. N = Uralic” like crazy. You only need common sense to understand that this is not how this works. Amateur genomics can’t get more embarrassing than the current “let’s look for ‘Siberian ancestry’ in every individual of haplogroup N” trend. Or maybe it can, and it will, but I can’t see it yet.

If one were to insist on looking for ‘foreign’ contributions among Iron Age Estonians, though, I think one should also check out first archaeology, and then the PC3 (or, more graphically, a 3D plot), to understand what might be happening with the many Uralic clines derived from Corded Ware, before starting to play around with bioinformatic tools to discover a teeny tiny 1% admixture of the wrong population, and rushing to build far-fetched narratives. Apparently, one of the different clines formed roughly between southern (steppe – forest-steppe) and northern (tundra-taiga) populations in Uralians is also seen in some Iron Age Estonian individuals – especially in some late samples from Ingria…This is not my main interest, so I will leave this here for others to keep wasting their time chasing the white whale of the 0.5% of True Uralic™ ancestry in ancient Baltic samples of hg. N.

pca-3d-estonians-iron-age-boo-samic
Still images of the 3D plot of Eurasian samples. Typical PC1 vs. PC2 visualization to the left, and shift of the view to PC3 on the right image. See full PCA and more related files.

An exclusive Volga-Kama homeland for Disintegrating Uralic?

Since I don’t believe in macro-regions of largely continuous ethnolinguistic communities, as I have often said about Slavic (naively associated with prehistoric tribes of Eastern Europe) or Germanic (absurdly considered to be represented by Battle Axe), it is difficult for me to believe that Battle Axe-derived cultures remained of the same Finno-Samic dialects since the Corded Ware expansion…unless we live in Westeros, where everything happens “for thousands of years”.

I have to admit, then, that the now prevalent identification among Uralicists has become quite attractive:

  • Fatyanovo-Balanovo as Finno-Permic:
    • Fatyanovo/Netted Ware with West Uralic (also called Finno-Mordvinic).
    • Balanovo/Chirkovo-Kazan with Central Uralic (Mari-Permic).
  • Abashevo, into the Andronovo-like Horizon through the Seima-Turbino phenomenon, with East Uralic (also Ugro-Samoyedic).

Exactly like the identification of Yamna Hungary – Bell Beaker transition as the North-West Indo-European homeland, it gives us simplicity and small and late ethnolinguistic communities, away from the traditionally overused big and early language territories.

This late homeland would be supported, among others, by:

  • The presence of Indo-Iranian loanwords in Finno-Permic and Ugric (probably also in Samoyedic, either lost, or – much more likely – underresearched), compatible with the immediate contact between Abashevo – Sintashta-Potapovka-Filatovka and Fatyanovo-Balanovo.
  • The supposed expansion of Netted Ware from Fatyanovo to the north-west, which may be explained as the split and expansion of Balto-Finnic and Samic ca. 1900 BC.
  • A longer-lasting Finno-Permic (West+Central Uralic) community contrasting with the early separation of East Uralic.
  • The compatibility of this late expansion with the late expansion of Pre-Germanic from Denmark with the Dagger Period, and of Balto-Slavic with Trzciniec, which puts all three dialects reaching the Baltic Sea in the EBA.

NOTE. I meant to update the linguistic text to include the most recently favoured phylogenetic tree of Uralic languages after Häkkinen (2007, 2009, 2014), which has very quickly become the new normal among Uralicists, but I don’t think I will have enough time to review the necessary papers for that. I am rushing to publish a printed edition, so the text will wind up being a mixture of “traditional” (meaning, basically, pre-2010s) description of Uralic dialects but using modern divisions; say, “West Uralic” instead of “Finno-Samic”. By the way, I am still amazed that none of my reader-haters (or any online user discussing Uralic migrations, for that matter) have come up with the questions that the new division pose, and it supports my suspicion about the complete lack of interest in linguistics of most (a)DNA fans, except for the occasional use of old and free PDFs Googled to support new narratives invented expressly for some qpAdm results…

textile-ceramics-europe-bronze-age
Textile ceramic styles and influence of Bronze Age cultures divided in clusters.

Problems with this Parpola-Carpelan’s (2012-2018) interpretation include:

  • The differentiation between Fennoscandian Textile Ceramics vs. Netted Ware, which is not warranted in archaeology. The assumption that Netted Ware expanded to the Baltic Sea (as Kallio does, following the traditional view) is thus weak, and it was probably a question of cultural contacts coupled with short-distance population movements/exchange in both directions (from the Baltic to the Volga and vice versa). In fact, the culture division relies on some fairly common and technically simple ornamentation patterns, widespread all over northern Europe, even before the Corded Ware expansion, and it is very difficult to separate certain neighboring Textile Ceramics from Netted Ware groups in southern Finland (i.e. Sarsa-Tomitsa groups).
  • The strict and radical direction described for the Netted Ware by Carpelan, as an eastward and northward expansion, within a very short time frame (ca. 1900-1800 BC), based on few radiocarbon dates, which seems to me like a very risky assumption. We know how this kind of descriptions of direction of culture expansion based on radiocarbon dates has turned out in much more complex “packages”, like the Bell Beaker culture… In fact, the earliest dates for Textile Ware are from the East Baltic, earlier than those of Netted Ware.
  • The assumption that Balto-Finnic traits shared with Mordvinic are a) late and b) meaningful for dialectalization of two closely related dialects, when it is clear that both dialects separated quite early. Phonologically Finnic is more conservative, morphologically less so, and the shared traits include a handful of non-Uralic substrate words which can’t be traced to a single common source, hence they were adopted when both languages had already separated… All in all, Finnic – Mordvinic correspondances are not even close to Italo-Celtic ones, which is clearly fully incompatible with a proposal of a Finnic separation from Mordvinic coinciding with the LBA-IA transition.

Especially problematic for Parpola’s model is the lack of genetic impact in Bronze Age or Iron Age Estonians, not reaching a significant level under any possible statistical threshold – which I am sure was quite disappointing for some of my readers -, but is in line with major archaeological continuity of groups the from region, only disturbed in cultural (and Y-chromosome) terms by the expansion of Akozino warrior-traders all over the Baltic Sea. Any proposed population movement will be very difficult to support in genetics, given the Corded Ware-derived populations that we will see in both regions, and the continued Baltic-Volga contacts since the Corded Ware expansion.

Problems with an interpretation of such a small impact in population genomics includes the similarly weak impacts and haplogroup infiltrations that can be seen among populations basically everywhere in Eurasia, during any given period, and much greater genetic impacts that are supposed to be (or that were certainly) followed by ethnolinguistic continuity.

akozino-malar-axes-fennoscandia
Distribution of the Akozino-Mälar axes according to Sergej V. Kuz’minykh (1996: 8, Abb. 2).

The Battle Axe question

From Kallio (2015), about choosing a tentative homeland for Proto-Uralic:

(…) linguistically uniform Proto-Uralic would have been spoken in the Volga-Oka region until the mid-third millennium BC when the Proto-Uralic-speaking area would have expanded to the Volga-Kama region as well. By the end of the same millennium, this expansion would have led to the earliest dialectal splits within Uralic into Finno-Mordvin, Mari-Permic, and Ugro-Samoyed. The splitting up of these three soon followed during the early second millennium BC when the Uralic-speaking area finally stretched from the Baltic Sea in the west to the Altai mountains in the east. Indeed, no matter where Proto-Uralic was spoken, the branching into the nine well-attested subgroups (viz. Finnic, Saami, Mordvin, Mari, Permic, Hungarian, Mansi, Khanty, and Samoyed) must have taken less than a millennium, because their shared phonological and morphosyntactic isoglosses are rather limited (see Salminen 2002). The traditional view that all this branching would have taken several millennia violates everything linguistic typology teaches us about the rate of language change.

The basic problem of this identification of Fatyanovo-Balanovo as West-Central Uralic and Abashevo as East Uralic is the nature of the Battle Axe culture, including the Bronze Age East Baltic and Gulf of Finland area. Even if it is accepted that Fatyanovo-Balanovo represented all Western groups, Battle Axe must have represented West Uralic-like dialects.

The ethnolinguistic identification of Battle Axe depends ultimately on the nature of contacts of Fatyanovo/Netted Ware with Battle Axe/Textile Ceramics. If both groups were close and interacted profusely, as it seems, it doesn’t seem granted that we will be able to distinguish a close Para-West Uralic dialect of Scandinavia from the actual expanding Balto-Finnic and Samic dialects, if they were actually linked to the Netted Ware expansion. Also from Kallio (2015):

No doubt the most convincing substrate theory has recently been put forward by the Saami Uralicist Ante Aikio (2004), who has not only rehabilitated but also improved the old idea of a non-Uralic substrate in Saami. His study shows that there were still non-Uralic languages spoken in Northern Fennoscandia as recently as the first millennium AD. Most of all, they were not only genetically non-Uralic but also typologically non-Uralic-looking, bearing a closer resemblance to the so-called Palaeo-European substrates (for which see e.g. Schrijver 2001; Vennemann 2003).

In comparison, the case of Finnic is much more difficult. The fact that Proto-Uralic was not spoken in the East Baltic region means that this area must have originally been non-Uralic-speaking, but so far the evidence for a non-Uralic substrate in Finnic has consisted of appellatives and proper names with no etymology (cf. Ariste 1971; Saarikivi 2004a). Contrary to the proposed substrate words in Saami, those in Finnic show no structural non-Uralisms, as if they had indeed been borrowed from some genetically related or at least typologically similar languages, as I suggested above. Also none of them is more recent than the Middle Proto-Finnic stage, which makes them at least two millennia old. All this agrees with archaeological evidence discussed earlier that the Uralicization of the East Baltic region occurred during the Bronze Age (ca. 1900–500 BC).

The discussion of the paper continues with an unsuccessful attempt to find a hypothetical ancient Indo-European substrate that Kallio believes must be associated with the expansion of Corded Ware, in line with the traditional belief. For example, the often mentioned – almost folk etymology-like, unsurprisingly popular among amateurs – ‘Neva’ as derived from IE “young” is logically rejected…Unlike Parpola, Kallio’s view seems to be confident that Netted Ware (as Textile Ware) expanded into the East Baltic, on both sides of the Gulf of Finland, already during the Bronze Age.

As it has become apparent in population genomics, none of them was right, and Textile Ceramics will essentially show – like Netted Ware – a large genetic continuity of Corded Ware peoples in the whole north-eastern European forest zone – despite small regional population movements, obviously -, which necessarily implies that the whole Corded Ware culture – and not only Fatyanovo-Balanovo and Abashevo – were Uralic-speaking territories.

The similarities in terms of culture and Y-DNA bottlenecks between Battle Axe and Fatyanovo-Balanovo also imply that the linguistic differences between these groups were probably not many, and became strongly divided only after their territorial division. Continued contacts between Battle Axe- and Fatyanovo-derived groups can explain the proposed contacts (Finnic with Samic, Finnic with Mordvinic) after their linguistic-but-not-physical separation.

east-european-fatyanovocwc
East European movement directions (arrows) of the representatives of the Central European Corded Ware Culture (according to I.I. Artemenko).

Battle Axe spoke “Para-Balto-Finnic”?

The Balto-Finnic-speaking nature of Battle Axe is thus supported by:

  • The lack of non-Uralic substrates in Balto-Finnic territory (Kallio 2015).
  • The early separation of Samic and Finnic from Mordvinic, and the virtual identity of Proto-West-Uralic and Proto-Uralic, which suggests that Proto-Uralic spread fast (Parpola 2012).
  • The scarce non-Uralic topo-hydronymy in the East Baltic and around the Gulf of Finland (Saarikivi 2004), comparable to that on the Upper Volga region.
  • The strong influence of a Balto-Finnic-like substrate on Pre-Germanic (or, in Kallio’s opinion, the same Scandinavian substrate influencing both Germanic and Balto-Finnic at the same time), and the continued influence of Balto-Finnic on Proto-Baltic and Proto-Slavic.
  • The continued influence of Corded Ware-derived groups in central-east Sweden in Finland and the East Baltic in terms of agricultural innovations appearing in the LBA, compatible with Schrijver’s proposal of intermediate Germanic-shifted Balto-Finnic groups and Balto-Finnic groups influenced by their pronunciation.
  • The intense Palaeo-Germanic and late Balto-Slavic / early Proto-Baltic superstrate on Balto-Finnic, which place all three dialects around the Baltic Sea since the Early Bronze Age.
  • The easy replacement of a hypothetic Para-Balto-Finnic dialect by incoming Proto-Balto-Finnic-speaking peoples (say, with textile ceramics), without much linguistic impact.

In fact, the continuous contacts of the East Baltic with the Volga, and especially the close interaction with Akozino warrior-traders just before the Tarand-grave period, could be the actual origin of the recent (if any) Finnic-Mordvinic connections that need to be traced back to the LBA-IA (maybe here the number ‘ten’), since most of them can be related to a Pit-Comb Ware culture substrate and earlier contacts through the forest zone, which Samic (due to its early split and presence to the north of the Gulf of Finland during the BA) does not share. In fact, some of them can be traced back to Balto-Finnic first

These are the most often mentioned, in order of descending relevance for a shared ancient community:

  • Noun paradigms and the form and function of individual cases.
  • The geminate *mm (foreign to Proto-Uralic before the development of Fennic under Germanic influence) and other non-Uralic consonant clusters.
  • The change of numeral *luka ‘ten’ with (non-Uralic) *kümmen.
  • The presence of loanwords of non-Uralic origin, related to farming and trees, potentially Palaeo-European in nature.

It’s not only a question of quantity. Are these shared Mordvinic – Balto-Finnic traits really more relevant than, say, those between Italo-Celtic, which are supposed to have formed a community for a very short period at the end of the 3rd millennium around the Alps? Are these traits even sufficient to propose a common early Mordvinic-Finnic group within West Uralic, rather than loose Mordvinic – Balto-Finnic contacts, i.e. contacts between East Baltic (Textile Ceramics) and Volga-Kama (Netted Ware)?

Based on the alternative (Kallio’s) view of continued contacts between Textile Ceramics groups, even without knowing anything about linguistics, you can guess that Parpola is spinning very thin when assuming that these changes suggest that Balto-Finnic may have expanded with Akozino warrior-traders, separating thus ca. 800 BC from Mordvinic…

Genetic findings now clearly help dismiss any meaningful population impact in the LBA-IA transition, although any linguist can obviously argue for linguistic change in spite of major genetic continuity. But then we are stuck in the pre-ancient DNA era, so what’s ancient DNA for.

netted-ware-textile-ceramics
Middle Bronze Age cultures of Eastern Europe.

Genetic continuity = language continuity?

In the end, it’s very difficult to say how much language continuity there is around Estonia since the arrival of Corded Ware peoples. Looking at Modern Estonians, they have been clearly influenced by recent contacts with Baltic- and Germanic-speaking peoples clustering to the south-west in the PCA. They seem to have also received contacts from north(-east)ern peoples, likely from Finland, evidenced by their shifts toward the modern Estonian cluster during and after the Middle Ages, with a slight increase in Siberian ancestry and N1c subclades associated with Lovozero Ware. How much language change did these contacts bring? Maybe an expansion of Gulf of Finland Finnic (Northern Estonian) over Inland Finnic (Southern Estonian) and Gulf of Riga Finnic (Livonian)? Difficult to know, exactly, but, in the traditional view of Balto-Finnic dialectal distribution among Uralicists like Kallio, possibly no change at all.

So, if the obvious changes in the Estonia_MA cluster relative to Estonia_IA cluster and Estonia_Modern relative to Estonia_MA do not represent radical language change…Why would Estonia_IA represent a change relative to Estonia_BA, when it is statistically basically the same? Or Estonia_BA relative to CWC_Baltic? Because of the infiltration of haplogroup N1c around the whole Baltic? Because of the occasional 1% “Siberian” ancestry in some non-locals of varied haplogroups across the whole Baltic area?

In spite of all this, the amount of special pleading we are seeing among openly Nordicist amateurs when discussing the Uralic homeland relative to the Indo-European question in genetics has become a matter of plain willful ignorance. Like the living corpses of the Anatolian homeland, the Armenian homeland, the OIT proponents, or the nativist Basque R1b association, the personal involvement in the revival of “R1a=Indo-European” and “N=Uralic” trends is just painful to watch.

[Next post in this line, if I manage to make time for it: “Genetic (dis)continuity in Central Europe“. Let’s see if early Balts and early Slavs, as well as Germanic peoples, show a cluster closer to Danubian EBA (viz. Maros), Hungary-Balkans BA, and Urnfield-related samples than their predecessors in their areas, i.e. away from East Corded Ware groups… If you want, you can enjoy for the moment the new PCAs I could get done and the tentative map of languages in the Early Bronze Age, that will probably give you the right idea about early Indo-European and Uralic population movements]

bronze-age-early-indo-european
European Early Bronze Age: tentative language map based on linguistics, archaeology, and genetics. See full map.

Related