Consequences of Damgaard et al. 2018 (II): The late Khvalynsk migration waves with R1b-L23 lineages


This post should probably read “Consequences of Narasimhan et al. (2018),” too, since there seems to be enough data and materials published by the Copenhagen group in Nature and Science to make a proper interpretation of the data that will appear in their corrected tables.

The finding of late Khvalynsk/early Yamna migrations, identified with early LPIE migrants almost exclusively of R1b-L23 subclades is probably one of the most interesting findings in the recent papers regarding the Indo-European question.

Although there are still few samples to derive fully-fledged theories, they begin to depict a clearer idea of waves that shaped the expansion of Late Proto-Indo-European migrants in Eurasia during the 4th millennium BC, i.e. well before the expansion of North-West Indo-European, Palaeo-Balkan, and Indo-Iranian languages.

Late Khvalynsk expansions and archaic Late PIE

Like Anatolian, Tocharian has been described as having a more archaic nature than the rest of Late PIE. However, Pre-Tocharian belongs to the Late PIE trunk, clearly distinguishable phonetically and morphologically from Anatolian.

It is especially remarkable that – even though it expanded into Asia – it has more in common with North-West Indo-European, hence its classification (together with NWIE) as part of a Northern group, unrelated to Graeco-Aryan.

The linguistic supplement by Kroonen et al. accepts that peoples from the Afanasevo culture (ca. 3000-2500 BC) are the most likely ancestors of Tocharians.

NOTE. For those equating the Tarim Mummies (of R1a-Z93 lineages) with Tocharians, you have this assertion from the linguistic supplement, which I support:

An intermediate stage has been sought in the oldest so-called Tarim Mummies, which date to ca. 1800 BCE (Mallory and Mair 2000; Wáng 1999). However, also the language(s) spoken by the people(s) who buried the Tarim Mummies remain unknown, and any connection between them and the Afanasievo culture on the one hand or the historical speakers of Tocharian on the other has yet to be demonstrated (cf. also Mallory 2015; Peyrot 2017).

New samples of late Khvalynsk origin

These are are the recent samples that could, with more or less certainty, correspond to migration waves from late Khvalynsk (or early Yamna), from oldest to most recent:

  • The Namazga III samples from the Late Eneolithic period (in Turkmenistan), dated ca. 3360-3000 BC (one of haplogroup J), potentially showing the first wave of EHG-related steppe ancestry into South Asia. Not related to Indo-Iranian migrations.

NOTE. A proper evaluation with further samples from Narasimhan et al. (2018) is necessary, though, before we can assert a late Khvalynsk origin of this ancestry.

  • Afanasevo samples, dated ca. 3081-2450 BC, with all samples dated before ca. 2700 BC uniformly of R1b-Z2103 subclades, sharing a common genetic cluster with Yamna, showing together the most likely genomic picture of late Khvalynsk peoples.

NOTE 1. Anthony (2007) put this expansion from Repin ca. 3300-3000 BC, while his most recent review (2015) of his own work put its completion ca. 3000-2800. While the migration into Afanasevo may have lasted some time, the wave of migrants (based on the most recent radiocarbon dates) must be set at least before ca. 3100 BC from Khvalynsk.

NOTE 2. I proposed that we could find R1b-L51 in Afanasevo, presupposing the development of R1b-L51 and R1b-Z2103 lineages with separating clans, and thus with dialectal divisions. While finding this is still possible within Khvalynsk regions, it seems we will have a division of these lineages already ca. 4250-4000 BC, which would require a closer follow-up of the different inner late Khvalynsk groups and their samples. For the moment, we don’t have a clear connection through lineages between North-West Indo-European groups and Tocharian.

Early Copper Age migrations in Asia ca. 3300-2800, according to Anthony (2015).
  • Subsequent and similar migration waves are probably to be suggested from the new sample of Karagash, beyond the Urals (attributed to the Yamna culture, hence maintaining cultural contacts after the migration waves), of R1b-Z2103 subclade, ca. 3018-2887 BC, potentially connected then to the event that caused the expansion of Yamna migrants westward into the Carpathians at the same time. Not related to Indo-Iranian migrations.
  • The isolated Darra-e Kur sample, without cultural adscription, ca. 2655 BC, of R1b-L151 lineage. Not related to Indo-Iranian migrations.
  • The Hajji Firuz samples: I4243 dated ca. 2326 BC, female, with a clear inflow of steppe ancestry; and I2327 (probably to be dated to the late 3rd millennium BC or after that), of R1b-Z2103 lineage. Not related to Indo-Iranian migrations.

NOTE. A new radiocarbon dating of I2327 is expected, to correct the currently available date of 5900-5000 BC. Since it clusters nearer to Chalcolithic samples from the site than I4243 (from the same archaeological site), it is possible that both are part of similar groups receiving admixture around this period, or maybe I2327 is from a later period, coinciding with the Iron Age sample F38 from Iran (Broushaki et al. 2016), with which it closely clusters. Also, the finding of EHG-related ancestry in Maykop samples dated ca. 3700-3000 BC (maybe with R1b-L23 subclades) offers another potential source of migrants for this Iranian group.

NOTE. Samples from Narasimhan et al. (2018) still need to be published in corrected tables, which may change the actual subclades shown here.

These late Khvalynsk / early Yamna migration waves into Asia are quite early compared to the Indo-Iranian migrations, whose ancestors can only be first identified with Volga-Ural groups of Yamna/Poltavka (ca. 3000-2400 BC), with its fully formed language expanding only with MLBA waves ca. 2300-1200 BC, after mixing with incoming Abashevo migrants.

While the authors apparently forget to reference the previous linguistic theories whereby Tocharian is more archaic than the rest of Late PIE dialects, they refer to the ca. 1,000-year gap between Pre-Tocharian and Proto-Indo-Iranian migrations, and thus their obvious difference:

The fact that Tocharian is so different from the Indo-Iranian languages can only be explained by assuming an extensive period of linguistic separation.

Potential linguistic substrates in the Middle East

A few words about relevant substrate language proposals.

Euphratic language

What Gordon Whittaker proposes is a North-West Indo-European-related substratum in Sumerian language and texts ca. 3500 BC, which may explain some non-Sumerian, non-Semitic word forms. It is just one of many theories concerning this substratum.

Diachronic map of Eneolithic migrations ca. 4000-3100 BC

This is a summary of his findings from his latest writing on the subject (a chapter of a book on Indo-European phonetics, from the series Copenhagen Studies in Indo-European):

In Sumerian and Akkadian vocabulary, the cuneiform writing system, and the names of deities and places in Southern Mesopotamia a body of lexical material has been preserved that strongly suggests influence emanating from a superstrate of Indo-European origin. his Indo-European language, which has been given the name Euphratic, is, at present, attested only indirectly through the filters of Sumerian and Akkadian. The attestations consist of words and names recorded from the mid-4th millennium BC (Late Uruk period) onwards in texts and lexical lists. In addition, basic signs that originally had a recognizable pictorial structure in proto-cuneiform preserve (at least from the early 3rd millennium on) a number of phonetic values with no known motivation in Sumerian lexemes related semantically to the items depicted. This suggests that such values are relics from the original logographic values for the items depicted and, thus, that they were inherited from a language intimately associated with the development of writing in Mesopotamia. Since specialists working on proto-cuneiform, most notably Robert K. Englund of the Cuneiform Digital Library Initiative, see little or no evidence for the presence of Sumerian in the corpus of archaic tablets, the proposed Indo-European language provides a potential solution to this problem. It has been argued that this language, Euphratic, had a profound influence on Sumerian, not unlike that exerted by Sumerian and Akkadian on each other, and that the writing system was the primary vehicle of this influence. he phonological sketch drawn up here is an attempt to chart the salient characteristics of this influence, by comparing reconstructed Indo-European lexemes with similarly patterned ones in Sumerian (and, to a lesser extent, in Akkadian).

His original model, based on phonetic values in basic proto-cuneiform signs, is quite imaginative and a very interesting read, if you have the time. His account hosts most of his papers on the subject.

We could speculate about the potential expansion of this substrate language with the commercial contacts between Uruk and Maykop (as I did), now probably more strongly supported because of the EHG found in Maykop samples.

NOTE. We could also put it in relation with the Anatolian language of Mari, but this would require a new reassessment of its North-West Indo-European nature.

Nevertheless, this theory is far from being mainstream, anywhere. At least today.

NOTE. The proposal remains still hypothetic, because of the flaws in the Indo-European parallels – similar to Koch’s proposal of Indo-European in Tartessian inscriptions. A comprehensive critic approach to the theory is found in Sylvie Vanséveren’s A “new” ancient Indo-European language? On assumed linguistic contacts between Sumerian and Indo-European “Euphratic”, in JIES (2008) 36:3&4.

Gutian language

References to Gutian are popping up related to the Hajji Firuz samples of the mid-3rd millennium.

The hypothesis was put forward by Henning (1978) in purely archaeological terms.

This is the relevant excerpt from the book:

(…) Comparativists have asserted that, in spite of its late appearance, Tokharian is a relatively archaic form of Indo-European.3 This claim implies that the speakers of this group separated from their Indo-European brethren at a comparatively early date. They should accordingly have set out on their migrations rather early, and should have appeared within the Babylonian sphere of influence also rather early. Earlier, at any rate, than the Indo-Iranians, who spoke a highly developed (therefore probably later) form of Indo-European. Moreover, as some of the Indo-Iranians after their division into Iranians and Indo-Aryans4 appeared in Mesopotamia about 1500 B.C., we should expect the Proto-Tokharians about 2000 B.C. or even earlier.

If, armed with these assumptions as our working hypothesis, we look through the pages of history, we find one nation – one nation only – that perfectly fulfills all three conditions, which, therefore, entitles us to recognize it as the “Proto-Tokharians”. Tis name was Guti; the intial is also spelled with q (a voiceless back velar or pharyngeal), but the spelling with g is the original one. The closing -i is part of the name, for the Akkadian case-endings are added to it, nom. Gutium etc. Guti (or Gutium, as some scholars prefer) was valid for the nation, considered as an entity, but also for the territory it occupied.

The text goes on to follow the invasion of Babylonia by the Guti, and further eastward expansions supposedly connected with these, to form the attested Tocharians.

The referenced text by Thorkild Jakobsen offers the interesting linguistic data:

Among the Gutian rulers is one Elulumesh, whose name is evidently Akkadian Elulum slightly “Gutianized” by the Gutian case(?) ending -eš.40 This Gutian ruler Elulum is obviously the same man whom we find participating in the scramble for power after the death of Shar-kali-sharrii; his name appears there in Sumerian form without mimation as Elulu.

The Gutian dynasty, from ca. 22nd c. BC appears as follows:


I don’t think we could derive a potential relation to any specific Indo-European branch from this simple suffix repeated in Gutian rulers, though.

The hypothesis of the Tocharian-like nature of the Guti (apart from the obvious error of considering them as the ancestors of Tocharians) remains not contrasted in new works since. It was cited e.g. by Gamkrelidze and Ivanov (1995) to advance their Armenian homeland, and by Mallory and Adams in their Encyclopedia (1997).

It lies therefore in the obscurity of undeveloped archaeological-linguistic hypotheses, and its connection with the attested R1b-Z2103 samples from Iran is not (yet) warranted.


Early Indo-Iranian formed mainly by R1b-Z2103 and R1a-Z93, Corded Ware out of Late PIE-speaking migrations


The awaited, open access paper on Asian migrations is out: The Genomic Formation of South and Central Asia, by Narasimhan et al. bioRxiv (2018).


The genetic formation of Central and South Asian populations has been unclear because of an absence of ancient DNA. To address this gap, we generated genome-wide data from 362 ancient individuals, including the first from eastern Iran, Turan (Uzbekistan, Turkmenistan, and Tajikistan), Bronze Age Kazakhstan, and South Asia. Our data reveal a complex set of genetic sources that ultimately combined to form the ancestry of South Asians today. We document a southward spread of genetic ancestry from the Eurasian Steppe, correlating with the archaeologically known expansion of pastoralist sites from the Steppe to Turan in the Middle Bronze Age (2300-1500 BCE). These Steppe communities mixed genetically with peoples of the Bactria Margiana Archaeological Complex (BMAC) whom they encountered in Turan (primarily descendants of earlier agriculturalists of Iran), but there is no evidence that the main BMAC population contributed genetically to later South Asians. Instead, Steppe communities integrated farther south throughout the 2nd millennium BCE, and we show that they mixed with a more southern population that we document at multiple sites as outlier individuals exhibiting a distinctive mixture of ancestry related to Iranian agriculturalists and South Asian hunter-gathers. We call this group Indus Periphery because they were found at sites in cultural contact with the Indus Valley Civilization (IVC) and along its northern fringe, and also because they were genetically similar to post-IVC groups in the Swat Valley of Pakistan. By co-analyzing ancient DNA and genomic data from diverse present-day South Asians, we show that Indus Periphery-related people are the single most important source of ancestry in South Asia — consistent with the idea that the Indus Periphery individuals are providing us with the first direct look at the ancestry of peoples of the IVC — and we develop a model for the formation of present-day South Asians in terms of the temporally and geographically proximate sources of Indus Periphery-related, Steppe, and local South Asian hunter-gatherer-related ancestry. Our results show how ancestry from the Steppe genetically linked Europe and South Asia in the Bronze Age, and identifies the populations that almost certainly were responsible for spreading Indo-European languages across much of Eurasia.

NOTE. The supplementary material seems to be full of errors right now, because it lists as R1b-M269 (and further subclades) samples that have been previously expressly said were xM269, so we will have to wait to see if there are big surprises here. So, for example, samples from Mal’ta (M269), Iron Gates (M269 and L51), and Latvia Mesolithic (L51), a Deriivka sample from 5230 BC (M269), Armenia_EBA (Z2103)…Also, the sample from Yuzhnyy Oleni Ostrov is R1a-M417 now.

EDIT (1 APR 2018): The main author has confirmed on Twitter that they have used a new Y Chr caller that calls haplogroups given the data provided, and depending on the coverage tried to provide a call to the lowest branch of the tree possible, so there are obviously a lot of mistakes – not just in the subclades of R. A revision of the paper is on its way, and soon more people will be able to work with the actual samples, since they say they are releasing them.

Nevertheless, since it is subclades (and not haplogroups) the apparent source of gross errors, for the moment it seems we can say with a great degree of confidence that:

  • New samples of East Yamna / Poltavka are of haplogroup R1b-L23.
  • Afanasevo is confirmed to be dominated by R1b-M269.
  • Sintashta, as I predicted could happen, shows a mixed R1b-L23/ R1a-Z645 society, compatible with my model of continuity of Proto-Indo-Iranian in the East Yamna admixture with late Corded Ware immigrants.

With lesser confidence in precise subclades, we find that:

  • A sample from Hajji Firuz in Iran ca. 5650 BC, of subclade R1b-Z2103, may confirm Mesolithic R1b-M269 lineages from the Caucasus as the source of CHG ancestry to Khvalynsk/Yamna, and be thus the reason why Reich wrote about a potential PIE homeland south of the Caucasus . (EDIT 11 APR 2018) The sample shows steppe ancestry, therefore the date is most likely incorrect, and a new radiocarbon dating is due. It is still interesting – depending on the precise subclade – for its potential relationship with IE migrations into the area.
  • New samples of East Yamna / Poltavka are of haplogroup R1b-Z2103.
  • Afanasevo migrants are mainly of haplogroup R1b-Z2103.
    • The Darra-e Kur sample, ca. 2655, of haplogroup R1b-L151, without a clear cultural adscription, may be the expected sign of Afanasevo migrants (Pre-Proto-Tocharian speakers) expanding a Northern Indo-European (in contrast with a Southern or Graeco-Aryan) dialect, in a region closely linked with the later desert mummies in the Tarim Basin. Its early presence there would speak in favour of a migration through the Inner Asian Mountain Corridor previous to the one caused by Andronovo migrants.
  • Sintashta shows a mixed R1b-Z2103 / R1a-Z93 society.
    • Later Indo-Iranian migrations are apparently dominated by R1a-Z2123, an early subclade of R1a-Z93, also found in Srubna.
    • R1b is also seen later in BMAC (ca. 1487 BC), although its subclade is not given.
  • There is also a sample of R1a-Z283 subclade in the eastern steppe (ca. 1600 BC). What may be interesting about it is that it could mark one of the subclades not responsible for the expansion of Balto-Slavic (or responsible for it with the expansion of Srubna, for those who support an Indo-Slavonic branch related Sintashta-Potapovka).
  • A sample of R1b-U106 subclade is found in Loebanr_IA ca. 950 BC, which – together with the sample of Darra-e Kur – is compatible with the presence of L51 in Yamna.

NOTE. Errors in haplogroups of previously published samples make every subclade of new samples from the supplementary table questionable, but all new samples (safe for the Darra_i_Kur one) were analysed and probably reported by the Reich Lab, and at least upper subclades in each haplogroup tree seem mostly coherent with what was expected. Also, the contribution of Iranian Farmer related (a population in turn contributing to Hajji Firuz) to Khvalynsk in their sketch of the genetic history may be a sign of the association of R1b-M269 lineages with CHG ancestry, although previous data on precise R1b subclades in the region contradict this. (EDIT 11 APR 2018) The sample of Hajji Firuz is most likely much younger than the published date, hence its younger subclade may be correct. No revision or comment on this matter has been published, though.

Modeling results. (A) Admixture events originating from 7 “Distal” populations leading 538 to the formation of the modern Indian cloud shown geographically. Clines or 2-way mixtures of 539 ancestry are shown in rectangles, and clouds (3-way mixtures) are shown in ellipses.

Also, it seems that the Corded Ware culture appears now irrelevant for Late Proto-Indo-European migrations. Observe:

In the text, a consistent terminology of Yamnaya or Yamnaya-related Steppe pastoralists, discarding the relevance of previous migrations from the North Pontic steppe in spreading Late Indo-European:

Our results also shed light on the question of the origins of the subset of Indo-European languages spoken in India and Europe (45). It is striking that the great majority of Indo-European speakers today living in both Europe and South Asia harbor large fractions of ancestry related to Yamnaya Steppe pastoralists (corresponding genetically to the Steppe_EMBA cluster), suggesting that “Late Proto-Indo-European”—the language ancestral to all modern Indo- European languages—was the language of the Yamnaya (46). While ancient DNA studies have documented westward movements of peoples from the Steppe that plausibly spread this ancestry to Europe (5, 31), there has not been ancient DNA evidence of the chain 488 of transmission to South Asia. Our documentation of a large-scale genetic pressure from Steppe_MLBA groups in the 2nd millennium BCE provides a prime candidate, a finding that is consistent with archaeological evidence of connections between material culture in the Kazakh middle-to-late Bronze Age Steppe and early Vedic culture in India (46).

EDIT (1 APR 2018): I corrected this text and the word ‘official’ in the title, because more than rejecting the role of Corded Ware migrants in expanding Late PIE, they actually seem to keep considering Corded Ware migrants as continuing the western Yamna expansion in the Carpathian Basin, so no big ‘official’ change or retraction in this paper, just subtle movements out of their previous model.

Modeling results.(B) A 540 schematic model of events originating from 7 “Distal” populations leading to the formation of 541 the modern Indian cline, shown chronologically. (C) Admixture proportions as estimated 542 using qpAdm for populations reflected in A and B.

NOTE. If they correct the haplogroups soon, I will update the information in this post. Unless there is a big surprise that merits a new one, of course.

EDIT (1 APR 2018): Multiple minor edits to the original post.

EDIT (2 APR 2018): While I and other simple-minded people were only looking to confirm our previous theories using Y-DNA haplogroups, and are content with wildly speculating over the consequences if some of those strange (probably wrong) ones were true, intelligent people are using their time for something useful, interpreting the results of the investigation as described in the paper, to offer a clearer picture of Indo-Iranian migrations for everyone:

Visit the beautiful interactive map with samples: with their location, PCA, ADMIXTURE and haplogroups (still with those originally given):!/vizhome/TheGenomicFormationofSouthandCentralAsia/Fig_1

Featured image, from the article: “A Tale of Two Subcontinents. The prehistory of South Asia and Europe are parallel in both being impacted by two successive spreads, the first from the Near East after 7000 BCE bringing agriculturalists who mixed with local hunter-gatherers, and the second from the Steppe after 3000 BCE bringing people who spoke Indo-European languages and who mixed with those they encountered during their migratory movement. Mixtures of these mixed populations then produced the rough clines of ancestry present in both South Asia and in Europe today (albeit with more variable proportions of local hunter-gatherer-related ancestry in Europe than in India), which are (imperfectly) correlated to geography. The plot shows in contour lines the time of the expansion of Near Eastern agriculture. Human movements and mixtures, which also plausibly contributed to the spread of languages, are shown with arrows.”