The Genomic Formation of South and Central Asia – Some thoughts, Part 2

Continuing with some thoughts about Narisimhan et al 2018 (preprint) from the previous post…

The steppe people going south

Now turning our attention to how the steppe people interacted, influenced or migrated to SC Asia, let’s first look at the Andronovo-BMAC interactions and then we’ll move further south onto the Indus Valley. I will follow the freely available (here) paper from Frachetti and Rouse 2012, Central Asia, the Steppe, and the Near East, 2500–1500 BC.

By the time of the formation of the Sintashta and Andronovo cultures, the situation further south in Asia was that of an already productive network of cultural and material exchange across a vast territory:

In the mid- to late 3rd millennium BC, centuries before the start of the BMAC, the antecedent framework for wide-scale connections between Mesopotamia, Elamite Iran, the Persian Gulf, Oman, Central Asia, and the Indus Valley was taking shape. A pioneer trade network, which Possehl (2007) has called the “Middle Asian Interaction Sphere” [MAIS] (see also Ch. II.40; Tosi and Lamberg-Karlovsky 2003) appears to have fostered the transmission of select innovations and ideologies far across Asia before the more substantial political economic formations of the 2nd millennium BC.¹

In addition to this, there was another, less known until very recently, trade network connecting BMAC with China, mediated by some nomadic pastoralist groups from the Inner Asia Mountain Corridor, found at sites like Tasbas or Begash². This Central Asian population was unrelated to the steppe people (Yamnaya, Afanasievo or later Andronovo), as evidenced by the ancient DNA sample I3447 (Dali_EBA) featured in this new preprint.

The arrival of the Sintashta/Andronovo people to the area, with their mobile pastoral economy and fast transport, found an excellent match in the urban society of the BMAC, always in need of importing raw materials and exporting manufactured products.

The links between these areas cross regions that have historically been utilized by mobile pastoralists (cf. Ratnagar 2004). If these routes were similarly exploited in the Bronze Age, mobile pastoralists would have been ideally placed to broker BMAC trade through a down-the-line exchange network (Vinogradova 1993; Christian 1998). By acting as middlemen, mobile pastoralists could have diversified their income sources without having to sacrifice a pastoral lifestyle or significantly alter their productive strategies, and, importantly, could thus establish themselves as peers rather than subjects of the BMAC communities.¹

And archaeology provides good evidence that the mobile pastoralists from the steppe were in close contacts with the BMAC:

Hiebert’s excavation at a discrete scatter of predominantly ICW ceramics, located c.1 kilometer southwest of the fortified BMAC site of Gonur South, recovered pottery belonging to both ICW and BMAC traditions and dated stylistically to the Late Bronze Age (1800–1500 BC). Hiebert considered these forms consistent with the preparation, storage, and consumption of liquids (Hiebert and Moore 2004). On the basis of these observations and the proximity of Gonur-N to Gonur South, Hiebert concluded that Gonur-N represented a short-term, mobile pastoral encampment where members of independent mobile pastoralist and BMAC communities feasted together as part of negotiations over land use (Hiebert and Moore 2004).¹

This mutually beneficial relationship was the base of the interactions between the steppe nomads and the agricultural societies from the south. No evidence of war, conquer or animosity is found, contrary to some unfounded speculations (not in academic circles, fortunately).

So what’s the cultural and genetic impact that these interactions had in the steppe to BMAC direction? In this early period (2100-1500 BCE), not much as far as we know. Apart from the exchange of some goods (Horses? Spoked wheels?), there’s not much evidence of cultural shift towards the steppe in the agricultural societies from SC Asia. From a genetic point of view we don’t see much influence either in the samples available. Dzharkutan1_BA, Bustan_BA, Sumbar_LBA or Parkhai_LBA  (from around the mid 2nd mill.) don’t show any significant steppe admixture nor any Y-DNA R1a, which was almost fixed in the steppe groups. A sample that does show some small amount of steppe admixture is labelled as an outlier (I6667, Parkhai_LBA_o, 1497-1413 calBCE):

Bustan_BA (4 samples)
Tepe_Hissar_ChL    47.9%
Shahr_I_Sokhta_BA2    17.7%
Gonur1_BA    17.1%
Hajji_Firuz_ChL    12.5%
Dali_EBA    4.8%
Sintashta_MLBA    0%

Distance 1.6209%


Dzharkutan1_BA (7 samples)
Tepe_Hissar_ChL    53.2%
Gonur1_BA    31%
Shahr_I_Sokhta_BA2    8.3%
Dali_EBA    6.9%
Hajji_Firuz_ChL    0.6%
Sintashta_MLBA    0%

Distance 1.0332%


Parkhai_LBA_o (1 sample)
Tepe_Hissar_ChL 76.1%
Dali_EBA 17.3%
Sintashta_MLBA 6.6%
Gonur1_BA 0%
Hajji_Firuz_ChL 0%
Shahr_I_Sokhta_BA2 0%

Distance 2.8554%

(Other outliers like the two Dzharkutan2_BA samples have steppe admixture, but probably acquired around the Caucasus, where they seem to come from).

However, we do know that there were steppe people moving and settling south:

As far south as Tajikistan, settlements such as Kangurt-tut and burial sites such as Zardcha Khalifa and Dashti Kozy exhibit ceramics and metallurgy with obvious parallels to late Bronze Age sites in the steppe zone (Bobomulloev 1998). These sites, roughly dated to the late 2nd millennium BC, suggest that the extension of steppe networks was intertwined with exchange vectors of southern Central Asia.¹

And we have samples from them:

Dashti_Kozy_BA (3 samples, c. 1500 BCE, all 3 females)
Sintashta_MLBA 83.6%
Dali_EBA 6.7%
Tepe_Hissar_ChL 6.1%
Shahr_I_Sokhta_BA2 3.6%
Gonur1_BA 0%
Hajji_Firuz_ChL 0%

Distance 1.4058%


Kashkarchi_BA (2 samples, c. 1200-1000 BCE, both R1a-Z93 males)
Sintashta_MLBA 89.7%
Dali_EBA 7.2%
Shahr_I_Sokhta_BA2 3.1%
Gonur1_BA 0%
Hajji_Firuz_ChL 0%
Tepe_Hissar_ChL 0%

Distance 1.6114%


The Indus Valley Civilization and the later Swat Valley culture

For the first time we got a sneak-peek into what the Indus Valley people could look like genetically. This comes from 3 samples found in the periphery of the IVC that harbour and unusual amount of Ancient Ancestral South Indian (AASI) admixture, while lacking any Anatolian related one, which makes them probable migrants from the IVC to those other places. The samples labelled as Indus_diaspora are:

  • Gonur2_BA, I2123, Gonur Tepe, Turkmenistan, 2452-2140 calBCE
  • Shahr-i-Sokhta_BA2, S8726, SE Iran, 3200-3000 BCE, mtDNA U2c1, Y-DNA J2a1h
  • Shahr-i-Sokhta_BA3, S8728, SE Iran, 2550-2450 BCE, mtDNA R7, Y-DNA J2a

The first two of them are genetically similar, with around 15% AASI admixture, while the third one has clearly higher AASI admixture at around 40%. This makes it difficult to say which is more representative of the average, so we’ll have to wait for more samples to know about the levels of AASI in the area. But the West Eurasian part is more or less clear, as something that can be modelled as Iran_Neolithic + ANE and lacking Anatolian-related admixture in any significant degree.

Now let’s compare these samples with the later ones from the Swat Valley.

Udegram_IA (15 samples)
Gonur1_BA 32%
Shahr_I_Sokhta_BA3 31.8%
Shahr_I_Sokhta_BA2 15.4%
Sintashta_MLBA 12.7%
Dali_EBA 5.4%
Naxi 2.7%
Hajji_Firuz_ChL 0%
Tepe_Hissar_ChL 0%

Distance 0.9495%

The Swat Valley samples are from 2 periods, the first one being around 1200-800 BCE and the second one around 500 BCE – 1 AD.  The above Udegram_IA ones belong to the first period and are more or less in the average of the rest. Variation of steppe admixture between individuals is high, ranging between 0 and 20%. You can check a similar model for them here.

From the 41 samples from the first period, 21 of them yielded Y-DNA. None of them belongs to the steppe_MLBA marker R1a-Z93, and there is one R1b (but could be dubious given other R1b’s assigned in the supplements). From the later samples, 15, 6 of them yielded Y-DNA, and one of them does belong to R1a. On the mtDNA side there are more matches with steppe populations.

The burial site at Udegram from where the above samples come from is described is described in this paper³ by Massimo Vidale and Roberto Micheli. The majority of the burials had a primary and a secondary interment.

This funerary pattern was also common in the cemeteries of Katelai I, Loebanr I and Butkara II. Thanks to the anthropological analysis by Maria Letizia Pulcini, we know that at Udegram, in the majority of cases, the primary burial was a mature adult female, resting directly on the floor of the grave, followed by a secondary burial of a male, of the same age or slightly younger, but not young enough to be considered her son. In some cases, the coxal and long bones of the secondary burial had indisputable cut marks from defleshing, left by metal blades that included cleavers.

All of this suggests a female-centred pattern: women were buried in megalithic graves, thereby alluding to their leading role in the household, while the remains of a male relative (?) played an important but secondary role as an ‘accessory’, joining the female occupant after a divergent and more transformative funerary process that may have involved exposure, defleshing and sometimes bone-chopping.

So where does this leaves us in terms of the steppe migration theory? The archaeology and the genetics tell us that this population from the Swat Valley didn’t descend from the steppe nomads. That part is clear. They do have some steppe admixture, but it’s not much (~12%?) and mostly (or totally) from the female side. Was it due to political marriages from the steppe people involved in trade in order to secure their business (marrying their daughters to locals looks like a safest way to avoid conflicts, robberies, etc… in the risky travel across the mountains passes)? Or was it something else?

Whatever the case, the problem is that the Swat Valley is a small and isolated area, not that representative of the who north of India and Pakistan. But these samples are what we have for now. How much can we extrapolate from them and modern ones?

We can look at Figure S4.2, where the Swat Valley samples (SPGT) are plotted together with modern SC Asian populations:

What this figure shows is that the ancient samples have a significantly lower amount of steppe ancestry than expected given their amount of West Eurasian ancestry (when compared to all modern populations). Does this mean that they are an extinct population? No, not really. Some relic populations from the area (like the Kalash) do show continuity with these ancient samples, though their steppe admixture also seems to have increased (about doubled) in the last 2000 years. How? Well, we don’t really know, though it seems that all the movements from historical steppe people have been accumulating steppe ancestry in SC Asia ever since those first movements from the middle Bronze Age. Or it could be something else, I guess.


The case of R1a-Z93

From a genetic point of view, the presence of early Y-chromosome haplogroup R1a-M417 in the steppe, and then it’s daughter (or grand daughter) clade Z93 has been the single strongest argument in favour of a migration from the steppe to India. India has a high incidence of this haplogroup, just like Eastern Europe, and they both are sister clades of the Z645 parent node (Z283 in Eastern Europe and Z93 in SC Asia). Sintashta and Andronovo (and Srubnaya, Potapovka,…) are almost unanimously R1a-Z93 in their male lineages, and many modern SC Asians belong to this clade.

So a movement of people from the steppe to India is undeniable based on this fact, right? (Unless R1a-Z93 was found in SC Asia before 2100 BCE, but that was unlikely and haven’t happened so far). Well, yes, but what about the details?

The first important detail was already mentioned above: The Swat Valley samples from the Iron Age don’t have a single instance of R1a-Z93, and only one sample has it in the historical period. The paper mentions the R1a connection, but remains silent about the lack of R1a in the ancient samples.

The second important detail is that all the steppe samples belong to a subclade of Z93, that is, Z2124. We have lots of samples across space and time in the steppe, but not a single one with the sister clade L657 has been found so far. The latter is the most common clade in South Asia today. What this means is that L657 did probably arrive together with Z2124, but it was a very rare clade among the steppe populations. Its high incidence today in South Asia looks much more like local expansion of a rare but lucky lineage than a marker for any large migration, making the overall incidence of R1a in South Asia hardly relevant for historical purposes.

The marker that should be more directly linked to steppe migrations is clearly Z2124, which is not too high in South Asia (but it is high in Central Asia). Even then, the details of when and how it impacted those populations is unclear. Was it mostly during the Bronze Age or mostly during historical times? (And then there’s the founder effects in many populations, with Kyrgyz being a prime example).

You can check models with modern SC Asian populations here. The first sheet uses an AASI ghost, while the second one is the same but using Mala instead, for those who prefer avoiding a ghost (just keep in mind that Mala appears in the first sheet as 9.2% Sintashta_MLBA, so that should be accounted for on the second sheet somehow).


The Rigveda

In the absence of any convincing archaeology, the main source for providing support to the Aryan Migration Theory (BTW, I don’t care much about semantics, but the often (mis)used work “invasion” instead of “migration” is really out of place. An invasion is what Germany during the 3rd Reich did with Poland or France: a military offensive in order to control a territory. This was not a case of Germans migrating to Poland and France to settle there. See the difference?) has been the early Vedic texts, chiefly the Rigveda. This is a book of mystical chants written in an intentionally obscure and symbolic language, for those who “understand”. It’s use as a source of historical information is limited, and the interpretations of it for historical purposes have mostly been an abuse more than anything else.

As an example, the detailed horse sacrifices have been interpreted as a prove of the steppe migrations. Leaving aside other archaeological matters (like equid burials in Mesopotamia, etc..), one question raised from within those same detailed descriptions has been the fact that the sacrificed horses, to be true horses, should have 17 pairs of ribs. From modern horses, only the Arabian Horse has (most often) 17 pairs of ribs (the rest having most often 18). So how many pairs of ribs did the Sintashta horses have? If they had 18, then the Sintashta provenance would be disproved, right? Well, here’s a counter argument I’ve heard (can’t remember where exactly right now): those 17 pairs of ribs are a symbolic reference to some constellation and not literally refer to the horse anatomy. Fair enough, but if one is going to interpret such a precise detail as a symbolic and non-realistic reference to some astronomical phenomenon, shouldn’t the same principles apply to the rest of the details?

My take on this is that we should leave the Rigveda alone, except for the linguistic analysis and very few other details that are unambiguous. Otherwise anyone can abuse the interpretation of the text to fit their own hypothesis leading to absurd and contradictory conclusions. Let’s stick to more objective data.



So all in all, what’s the verdict from all this data? The jury is still out. I think that the most clear answers will come when we get ancient DNA from those regarded as early Vedic Cultures (Cemetery H, Ochre Coloured Pottery culture,…) around the Punjab and Haryana c. 1900-1400 BCE. If those turn out to be very Sintashta-like and mostly R1a-Z93, the Aryan migration from the steppe theory will get strong support (or proved, as some will prefer to say). I will leave up to each one to decide how likely that option seems when looking at the current data.



1 – Frachetti and Rouse 2012, Central Asia, the Steppe, and the near East, 2500–1500 BC, in A Companion to the Archaeology of the Ancient Near East

2 – Spengler et al 2014, Early agriculture and crop transmission among Bronze Age mobile pastoralists of Central Eurasia

3 – Vidale, M. and Micheli, R, 2017, Protohistoric graveyards of the Swat Valley, Pakistan: new light on funerary practices and absolute chronology

28 thoughts on “The Genomic Formation of South and Central Asia – Some thoughts, Part 2

  1. I would like to make a new case for Indo-Iranian. After the dog sacrifice comment on Eurogenes and a few of Rob’s controversial comments that the Steppe might not even be IE in the first place let alone LPIE, I’ve decided to go back to read some of Anthony’s and Kuzmina’s works. Looking at linguistics this time, which I did not care much about, It now baffles me how the Steppe homeland is mainstream according to linguistics.

    “All models cited above acknowledge that the Proto-Indo-Europeans possessed
    an economy based on domesticated livestock and domestic cereals.”
    “As Anthony remarks in this symposium, there is really no serious evidence for arable agriculture (domestic cereals) east of the Dnieper until after c 2000 BCE (see also Ryabogina & Ivanov 2011; Mallory, in press:a). This means that there is also no evidence for domestic cereals in the Asiatic steppe until the Late Bronze Age (Andronovo etc). ”

    This kinda rules out almost every steppe culture.

    Soma is derived from sewh (to push), ushtra(camel) from PIE (uks) or in case of Mallory *usr- (bull), and Indra is similar to the Hittite Innara. I am not sure why Anthony turned them into non-IE loanwords. Let’s go to Kuz’mina shall we?

    “The earliest layer of Indo-Iranian borrowings is defined by common Indo- Iranian, Indo-Aryan and proto-Iranian terms relating to three spheres of culture: productive economy, social relations and religious beliefs. Words pertaining to the productive economy include domestic animal names (‘sheep’, ‘ram’, ‘Bactrian camel’, ‘stallion’, ‘colt’, ‘piglet’, ‘calf’); terms connected with stockraising products or processing (‘udder’, ‘skin’, ‘wool’, ‘cloth’, ‘spinner’); farming terms (‘grain’, ‘awn’, ‘beer’, ‘sickle’); names of tools (‘awl’, ‘whip’, ‘horn’, ‘hammer’ or ‘mace’), probably the word for ‘ladder’ (or ‘bridge’), and finally, ‘metal (ore)’. A large group of words reflects personal, family and social relations (e.g., ‘man’, ‘sister’, ‘orphan’, ‘name’), important Indo-Iranian social terms, e.g., dāsa (‘non-Aryan, alien, slave’) and asura (‘god, rich, master, hero’).”

    So, loanwords for camels and farming terms into Finno-Ugric? Sounds like the BMAC-Andronovo interactions. If anything, this supports the idea that BMAC was I-I and Andronovo Finno-Ugric.

    “Between about 1800 and 1600 BCE….Eventually the simple incised pottery of the steppes gave way to new ceramic traditions, principally gray polished wares in Margiana and the Kopet Dag, and painted wares in Bactria and eastward into Tajikistan.” – Anthony

    Burnished Grey ware and weaponry is what links Medes, Mitanni Aryans and Indo-Aryans in India. However, Anthony is wrong here. The grey and painted wares have their roots in Iran. Burnished Grey Ware was a chalcolithic development found in Ghabristan (Majidzadeh) and Painted Ware in Sialk III. Western Grey Ware culture of Iran, which evolved from the Gurgan Grey Ware of Hissar, is considered to be the Mitanni Aryan culture. Hissar IIIC which Parpola puts the first arrival of Indo-Aryans in is actually dated to 2170-1900 cal BCE (Erich Schmidt). If anything, that shows a movement from Iran to BMAC.

    “The simple socketed axe-adze from Mohenjodaro closely resembles ones from Hissar (Schmidt 1937:P1.52) and Shahi Tump (Stein
    1931). At Hissar this axe was found in IIIC. A macehead from Mohenjodaro, dated to ca. 2000 B.C. (Piggott 1947 : 3 1) , also is typologically similar to one
    from Hissar IIIC (Schmidt 1937:P1.52). Similar maceheads are known in Luristan, and dated there to cu. 1400 B.C. (Piggott 1947:39-41). ”
    & “A bronze sword with a strong medial midrib, the Fort Munro sword, found in the Punjab and of unknown date, is unique within the entire Indian metallic inventory. The peculiar hilt finds its closest parallel in the “Luristan” words of Iran. La1 (quoted in Joshi 1962 : 18) claims that this sword, like the axe of Mohenjodaro and the socketed axes of Shahi Tump, has no relation to any Indian culture and is clearly intrusive”

    MLBA Luristan was ruled by Kassites who had Indo-Aryan elite like the Mitanni. The link is clear and It is much more plausible to trace the Indo-Aryan culture to Hissar III and Ariana. Indeed there is a lack of Andoronovan artifacts in BA Iran.

    I really think that the Rigveda is very useful if one reads it objectively and can be useful in dating Sanskrit.

  2. Hi, nice summary!

    I would say the (provisional) verdict from Narasimhan et al. (2018) is that Indo-Aryan may have expanded ‘small and late’ (or ‘smaller and later’ than expected), from the potential interesting sites you propose, which is OK with linguistics.

    I have one question, about your comment:

    “Other outliers like the two Dzharkutan2_BA samples have steppe admixture, but probably acquired around the Caucasus, where they seem to come from”

    The paper states:

    “Finally, we examined our Swat Valley time transect from 1200 BCE to 1 CE. While the earliest group of samples (SPGT) is genetically very similar to the Indus_Periphery samples from the sites of Gonur and Shahr-i-Sokhta, they also differ significantly in harboring Steppe_MLBA ancestry (~22%). This provides direct evidence for Steppe_MLBA ancestry being integrated into South Asian groups in the 2nd millennium BCE, and is also consistent with the evidence of southward expansions of Steppe_MLBA groups through Turan at this time via outliers from the main BMAC cluster from 2000-1500 BCE. Later samples from the Swat time transect from the 1st millennium BCE had higher proportions of Steppe and AASI derived ancestry more similar to that found on the Indian Cline, showing that there was an increasing percolation of Steppe derived ancestry into the region and additional admixture with the ASI through time.”

    The outliers from the main BMAC include the Dzharkutan2_BA samples (ca. 2100-1800 BC), and they suggest that BMAC was surrounded by steppe pastoralist ancestry between 2100-1700 BC.

    In fact, we expect the early Indo-Aryan-speaking ancestors of the Mitanni to have traversed rather early in the opposite direction, into the Iranian Plateau… So why would that steppe ancestry (and these two samples) be *from* the Caucasus?

  3. When ancient L657 in India and pre-L657 outside of India are discovered, we are a lot wiser. Things are still very much up in the air.

    I have on my wish list an autosomal analysis and a more detailed yDNA analysis of Bronze Age Xiaohe samples from Tarim Basin which are dated c. 2000-1500 BC. It has been said that they are R1a1(xZ93). Mtdna of these BA Tarim samples is mostly C4. They do not seem to be EBA or MLBA. These samples could shed more light on the dispersal pattern of R1a1 in Asia.

    As Uralic languages have a close relationship with Indo-Iranian languages, it would be desirable to find ancient samples in a culture that acts as a bridge between Indo-Iranian and Uralic languages. Narisimhan et al did not yield any yDNA N, let alone Uralic yDNA N in Central Asia. We have even 32 yDNA samples from Sintashta and most of them are R1a1-Z93 + some R1b and Q. Therefore, I would exclude Kazakhstan and Altai from the potential areas of interaction between Uralic and Indo-Iranian languages.

    Instead, both R1a1 and N1c were found together in Iron Age Sargat culture in West Siberia. It would be very useful to know if those R1a1 samples are Z93, as these samples should be related to the Indo-Iranian influence in the Ugric area.

    On a Finnish site, analyses of presumed ancient Saami and Finnish genomes have been given. According to them:

    CWC_Baltic, 63.4
    UstIdaBA, 13.2
    Comb_Ceramic, 12.8
    ShamankaEN, 10.6

    Anc_Saami, 35.4
    Baltic_IA, 34.6
    Anc_Finn, 23.2
    Srubnaya, 5.8
    Comb_Ceramic, 1

    Genetically, the area of origin of Baltic Finns and Saami seems to be in a broader Fatyanovo horizon in Western Russia. Steppe ancestry in Finns seems to be rather western, i.e. of Srubna type. The yDNA of Srubna is 4xR1a1-Z93, 2xR1a1-Z2123, 1xR1a. I expect that Abashevo DNA will show signs of interaction between Uralics and Indo-Iranians.

  4. “All models cited above acknowledge that the Proto-Indo-Europeans possessed
    an economy based on domesticated livestock and domestic cereals.”

    Please can you give the reference to this groundbreaking ( I was not aware of this) sentence.

  5. @Carlos Quiles

    Re: your question about why those two Dzharkutan2_BA samples and their steppe ancestry be from the Caucasus, the answer is in their autosomes. They cluster with samples from the Caucasus and far away from SC Asian ones. If you look at the models with proximal sources for them in the paper (Table S3.41 in the supplements) you’ll see that all the models are with Armenia_EBA as a source. And I’ve found that including Armenia_EBA (or CHG) is the only way to model them successfully too. So their origin in the Caucasus (or near) seems unavoidable.

    About their steppe admixture, we know that steppe ancestry entered the south Caucasus a few centuries before it did SC Asia. For example, we have the Hajji_Firuz_BA sample (I4243) from 2465-2286 calBCE with a large amount of steppe ancestry, so by 2100-1800 that those Dzharkutan2_BA samples are dated I suppose that steppe admixture was widespread in the south Caucasus (we’ll probably see steppe admixture and R1b-Z2103 in Hurrian samples, unlike in Hittite ones, contrary to what some people are expecting, but I digres…). The steppe ancestry in the Caucasus is from Steppe_EMBA and not from Steppe_MLBA, though, so one way to test if these migrants from the Caucasus brought their steppe admixture with them or if they were “steppe-unadmixed” Caucasus people who migrated east and got their steppe ancestry already in SC Asia would be to test that steppe ancestry:

    Armenia_EBA 80.1%
    Yamnaya_Kalmykia 14.7%
    Naxi 5.2%
    Sintashta_MLBA 0%
    Dali_EBA 0%
    Gonur1_BA 0%
    Levant_BA 0%
    Hajji_Firuz_ChL 0%
    Tepe_Hissar_ChL 0%
    Shahr_I_Sokhta_BA2 0%
    Shahr_I_Sokhta_BA3 0%

    Distance 3.512%

    This does not prove it, but it just seems like the most simple explanation to their steppe admixture to me.

  6. @Vara

    Thanks for that contribution. I still don’t have a strong opinion about the origin of Indo-Iranian, but the way I see the data so far it does make me quite sceptic to attribute it to Sintashta. So all other suggestions are interesting for me.

    Regarding the loanwords mentioned by David Anthony just wanted to mention that he’s following Lubotsky there, so whether right or wrong I wouldn’t blame it on him this time.

    About the Indo-Iranian and Finno-Ugric interactions taking place between BMAC and Andronovo, I’m not too convinced it was at that place and time. I’d go for a later interaction between the steppe (Indo-Iranian after their interactions with BMAC, no matter who was IE speakers first) and the forest steppe people. Probably what Kristiina suggests about the Sargat Culture makes more sense.

  7. @Kristiina

    Yes, I’m waiting too to see where R1a-L657 first pops up. But I’m not really expecting a surprise. I’m inclined to think it was just a minor lineage from Sintashta/Andronovo. But let’s see what ancient DNA says when it comes.

    Also waiting for the Tarim Basin mummies DNA. If it gets confirmed that they are R1a(xZ93) it will be very interesting to see their autosomes. But that only comes from a comment by one of the authors of the original paper and it was never published, so I don’t know how reliable it is.

  8. @Vara

    Forgot to say that regarding the usefulness of the Rigveda, I agree about the linguistic aspects. And a few details can be useful for the geographical area. And probably being really careful one can extract some interesting conclusions, but it’s a dangerous thing. Remember that a careful reading of Vedic texts took an Indian scholar (?) to place the origin of the Aryans in the Arctic.

    So I think it’s better to stick to more tangible data for the most part and not to abuse the interpretations of religious texts.

  9. @Alberto

    I agree that your hypothesis is more plausible, especially since we have Cimmerian names that can be interpreted as a mix of Indo-Aryan and Iranian names, e.g Sandaksatru. I believe a Scythian group is the reason for the Indo-Aryan toponyms in Circassia and the Altai. However, If Andronovo was not Tocharian then it was most likely some Para-Uralic group.

    It’s still an interesting paper even though Altaic is kinda discredited. So much for “consensus” btw.

    I disagree with you on the Rigveda, though. Sure there are a few metaphorical things but It still can tell us of the lifestyle, the tools used, the diet…etc. My whole interest in the PIE debate itself began with Islamic Studies where I then learned about Zoroastrianism and from there I wanted to know where the homeland of Zoroaster was. Anyways, there was this Russian scholar who believed that the Mecca of the Quran was in fact modern day Moscow. To take this to the extreme some pseudo-scientists can look at the pyramids and yell about aliens . That just means you have to look at these text objectively. You simply cannot study Indo-Iranians without the Rigveda the same way you can’t study Islam without the Quran. The problem is nitpicking certain parts of the Rigveda to ascertain certain models that can be disproved by looking at the next section of the book.

    Or we are making the same point?

  10. @Vara

    The problem I see with Sintashta/Andronovo being some form of Uralic is mostly because they were almost universally R1a-Z93, which is an irrelevant haplogroup for Uralics. On the other hand, Scythians were direct descendants of the Andronovo culture (with strong BMAC influence) and mostly R1a-Z93, and they spoke some form of early Indo-Iranian.

    On the other hand, if a word for Camel was borrowed into Uralic at an ancient time, it might mean that the contact zone was further south than the forest steppe, because why would someone borrow a word for an animal they’ve never seen? But my knowledge about the timing of these borrowings is lacking, so maybe someone else can have a better answer for this.

    Re: the Rigveda, I think we’re not that far in agreement. What you say makes sense in theory, but you know how it has worked in practice. I wouldn’t be against solid and reasonable interpretations even of religious texts, but that seems not to work well for one reason or another. Most of the Aryan Invasion Theory (and they mean invasion) is based on the Rigveda, since there is no archaeology to support it. When we get samples from those early Vedic cultures we’ll see how plausible is that interpretation.

  11. We cannot get new insights while dredging the same analysis whether linguistics(e.g. isolated loan words) subsistence technology/social customs. These are coarse traits or highly variable.

    We can look new disparate lines of evidence that lead to statistical convergence. Its a slow process but I think theres a huge amount of stuff that has not been looked at

  12. sanskrit has an “m” ending in the dative case that matches old germanic dative at least in the masculine. In English its there as him, whom, them etc..
    theres also the genitive ending “‘s” in German and English.
    both are not there modern North Indian languages.

    How do other IE languages compare?

  13. @ Alberto

    Great analysis, not much else I can think of. I guess the major issue with the paper is the disjunction between the conclusions and data (e.g. the lack of an R1a link; the non-correlation between Iron Age Swat culture and Andronovo – meagre autosomal impact, no cultural impacts). But when we have to envisage ‘magic bullets’ i guess it’s not surprising.
    Of course more data might clarify things.

  14. Has anyone read the archaeological supplement to the Daamgard et al paper ?

    The last concluding paragraph is worth quoting in full,

    “This survey of the archaeological and biological record of southern Central Asia yields four important findings. First, contacts between the sedentary food-producing populations of the Namazga culture populations residing in Kopet Dagh piedmont and Geokyur oasis of southern Turkmenistan who likely established the outpost at Sarazm had little to no contact with populations residing in the southern steppe zone. Second, contacts between Bronze Age steppe populations and NMG V and BMAC populations appears to have been one in which the dynamic of cultural influence was stronger on the side of the well-established sedentary food- producing populations, and this resulted in the partial assimilation of these initial newcomers to the region both culturally and, to a lesser degree, biologically as well. Third, not all of those who emigrated from the north turned to farming but may have continued a semi-nomadic existence in the highlands, which were unsuitable for the kind of intensive farming practiced in the BMAC homelands or in the regions of Khorezm. Fourth, if there was any Central Asian influence on South Asian populations, that influence likely long predated any development of Iranian, let alone Indo-Aryan, languages, and most likely occurred during the late NMG IV to early NMG V period (ca. 2800–2300 BCE) and even earlier during the Eneolithic from Kelteminar culture groups (4000–3500 BCE).

  15. @Jaydeep

    Thanks, I had not read that yet and it was really an interesting one. It seems that this team did a much better job with archaeology and linguistics, but their genetic part is much centred in Yamnaya-Botai (even Siberia) which is hardly a good companion for those other supplemental materials. And their lack of samples from IVC or related limited their possible analysis.


    Yes, I didn’t want to spend much time with the main text and conclusions, in the hope that it’s just a first and rather unfortunate draft. Their genetic data is much more interesting than their analysis of it (let alone their linguistic and -lack of?- archaeological remarks).

    If these 2 teams had work together they might have produced a much more interesting paper overall, with solid genetic, archaeological and linguistic data.

  16. Wow, Alberto, just found this blog, I am always too slow with these things lol.

    Really excited to have this place to thrash out our ideas. Thanks

  17. What do you mean with this word ‘camel’ borrowed into Uralic? Do you mean the word ‘ushtra’?

    If you mean that PIE *uks, or, in case of Mallory, *usr- (bull), are borrowed into Uralic, it does not refer to camel. For example, the Volgaic words: skal (Erzya), uškál (Mari) and (i)skal (Udmurt) ‘cow’ are rather related to Tocharian word okso ‘cow’, ‘bowid’ than to ushtra ‘camel’. The comparison between OHG ūro `Aurochs’ and Old Indian úṣṭra ‘buffalo’, ‘camel’ (áuṣṭraka — , auṣṭrika — ; uṣṭrapāla — ; *jarōṣṭra –) and Mordovian vašo and North Saami vāǯâ -ččâm- ‘full-grown female reindeer is dubious and not related to ‘camel’ in Uralic.

  18. @Kristiina

    Thanks for clarifying that. It looked a bit strange to me that PU or PFU would borrow the word for camel, but I don’t know enough to argue such thing against what Kuzmina or whoever might have stated in a published book.

  19. @ Alberto

    The Frachetti article you linked importantly describes the character of steppe-pastoralist viz. BMAC interaction. “Gonur – N represented a short – term, mobile pastoral encampment where members of independent mobile pastoralist and BMAC communities feasted together as part of negotiations over land use (Hiebert and Moore 2004 ). In this scenario of mobile – sedentary interaction, contact between the two groups was limited to marginal areas, and though interactions may have been formalized (through feasting), they were not necessarily regular or seen as especially essential to the survival of either group”.

    So it was occasional and marginal interaction. Indeed, we see some very limited genetic exchange, which is understandable.

    After the collapse of BMAC, we are supposed to see a significant intrusion of steppic groups, but this doesn’t seem to be the case neither in the post-BMAC area nor in the Indus.

    I therefore think that the more likely scenario is that the collapse of BMAC and IVC networks led to the differentiation of Indo-Iranian networks used for hundreds (?thousand ) of years earlier along that interaction zone.
    This is of course consistent with linguistics, because all the secondary products reconstructed in paleolexicon* were present throughout Eurasia progressively from the 6th mill BCE (and indeed last of all on the steppe).

    *if we are to take the methodology of “paleolexicology” as legitimate science

  20. @Kristiina

    That was from Kuzmina’s Origins of the Indo-Iranians. Should’ve double checked myself but at least Uralic Andronovan theory is laid to rest.


    I was suggesting that the language of Pre-BMAC influenced Andronovo was Uralic but it seems unlikely now.

    Last one on the Rigveda. I see what you are saying, the scholars are treating the Vedas the same way an Imam would the Quran, with bias . The Aryan Invasion theory is another example of the conformation bias. The Dasyus and Dasas or enemies of Indra either have Iranian names (Toporov), though I’m not sure if Toporov’s Iranian names are actually Iranic. E.g, the destruction of the Vrcivans by Indra near Hariyupiya (Hariab) which according to Toporov is an Iranian hydronym. Even the Gathas supports infighting between different Aryan groups led by a Kavi. You are correct, If one looks at BMAC and IVC their decline was because of the severe climate rather than a conquest. However, If you look at the sites of the 3rd millennium of the Iranian plateau there seems to be signs of battles and conquests, Hissar IIIB was burnt and Shahr E Sukhteh burned and abandoned.

  21. I highly recommend those interested in the BMAC-Steppe interaction pick up a copy of Shahnameh and read through that.

    Shahnameh is the traditional epic of Iran, which in the book is depicted as South East from the Amu Darya (Oxus) and extending to Mazandaran. In the South East is Hind, inc Kabul, which they seem closely associated with.

    In the main part of the book, the main dynasty is that of the Kayanids, based in Iran, and they are often helped out by the main hero in the story, Rustam, whose kingdom is Zabulistan, which seems distinct, but related to Iran.

    So this geography has the BMAC region as Iran, with Kabul and Zabul both being outside of this ‘state’ and somewhat closer to Hind.

    North of the Oxus are the Turanians, a group sharing much traditions and culture with the Iranians, but who are also their major rivals/enemies. Though at times there is peace and cooperation between the two.

    For instance, when the Iranians and Turanians go to war, it is usually at some fixed place at a fixed time, outside of the major settlements, rather than a siege and destruction of a town.

    There is lots of feasting in the story, I mean stuff like feasting for 12 days non-stop when visiting some local nobleman. Lots and lots of drinking too. Infact, it the was the post by Rob that inspired to write this.

    Anyway, I found it a really interesting book. I do think it relates the society anywhere from 3000-1500 BC, it wouldnt be prudent for me to put a more specific dating on it as there are likely traditions from different early subcultures woven into a singular narrative.

    I think this is main narrative of the BMAC culture. Iran are the majority of BMAC settlements, some of the Northeastern ones maybe Turanian, along with Sintashta and Andronovo. Afrasiyab, the mythical Turanian King, is now the name of a Sogdian settlement North of the Oxus.

    It does frustrate me though how it is completely ignored by Western Academic scholarship. Fair enough you cannot draw conclusions from it, but this is the major historical tradition from South Central Asia, it should be acknowledged somewhere.

    Also, the social/tribal/geographic place names mentioned in different early Iranian texts vary greatly (Avesta, Shahname, Vendida, Bundesahin) , which to me implies a great distance in time between the different texts, no way can they all start post 1500BC.

  22. @Robert

    Yes, I agree. If Indo-Iranian was in SC Asia prior to 2000 BCE (which to me looks more likely than not), the initial split would probably start with the collapse of the IVC and its trade network, c. 1900 BCE. This would be consistent with the Rigveda dating to c. 1500-1700 being a very early form on Indo-Aryan, while the Scythian language having its genesis around the same time but much further north being a very early form of Iranian. Avestan is a few centuries younger, so more clearly Iranian, but still quite archaic and not too far from proto-Indo-Iranian.


    That’s probably an interesting book for the history of Iranians, but being written c. 1000 CE, how accurate can it be about events that happened 3000 years earlier? I don’t know how much oral tradition or other ancient texts might have inspired the author about those events and how much is of his own. But probably someone else knows much more about this and can give a more informed opinion.

  23. @Bronze

    This is not the place to come to insult people. If you want to discuss about Davidski do it in his own blog, not here.

    Everyone has an opinion, but opinions should just be taken as that. In the end the facts are what matter, and they will be what they will be. So no one worry about opinions too much.

  24. @mzp1

    There are 3 versions of Iranian myths from Media, Bactria and Drangiana. All three seem to come back to Drangiana, eg. in all three versions Vishtaspa celebrates the religion around lake Hamun and Saoshyant rises from there as well. Even in the Bactrian versions which are prominent in the Shahnameh most of BMAC wasn’t Iranian. In the Shahnameh Bactria was Turanian and Isfandiyar conquered it for his father who then turned it into his capital. These Isfandiyar myths were rewritten during the late Sassanid period, that’s why he battles Yabqu Khaqan and other Gokturk figures in the Shahristan. You caught something few people catch, the House of Sam is not actually Iranian but Turanian and Turan proper is eastern Zabolistan. In the Bundahishn and even the other namas like Garshaspnama they were the sons of Tur. In the early Sassanid period the the province of Turan was north of Makran and south Kushan.

    According to the Shahnameh the capital of Kay Kavad was Istakhr so the book shouldn’t be taken literally, but it is a very important part of the IE literature and I agree with you it seems that western scholars do not respect it. It is a must read for those who wish to study the Sassanid period. The Pishdadian part seem to reflect the PIE myths with the story centering mostly around Treto and the dragon and very similar to the reconstructed PIE myths, the first Kayanian dynasty Indo-Iranian for the last Kavi shared by both Iranians and Indo-Aryans is Kavi Sushravas/Husrova and the second Kayanian dynasty early Iranian then cuts off into the Achaemenid period thanks to the Sassanids revising Iranian history.

    I do agree with you that post 1500 BCE is too late. Indo-Aryans reached Mesopotamia around 1800 BCE (I believe earlier based on Baghdadu) and are derived from Hissar III 2400-2000BCE. Also, It’s kinda unlikely that people from India, west Iran and Syria speak Indo-Aryan as soon as the split happens.


Comments are closed.