The debate over the location of the so called Indo-Anatolian homeland won't be decided by the persistence of any type of genetic ancestry in ancient Anatolia.
It'll be decided by a multidisciplinary study on the interactions between the ancient peoples of the North Pontic steppe, the eastern Balkans, and western Anatolia.
If such a study finds a pulse of steppe-related gene flow from the Balkans into Anatolia sometime during the early metal ages, it'll corroborate the linguistic hypothesis that a language ancestral to Hittite, Luwian and related tongues moved into Anatolia from Eastern Europe.
Why do we only need a pulse of gene flow, you might ask? Obviously, because:
- language and genetic ancestry can start with a strong association but, since they're not linked, they can eventually follow very different trajectories
- the dilution of genetic ancestry is an important factor, especially in ancient West Asia, and it must be taken into account in models of language spread, rather than ignored in favor of simple, elegant models that do not reflect reality.
Here's my favorite quote from the recent
Lazaridis, Alpaslan-Roodenberg et al. paper, because, probably unbeknownst to the authors, it's exceptionally revealing about the spread of a wide range of Indo-European speakers into Anatolia.
However, in individuals from Gordion, a Central Anatolian city that was under the control of Hittites before becoming the Phrygian capital and then coming under the control of Persian and Hellenistic rulers, the proportion of Eastern hunter-gatherer ancestry is only ~2%, a tiny fraction for a region controlled by at least four different Indo-European–speaking groups.
Indeed, this is exactly what the Lazaridis, Alpaslan-Roodenberg et al. paper should've been about.
That is, the authors should've given us a painstaking account of the spread of different ancient Indo-European speaking groups into Anatolia and explained how, overall, their DNA was rapidly diluted to a trace amount.
However, instead they treated us to a make-believe tale about a so called Indo-Anatolian homeland in what is now Armenia.
See also...
Dear Iosif...Yamnaya
But Iosif, what about the Phrygians?
Dear Iosif...
Dear Iosif #2
Dear Iosif #3
267 comments:
1 – 200 of 267 Newer› Newest»I'd like to see the relevant scientists redeem themselves in the near future. Here's a random idea:
A multidisciplinary study on the interactions between the ancient peoples of the North Pontic steppe, the eastern Balkans, and western Anatolia.
By the way, looking forward to Genos Historia's new video about the Lazaridis, Alpaslan-Roodenberg et al. paper.
I'll link to it in a new blog post.
Hittites before becoming the Phrygian capital and then coming under the control of Persian and Hellenistic rulers, the proportion of Eastern hunter-gatherer ancestry is only ~2%, a tiny fraction for a region controlled by at least four different Indo-European–speaking groups.
Who are they talking about as having 2% EHG? Is it people in anatolia after the time of the persian takeover?
We know for a fact that the people of anatolia spoke languages that came from people with levels of EHG much higher than 2%. The Greeks and Persians both had higher levels of EHG than 2%. So why wouldnt the original speakers of anatolian have higher levels of EHG like the greeks and the persians. Thats the ultimate concern right? We want to know how much steppe the original speakers of anatolian had.
All they found was ~2% EHG in Early Medieval Central Anatolia.
However, as I demonstrated, if they actually bothered to look at their samples in more detail, they would've found that some of these samples had minor but significant Eastern Europe ancestry while others didn't.
https://eurogenes.blogspot.com/2022/09/but-iosif-what-about-phrygians.html
It's these sorts of details that matter, including obviously accurate and relevant time slices with ancient DNA to test how the European ancestry was diluted over time.
Some qpAdm models I've posted on Anthrogenica. Steppe signal is very apparent during the Chalcolithic.
https://media.discordapp.net/attachments/998736754804785264/1016668146565906482/unknown.png
@Davidski
Could Khasi samples from this paper be processed for G25?
https://evolbio.ut.ee/munda/
@Davidski
Anatolian Greeks from Central Anatolia are ~9-10% steppe. If the steppe admixture since the bronze age is diluted, what do you think about their steppe source?
Target: Greek_Central_Anatolia
Distance: 1.4204% / 0.01420412
76.4 TUR_Kaman-Kalehoyuk_MLBA
9.8 Yamnaya_RUS_Samara
9.2 IRN_Hajji_Firuz_C
4.6 RUS_Maykop_Novosvobodnaya
Target: Greek_Cappadocia
Distance: 1.7855% / 0.01785510
69.4 TUR_Kaman-Kalehoyuk_MLBA
21.2 IRN_Hajji_Firuz_C
9.2 Yamnaya_RUS_Samara
0.2 RUS_Maykop_Novosvobodnaya
@Norfern-Ostrobothnian
Not possible.
@baickoz
No idea. Maybe some recent Balkan Greek ancestry?
@kaksoipiste
I do not know your model specifications. Lets run G25, first without Kura Araxes.
Target: TUR_Marmara_Barcin_C:I1584
Distance: 2.3407% / 0.02340744
84.2 TUR_Marmara_Ilipinar_C
9.6 RUS_Progress_En
6.2 TUR_Tell_Kurdu_MC
0.0 BGR_C
0.0 BGR_N
0.0 BGR_Varna_C
0.0 TUR_Marmara_Barcin_N
0.0 TUR_Marmara_Ilipinar_N
0.0 TUR_Tell_Kurdu_EC
Looks like your model. Distance of 2.34%.
Now adding Kura_Kaps, removing progress and limiting to 3 sources to compare to above model with same number of sources.
Target: TUR_Marmara_Barcin_C:I1584
Distance: 2.1330% / 0.02132964 | R3P
53.0 Kura-Araxes_ARM_Kaps
35.6 BGR_C
11.4 TUR_Tell_Kurdu_MC
53% KA_Kaps. Distance is lower at 2.13%.
Even if you just want to compare two-source models.
Target: TUR_Marmara_Barcin_C:I1584
Distance: 2.3575% / 0.02357541 | R3P
90.8 TUR_Marmara_Ilipinar_C
9.2 RUS_Progress_En
vs.
Target: TUR_Marmara_Barcin_C:I1584
Distance: 2.2795% / 0.02279473 | R3P
59.6 Kura-Araxes_ARM_Kaps
40.4 TUR_Marmara_Barcin_N
The 2nd model is (insignificantly) better than the 1st. Also, progress ancestry cannot admix in its pure form without mixing with intermediate populations. What you have caught here, is excess affinity to CHG in Barcin_C which
If there was widespread yamnaya ancestry in west anatolia post 4000bce, why is there noise level progress/yamnaya ancestry in the 3500bce Ilipinar_C sample I10542? This site is 50km from Barcin.
Target: TUR_Marmara_Ilipinar_C:I10542
Distance: 1.9033% / 0.01903324
30.4 TUR_Tell_Kurdu_MC
29.6 Kura-Araxes_ARM_Kaps
27.0 TUR_Marmara_Barcin_N
10.6 TUR_Tell_Kurdu_EC
2.0 Yamnaya_RUS_Samara
0.4 BGR_C
0.0 BGR_N
0.0 BGR_Varna_C
0.0 RUS_Progress_En
0.0 TUR_Marmara_Ilipinar_C
0.0 TUR_Marmara_Ilipinar_N
I have run qpAdm rotating models on Barcin_C before, yamnaya is not needed.
@Davidski
I don't think so. Anatolian Greeks seem to be more ''eastern'' than Bronze Age Anatolians. Meaning they have less Anatolia_N and slightly more CHG/IranN + steppe (according to a quick G25 run). If they had come from the Balkans I think Anatolian Greeks would have less CHG/IranN than their predecessors.
Optimal models need relevant source pops. Eg Barcin N didn’t exist in local form, so no point using it, and it’s hides Balkan related flow. The KA like ancestry is from a different population than KA itself.
A more local form of steppe admixed group rather than Steppe-En etc
Just randomly putting pops into a auto-machine doesn’t really tell us much
@vAsiSTha
>If there was widespread yamnaya ancestry in west anatolia post 4000bce, why is there noise level progress/yamnaya ancestry in the 3500bce Ilipinar_C sample I10542?
Same reason some lower caste Dalits and Brahmins inhabiting in the same region could have differing levels of Steppe ancestry. But y'know, also due to the fact that Ilipinar_C samples are dated between 5000-4000 BCE and the Barcin sample between 3946-3660 BCE.
As for your G25 models, apart from the fact that they are historically inaccurate; it should still be picking up additional (+~4-5%) Eneolithic steppe even if you add Kura Araxes to the model— which it does (even with the addition of BGR_C).
Target: TUR_Marmara_Barcin_C
Distance: 2.2043% / 0.02204338
31.2 Kura-Araxes_ARM_Kaps
31.2 TUR_Marmara_Ilipinar_C
21.6 BGR_C
7.2 TUR_Tell_Kurdu_MC
5.0 RUS_Vonyuchka_En
3.8 AZE_Caucasus_lowlands_LN
Target: TUR_Marmara_Barcin_C
Distance: 2.2108% / 0.02210832
33.0 Kura-Araxes_ARM_Kaps
31.6 TUR_Marmara_Ilipinar_C
21.2 BGR_C
7.2 TUR_Tell_Kurdu_MC
4.2 RUS_Progress_En
2.8 AZE_Caucasus_lowlands_LN
Also I can't run rotating models since the dataset I have on my hand seems to have been misformatted. Would you be kind enough to upload your merged dataset on Google Drive or somewhere else? I'd really appreciate it.
Essentially its a 2-way admixtue of central-eastern Anatolia & Balkan sources
Getting between 20-40% admixture. That's big enough to have mustered an elite conquest of central Anatolia
NB: Kura-Araxes is pretty irrelevant.
These posts by Davidski addressed to Iosif feels like a 5 year old boy writing angry letters to his daddy...
@Johan A
Unfortunately, let's be honest here, the Science paper by Iosif et al. looks like it was written by 5 year olds.
Population density is the obvious hurdle for Anatolia given that it is the mother ship of endless Neolithic out-migration and home to clusters of urban centers. It's city states weren't exactly soft targets.
I think expectations in seeing admixture levels like South Asia are unwarranted. Apples and oranges
South Asia is almost as complex as Anatolia, with a lot of steppe admix in parts of the north, and very little to none further south.
I think the authors expected to see the equivalent of the Corded Ware expansion in Anatolia, failed to find this, and that's how their make-believe tale of the Armenian homeland came about.
But like I said in my other blog post, if they couldn't make sense of the Phrygian migration, even though they had samples from the Phrygian capital, then how could they make sense of the Hittites and Luwians?
I still can figure out where the Tocharians came from, their relationship with Saka, Wusun, Paszeryk, Tagar snd other Europoid cultures, are they descendants of Andronovo or Afanasievo and many other dead ends.
here's the thing. As ive pointed out before, there was a population crash in western Anatolia after ~ 4700 bc. So density itself isnt the issue
''Of all post-Neolithic periods in northwestern Anatolia the late Middle Chalcolithic is the least known. ..the period between 4500 and 4000 BC still appears as a major chronological gap or lengthy occupational hiatus. A similar situation can be seen in the eastern contact zones, especially in central Anatolia; and since the few contemporary settlements in southeast Turkey share distinctive elements with those of the Syro-Mesopotamian cultural sphere, it seems that the vast Anatolian land-mass from Cilicia in the east to Turkish Thrace in the west was temporarily uninhabited (Fig. 1).(!)
All the more unexpected, therefore, is a cross-cultural comparison in the mid-5th ‒ early 4th millennium BC between Anatolia and the Chalcolithic of southeast Europe – as represented by the complexes Karanovo V-Marica and Kodžadermen-Gumelniţa-Karanovo VI, with their huge burial mounds, extremely rich graves, their high-level metallurgy and hierarchically structured societies.''
That is why you cant use Barcin_Neolithic in models, or other distal sources.
Ergo, my 2-way model above.
then rapid demographic recovery of eastern provenance, after 4000 BC, with prot-Urbanism by 3000 BC. Sometime at this point, steppe-Balkan people arrive into an already densely populated region
@All
I've updated all of the G25 datasheets.
https://eurogenes.blogspot.com/2019/07/getting-most-out-of-global25_12.html
Please report any errors here or via email.
@Rob
"But y'know, also due to the fact that Ilipinar_C samples are dated between 5000-4000 BCE and the Barcin sample between 3946-3660 BCE."
Check properly, I have used the sample I10452.
I10542 Adult; 40+ yrs. Direct: IntCal20 5461 83 3630-3377 calBCE (4730±25 BP, PSUAMS-7729) TUR_Marmara_Ilıpınar_ChL
Kaksoipiste was talking to you about Ilipinar C
But this group already has ~20% Balkan ancestry . Then on top of that MBA have extra 20%
@Davidski
I reiterate what I wrote on the other thread, Chile's initials have only "C" instead of "CHL".
These samples should be named as follows:
CZE_Unetice_C:CHL001
CZE_Unetice_C:CHL006.merged
CZE_Unetice_preC:CHL002
CZE_Unetice_preC:CHL003.merged
CZE_Unetice_preC:CHL005.merged
CZE_Unetice_preC:CHL007.merged
CZE_Unetice_preC:CHL008
POL_Chlopice_Vesele_Culture:I6531
POL_Chlopice_Vesele_Culture:I6537
JOR_PPNC:I1708
Levant_Natufian_EpiP:I1072
Levant_Natufian_EpiP:I0861
@ Andrzejewski
'I still can figure out where the Tocharians came from, their relationship with Saka, Wusun, Paszeryk, Tagar snd other Europoid cultures, are they descendants of Andronovo or Afanasievo and many other dead ends.'
I think given the linguistic dating of pre-proto-Tocharian branching away from nuclear IE, Afanasievo makes a lot of sense. It's definitely not Iranic so I don't think any known-to-be-Iranic groups makes sense.
Afanasievo -> Dzungaria_EBA -> Tarim basin at some point, is what my money's on. I think Iranic groups entered the Tarim basin at some point, as well, for sure, but the Tocharian language obviously persisted in some capacity, and Afanasievo just makes sense to be the pre-proto-Tocharian-speaking group in terms of chronology and geography, I think.
RE: Tocharian i think more Iron Age Tarim data is needed. The Afnasievo-Chemurchek scenario is growing on me, but it's still too primordialistic - Long-time gap and doesnt account for the influence of so-called Saka groups in the region.
I recall C'Axe said that Xianshuiquangucheng and Baiyange might be Tocharian (C) sites . If so, they have Andronovo-related ancestry as well as Chemurchek, although I havent checked them personally with qpADM. I had been leaning toward 'eastern Scythian/ western Xiongnu' type scenario because I don;t think Sintashta wwas yet differentiated into Indo-Iranian. You cant have the latter without BMAC-related ancestry, as such the Irani identity of Karasuk etc seems presumptuous
It's faulty if Lazaridis, et. al. were not accounting for these timing:
pnuadha said... "Who are they talking about as having 2% EHG? Is it people in anatolia after the time of the persian takeover?"
Davidski said... "All they found was ~2% EHG in Early Medieval Central Anatolia"
These samples are basically meaningless. The early Indo-European people were extremely mobile. Why should we assume they came in to any one place, replaced many of the locals and then stayed in place? This may have happened in some places, like NW Europe but I don't see why we would expect the crossroads that Anatolia was, to be have a static environment once Indo-Europeans arrived.
So Andronovo giving 15-20% to NW india makes it indo iranian, but the presence of 30-50% andronovo ancestry inside Tarim basin post 1500bce does not?
Afanasievo is too old to be relevant to anything. Its ancestry also dies out for the most part to be replaced by andronovo heavy ancestry. Tocharian, as per Douglas Adams, is closest to Germanic and Andronovo being the eastern variant of corded ware makes it the best contender for Tocharian in my eyes.
The iranian loans are from the bmac related ancestry.
One of the results in recent paper of Lazaridis et al. (The genetic history of the Southern Arc: A bridge between West Asia and Europe) is lack of EHG in Bronze age anatolian samples. After I checked suplement witch samples I found some problems with this assumption.
In early bronze age cremation was rare in Anatolia and was found only in the western part of it. In the second part of 3rd and in 2nd millenium BC cremation has spread considerably in western and central Anatolia, in several Hittie cemeteries it was dominant form of burial (Ilica 126 cremation for 131 graves, Osmankayasi near Hattusa 69 cremation for 91 graves), so many of buried individuals will not provide material for research.
Another problem is geographical distribution of bronze age samples. There is very small amount of samples from territories where indoeuropean anatolian speakers lived, I found that there are only few from central Anatolia (Çamlıbel Tarlası - 10 samples from first half of 4th millenium BC, Kalehöyük - 5 samples from first half of 2rd millenium BC, Ovaören - 3 samples from first half of 3rd millenium BC), there are also some samples from south-western Anatolia, Harmanören - Göndürle Höyük and from Yassitepe. We must remember that in early 2nd millenium BC Hittites still did not occupy territory which was their core in later times, Hattusa was Hattian site until Anitta destroyed it in 18th century BC.
Most of samples from bronze age Turkey are from Turkish-Syrian border (Oylum Höyük, Tilbeşar Höyük, Tell Atchana/Alalakh,Titriş Höyük), Turkish-Syrian-Iraq border (Güçlükonak), northern Turkey (Suluova, İkiztepe where localizedare nonindoeuropean Kaskans), some from eastern shores of Marmara sea (Barcin, Orhangazi), eastern Turkey (Arslantepe) and Turkish-Iran border (Muradiye), in general not from the territory where indoeuropean anatolians lived in bronze age.
The problem of small amount of samples in the territories where indoeuropean anatolian speakers have been detected in bronze age (which is not a surprise, cemeteries are rare in 2nd millenium BC Anatolia and cremation was common burial practice) makes me suspicious about results of the study.
@Zarrr
I agree with your overall assessment, except that TUR_Aegean_Izmir_Yassitepe_MBA:I5737 actually does have East Euro/steppe ancestry and belongs to European Y-haplogroup I2a-P78.
So that's a pretty good strike rate considering the lack of relevant samples in the paper.
@vAsiSTha
“Tocharian, as per Douglas Adams, is closest to Germanic and Andronovo being the eastern variant of corded ware makes it the best contender for Tocharian in my eyes.
The iranian loans are from the bmac related ancestry.”
No, Mallory and Adams theories and dictionaries are outdated.
From recent Indo-Iranic-related talks of #IGAT2022:
“Stefan Norbruis presented some morpho-lexical innovations shared by Tocharian, Indo-Iranic and Balto-Slavic arguing for an eastern location of Tocharian in the (post-Anatolian) IE dialect continuum in the homeland.”
I think that Tocharians had contact with Andronovo, which was Indo-Slavic and this explains similarities of Tocharian with Balto-Slavic, and later with Indo-Iranians who were formed in Central Asia with BMAC influence.
@Rob
"Tocharian i think more Iron Age Tarim data is needed. The Afnasievo-Chemurchek scenario is growing on me, but it's still too primordialistic - Long-time gap and doesnt account for the influence of so-called Saka groups in the region"
If you mean historical Saka, their presence within the Tarim Basin is in my eyes a modern invention. I made a blog entry about it not too long ago: https://musaeumscythia.blogspot.com/2022/07/did-saka-migrate-to-tarim-basin-and.html
If you mean nomads from the Scythian horizon, this influence goes back to the early iron age. Interestingly some of it may have come directly from siberian populations rather than Scythians themselves, as evidenced by the early iron age Liushui sample that has 0% steppe_mlba yet is a mix of DSKC types and Aigyrzhal types.
I'm not sure why this would be in the way of an Afanasievo origin, it has always been known that Tocharians were always on the periphery of steppe cultures, with some of the Han period descriptions nearly describing them as steppe nomads, and Tocharian city states were part of the Xiongnu empire since its inception. Loulan was mentioned as one of the "people of the bow" that were now united as Xiongnu by Modu Chanyu to emperor Wen.
As far as samples go you can check C4127 and C3616, as these seem to produce models that show a steppe_emba > mlba preference. Some samples on G25 will show a EMBA>MLBA presence that isn't strongly there, like one of the Xiongnu samples whose ID I forgot (SKT009 perhaps?). F004 from Shirenzigou might be another candidate but it is quite low quality.
Tocharian, linguistically, requires a split relatively early from other IE languages. This is something every Tocharian expert agrees on. Then you have the more speculative parts, but most linguists think for a while there was strong contact with "Siberian" languages and then by the later bronze age you must have Iranian contacts evidenced by loanwords from Iranian languages that were not Khotanese derived. It is possible these came from nomads but they could also be from Andronovo derived people on the most eastern points of Xinjiang.
Also the Rigveda and Avesta were both composed in the later bronze age and yet were already clearly differentiated from one another, they can't be more than 700 years removed from Sintashta so I don't see how you come to the conclusion that Indo-Iranian wasnt around back then. Maryannu is clearly derived from an Indo-Aryan term and found its way into attestations by 1751 BC.
Alakul groups probably were the vector for Iranian languages as you can find them all exactly where you need them during the bronze age, that is southern central Asia, the Ural region, and the steppes around the Altai.
@vAsiSTha
The entire western half of the Tarim Basin was Iranic speaking, as were the most eastern points where Dzungaria, the basin and western Gansu connect again. Tocharian is only on the northern and southern rim of the eastern half of the Tarim Basin. Furthermore most of the ancient DNA we have is not from the Tarim Basin, nearly all samples were from the Ili region which is separated from the Tarim Basin by the Tian Shan mountains. That is the area where the Iranian-speaking Saka lived according to Han era records, so yes those 50% andronovo samples were Iranian speaking due to their steppe_mlba.
@Davidski
If you want, you may add these samples to the spreadsheet.
https://drive.google.com/file/d/13zfclp2hFvTAECLJVLXTH1bxEj851NuV/view
https://drive.google.com/file/d/1BB9m_FPBS0fG9PYnLfbE4VzXc7xzVigE/view
@ Copper Axe
I get what you're saying- that's the default hypothesis: Afansievo = Tocharian and Andronovo-Indo-Iranian. With Afansievo clearly present as far as Mongolia & Dzhungaria, it's looking reasonable, But I suspect there is more to it though
As with someone presenting their qpAdm results, or an archaeologist reconstructing a theory based on pottery shapes, linguistics needs to be carefully analysed, critiqued and limitations understood. There is an issue with too much reliance on 'treeness'', reconstructions or loans and the belief they dictate what 'has to have been'.
In reality, languages develop over long periods subject to multiple influences, and this could include two streams of IE speakers as well as whatever non-IE influences were present.
For example, Germanic emerged from BAx and northern Beaker IE groups, as well as a couple of pre-IE groups. Similarly, the differentiation you speak of for Indo-Iranian is probably not solely due to bifurcation from a Sintashta Ur-Volk. Indeed, people speak of specific Indo-Slavic links on the one hand and Iranian -Armenian-Greek glosses on the other. And this is clearly born out with aDNA. So I dont think these Andronovo related groups are 'in the way' of an Afansievo link, but instead something which might add to it. I would think Andronovo as versatile and possibly dialectically fluid communities, and clearly groups like Bgazy Dandabai would be distinctive to those which ventured off and became relatively isolated in the Dhzungar passes already by 1500 BC. Add different forces of BMAC vs WSHG vs East Asian admixture, then you have a rather heterogeneous set of communities.
As for models, G25 is a good initial screen but I would defer to qpAdm & additional tests. I thought Altvred on your blog found a preference for steppe MLBA over Afansievo for C3613 ? Id have to check for myself tbh
I cannot do the topic justice here, but that is why the default hypotehsis, whilst not wrong, might be too primordialistic, and miss details. At face value at least, I am intrigued what the dominance of R1b-Ph155 and secondary presence of R1a-Z93 might tell us. We need more Tarim IA-medieval samples..
"@vAsiSTha
The entire western half of the Tarim Basin was Iranic speaking, as were the most eastern points where Dzungaria, the basin and western Gansu connect again. Tocharian is only on the northern and southern rim of the eastern half of the Tarim Basin. Furthermore most of the ancient DNA we have is not from the Tarim Basin, nearly all samples were from the Ili region which is separated from the Tarim Basin by the Tian Shan mountains. That is the area where the Iranian-speaking Saka lived according to Han era records, so yes those 50% andronovo samples were Iranian speaking due to their steppe_mlba."
I get the feeling that you have not seen the Vikas Kumar Xinjiang paper.
Here are all the sites sampled in that paper, and also the broad regions where Toch A,B,C were attested.
Specifically, the sites labeled Baiyanghe (BYH) inside Toch A region, XSQG & XKKD inside Toch B region. ZGLK near Toch C region. None of these sites are in western tarim, and none of these sites are in the Ili region.
The analysis of these samples include what Kumar has labeled Xinj_IA4-IA8_aEA. These have between 30-40% Andronovo ancestry, minor BMAC and rest east asian/shamanka as per their qpAdm. One has afanasievo ancestry.
The sites of Xianshuiquangucheng (XSQG) and Xikakandasayi (XKKD) from 200-500ad are on the border of Toch A/B region, near Agni. One sample has 50% andronovo ancestry as per their qpAdm.
Zhagunluq has falls in toch C region and has 6 samples dated to around turn of CE all of them modeled with 25-40% andronovo by the authors.
So in short there is enough andronovo ancestry smack inside the Tocharian speaking regions.
@Copper Axe “ Then you have the more speculative parts, but most linguists think for a while there was strong contact with "Siberian" languages and then by the later bronze age…”
What “Siberian” languages are you referring to? Paleo-Siberian like Kelteminar and Okunevo, or Finno-Ugric ones?
@Copper Axe
The whole linkage between Tarim Mummies, Tocharian languages and Afanasievo is sketchy at best.
Most TB Mummies were found to be R1a1, therefore Andronovo
@Vashista suggesting BMAC is Iranien and Indus Valley is Indo Aryan , you are truly taking drugs. Basal L657 found in Tajikistan population and there existe in sud Centrale Asia , Indo Aryan speakers, with dialects and cognate more proximite with archaic Sanskrit.
Con comme un balai.
Ran Vahaduo to identify the closest population average to all the ancient Turkish samples on G25:
Results:
https://imgur.com/a/HH3xCvU
Definitely not news, but you can really see the samples going from being closest to other ancients to tending to have their closest match as a present-day population.
Re; Yassitepe samples, what's quite interesting about them from this perspective is that:
A) I5735 and I5733 from 2900-2700 BCE present beth overall match with Hellenistic and Byzantine Period samples from NW Anatolia or the Aegean coast. So they represent a profile that was quite sticky in the broader region of Anatolia after at least this time.
B) The sample I5737 from 1000 years later (1900 BCE) has the best overall match to the average of the two GRC_Cycladic_EBA samples, from 2200 BCE, Kou01 and Kou03. People have suggested these two samples have some low level steppe ancestry. So maybe the I5737 might represent a further spread of people from the Aegean islands around Greece. Not too surprising. Or vice versa.
@vAsiSTha, related to BMAC, with the new samples from Mardin in SE Turkey and Iraq from PPN, do you think there is any case for that we can actually distinguish and actually BMAC having admixture from the centre specifically, rather than the NW Anatolian pool represented by Barcin?
@Rob “ For example, Germanic emerged from BAx and northern Beaker IE groups, as well as a couple of pre-IE groups.”
What pre-IE groups?
As far as we know Germanic was a merger of 2 Corded-Ware related groups: the CWC proper and Bell Beaker/SGC, although I suspect that “dog” and “boy” don’t have an IE etymology.
@Corded Axe TB Mummies were overwhelmingly R1a1 and a small minority was R1b. This is the main reason that if Tocharians = Tarim Mummies then they are a scion of the Andronovo horizon.
@Vasistha Since the majority of both BMAC’a and Elamites’ DNA derived from Iran_N or Hotu, their languages were most likely related.
@ Andrze
proto-Germanic acquired middle Neolithic ancestry from TRB and northwestern Europe (I1 and some I2a clades)
"The early Indo-European people were extremely mobile. Why should we assume they came in to any one place, replaced many of the locals and then stayed in place?"
The earliest attested Anatolian language speaking communities were not particularly mobile.
@vAsiSTha
"I get the feeling that you have not seen the Vikas Kumar Xinjiang paper."
Try not going too much on your intuition because clearly it isn't working too well. Obviously I have, quite extensively so I'd say. I've also read the necessary archaeology of the area quite a bit and was familiar with a whole bunch of the sites they sampled from. Let me tell you something about Zaghunluq for example: In terms of language it has always been considered a question mark because its on the border of the assumed linguistic area of Khotanese-Tumshuqese, and Tocharian C. But everyone who hypothesized them being potential Tocharian C speakers always took care because A; the boundaries of TochC are not too secure due to limited attestation and B: Zaghunluq is many centuries removed from its attestation. However the archaeology is a bit clearer in this regard; The grey ceramics of Zaghunluq connects it to the populations to their west, as the eastern parts of the Tarim Basin had a painted pottery tradition. So Zaghunluq being steppe_mlba derived is really a moot point for Tocharians, it fits in quite nicely with the archaeology actually.
Furthermore, a presence of steppe_mlba in Tocharian populations is not an issue, its to be expected considering the peoples to their west, east and north all had steppe_mlba. Even within the Tocharian zone not all inhabitants were Tocharian speakers, obviously. During the historical period we have tons of evidence of Sogdians for instance. In fact one of them became a rather famous mummy (Yingpan man). The question is if they also have input from other steppe populations, and the scant evidence we have seems to support this possibility.
As far as the models in the article goes, they modeled several samples with the Afanasievo/Chemurchek references (was it Xinj_BA3?) and for some of these I couldnt replicate it on G25 and it wasnt replicated on qpadm by others so far. Mind you that the authors themselves based on this data and their modell say that its either the Afanasievo route OR BMAC route for Tocharian. I went to Reich's lecture last week, he gave it 5:1 odds on it being Afanasievo over BMAC btw. Biggest discrepancy was Indo-Iranian, where he gave it 500:1 in favour of... well I don't even need to say it.
Also I would like a sample ID for the 50% Andronovo sample from XSQC/XKKD and the corresponding model.
@Andrzejewski
"@Corded Axe TB Mummies were overwhelmingly R1a1 and a small minority was R1b. This is the main reason that if Tocharians = Tarim Mummies then they are a scion of the Andronovo horizon"
Did you miss the article with full genomes where every EBA Tarim mummy had R-PH155? How does that even happen these samples were all over the mainstream news. The article you are referring to pertained to samples from Xiaohe and was done via short-tandem repeats. The new article has samples from the same layer as the R1a one, as well as sample(s) from a younger layer.
Keep in mind (god knows how many times I've said this) that the Tarim mummies were not a people, nor was it a tradition. It is a natural phenomenon. If you got lost hiking the Taklamakan and nobody would find you, you too would end up as a tarim mummy.
@Rob “ proto-Germanic acquired middle Neolithic ancestry from TRB and northwestern Europe (I1 and some I2a clades)”
Completely agreed. Particularly from GAC > TRB.
But I was referring to the linguistic aspect. There’s a scant evidence that non-IE had more than a non-negligible share in contributing to Proto-Germanic, although BBC and SGC had significantly more EEF genetically than CWC and modern Eastern Europeans.
''But I was referring to the linguistic aspect. There’s a scant evidence that non-IE had more than a non-negligible share in contributing to Proto-Germanic, although BBC and SGC had significantly more EEF genetically than CWC and modern Eastern Europeans.''
Yes there might be little in the way of lexical (vocab) inheritence or loans. But the preIE would be felt in other aspects of language - eg syntax, morphology & phonology. And even then it might be slight, despite - in population terms - the contribution was significant. This is what happens in cases of 'language shift' to new , socially dominant languages
Gandharva Grave Culture was Indo-Aryan, right?
Burusho people can be modeled as 15% Kalatasi_IA + 85% Swat_IA (SPGT)
Target: Kalatasi_IA:C1706
Distance: 2.7469% / 0.02746943
53.8 MNG_Slab_Grave_EIA_1
26.6 RUS_Sintashta_MLBA
11.0 UZB_Bustan_BA
08.6 MNG_Xiongnu_East_Asian
Target: Kalatasi_IA:C1705
Distance: 2.7699% / 0.02769858
49.4 MNG_Slab_Grave_EIA_1
28.0 RUS_Sintashta_MLBA
16.0 MNG_Xiongnu_East_Asian
06.6 UZB_Bustan_BA
Slab grave people were proto-Turks right? So these kalatasi people with 24% sintastha spoke Indo-Iranian or Proto-Turkic?
Burusho speak an unrelated language which is derived from neither their steppe heavy Mongolian ancestry nor from Swat which spoke Old Indic/Prakrit.
//I went to Reich's lecture last week, he gave it 5:1 odds on it being Afanasievo over BMAC btw. Biggest discrepancy was Indo-Iranian, where he gave it 500:1 in favour of... well I don't even need to say it.// This was indeed odd giving huge Bayesian factor 500:1 for claiming Sintashta as Indo-Iranian. btw Reich had to change his slide deck after coming to Leiden at the last moment and did not sleep properly for 48 hours before lecture according to his own admission. probably playing to the gallery ;-)
Modern Kurds(Iranian speakers) come from almost 100% from Hasanlu IA which had almost nil Sintashta ancestry is huge factor to be considered. not sure 500:1 would hold water.
Gandharva Grave Culture 1400-800bce is considered Indo-Aryan by allmost everyone.
Gandhari Prakrit was in use from 3rd century BCE to 4th century.
SWAT people on average carry 15% steppe via females and barely any R1a lines(1 or 3 samples). Mostly E1b1b, L1a and R2.
Fast forward medieval times and many Afghans made inroads in the region.
Today 90% of swat is Iranic Pashtoon who are largely R1a z2125 types.
"Other languages, mainly the Kohistani languages of Torwali and Kalami, are spoken by 7.35% of the population, and form the majority in the Swat Kohistan region of Upper Swat."
Would be interesting if this 7% IA speaking population has same profile as the swat IA. Pashtoons on the other hand have increased steppe ancestry ~35% or so.
Why did Tarim mummies, Tocharians and other ANE and IE like people go extinct? Were they slowly absorbed by numerically superior Turkic Mongolic and Sinitic groups or were they genocided in a similar fashion to what happened to the Dzungars in the 1800s? Its weird to me that euro~caucasoid phenotypes and Tocharian and Iranic language persists in that part of the world until roughly about the Viking age, at which point they all but disappear except in a few Pamiris or the occasional village Uyghurs, to the point where in Marco Polo's time the Chinese elite had never even seen a europesque looking person before.
@Muthu
There is Sintashta ancestry in Kurds because they have R1a-Z93.
So there's nothing controversial about Kurds.
Two very simple points. We know that in the Neolithic and the Chalcolithic, the farmer population of southeastern Europe was much larger compared to that of the North, the latter showing signs of population decline.
Conversely, steppe ancestry is much lower where the Neolithic population was higher. Mycenaeans spoke an IE language, yet steppe ancestry varied between 0-16%.
Why would this scenario not apply in Anatolia, which was even more densely populated than Greece? To be honest, its impressive that even 2% steppe survived there.
Genetics has shown that avalanche-type migrations are most frequent, not the migration of discrete, unmixed groups that replace the preceding population. This is what happened in Surope, Central Asia, and the Indian subcontinent.
Mongols!
Z2124 in Huns
Z93 in Arabs
@andrze
"@Vasistha Since the majority of both BMAC’a and Elamites’ DNA derived from Iran_N or Hotu, their languages were most likely related."
Lol. Hajji Firuz is 90% Iraq_PPN like, IranN disappeared from Zagros region by that time and moved east. By 4700bce at seh_gabi_c, 70% of the ancestry is like Arm_aknashen, only 20% iranN type remains. Aknashen is 50% iraq_ppn, 25% BracinN, 25% CHG. Also elamite/Dravidian hypothesis is bunch of horseshit, and almost no linguist accepts it apart from the 2 or 3. The only place where higher % of IranN persists is SC asia (allentoft models Monjukli_N 4700bce as 65% IranN 30% CHG 5% ANE).
@Caxe
"Also I would like a sample ID for the 50% Andronovo sample from XSQC/XKKD and the corresponding model."
C795 Xikakandasayi Xinj_HE6 1696-1544cal BP
modeled by authors in qpAdm as
Xinj_BA1_TMBA1 0.201, Andronovo 0.481, Han 0.318
In G25:
Target: CHN_Xikakandasayi_HE:C795
Distance: 2.6034% / 0.02603391
22.8 Han_Hubei
21.8 RUS_Sintashta_MLBA
21.0 RUS_Shamanka_EBA
18.8 PAK_Saidu_Sharif_H
15.6 RUS_Afanasievo
Already here, in 500ce, we see mixing between an Indian buddhist (Saidu_sharif_h) and a Tocharian.
If afanasievo is removed, distance doesnt worsen much (2.77%)
Target: CHN_Xikakandasayi_HE:C795
Distance: 2.7739% / 0.02773947
32.4 RUS_Sintashta_MLBA
22.0 Han_Hubei
21.4 PAK_Saidu_Sharif_H
20.0 RUS_Shamanka_EBA
4.2 CHN_Tarim_EMBA1
But if Sintashta is removed, distance worsens a lot to 3.18%
Target: CHN_Xikakandasayi_HE:C795
Distance: 3.1822% / 0.03182225
33.4 RUS_Afanasievo
23.8 Han_Hubei
23.0 PAK_Saidu_Sharif_H
19.8 RUS_Shamanka_EBA
@davidski
"There is Sintashta ancestry in Kurds because they have R1a-Z93.
So there's nothing controversial about Kurds."
Target: Kurdish_avg
Distance: 0.8963% / 0.00896332
95.6 IRN_Hasanlu_IA
2.8 TKM_IA
1.6 ARM_Bagheri_Tchala_EIA
Go ahead and prove that Kurds have any significant Sintashta ancestry. NW iran was already iranized by 1000bce, show presence of Sintashta ancestry in those samples.
I don't need to prove that Kurds are significantly Sintashta, because I know that Kurds have the same type of R1a that other Iranic speakers have, and this R1a arrived in Iran during the Iron Age when Iranian languages arrived there.
So everything fits. Haha.
Dinkha_B samples Dated between 1400-800bce. NW Iran, 25km from Hasanlu.
Target: IRN_DinkhaTepe_BIA_B
Distance: 0.8066% / 0.00806570
43.4 IRN_DinkhaTepe_BIA_A
34.8 UZB_Bustan_BA
13.2 ARM_Lchashen_MBA
5.8 RUS_Catacomb
2.8 RUS_Kubano-Tersk_Late
0.0 RUS_Sintashta_MLBA
0.0 TKM_IA
Here, this is proof that W Iranians came from BMAC. Aligns perfectly with the time of Iranian entry into the region (TC Young, Kuz'mina, Douglas Adams among others, based on attestation by Assyrian tablets).
Even Dravidian and arabic speakers have R1a-Z93, it's a great argument only if I were brain dead. R1a doesn't have some magical powers that if it so much as sneezes at someone they start speaking IE.
Yeah, and obviously both Dravidian and Arabic are closely related to Balto-Slavic in Eastern Europe where R1a-Z93 comes from.
Bwahaha...
Your own model
Burusho
IRN_Shahr_I_Sokhta_BA3 0.710±0.029
KAZ_Botai 0.048±0.024
RUS_Sintashta_MLBA 0.242±0.028
chisq 14.509
tail prob 0.206105
Full output
Burusho
IRN_Shahr_I_Sokhta_BA3 0.718±0.028
RUS_Sintashta_MLBA 0.243±0.028
RUS_West_Siberia_N 0.039±0.023
chisq 14.182
tail prob 0.223101
Full output
28% Burusho carry R1a.
So these people would spread both R1a and steppe ancestry but not Indo European language in future.
This just supports my contention that Indo-Iranian is a linguistic secondary convergence area. In V’s model, the Catacomb ancestry is also present in Kubano-Tersk & Armenia MBA. But looking at Iron Age is too late anyhow
Also Kurds have I2a-L68
@CopperAxe
Even the other sample, dated to 300ce, picks up sintashta ancestry, and minor buddhist admixture.
Target: CHN_Xianshuiquangucheng_HE:C2032
Distance: 1.5748% / 0.01574838
29.2 RUS_Shamanka_EBA
23.6 RUS_Sintashta_MLBA
16.4 UZB_Bustan_BA
10.2 CHN_Tarim_EMBA1
7.4 Han_Hubei
7.0 RUS_Afanasievo
6.2 PAK_Saidu_Sharif_H
0.0 KAZ_Dali_EBA
0.0 KAZ_Dali_MLBA
0.0 PAK_Saidu_Sharif_H_o
@vAsiSTha
"Already here, in 500ce, we see mixing between an Indian buddhist (Saidu_sharif_h) and a Tocharian."
By this time buddhism had spread into the Tarim Basin for several centuries, as had the brahmi script. South Asian influences predate the common era. Try making it a habit to actually inform yourself about a topic before speaking on it so confidently, because people will have a hard time taking you serious when you present yourself as a laughing stock to the world.
As for calling this sample an Indian-Mixed Tocharian:
Target: CHN_Xikakandasayi_HE:C795
Distance: 2.0891% / 0.02089099
32.2 CHN_Upper_Yellow_River_LN
27.0 RUS_Sintashta_MLBA
13.4 TKM_Gonur1_BA
13.0 RUS_Tyumen_HG
8.4 MNG_North_N
6.0 IRN_Shahr_I_Sokhta_BA2
Target: CHN_Xikakandasayi_HE:C795
Distance: 2.1618% / 0.02161792
41.6 KAZ_Kyzlbulak_MLBA2
31.8 CHN_Upper_Yellow_River_LN
9.4 RUS_Late_Xiongnu
8.2 KAZ_Otyrar_Antiquity
6.2 PAK_Saidu_Sharif_H
2.8 CHN_Tarim_EMBA1
Target: CHN_Xikakandasayi_HE:C795
Distance: 2.0102% / 0.02010201
40.8 CHN_Simutasi_IA1
28.6 CHN_Zaghunluq_IA2
22.4 CHN_Upper_Yellow_River_LN
7.8 PAK_Saidu_Sharif_H
0.4 CHN_Tarim_EMBA1
Burusho Y Haplogroup %age
n=97
C 8.2
E 0
F 1
G 1
H 4.1
I 0
J 8.2
K 0
L 16.5
N 0
O 3.1
P 1
Q 2.1
R 0
R1 0
R1a 27.9
R1b 0
R2 14.4
T 0
@Leonidas D
Its almost like an analogue to the colonisation of the Americas, where European settlers displaced the Native Americans almost entirely in North America, but in Mesoamerica and South America they were much more mixed with locals, even though Spanish colonialism was in many ways far more brutal and destructive than that of the British. I wonder if a similar occurence happened with the bronze age colonialism of the Indo-Europeans
Mitanni were proto-Iranian. So Iranians were already in Iraq-NW Iran by 1500bce.
@Copperaxe
You yourself are showing 7% South asian ancestry.
Do you deny that Indian buddhists were in this region? Do you deny that given the buddhist nature of Tocharian texts, it is likely that both admixed to some extent?
Do you deny that SAsian ancestry could have gotten diluted over time?
DO you deny the presence of significant andronovo admixture in those samples?
"Mitanni were proto-Iranian. So Iranians were already in Iraq-NW Iran by 1500bce."
But muh R1a. reee
@vAsiSTha
You should take your meds because you are fighting figments of your own mind. Where did I fight these points? All I pointed out was your ill-fitting models were overblowing South Asian ancestry in that sample.
@Germ said "even though Spanish colonialism was in many ways far more brutal and destructive than that of the British"
Really? Whenever you want, we talk about British colonialism
North America was also colonized by the Spanish-Northern California, New Mexico, Nevada, Arizona, Utah, Colorado, Texas, Louisiana, Florida etc. ....
@Leonidas D said "Genetics has shown that avalanche-type migrations are most frequent, not the migration of discrete, unmixed groups that replace the preceding population. This is what happened in Surope, Central Asia, and the Indian subcontinent"
When exactly did these mass migrations occur in Greece?, what were the uniparental markers involved?
Andrew said... "The earliest attested Anatolian language speaking communities were not particularly mobile."
My comment about population change was relative to these two findings:
pnuadha said... "Who are they talking about as having 2% EHG? Is it people in anatolia after the time of the persian takeover?"
Davidski said... "All they found was ~2% EHG in Early Medieval Central Anatolia"
There early Indo-Europeans were extremely mobile. I don't know about the Hittites, Luwians, etc. but these samples are not decades or centuries later than the origin of Indo-Anatolians, they are millennium later. It doesn't take a lot of mobility or a lot of immigration to change populations over this much time.
It seems like many folks want to assume populations stayed in place for long periods. That's been proven wrong time and time again.
@ Davidski
Good to see you've labelled I6184 as 'BA"
Harvard's balkan boy, Dusan Boric, who supplied the sampled still called it Neolithic.
So this guy is R1a and looked like GCW - GAC + Yamnaya. But in strict terms CWC wasn;t in Moldavia. So this could be from the famed Budhzak culture
@MikeW
It seems like many folks want to assume populations stayed in place for long periods. That's been proven wrong time and time again.
This is one of the several wrong assumptions that the Lazaridis et al. conclusions rest on.
I believe we also have the earliest "Slavic-related" I2a - L621 in Monteoru culture, eastern Romania
not carbon dated, but doesnt seem to be outlier. Monteoru seems to be I2a dominated, but has a G2a and R1a-CTS1211 there
@ copper axe
You’re not exactly helping your argument by quoting Reich
@All
Need some serious feedback for a future blog post:
- is the Lazaridis, Alpaslan-Roodenberg et al. paper actually worse than Gray, Atkinson 2003?
https://www.nature.com/articles/nature02029
- what's the fundamental reason why these sorts of papers appear in major journals?
Off-topic: https://www.eshe.eu/static/eshe/2022/ESHE_2022_Schedule.pdf - 22/09/2022
Genetics related abstracts:
Vanessa Villalba-Mouco - Genetic link between Aurignacian and Magdalenian in Ice Age western Europe
Owen Alexander Higgins - Biological profile and early life history of the Upper Palaeolithic infant of Grotta delle Mura (Puglia, Italy): integrating dental histology, aDNA, and time-resolved geochemical signals
Sophy Charlton - Dual genetic ancestries of the Late Glacial Palaeolithic in Britain
Cosimo Posth - The genomic ancestry of European hunter-gatherers
New paper
Genetic continuity, isolation, and gene flow in Stone Age Central and Eastern Europe
https://assets.researchsquare.com/files/rs-1966812/v1_covered.pdf?c=1662993174
@Andrzejewski& Rob
Some have speculated that the Germanic soundshift was the result of a non-IE substrate, like the TRB. But this is impossible, because the Germanic soundshift didn't occur before 500 BC. Loanwords that didn't reach northern Europe before 500 BC, and which yet have undergone the Germanic soundshift are evidence for this. Like Skythian *kanba -> Germanic *xanapiz (hemp) or Skythian *baita (a sheperd's skirt) -> Germanic *paitho. For comparison, none of the Latin loans that reached central Europe in the 1st century has undergone the Germanic soundshift. Also the fact that the three classes of occlusives, that were affected by this soundshift are preserved unchanged and not mixed up in the early attested Germanic texts is evidence against an early date of the soundshift.
@Davidski
Gray & Atkinson is by far the worse of the two. They are are applying a technique that simply is not applicable to the field of study. Lazaridis just brought some opinions or perhaps just one big one to the paper.
@copperaxe
"You should take your meds because you are fighting figments of your own mind. Where did I fight these points? All I pointed out was your ill-fitting models were overblowing South Asian ancestry in that sample."
The ill-fitted models were made to show presence of andronovo ancestry in the tocharian region samples and compared to afanasievo related ancestry. south asian ancestry (whatever %) was a discovery. Of course, using other proximal chinese samples would have made the fit better, that wasn't the point.
The overall point from me here is that that the south asian ancestry (diluted) is seen in tocharian regions very close to the time the tocharian texts were discovered. And most importantly, this ancestry is admixed with western ancestry of the andronovo kind rather than the afanasievo kind. Which makes andronovo a likely vector of Tocharian as compared to Afanasievo. There's no 3rd option.
There is one Eneolithic sample from the Dnieper 3500 BC. She’s Yamnaya like. Volga EHG/ CHG mammas
@ SimonW
Yeah that’s the old Germanic Substrate hypothesis , but it wasn’t what I’m referring to
Sorry not Yamnaya like, but shifted toward it on PCA
@Rob, Ukr104 "Sredny Stog" 500 years before Yamnaya interesting. Like Yamnaya+33% Ukraine_N. Somewhat anachronistic or reversed model. Need more CHG related ancestry to get to Yamnaya. Would CHG+Anatolian mix work? Leaked PCA by David Anthony suggested "early Sredni Stog" 1x sample would be north shifted compared to later SS and Yamnaya.
@Kouros
"Basal L657 found in Tajikistan population and there existe in sud Centrale Asia , Indo Aryan speakers, with dialects and cognate more proximite with archaic Sanskrit."
R-Y3 has been found at Alai Nura dated to 100 or 200ce iirc. Y3 was born 2500 years before that. R1a-L657+ has been found in Xinjiang too at Guanjingtai dated to 100ce or so. L657 wsa born 2300 yrs before that. Not to be a spoilsport, but this is due to migration from India during the buddhist times.
@Matt
"@vAsiSTha, related to BMAC, with the new samples from Mardin in SE Turkey and Iraq from PPN, do you think there is any case for that we can actually distinguish and actually BMAC having admixture from the centre specifically, rather than the NW Anatolian pool represented by Barcin?"
Indeed, I havent used the Mardin samples, but IRQ_Nemrik-PPN samples are a game changer. Latest post on my blog is my best effort related to mapping the ancestries from anatolia to SC Asia and IVC and everything in between from the HG/Neolithic onwards. Have a look.
@Rob & @SimonW Germanic was a melange of CWC and SGC - hence it’s unique spot on the IE tree. No substrate whatsoever.
But the same applies to almost any IE language, with the exception of perhaps Greek.
The no non-IE substrate, that is.
Also, the idea that Tocharian language(s) was/were influenced by Samoyedic languages are bunkers because if T is indeed an Afanasievo derived branch, it split from PIE 5 kyo but Proto-Uralic, ancestral to Samoyedic wasn’t influenced by Sintashta Indo-Iranian until 1,500 years later.
@Davidski
//There is Sintashta ancestry in Kurds because they have R1a-Z93.//
not really.
20% of Modern kurds (in Iran) have R-M17/M198 (R1a1a), Grandfather HG of R1a-Z93(Sintashta), even R1a-M417(Fatyanovo–Balanovo)!
https://journals.plos.org/plosone/article/figure?id=10.1371/journal.pone.0041252.t001
//So there's nothing controversial about Kurds.//
in fact, big question mark regarding Sintashta ancestry in Kurds, both autosomal and Y
@Muthu
The paper you cited is from 2012, so it's outdated. The authors didn't even know that R1a-Z93 existed.
The ~20% of R1a in Kurds is mostly R1a-Z93.
I've tested plenty of Kurds. They have both Z93 and Sintashta autosomal DNA.
Quit posting crap in my comments and educate yourself.
"The ~20% of R1a in Kurds is mostly R1a-Z93."
Plz show references for both, 20% number and Z93.
"I've tested plenty of Kurds. They have both Z93 and Sintashta autosomal DNA."
How much? 4% like in anatolia? Care to plot a simple pca with Hasanlu_IA and Kurds? Or do a simple F4 test wrt Hasanlu_ia and Kurds?
There's 10-20% R1a in Arab speakers too. I guess steppe superpowers ended before reaching Arabs. R1a magically only becomes relevant when some IE speaking pop is involved lol.
There are plenty of good reasons why there's Indo-Iranian R1a-Z93 in Arabs.
For one, Arabia is just across the Persian Gulf from Iran and Pakistan.
You dipshit.
A bit off tangent but related somewhat to language shift in Anatolia ; )
"Some have speculated that the Germanic soundshift was the result of a non-IE substrate, like the TRB. But this is impossible, because the Germanic soundshift didn't occur before 500 BC."
@Simon,
I would agree on a late Germanic sound shift, but that does not exclude the possibility that the old TRB language in the vicinity of the pre-Jastorf was continually spoken throughout the entire Bronze Age, despite peoples there being mixed with or mostly IE. This was the case in much of Iberia and Southern France apparently, and I suspect the Irish Sea and Southern Ireland as well.
It's probably worth pointing out that Finland is especially disjointed between language and culture. Although that may have been a different set of later circumstances
Arabs don’t have 20% R1a, apart from small subgroups
R1a-Z93 in Arabs and Jews is a recent founder effect. Just check their tmrca
@ Matt
''. Need more CHG related ancestry to get to Yamnaya. Would CHG+Anatolian mix work? Leaked PCA by David Anthony suggested "early Sredni Stog" 1x sample would be north shifted compared to later SS and Yamnaya.''
Im not 100% sure, as we would need to check the data for ourselves when it's out.
"There are plenty of good reasons why there's Indo-Iranian R1a-Z93 in Arabs.
For one, Arabia is just across the Persian Gulf from Iran and Pakistan.
You dipshit."
Lopez et al 2017 studied 938 Iranians.
10% R1a in non zoroastrian iranians'. 5% in lay zoroastrians. 12% in zoroastrian priests.
So far, from 10000-600bce, not a single R1a found from Iran. Neither is there a single sample with Sintashta ancestry. So when did the Iranians enter iran? I'm sure you will say whenever the first R1a is seen, archaeology and literature be damned. That's because you're a propagandist,
@rob
Kuwait, Oman has basal R-Y3 and R-Y2 moderns. Check yfull.
Here's a bit of background reading about the time when Iranians entered Iran.
https://www.cambridge.org/core/journals/antiquity/article/palimpsest-grave-at-the-iron-age-cemetery-in-estarkjoshaqan-iran/6B08213F95174AF5BB233B5EB77F1252
And here's my blog post about an early Iranian. Still pure gold three years later...
https://eurogenes.blogspot.com/2018/07/an-early-iranian-obviously.html
@ Vasistha
You nee to understand the difference between basal, and collateral &/or unresolved or unmatched lineages. it's the same with R1b-PH155, its not basal , but a minor sidebranch of the main R1b bulk.
Then, you need to know the personal history of those Omani gentlemen, which as I said, are a minor Arab subgrup, and in no way representative of the entire Arab world. Do you know their geneaologies ?
what do think about the higher resolution Smyadovo_Chl outlier ?
~ 4600 BC, 0.1 x, > 100 K SNPs, no contamination
He's about 70% MNE 30% Yamnaya
They look like east Iranian speaking dahae or massagatae.
Kurds speak West Iranian languages.
"Old Azeri, Gilaki, Mazandarani, Zaza–Gorani, Kurdish, and Baluchi."
The Iranic ancestors of the Kurds are also from Airyanem Vaejah, that's why Kurds have R1a-Z93.
Pure gold this...
https://eurogenes.blogspot.com/2018/07/an-early-iranian-obviously.html
There were Indo-Iranians in Kurdish inhabited lands in 1500bce before you forget.
The Mitanni were Indo-Aryans you moron.
And these Mitanni kings had names Arta-shumara and Arta-tama instead of Indo-Aryan "rta".
Arta is an old Persian variant of Vedic rta and avestan (east Iranian) asha.
Biryamashda instead of birya medha.
Mitanni worshiped Indo-Aryan God's that means they had an Indo-Aryan clergy attached to them.
Artabanus
Artaxerxes
Artasyrus
Artabazos
Your own well lady from 1500bce alalakh derive her ancestry from BMAC which is a common feature across Iranian speakers :)
And then the similarities between Mitanni and BMAC is already well established through archaeology.
I’m just a passerby commenting as an impartial observer and not rubber stamping anyone’s position. I also appreciate the time and effort put into the “Southern Arc” latest papers. I also totally get why EHG was used in lieu of Steppe-EMBA or MLBA to estimate steppe input into various Southern Arc populations. The EHG estimates will be off for some populations for the following reasons:
1- An important assumption that’s always made when using any type of admixture calculator whether it be Admixture or qpAdm or f4s is the sources used are actually ancestral to the target tested. However for some of the targets they tested had substantial Iran-N ancestry which they failed to use as a source. Instead CHG was uniformly used as a source. The reason this is important is because if they had used Iran-N in addition or in lieu of CHG, the EHG qpAdm estimates would have increased.
2- It’s very important to do a sanity check using qpWave in Rank 0 2 source mode to check the cladeliness of EHG and a Southern Arc. Iosif, Patterson and I are usually on the same page when it comes to this and I intend to share my results that I have compiled over the past couple of weeks using qpWave and qpAdm with them as well as publishing at EurasianDNA.com
@vAsiSTha
1- You can’t take the lazy road and use one calculator result to conclude that contemporary Kurds are 97%+ 2700 year old Hasanlu-IA. Our research has shown that except for really isolated populations such as Andamanese and such modern populations such as Kurds who number in the millions and cover thousands of square miles in such a crossroads area are all 97% a 2700 year old sample from one point. In the case of Hasanlu and Dinka-B we have confirmed this with qpWave and qpAdm and have found other Iron Age populations that form a much tighter clade with Kurds and when combined with Hasalu or Dinka cause their contribution to go way down. We will be publishing all this soon
2- We have also tested most of the recent Southern Arc and Simons WGS including Kurds for cladeliness with EHG using qpWave and will be sharing this also
"Here's a bit of background reading about the time when Iranians entered Iran.
https://www.cambridge.org/core/journals/antiquity/article/palimpsest-grave-at-the-iron-age-cemetery-in-estarkjoshaqan-iran/6B08213F95174AF5BB233B5EB77F1252
And here's my blog post about an early Iranian. Still pure gold three years later...
https://eurogenes.blogspot.com/2018/07/an-early-iranian-obviously.html"
That's cool and all but the burials predate the TKM_IA(~855) guy by a few centuries. You know TKM_IA was around the time the Medes and Persians formed regional powers on the borders of Assyria.
But here's the shocker David cremation has been found in that region since the neolithic!
https://www.researchgate.net/publication/271518136_Evidence_of_Late_Neolithic_cremation_at_Tepe_Sialk_Iran
You know David that R1a could have reached Kurdistan with the Cimmerians and Scythians independently of the Proto-West Iranians who we have a good evidence for in Hasanlu now right?
It's okay David you can continue to shift the goalpost. I guess the Hasanlu Mitanni and Iranians were just twitter fanboys hahahahaha.
From Syria.
"Dated to the period between 2.5 Kyrs BC and 0.5 Kyrs AD the studied individuals carried mtDNA haplotypes corresponding to the M4b1, M49 and/or M61 haplogroups"
No wonder there are Indo-Aryan loans :)
So CHN_Xikakandasayi_HE:C795 is a Tocharian? Cool! With this sample it is possible to model Tocharian as a Corded Ware offshoot separate from Indo-Iranian.
Target: CHN_Xikakandasayi_HE:C795
Distance: 1.9005% / 0.01900547
24.2 CHN_Western_Liao_River_MN
23.0 Corded_Ware_CHE
19.8 IRN_Shahr_I_Sokhta_BA1
17.6 RUS_Sosonivoy_HG
15.4 CHN_Upper_Yellow_River_LN
From the perspective of linguistics that would make a lot of sense. I see Uralic as Para-Indo-European. If they got that through Afanasievo via a kra001-like population, I don't think Afanasievo can also be the source of the Tocharian languages.
Rob: Im not 100% sure, as we would need to check the data for ourselves when it's out.
Yeah, no doubt.
These samples who are rich in Ukraine_N (this woman, "early_Sredny_Stog Ukraine" in David Anthony's PCA if that's a different sample and the Don HG from 5500 BCE from "Genomics of Stone Age...") pose the question of whether they contributed at all to Yamnaya / early_CWC... in which case they would require some CHG-rich supplementary source that's richer in CHG than Progress or the maximum of the Khvaylnsk cline (unless they made very small contribution) - https://imgur.com/a/WZl54AY (maximum of the Khvalynsk cline looks like Berezhnovka site on that Anthony PCA).
But if they were simply totally replaced without admixture, no problem.
@ Dilawer
There’s more to archaeogenetics than admixtools. Ydna , archaeology , etc all need to be considered alongside a (rather questionable) series of distal models
Otherwise, all what you get is a caricature instead of something inciteful
To give you a context R1a and steppe rich Iranians (scythians) left behind inscriptions in Sanskrit. Nothing different from their Mitanni ancestors.
"dharmma-vijayina Śaka-Nanda-putrena Maha-danda-nayakena Śakena Sridharavarmmana var[shsha]-[saha]sraya sva-rajy-abhivriddhikare vejayike sarhvatsare trayodasa[me]"
@bellbeakerblogger
@Andrzejewski
@Simon_W
@Rob
'I would agree on a late Germanic sound shift, but that does not exclude the possibility that the old TRB language in the vicinity of the pre-Jastorf was continually spoken throughout the entire Bronze Age, despite peoples there being mixed with or mostly IE. This was the case in much of Iberia and Southern France apparently, and I suspect the Irish Sea and Southern Ireland as well.
It's probably worth pointing out that Finland is especially disjointed between language and culture. Although that may have been a different set of later circumstances'
Yeah Finnic language is definitely a different story - We know that it came with LBA-EIA migrations from Uralic-speaking populations, making it "post-IE" rather than "pre-IE" like those languages of Iberia and Southern France (and Etruscan!).
I think you raise a very good point and very interesting possibility, that we can't completely discount that there may have been pre-IE languages spoken in North-western Europe throughout the bronze age. Some of the Tollense valley battlefield samples look like they're from some population which is richer than the others in (WHG-rich) EEF ancestry, so I wonder if these guys could have been from a putative non-IE-speaking northern group.
But this is so much "maybe this could have happened" speculation - While interesting, and possible, we lack positive evidence. The Germanic substrate hypothesis - someone correct me if I'm wrong - seems to have been really plucked apart since its inception and we seem to have gone from "1/3 of proto-Germanic vocabulary appears to have a non-IE origin" to "there appears to be a small handful of words in proto-Germanic with no clear PIE origin", which is probably not a state of affairs unique to Germanic.
With regards to Y-lineages like I1, since someone mentioned it, we still don't really know where it even comes from. Despite all of the mesolithic and neolithic samples we have, it just appears out of thin air in late neolithic Scandinavia, in a single sample rich in steppe ancestry, and then it "properly appears" in a few more samples in what seems to be a bronze age founder effect in Scandinavia. I'll bet that if we had more bronze age samples from the north-western corner of the continental mainland, we'd see it among them, too. But what's notable is that we haven't found a single earlier sample with a subclade of I1 with any chance of being ancestral to I1-M253, hence my comment of it appearing out of thin air. It's very odd indeed.
@dilawer
"1- You can’t take the lazy road and use one calculator result to conclude that contemporary Kurds are 97%+ 2700 year old Hasanlu-IA. Our research has shown that except for really isolated populations such as Andamanese and such modern populations such as Kurds who number in the millions and cover thousands of square miles in such a crossroads area are all 97% a 2700 year old sample from one point. In the case of Hasanlu and Dinka-B we have confirmed this with qpWave and qpAdm and have found other Iron Age populations that form a much tighter clade with Kurds and when combined with Hasalu or Dinka cause their contribution to go way down. We will be publishing all this soon".
Cool, look forward to seeing these Iron age samples published.
Frankly, the ancestry of Modern Kurds should have been irrelevant to this conversation but it is just additional proof to be served to show no excess steppe ancestry entered this region.
Fact is that latest by 1000 BCE, NW Iran was already teeming with Iranian speakers from BMAC. The archaeology (and literature) based consensus predates the southern arc samples by well over 20 years. And like good science does, it proves that the entrants into the region at this consensus time were the BMAC people, along with IVC ancestry in some samples (including L1a2 swat subclade). Sintashta ancestry or R1a are nowhere to be seen. This should actually be the end of the matter.
Even after its proven that only BMAC like people entered this region at the consensus time of Iranian entry, propagandists like Davidski will still screech about Sintashta ancestry, R1a and 850bce TKM_IA and claim that whenever Sintashta ancestry entered this region is when they become Iranians (because conclusion comes first, not facts).
That's ok, because F4 shows Sintashta has more affinity to Hasanlu_IA than to Kurds(and all Iranians) showing that no external steppe rich ancestry entered later.
This scenario is the reverse for North Indians, where clear excess sintashta affinity is seen in North Indians after IVC age.
@Davidski
"And here's my blog post about an early Iranian. Still pure gold three years later..."
Lol, its obvious garbage.
Target: Kurdish
Distance: 0.7429% / 0.00742891
95.6 IRN_Hasanlu_IA
4.4 MDA_Cimmerian
0.0 ARM_Bagheri_Tchala_EIA
0.0 ARM_Black_Fortress_LBA
0.0 RUS_Kubano-Tersk
0.0 RUS_Kubano-Tersk_Late
0.0 RUS_Sintashta_MLBA
0.0 TKM_IA
0.0 UKR_Cimmerian
0.0 UKR_Cimmerian_o
@ Matt
my reservation was that there must have been at least some continuity not detected by admixture modelling (which isn't perfect), because of the Y-hg I2a in Yamnaya found in the Dnieper sites. But if it was also found in the Don, along with R1a, then Don foragers might have been the prominent set
@ H₂ŕ̥ḱtos
Nice comments, but to clarify I was not talking about vocabulary, but at a population level for Nordic Bronze Age, the putative proto-Germanics. When a population shifts to that of another, whether by force or sheer charisma, you dont get blending of vocab. Instead, the elements of the langauge of shifting speakers can be preserved in syntax & phonology. But when that language is not even attested, then its virtually impossible to know. So yeah it's all theoretics
To me, I1 isnt a great mystery. The chronologically most proximate I1 is in France, and the extesnvie flint-dagger trade network which stretched from SW France to Scandinavia is a hallmark of this, but I1 itself probably moved within the BB network which acquired it.
Earliest Proto NW Iranian. 1350bce.
Target: IRN_Hasanlu_LBA_A:I4097
Distance: 1.4178% / 0.01417807
61.4 UZB_Bustan_BA
19.8 IRN_Hasanlu_LBA_B
9.0 RUS_Catacomb
3.4 ARM_Black_Fortress_LBA
3.2 ARM_Lchashen_LBA
3.2 RUS_Kubano-Tersk
0.0 ARM_Lchashen_MBA
0.0 IRN_Hasanlu_MBA
0.0 RUS_Kubano-Tersk_Late
@Assuwatama “ Your own well lady from 1500bce alalakh derive her ancestry from BMAC which is a common feature across Iranian speakers :)
And then the similarities between Mitanni and BMAC is already well established through archaeology.”
BS. Hogwash.
Assuming that Indo-Iranian has any discernible BMAC (or WSHG for that matter either) is just like assuming that Proto-Germanic has 1/3 of its vocabulary being non-IE.
Both are crazy conspiracy theories.
On the same vein, can any geneticist here with a brain find out the origins of the Burusho and whether their language isolate is the IVC language? Maybe Burushaski was BMAC related or the language of the Hotu Iranians HG while Dravidian languages were actually the languages of the South Asian Onge-like HG of South India?
Who are the other language isolate speakers of the Indian subcontinent- the Nihali and the Kusunda?
Claiming that PIE is Out of India is just as insane as those theories I read online (written by African-American professors!) that attribute the evolution of country music to 17th century slave ships, just because the banjo was brought over from West Africa, where in reality it evolved from Celtic (in particular Appalachian Scots-Irish), and even has similar motives with Cajun music, Polka music and Kleyzmer music.
@Assuwatama and Vasitha
Without letting the cat out of the bag we have found a Parthian era sample younger than Hasanlu and Dinka and TKM-IA with good coverage sample in their dataset that’s likely mislabeled and is a Parthian in disguise.
This sample is from the eastern borders of the Parthian empire so a little east of TKM-IA. It’s slightly less EHG shifted than TKM-IA with about the same amount of EHG shift as modern Brits so plenty of Steppe. It forms a much tighter clade with Kurds than Hasanlu, Dinka or TKM-IA. Unlike Hasanlu or Dinka Kurds can be modeled 100% with this sample with good p-values.
When combined with Dinka or Hasanlu it causes P-Values to shoot up in qpAdm and Hasanlu/Dinka values to shoot down to around 50% for Kurds (30X WGS) while at the same maintaining decent standard errors. This is consistent with linguistics, archeology and history where Parthians form a very substantial layer on top of Hasanlu era Iranics in the ethnogenesis of Kurds.
Inspite of our numerous attempts by us to cause qpAdm p-values to fail by adding more and more pright outgroups (we used 15 outgroups from across Eurasia many Neolithic and younger which usually does cause p-values to fail) they wouldn’t fail.
Details are forthcoming
Some interesting place names mentioned in Mitanni-Hittite treaty:
City of Arahati
Mount Karna
land of Isuwa
land of Kalasma
Mount Harana
land of Tegarama
city of Arawanna
What is the native American ancestry made of?
East Eurasian and Ancient North Eurasian, right?
Native American languages
There are approximately 296 spoken (or formerly spoken) Indigenous languages north of Mexico, 269 of which are grouped into 29 families (the remaining 27 languages are either isolates or unclassified).
If true how do you explain that?
Given the strict views Zoroastrianism has on cremation I doubt those cremation burials have anything to do with the migrating Iranians, who definitely would have been majority Zoroastrian in those periods.
@Dilawer
Which sample ID does this Parthian have?
@Dilawer
Al Jazira and Adiabene has been settled by various different nomadic groups over the centuries. Kurdish itself is related to Parthian and may have some "eastern influence" so it should not be a suprise.
However, that has little to do with the supposed first contacts of the west Iranians with Assyria 1100 BCE (mind you the first attested specifically west Iranian name is found not far from Kurdistan 1300 BCE).
So Vasistha is actually right TKM_IA has nothing to do with Proto-Iranians who predate him by centuries.
@CopperAxe, saw on Anthrogenica you had a question about whether Dzudzuana input into Euro HG is compatible with what we know of Basal Eurasian levels... Prompted me to try and do a quick plot to explore it: https://imgur.com/a/yAF5fCA
So the Y-axis is the stat of f4(Cameroon_SMG,Ust_Ishim;Kostenki14,X), which should index the level of Basal Eurasian, since affinity to Ust_Ishim is how we define this now that geneflow from East Eurasia confuses using affinity to ENA groups.
While the X-axis is affinity to early West and East Eurasians, Kostenki14 and Tianyuan.
There's probably some noise but judging but that Pinarbasi_HG from Turkey would be estimated to have no more Basal Eurasian than Serbia_IronGates_Mesolithic, Ukraine_Mesolithic or Karelia_HG. While Boncuklu would have only a little more.
However Turkey_N should have quite a bit (although not much more than CHG).
Which is a roundabout way of saying it could possibly depend on whether the Dzudzuana behaves more in these stats more like Pinarbasi_HG and Boncuklu, or more like Turkey_N.
(minimum SNPs for these stats is 400k, but typically more like the full 1000k)
(As an aside it seems like the lower Basal Eurasian like affinity in West Euro HG compared to EHG balances out with greater Basal Eurasian in Turkey_N, such that Globular Amphora, Progress_N and Afanasievo score more or less exactly the same on the Basal Eurasian proxy f4 stat.)
@Copper axe
To be clear it’s a 2000 year old Parthian era sample we believe is mislabeled based on analysis and how much tighter of a clade it forms with Kurds and to a slightly lesser extent other regional Iranic samples than 2700 year old Hasanlu and Dinka Tepe B.
Sample details along with one Kurdish 30X WGS who we have obtained release permission from will be made available so that individuals can easily verify it’s a mainstream Kurd and not outlier upon our publication
@Vara
No one is contesting contacts with Assyrians. In fact based on what we see Dinka Tepe A maybe the result of such contacts. No one is saying that Iranics such as Kurds don’t have an Iran-Chl related base. TKM-IA related clearly has some contribution on Iranics as well as the later Parthians on different levels
@Dilawer
I agree with Vara on this. Iranian entry into NW Iran earliest by 1500bce, latest by 1000bce has been archaeological consensus for decades. A lack of R1a changing this consensus archaeological opinion is bad science/data fitting. Later dated samples wont affect this.
"When combined with Dinka or Hasanlu it causes P-Values to shoot up in qpAdm and Hasanlu/Dinka values to shoot down to around 50% for Kurds (30X WGS) while at the same maintaining decent standard errors. This is consistent with linguistics, archeology and history where Parthians form a very substantial layer on top of Hasanlu era Iranics in the ethnogenesis of Kurds."
When will this sample be published?
In any case, if what you say is correct that Kurds are 50% Hasanlu + 50% TKM_IA_new, given that f4(Chimp, Sintashta, Hasanlu_IA, Kurd) is -ve, I dont expect this new sample of yours to have more steppe ancestry than what Hasanlu_IA has, which is about 10% Yamnaya.
Also, the 18 whatever Hasanlu_IA samples are very varied in BMAC ancestry and do not cluster tightly like Sintashta_MLBA. Half of them are like older Hasanlu_MBA sample; ~half are pulled towards BMAC & 1 towards IVC. Given that on PCA the Kurds cluster inside Hasanlu_IA, while modeling Kurds we shouldn't use the Hasanlu_IA label, rather break it up into individuals and test.
Anyway, waiting for this sample of yours to come out.
@All
What can we say about steppe ancestry in Kura-Araxes versus what Lazaridis et al. said on the topic?
Can’t recall what Harvard claimed, but there’s some residual and patchy Steppe eneolithic ancestry in KA individuals
@Davidski
It’s very clear based on what I’m seeing that Kura-Araxes-EBA and IRQ-Nemrik-LBA are significantly more related to EHG than Iran-NW-Chl/IA are to EHG. What’s not clear at the moment what causes this relationship to be closer. Kura-Araxes is very important to present day Armenians not so to Iranics such as Kurds based on what I see. We’re diving alot deeper at the moment based on the new Southern Arc samples and should have more to say on this
@Vasitha
The sample is buried in the Southern Arc dataset. Kurds can’t be modeled as 50% Hasalu+50% TKM-IA by any stretch. However they can be modeled as 100% this sample which we strongly suspect to be Parthian or 50/50 Hasanlu/Parthian even with 15 pright diverse W/C Asian and European outgroups. Best p-values and SE though are achieved with DinkaB/Parthian
Edit: Meant DinkaA/Parthian suspect at least for Iraqi Kurds, who although are much more related to DinkaB than DinkaA overall but do prefer the added Mesopotamian admixture DinkaA has over DinkaB in models consisting of Iran-NW/IA + Parthian
@Assuwatama
Either you were deprived of oxygen at birth or just a full retard. The Mitanni related samples is not just Alalakh_outlier but those at Megiddo , one of whom was R1a and the other older individual has Steppe MLBA and WSHG ancestry each having close to 25-30% Steppe MLBA. The late BMAC only became Indo Iranian because they were taken over by essentially horse people from the Kazakh Steppe who integrate with them and diffuse in a checkered manner southwards into the Iranian plateau and into South Asia. IF you and the other village idiot Vashista ,think Indo Aryans came from India, you both better book your therapy sessions now. One of the Hasanlu samples 4097 , is essentially teleported from Central Asia and he has definitive Andronovo/Fedorovo ancestry.
Pre-Median Indo-Iranian Tribes in Northern Iran
I. M. DIAKONOFF
left pops:
IRN_Hasanlu_LBA_4097
UZB_Sappali_Tepe_BA 0.839
KAZ_Shoindykol_MLBA_Fedorovo 0.161
tail: 0.857
model with Yamnaya
left pops:
IRN_Hasanlu_LBA_4097
RUS_Yamnaya_Samara_EBA 0.16
UZB_Sappali_Tepe_BA 0.84
tail: 0.208
His closest relative is TUK001 , another Central Asian .
@ Dilawar
That would make sense Parthians were NW Iranic speakers like Kurds.
@Davidski,
It is not patchy.
Most Kura Araxes samples have about 10% Steppe admix. It is uniform. Looks like they were a stable population with Steppe admix.
@Samuel
Most Kura Araxes samples have about 10% Steppe admix. It is uniform. Looks like they were a stable population with Steppe admix.
And Lazaridis didn't find any steppe ancestry in KA?
@Assuwatama ,
Mitanni are not Iranic, which David already mentioned, they are distinctively early Indo Aryans before the split with Dards and Vedic Indo Aryans , Medes ie NW Iranics arrived later, from the same region which Parthians habited between the Caspian and Kopet Dag and consistent with what Dilawer is saying wrt to Kurds. Its also noteworthy that Hasanlu 4097 sample works exceedingly well with Fedorovo which is tied with Indo Aryans of the Steppe, he is clearly part of the Mitanni expansion from Central into West Asia.
@All
I'm hoping that historical linguists aren't fooled by the Lazaridis, Alpaslan-Roodenberg et al. paper.
But it seems that James Mallory has already made a fool of himself thanks to it.
Your typical Indo-Aryans of swat valley from 1300-0bce :)
Aligrama_IA
970-550bce
G2a2a, R2a3a, L1a
Barikot_IA
1000-800bce
U2e1, H2Oa, J1b1b, U8b1a1, M65a
H1a1
Butkara_IA
200-0bce
M30b, U2a, HV
J1
Katelai_IA
1000-800bce
U4d, J1d, M35b, Z31a1
J2a1, R2a
Leobanr
1300-1000bce
R30b1
L1a
1000-800bce
W3a1b, U2e, M4, M5, U7a, T2g1, U2c1, U3b
L1a x4, R1b, R2a, C1b, Q1b2
Saidu Sharif
500-300bce
H, H15a, M30, K1b1, M52a, C4a, H13a, R5a2, U2b2
R1a1a1b, L1a, Q1b2, A
Udegram_IA
1200-800bce
M65a1, T2a1b, U8b, R30a1b, U7a, U1a, H2a2, M, U4c, W3a1, U1a1, H14a,
E1a, E1b x8, A0T, CT, H1a1a
@Assuwatma
When more relevant Iron Age genomes from South Central Asia come out , you and those nutjob nationalist trolls will be eating crow. What I find hilarious is that Dravidian-speaking Indians actually understand everything and are more based and can see the delusions of grandeur people with your retarded viewpoint hold.
The fact Bayesian analysis of the RV dates to 1200-1100 BC and Steppe admixture enters high Steppe groups like Brahmins during that time , is not mere coincidence its common sense. Also using SPGT uniparental markers and plastering and spamming posts does not change reality.
What your not posting is how different many of these modern PNWI (Pakistani and NW Indian ) populations are and they are already all forming a strong cline with very Steppe rich Central Asian populations from the Iron Age.
Here Model for us Kashmiri Brahmins
left pops:
Kashmiri_Pandit.DG
Iran_ShahrISokhta_BA2 0.355
Kazakhstan_Kangju.SG 0.433
Irula.DG 0.211
p-value 0.21
Model for Kangju
Target: KAZ_Kangju:DA206
Distance: 1.8255% / 0.01825521 | R3P
52.0 RUS_Sintashta_MLBA
41.4 TKM_Gonur1_BA
6.6 RUS_Shamanka_EBA
Roughly ~22%. Steppe ancestry.
History!
Not surprising to us since a lot a steppe rich tribes migrated into India 200bce-500ce. Scythians, Kushans, Parthians etc
Today Swat is 90% Pashtoon who are a lot more steppe and overwhelmingly R1a.
Another model for us Kashmiri Brahmins
Kangju 35%
Swat_nosteppe 45%
Irula 20%
p-value 0.05
Brahmins carry a lot less steppe than many NW groups like Jats.
@Davidski Hi David. I analyze modern samples between G25 and qpAdm and it greatly speeds up my work. However, there seem to be some samples that are not in the G25 spreadsheet, either due to being outliers or you maintaining the list as to not get cluttered. Some of the samples also have different master IDs than what's on the official aadr list, so I am unable to cross reference them. Is possible for you to convert these samples to G25 for my own personal use? I've already put them in plink format so you don't have go around compiling them:
https://drive.google.com/file/d/1irdtSokJkyQos4ahvFv7nyycPbsvxJHb/view?usp=sharing
@Assuwatama you really are a lost cause, none of those groups you mention carry L657 and by the 1th century AD, South Asia has like 50-60 million people its not going to make a dent if some new nomads show up. Also who models PNIW like groups with Irula and tribals???
Here are a group of SI Brahmins
left pops:
Nambudiri
ITU.SG
Turkmenistan_IA.SG
best coefficients: 0.774 0.226
tail: 0.9738
Here with a Steppe rich Gujarat Brahmin (25-27% Steppe MLBA) , the Xinjiang sample from 700 BC is preferred , who is Z2123 which is easily found in modern Indian populations. Steppe admixture dates for these people is during this time frame not in the Classical period with the likes of Sakas or Kushans, neither of which of works as sources.
left pops:
Gujarat_Brahmin.SG
ITU.SG
ABST_IA3
best coefficients: 0.470 0.530
tail: 0.462
Just looking at these huge proportions , it is quite obvious a significant migration took place and explains why Indo Aryans could rapidly eradicate the local post Harappan languages and replace it with Sanskrit based ones in a short spate of time. Food for thought!
@Istakhr
I'll have a look at that dataset in the next few days, but I might not be able to get too many of those samples into the G25 for different reasons.
You don't have scythian samples from India. Scythians in good number settled in sakastan West of Indus. Kushans too left Xinjiang and end up in Bactria before final ingress into India.
They weren't 10-20 people that came and ruled India for years. They came in good number.
Andronovians also weren't l657.
Random 800ce samples
I8726 31%
Sintashta 34%
Irula 34%
p-value 0.8420
Y haplo - J2b2b
I8726 37%
Sintashta 24%
Irula 29%
p-value 0.8163
Y haplo J2b2b
I8726 32%
Sintashta 22%
Irula 46%
p-value 0.0375
Y haplo E1b1b
R1a1a 0.2905 100% irula
R1a1a1b2a1a1a1f~ 0.1593 100% irula
Average steppe ancestry in India is 12%
It's only in the NW which recieved multiple waves that we see on average 25% steppe.
There is insignificant steppe in Indo-Aryan speakers in Eastern India where we attest the 1st Indo-Aryan language and In Gandhara where it is female mediated 15% steppe.
Even the central Indians and Indo-Aryans in Sri Lanka and Maldives have insignificant steppe.
Here models of David :)
Burusho
IRN_Shahr_I_Sokhta_BA3 0.710±0.029
KAZ_Botai 0.048±0.024
RUS_Sintashta_MLBA 0.242±0.028
chisq 14.509
tail prob 0.206105
Full output
Burusho
IRN_Shahr_I_Sokhta_BA3 0.718±0.028
RUS_Sintashta_MLBA 0.243±0.028
RUS_West_Siberia_N 0.039±0.023
chisq 14.182
tail prob 0.223101
Full output
Don't speak Indo European.
@dilawer
"The sample is buried in the Southern Arc dataset."
Which sample id?
Correction
Not J2b2b but J2b2a 🙏
@Blue Caviar
"IF you and the other village idiot Vashista ,think Indo Aryans came from India, you both better book your therapy sessions now. One of the Hasanlu samples 4097 , is essentially teleported from Central Asia and he has definitive Andronovo/Fedorovo ancestry.
left pops:
IRN_Hasanlu_LBA_4097
UZB_Sappali_Tepe_BA 0.839
KAZ_Shoindykol_MLBA_Fedorovo 0.161
tail: 0.857"
Here are all 65 2-source models for Hasanlu_LBA_A.
Any dipshit like you can make some biased qpAdm models without checking for better sources.
Any Sintashta/Fedorovo Related sources fail. The steppe required is from Caucasus, labels like Arm_LBA/MBA, as is also concluded by Lazaridis and proved by R1b presence.
Here are multiple models, all compared with rotating sources/references.
Best models for Hasanlu_LBA_A:I4097 ~1350BCE
P-value Source 1 Source 2
0.597183 Tajikistan_C_Sarazm IRN_Hasanlu_LBA_B
0.233186 Tajikistan_C_Sarazm IRN_Hasanlu_MBA
0.187023 Tajikistan_C_Sarazm IRN_DinkhaTepe_BIA_A
0.140938 Tajikistan_C_Sarazm ARM_Lchashen_LBA
0.0918024 ARM_Lchashen_LBA Turkmenistan_Gonur_BA_1
0.0666421 Iran_BA1_ShahrISokhta ARM_Lchashen_LBA
Bad Models:
0.0188253 Uzbekistan_SappaliTepe_BA Kazakhstan_Shoendykol_MLBA_Fedorovo
The Alalakh well lady also has similar ancestry, no sintashta related ancestry.
the 165 2-source models for here are found here
@Blue Caviar
Thank you for acknowledging that The Hasanlu LBA_A sample was Iranian. But it seems you have shot your own foot because of your stupidity. It is clear that NW Iran had archaeological material from its north as well as east, corresponding best with the Sarazm/BMAC/ShahrSokhta + Arm_MBA/LBA/Catacomb model.
Here's Kuz'mina final take under the heading "Genesis of West Iranians"
"Thus the archaeological materials of Iran analyzed by M. N. Pogrebova and myself correspond well with the historical data on the migration of the Iranian speaking people to Iran. The contradiction between the hypothesis of E. A. Grantovsky about the Caucasian route and M. D’yakonov’s opinion on the migration through Central Asia is thus eliminated: it is obvious that both routes were used. The beginning of the migration may be believed to have happened in the last quarter of the 2nd millennium BC."
Now unless you want to argue that Iranians came from Armenia, you are toast.
@Assuwatama,
Those Roopkund results are crap models jacked from some noob's twitter account. They are completely nonsensical. Instead of always copying and pasting , take a moment to realize if your going to post something today which will make you look a bigger dunce than you already are.
Burusho have far more Tibetic related ancestry than that and remained isolated and inherited through slow protracted mixing with Pamiri, Tibetic and Dardic related populations adjacent to them, which would further mean they need significant Turan related ancestry.
Burusho
South Indian 23%
Baloch 41%
Caucasian 12%
NE Euro 10%
NE Asian 8%
Siberian 3%
Burusho people can be modeled as 15% Kalatasi_IA + 85% Swat_IA (SPGT)
Kalatasi_IA has the following rough ancestry profile -
Target: Kalatasi_IA:C1706
Distance: 2.7469% / 0.02746943
53.8 MNG_Slab_Grave_EIA_1
26.6 RUS_Sintashta_MLBA
11.0 UZB_Bustan_BA
8.6 MNG_Xiongnu_East_Asian
Target: Kalatasi_IA:C1705
Distance: 2.7699% / 0.02769858
49.4 MNG_Slab_Grave_EIA_1
28.0 RUS_Sintashta_MLBA
16.0 MNG_Xiongnu_East_Asian
6.6 UZB_Bustan_BA
17 million bhills speak Indo-Aryan language
Harappa world
South Indian 67%
Baloch 24%
Caucasian 0%
NE Euro 2%
Guess that will translate to insignificant steppe ancestry.
Bengali
226 million native speakers
South Indian 47%
Baloch 27%
Caucasian 2%
NE Euro 4%
SE Asian 7%
NE Asian 6%
@VAsiSTha I've been reading your blogs since a while. I'm trying to do rotating qpAdm rotation. Is that possible in linux version of Admixtools?
@ your just being a troll now , again piss poor models, there is no local ancetry, he is not a local.
Its obvious from the distances , interesting he is very similar to TUK001 which is clear from the qpAdm result I posted, as well as the pca. Your right pops are all a joke as well.
MNG TUK001 Average
3.27
Sappali Tepe BA Average
5.07
Alalakh MLBA o ALA019
5.61
@Vashishta
The village idiot talking about other's stupidity is rich. I didn't bother reading your post because its literally the battle of the dunce cap between you and Assuwtma, I guess he will be runner up this year. The very fact you think Sanskrit came from India, shows the level of idiocy. You can fool your countrymen with your call center scams but good thing is many Indians do see past the kind of BS and disinformation your spewing. You should also invest in a world map. The BMAC is not in the South Caucasus and the impetus largely driving movements to the Iranian plateau in that time , were Steppe Indo Aryans , if you can familiarize yourself with PCA you will see this sample sits with TUK001.
If you know Iranian history , Medes and NW Iranics do not arrive in Urmia until the Iron Age. Any Iranian or Kurd will laugh at the BS your spewing. I am stunned David even allows such trolls, but keep reaching , you still won't get anywhere. Like a 2 year old trying to open a top kitchen cupboard but never can.
@ Assuwatma , jacking off Vashishta's doctored rotating qpAdm models does not say much, we know his bogus qgraphs, the very fact he modeled what clearly is a Central Asian with Bronze Age Armenians shows his epic desperation. Using Sarazm which is a distal population from the late Neolithic, Jesus. As aptly Rob mentioned said you don't know anything about using populations. Vashishta can fool his legion of nationalist incels since he is one himself and he should stick to his day job , if he has one that is.
@Vashishta/Assuwatma
The rotating model is shit btw, you think Neolithic Central Asians admixed with LBA Armenians?? Seriously?? What you showed was a doctored distal model basically. Truly pathetic, but I am not surprised at all you stoop that low.
Also if you think Iranics came from the South Caucasus you definitely had a lobotomy at some point. Just like Indo Aryans came from Goojrat right? LMAO
@Dilawer
What are the proportions of the Dinkha_A + Parthian model for Kurds on qpAdm? Also, is this sample similar to TUK001?
@Blue Caviar “Burusho have far more Tibetic related ancestry than that and remained isolated and inherited through slow protracted mixing with Pamiri, Tibetic and Dardic related populations adjacent to them, which would further mean they need significant Turan related ancestry.”
It will be interesting to find out where the Kusunda and Nihali came from.
It’s funny how Assuwatama consistently crops out the 2x I2aL68 from Swat samples
That’s 4 lineages (Ie fairly straightforward evidence) of a link to Eastern Europe
As for models, maybe he should start using Vindaja Neanderthals as well , to prove his point
@Vashista ,
Looking at your models it shows what a scam artist you truly are. We all know you doctor your results but your sleight of hand approach ( your qgraphs was the center piece for that) have gaping flaws because they are based of nonsense and a steaming pile of horse shit. The scary part is you have a contingent of incels in your country lapping this shit up, you should be ashamed of yourself.
Here is a model with Hajji Firuz BA with an even higher tail than your model, but its still nonsensical given this individual is primarily BMAC derived and does not match with the archaeology and history of the region, a subject you probably failed growing up.
https://pastebin.com/6J25KEzY
it still does not best the Steppe MLBA based model.
https://pastebin.com/gGXG3vME
Even on the PCA it sits on a cline with other Steppe MLBA enriched Central Asians.
https://i.imgur.com/WC7LFCF.png
Looking at the distal its clear 4097 some Central Asian immigrant with Steppe MLBA ancestry.
sample: Hasanlu_4097
distance: 3.5677
Ganj_Dareh_N: 51
Barcin_N: 21.5
GEO_CHG: 14
Volga-Kama_N: 8
Vologda_Veretye_Meso: 5
LAO_Hoabinhian: 0.5
sample: MNG TUK001:Average
distance: 3.7784
Ganj_Dareh_N: 49
Barcin_N: 23
GEO_CHG: 10
Volga-Kama_N: 12
Vologda_Veretye_Meso: 4
LAO_Hoabinhian: 0
Devils_Gate_Cave_N: 2
Its clear from the distances as well
Distance to: IRN_Hasanlu_LBA_A:I4097_all___J-ZS6637___BC_1354___Coverage_86.75%
0.03265890 MNG_TUK001:TUK001
0.05019587 IRN_DinkhaTepe_BIA_B:I4340___female___BC_1341___Coverage_60.29%
0.05615369 TUR_Alalakh_MLBA_o:ALA019
0.05627587 IRN_DinkhaTepe_BIA_B:I4238_all___J-Z7515___BC_1162___Coverage_81.51%
0.05762061 IRN_Hasanlu_IA:I6430___R-Y23838___BC_900___Coverage_74.99%
0.05834498 IRN_DinkhaTepe_BIA_B:I3914_all___female___BC_1078___Coverage_85.39%
0.06053626 IRN_Hasanlu_IA:I6432___female___BC_1178___Coverage_74.75%
0.06117189 IRN_Hasanlu_IA:I6428___J-BY40968___BC_1027___Coverage_77.02%
0.06220058 IRN_Hasanlu_IA:I4269_all___R-Y88647___BC_803___Coverage_84.96%
0.06470890 IRN_DinkhaTepe_BIA_B:I3915_all___G-CTS11562___BC_927___Coverage_85.08%
0.06715505 IRN_DinkhaTepe_BIA_B:I3912_all___G-Z3353___BC_1793___Coverage_84.84%
0.06734644 IRN_Hasanlu_IA:I4232_all___R-M12149___BC_874___Coverage_82.58%
0.06759286 IRN_DinkhaTepe_BIA_B:I4273_all___female___BC_1017___Coverage_78.33%
0.07167512 IRN_Hasanlu_IA:I6431___R-FGC14590___BC_1072___Coverage_44.79%
0.07494892 IRN_DinkhaTepe_BIA_B:I4236___female___BC_1500___Coverage_74.37%
0.07847426 IRN_Hasanlu_IA:I4357___female___BC_871___Coverage_67.43%
0.08817144 IRN_Hasanlu_IA:I4280_all___R-FGC14590___BC_787___Coverage_80.40%
0.08817453 IRN_Hasanlu_IA:I6388___R-Y19434___BC_1065___Coverage_66.92%
0.09095997 IRN_Hasanlu_IA:I4353___J-PF5252___BC_961___Coverage_65.91%
0.09327439 IRN_Hasanlu_IA:I4100_all___female___BC_961___Coverage_88.31%
0.09495759 IRN_Hasanlu_IA:I4355___R-M12149___BC_948___Coverage_51.88%
0.09741795 IRN_Hasanlu_IA:I4099_all___female___BC_1201___Coverage_86.35%
0.10288792 IRN_Hasanlu_IA:I4233_all___L-Y28524___BC_856___Coverage_84.22%
0.10311437 IRN_Hasanlu_IA:I4356___R-Y4364___BC_1017___Coverage_65.63%
sample: MNG TUK001:Average
distance: 1.9903
Sappali_Tepe_BA: 73.5
Sintashta_MLBA: 23.5
Slab_Grave_EIA_1: 3
sample: Hasanlu_4097
distance: 1.3971
Sappali_Tepe_BA: 76
Sintashta_MLBA: 20.5
CG_IVCp: 3.5
They are pretty similar, difference is 4097, has as slight South Asian shift which is lacking in TUK. I am sure you will spin this too down the road. Maybe he had a sojourn in Goojrat?
I strongly suggest you find a day job if possible instead of constantly trolling David, Rob , and others who offer a lot of insight, or better yet crawl back into the dumpster you came out of and don't forget to close the lid .
Congrats you got dissed and dismissed today!
@ Davidski
What did Mallory write.
Apparently he did a presentation claiming that Armenia was the Indo-Anatolian homeland based on the latest, super duper ancient DNA evidence.
But I don't know the details.
I almost don’t like critiquing these issues because it might come across as if we have something against Iran or Armenia. But it’s always been about methodology
In some cases they completely overestimate steppe migrations in others they miss it because they’re expecting 90% Yamnaya R1a or R1b, but forget about unsampled preYamnaya groups or balkan related stuff
There’s no consistency, because of this form of ‘distal statistical pseudo-anthropology’ which seems to have taken hold amongst most teams
There's definitely a problem with the interpretation of the raw numbers, some of which may not even be very accurate.
I think that by and large these scientists have been staring at their little pie charts of autosomal admixture proportions for way too long.
So much so, that they're imagining things that never happened (like waves and waves of migrations into the steppe from the Near East) and ignoring the details.
btw check this from Mattila et al paper (SredniStog, etc)
Rom_047
From Decea Mureşului. This cemetery is considered a local variant of the Suvorovo-Skelja horizon. I.e. early Sredni Stog.
Usual grave goods: ochre coverings, unio shells, copper beads, flint knives, maceheads. In other words - this would look typical in Khvalynsk
But this guy was a ''Farmer''; although I cant make out detail in PCA, he plots with Eneolithic Balkan people, and is Y-hg H2 (although citizen Y-callers should double check that)
Against that, you have that unbpublished individual from nearby Csongrad who was Y-hg Q !
So this all just proves what local specialists have long pointed out - this horizon formed due to farmer-forager interaction which extended deep into the Balkans, which is the start of the steppe. The reason why Farmer admixture did not make it out to the Volga was because the interaction occurred in the Carpatho-north Balkan region, not out in the Volga, and ideas were imported back by the forager-come-pastoralists themselves who were undertaking cyclical migration. Of course, some teams have no clue about this (not referring to the Swedes, of course), and even with data they'd ignore or not notice it. Hence need to invent 'mystery invaders from Asia' to account for Yamnaya popping out of the ground suddenly.
One minor critique on the paper is that theyre too quick to claim continuity in the DNieper. This data shows the same thing as Mathieson, slight WHG shift in the Meso - Neolithic transition, again supported by shift from R1a, and now a Q1a, to I2a & R1b-V88. This was the first phase, or pre-Sredni Stog.
@Davidski
Thanks, really appreciate the time. Just to clarify, I'm just looking to study them personally (even the outliers), but if you want to add any of them to spreadsheet, by all means feel free to do so.
@Blue Caviar
Seethe more lol. You can throw as many racist and personal remarks, i dont give a shit, Im used to that here.
'Oh your right pops are shit because they don't pass the model I like best' - Lmao, what a fool.
Lazaridis and friends simply have an agenda that IE must originate outside Europe. They will jump through all and any required hoops to come up with a supporting theory.
@ Vasistha
There issue is you make statements like
- ''there's no Andronovo related ancestry in Turan, because Dali_MLBA is a better fit''... but we know that Dali_MLBA is ~ 85% Andronovo derived
- ''haha fools there's only 4% Yamnaya-Volga ancestry in Anatolia''.. although Yamnaya didnt come from Samara, and neither did the Hitto-Luwians. You have not been gracious enough to admit you're wrong because there clearly was a Balkan migration into Anatolia, and of a fairly substantive character.
- "Y-DNA doesnt correlate with language'' but ''Haplogroup J2 attests to the expansion Indo-Iranian groups into Europe''
These sorts of exercises might work for the people on your blog, who simply want to hear that 'cultured southerners'' civilized barbarians in Europe,. Heck it might even land you a professorship at Harvard. But it doesn't work well amongst Free Thinkers. You need to be honest, and people will respond to you better
@Rob
“These sorts of exercises might work for the people on your blog, who simply want to hear that 'cultured southerners'' civilized barbarians in Europe,. Heck it might even land you a professorship at Harvard. But it doesn't work well amongst Free Thinkers.”
Exactly, they cannot stand that the cradle of our civilization was in Eastern Europe and not in India or the Middle East as it was believed in the XIX century.
I don't really see the reason for such angry dispute about Hasanlu_LBA_A.
If you model this single sample with local ancestry, it's obvious the best model will be using Sarazm reference. That the shift is most in this direction is even obvious in the distance difference in G25 - https://imgur.com/a/evQVxhy
There isn't enough data in Central Asia to determine with a Sarazm rich people still persisted at this time.
Is there any reason in isotope data or archaeological description or y or mtdna haplogroups to prefer one scenario or the other? If so, just say so and there's no need to shout like a 4channer (which is likely to impress or intimidate other 4channers only, who are usually pot-calling-the-kettle for virtually anything they might say).
I can see nothing in the supplementary material. No extended archaeological description for this sample, and no obviously unusual haplogroups; y dna is J1a2a1b1a~ while mtdna of R2 does point to Central Asia but is not going to greatly influence whether the individual is admixed between either source.
Although such a source need not be unadmixed Sarazm_CA: https://imgur.com/a/dmCxoDb . Although this is just a G25 model, it finds that with some local ancestry, some Central Asian mix with low levels of Steppe_MLBA related ancestry and some levels of Sarazm_EN/CA like ancestry and some post-BMAC like ancestry seems to work well. It need not be one of the other of either having no ancestry from Sarazm_CA or all of it. We're just limited in the references from Central Asia we have at the moment, and it would be no surprise if Steppe_MLBA+BMAC/post-BMAC like ancestry didn't really work without extra contributions from Central Asian Eneolithic/CA populations from further north/east.
Would also note that in Davidski's G25 data via Vahaduo, whilst a 2 way model of Sappali_Tepe and Federovo might superficially work in the West Eurasia PCA (https://i.imgur.com/wwD3vhX.png) it has obviously slight problems when look at the North Eurasia PCA that is more specific to Central Asia (https://i.imgur.com/2T4jhfV.png).
We obviously cannot hang any revision of Indo-Iranian origins on "Oh, there is this one sample that is at this site, which can possibly be modelled multiple ways", particularly one with no obvious impact on later genetics, which are simply largely continuous with the MBA.
@Rob
"There issue is you make statements like
''there's no Andronovo related ancestry in Turan, because Dali_MLBA is a better fit''... but we know that Dali_MLBA is ~ 85% Andronovo derived"
This is a lie. I very explicitly have claimed that Dali is Sintashta derived, but have made the distinction that Dali is quite far from Sintashta and it is not obvious or apparent that the vehicle culture at Sintashta is present at Dali. I do not deny presence of Sintashta derived steppe_MLBA ancestry in Indians or Swat.
"- "Y-DNA doesnt correlate with language'' but ''Haplogroup J2 attests to the expansion Indo-Iranian groups into Europe''
Another lie. I have only said that Y-dna need not correlate with language. I have not claimed that J2 attests to expansion of indo-iranians into europe. On being pressed about alleged absence of iranian uniparental markers in steppe even though autosomal ancestry is clearly present, I have only said that presence of J markers in karelia, khvalynsk etc attests to migration from the south. I never claimed that presence of J in itself proves male mediated IE migration to steppe. For me, autosomal ancestry >>> Y markers.
"although Yamnaya didn't come from Samara, and neither did the Hitto-Luwians. You have not been gracious enough to admit you're wrong because there clearly was a Balkan migration into Anatolia, and of a fairly substantive character."
Don't care about Balkan migrations because they did not bring Anatolian languages, eastern ancestry into Anatolia did. As for origin of Yamnaya related ancestry from Samara or Sredni Stog, don't really care about it for now. new data will clarify this question soon, I don't need to spend my time on it prematurely.
@Matt
There is archaeological data, maybe not from this paper, but from NW Iran in general. It is clear that it has material from both Caucasus as well as from east in early - late IA (Ref. Kuz'mina, Grantovsky, Pogrebova). This is entirely consistent with many R1b's (from Caucasus) found at Hasanlu_IA, older Hasanlu_IA.SG etc. as well as BMAC related autosomal ancestry and minor steppe related ancestry.
The same is concluded correctly by Lazaridis et al, although they go way off mark in interpreting this data to mean non Iranians at Hasanlu.
Wrt G25 models, we always have to be wary of overfitting, and qpAdm should be relied on. qpAdm consistently favours 2 source models with Sarazm, and even Shahr-Sokhta/BMAC along with either local Hasanlu MBA ancestry or with Armenia MBA/LBA.
The reason Sarazm is selected consistently is because it has excess ANE ancestry over Bmac as well as excess CHG ancestry (and maybe something else which we haven't figure out yet) which gives it affinity to Kumsay_EBA and yamnaya related pops as well.
"y dna is J1a2a1b1a~ while mtdna of R2 does point to Central Asia but is not going to greatly influence whether the individual is admixed between either source."
As for J1a2a1b1a, there are 4 samples from Geoksyur and 1 from Gonur with this subclade. R2 is found at IranNTepeAbdulHosein as well as in Gonur and Gonur IVC outlier.
"Oh, there is this one sample that is at this site, which can possibly be modelled multiple ways", particularly one with no obvious impact on later genetics, which are simply largely continuous with the MBA."
Completely disagree here. LBA_A sample is quite important because the best model in qpAdm for Hasanlu_IA = 25% Hasanlu_LBA_A + 75% Hasanlu_LBA_B.
I also dont agree with your conclusion that later genetics is continuous with MBA. Sure, half the samples are like MBA, but other half are not, and are quite pulled towards BMAC.
G25 based on Pop averages, scaled.
Target: IRN_Hasanlu_IA_I6428:I6428
Distance: 2.4035% / 0.02403475
32.4 IRN_Hasanlu_LBA_B
29.4 IRN_Hasanlu_LBA_A
21.4 IRN_Hasanlu_MBA
13.0 ARM_Lchashen_MBA
3.8 IRN_Shahr_I_Sokhta_BA2
0.0 KAZ_Dali_MLBA
0.0 RUS_Kubano-Tersk
0.0 RUS_Kubano-Tersk_Late
0.0 TJK_Dashti_Kozy_BA
0.0 TJK_Sarazm_En
0.0 UZB_Sappali_Tepe_BA
I6430 being pulled towards IVC.
Target: IRN_Hasanlu_IA_I6430:I6430
Distance: 1.7894% / 0.01789448
36.4 IRN_Hasanlu_LBA_B
22.4 IRN_Hasanlu_MBA
17.8 ARM_Lchashen_MBA
13.6 IRN_Shahr_I_Sokhta_BA2
5.6 UZB_Sappali_Tepe_BA
4.2 IRN_Hasanlu_LBA_A
0.0 KAZ_Dali_MLBA
0.0 RUS_Kubano-Tersk
0.0 RUS_Kubano-Tersk_Late
0.0 TJK_Dashti_Kozy_BA
0.0 TJK_Sarazm_En
Caveat: G25 models are only indicatory and may be overfitted. Best qpAdm model for I6430 is simply 79% Hasanlu_MBA + 21% IVCp.
@ Vasistha
They obviously did ;)
@Matt
I forgot to add. "There isn't enough data in Central Asia to determine with a Sarazm rich people still persisted at this time."
Completely agree with this. Someone contacted me for modeling their ancestry, they were a Punjabi Khatri-KashmiriPandit mix. I could not model them at first at all, and only after looking at gendstats to figure out what was wrong, I had to add Sarazm to the sources and then it worked. Here are their models.
Even modern Kalash need Sarazm like ancestry. Eg see
Target: Kalash
Distance: 1.3573% / 0.01357286
30.2 TJK_Sarazm_En
17.4 IRN_Shahr_I_Sokhta_BA2
16.0 KAZ_Dali_MLBA
11.6 Irula
10.0 UZB_Sappali_Tepe_BA
9.6 KAZ_Kangju
5.2 RUS_Sintashta_MLBA
0.0 ARM_Lchashen_MBA
0.0 IRN_Hasanlu_LBA_B
0.0 IRN_Hasanlu_MBA
0.0 KGZ_Aigyrzhal_BA
0.0 RUS_Kubano-Tersk
0.0 RUS_Kubano-Tersk_Late
0.0 TJK_Dashti_Kozy_BA
(Scaled averages for all labels).
@vAsiSTha: OK I can see why you prefer the 2-source models for parsimony, but I really won't be surprised if we end up with some sampled populations from TJK with Sarazm+Steppe_MLBA admix and they offer a comparable or better degree of fitting here. I do agree that the modelling of the LBA_A as a Steppe_MLBA population + the BMAC/post-BMAC from UZB or TKM seems like it has problems from the point of view of West_Siberia_HG/Tarim_EMBA related ancestry.
"As for J1a2a1b1a, there are 4 samples from Geoksyur and 1 from Gonur with this subclade. R2 is found at IranNTepeAbdulHosein as well as in Gonur and Gonur IVC outlier".
Interesting, so both mtdna and y may link us into Southern Central Asia?
"Completely disagree here. LBA_A sample is quite important because the best model in qpAdm for Hasanlu_IA = 25% Hasanlu_LBA_A + 75% Hasanlu_LBA_B.
I also dont agree with your conclusion that later genetics is continuous with MBA. Sure, half the samples are like MBA, but other half are not, and are quite pulled towards BMAC."
Well, I mean it clearly seems fair to call it largely continuous in an overall sense if most of those two populations are composed largely of Hasanlu_MBA.
E.g. https://imgur.com/a/xZkYYc0
And the overall shift to IA in a two-way model of Hasanlu_MBA is only around 8% Sarazm_EN (incidentally somewhat comparable in single-digit size to the likely shift in Yamnaya ancestry shown at the Aegean Coast Yassitepe samples).
But yes, there's an argument about definition of continuity there, and you could say that proximally there's discontinuity.
You could argue that with the right chain of proximal populations there could be some turnover.
But like talking about Anatolian and the Balkans, this would seem to me like it requires more sampling than we have, to establish a chain.
Also, in terms of fitting IA, in the case of the MBA sample, and the two LBA samples we're reliant on essentially 3 single samples, two of which are close together and one which most likely has the good deal of its ancestry from a population including the other two, and don't give a basis for really a robust idea of how much variation there was around there that could explain the later patterns.
It seems a bit premature to go effectively "Well, we have three samples and its easier to model these later samples with this two of them... rather than this one of them or this two or this three of them". Here you need to use qpWave and investigate first whether rank=0 of Hasanlu_MBA and Hasanlu_IA is not viable, or Hasanlu_MBA+Hasanlu_LBA_B, but even then it's probably premature on the basis of sample size. It's good to use proximal modelling but if you're using these two or 3 samples where two are relatively derived from the MBA population, I'm not certain that you wouldn't get just as good fits with, like 3 samples from the MBA population or something. And in the G25 version of fitting the same thing, for individual samples it doesn't look like a wave that happens at a specific time or anything like this, just random.
Perhaps using the DinkhaTepe samples to increase the set size could make a more persuasive argument for whatever a wave would have to be? Though we would need to look at the whole set and not arbitrarily select the more Near Eastern subcluster (A).
@Matt
"Well, I mean it clearly seems fair to call it largely continuous in an overall sense if most of those two populations are composed largely of Hasanlu_MBA."
Well this is about semantics then. Swat_IA has avg 15% steppe ancestry, same can be said about that.
"Interesting, so both mtdna and y may link us into Southern Central Asia?"
Yes but mtdna R2 has also been found in Zagros neolithic apart from SC asia. so could be local too.
"Perhaps using the DinkhaTepe samples to increase the set size could make a more persuasive argument for whatever a wave would have to be? Though we would need to look at the whole set and not arbitrarily select the more Near Eastern subcluster (A)."
Indeed, all of Dinkha_B samples are pulled away from Dinkha_A. All with substantial additional BMAC + yamnaya/Armenian ancestry.
qpAdm for Dinkha_B gives 53% Dinkha_A, 10% Yamnaya, 36% Bustan, no other model works. I'm sure we wouldn't call this situation similar to Yassitepe :)
G25 based on averages:
Target: IRN_DinkhaTepe_BIA_B
Distance: 0.8545% / 0.00854536
41.2 IRN_DinkhaTepe_BIA_A
19.2 UZB_Sappali_Tepe_BA
13.8 ARM_Lchashen_MBA
13.0 TJK_Sarazm_En
7.0 IRN_Hasanlu_MBA
3.4 KAZ_Dali_MLBA
2.0 RUS_Kubano-Tersk_Late
0.4 KGZ_Aigyrzhal_BA
0.0 IRN_Shahr_I_Sokhta_BA2
0.0 Irula
0.0 KAZ_Kangju
0.0 RUS_Kubano-Tersk
0.0 RUS_Sintashta_MLBA
0.0 TJK_Dashti_Kozy_BA
G25 averages, reduced to 4 pops
Target: IRN_DinkhaTepe_BIA_B
Distance: 0.9688% / 0.00968753 | R4P
50.8 IRN_DinkhaTepe_BIA_A
27.0 UZB_Sappali_Tepe_BA
14.0 RUS_Kubano-Tersk_Late
8.2 TJK_Sarazm_En
Reduced to 3 pops
Target: IRN_DinkhaTepe_BIA_B
Distance: 1.0195% / 0.01019485 | R3P
47.2 IRN_DinkhaTepe_BIA_A
37.0 UZB_Sappali_Tepe_BA
15.8 RUS_Kubano-Tersk_Late
Almost same as qpAdm.
@Alkabir
All this and much more will be detailed in our upcoming article on our website within 2 weeks comparing and contrasting population histories of Armenians, Kurds, Persians and Turks
The qpAdm results using our suspected Parthian are:
IRQ-Kurds= 73% Parthian + 27% IRN-DinkaA. P-value 0.79 SE 0.10 which is remarkable p-value using all these pright pops which easily discriminate various types of E/C/W Asian ancestries
Since p-value measures the probability of obtaining the observed results, assuming that the null hypothesis is true the higher the p value the better the confidence in the model
right pops:
Mbuti.DG
CHG
IRN_Ganj_Dareh_N
ISR_PPNB
MAR_Taforalt_EpiP
SRB_Iron_Gates_HG
ANE
TUR_Marmara_Barcın_N
WHG
RUS_Yana_RHS
MNG_Khovsgol_LBA
RUS_DevilsCave_N
RUS_Shamanka_EN
Loschbour_WHG
UZB_Sappali_Tepe_BA
RUS_Siberia_Lena_EBA
DinkaB is quite different from DinkaA and the Parthian % goes down to 60%. Same with Hasanlu but the p-values with DinkaA and Hasanlu are slightly lower and SE higher. DinkaA has more Assyrian/Mesopotamian type admixture which Kurds like but overall they are much more similar to DinkaB than DinkaA
"Exactly, they cannot stand that the cradle of our civilization was in Eastern Europe and not in India or the Middle East as it was believed in the XIX century."
Haha ,the cradle of our civilization is in Pelasgo-Minoans who in turn were influenced by Anatolians , Levantines , Egyptians.The closest thing to the "steppe civilization" is something like Islam and the Ottomans.
Davidski said...
"@All
I'm hoping that historical linguists aren't fooled by the Lazaridis, Alpaslan-Roodenberg et al. paper."
Let me sketch you the issue from the perspective of historical linguistics. Because from that perspective it is really clear (in my opinion).
Language always changes. We don't talk like our grandparents. The change in the way we talk may be very small. But over time these changes become big. Just look at Beowulf in its original Old English: "Hwæt. We Gardena in geardagum, þeodcyninga, þrym gefrunon, hu ða æþelingas ellen fremedon.". This is English from 1000 years ago. Do you understand any of it?
And when people don't need to talk much to each other, because they live in different places, their language tends to also become different from the language which is spoken at other places. One language splits into dialects, the dialects grow apart and become different languages.
So Non-Anatolian Indo-European started splitting up around 2700 BCE, right?
And then around 1500 BCE we get the first taste of Indo-Iranian. The Indo-European sounds l and r had merged into r. And the vowels e, o and a all became just a. And there were a lot more sound changes in Indo-Iranian.
Then around 700 BCE we get a taste of Ancient Greek. The old s and w sounds were gone in many of the cases. The old labiovelar sounds (kʷ etc.) had all became something else (t, p or k, etc. depending on the vowel they were next to).
The vowels were still mostly unchanged in the earliest Greek, but that did not last very long. Old ā merged with old ē in Attic, but with lots of exceptions. And later this ē, together with u (later y) merged into i.
When the Romans conquered Gaul around 50BCE they spoke classical Latin. At that time the letter C was always pronounced as k. The letter H was pronounced as h. And V was pronounced as w.
At the end of the empire (around 450CE) C had become something like ts or tsh before e and i, depending on the dialect (Sicilians kept the k sound). V was pronounced as v. And H was no longer pronounced at all.
So what is the difference historical linguists see between the phonemes (the relevant different sounds in a language like vowels and consonants) in Indo-European without Anatolian and the version with Anatolian?
*NOTHING*
But the DNA experts from Reich labs tell us that Anatolian Indo-European and non-Anatolian Indo-European must have split *thousands* of years before non-Anatolian Indo-European split into Greek, Indo-Iranian etc. Thousands of years apart, and still mostly the same language, mutually understandable. That is a beautiful linguistic fairy tale.
Now there are some caveats. It is possible that there were different dialects on the steppe before Indo-Europeans spread far and wide. And that would mean that the formal split-up of Indo-European as *one language* was earlier than 2700 BCE.
In that case Anatolian was just a dialect that left the steppe early. So Anatolian would not get some common innovations that all the other Indo-European dialects still on the steppe adopted. For example, the formation of a female noun-class. All Indo-European branches have male, female and neuter nouns, except Anatolian. Anatolian only had common and neuter nouns.
@Dilawer
Thank you. Bear in mind they do advise against P-value ranking in qpAdm, so just because the pvalue is slightly higher does not mean it is necessarily more correct. If such an individual indeed exists, and the models are true, then the choice should be on the most historically sound model. Dinkha-B is possibly correct too, but I think Hasanlu IA makes most sense. Does this model work for Lors, and do they work as pure Hasanlu iA?
@ Vashista ,
when intellectually cornered play the race or victim card , because you have no rebuttals, thats a pathetic copout but being pathetic and playing victim is your only way to justify your BS. Unfortunately for you, there are people , as Rob put it best who are free thinkers and not beholden to BS theories and tropes which you spew out on your twitter handle. You can scam incels in your country but not with the outside world. You literally live on David's comment section because , even with the sycophants you have on Twitter, its here you get a reality check.
Even the way you model is bizzare because your just hell bent on proving Indo Aryans came from India or somewhere in Goojrat, just like that other moron who was on here few years ago. Nobody takes you seriously and the one or two which do, are just being tactful and polite in the same way a school teacher would be with a kid with down syndrome.
FYI, that 6430 sample works best with Udegram/Loebanr , approx 30-35%, but unfortunately means it has Steppe MBLA ( Heaven forbid).
Your just like this Afrocentrist who claims R1a was from Kushites or this lunatic who believe Scythians and R1a are from India ( I am sure you probably follow him )
https://www.researchgate.net/publication/49582995_The_Kushite_Spread_of_Haplogroup_R1-M173_from_Africa_to_Eurasia
https://www.youtube.com/watch?v=e5GPo_FtwCw
@Assuwatama, vAsiSTha, and CopperAxe
Mr. Dilawer is referring to Kushan sample I12292 from Tajikstan. This sample is nothing new and has been known to exist for years. It is not confirmed to be Parthian or any other ethnicity for that matter. It generally clusters half way between western and eastern Iranics, but that shouldn't necessarily be interpreted as the sample being of Parthian background. Based on some preliminary qpAdm models I ran, it could also be a local Eastern Iranic with mixed Western Iranic or Mesopotamian ancestry. The exact background of this individual has been debated before and will continue to be debated until we get more ancient Iranic samples
I do not know whether the suspected Parthian sample is it but both I12292 and its first degree relative I12260 seem like a mixture of DA382 (TKM_IA) and Dinkha_B-like populations. It does not seem like a singleton; so a Parthian origin is possible considering the difference between them and I12294 (I think it is more likely for Kushans to have considerable East Asian ancestry). Also they are genetically close to TUK001 which might also turn out to be a Parthian sample since we know that there are Parthian textiles discovered in Xiongnu burials.
Kushan era samples from Rabat and Serkharakat should be examined. If they cluster with I12292 and not with I12294 then perhaps it is more likely for it to be a Kushan sample instead of a Parthian one.
@Alkabir
Historically sound models should always be given importance. Wrt the p-values and LRT used in qpAdm, this Is very confusing for most people and very poorly understood that’s why I tried to explain it about 4 years ago at https://eurasiandna.com/impact-of-the-iron-age-saka-and-scythians-on-south-west-asian-demography/.
I remember having a little debate on this with Patterson some years back, however I was glad to see that they subsequently took my position on it at https://www.biorxiv.org/content/10.1101/423079v1.full.pdf where they write:
“Multiple models may fit some populations; we show here the one with the highest P-value”
Briefly, the Null hypothesis in qpAdm is the constrained model in pleft. The P-value is calculated using a likelihood ratio test in which the constrained model is the null hypothesis and the unconstrained model is the alternative hypothesis. By definition likelihood is a probability of the Null hypothesis being true ( modeling the target with the listed sources).
As the P-value becomes smaller, it becomes less likely that the NULL hypothesis is true.If you do this long enough you’ll see that on average the higher p-value models make more sense tgan the lower ones.
Wrt Hasanlu-IA I’m not sure with Lurs but I know with Kurds the Null hypothesis Kurds=100% Hasanlu-IA can always be rejected given a high enough number of diverse references . This said Hasanlu-IA is quite similar to Kurds, more so than to Persians and Armenians. However HajiFiruz-Chl seems equidistant to Armenians and Kurds. With populations younger than 6000 years old they are disimiliarly related
@Blue Caviar
" Unfortunately for you, there are people , as Rob put it best who are free thinkers and not beholden to BS theories..."
Rob, post from your own account. Don't create new accounts (created Sep 2022) just to throw personal insults and for fear of losing face.
You keep whining about steppe_mlba ancestry in hasanlu_LBA_A, well Dinkha tepe_B is from the same period. Where is the steppe_MLBA huh? Its clear that the steppe is from Caucasus, like archaeologists had already predicted and R1b makes it very clear as well.
Target: IRN_DinkhaTepe_BIA_B
Distance: 1.0195% / 0.01019485 | R3P
47.2 IRN_DinkhaTepe_BIA_A
37.0 UZB_Sappali_Tepe_BA
15.8 RUS_Kubano-Tersk_Late
"FYI, that 6430 sample works best with Udegram/Loebanr , approx 30-35%, but unfortunately means it has Steppe MBLA ( Heaven forbid)."
Im sorry to tell you that I would have been ok if that was the case, but it just doesn't work out. Aligrama (low to no steppe) is always selected, both by Vahaduo as well as qpadm.
Target: IRN_Hasanlu_IA_new:I6430
Distance: 1.8034% / 0.01803353
36.2 IRN_Hasanlu_LBA_B
32.4 IRN_Hasanlu_MBA:I4354
16.2 IRN_Hasanlu_LBA_A
15.2 PAK_Aligrama_IA
0.0 PAK_Udegram_IA
0.0 IRN_Shahr_I_Sokhta_BA2
0.0 PAK_Loebanr_IA
(scaled averages)
The same is confirmed by rotating models.. No excess steppe_MLBA ancestry is needed over Aligrama_IA or IVCp.
@istakhr
"Mr. Dilawer is referring to Kushan sample I12292 from Tajikstan. "
Thanks, I was wondering which sample he was talking about. No SC asian sample was published by Lazaridis. I checked on vahaduo, 2 of 11 Kurdish samples show 30-35% from I2292, others dont. So some later admixture from them is indeed possible.
@ vasistha
I’m not blue caviar . You must live in a your OiT blog bubble because you seem to have missed the last few threads with multiple people have pointed you’re out of your depth and are a fool
I personally haven’t looked into Hassanlu IA but do t see the issue given that there’s steppe ancestry in NW Iran already 2500 Bc; and I already made a point about catacomb on this very thread
So you’re confused again
I suppose you think R1b is from India or Southeast Asia
A sample being labelled "Kushan" does not mean the sample itself was an ethnic Kushan, it just means it was from the period of Kushan rule. Many cases like this.
Anyways one of them is a two way mix of west asia and southern central asia, the other is a mix of south central asia and the steppes. Probably more Saka than Yuezhi but who knows.
Aa for calling samples Parthian I think you'd need a good amount of genomes from Parthia before making that conclusion. Central Asia was the recipient of Persian related geneflow for centuries so calling every mixed sample a "Parthian" is way too trigger happy. Many of the medieval Sogdian samples from Mongolia show west asian geneflow and a few centuries ago this must've involved ancestors who were mixed 1-to-1. TUK001 was a random tooth of some woman without context during the Xainbei-Rouran period, and could have originated from a myriad of places.
@ vasistha
Sorry to burst your bubble but many people think that you’re a clown, so I’m not blue caviar.
Given that there’s steppe ancestry in NW Iran in the third mill., your debates about Hassanlu are irrelevant
Moreover, I don’t get how you managed with to get yourself confused (once again) when I made reference to catacomb and Z2103 on this very thread
@Vasistha; Does seem like those DinkaTepe samples strengthen the case for a shift in a particular direction.
I divided them into three clusters (basically the same as the paper, but I thought I4237 was intermediate enough and late enough to go into its own cluster 3), and it does seem like the direction of Cluster 2 (B) relative to 1 (A) is this direction towards Hasanlu_LBA_A, TKM_Parkhai_LBA_o and TJK_Sarazm_CA, : https://imgur.com/a/F0UL1rD
So I think that strengthens a particular directional shift and not so much just one outlier.
On the other hand I'm not totally certain or convinced one way or the other if whatever shifts between those subclusters does or does not involve some level of Sintashta like ancestry. More samples from Central Asia needed.
Interesting that Dinkha-Tepe doesn't show any of the R1b-M269 that appears at Hasanlu down to the IA, purely J and G. The site is, per the archaeological supplement, viewed as a satellite of Hasanlu, but if so there's no male line patrilineage flow evident (and the Cluster A/1 seems to have no sign of either Steppe_EMBA/MLBA related ancestry).
Well, TUK001 has probable Sassanian context.
Jeong 2020 notes
"One individual in this study (TUK001) at the site of Tamiryn Ulaan Khoshuu (Burkhan Tolgoi) dates to the era of Xianbei power in Inner Asia; however, there is no cultural context that could affirm affiliation with the Xianbei or other groups of northeastern China. Instead, recent excavations at this site have yielded artifacts, such as pottery from the Kwarezm oasis cultures near the Aral Sea and coins of the Sassanian Persian empire, that indicate significant interactions with areas in Central Asia and much farther west."
Target: MNG_TUK001:TUK001
Distance: 1.5561% / 0.01556143
56.2 TKM_Gonur1_BA
24.4 ARM_Black_Fortress_LBA
8.0 RUS_Kubano-Tersk
7.8 RUS_Sintashta_MLBA
3.6 IRN_Hasanlu_IA
0.0 IRN_DinkhaTepe_BIA_A
0.0 IRN_DinkhaTepe_BIA_B
0.0 IRN_Hasanlu_IA_new
0.0 IRN_Hasanlu_LBA_A
0.0 IRN_Hasanlu_LBA_B
0.0 IRN_Hasanlu_MBA
0.0 RUS_Kubano-Tersk_Late
0.0 TKM_IA
(scaled individuals, not averages).
Re; historical linguistics, Lazaridis on twitter linked up a recently published review of IE by Thomas Olander: https://www.cambridge.org/core/books/indoeuropean-language-family/4B44B5ACF0D3BBA89B9408050F112A52
Some quotes:
Introduction:
"Interestingly, the evidence for a subgroup consisting of Indo-Iranian and Balto-Slavic, occasionally discussed in the literature (Søborg 2020: 52; cf. Ringe, Warnow & Taylor 2002: 103–4), is considered to be insufficient by both Kümmel (Chapter 14) and Pronk (Chapter 15)."
In fact the "neo-traditional" model tends to place all core IE on a single unstructured star-like dispersal (indicative that the separation is too fast for many meaningful shared linguistic innovations to occur that are shared between branches due to ancestral reasons and not later contact, even if on some level it was more structured). See - https://imgur.com/a/8TSvRNd . Perhaps unusual that lay-people have taken the solidity of the idea of "Balto-Slavic-Indo-Iranian" further than specialists do.
On Anatolian by Alwin Kloekhorst:
"Although it is difficult to say anything certain about the absolute dating of reconstructed ancestor languages, in the case of Proto-Anatolian we have seen that its two best-known branches, Luwic and Hittite, have proto-languages that are roughly contemporaneous: Proto-Luwic can be approximately dated to the twenty-first–twentieth century BCE, and Proto-Hittite to c. 2100 BCE. The difference between the two is quite sizable, and elsewhere (Kloekhorst in press) I have therefore argued that they may have been a millennium apart from each other, which would mean that Proto-Anatolian started to diverge sometime around the thirty-first century BCE."
"Melchert himself thinks that such common innovations between Anatolian and “western” languages may indeed exist, but, to his mind, they would rather prove “post-divergence contact between Anatolian and the western dialects” (2016: 300), and thus have no bearing on the genealogical position of Anatolian. Although space limitations do not allow me to examine the four examples treated by Melchert (2016), it is quite clear that none of them can withstand scrutiny. There is thus no reason to assume that Anatolian shares any innovations, either contact-induced or caused by a genetic relationship, with “western” Indo-European languages or, for that matter, with any of the other Indo-European languages"
"With the Tocharian split commencing around 3400–3300 BCE, the Anatolian split may be dated to the period between 4400–4100 BCE. If Proto-Anatolian indeed first broke up into its daughter languages around the thirty-first century BCE (see Section 5.3.5), it would mean that it had some 1,300–1000 years to undergo the specific innovations that define Anatolian as a separate branch (see Section 5.2). Since these innovations include some large restructurings of especially the verbal system (loss of the subjunctive and optative mood, merger of the present and aorist aspects, creation of the ḫi-conjugation on the basis of the PIE perfect), such a time span would certainly be fitting."
Or, another rough model with better distance.
Target: MNG_TUK001:TUK001
Distance: 1.0315% / 0.01031520 | R4P
54.2 TKM_Gonur1_BA
23.0 IRN_Hasanlu_IA
12.2 Saka_Tian_Shan
10.6 RUS_Sintashta_MLBA
Post a Comment