Saturday, October 8, 2016
Dead cat bounce
Max Planck's Johannes Krause featured this curious map at a recent talk in Moscow on the Proto-Indo-European homeland debate (two hours into the clip here).
It appears to be an attempt to merge several of the main competing PIE homeland hypotheses into a single "hybrid model" under the umbrella of a modified version of the recently much maligned and now practically dead Neolithic Anatolian PIE hypothesis.
Make no mistake, it is an exceedingly strange effort that will be torn apart on several levels if it ever gets published.
See also...
Big deal of 2018: Yamnaya not related to Maykop
Late PIE ground zero now obvious; location of PIE homeland still uncertain, but...
Subscribe to:
Post Comments (Atom)
275 comments:
1 – 200 of 275 Newer› Newest»Actually guys its a Chalcolithic west Asian highland theory. Not Anatolian farmer.
Dave,
Re-write your Post, and simply say "The Kurgan hypothesis is dead" :D.
https://www.youtube.com/watch?v=JTY9K1Q_Sbg
Is this the same one?
@postneo
no it isn't, your link from Aug 5th has the final slide as "The Neolithic Revolution: Pots, People, Genes, and Diseases"
This more recent talk has the map and the final slide as: "The Neolithic Revolution: Pots, People, Genes, Languages, and Diseases"
It looks like he changed it in the last month. I wonder if this hybrid spread hypothesis has been in the work for awhile or if some new unpublished data is driving the theory.
It seems like maybe they have ancient DNA from India, and if so, then South Asians from ~4,500 years ago look like they're mostly ASI.
But that just means we need really big population movements into South Asia from the west after ~4,500 YBP to account for modern South Asian genetic structure.
This is a very solid model, which I've suspected as an alternative to my current stance.
Just waiting on Rakhigarhi and BMAC dna.
But this is very, very solid.
Wait a sec, how can Indus Valley be mostly ASI when it resembles Iranian and Middle Eastern Neolithic a lot in terms of material culture?
ANI label is next to IVC, it rather looks like they think it was mostly Iran_N.
@Atriðr
But this is very, very solid.
It's stupid and weak.
@George Okromchedlishvili
Wait a sec, how can Indus Valley be mostly ASI when it resembles Iranian and Middle Eastern Neolithic a lot in terms of material culture?
Elite, mostly male, dominance from the land of Elam?
@Shaikorth
ANI label is next to IVC, it rather looks like they think it was mostly Iran_N.
The North Indian pie graph is mostly ASI, no?
Looks like it, but why put "ANI" next to the pie if their Iran_N/ASI proportions were similar to tribals of Kerala? WHG and EHG also have identical pies even though they're not identical.
Krause, Haak & co. should know what ANI is broadly supposed to be.
@Shaikorth
Krause, Haak & co. should know what ANI is broadly supposed to be.
They should know exactly what it's supposed to be, and that pie chart obviously isn't it, because it looks like a Paniya sample or something similar.
Maybe they're just trying really hard to accommodate Heggarty and Gray?
But don't ever underestimate the ignorance of people supposedly working at the highest levels in their fields with even really basic stuff.
David
These guys are highly intelligent , and already know the results from India , BMAC
So take a chill pill
From what he actually says in the lecture, this hybrid model was mostly made to account for the Anatolian branch of Indo-European being the first to diverge.
So they say that maybe the Proto-Indo-European homeland was in Northern Iran.
Mostly I think this is just a way to keep the "controversy" going for a bit longer, and encourage people to think about how the known historical languages could have actually gotten where they were/are.
Obviously there were waves of language and genetic expansion, so the flaw is mostly in assuming a single accurate tree structure can really be produced for all of the Indo-European languages.
The wave model is a much better way to think about it, as the DNA shows. People went in all directions, and ever faster with better technology.
No tree model will ever be accurate.
Where's the arrow from the steppe to India gone ?
Where's the arrow from the steppe to India gone?
There's just a very faint one if you blow up the image and look really closely.
But if they're actually using ~4,500 year old Harappan remains as a proxy for the earliest Indo-Aryans in India, then it's not surprising that they didn't find any steppe influence, because the Indo-Aryans expanded from the steppe via Central Asia to India only ~3,500 years ago.
"Where's the arrow from the steppe to India gone ?"
Clearly this image was just thrown together at the last minute.
There will is no doubt that the Indo-Iranian languages are most closely related to Balto-Slavic languages. And there was no population movement from Iran into Latvia.
I never understood why the Anatolian Hypothesis ever got the kind of attention that it did. On the OIT which at one time seemed a promising alternative on the other hand was rubbished without much consideration. Given the available data it seems that its nearly time both these theories are consigned to the garbage heap once and for all. As for this map ancient DNA or not I shouldn't think we ought to give this map too much credit, the chronology seems way out of place atleast as far as India is concerned What expansion are they pointing to around 1500 ybp ? Indo-Aryan Kingdoms had been established in eastern and central India well before the Buddha (5th century BCE) The Dravidian Kingdoms of the Sangam age had already received a fair amount of Indo-Aryan Influence prior to the common era. Like wise 2500 BCE only marks the peak pf the mature Harappan period any Intrusion would have to be placed nearly a millenium earlier when the indus basin saw the beginning of a phase of industrialization
@davidski
Why specifically the 1500 BCE date? Is it just a round figure. I'm think a date a few centuries earlier might be more meaningful. Given that the Rg veda completed by 1200 BCE when the iron age began has no memories of central asia at the same time alludes to a healthy flow in the Saraswati a date closer to 1900 BCE might be more accurate
As for this map ancient DNA or not I shouldn't think we ought to give this map too much credit, the chronology seems way out of place at least as far as India is concerned.
There are so many things deeply wrong with this map that I'm mostly wondering how this stuff gets vetted and by whom at Max Planck?
Do they just slap together any old crap for these sorts of presentations at the last minute or what? Very worrying.
Why specifically the 1500 BCE date? Is it just a round figure.
Yep, just a round figure. I won't be perturbed if the real date turns out to be 1900 BCE.
A number of unusual features to that model. What is this red dot meant to signify? Clearly not Natufian / Levantine Neolithic, which at any rate shouldn't form a unified cluster with the Anatolian Neolithic I don't think... Somewhat simplified in Europe to only include the Early Neolithic as well. Let's see in future if any of it matches with any adna (hard to see how it can!).
Let me clarify. They have tons of unpublished data . So they have reasons to create a map to parallel their findings. However, India IMO was still ANI dominant by 2000 BC.
^ North India .
This has nothing to do with accommodating new ancient DNA findings, and everything to do with accommodating Gray's wonky language models and Heggarty's focus on making early farmers the proto-Indo-Europeans.
So in fact they don't have any good reasons.
So they have gone mad?.
OK, guess what, that pie chart in North India likely does represent ancient samples from the Indus Valley dated to ~2,500 BCE.
The North Indian pie chart is around 85% ASI and 15% Iran Neolithic. That's their version of ANI apparently based on ancient Indus Valley samples.
So they're simply not yet showing the steppe admixture
Its not sure what they mean until we see the data . I can bet that N India had ANI dominance still by 2000 BC .
However , they are suggesting IE languages were there in India already in the mature harappan period, So I guess they have data to demonstrate that . So this is the end Dave.....
Yes, it is the end for some things if IVC samples are indeed 85% ASI 15% ANI, like it seems based on that pie chart.
But you know the real thing is perhaps from here destroyed. Lots of academics perhaps even will cry if its true out of depression..
You don't have a clue what you're talking about.
Stop and think carefully: what does it mean if IVC samples from 2500 BCE are like modern day South Indian Dravidians?
You think carefully . Indo-Europeans were present in India in 2500 BC AKA mature Harappan period .
Its over Dave.
"they are suggesting IE languages were there in India already in the mature harappan period, So I guess they have data to demonstrate that"
This thread is the most conspiracy theorist in a while.
As if they had proof that IE languages were used in the harappan period...
How?
They found some secret writings in the DNA? They deciphered the Indus texts on accident?
Seriously?
Seriously are you blind and troll together . Its the main thing , IE folks were there in India as per their version . So its quite convincing they have data of 2500 BC remains .
Of course the final and decisive data is from script but until we have it , its the aDNA and others .
@nirjhar
Only two possible linguistic inferences from the harappan DNA
1) Harappans lack the steppe like ancestry present in modern indians (if the pie charts are correct, they have limited iran neolithic like ancestry as well making them simillar to isolated SI tribes). This would make an IA intrusion in the 2nd millenium BCE very likely pretty much tallying with the mainstream theory
2) They pretty much resemble northern upper caste Indians. This would mean that the possibility of IE as the language of the IVC exists but in no way ruling out language change divorced from pop movmts brought by small nos of an IE speaking elite from central asia
@ Davidski, using your AdmixQ13, if you ran with that 15:85 Iran_N:ASI as the pre-IE IVC contribution to Pathan and Santhal and Gond as proxy for ASI, and Iran_N as itself, then the post-Indus input population would seem to be, at around 60% in Pathan, approx: Iran_Neolithic 52%, Steppe EMBA 27%, Anatolia 8%, Bedouin 8%, Villabruna 4%, and misc other.
GujuratiD at a clip of 30% post-Indus gives effectively the same result, approx: Iran Neolithic 52%, Steppe EMBA 28%, Bedouin 11%, Anatolia 6%, Villabruna 3% and misc others.
This "ANI" / post-Indus would seem to most strongly correlated in proportions in that spreadsheet to Avar, Tabasaran, Lezgin, Tajik Yaghnobi, Chechen.
On the one hand that, and the AdmixQ13 itself, actually pretty difficult to reconcile with what Lazaridis 2016 estimated proprtions for Indian populations. But seems pretty perfectly OK at reconciling with what ANI would've been estimated to be close to by prior works.
Probably Neolithic movements to India were larger than 15% but if they have no adna from South India maybe that is how it will look.
Jijnasu,
They are a venerated scientific institute . They are saying that IE languages with ANI arrived in India around 2500 BC. This means the theory of steppe origin of Aryans is dead as it was Archaeologically,Anthropologically and Ancient Text Wise . Their map don't show anything coming out from Steppe to India . They show a movement from close to the area that I proposed for PIE.
Ah, bummer if those aren't the IVC samples. Apparently, Harvard does have some now...
http://www.tatkalnews.com/news/110060-Rakhigarhi-gets-scientists-from-Cambridge-Harvard.aspx
@ Davidski
You said to me that I wouldn't have to speak of conspiracy theory. OK. But whay to say about:
1) Natufians in Israel 44000 (?) years ago would be the ancestors of "Early Farmers of Anatolia" 9500 ya, of the great part of Central Europeans of 7500 ya, the same of Cantabrians, perhaps Basques. The 100% WHG of Western Europe would have contrinuted a little to Modern Western Europeans except more in the North and North East.
2) What has this to do with what we know: WHG of the Franco-Cantabrian Refugium went extint after the Younger Dryas and only the Villabrunas peopled Europe (and Anatolia, Middle East etc, at least in part).
3) Hg. R1b (but I think also J beyond I and some subclades of G) expanded from there: no R1b has been found either in Natufians, Anatolians, Iranians, only in Samara, and only around R-L23 I think come from west.
4) What has all that to do with the IE languages? The date of 8000 YBP are reliable, but what does demonstrate that they expanded from Zagros if not the Gamkrelidze-Ivanov theory and the Euphratic of our Nirjhar007? This is out of any reasoning. Their deepest link is with Uralic languages and the link with Finnic is due not to that old time but to an Indo-Iranian satem language that should be then in the Russian plain.
5) The possible link of IE with Semitic, inquired from Dolgopolkij and caressed from Nirjhar007, is out of any reality, if not in the shit theory that every language derive from Hebrew as in other times someone said that any IE language derived from Sanskrit. Full myth.
If its a bummer and they don't have any N Indian samples . Which I think its not the case. Then I think it will be a remarkable feat .
@nirjhar Reputed institutions and academics have been publishing their theories on this issue for decades now without concensus. wrt these questions regarding historical linguistics and human prehistory 100% certainty is not a possibility. The kurgan hypothesis despite its defects explains the indo-european language expansion the best despite several defects of its own. Since PIE is primarily a linguistic question the an equally effective alternative model will havr to be developed if the kurgan hypothesis is to be displaced. Therefore if the DNA findings are favourable the a steppe homeland is ( almost certainly ) correct. On the other hand even if unfavourable it is not ruled out but only leaves the door open for other possibilities
There is alternative models . You can understand it by your own if you read more . Steppe theory works for Eastern and Central Europe somewhat . But rest of Eurasia is mostly gibberish.
@ Davidski
As I demonstrated that F38 found in Iran derived from the R-L23-M584 from Samara, and from the same haplotype derived an Englishman, and a Lebanese from a Crusader linked to him, and very likely from Visigoths derived Iberian Grijalba ancestor of the Jewish Spyra family, so I can demnstrate that (seen that we spoke about R-Z2109*):
N83705 Southern Iraq, Arab [Arab] Iraq R-KMS75
12 24 14 10 11-15 12 12 12 13 13 29 17 9-9 11 11 24 16 19 31 14-15-17-17 11 11 19-21 15 15 17 17 36-36 12 12 11 9 15-16 8 10 10 9 10 11 12 23-23 16 10 12 12 14 8 12 22 20 13 12 12 13 11 11 12 12
derived from Samara R-Z2109-KMS75 with DYS534=14
329335 Mr. Raiman Russian Federation R-KMS75
12 24 14 10 11-14 12 12 12 13 13 30 16 9-10 11 11 25 15 19 30 15-15-16-17 11 11 19-23 16 15 20 16 37-37 13 12 11 9 15-16 8 10 10 8 10 11 12 23-23 16 11 12 12 14 8 12 24 20 13 12 12 13 11 11 12 12 34 15 9 16 12 23 26 20 13 11 13 12 10 9 12 12 10 11 11 30 12 13 25 13 10 10 20 15 19 13 24 17 12 15 24 12 25 19 10 14 17 9 11 11
Thus no doubt to me that there was a migration from Samara to Middle East (Iran/Iraq).
These are exact proofs that no conspirasy will be able to abolish.
Poss of interest to some, I had a few bored moments on my time, so interested by trying to compare the Admix13Q ADMIXTURE values to the Lazaridis 2016 values for South Asians, I made some graphs comparing them:
http://i.imgur.com/4yEHTVk.png
(Note, I adjusted the Steppe EMBA values from Admix13Q to be higher than in the basic spreadsheet Daviski gave, as I noticed that Yamnaya had about 11% Iran_N, so stole some IranN to roll into Steppe EMBA).
It does look like there's a difference. Lazaridis values have South Asians having a base of high level of Steppe EMBA, then accumulating more so it's about half by Kalash, and also a higher accumulation of Onge+Andaman components into less Steppe EMBA populations. Meanwhile the Admix13Q shows a higher baseline of East Asian and Siberian+Amerind+Beringian components in all South Asian populations and in southern South Asian populations.
I'm wonder if this is an indicator of excess Ancient North Eurasian in South Asians that's offsetting the four populations in Lazaridis's model, and so giving a higher level of Steppe EMBA across the board than in reality?
Besides what's shown, in Admix13Q, also an accumulation of BedouinB and Anatolia components, and some SSA and Villabruna components.
Dienekes and Olympos Mons are taking their terrible revenge! :-)
I tried to figure out how this model could work. First of all, any model should account for the fact that Indo-Iranian and Balto-Slavic languages are quite close and should not represent the first split.
The years given should be years before present.
So, it looks like the Proto-IE language was spoken in Shulaveri-Shomu culture (circa 6,000-4,000 B.C.E.).
Anatolian languages are well explained as there are no difficulties in explaining their archaic character and migration route. Then the model proposes a Centum migration over the Caucasus to Southeast Europe. This is quite okay as it would explain why Greek, Albanian and Italo-Celtic languages are so different and divergent. Albanian is a special case as it is classified as a Satem language with incomplete satemisation. The impulse of this first wave seems to be metal working and its expansion from Caucasus to the Balkans. One innovation of this model is that it suggests that proto-Satem was spoken in the area to the south of Caucasus until the kurgan migrations began. Indo-Iranian languages found their way to IVC c. 2500. It looks like the Yamnaya Culture would be the northern branch of Satem languages that gave rise to Balto-Slavic languages. The contact zone between Uralic languages and IE languages would be between Yamnaya and Volosovo which is okay but the contact is restricted to Balto-Slavic branch. It must be noted that Yamnaya dates are older near the Urals than in the south but this can probably be explained out somehow.
The good thing here is that there is no need for any Armenian migration as it could have diverged from the protolanguage as a result of contacts with the Caucasian languages there where it is. On the map, Tocharian would preferably be an upshoot from Iran to Xinjiang which is better than from Yamnaya to Xinjiang as Tocharian is not a Satem language.
Another good thing about the model is that it is not stuck on yDNA question. The model is probably based on multiple y lines and does not care about the fact that Indians are R1a1 and Yamnaya is R1b. However, I wonder if they have detected R1a1 in IVC c. 2500 BC and these guys are CHG and not steppe like.
This is an exciting way of explaining the deep roots of IE languages. I had a good model in my head but it did not explain Anatolian, Armenian or Tocharian languages. On the other hand, I still doubt that a language to the south of the Caucasus Mountains could be structurally so close to languages in northern Eurasia such as Uralic and Turkic languages. Maybe the model suggests that these languages were part of this modernization process of the whole Eurasia.
@David
It's stupid and weak.
No, it's not. It's also called the dispersal of I-E languages, which is exactly what the model should be.
I can explain more if you need.
It's still not my primary model, but they might have access to info I do not; which is why it'd be my alternative depending on 2 variables.
@Kristiina
First of all, any model should account for the fact that Indo-Iranian and Balto-Slavic languages are quite close and should not represent the first split.
Yes, yes, yes. This is key and often just glossed over. There are some ways around. Also, I'd also say recognizing elements Graeco-Aryan.
@Atriðr
I can explain more if you need.
It's still not my primary model(...)
Can you say how your primary model looks like? I wonder what you've found.
Judging by this:
This is key
you're on the right track.
The problem with this model is: There is no IE stem for things like tiger. There is for things like birch. The other problem is that the link between proto-Uralic seems *deep* (Yes, Rob, I know its a minority point of view ;)
@Kristiina
"I wonder if they have detected R1a1 in IVC c. 2500 BC and these guys are CHG and not steppe like."
I don't think so at all, but what an amazing coincidence that would be?!!!
@Rob
Or he did for the PIE Urheimat hypothesis what Fred Hoyle did for the Big Bang: Deliberately create an alternative to keep the discussion going and sharp. Really, there should be a law against theoretical monopoly as well ;)
@Rob
PIE also doesn't have reconstructed *words for hands, but does for feet, doesn't for rain but does for snow. I guess the PIEs came from some bizarre alternate dimension via a matter transporter.
LOL!
However, some of the environmental probing into what actually can be found in PIE came to about the same Urheimat as the Yamnaya Urheimat theory.
@Rob
PIE also doesn't have reconstructed *words for hands, but does for feet, doesn't for rain but does for snow. I guess the PIEs came from some bizarre alternate dimension via a matter transporter.
Thanks for this witty come back. Makes discussing with you a pleasure, sir.
@Kristiina
They've perhaps found Iran_N in IVC, but not CHG.
Look at subtable 3. Outside Caucasus, Anatolia and one of the Greek samples CHG(KK1) never appears as a donor without Loschbour which also proxies for EHG, and every population in North India prefers WC1 (Iran_N) to KK1.
http://science.sciencemag.org/highwire/filestream/681447/field_highwire_adjunct_files/3/Table_S24.xlsx
@ Epoch
It's actually an adapted partial quote from a famous old linguist. I thought you might get it, but obviously not
Have some Canomile tea
@Rob
I am getting the impression we miss each others wit and points. I did like the quote though and get that we might hypothesize what we want, but we don't know what is lost so we don't know what we missed. However, as circumstantial evidence such probes may serve as supporting evidence. Or may not, whatever.
As for the camomile tea: As I am on Belgian beer tonight I'm afraid I'll skip this offer. Fancy a beer?
@ Kristiina
"On the other hand, I still doubt that a language to the south of the Caucasus Mountains could be structurally so close to languages in northern Eurasia such as Uralic and Turkic languages"
@ epoch
"The other problem is that the link between proto-Uralic seems *deep* (Yes, Rob, I know its a minority point of view)".
I think that an Uralic language speaking like Kristiina should get many doubts about this theory. IE has to be born in the Siberian corridor. If this hypothesis were true, we should disconnect IE from hg. R1 (a and b) and link it to hg. J, which expanded to India long before, but also J came from WHG like R1 (this is my thought, in fact I expect that the oldest J1 is found in Palaeolitic Italy or Western Europe).
Absolutely ! Belgian is the best- esp. Chimay blue
But it's early morning here, which is probably why I might not be articulating well
@ Atrior
"There are some ways around. Also, I'd also say recognizing elements Graeco-Aryan."
On last readings, it was doubted that G-A even exists (?)
"PIE also doesn't have reconstructed *words for hands, but does for feet"
Is PIE reconstruction infallible? There is hasta, hAtha etc (indic) and hand(germanic). The criteria that says cognates have to work in all branches should be reexamined. I think there is a common lost indo-germanic lexicon lost in later intervening branches.
Not all branches are guaranteed to preserve the same lexicon especially if they peeled off at different times. The sound shits governing one set need not be the same as another.
As for this map, its nothing much.... just a pictorial representation of the Lazadis paper. Also its such crappy resolution, I am surprised people can comment and see so much. does someone have a better image??
there is no bronze age pie chart from peninsular india, so both the north indian and peninsular pie charts are hypothetical place holders to be consistent with a iran chalc to yamnaya arrow and then yamnaya to IVC arrow.
the arrows are very weakly drawn and fizzle out in intervening regions?... indicating uncertainty.
Lets just face it, they have no data from asia for regions of interest exept Hotu, kotias and satsurblia. The problem is that these are from 9000 to 3000 years older than the period of interest.
PonstNeo- dont take my comments too literally, & and it would be silly to suggest anyone expects any method to be infallible. But the moral of the story is many linguists have fabricated evidence from where there is none, ar at least, have become a little "creative".
I found the original quote: from A Keith, via S Piggot. "taking the linguistic evidence too literally, one could conclude that the original IEs know butter but not milk, snow and feet but not rain and hands.."
@ Karl_K
"No tree model will ever be accurate."
No, but there is model which can markedly improve on the current trees, which fall short
@arza
what allowed me to find is this: listen to everything, be attached to nothing, and never disregard the simple and evident.
Tangential to this (and I've said this here before), the similarities (and cognates) between Russian (of course, can use Lithuanian too) and Sanskrit are in too many cases identical. Moreover, there is an element I call "flow." When you speak Sanskrit, the flow is like in Russian, or rather vice-versa. There is more than meets the eye in this comment.
Also, anyone familiar with Slavic myths and folk stories, might find a trail of bread crumbs to follow.
This also allows me to segway to this comment of @Kristiina that I like:
This is an exciting way of explaining the deep roots of IE languages. I had a good model in my head but it did not explain Anatolian, Armenian or Tocharian languages. On the other hand, I still doubt that a language to the south of the Caucasus Mountains could be structurally so close to languages in northern Eurasia such as Uralic and Turkic languages. Maybe the model suggests that these languages were part of this modernization process of the whole Eurasia.
@Rob
Curious. I was wondering this might have been your model ;)
Some elements of it to be sure. ;)
Re: G-A. To be clearer, I meant aspects of, not the fill paradigm itself.
@Gioiello
IE has to be born in the Siberian corridor. If this hypothesis were true, we should disconnect IE from hg. R1 (a and b) and link it to hg. J, which expanded to India long before
Here lies the mystery.
@ Atrior
"what allowed me to find is this: listen to everything, be attached to nothing, and never disregard the simple and evident. "
Wise words!
This model is what I and some other have been proposing for some time. The Steppes being secondary homeland for major part of the Indo European expansion towards Europe, but the more archaic Indo European groups actually evolving further south.
It looks to me like they have new data at hand that brought them to this conclusion.
@Kurti
This model is what I and some other have been proposing for some time.
But the model doesn't work at a very basic level. To make it work they had to pretend that Neolithic farmers from Iran were Indo-Europeans and the steppe was Indo-Europeanized by Caucasus Hunter-Gatherers.
At some point you have to start thinking more clearly. Try and be more objective and reasonable. Just because you want this model to be true, and just because some academics at Max Planck do too, doesn't mean it's true.
It looks to me like they have new data at hand that brought them to this conclusion.
They're just desperate and trying to make one last go of a homeland in the Near East.
It won't work though.
Kurti,
Yes Kurti this model goes mostly to which I and others suggested for sometime .
It will be great to see the hard data behind the conclusions . It also somewhat matches the description and possibilities of Moorjani et al regarding the case of S Asia and ANI ASI admixture .
@ Davidski
"They're just desperate and trying to make one last go of a homeland in the Near East".
Thank you. So long ago I called them desperate scullery-boys!
Some commentators seem to assume that the Satem languages somehow form a clade distinct from the kentum languages. This is rather doubtful. Likewise Greek and Indo-Iranian share several features but these may not imply that they form a distinct clade either
Is there any archaeological evidence for the existence of PIE language in neolithic Iran?
"Is there any archaeological evidence for the existence of PIE language in neolithic Iran?"
Is there archaeological evidence for PIE language in Neolithic steppe ?
@Rob
No, but one could imagine more finds preserved from Iran. I don't know what exactly. Loanwords in first written languages? Anything.
There's nothing pointing to Iran being the PIE homeland. It just gets picked by total dilettantes to the debate because of its central geography.
Iran is the Elamite homeland. The Indo-Europeans got there quite late in the game.
@Rob
I mean if PIE originated there you may reasonably find some traces of it in the earliest writings in the fertile crescent, their neighbours. Even if that doesn't prove anything I am curious if this is the case.
Is there archaeological evidence for PIE language in Neolithic steppe?
Sure, the right male dominated culture and the right DNA at the right time.
@ Epoch
I'm not advocating an Iran Neol. PIE homeland, but consider the flip side.
"Loanwords in first written languages? Anything."
I think some scholars do point to Sumerian, Kartvelian loans. Remember, that most languages even in West Asia became extinct, and several langauges expanded during the Bronze & Iron Ages, such as Semitic, probably from southern Mesopotamia. So it would have erased all language diversity and many of the contact languages for which might have once had loans.
So who knows ?
The steppe looks to have been demographically volatile, groups coming and going, large borrowing of culture and technology from neighbours, and, it would seem, a lot of exogamy. This does not create linguistic stability, but constant linguistic flux. So the chances of PIE actually originating on the steppe are actually quite minimal, but in terms of spreading from there - yeah absolutely. .
@ David
"Iran is the Elamite homeland. The Indo-Europeans got there quite late in the game"
Agree. However, Elam was in southwest Iran.
"Sure, the right male dominated culture and the right DNA at the right time."
Not really Dave. In the neolithic steppe, females appear to have been the vehicles of elite display. Sexually dimorphic burials, an male chiefdomship had origins outside the steppe.
Not really Dave.
During the Bronze age a certain type of people moved out of the steppes with a unique culture and DNA. Unless they were mutes, then they look like the Proto-Indo-Europeans. They were very different from Neolithic farmers and their movements are now easy to track.
And that's the PIE homeland debate won in a couple of sentences.
Post by arza deleted.
Way too much bullshit. Try and stick to reality next time.
I don't want to read too much into a graph that (as others have noted) seems to be a very sloppy visualization, definitely something they put together last-minute. On top of that, the theory underlying this visualization seems rather problematic.
Still, focusing on South Asia, and playing devil's advocate, if it truly turns out that IVC samples are, in genetic terms, similar to Paniya-like South Indian tribal populations, this would have some very radical implications.
For one, as David noted, we would be forced to postulate massive expansions into South Asia, from both Iran and Central Asia/the steppe. I mean, the Balochistanis are overwhelmingly Neolithic Iranian, in terms of genetic ancestry (with the addition of Bronze Age steppe admixture, and whatever ASI happens to be). Neolithic Iranian ancestry is an essential/very heavy genetic component for all Afghan/Pakistani/North Indian populations, not to mention non-Turkic Central Asians. In addition, the levels of steppe admixture throughout the region are quite high.
Basically, if the IVC population was Paniya-like, at least one rather radical implication would be that IVC ancestry is a minor genetic component in this region, the very region it once thrived in.
We can easily test how South Central Asian populations would look, in this scenario. I think nMonte is good for that sort of analysis.
For what it's worth, I've been looking at different nMonte sheets, and found that PCA and d-stat based nMonte sheets yield identical output, when using identical references. This is a very good sign, in terms of how robust these results truly are.
By contrast, nMonte used with ADMIXTURE output yields results which are quite different from both PCA and d-stat based analyses. Obviously, that isn't surprising, considering the nature of the information which is being inputted. So, I decided to stick with PCA/d-stats nMonte sheets.
Here are some models of South Central Asians, using steppe references (Afanasevo, Sintashta, Srubnaya, etc), ancient Iranian references (Neolithic Iran, Iran_Hotu, etc), the Paniya (our theoretical stand-in for IVC), and the Ulchi (to track non-IE influences from the steppe/Siberia):
Kalash sample
46.55% Neolithic Iranian
37.40% Bronze Age Steppe
16.05% Paniya
Pakistani Pashtun sample
42.55% Neolithic Iranian
37.55% Bronze Age Steppe
19.80% Paniya
Pakistani Brahui sample
73.10% Neolithic Iranian
21.45% Bronze Age Steppe
5.45% Paniya
Pamiri sample (Ishkashim)
47.35% Bronze Age Steppe
37.50% Neolithic Iranian
8.90% Paniya
6.25% Ulchi
Pamiri sample (Shugnan)
56.2% Bronze Age Steppe
36.95% Neolithic Iranian
5.85% Paniya
1.00% Ulchi
If the IVC population was like the Paniya, it would mean that the Brahui, a relic Dravidian population, the closest living people we probably have to the ancient population of Elam, are only 5%-10% IVC-admixed!
And Pashtuns/Kalash would only be 15%-25% IVC-admixed, while the Pamiri peoples would only be 5%-15% IVC-admixed.
If true, this would be some shocking stuff. We'd be dealing with massive genetic discontinuity in South Central Asia, some very serious population replacement.
Sidenote: When it comes to the Paniya themselves, I'd say it's a good bet that they are 60% ENA (their ENA, like most South Asians, is intermediate between the Onge and East Asians proper), and 40% Neolithic Iran/Iran Hotu.
Also, I included MA1/Afontova Gora 3 in all my fits, to account for extra ANE. None of these South Central Asian populations needed it, Neolithic Iran + Bronze Age Steppe was enough.
Sure, Yamnaya people were dark as Near Easteners are today.
This model for IE languages is probably based on cultural evolution and not on race, mass migrations and replacements. So, it is antiracial and antiracist which I like. Of course, it is just a model.
It is true that there was a lot of IE linguistic diversity in Anatolia: Hittite, Lydian, Luwian and Palaic and their subdivisions. and some old texts are available. According to Wikipedia most of the records are dated to the 13th century BC. In any case, there is an old presence of IE languages from Anatolia via Armenia to India (Indo-Iranian).
Words (excluding maybe some less used basic terms) often spread as waves. A tree model is unable to catch it. Also languages that separated a long time ago can start converging and can continuosly change new words. For example, the Alexander the Great’s empire created a channel for exchange from Greece to India at a time when all these languages were surely well established. However, more long lasting kingdoms and commercial networks surely had a bigger impact in the long run. I am sure that the Roman Empire is a great example of this kind of "exchange network".
@Seinundzeit
Why would this Paniya-like population not rebuild their civilization or something like it below the Vindhya Range where the Paniya people are found?
The IVC was a massize civ and it wasn't leveled to the ground. The dancing girl (artifact) might resemble a Paniya, but not the bust of the priest-king (artifact).
Or are you suggesting something else by:
If the IVC population was like the Paniya
@ Dave
"During the Bronze age a certain type of people moved out of the steppes with a unique culture and DNA. Unless they were mutes, then they look like the Proto-Indo-Europeans. They were very different from Neolithic farmers and their movements are now easy to track."
No arguements there
@Atriðr
I'm just taking the graph at face-value. The most ASI-rich living population are the Paniya, and that graph suggests that IVC was, in genetic terms, mostly ASI.
If so, IVC ancestry would be a rather minor component in modern South Central Asia.
And if IVC was even more ASI-admixed than the Paniya (who seem to be anywhere from 30% to 50% West Eurasian), the levels of IVC ancestry would be even lower, like only 10%-15% among Pashtuns/Kalash (with Paniya, Pashtuns/Kalash seem to be at around 15%-25% for this sort of ancestry).
This is very unexpected, to say the least. Looking at the modern picture of genetic variation in this region, one would expect IVC to be overwhelmingly Neolithic Iranian, in genetic terms.
This is definitely something to look forward to, can't wait to see what all of this means.
Also, IVC wasn't leveled to the ground. Rather, it collapsed pretty hard, the cities were abandoned, and it seems population levels went drastically down. I'm no archaeologist, but natural causes for the collapse would make sense.
If I'm not mistaken, this is academic consensus.
epoch2013
"I mean if PIE originated there you may reasonably find some traces of it in the earliest writings in the fertile crescent, their neighbours. Even if that doesn't prove anything I am curious if this is the case."
Traces actually do exist in the earliest writings. In fact the earliest attests of anything Indo European in texts is from Western Asia. I thought that should be widely known fact by now. A major reason why many people (including me) have a hard time believing that this all is due to migration from outside and not at least a part of it is ancient in the region. Indo European names are mentioned already in Sumerian texts and predate any known supposed Indo European "migration" into the region.
@Seinundzeit
can't wait to see what all of this means.
It's right there. The meaning. In your comments.
academic consensus
Changes like the wind.
@Atriðr
Bit of a misunderstanding here. That's not what I meant.
Rather, I can't wait to see whether that graph is based on an actual IVC sample, and whether its tribal South Indian-ness isn't a fluke, some by-product of a problematic method that might have been used to analyze it. That's what I was implying by "can't wait to see what all of this means".
If it is based on an actual IVC sample, and if the analysis itself isn't problematic, then yeah, very shocking stuff, massive population replacement in South Central Asia. A huge genetic gulf exists between Paniya-like people and modern South Central Asians, you just can't derive the modern population of the region from ancient people similar to modern tribal South Indians.
Also, when you say academic consensus changes like the wind, in this case, do you personally think the IVC never collapsed? Just curious.
@Seinundzeit
Also, when you say academic consensus changes like the wind, in this case, do you personally think the IVC never collapsed? Just curious.
The IVC was numbered in the millions. Why would it collapse? Desiccation? And no one rebuilds elsewhere? Massive Earthquakes where the buildings still stand? No, it didn't collapse.
Atriðr
The Paniya are a tribe of hunter-gatherers residing in the south western extremity of the Indian peninsula. Even if it so happens that the IVC inhabitants resembled the Paniyas in terms of a being an ASI heavy population, it does not in any way imply that they were the direct ancestors of or even closely related culturally to the Paniyas. It just means that both of them share common ancestry in the distant past from the hunter gatherers of pre-Neolithic south asia
@Atriðr
Fascinating, the first time I've ever read someone claim that IVC never went through a rather brutal decline/pretty much collapsed.
For example, late Harappan skeletal material has demonstrated, to all the archaeologists I've read, a great increase in violent deaths, and an increase in death via infectious disease, compared to older Harappan levels.
On top of that, it's well understood that nearly all IVC cities were abandoned, during approximately the same time period, if I'm not mistaken.
If that isn't a civilization collapsing, I suppose we have different definitions of "civilizational collapse".
Perhaps, by collapse, you mean a literal collapsing of buildings/walls, like what the Book of Joshua claims happened to Jericho? (Lol)
As for the collapse of the IVC. There's no evidence of any millitary invasion or arson. A natural population decline brought on by drought combined with the gradual decline in flow through the ghaggar-hakra system (possibly the Vedic Saraswati) might have been an important cause. I however find this scenario of them being predominantly ASI unlikely as it would mean a very minor contribution to succeeding populations. Archaeology so far suggests a significant degree of cultural continuity in the region despite the de-urbanization. My own theory is that the indo-aryan speaking immigrants entered the subcontinent around the time the IVC began to decline, taking up local female as partners and incorporating local artisans into their community. I would expect the people of the IVC to be more baloch like though probably speaking some language isolate
Dear Sein,
Thanks for the analysis. What this suggests is that if N_Iran-like population i.e. ANI moved into South Asia after 2000 BC, then there would have been a massive level of population turnover in South Asia. This in addition to the suppossed movement of steppe people who also apparently admixed with all South Asians including remote South Indian tribals.
This is impossible and therefore aDNA from South Asia will never support such a scenario. Let me just inform you if you're not already aware - While the N_Iran from sites such as Ganj Dareh have their closest modern groups in people like the Baloch, one of the earliest sites of Neolithic South Asia, Mehrgarh is also in Balochistan and its archaeology has already shown that it is very closely related to Iranian Neolithic sites such as Ganj Dareh. The inference ? It is most likely that from the earliest period of Neolithic in South Asia, people here were already quite closely related to Neolithic Iranians.
Hence a latter day movement from Iran into South Asia - around 2000 BC - causing massive population turnover looks very unlikely.
What is more likely is that there was some sort of population movement from South Asia into Iran - the South Asian Zebu and South Asian buffalo are an integral part of Iranian farmers' lives and these animals seem to have been present in Iran from atleast the 4th millenium BC. While Zebu bones have been discovered from the Helmand site of Shahr-I-Sokhta, buffalo and probably South Asian rice was already used by the Sumerians.
@Seinundzeit
to all the archaeologists I've read, a great increase in violent deaths, and an increase in death via infectious disease, compared to older Harappan levels.
You mean like Detroit? Or bubonic-plagued Europe? As skillful as your snark was, I think you should account for millions as opposed to whatever vestiges archaeologists have found. Nevermind the geographic breadth of the cities.
Mind you, I never discounted conquest.
@Jijnasu
I know what the Paniya are, which is why I referred to the Vindhya Range to begin with.
@Kristiina
The devil is in the details, and this model is going to hell.
Have a closer look at what they're suggesting; language transfers between groups that don't look like they've had any contacts since the Paleolithic.
Or am I wrong? Can you find any direct links between Yamnaya and Neolithic or even Chalcolithic Iran apart from some decent statistical fits?
Take a look at the uniparental markers of these groups, see anything that matches? No? So how did they exchange culture and language if Yamnaya doesn't look like it has any direct admixture from the South Caspian? Where's the Y-DNA G and mtDNA U7 or other classic Near Eastern markers?
Oh wait, maybe the languages went from Iran Neolithic to the Neolithic descendants of Caucasus Hunter-Gatherers, and then onto the steppe? And then what, Balto-Slavic languages came from the steppe, while Indo-Iranian from Iran?
We need to be practical and brutal here. This convoluted model isn't even hanging on a thread. In fact, it's not even worth discussing unless someone actually publishes a bit of ancient DNA data that backs it up.
Currently, as things stand, we not only need some relevant data, but this data will have to be strong enough to reverse some very robust trends that have already appeared in the data published to date. I don't see that happening.
As for the IVC collapse, it is not so clear cut.
The collapse narrative suits those who want to posit some sort of Indo-Aryan migration into South Asia around 1500 BC. Hence it is bandied about a lot by these groups.
But a more naunced take on the issue seems to suggest, to my mind, a slightly different outcome - It is most likely that the mature IVC phase represents, as many archaeologists have suggested, a sort of centralisation of power while the earlier era was that of regionalisation. Centralisation is a process that cannot by its very nature last foreover. Hence the late Harappan phase is most likely the period where the Centralisation of power collapsed and it was replaced by smaller regional powers. Such a situation would indicate some massive upheaval in society but it is highly doubtful that it would lead to a major decrease in population.
@ Davidski
"We need to be practical and brutal here. This convoluted model isn't even hanging on a thread. In fact, it's not even worth discussing unless someone actually publishes a bit of ancient DNA data that backs it up.
Currently, as things stand, we not only need some relevant data, but this data will have to be strong enough to reverse some very robust trends that have already appeared in the data published to date. I don't see that happening".
I adfirm here what has been my thought in these ten years: not only they don't have what you say, but very likely they have all opposite data. For that your adjective "desperate" was the best, I think.
(I didn't speak of "conspiracy" as you forbade me to do).
I'm seeing two 'blips' on the north India pie. One green one blue - they're just both small, and right next to each other.
That's a negative IMO. Compare the dark portion of the North Indian pie chart to the Yamnaya blue.
https://drive.google.com/file/d/0B9o3EYTdM8lQZjlNU2FXeEdQYTg/view?usp=sharing
That portion looks Neolithic Iranian teal.
To everyone,
Although a huge genetic gulf does exist between South Central Asians and ASI-rich South Indians (which is why you can model Pashtuns/Kalash as only 15%-25% Paniya, and Balochistanis + Pamiris as only 5%-10% Paniya), this doesn't necessarily hold for some Pakistanis and North Indians. For example, you can model Lahori Punjabis as so (this is the same setup that was used on the South Central Asians):
60.70% Paniya
20.95% Neolithic Iranian
18.35% Bronze Age Steppe
Although Paniya was a minor component for Pashtuns/Kalash, it's the main component for these Punjabis.
Also, here is how Chamar from Uttar Pradesh stack up:
86.85% Paniya
13.15% Neolithic Iranian
0% Bronze Age Steppe
The Chamar are overwhelmingly Paniya-like, and they don't show any steppe ancestry (as a reminder, with these models, Pashtuns/Kalash are 40% Steppe-admixed, and the Pamiri are at 50%)!
So, if we assume that the IVC population was Paniya-like, that doesn't mean they disappeared into thin air. There are still Pakistani populations who could, in such a scenario, be 60% IVC-admixed, and there are still North Indian populations that could, in such a scenario, be 85%-90% IVC-admixed.
So, IVC being a very minor genetic component actually only applies to Tajik + Afghan + all western Pakistani + perhaps some eastern Pakistani populations, while great swathes of people in North India (and some populations in eastern Pakistan) could still be anywhere from 60% to 90% IVC-like.
But again, this is all based on a very sloppy, vague, tentative graph. Hell, we don't even know if they have an IVC sample. Everything I'm saying is contingent on IVC truly being ASI-rich, which is something we are only gleaning from this solitary graph.
We just have to wait for the paper.
Still, considering the Punjabi and UP Chamar scores for Paniya-like admixture, I think IVC being Paniya-like sounds much more reasonable now.
Yeah, I was gonna say, some of the Chamar samples that I have are almost like Paniyas. And those Punjabis from Lahore and GujaratiD individuals are indeed largely like Paniyas. All of these groups are from close or among former IVC sites.
Didn't think it needed saying, but apparently it does. The relevance here is the Iran_N.
Or to be fair, the admixture of Iran_N + ASI, whose fractions tell us a story.
What is the relevance to the PIE homeland debate of admixture from the Neolithic Zagros in a putative IVC sample?
Isn't that more relevant to, say, the Elamo-Dravidian debate?
David,
Reexamining everything, you're absolutely right, it actually makes sense.
Originally, I was only testing South Central Asian populations, not South Asians proper, which was giving me a highly skewed/inaccurate perspective.
Looking now at the South Asian populations you mentioned, it really wouldn't be shocking at all, for an IVC sample to be Paniya-like.
In fact, taking into account these Punjabi, Chamar, and GujaratiD individuals, perhaps we should always have expected IVC to be rather Paniya-like.
So if that pie chart really represents an IVC sample, and I'm hearing that it does, I guess that finally settles it.
Yamnaya does have admixture from the South Caspian. Almost half of their ancestry comes from the south.
What are the direct links between Yamnaya and Khvalynsk? Yamnaya yDNA may turn out to be common in ancient South Caucasus. The Khvalynsk R1b was covered with metal objects and metallurgy came from the south. I quote your own text:
”Male (confirmed genetically), age 20-30, positioned on his back with raised knees, with 293 copper artifacts, mostly beads, amounting to 80% of the copper objects in the combined cemeteries of Khvalynsk I and II. Probably a high-status individual, his Y-chromosome. Haplotype, R1b1, also characterized the high-status individuals buried under kurgans in later Yamnaya graves in this region, so he could be regarded as a founder of an elite group of patrilineally related families”
The circle on the map is in northwest Iran very close to the area of Shulaveri Shomu Culture in southern Caucasus in Georgia, Azerbaijan and the Armenian Highlands and we do not have any DNA from there. R may appear again, even R1a. It is true that ancient Levant yDNA is very different, but R was present in Iran and Caucasus in the latest papers: Iran Neolithic was P1(xQ, R1b1a2, R1a1a1b1a1b, R1a1a1b1a3a, R1a1a1b2a2a) and Armenia EBA was R1b1-M415. MtDNA is not a problem because y lines often spread independently from mtDNA.
The fate of this model is decided by the ancient DNA. The model could be interpreted in a way that Yamnaya spoke Proto-Balto-Slavic-Germanic and you should be happy about it and celebrate it, because there you have steppe ancestry and kurgans in the same package.
How does the ratio of steppe to iran Neolithic vary across different populations? Is it nearly constant or is there a caste/regional gradient?
@Kristiina
Yamnaya does have admixture from the South Caspian. Almost half of their ancestry comes from the south.
Nope, definitely not from the South Caspian. See here.
It's also important to note that the uniparental marker data in Lazaridis et al. firmly back up my TreeMix output, with the Steppe EMBA groups showing starkly different Y-chromosome and mitochondrial (mtDNA) haplogroups from the ancient samples from Iran.
Indeed, mtDNA haplogroup U7 is an excellent diagnostic marker for ancestry from the southern Caspian region, and, sure enough, it appears in the Iranian Chalcolithic set. Conversely, it's conspicuous by its absence from all Bronze Age steppe remains tested to date.
http://eurogenes.blogspot.com.au/2016/06/yamnaya-eastern-hunter-gatherers-iran.html
http://eurogenes.blogspot.com.au/2016/07/modeling-steppeemba.html
As things stand, there is no evidence of Yamnaya having any ancestry from the putative PIE homeland on that map.
Jijnasu,
I think it would take manual testing of many South Asian populations, from all over India, to get a clear picture of that sort of thing.
But for what it's worth, we can compare UP Brahmins (obviously, highest caste in their province) to Chamars from the same sampling locations:
53.35% Paniya
26.10% Bronze Age Steppe
20.55% Neolithic Iranian
These Brahmins are way more Paniya-shifted in comparison to South Central Asians (around 55% Paniya compared to around 15%-20% Paniya in the Pashtun/Kalash samples I ran).
Still, these Brahmins are substantially different from the Chamars of this very same province (25% Steppe admixture versus 0% Steppe admixture, 55% Paniya-like versus 85% Paniya-like).
But looking at your question, I guess it is of significance that UP Brahmins have more Iranian Neolithic ancestry compared to UP Chamars.
Perhaps, the Indo-Aryans had extra Neolithic Iranian ancestry, via BMAC, the effect of which we see in this Brahmin versus Chamar comparison. That's the best I can come up with.
@Jijnasu
Steppe ancestry in South Asia shows a strong correlation with language, geography and caste, more or less in that order.
The most steppe admixed South Asians are high caste Indo-European speakers from the northern parts of South Asia. Also, Indo-European speaking South Asians usually have higher levels of steppe admixture than their Dravidian speaking neighbors, and high caste Indians usually have higher levels of steppe admixture than their lower caste neighbors.
I'm not aware of any Dravidian groups that have clearly more steppe admixture than nearby Indo-Europeans, although such cases are possible. And indeed, Dravidian speaking Brahuis have higher levels of steppe ancestry than many Indo-European speaking Indian groups from Central and South India.
On the other hand, Neolithic Iranian ancestry shows a very high ratio relative to steppe ancestry in all Dravidian speaking groups, and in some cases I reckon all of the West Eurasian admixture looks Neolithic Iranian.
@Davidski
Isn't that more relevant to, say, the Elamo-Dravidian debate?
Potentially yes. But the alternative is also a possibility.
Though as I mentioned earlier, linguistically, there is no strong Sanskrit connection in near east/middle-east other than a few names of gods and kings here and there (which an elite model could explain).
And Sanskrit is key.
@Seinundzeit @Davidski
If you really want to figure out Brahmins, you need some 500 BC samples.
Alternatively, find some South Indian Brahmin samples.
Or even more ideal, find some samples from the South Indian Brahmins who teach the Vedas by oral tradition.
Till then, you don't really know what you're looking at.
@Atriðr
South Indian Brahmins often show almost as much steppe admixture as UP Brahmins.
https://drive.google.com/file/d/0B9o3EYTdM8lQVEJUcDVmOUVNYjQ/view?usp=sharing
@Atriðr
On the contrary, what we're looking at actually seems rather clear.
Long before we had aDNA, it was well known that Brahmins are always more European-shifted, compared to lower caste people in the same region, with something as basic as ADMIXTURE.
I mean, just look at the Uttar Pradesh comparison, the Brahmins model as 25%-30% Bronze Age Steppe-admixed, while the Chamar model as 0%!
This sort of rather drastic difference, similar ratios, probably constitute the pattern in much of India.
The only region where non-Brahmins have more steppe admixture compared to Brahmins is in the Indian Punjab + Haryana. There, the Jatt people have more steppe ancestry compared to Punjabi + Haryanvi Brahmins.
But that's because Jatts obviously have Scythian ancestry, or at any rate something from the steppe which came in the known historical period (Kushans, etc), stuff that clearly postdates the Indo-Aryan genetic influx.
Let's put it this way. It's perfectly reasonable to think that Iran Neolithic and middle eastern admixture could have reached India in many ways that don't involve a linguistic and cultural expansion. In the region there are plenty of similar examples, an obvious one is the seemingly smooth transition from EEF to CHG. Actually, it would be shocking if wasn't the case. On the contrary, it's completely unreasonable to think that european admixture will have ever reach India independently, in addition to that the russian steppe admixture follows a very precise path that was already established by archaeologist and linguist. It would be a very extraordinary coincidence. Also other theories can't really explain why IE ended up in Europe if they were originally from a very "southern" region. It is such and hard task that they came up with an "hybrid model", but IE as an hybrid family is an obvius "ad hoc" argument. The biggest problem is that outside the caucasus and the steppe there is no culture that is even a candidate for PIE, which Iranian material culture was PIE? That's why there is an anatolia hypothesis and not an Iranian hypothesis, but since genetic studies make the agriculture=IE paradigm so ridicoulous that even the proponents switched to an "hybrid model", there is really only one "actual" theory left. If you have a theory that don't make any prediction, it's unfalsifiable, and it's so vague that the places and times can vary by thousand years and thousand of miles, that's not a really a theory. Which ancient iranian culture is the anti-yamna? There is not even an answer to that.
I presume that probably also the Silk Road continuosly brought steppe DNA to India (c. 114 BCE – 1450s CE). The upper castes were surely more affected by this gene flow than lower castes. When that period ended, the Mughals came shortly thereafter, and the dynasty though ethnically Turco-Mongol, was Persianate in terms of culture (Wikipedia). Mughal commercial and social networks must also have brought more steppe DNA to India. Therefore, it would be interesting to see how much steppe ancestry north Indians carried c. 500 BC compared to the present day.
The Silk Road doesn't explain the Kalash, who are a long term linguistic, religious and genetic isolate in the Hindu Kush, let alone groups deep in India like the Tiwaris from east Central India who have a lot of steppe admixture.
@Kriistina
In that time frame eastern europeans and russian were modern european like, hence they had loads of WHG, and it's not the case at all for India. Also yours is a completly ad hoc argument, plus the silk road didn't move a single person from China to Europe or from India to Europe, or from Europe to China, but load of people from an already dead and gone Yamna culture to India. It's just stupid sorry.
Let's think as we wish. I do not mind, and whatever we think, it won't affect the picture as it will emerge in the light of the future samplings.
However, these Simon's recent postings on Anthrogenica are interesting in this respect: http://www.anthrogenica.com/showthread.php?8193-ancient-DNA-in-the-Gedrosia-Near-East-Neolithic-
K13
Now for some Yamnaya samples...
RISE552
Admix Results (sorted):
# Population Percent
1 EHG 73.64
2 IRAN_NEOLITHIC 12.26
3 CHG_EEF 9.55
4 KARITIANA 1.88
5 SHG_WHG 1.13
6 SUB_SAHARAN 0.95
7 POLAR 0.59
I0231
Admix Results (sorted):
# Population Percent
1 EHG 68.86
2 CHG_EEF 14.69
3 IRAN_NEOLITHIC 7.15
4 SHG_WHG 3.83
5 ANCESTRAL_INDIAN 2.27
6 KARITIANA 2
7 PAPUAN 0.91
8 POLAR 0.17
9 SUB_SAHARAN 0.14
I0443
Admix Results (sorted):
# Population Percent
1 EHG 69.15
2 CHG_EEF 15.11
3 IRAN_NEOLITHIC 7.53
4 SHG_WHG 3.95
5 ANCESTRAL_INDIAN 2.15
6 SIBERIAN 0.67
7 KARITIANA 0.65
8 SUB_SAHARAN 0.26
9 POLAR 0.25
10 NATUFIAN 0.2
11 PAPUAN 0.08
They all have some Iran_Neolithic in addition to their CHG_EEF besides their predominant EHG.
Some Corded Ware samples from Germany:
I0103
Admix Results (sorted):
# Population Percent
1 EHG 37.92
2 CHG_EEF 35.38
3 SHG_WHG 13.39
4 IRAN_NEOLITHIC 6.62
5 ANCESTRAL_INDIAN 3.2
6 NATUFIAN 1.88
7 ANATOLIA_NEOLITHIC 0.78
8 POLAR 0.55
9 SUB_SAHARAN 0.16
10 KARITIANA 0.08
11 PAPUAN 0.05
I0104
Admix Results (sorted):
# Population Percent
1 CHG_EEF 44.1
2 EHG 38.52
3 SHG_WHG 9.69
4 ANATOLIA_NEOLITHIC 2.76
5 IRAN_NEOLITHIC 2.67
6 ANCESTRAL_INDIAN 0.84
7 KARITIANA 0.82
8 PAPUAN 0.59
9 NATUFIAN 0.02
Bell Beakers and Corded Ware sample from Estonia did not have any Iran Neolithic. Yamnaya and Corded Ware have even some Ancestral Indian.
Completely irrelevant.
Like I told you, Yamnaya doesn't have any ancestry from Neolithic Iran, and probably not even from Chalcolithic Iran, which was the favored model in Lazaridis et al.
By the way, the R in Neolithic Iran is R2. Not a single Kurgan burial from the Bronze Age has R2 or even J or G.
The links you're seeing are not there.
Well they say that Indo-Europeans were there in N India 2500 BC .
But after dealing with real academics , I am getting the info that they are not suggesting this on the basis of any Indian aDNA . That's the most likely possibility for now . I will be glad if its not the case though.
And those faint lines that they sketch there with question marks , they are actually the hypothetical routes , the green and red ones are the ''established'' ones .
Well they say that Indo-Europeans were there in N India 2500 BC.
And this is based on affinity between South Asians and Neolithic Iranian samples from the Zagros, which is the Elamite homeland.
Comical to say the least.
AND I repeat again that IVC will not be like Paniya in general . No way...
I think IVC people will be like modern Chamar, but without any hints of steppe ancestry.
Dave, You have no idea what the word Geography means it seems , anyway .
I don't think there will be many that deny that CHG/Iran_Chalco works far better that EHG as the PIE component .
You have to put every section of data before making a suggestion . Its not matter of creating one section of data as the most vital . You need a collaboration after verifying them . If they all point towards something similar and more or less creates a pattern , then you have something dependable .
It's going to be a laugh watching how these people try and explain a highly patriarchal culture like the Proto-Indo-Europeans originating in Iran, but with paternal ancestry entirely from the steppes.
I'll enjoy that.
Northern Iran is the area I chose but also had C Asian area attached . There were various reasons for that , anybody who have read the draft knows it .
Anyway, sometime ago I was debating that the CHG type ancestry which penetrated Yamnaya was not farming centric but herding . I think Majkop aDNA will be massive in this regard , since its the closest and nothing beats it . Iran on the other hand had developed pastoralism as early as ~4500 BC.
The Lazardis paper stated that europeans have no ancestry from that zagros neolithic farmers, hence no, Iran has nothing to do with IE. That k13 calculator is a joke, it has a CHG/EEF component, it can't even distinguish EEF from CHG.
@ Nirjhar
Which (material) culture is the "Iranian Yamna", so to say the counterpart to the yamna for the "iranian hypothesis"? Those people didn't have a particular pottery, some kind of burials? Nothing?
@ Kristiina @ NIrjhar007
The Levantinists discussed about the origin of number "7" in IE and its possible link with Semitic. Also our Nirjhar007 caressed that theory, functional to his Euphratic and now to this revival of IE from Middle East.
Alfredo Trombetti, in his Elementi di glottologia, Bologna 1923, not only treated IE and Uralic together, but about the numeral "7" wrote (I apologize for diacritic marks):
"Ostjaco tabet, tlabet per *sabet, Vogulo sat per *sapt, Finn seitsema-n, Samojedo T. sjaibua, Jac. sat-ta ecc.: Indoeur. septe-m, Gotico sibu-n (a. Ted. sibu-n-to per sibu-m-to 7° = Samojedo Jur. sivi-m-d-, Jen. se'odde per *sebo-m-de id.)" (page 152).
See now Starostin:
Number: 1771
Proto: *säptV
English meaning: seven
German meaning: sieben
Khanty (Ostyak): läwǝt (V), jäwǝt (Vj.), tȧpǝt (DN), lȧpǝt (O) 'sieben; Woche', tapǝt (Šerk.) 'sieben'
Mansi (Vogul): sǟt (TJ), sɔ̄̈t (KU), sāt (P So.) 'sieben; Woche'
Hungarian: hét (acc. hetet)
@Rob
I think some scholars do point to Sumerian, Kartvelian loans. Remember, that most languages even in West Asia became extinct, and several langauges expanded during the Bronze & Iron Ages, such as Semitic, probably from southern Mesopotamia. So it would have erased all language diversity and many of the contact languages for which might have once had loans.
So who knows ?
But Sumerian has a few suspected loans and a few suspected words that might have been loaned by PIE. I read an article stating roughly 30. A number could be traced to intermediaries.
That is not much for a language that was a neighbour for so long and preceded a number of the upheavals you mentioned. And remember, compared to the steppe neolithic Iran was a less volatile region.
Again, this is not proof of anything. But I am curious how much is there. Furthermore, even if we consider parts of the proposed Uralic links "voodoo" than there is still a number of words Uralic loaned from PIE rather than any IE derivate (porćas e.g).
And this link at least from Proto-Uralic and Proto-Indo-European:
Number: 1581
Proto: *sisare ~ *sesare ~ *sasare
English meaning: sister
German meaning: Schwester
Finnish: sisar (gen. sisaren) 'Schwester', dial. sisär, siar id., sisarukset 'Geschwister, dial. Kettenfäden, die gleichzeitig an den Zettelrahmen geführt werden'; sisko 'Schwester', siso ( > Est. dial. siso) (möglicherweise Lallformen))
Estonian: sõsar (gen. sõsara, sõsare), dial. sösàr, söhàr, sisàr; sõsaritse (dial.) 'Schwestern untereinander; zwei falsch nebeneinander laufende Fäden im Gewebe'
Mordovian: sazor (E M) 'jüngere Schwester'
Mari (Cheremis): šǝ̑žar (KB), šüžar (U), šužar (B) 'jüngere Schwester', akak šužarak (B) 'Geschwister; Fadenbruch'
Udmurt (Votyak): suzer (S G) 'jüngere Schwester'
Komi (Zyrian): sozor (S) 'Fadenbruch (im Gewebe)', so.zør (PO) 'переслежина (в холсте)'
Addenda: Kar. sizar 'Schwester'; Olon. sisar, sizar, sisär, süzär; Lüd. siza͕ŕ, sizär; Veps. sizar; Wot. sezar; Liv. se̮zar, sizàr, süzàr 'Schwester'
@Kristiina
Hard to believe that the mughals contributed significantly to the ancestry of Indians besides a few elite Mughal families?
@Seinundzeit
Agree that a some amount of iranian Neolithic related ancestry might have been brought by IA from south-central asia but can that account for the large amount of Iran neolithic ancestry in modern Sindhis who occupy the region where several important IVC sites are located.
Also was wondering about the near 0% steppe ancestry in chamar's as per this model. The Lazaridis paper on the other hand seemed to indicate relatively larger amounts of steppe ancestry in lower caste groups. Could the actual value actual be something in between considering that its very unlikely that they remain completely unmixed with neighbouring populations. Strict caste endogamy was a much later practice not observed among the earlier Indo-Aryans
Frank .
Here is a direct link on Indo-European and Afro-Asiatic roots sharing .
https://www.scribd.com/document/223708789/Bomhard-Common-Indo-European-Afroasiatic-Roots-Supplement-I-1986
As I said I will add his references with of other linguists to the work on IE-Sumerian . Which have more amounts of words...
Oh Yes Kristiina , I know you will like the link as well :) .
Italian Jewel ,
I didn't say that the suggestion is the case . I said to you that it is a suggestion and those links you stated actually reflects Uralic borrowing from IE if I am not wrong .
But I will investigate the Semitic seven and IE seven for more clear and confident suggestion.
@Jijnasu,
That makes sense. I'm sure the Indo-Aryans had admixture from the people of the BMAC.
Although, when it comes to the Sindhi people, I guess we have to take into consideration their strong ethnic/tribal/cultural ties to Balochistan. In fact, there is a long history of gene-flow from the Balochistan highlands into contemporary Sindh.
Considering that the Balochistanis are around 70%-75% Neolithic Iranian, the genetic position of Sindhis in the region makes sense.
Also, I'm pretty sure not all Sindhis will be as strongly Iran_N shifted as the samples we've had since the HGDP.
Sindh is very diverse, I'm sure many Sindhi communities will turn out like those Lahori Punjabi people, who can be modeled as 60+% Paniya. The Mohanna people, on the Indus river, come to mind.
Having been to Sindh, the people look very diverse, depending on their geographic origins in Sindh + their tribal/caste affiliation. The phenotypic diversity probably maps unto actual genetic differentiation among the people living in the region.
And I think you make a good point. The modelling I tried has UP Brahmins at 25%-30% steppe-admixed and UP Chamars at 0%. But if, due to some confounding factors, the steppe ancestry is being underestimated in UP Brahmins with this method, I'm sure the Chamar would be around 10% steppe-admixed, which would make more sense.
In fact, if the Chamar have steppe admixture, then the Brahmins are probably more like 40% steppe-admixed, while Pashtuns/Kalash are more around 50%, while the Pamiri are more around 60%-65%.
Honestly, I think this would be much more reasonable.
Again, with the nMonte fits, Pashtuns/Kalash were at 40% steppe-admixed, and the Pamiri were at 50%-55%, so the Chamar get 0% for steppe ancestry. Since it's unlikely that the Chamar are 0% steppe-admixed, I'm sure nMonte is underestimating the steppe admixture in South Central Asians and Upper Caste North Indians as well. TreeMix and QpAdm give much higher percentages than 40%-55% for the former, and 25%-30% for the latter.
Jaydeep,
Good points . But I do not endorse OIT for PIE now . If something that comes up backing it , I will endorse .
About Cattle issue , they can check yamnaya cows for Zebu admixture . To see if there something interesting turns up .
Jijnasu, muslim rulers tended to have a harem. :-) In reality, I did not mean only family relationships but also commercial relationships between traders, and if rulers used mercenaries, they were often foreigners as Hazara who came from the steppe.
Gioiello, by way of curiosity, I note that the Yakut word for 'woman' is daxtar (дьахтар).
As for the Uralic question, the point is to find out if the shared words with Uralic are in the Indo-Iranian-Balto-Slavic-Germanic branch excluding Anatolian, Greek and Celtic (and Albanian) languages, as this new model predicts. However, the problem is that there are probably words that are substrate words that did not belong to IE protolanguage or to Uralic protolanguage, but however exist in some daughter languages of both groups.
As for Kalash, U2e1 and U4a1 account for 50% of Kalash mtDNA and they are old steppe haplogroups and detected already in the Mesolithic Western Siberia. With all probability, they did not arrive from Europe as they are not under any European subgroups.
Kristiina,
It seems caxtar . Or I am making a mistake?.
https://en.wiktionary.org/wiki/%D0%B4%D1%8C%D0%B0%D1%85%D1%82%D0%B0%D1%80
http://www.engyes.com/en/dic-content/%D0%B4%D1%8C%D0%B0%D1%85%D1%82%D0%B0%D1%80
Kristiina,
The Kalasha are an interesting case.
Unlike Pashtuns or North Indian Brahmins, R1a1a doesn't dominate (although, obviously, they still have a substantial amount).
Yet, unlike Pashtuns or North Indian Brahmins, they show some very serious steppe connections via their mtDNA.
One wonders if this is due to the Kalasha perhaps having substantial maternal ancestry from steppe populations (less male-biased migration, perhaps more of a whole folk-movement dynamic), while 40%-50% of Pashtun autosomal ancestry came via mostly male-biased gene-flow from later steppe populations, or whether this just boils down to drift/founder effects/etc?
I'm thinking the latter, but who knows, the former could be possible.
@Kristiina
I presume that probably also the Silk Road continuosly brought steppe DNA to India (c. 114 BCE – 1450s CE). The upper castes were surely more affected by this gene flow than lower castes. When that period ended, the Mughals came shortly thereafter, and the dynasty though ethnically Turco-Mongol, was Persianate in terms of culture (Wikipedia). Mughal commercial and social networks must also have brought more steppe DNA to India. Therefore, it would be interesting to see how much steppe ancestry north Indians carried c. 500 BC compared to the present day.
There is nothing in the historical record that suggests that the upper castes of India were more affected by the Muslim or other post-Indo-Aryan gene flows to India than the lower castes. Most of the post-Indo-Aryan invasions probably did not have a significant effect on the genetic structure of India.
Gioiello, by way of curiosity, I note that the Yakut word for 'woman' is daxtar (дьахтар).
дь represents the 'gy' phoneme, not 'd', in Yakut:
https://en.wikipedia.org/wiki/Yakut_language#Writing_system
In wikipedia, the Yakut sound is explained as voiced palatal stop ɟ̠ . The cyrillic letter 'ь' probably means that 'd' is palatalized as in the following Slovak example:
Slovak ďaleký [ˈɟ̟äɫe̞kiː] far
Yakut дьахтар [ɟ̟axtar] woman
Onur, I do not insist. This is not important for me, but if a Moghul ruler was c. 50% steppe and had 30 surviving children who got a good position in the society and had a lot of offspring, this 50% was not insignificant in the ruling class and its surroundings. And probably it was not only one Moghul man but also other male relatives got their fare share.
I still think that upper classes are more international than lower classes.
@Jijnasu
Broushaki 2016 has similar low steppe results for South Asians, everything more South Indian-shifted than GujaratiB (including their Lahore Punjabi sample) needs nothing but Ust-Ishim (presumably an ASI proxy) and Iran_N. This is probably confounded somewhat by lack of a high coverage EHG sample which forces CHG, which doesn't appear in any Indian or Pakistani population, and WHG proxy for European steppe (Iran_N doesn't appear in North and East Europe).
In any case the Lahore Punjabi sample's results based on ancient donors in that paper are such that it could be fitted as about 70% Mala 30% Baloch.
@Kristiina
Onur, I do not insist. This is not important for me, but if a Moghul ruler was c. 50% steppe and had 30 surviving children who got a good position in the society and had a lot of offspring, this 50% was not insignificant in the ruling class and its surroundings. And probably it was not only one Moghul man but also other male relatives got their fare share.
I still think that upper classes are more international than lower classes.
The upper classes of the Hindu caste system did not mix with the Muslim invaders unless they converted to Islam, which would take them out of the caste system.
@ Onur
"Gioiello, by way of curiosity, I note that the Yakut word for 'woman' is daxtar (дьахтар).
дь represents the 'gy' phoneme, not 'd', in Yakut:
https://en.wikipedia.org/wiki/Yakut_language#Writing_system"
@ Kristiina
"In wikipedia, the Yakut sound is explained as voiced palatal stop ɟ̠ . The cyrillic letter 'ь' probably means that 'd' is palatalized as in the following Slovak example:
Slovak ďaleký [ˈɟ̟äɫe̞kiː] far
Yakut дьахтар [ɟ̟axtar] woman"
Thus the "c" is the Turkish one. But it seems to me that this case is completely different from mine about the word for "sister". This case seems isolate, and it would be anyway interesting to understand from which IE language Yakuts took their word (if it is really the IE word). The IE word for "sister", if it is a loanward from IE, entered very likely long time ago, before the separation of the Uralic languages and very likely also IE was an unique language, and that has to have happened only in the Siberian corridor. Of course the links between IE and Uralic are at the deepest level not only of vocabulary but of the same morphology.
@ Kristiina
"As for Kalash, U2e1 and U4a1 account for 50% of Kalash mtDNA and they are old steppe haplogroups and detected already in the Mesolithic Western Siberia. With all probability, they did not arrive from Europe as they are not under any European subgroups".
I wouldn't be so sure that Italy hasn't those hgs. Many years ago Palanichamy asked me to test an Italian U2e1 I found on SMGF. Unfortunately those samples were anonymous, and now SMGF is completely out. I have found also U4a1 in Italy, and very old, not come recently, and Italy is low tested.
@Seinundzeit
On the contrary, what we're looking at actually seems rather clear.
You missed the point then precisely go on to enumerate my point.
On the other hand, @Davidski reminded of South Indian samples, which further minimizes Iran_N being I.E., which was no surprise to begin with, but still doesn't see the value of pre-Scythian movements into the subcontinent.
As for the Kalash, they are a third branch of I-I, and have some elements of Vedic worship in their belief system. This is where positions should start, from them.
Gioiello, sorry that I was not clear with that word. It is not related to the word sister! I added it by way of curiosity because it looks like being another IE root, daughter, *dhuga-t-er-.
Onur, sure. However, I had a look at the admixture graph here: https://drive.google.com/file/d/0B9o3EYTdM8lQVWFIOXR0c1ZTazg/view
Punjabi persons seem to have very little light blue Steppe component, maybe 5-7%. If the Indo-Aryans according to this new model were like MLBA Armenians, part of the Punjabi Steppe component would come from there. Then, we need very little, if any, steppe extra to get the modern levels. However, all is still very uncertain and let’s wait for the paper to be published.
Using the ADMIX13Q spreadsheet, and the same four populations (Iran_N, Steppe, Onge, Han) and 4mix, the values for South Asian populations are like this:
Balochi - Iran_N 0.6, Yamnaya 0.21, Onge 0.08, Han 0.11
Bengali - Iran_N 0.37, Yamnaya 0.22, Onge 0.17, Han 0.24
Brahmin - Iran_N 0.38, Yamnaya 0.29, Onge 0.14, Han 0.19
Brahui - Iran_N 0.61, Yamnaya 0.2, Onge 0.08, Han 0.11
Burusho - Iran_N 0.42, Yamnaya 0.28, Onge 0.09, Han 0.21
GujaratiA - Iran_N 0.42, Yamnaya 0.3, Onge 0.12, Han 0.16
GujaratiB - Iran_N 0.4, Yamnaya 0.28, Onge 0.14, Han 0.18
GujaratiC - Iran_N 0.41, Yamnaya 0.24, Onge 0.15, Han 0.2
GujaratiD - Iran_N 0.4, Yamnaya 0.22, Onge 0.17, Han 0.21
Kalash - Iran_N 0.37, Yamnaya 0.4, Onge 0.05, Han 0.18
Kusunda - Iran_N 0.13, Yamnaya 0.1, Onge 0.14, Han 0.63
Makrani - Iran_N 0.63, Yamnaya 0.19, Onge 0.07, Han 0.11
Pathan - Iran_N 0.48, Yamnaya 0.29, Onge 0.1, Han 0.13
Punjabi - Iran_N 0.35, Yamnaya 0.24, Onge 0.18, Han 0.23
Sindhi - Iran_N 0.49, Yamnaya 0.25, Onge 0.11, Han 0.15
(I haven't included above how well these fit).
Cross plot of the values with Lazaridis 2016:
http://i.imgur.com/NpP7QEi.png
The Paniya like groups didn't fully disappear from modern day Pakistan. Just because Pashtuns and Baloch aren't super high in Paniya doesn't mean that Paniya populations don't exist in Pakistan(there are many theories that those 2 groups are recent in the Pakistan area anyways). If you look at the Punjabi Lahore samples for example, they are 50-55% south Indian(Paniya like) on average. I'm pretty sure many " low castes " that form that majority in many Pakistani regions are gonna score high south Indian as well. even groups like Pakistani jatts get 30-40% South Indian (Paniya like). Also, most IVC major centers were pretty far from modern Pashtun areas.
I am staying away from this echo chamber, and let’s see if this comment also gets deleted.
But.. .Couldn’t help to come here and say this: Why does the slide shows 6900 years? why not 7000?
6900 ybp is the date when Mentesh Tepe fell... and the Shulaveri started to disappear from the south Caucasus and some, many, flee to ansueli and then up the black sea shores to Steppe via places such as the Kuban river. To me 4900 Bc is when the serpent people (ophideans) came and created the first mass human dispersal that I know of…
How much CHG those guys must have had after 2000 years in the land of Kotias and Satsurblia?
Just wondering why he wrote 6900 …. :-=)
Well let's see whether you can figure out that puzzle. Here's a clue...
http://eurogenes.blogspot.com.au/2015/11/the-khvalynsk-men.html
@ Kristiina
"Gioiello, sorry that I was not clear with that word. It is not related to the word sister! I added it by way of curiosity because it looks like being another IE root, daughter, *dhuga-t-er-".
How may you say that Uralic *sisare/*sesare/*sasare isn't from IE *swesor?
There is an Uralic word from Nostratic linked:
Eurasiatic: *SäwCV
Meaning: sister, woman?
Borean: Borean
Indo-European: *su̯esor
Uralic: *säće
Kartvelian: *ćoź-
@Matt,
Using that same ADMIX13Q spreadsheet, with nMonte, I get these kinds of models for the Pashtuns/Kalasha, when I get more time I could also try some South Asian populations:
A Pashtun sample
39.48% Steppe (mostly Scythian_IA + Andronovo)
39.00% Iran_Neolithic
14.85% ENA (Andamanese + Ami + Papuan)
5.70% Bedouin
1.00% MA1
A Kalash sample:
39.55% Steppe (Scythian_IA + Andronovo)
27.85% Iran_Neolithic
12.35% MA1
10.80% Bedouin_B
9.35% ENA (Ami + Papuan)
Despite being based on the same data, the results look quite different from the 4-mix output.
Basically, I don't think the 4-mix method is a good choice, just too limited. By contrast, this output broadly agrees with the PCA + d-stat based nMonte results.
I suppose the only advantage with 4-mix would be speed, when it comes to examining multiple populations?
Regardless, I tried an nMonte fit with similar reference populations to what you tried, and again using the same ADMIX13Q spreadsheet, just for fun:
A Kalash sample
34.70% Sintashta
30.05% Neolithic Iranian
11.85% BedouinB
11.15% MA1
7.95% Ami
4.30% Papuan
0% Yamnaya
0% Onge
Again, quite unlike the 4-mix fits, although still very odd (it's that Bedouin component).
Truth be told, I don't think it's wise to use even nMonte with ADMIXTURE results, if we are trying to look at ancient ancestry in living populations. The very nature of ADMIXTURE clusters, and the manner in which ADMIXTURE works, militates against this.
With PCA data, it just seems a more natural/accurate way of modelling populations, and the same goes for d-stats. The fact that nMonte output agrees between the PCA and d-stats sheets is very encouraging. For what it's worth, I posted those models, with the Paniya, further up-thread. All the fits were extremely good, looking at the distances.
Of course, the assumption in those models is that a putative IVC sample is genetically similar to the modern Paniya, which is very far from certain.
@Shaikorth
Such a marked difference between ethnic groups from the same area would by surprising. I am not sure how good a proxy for the ASI the Mala are though. I think this discussion will be more meaningful once the actual south asian DNA comes out
@Atriðr
The Kalash are Indo-Aryan speaking. They speak a Dardic language. However they resemble the Nuristanis closely culturally and may have switched languages at an early period
@ Sein, I guess, you could run models with different populations as well, my interest there was really mainly using the comparable populations from Lazaridis's model. Including other populations like late BA and Iron Age steppe than Steppe_EMBA are waaaay less transparent if I'm trying to compare to what Lazaridis modeled (model how much ancestry is early steppe or EHG).
Some model with >4 populations would no doubt be more optimal. I do think ADMIXTURE is particularly worse for nMonte or 4mix fits when we get to very high dimension. Hard for me to compare the models I've run to yours since you're looking at individual samples and I've grabbed the population means. Only thing I would say is MA1 would be really unsuitable in any model using ADMIXTURE.
@ All - Has anyone mentioned or chatted about this conference by Max Planck tomorrow - http://www.shh.mpg.de/105110/lag_conference? I've missed it upthread if so.
lol that was last year. I have poor reading comprehension...
@Matt
For what it's worth, if I restrict it to only the populations that were used in Lazaridis, with the same ADMIX13Q spreadsheet, I get an extremely poor fit, the distance is just too bloated. It's just too bad of a fit, in order to correspond to the reality of how these populations actually stack up.
Which I guess brings us back to the very problematic nature of using high K ADMIXTURE output with nMonte (again, PCA + d-stat data is way better for this), the fact that 4-mix is even worse for this case, and the fact that those four references just aren't adequate at all.
But getting back to that graph, I wish someone could say with certainty whether they based it on an actual IVC sample, or not.
@Jijnasu
There is definitely a drastic difference between PJL and Sindhi, Baloch and Brahui are even more distant. It's probably safe to assume that using modern populations provides a preferred model in this case, and then the tables have among others a fit with over 60% East Bengali, rest mostly Sindhi, for the Lahore Punjabi sample.
According to Moorjani and Reich papers, Mala were supposed to be 60-65% ASI. Paniya 78-83% though that was with modern proxies, not Iran_N etc. The Bangladeshi sample is less ASI than Mala but has more Southeast Asian.
@Jijnasu
Yes, you are correct. Kalash is Indo-Aryan. I was thinking Nuristani, which is the third I-I branch.
But yes, Nuristani and Kalash are basically same group. Nuristani is a recent, made-up ethnonym as we both likely know.
@Kristiina
Whether the authors know this or not, their model hangs on two assumptions.
- that R-M417 and R-M269 are not EHG markers but arrived on the steppe from Iran
- that steppe admixture in South Asia dates to the historic period.
Both of these assumptions are wrong, and everything else is background noise.
@Jijnasu its not surprising to see differences along caste/tribe lines in same region in south asia. If you check 20 individual Pathan HGDP samples results from FATA/Pakistan then one can see their Paniya like admixture in 4-5 samples is as high as 35-36% and in others as low as 20%. This is despite the fact that samples were taken from one ethnic group from same region which is quite isolated area of FATA. Same is the case with Sindhi HGDP samples, some are more Paniya shifted then others. I guess Sindhis with low Paniya are likely Baloch admixed. While the reason for pathans with high Paniya could be because they used to pick up low caste slaves from south asia and go back to mountains.
It's quite clear to me that the Yamnaya culture is only a branch in the history of the Indo-European language expansion, and not its homeland. The ANE component of Yamnaya was probably introduced to the PIE from their Near Eastern component. It isn't correct to assume that where a language goes, it must have been spread by people of the same ancestry.
The ANE component of Yamnaya was probably introduced to the PIE from their Near Eastern component.
No, Yamnaya got its ANE from both it's Eastern European HG and Caucasus HG ancestors, but more from EHG, which has a much higher ratio of ANE than CHG does.
More importantly, Y-DNA R1 and its young subclades are associated with EHG and rapid expansions from the steppe during the Bronze Age that shaped the genetic and cultural landscape of Europe and much of Asia.
These were the Proto-Indo-Europeans.
@ Epoch
" Furthermore, even if we consider parts of the proposed Uralic links "voodoo" than there is still a number of words Uralic loaned from PIE rather than any IE derivate (porćas e.g).'
I take your point, but these are a very small corpus, which could have been mediated by early derived forms. I.e. early Indo-Aryan could still have harboured proto-IE forms. Sound changes spread unevenly and in a protracted manner even within one language community. This is obvious, eg, if you study early Slavic inscriptions and names from East Roman sources, and the treatment of the assibilation of the metathasis.
@ Kristiina
I agree. We need some pre-Antiquity aDNA from south asia, to remove any potential confounding effect from Turkic admixture.
Kristiina
"The Khvalynsk R1b was covered with metal objects and @@@metallurgy@@@ came from the south. I quote your own text:
”Male (confirmed genetically), age 20-30, positioned on his back with raised knees, with 293 copper artifacts, @@@mostly beads@@@, amounting to 80% of the copper objects in the combined cemeteries of Khvalynsk I and II."
Nuggets of copper (like gold) can be found naturally in copper-rich regions in rivers etc and it's soft enough to cold hammer into simple shapes like beads.
So smelting etc may well have started further south but the Khvalynsk dude doesn't require any of that to have what he had - just needs a few villages in the giant Kargaly copper field specializing in making and trading copper beads.
Dave,
Its only a matter of time . By the end of this year I presume .
Kristiina ,
Thanks for the clarification .
BTW all,
Did any of you look at the afro-asiatic root sharing link I provided?.
Grey
The copper in the Khvalynsk Chief's grave is of Balkan provenance, not Urals
@ Nirjhar007
"Did any of you look at the afro-asiatic root sharing link I provided?".
I'd be glad to do, but the paper wasn't for free, anyway I wasn't able to download it.
Dear Nirjhar007, I wasn't able to download your article of Bomhardt, but it seems also old, thus I read something more recent as Gàbor Takàcs, A. Dolgopolsky's Nostratic Dictionary and Afro-Asiatic (Semito-Hamitic), and I was going to say, after having read some etymologies, that this people is accustomed to write Bibles, that, not being verifiable nor falsifiable, are in what in my language I'd claim a "puttanata", but after I downloaded the Dolgopolsky's Dictionary of Nostratic, and, in respect of the monstrous work he and all his colleagues did, I am waiting for studying it more deeply.
I found only this about your "seven": "for S *-t- ı K *-d- cf. S *s7ab÷at- 'seven' ı K *s7wid- 'seven', w K2 251 ¶ KB 48O, 1728, KBR 5O5, GB 368, Js. 627, Dlm. 2O1, Sl. 567, Br. 352, Sd. 518, CAD VIII 61O, K 1O7, K2 87-8, FS K 166-7, FS E 181-2, LH 211 ¶ ".
Anyway I would want to say that in genetics someone may do some hypothesis, but after there has to come the "proof", and remember that I was the unique to announce Villabruna.
Wonder what R1b(V88) could tell us about some similarities between Afro Asiatic and IE ? MtDNA Haplogroup U6 was found in Romania very early and is now common in the Sahara just like R1b(V88).The oldest R1b sofar in Villabruna . Oldest MtDNA U6 in Romania . What is between the two? The Balkan.....
Anyway. Do you remember the famous proverb from the Bible: "be-leb yam", in the heart of the Sea/in the middle of the Sea. Thus
1257. * * * *L L L Li i i ib b b bV V V V 'heart' ([in descendant lgs.] ‘ 'life', 'middle')
would be linked with
IE: NaIE {P} *lei\b≈- or {WH} *lei\b[≈]- 'live'
Thus by using ad libitum the "signifiant" or the "signifié" a link may always be found.
Any nostalgia for the great Ferdinand de Saussure?
"Wonder what R1b(V88) could tell us about some similarities between Afro Asiatic and IE ? MtDNA Haplogroup U6 was found in Romania very early and is now common in the Sahara just like R1b(V88).The oldest R1b sofar in Villabruna . Oldest MtDNA U6 in Romania . What is between the two? The Balkan..."
Even though you probably know that I don't like unknown people, I answer you:
U6* was found in Romania more than 30000 years ago, and that disproves another "puttanata" of the Levantinists that U6 wasn't European (whereas pretty all the other hg. U were) but from Middle East.
R-V88 comes certainly from the Italian Refugium of the Villabrunas and expanded to Sahara not before 7000 years ago, thus I think that there is no link about their expansion.
Gioiello yes but this points to two probable migrations from the Circuit Balkan area with the same destination. The Sahara Subpluvial started +-9000ybp and cattle were domesticated not long before that in the Taurus Mountains through which R1b must have migrated into the Lavant and eventually the Sahara...Could R1b(V88) have spoken a very early Proto-Proto-IE which could have influenced the formation of Afro Asiatic?
@ Ric Hern
Of course I think that dates are important, both for genetics and for linguistics. That IE evolved amongst the Villabrunas is just my theory once more against all.
We have to understand at which phase the migrations happened. I have been against the kurganists and the theory that everything about R1b and IE did come from Southern Russia, but I think that the above all R-L23 of Samara came from West and the IE spoken there was a satem IE, but that centum expanded from Italy or nearby (of course the Balkans are part of all these facts).
Which language did the Villabruna bring wity them when expanded to Anatolia or Middle East? May we think that Hittite was their language? And the R-V88 and others who migrated to Iberia 7500 years ago (see Zilhao) which language did speak? Of course the links between IE and Afro-Asiatic (if there are) are so old that we cannot think that a few people of 7000 years ago may have influenced it, if not for some loanword. There are words so diffused that witness these links (and of course may be brought from other groups into Europe too).
I am glad when I may do an exact link as amongst F38, the R-L584 from Samara, the Visigoth Grijalba and the Jewish family of the Spyras.
@ Kristiina
Not only, but I'd say that both for Uralic *sisare/*sesare/*sasare and for Yakut дьахтар we could hypothesize that there has to have been an extict branch of IE expanded to Central Asia from which the loanwords came. It could be linked to Sanskrit for the derivative -a- of all the IE vowels and for the uncertain one of IE *swe-.
I think that Sanskrit and Indian IE came from the Easternmost expansion of the kurgans before going to South.
Gioiello no I don't think R1b(V88)s language was Hittites. Hittites look like originating from the Ezero Culture. Looking at the Neolithic Anatolian DNA we see no Haplogroup R anywhere so I think there was a displacement most probably from the direction of the Caucasus which pushed R1b(V88) Southwards into the Levant and Westwards from where their forefathers came and this is how those R1b(V88) landed up in Spain..The R1b class that did not migrate Eastwards earlier migrated Northeast into the Steppe hugging the Black Sea coast and Rivers with some coming into contact with the Southern Urals Yangelskaya Culture which show a striking resemblance to the Zarzian Hunters of the Southern Caspian who probably was the CHG contributors....
Could the CHG component in Yamna have come from the Yangelskaya Culture of the Southern Urals who show remarkable similarities to the Zarzian Hunters of the Southern Caspian?
@ Ric Hern
1) I didn't speak of the Hittite as linked to hg. R-V88, but to some hg. coming from Villabruna after the Younger Dryas to Anatolia, and the Villabrunas weren't only hg. R, but there are in Northern Anatolia many hgs. that may have come from Villabruna (both Y and mt).
2) "so I think there was a displacement most probably from the direction of the Caucasus which pushed R1b(V88) Southwards into the Levant and Westwards from where their forefathers came and this is how those R1b(V88) landed up in Spain..."
That R-V88 came from Middle East is the undemonstrated and now disproved theory of the Levantinists. R-V88 is enterely from Italy both for the oldest subclades (R-V88*, R-M18, R-V35) and for the sample found in Iberia 7100 years ago.
3) All what you say is undemonstrated. We have Villabruna 14000 years ago and the samples around the Caucasus at least 7000 years later. Show me the samples and I'll believe you.
Could the CHG component in Yamna have come from the Yangelskaya Culture of the Southern Urals who show remarkable similarities to the Zarzian Hunters of the Southern Caspian?
Yeah, maybe. My choice for now is the North Caucasus and the eastern end of the Black Sea. There are cemeteries in this region featuring skeletons with exaggerated North Europid and South Europid features. Maybe they represent mostly EHG and CHG people?
But I won't be perturbed if the CHG-related ancestry moved into the steppe via the southern Urals.
I also think Yamnaya and Corded Ware have some Balkan farmer ancestry. Hopefully Mathieson tested this angle for his upcoming ASHG presentation and new paper.
Thank you.
According to D-stats Yamnaya is definitely not part IranNeolithic or IranChalcolithic. Neither can explain Yamnaya's affinity to CHG, both give too high of estimates in relation to themselves, and they miss represent Yamanya's relation to other outgroups.
@Samuel Andrews
According to Lazaridis 2016, Chalcolithic Iran+EHG is the best two-way fit for Steppe_EMBA when Ust_Ishim, Kostenki14, MA1, Han, Papuan, Onge, Chukchi, Karitiana, Mbuti are the outgroups and the only decent two-way fit when Anatolia_N, Levant_N, Natufians, and WHG are added to outgroups.
However, Yamnaya isn't necessarily a two-way fit (these papers have a tendency of going with the simplest statistically fitting models) which brings CHG back into play. If we had >10x coverage steppe sample, EHG sample and a Chalcolitic Iranian sample we could do a haplotype model and resolve it immediately, as certain fits allowed by tests based on unlinked SNP's are ruled out then. That modern North&East Europe is best fitted as WHG+CHG without any Iran_N may already rule Iran_N type pops out even with our limited samples, Iran_CHL is another matter since it can be modeled with a major CHG component.
Gioiello no I said West wards from where their forefathers came. In other words the Balkans,Italy etc. And I mentioned the Taurus Mountains near where cattle were domesticated not the Caucasus. I refer to the Caucasus as an origin of other peoples that could have pushed R1b(V88)from Anatolia into the Levant. Mesolithic DNA from Anatolia will come in handy...
@Rob
I agree. We need some pre-Antiquity aDNA from south asia, to remove any potential confounding effect from Turkic admixture.
Yes, the point I was making above to Seinundzeit.
Mostly to date Scythian ancestry (Turkic as term relatively recent).
And the main reason I mentioned 500 BC Brahmins, was for the Y, not the aDNA.
But linguistically, we know the answer.
I'm thinking you have the right suspicions as well.
@ Ric Hern
"Gioiello no I said West wards from where their forefathers came. In other words the Balkans,Italy etc. And I mentioned the Taurus Mountains near where cattle were domesticated not the Caucasus. I refer to the Caucasus as an origin of other peoples that could have pushed R1b(V88)from Anatolia into the Levant. Mesolithic DNA from Anatolia will come in handy..."
I understood your post, and certainly people went and gone in all times, what my opponents denied when I said that, because only Middle East was the "lux" (Ex Oriente lux). But this time I think we don't need to put Middle East in the play again: from aDNA there is no R1b there, neither in Anatolia, less in Natufians and Iran. R-V88 went clearly from Sardinia/Italy to Iberia 7500 years ago with the migration hypothesized from Zilhao. To Sahara may have gone from Iberia or directly from Italy. Don't forget the paper on the "Tuderela sulcata", which demonstrated this pathway already many years ago.
About the few R-V88-M18 hypothesized in Lebanon, we'll see if they will be tested and submitted to YFull: don't forget that Shardana were mercenaries during the second millennium BC in Middle East and Egypt, and, as the mt U5b found in a Carthaginian demonstrated, it isn't said that Phoenicians gave their DNA: it is more likley that they received it (in spite of whom is paying for that).
Could be. Some Tenerian and Kiffian Y-DNA from the Sahara Subpluvial will surely be welcome.
@Gioiello
The Lebanese M18 is SNP tested, which is not what I'd call hypothetical. Do you think for some reason that they are an error, the haplotypes better match something else? (I don't have any opinion about the origin of V88, I am just curious.)
Do we have any ancient DNA from early speakers of Anatolian IE languages? That would be the easiest way to resolve things I'd think. If they have significant EHG admixture then Anatolian hypothesis would be even more dead.
I haven't engaged fully in the discussion about "Race" brought up in the previous threads, which emerged in response to a a rather "dry" comment about principal components in the Ashkenazi Jewish population and whether Tuscans or another group are the best proxies for their southern European ancestry. Whether this was "OT" is irrelevant now, since it seemed to be a key point for a lot of people here. It certainly became "On Topic" without any effort or intent on my part, and as they say "You broke it, you own it."
I think that many of you, and The "Generalissimo" ;) may want to comment on this article on Vice News about the "Alt-Right", DNA testing, ancestry heritage "percentages" and White Nationalism:
White Nonsense:
Alt-right trolls are arguing over genetic tests they think “prove” their whiteness
While this blog is certainly not The Apricity Forum or Forum Biodiversity (a reference to the "Human Biodiversity" / HBD movement), this blog in particular provides much of the (increasingly accurate!) raw data analysis that informs these discussions, and that's why these issues emerge here.
A quote:
"TheApricity.com is not explicitly political but rather a “European cultural community.” Like on many older message boards, members can list age, gender, and location, as well as several ways of describing their ethnicity, various DNA markers, and supposed ethnic phenotype. Not every member is white, and not every post implies white supremacy. But there are countless threads classifying people, usually women, into various European phenotypes, and many about whether a particular person or ethnic group could “pass” as white."
@Shaikorth,
The analysis by Lazardis 2016 isn't the best type of analysis. D-stats with better outgroups are a better analysis. They irrefutable say Yamnaya's relation to CHG is much too close for it to be of significant IranNeolithic or IranChalcolithic. The debate is over.
@Samuel Andrews
What outgroups and D-stats? If this is nmonte I wouldn't go putting it over their methods very lightly.
@ capra internetensis
"@Gioiello
The Lebanese M18 is SNP tested, which is not what I'd call hypothetical. Do you think for some reason that they are an error, the haplotypes better match something else? (I don't have any opinion about the origin of V88, I am just curious.)"
My hypothesis was referred to the origin of those R-V88-M18 (I searched for their haplotypes already many years ago but I didn't find anything), if old in Middle East or introgressed from the Sardinian ones, for that I asked that they were tested and submitted to YFull. But Italy has much more:
1) the oldest R-V88*, found in the Isles (Sexton/Saxton) and in Italy. I remember you that Marchesi was tested from FTDNA as M269 and only after many letters of mine to the administrator of the "Italy FTYDNA project" was recognized as R-V88 "in base to his haplotype" (under my suggestion) and his SNP changed "d'amblée" from M269 to M343.
2) R-V88-M18
3) R-V88-M35
4) The oldest aDNA of R-V88 in Iberia I think come from Italy etc.
@Atriðr
Linguistics certainly do match genetics. Have a look at the map at the bottom of this page.; it matches the DNA substructures in modern South Asians, even though the author knows nothing about them.
https://sites.google.com/a/sudiptodas.com/www/thearyantrail
@Gioiello
Both Italy and the Middle East have M18 and Y7777, so that doesn't swing it one way or other.
The only Y haplotypes I know of for the Lebanese M18 are from Zalloua et al 2008, "Y-Chromosomal diversity in Lebanon is structured by recent historical events". There are 3 samples, listed as R1b1a in the supplemental table. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2427286/
The Italian ones I know of are from Contu et al 2008, "Y-Chromosome based evidence for pre-Neolithic origin of the genetically homogeneous but diverse Sardinian population: inference for association scans". There are 8 samples falling into 4 haplotypes. 7 belong to one cluster, the other to a quite different lineage. In Francalacci et al 15/18 M18 samples belonged to one branch and 3/18 to the other, so it makes sense that there is a major branch and a minor one in Sardinia.
I am no good with STRs but neither seems clearly derived from the other.
@Kristiina, Atriðr et al
It's pretty easy to figure out Turkic ancestry is quite insignificant in India, even in groups like Uttar Pradesh Muslims.
Moorjani et al. 2013 didn't find Turkic-looking outliers using PCA, so either Turkic ancestry is the same everywhere in India (hardly likely) or it isn't there. In Broushaki et al. 2016 we have several Indian populations and can see this too.
Uzbeks are modeled as 21% Loschbour, 41% Iran_N 35% Han rest self-copy
Turkmens as 2% Kotias 19% Loschbour 56% iran_N 14% Han rest self-copy
Now an Indian population hardly likely to have Turkic influence from Tamil Nadu is fitted as 38% Ust-Ishim 62% Iran_N. A combination of just these two populations appears for Ksharityas and Brahmins from Uttar Pradesh, Meenas, Meghawals and even Uttar Pradesh Muslims in varying proportions. For Scythians we probably need ancient DNA to be sure but the Turkic case looks clear.
It would be very hard for the recent medieval turkic migrants to integrate seamlessly in such a fiercely endogamous place in india. Every individual, clan was spoken for and caste politics always ran deep.
Only "naive foreigners" can hold these views.
Mughal descendents are miniscule and well known by names like chugtai. There is a large wannabe/fake turko-persian population who adopted generic names islamic names, seeking upward mobility, avoiding pagan tax and persecution by converting to islam. Among this massive population a few genuine turkics might be found.
The other factor is uzbegs and their immediate neighbouring south asians were not highly differentiated to start with.
You can't polish a turd.., that is all
@OpenGenomes:
This is my favorite excerpt from that article, concerning one of the wonderful white supremacists:
"Promethease showed he had a gene that limits oxytocin, a hormone that helps humans bond, thought to have evolved among humans in Northern Europe, where the mortality rate was high. It made perfect sense to Mylott. That’s why he “can very easily disengage from relationships if they’re not going well, and the next day be whistling like nothing happened,” he said. “The only reason I have this characteristic, which has affected until this day in 2016 my relationships, is because for thousands of years my ancestors lived in an environment that was covered in ice!”
When ASD continues to sky-rocket, they will *own* that "happily" as well...
Anyway, so what that lots of people are racist? There are lots of non-white racists too who are extremely attached to the endogamy that supposedly created them.
It's still better that more people know what the truth is data, that data and analyses are available, etc.. Otherwise, theories will still exist, but just backed by crappier data instead.
BTW: If I a person is a "hybrid-vigor supremacist", can that make one racist too (I hope not! j/k)?
Have you thought about starting an "anthro-racism-watch" kind of site yourself?
@Davidski
I was expecting your link to be to a linguistics paper.
I'm not sure what you were intending with the link. It seems to cater to OIT proponents which is certainly not my position; tbh, I hardly read it.
As all my previous commnets, I've always proposed a much closer connection between the Balto-Slavic and I-I branches that currently popular models.
The main issues are not between those branches.
@Shaikorth
I mentioned Scythians (all groups), not Turkics.
@Atriðr
The map at the bottom of the page outlining the main Indo-European migrations into South Asia matches the genetics of South Asia as per the latest findings based on ancient DNA from the steppes and Iran.
And this map is based on language and historical data, not DNA.
@Davidski
Several months ago, I mentioned elsewhere in the comments of your blog, a similar path.
I do have certain differences though.
And these days, I have even more differences.
Specific to the map of the link you sent, the dates don't match with key variables - Lazaridis Steppe EMBA being one of them.
How can Lazaridis' Steppe_EMBA not match when it includes samples from the EBA, MBA and LBA steppe, from Afanasievo, Yamnaya, Potapovka and Srubnaya?
In what way doesn't it match?
Post a Comment