search this blog

Thursday, March 2, 2017

Confirmation bias


Every time I put up a thread that is even remotely linked to the Proto-Indo-European (PIE) homeland debate it gets hijacked by people who appear to have a pathological hate for the Kurgan PIE theory.

You'd think that considering the latest ancient DNA results from across Eurasia, which have thus far been very favorable to the Kurgan theory, these people would pipe down a little, at least for the time being, until something shows up that genuinely supports their stance. But nope.

The amount of confirmation bias in such threads is phenomenal. I'm going to start blocking and deleting the worst examples of this nonsense from now on. I'd also urge all reasonable and objective commentators here to try and ignore such comments, so that the offenders are left with no one to talk to.

If you're not quite sure how to spot an off-the-dial confirmation bias effort, here's an example from the last thread. I couldn't be bothered replying to this claptrap initially, but I will now, just to illustrate how off the mark it really is.

The truth is coming out.

Wonderful. Let's hear it.

EBA in South Asians is closer to Afanasievo than to Andronovo.

Maybe, but Afanasievo and Andronovo genomes aren't all that different, and at the moment we only have four Andronovo individuals, presumably from elite burials.

Who knows what more sampling from the territory of the vast Andronovo horizon, and a wider cross section of the Andronovo population, is going to uncover? We might find Andronovo samples that are perfect proxies for the Early Bronze Age (EBA) steppe admixture in South Asians; better than Afanasievo even.

Andronovo is different from Afanasievo because of Western/Caucasus expansions.

No, actually, Andronovo has Middle Neolithic farmer admixture from deep in Europe that Afanasievo lacks, or at least has much less of. Considering the preponderance of Y-haplogroup R1a in Andronovo remains, this admixture was probably mediated via female gene flow at the western edge of the Western Steppe, not near the Caucasus.

Afanasievo is leaning exclusively (for now) R1b. South Asians do not have Andronovo DNA.

As per above, South Asians may well have Andronovo ancestry. Or they may have ancestry from an R1a-rich sister group to the early Corded Ware population, which, based on an early Baltic Corded Ware genome, was in all likelihood basically identical to Afanasievo and Yamnaya in terms of genome-wide genetic structure (see here).

Ergo, EBA in South Asians did not come from Afanasievo. And R1a did not arrive in South Asia with Yamnayans or Andronovo/Afan.

But like I say, R1a may have arrived in South Asia with a Corded Ware-related population basically identical to Afanasievo and Yamnaya. Or with an Andronovo group basically identical to Afanasievo and Yamnaya.

This leads to at least three options:

1) PIE did not come from R1a or R1b, but J2. @Nirjhar J2 is present is UC Indians.


Y-haplogroups don't speak languages, so there's that. But we might be able to say, with a high degree of confidence, which Y-haplogroups were common in the PIE community based on their frequencies in ancient and modern-day populations.

Clearly, as things stand, the best candidates for so called PIE markers are R1a and R1b, and probably more specifically R1a-M417 and R1b-M269. J2 is a poor candidate for an PIE marker. See here: Y-hg J2 cannot be a Proto-Indo-European marker

2) R1a is source of PIE and did not come from Yamnaya, Andronovo, Afanasievo.

Well, most of the R1a in the world today may well be from an as yet unsampled Yamnaya population from the Pontic Steppe north of the Black Sea.

3) PIE came from several haplogroups, likely Caucasus area.

But why, considering the predominance of R1a and R1b on the ancient steppe and in modern-day speakers of Indo-European languages?

There are two potential locations for R1a. And I'm fine with both because they've made sense from the very beginning, unlike Yamnaya.

You mean the two potential main expansion points for most of the R1a in the world today? Surely either the Pontic Steppe or the Caspian Steppe?

One thing certain, R1a does not come from anywhere near Europe.

Hard to say where R1a comes from originally. My bet is that it was born in Upper Paleolithic Siberia, on the Mammoth Steppe that straddled Europe and Asia, possibly at a location very close to Europe.

R1a was the invasion that pushed R1b to the oceans of the Atlantic.

Actually, it seems that R1a-rich and R1b-rich steppe clans did their own thing when expanding into Asia and Europe during the Eneolithic and Early Bronze Age, and there's no evidence that they got in each other's way.

It's only during the Middle and Late Bronze Age that we see a population shift from R1b-rich to R1a-rich groups on the Caspian Steppe. This shift may have been accompanied by violence and language change, but if so, it's likely that one Indo-European language replaced another.

Also, please note that this R1a-rich population came from somewhere west of the Caspian Steppe, possibly the Pontic Steppe or the nearby forest steppe, because it had a higher level of Middle Neolithic European farmer admixture than the R1b-rich population that it replaced. So it's impossible to posit that this was an invasion from Asia, that pushed the R1b-rich population to the Atlantic.

This reply didn't take me long to put together. All I did was knock down the proverbial straw man over and over again. Easy work. But, at the same time, irritating and depressing.

Main take away point: if someone claims to know the "truth", chances are they're full of shit.

83 comments:

Josep Coderch said...

The main problem I see is that the people who don't support the kurgan hypothesis aren't consistent with their claims where PIE originated.
They basically follow the indo-european trail but going the reverse route, and every time any discovery reinforces the kurgan hypothesis they push their viewed PIE homeland northward, without realising it will eventually end up in the steppe.

Having said this there are still many things that aren't clear, but using them as a way to discredit the kurgan hypothesis doesn't help when all the other options make much less sense.

Open Genomes said...

Afanasevo culture 3300-2500 BCE R1b-PF7580 & R1b-CTS7763
contemporary with Yamnaya R1b-KMS67 - the Tocharians?


The Indo-European Tocharian language family is one of the very earliest branches of Indo-European. It was spoken in the Tarim Basin in the first millenium CE. Based on the vocabulary of the Tocharian languages, linguists have determined that the early Tocharian speakers left the Proto-Indo-European homeland (Urheimat) around 3,300 BCE to go east, and lost contact with other Indo-European speakers. Tocharian is a "centum" language like Italio-Celtic and Germanic as opposed to the "satem" languages like Balto-Slavic and Indo-Iranian.

Tocharian languages

We know that the Yamanaya culture kings buried in the kurgans (mounds) of the steppes were
R1b-Z2105 ⟹ R1b-Z2106 ⟹ R1b-Z2108 ⟹ R1b-KMS67

However, two other branches of R1b-Z2105 are found in East Asia:
• R1b-Z2106 ⟹ R1b-CTS7763 among Han Chinese from Beijing and Bhutanese, and
• R1b-L584 ⟹ R1b-PF7580 ⟹ R1b-SK2094 among Uyhgurs.

The Uyghurs live in the Tarim Basin of Xinjiang in Western China, and we know that the Turkic Uyghur language replaced the Tocharian languages around the year 1000.

Since R1b-M269* isn't found in East Asia, it seems likely that the Afanasevo R1b-M269s were in one or both of the two East Asian R1b-Z2105 clades.

The YFull tMRCAs for these R1b-Z2015 clades are 2,700-2,500 BCE which is within the time frame of the Afanasevo culture.
This corresponds well to the split between the early "centum" branches of Indo-European, which occurred before the "satem" languages developed:
YFull tree R-Z2103

The tMRCA of R1b-L51 is 2,800 BCE, which is about the same time that these East Asian R1b-Z2015s split from their closest relatives.
In combination with the other ancient R1b-M269 samples from Europe and the steppes, I think we can say that R1b-L23 developed in a place where most branches could travel both east and west, and also northward to Scandinavia (R1b-U106*) and south to the Caucasus.

This would have to be somewhere around the the home of the Yamnaya culture, in the steppes north of Black and Caspian Seas.

John Thomas said...

What would this site be without the unhinged quibbling and chauvinism?

Onur Dincer said...

Also, please note that this R1a-rich population came from somewhere west of the Caspian Steppe, possibly the Pontic Steppe or the nearby forest steppe, because it had a higher level of Middle Neolithic European farmer admixture than the R1b-rich population that it replaced. So it's impossible to posit that this was an invasion from Asia, that pushed the R1b-rich population to the Atlantic.

Not to mention the fact that eastern Yamnaya R1b belonged to the Z2105 subclade of L23 rather than L51 that western Europeans have. Assuming that western Yamnaya was rich in L51 rather than Z2105, this would cause a serious problem for the push to the Atlantic hypothesis as it would imply that only western Yamnaya was pushed to the Atlantic by R1a people supposed to come from Asia and that eastern Yamnaya was pushed to the south towards Anatolia and the Balkans instead. Such a selective push of various Yamnaya groups by R1a people supposedly from Asia is highly implausible.

Nirjhar007 said...

Let me ask something here, you mean in Andronovo itself an Yamnaya type autosomal population existed?.

That sounds really interesting . Also to note that Andronovo can't be archaeologically responsible as I have pointed a 1000 times here .

Ulug Depe already looks EBA type Yamnaya . So in this scenario the Southern Portion of the Andronovo horizon should be EBA Yamnaya type . Which overlaps with Ulug Depe and Indian pops .

Jaydeep said...

OG,

Before the Uighurs descended from the North, the Tarim Basin was divided into several city states across the margins of the Basin. Khotan dominated the southern route and Kucha the northern route. Kucha and other states like Karashahr & Turfan most likely spoke Tocharian. But the majority of the city states actually spoke East Iranian dialects that are closely related to East Iranian dialects in Tajikistan & Afghanistan.

Hence a straight forward link to Tocharian and R1b is not so easily established. If I remember correctly, R1b is also relatively high in Central Asian countries just above the Himalayas, which are speakers of East Iranian languages. A genetic influence coming from the southeast from Central Asia in a very real possibility considering the presence of east Iranian languages and also due to the heavy Indian influence due to Buddhism in the early centuries CE.

----------------

P.S. what are the R1b subclades present in Central Asia ?

Davidski said...

Let me ask something here, you mean in Andronovo itself an Yamnaya type autosomal population existed?

I don't know. Maybe?

What happened to the EBA migrants from the Western Steppe to Central Asia, like Afanasievo? Maybe their descendants were incorporated into early Andronovo?

Ulug Depe already looks EBA type Yamnaya.

But the Ulug Depe population that shares mtDNA haplogroups with EBA Europe is from the Iron Age. Andronovo is from the Bronze Age.

Genetic data suggest a close relationship between Yamnaya related populations and Iron Age Ulug Depe population.

http://eurogenes.blogspot.com.au/2016/08/maybe-first-direct-hints-of-yamnaya.html

Nirjhar007 said...

And another thing . You say that for millennia R1a and R1b 'IE Groups' it seems that R1a-rich and R1b-rich steppe clans did their own thing when expanding into Asia and Europe during the Eneolithic and Early Bronze Age .

Did their own thing okay. So according to this, perhaps the R1a and R1b blokes , did dna testings and after a meeting they planned to 'expand' y-dna wise?.

Is that suggestion more reasonable or the suggestion that , they were of different linguistic groups?. Endogamy played the role yes and perhaps some inter ethnic bride exogamy . But its simply not possible to say that they both belonged to same language i.e. IE . Its not possible to suggest that IE comes from R1 bifurcation . And at later stage R1a IEs started to dominate R1b, the bad asses in your language , the Sintashta Types . Possibly due to population expansion and invasions.

Remember that this idea that R1a is the PIE , is not and old 'pre-conceived' notion of mine , I had the doubts of PIE being R1a+R1b since R1b didn't exist in notable manner in S Asia, and R1a is the most common clade throughout IE groups worldwide or almost a universal IE clade .

Davidski said...

Nah, that's just another straw man.

We already know that during the Eneolithic the steppe was home to populations with both R1a and R1b, like Khvalynsk, that may have spoken Proto-IE.

It's from groups like these, speaking the same language, that the R1a-rich and R1b-rich clans formed due to patrilocality, endogamy, founder effect, etc.

Nirjhar007 said...

I don't know. Maybe?

Maybe can be many things mate :) .

What happened to the EBA migrants from the Western Steppe to Central Asia, like Afanasievo? Maybe their descendants were incorporated into early Andronovo?

Or they simply failed to create any impact and went extinct , and IF they were assimilated with Andronovo family , why they are absent in India?.

But the Ulug Depe population that shares mtDNA haplogroups with EBA Europe is from the Iron Age. Andronovo is from the Bronze Age.

Yes of course , but thats the point , if even in Iron age a SC Asian population was Yamnaya like , then how the Andronovo fits into all this?. with the absence of the 'Yamnaya Type' Andronovo you speak of , that itself must be in South , close to BMAC and N Indian area . Yes, but also Kalash for instance show W Eurasian type Mtdna , there is every chance that same will happen for N Indian +BMAC .

Onur Dincer said...

@Nirjhar007

What language did the Khvalinsk, Yamnaya and Afanasievo peoples speak according to you?

Davidski said...

Or they simply failed to create any impact and went extinct, and IF they were assimilated with Andronovo family, why they are absent in India?

They're not absent in India. There's Bronze Age steppe Y-DNA, mtDNA and genome-wide DNA in India today.

Yes of course, but thats the point, if even in Iron age a SC Asian population was Yamnaya like, then how the Andronovo fits into all this?

The abstract says Yamnaya-related mtDNA. Andronovo is ~75% Yamnaya-like and with a lot of typically Yamnaya mtDNA. So Andronovo is Yamnaya-related too.

Nirjhar007 said...

correction, I meant :

Remember that this idea that R1a is the PIE , is not an old 'pre-conceived' notion of mine .
Dave,
We already know that during the Eneolithic the steppe was home to populations with both R1a and R1b, like Khvalynsk, that may have spoken Proto-IE.

It's from groups like these, speaking the same language, that the R1a-rich and R1b-rich clans formed due to patrilocality, endogamy, founder effect, etc.


Okay but Khvalynsk R1a1 can be intrusive to the R1b area , which it was apparently from the samples we are having . Yes your suggestion is relevantly and technically possible , but still is a bit superficial , so we just wait for more proof on things.

They're not absent in India. There's Bronze Age steppe Y-DNA, mtDNA and genome-wide DNA in India today.

But first , just like for your 'EBA Type' Andronovo folks we need to check the Indian aDNA first with the BMAC one. But by absence, I meant the R1b, so even if we are to consider that Afanasievo was assimilated by Andronovo folks, they failed to leave Y-dna impact in India and near by . Also Indian L-657 is absent . And there is a big modern paper also coming , as you may have heard .

I truly hope this year will answer most of the fascinating and deep impacting questions .

Nirjhar007 said...

What language did the Khvalinsk, Yamnaya and Afanasievo peoples speak according to you?

Not sure , but I am at the moment not favoring them to be of IE . Can be Proto-Turkic , Uralic related .

Karl_K said...

"Okay but Khvalynsk R1a1 can be intrusive to the R1b area"

What language did that Malta guy speak?

There is no reason to either deny or support that most R1a and R1b populations spoke different languages at some point back 15,000 years ago.

But so what? That isn't the question. The R1a and R1b people at the right time and place to actually be Proto-Indo-Europeans look extremely closely related autosomally. That means there was a lot of interbreeding for quite some time.

Karl_K said...

@Nirjhar007

What language did Corded Ware speak?

Onur Dincer said...

@Nirjhar

Not sure , but I am at the moment not favoring them to be of IE . Can be Proto-Turkic , Uralic related .

Proto-Turkics lived during the Iron Age, so you should have meant pre-Proto-Turkics instead. How realistic is it to assume that pre-Proto-Turkics were rich in R1b-Z2105? If they were rich in a R1 sub-haplogroup, it would prolly be R1a-Z2124 given the modern distribution.

As for Proto-Uralics, I don't think there is any R1 sub-haplogroup which can plausibly be conceived as high in Proto-Uralics.

Unknown said...

I would like to pose a question based on the TMRCAs of R1b-L23 vs. R1a-M417 as reported by YFull:
If L23 has an expansion time of say 6200 ybp (Z2103--6100 ybp and L51--5800 ybp) and M417 has an expansion time of 5500 ybp (Z645-5000 ybp and Z93-4700 ybp) then how can these R1a and R1b lineages be part of the same expansive culture with the a common IE language? From an anthropological perspective, you'd expect these R1a and R1b lineages to have nearly identical expansion times if they spoke a common language and subscribed to a common culture. I think that the most plausible scenario is that either R1b was proto-IE and the R1a's acquired the language or the reverse--R1a was proto-IE and R1b acquired the language.Throw me a bone here!

Arza said...

@ Open Genomes

The Indo-European Tocharian language family is one of the very earliest branches of Indo-European.
No proof for that.

It was spoken in the Tarim Basin in the first millenium CE.
Indeed. Attested shortly before it dissolved completely.

Based on the vocabulary of the Tocharian languages, linguists have determined that the early Tocharian speakers left the Proto-Indo-European homeland (Urheimat) around 3,300 BCE to go east, and lost contact with other Indo-European speakers.
Glottochronology is a pseudo-science.

Tocharian is a "centum" language
No, it's not.

https://en.wikipedia.org/wiki/Centum_and_satem_languages
While Tocharian is generally regarded as a centum language,[5] it is a special case, as it has merged all three of the PIE dorsal series (originally nine separate consonants) into a single phoneme, *k. According to some scholars, that complicates the classification of Tocharian within the centum–satem model.[6] However, as Tocharian has replaced some Proto-Indo-European labiovelars with the labiovelar-like, non-original sequence *ku; it has been proposed that labiovelars remained distinct in Proto-Tocharian, which places Tocharian in the centum group (assuming that Proto-Tocharian lost palatovelars while labiovelars were still phonemically distinct).[5]

In other words - Western linguists proposed and assumed that Tokharian is a "centum" language, because it was very convenient for them. But it doesn't really matter because:

as opposed to the "satem" languages like Balto-Slavic

Balto-Slavic is not a "satem" language, nor "centum". There is no such division.

The only division we can make is:
a) languages that maintain rich set of PIE sounds (Polish - [c] gęś, [s] słyszeć, [c] gard, [s] żerdź, Lithuanian - [s] žąsìs, [c] klausyti, [c] gaȑdas, [s] žárdas)
b) languages that underwent sound shifts to a degree which caused the collapse of declination, often connected to secondary expansions and IE speakers being initially a minority.

Guess which are the so called "centum" languages.

Gioiello said...

@ Roy King
"If L23 has an expansion time of say 6200 ybp (Z2103--6100 ybp and L51--5800 ybp) and M417 has an expansion time of 5500 ybp (Z645-5000 ybp and Z93-4700 ybp)"

All the YFull tree dates are based upon two (undemonstrated) assumptions:
1) That the most part of the Big Y submitted may give a reliable quantity of SNPs. They are 20 to 1 as to the more reliable Y Elite or Full Genome with 15000 no calls out of about 100000 SNPs.
2) That the separation from A00 and A0-T is 235900 years old, whereas Poznik calculated it at 275000 and Shi Huang and colleagues calculated it at more than 300000.
3) Thus the YFull dates may be underestimated for an 1.17 or an 1.26 factor (at least).

Grey said...

Roy King

"I think that the most plausible scenario is that either R1b was proto-IE and the R1a's acquired the language or the reverse"

it could be but when starting from the single origin option for the sake of argument the model i have in my head for this question is a kind of reversal in mobility vs military power over time

when you read about historical steppe nomad incursions from later periods one of the things you notice is the size of the herds they needed e.g. each horseman having up to five remounts plus herds of cattle, sheep etc for food

so i picture:

"wagon IE" - earlier, with horses etc but not on the same scale as later and less well developed military power so they're less mobile in one sense but more mobile in the sense of "where" they can go because they need less fodder on the march e.g. lightly populated forests

and

"cavalry/chariotry IE" - a later, fully developed horse culture able to directly confront settled populations but traveling with much bigger herds so limited to regions which produced enough fodder for their animals

Karl_K said...

@Roy King

Very small differences in the expansion rate, or lineage extinction in the first few centuries before or at the start of expansion could explain the difference.

Actually, they are quite similar.

Imagine you had two groups, 'A' and 'B'.

At some exact point in time, they each split into A1/A2 and B1/B2 and then maintained the same very low levels for 1000 years.

Then, in a bad year, all of A2 became extinct for a random reason.

Then the next year, all 3 remaining lineages expanded rapidly by several thousand fold.

Later, it would appear that 'B' was an older expansion than 'A'.

Unseen extinctions.

postneo said...

Tocharian is a medieval language. Its useless to talk of their genetics. Why not sample actual tocharians then people thousands of years before them.

@Roy so you are saying the language expansion is completely synced with yDNA mutations.
I don't think either R1a or R1b was exclusively IE.

capra internetensis said...

@Roy King

These are point estimates though, and the true TMRCA can fall into a wide range around them. Y-Full gives R1a-M417 at 6300-4800 years and R1b-L23 at 6900-5500 years, which probably underestimates the real uncertainty. We will need to calibrate with high-quality ancient genomes to get more solid dates.

Also, a coalescence date is not an expansion date. L23 does not give rise to a star-like expansion: both branches descending from it have many mutations (=centuries) before they split again, so there is no reason to think anything especially demographically interesting was happening in this period. L51 is 5 SNPs down and then Z2118 and L11 are 8-12 SNPs further down, both about 5000 years old.

Y-Full adjusts the reported TMRCAs upward for consistency when subclades are calculated to be older than parent clades (I wish they wouldn't), which means an anomalous subclade TMRCA will screw up the dates above it. In this case under the 'info' tab you can see that Z2103 ultimately takes its date from Z2110. Z2110 has a very high estimated TMRCA (probably because it is an average of 2 single-sample lineages that happen to be very long branchs with 1 main subclade of normal branch length).

In fact Z2103 has 10 SNPs under L23 and then 7 more down to Z2110, which according to the dates given took place in 100 years, while 4 more SNPs then occurred over the 1300 years leading to CTS7556, a 55-fold difference in branch lengths. The uncorrected date for Z2103 is 5300 years, which is much more reasonable.

Anyway, to sum up, while L23 is probably older than M417, the actual expansion periods beneath these nodes are probably not statistically distinguishable.

Gioiello said...

@capra internetensis
"In this case under the 'info' tab you can see that Z2103 ultimately takes its date from Z2110"

YF02873 am I, and, having had Full Genome, have only 591 no calls (0,58%) out of 101916 SNPs at the date of 3.2.2017.
I have 89 reliable SNPs as to YFull, but now reduced to 86 (71+3+12: the ambiguous quality have been thrown away) (but I have more) and of course YFull didn't consider all of them: I have 12 SNPs with only 1 call but completely reliable, that have been excluded from the calculation. For that I say that instead of 6100 years my haplotype may have 7200 or even 8100 years.
From the smal's tree at FTDNA I have 16 SNPs in common with a French Basque and and Englishman (Winterowd/Barnum) and am waiting that an "Arab" is tested also from YFull, but these other people aren't tested at my level, so I don't know if the SNPs in common may be more, thus I cannot calculate when our lines separated. It seems that this "Arab" has only one SNP in common with me and the others (FGC24408), thus we separated at the beginning of our subclade. Anyway my Y is the oldest of all the survived ones, more than R-M269-PF7562 and all the others. Only R-L51 may be at my level. This is one of the reasons of my "Italian Refugium" (amongst infinite others).

Ryan said...

I'm telling you - secret Inuit David. PIE are secret Inuit.

Ryan said...

There's actually (this part isn't a joke) linguistic evidence for an Inuit-PIE link at least. Wake up sheeple!

Gioiello said...

@ Ryan
"There's actually (this part isn't a joke) linguistic evidence for an Inuit-PIE link at least. Wake up sheeple!"

IE is linked with all the Nostratic languages and with the Caucasian-NaDené-Sino-Tibetan had contacts. I demonstrated more than 30 years ago the link between IE "6" and Sino-Tibetan. The paper wasn't published, but was known from many linguists of the "Scuola Normale Superiore" of Pisa.

capra internetensis said...

@Gioiello

I thought that one might be you, throwing everything off with your excessive number of mutations. ;)

Gioiello said...

@ capra internetensis
"I thought that one might be you, throwing everything off with your excessive number of mutations"

You should understand if my mutations are excessive or if the others are underestimated (as I think) or even not catched.

Al Bundy said...
This comment has been removed by the author.
Ryan said...

@Gioello - I'm referring to the study David posted a while back that showed ?Indo-European and Chukotko-Kamchatkan forming a clade to the exclusion of all other languages, and both forming part of a Siberian/Steppe Eurasiatic macrofamily.

http://www.pnas.org/content/112/41/12752.full

I'm not taking that to the bank, but it is interesting.

I'm assuming Chukotko-Kamchatkan is closely related to Eskimo-Aleut, but that's by no means certain either.

Ryan said...

There are weird ties between Chukotko-Kamchatkan and Scots Gaelic and Manx oddly enough in that paper. Sounds crazy, but perhaps not the craziest in the world if there's either a Dene-Caucasian substrate in both (if Dene-Caucasian is valid as a theory), or if there's an Afanasievo substrate in Chukotko-Kamchatkan (which is quite possible).

Atriðr said...

Despite your consistency in claiming that I’m filled with waste material, thanks for the honor. Allow me to rebut.

You say:
Maybe, but Afanasievo and Andronovo genomes aren't all that different, and at the moment we only have four Andronovo individuals, presumably from elite burials.
Do you think I or anyone disagrees with this? However, you know perfectly well what the Lazaridis (who also wants more samples) paper demonstrated. No use to be coy. Why pretend that from current samples, EBA in South Asians is not closer to Afanasievo than Andronovo? Why not accord the point? In light of new data, we adapt. But we don’t have the new data, so no use to talk about it. Also, if you were more thorough in your homework, you’d see that I used to promote and support Andronovo for I-Ir on your own blog. Till Lazaridis.

At any rate, everyone wants more samples:

P.128 Hollard, Clémence 2014: Le faible échantillonnage observé par groupe ainsi que le processus de recrutement funéraire peut bien évidemment avoir biaisé une partie des analyses. C’est pourquoi, afin de discuter ces hypothèses, il est indispensable d’envisager d’autres analyses paléogénétiques. P. 128 Hollard, Clémence 2014
You say: No, actually, Andronovo has Middle Neolithic farmer admixture from deep in Europe
Davidski, Europe is WEST of Afanasievo, you know that. As I stated: Western/Caucasus expansions. In this statement, you focus on my inclusion of Caucasus (N.) to try and deride my point? That is disingenuous.

As per above, South Asians may well have Andronovo ancestry.
They might - but no proof of it yet. See my point, two above.

But like I say, R1a may have arrived in South Asia with a Corded Ware-related population basically identical to Afanasievo and Yamnaya.
CWC mostly R1a. Yamnaya, mostly R1b. How many samples of R1a in Afanasievo? What is the CWC related population? No need to invent a people. Name where they are and avoid teleporting populations. Anthony teleports populations too, so don’t feel too bad.

I say: This leads to at least three options
How closed-minded of me, listing AT LEAST, three options.

Y-haplogroups don't speak languages, so there's that.
This one is rich. As far as I know, only @Rob, @Kristina and I are actually open to this idea. But coming to the point, most of your own blog posts don’t even believe this. The entire polemic is based precisely that they do. The entire usage of Indo-European to refer to haplogroups movements is based that they do. Nick Patterson’s current quest is based that they do. And the paper of your last posting is guilty of the same:

P. 128 Hollard, Clémence 2014: L’haplogroupe (R1a1a (xM458)) pourrait ainsi être associé à l’expansion de la culture d’Andronovo et aux locuteurs de langues indo-iraniennes.

Very disingenuous.

But we might be able to say, with a high degree of confidence, which Y-haplogroups were common in the PIE community based on their frequencies in ancient and modern-day populations.
This is what I listed as a third option, but I’m guessing you want to ignore I’ve said as much and attack first.

J2 is a poor candidate for an PIE marker.
The Latin descendants (a.k.a. Romans) have J2 but not R1a. UC Indians have J2 but not R1b. So two of the oldest speakers, of two of the oldest Indo-European languages share J2 amongst each other (as well as some Albanians and Persians for the linguists) but yet the suggestion is preposterous?

Well, most of the R1a in the world today may well be from an as yet unsampled Yamnaya population from the Pontic Steppe north of the Black Sea.
May. Confirm first if you wish to insist on insulting others.

I say PIE came from several haplogroups and you reply But why, considering the predominance of R1a and R1b on the ancient steppe and in modern-day speakers of Indo-European languages? - which absolutely contradicts your own statements in your rebuke of my option 1.

Atriðr said...

I say R1a does not come from anywhere near Europe
You reply Hard to say where R1a comes from originally. My bet is that it was born in Upper Paleolithic Siberia
Well, at least we have common ground here.

Actually, it seems that R1a-rich and R1b-rich steppe clans did their own thing when expanding into Asia and Europe during the Eneolithic and Early Bronze Age, and there's no evidence that they got in each other's way.
This same bad (meek?) habit causes us to change the name of the Battle-Axe culture to the Corded-Ware culture. The affinities between R1a and R1b are rich, maybe brotherly, but to suggest that the various clean territorial splits, and the satem-centum split are simple happenstance... that’s sloppy.

It's only during the Middle and Late Bronze Age that we see a population shift from R1b-rich to R1a-rich groups on the Caspian Steppe. This shift may have been accompanied by violence and language change, but if so, it's likely that one Indo-European language replaced another.
Maybe, maybe not - but I thought haplogroups did not speak languages. Also, why ignore the possibility that R1b spoke a non-I.E. language? Or R1a in the remote past?

chances are they're full of shit.
How to make friends and assure the mutual exchange of ideas. And frankly, you have no clue where I stand. You just attack rabidly anyone who disagrees with the Kurgan stance.

Confirmation Bias? You think you are immune to this? You think Mallory is when he analyses the tartans of Tocharians, coincidentally making connections with his Irish roots? Or Gimbutas, injecting her own Lithuanian bias in her works? OIT-believers are almost exclusively Indians. Think this is chance? And here you promote an Eastern European rigid stance. I don’t know your background, but if you said you were Eastern European, then I’d say, it’s just normal. Just normal.

Ryukendo K said...
This comment has been removed by the author.
Davidski said...

@Roy King

The most likely scenario at this time is that the PIE community on the Eneolithic Western Steppe harbored various subclades of I2a, R1a, R1b and Q.

At some point during this phase, and at the expense of the other lineages, two steppe clans rich in M269 and M417, respectively, began expanding. Considering the uncertainties in estimating expansion times with modern genomes, it's possible that they started expanding at about the same time, but if not then it's not a big deal.

@Atriðr

It's not important where R1a was during the Upper Paleolithic. It might have been on the moon. This has nothing directly to do with what happened during the Eneolithic and Bronze Age.

And it's not important that Latins and Indians carry J2, because J2 is an old and complex haplogroup with a number of fairly popular subclades that show very early and rather late expansion dates that appear to mostly flank the Eneolithic and Bronze Age.

Indeed, it's not a coincidence that J2 is missing from the Eneolithic and Bronze Age Steppe. Most importantly, not a single skeleton from an elite Kurgan on the Bronze Age steppe belongs to J2. This West Asian haplogroup simply didn't show up to the party at the right time, which is actually a strong argument against a West Asian PIE homeland.

What is important is that Latins and Indo-Aryan Indians carry a lot of steppe admixture and a lot of M269 and M417, respectively, most of which they appear to have acquired at about the same time during the Bronze Age from more or less the same people, who probably spoke very closely related languages.

Jaap said...

Davidsky, please keep your great site as it is, and don't ban anyone. Please? So far you've been a great host, bearing with all sorts of visitors. Please keep up the good work!

Ryukendo K said...
This comment has been removed by the author.
Ryukendo K said...
This comment has been removed by the author.
bellbeakerblogger said...

What's with Kh22? Do you have access to that genome?

Ryu,
There's a paper out just now on Mexican mtdna @ DNAeXplained which supports your point. Mexican maternal ancestry is overwhelmingly Indian, yet they are autosomally 60/40 or thereabout. Suggests a very high degree of social bias toward European males which probably accounts for eventual paternal fixation (like R1b)

capra internetensis said...

@RK

If I understand you correctly that would also contradict a 'Hungarian' scenario for the origin of Basques, i.e. where a regular IE-speaking R1b-rich population shifted to the language of Vasconic-speaking elite without much genetic impact, and subsequently diverged from other such populations.

Ryukendo K said...
This comment has been removed by the author.
Alberto said...

@Ryu

We mentioned this when that paper about the Rathlin genome was out, I think. It looks like certain populations with low diversity (Druze, Basque, Sardinian, Lithuanian) get quite lower than expected amount of shared haplotypes than other more diverse ones (clearly those with SSA admix, but also for example Turkish), so that the former are consistently above the line and the latter below the line.

Basque sharing less haplotypes with Loschbour than Spanish seems to be caused by this effect, for example. Opposite effect from formal stats (so if instead of IBS -which is more neutral- you'd use f3 with outgroup, the overall differences could be much bigger than seen here).

Arza said...

@ Ryukendo K

Do I understand correctly the last chart that Lithuanians have the highest WHG-like ancestry, but when it comes to exact connections they are behind Western Europe?
What does this say about "WHG in Balts"?

Davidski said...

WHG in Lithuanians is mostly from Bronze Age Narva people, not from anywhere near the Loschbour cave in Belgium.

Olympus Mons said...

@ Ryukendo K

What if you got it wrong? What if the case was a small number off BB from a 3rd point of origin moved to be proto basque region and another group of that point of origin moved past the proto basque and into germany mixing with a different more central european population and becoming Rathlin... Like anthropology seems to indicate. What would it mean to your analisys? Nothing. It would still look the same, correct?

Rob said...

@ Grey

From last thread, your speculations about horse riding have no merit. For example, it is well observed that "e know they had domes- ticated horses and horse bones were also found in the con- temporary neigbouring societies in southeastern Europe, for example, likely to have been passed on by steppe pop- ulations no matter if Yamnaya or slightly earlier Pit-Grave groups. However they are always only found in small numbers questioning their importance beyond a mere prestige object" (V Heyd).

Nor does the Numidian cavalry offer any detour around the negative; as by that time the horse had been domesticated for thousand years or so; and light / accoutrement-less cavalry represents an intentional "devolution" rather an evolutionary steppe in progress.

Rob said...

@ Matt

"common Neolithic ancestry between BB and IE Europeans not shared by Basque and Sardinian. In other words, the difference between IE Europeans and Basque/Sardinian is not just in the Yamnaya/directly steppe-derived portion of the ancestry, IE Europeans are distinguished by similarities in Neolithic ancestry as well."

That is why I raised the issue a few threads ago when Davidski modelled Basques purely on Lengyel- it didn't seem historically sound (although this wasn't the main aim of his exercise). As such basques would have mostly actually middle Neolithic southern/ SW European ancestry.
Curiously, I don't think this is the case for southeastern Europe, even Greeks; where the first Barcin farmers might have left little legacy, with subsequent Neolithic ancestry coming from east-central Europe, and even Armenia/ Chalcolithic Anatolia.

Taymas said...

@Nirjhar

"Andronovo can't be archaeologically responsible as I have pointed a 1000 times here"

I've seen you claim this often but I haven't seen the full explanation. Would you mind fleshing out the argument or perhaps pointing me to a comment I missed where you have already done so? Thanks.

postneo said...

"I've seen you claim this often but I haven't seen the full explanation"

There is no archeological trail, its not a claim but evident in every excavation and paper on the subject.
Most proponents of Andronovo incursion already know this and have been looking "desperately?" for signs for some time.

Rather the hypotheses of Andronovo incursion is the claim not the other way around. Its built on two things:
There are some andronovo artefacts in late BMAC indicating some contact. remember BMAC and south asia are a great distance apart.
The supposed genetic turnover hypothesized based on modern DNA. The evidence there is not clear cut.

Ryukendo K said...
This comment has been removed by the author.
Rob said...

@ postNeo
Yes not much evidence at all apart from some "barbarian" ceramics in BMAC settlements nearer the steppe. They contain mike residue so it does show that pastoralists were importing secondary products to the BMAC citadels.

So it'll be one of the most fascinating evidences to see how Z93 ended up in South Asia, as the only evidence for incursion south into Iran and beyond is from BMAC material, not Andronovo

Taymas said...

@postneo

Evident how? You've begged the question. I'm not asking to be a dick, it's genuine. I don't spend a great deal of time reading archaeology papers. My hard-science brain finds them a little too dependent on implicit knowledge and thus tricky to verify.

Genetic turnover: what's the competing hypothesis explaining the genetic variation in South Asia? IIRC you've advocated for an origin in South Central Asia, right? So does that mean you'd predict some regions south of BMAC territory to be harboring, pre-Bronze Age, something genetically similar to what Davidski is calling "steppe" ancestry (which in this framework would be a slight misnomer)?

Or do you believe, as I think Nirjhar does (sorry if misrepresenting) that the steppe ancestry entered South Asia much more recently? In that circumstance, are you positing an Indo Iranian conquest of the steppe originating from South Central Asia, then a back-migration of some Indo-Iranian subset carrying steppe ancestry? Why did that late steppe-influenced wave (which between the non-steppe SCA ancestry and dilution in the back-migration must have been weakly steppe), have such a greater genetic impact than the later Turkish wave? Why haven't we seen much Iranian neolithic or ASI in our steppe samples thus far?

I just don't understand the anti-kurgan case very well but I'd like to. Thanks.

Gunther said...
This comment has been removed by the author.
Gunther said...

@Davidski

You already have your Yamnaya/Afanasievo-like R1a population, Davidski. Potapovka, 2/2 of them. They're in your own PCA plots.

Btw, why is Andronovo not a good proxy for steppe admixture in South Asians? No European farmer in South Asians or is it just d-stats/qpadm don't show them as a good match? What about Poltavka outlier(who had Z94), Srubnaya or Sintashta?

Nirjhar007 said...

The archaeological , Anthropological and Ancient text wise evidence , despite all kind of manipulation attempts , comprehensively suggests that , Steppe Invasion of South Asia , prior to actual Steppe Scythian migrations is a dream , a dream that many wants to come true .

So to wake up , since this blog is genetic comes first , a few Indian aDNA will be enough . That's it , that's the whole thing .

Taymas said...

Nirjhar, beg your pardon, but now you're begging the question. What evidence? What is it about the archaeology, anthropology, and ancient texts that make you think Andronovo could not be partially ancestral to Indo-Aryan culture? Thanks.

Taymas said...

Regarding the hypothesis that steppe-like ancestry came to South Asia via later Scythian invasion, I'll direct my questions for Postneo to Nirjhar as well:

In that circumstance, are you positing an Indo Iranian conquest of the steppe originating from South Central Asia, then a back-migration of some Indo-Iranian subset carrying steppe ancestry? Why did that late steppe-influenced wave (which between the non-steppe SCA ancestry and dilution in the back-migration must have been weakly steppe), have such a greater genetic impact than the later Turkish wave? Why haven't we seen much Iranian neolithic or ASI in our steppe samples thus far?

Thanks.

Alberto said...

@RK

Maybe in other cases yes. But in this particular one that doesn't seem to happen. Quite the opposite.

Basques do share a good amount of haplotype with Ballynahatty, but less so than Spanish, Tuscans or French (while they get higher score with IBS with Esperstednt_MN and Spain_MN). Same situation with BR2, and same situation with Loschbour (with Spanish and French).

Druze behaves in an analogous way vs. Syrians too. So there's something going on with these populations (which doesn't invalidate what you said, but I just wanted to point out that part of the phenomenon is due to "technical" reasons, whatever they are in this case).

This is why lately I've commented on why I prefer IBS over other methods for phylogenetic modeling of populations. Haplotype sharing is very informative (can tell you many things that formal stats can't), but it has its limitations (in this case, showing that Druze and Syrians are quite different, when we would expect them to be quite similar), plus we just can't use this method often due to lack of high coverage samples.

Formal stats would also say that Druze and Syrians are very different. While IBS will say that they are very similar (so for example, if you run an IBS for Druze against modern populations, Syrians would be at the top of the list, but if you run f3 instead, you'll find Icelendic, English_Cornwall, Lithuanians and basically every single European population above Syrians, which for me is a problem for accurate modeling). But this is a different debate altogether, I guess.

Gioiello said...

@ Alberto
"Formal stats would also say that Druze and Syrians are very different. While IBS will say that they are very similar (so for example, if you run an IBS for Druze against modern populations, Syrians would be at the top of the list, but if you run f3 instead, you'll find Icelendic, English_Cornwall, Lithuanians and basically every single European population above Syrians, which for me is a problem for accurate modeling). But this is a different debate altogether, I guess".

In fact we may have some doubt about the origin of Druzes. Many people who based upon them that mt Hg. K came from Middle East may be wrong, just because Druzes have close links with Europeans more than Middle East, thus I wouoldn't be so confident about IBS...

Shaikorth said...

@Alberto

If you run f3 for Syrians, are the Europeans above Druze?
Something similar has been seen before.

Alberto said...

@Shaikorth

I don't have specifically f3 for Syrians, but here is the list for double outgroup D-stats for Syrians in descending order (notice that there are Syrians too in the list somewhere down there - 2 randomly sampled groups, one the reference and one in the list):

Hungary_CA 0.3822
Anatolia_Neolithic 0.3804
LBK_EN 0.3799
Germany_MN 0.3798
Hungary_EN 0.3791
Sardinian 0.3791
Iberia_EN 0.3789
Basque_Spanish 0.3786
Italian_Bergamo 0.3783
Iberia_MN 0.3782
Italian_Tuscan 0.3779
Basque_French 0.3778
Spanish_Pais_Vasco 0.3776
Greek1 0.3774
Albanian 0.3772
Greek2 0.377
Icelandic 0.3767
Czech 0.3765
English_Cornwall 0.3763
English_Kent 0.3763
Norwegian 0.3763
Croatian 0.3762
Bulgarian 0.3761
Scottish_Argyll 0.3761
French 0.3759
French_South 0.3757
Hungarian 0.3757
Armenian 0.3756
Cypriot 0.3755
Hungary_BA 0.3753
Lithuanian 0.3752
Orcadian 0.3752
Spanish_Cataluna 0.3752
Belarusian 0.3751
Georgian 0.3751
Spanish_Valencia 0.3751
Spanish_Aragon 0.3748
Spanish_Cantabria 0.3748
Ukrainian_West 0.3747
Abkhasian 0.3744
Ukrainian_East 0.3744
Spanish_Castilla_la_Mancha 0.3743
Estonian 0.3742
Spanish_Baleares 0.3742
Georgian_Jew 0.3741
Iranian_Jew 0.3739
Italian_WestSicilian 0.3739
Chechen 0.3737
Lezgin 0.3736
Italian_EastSicilian 0.3733
Ashkenazi_Jew 0.3732
Iraqi_Jew 0.3732
Sephardi_Jew 0.3732
Spanish_Castilla_y_Leon 0.373
Unetice 0.3728
Spanish_Andalucia 0.3727
Spanish_Extremadura 0.3726
Kumyk 0.3725
Srubnaya 0.3724
Sintashta 0.3723
Finnish 0.3722
Spanish_Murcia 0.3722
Adygei 0.3721
North_Ossetian 0.3717
Druze 0.3716
Turkish 0.3716
Bell_Beaker_Germany 0.3715
Corded_Ware_Germany 0.3715
Mordovian 0.3715
Russian_Kargopol 0.3712
Maltese 0.3708
Spanish_Galicia 0.3708
Armenia_BA 0.3702
Potapovka 0.3702
Poltavka 0.3701
Andronovo_subset 0.3694
Andronovo 0.3687
Azeri_Baku 0.3684
Western_HG 0.3684
Yamnaya_Samara 0.3681
Iranian 0.3677
Moroccan_Jew 0.3677
Nordic_LNBA 0.3677
Afanasievo 0.3665
Syrian 0.3664
Saami 0.3662
Saudi 0.3662
Caucasus_HG 0.3652
Tajik_Rushan 0.3652
Tajik_Shugnan 0.3648
Yamnaya_Kalmykia 0.3648
Lebanese 0.3644
Yemenite_Jew 0.3643
Tajik_Ishkashim 0.3641
Nogai 0.3631
Mezhovskaya 0.363
BedouinB 0.3619
Kalash 0.3608
Pathan 0.3598
Palestinian 0.3596
Balochi 0.3595
Brahui 0.3591
Karelia_HG 0.3585
Jordanian 0.3574
Turkmen 0.3569
Makrani 0.3562
Karasuk_subset 0.3559
GujaratiA 0.3558
Sindhi 0.3555
Burusho 0.3548
Uzbek 0.3521
BedouinA 0.3519
Karasuk 0.3519
Mansi 0.3519
GujaratiB 0.3517
Altai_IA 0.3504
GujaratiC 0.3501
GujaratiD 0.3486
MA1 0.3484
Uygur 0.3476
Punjabi_Lahore 0.3468
Okunevo 0.3458
Tubalar 0.3449
Kyrgyz 0.3423
Dravidian_India 0.3381
Karasuk_outlier 0.3335
Yakut 0.3333
Itelmen 0.3321
Nganasan 0.3289
Ulchi 0.3261
Mozabite 0.3255
Dai 0.3253
Kharia 0.3252
Atayal 0.3244
Onge 0.318
Ust_Ishim 0.3109
Papuan 0.3088
Masai_Kinyawa 0.1847
Esan_Nigeria 0.1092

Alberto said...

@Gioiello

Druze was just an example, it could be said the same about any other Lebanese group, or Jordanians or whatever. So it was not a point specifically about the Druze (except that the Druze have low diversity and it shows in those haplotype sharing stats that they get depressed values).

Karl_K said...

@Ryukendo K

"Hmm If I'm not wrong, small and drifted populations tend to have more IBD sharing with other populations, not less, as the diversity of haplotypes is lower and more homogeneous in general."

You are right! There is someone among us who actually understands how to properly analyze IBD and IBS.

IBD is mostly about diversity if POSSIBLE haplotypes. To greatly simplify, we can imagine segments as being similar to Y and mt haplogroups, but with the added fact that they cease to exist with even a bit of admixture with a distantly related group.

So if the breeding population ONLY contains a certain haplotype, then it could remain unaltered for hundreds of generations (especially at a SNP-chip level of comparison).

But if even a single person from an unrelated group joined the breeding population, it starts to fall apart. The more successful their lineage is, the faster the IBS disappears from the entire population.

Ancient DNA compounds this problem by having low coverage. So that haplotypes from an inbred individual are much more accurate than haplotypes from a genetically diverse individual.

Shaikorth said...

@Alberto

Yeah expected numbers. Both Dstats and f3-stats suffer from this issue, some pops get elevated and others depressed numbers.

When modeled as mixes of modern populations based on their haplotypes, Druze look fully Lebanese (makes sense since they are a an endogamous subset of local diversity) while Syrians look like Lebanese + peninsular Arab + some Armenian, both of these fits make more sense than everyone being closer to Europeans compared to their neighbours. If modeled as mixes of ancient haplotypes, Druze have similar Iran_N/EEF ratio as the Lebanese on average but much more self-copying (drift) and less SSA while Syrians have more Iran_N.

Grey said...

@Rob

My contention is farming groups like C-T who'd originally encroached on the steppe could have been displaced by early PIE (through constant raiding rather than mass hunnic conquest) and the counter argument saying that was *impossible* cos no saddles, bridles etc is bogus as proved by 100s of youtube videos

https://www.youtube.com/watch?v=zcPqXfck9Hs

there may be better counter arguments - no horses would be a good one - but until that is shown then the argument that it was *impossible* for PIE to displace groups like C-T is unproven imo

(nb the raiding argument only applies on and near the steppe as hit and run won't work if you can't run away afterwards)

anyway, apart from an excuse to link lots of cool free riding videos

https://www.youtube.com/watch?v=0jBjg717TJc

the interesting part of this is *if* groups like C-T were displaced slowly by raiding then where did they go?

Greece and the Balkans are the obvious places but maybe it explains some of the loan words in Sumerian as well?

#

in terms of finding evidence *if* PIE did use halters rather than bridles then maybe they decorated them like this

https://s-media-cache-ak0.pinimg.com/736x/93/df/fa/93dffa3ed5cc8a026c304a26b6e656b7.jpg

or hung scalps from them

https://www.jstor.org/stable/25669050?seq=1#page_scan_tab_contents

Arch Hades said...

How can anyone say that R1a doesnt come from anywhere near Europe? lol It's been found in a Karelian Hunter Gatherer dated to like 5500 BC. None of the pre 3,000 BC ancient genomes we have of the Near East have been found to carry R1a. Is this guy expecting Neolithic or Paleolithic India to be full of R1a?

Atriðr said...

@Arch Hades
How can anyone say that R1a doesnt come from anywhere near Europe?
OP actually states the same thing I imply: My bet is that it was born in Upper Paleolithic Siberia, on the Mammoth Steppe that straddled Europe and Asia. Where has reading comprehension gone?

But forget everything I've said, and just remember this: the majority of the Steppe is not and has rarely been considered European. If you wish to change Europe's definition to Eurasia, let's just agree to call Eurasia, Eurasia.

Davidski said...

The Western Steppe is in Europe. It makes up a big part of Eastern Europe.

And it's the only relevant part of the steppe to the PIE debate, especially considering all of the latest ancient DNA data.

Atriðr said...

And it's the only relevant part of the steppe to the PIE debate, especially considering all of the latest ancient DNA data.
Precisely, which might finally lead you to see what I'm saying.

Atriðr said...

But I would add, that you cannot refer to Yamnaya as Eastern Europe in its ancient context. It is Eurasian.

Davidski said...

I can't be bothered getting into these sorts of semantics. But the division you're looking for is Europe vs European steppe, not Europe vs Eurasia.

So it's more correct to say that Yamnaya comes from the European steppe. The new Scythian paper follows this line of thinking.

"This further implies that carriers of the Yamnaya culture migrated not only into Europe [26] but also eastward, carrying west Eurasian genes—and potentially also Indo-European languages—to this region [17].

All Iron Age individuals investigated in this study show genomic evidence for Caucasus hunter-gatherer and Eastern European hunter-gatherer ancestry. This is consistent with the idea that the blend of EHG and Caucasian elements in carriers of the Yamnaya culture was formed on the European steppe and exported into Central Asia and Siberia [26]."

http://eurogenes.blogspot.com.au/2017/03/genetic-origins-and-legacy-of-scythians.html

Atriðr said...

Yamnaya and Afanasievo are the same group; those are the boundaries; that places it clearly as a Eurasian culture. But to each his own.

Regarding the latest (fantastic) paper on the Scythians which you quote (which I'm in midst of reading), it also states:

The northern Black Sea steppe was originally considered the homeland and centre of the Scythians3 until Terenozhkin formulated the hypothesis of a Central Asian origin4. On the other hand, evidence supporting an east Eurasian origin includes the kurgan Arzhan 1 in Tuva5, which is considered the earliest Scythian kurgan5. Dating of additional burial sites situated in east and west Eurasia confirmed eastern kurgans as older than their western counterparts6,7.

Do you start seeing certain contradictions? The paper should be wiser and avoid talking about Indo-European (ling.); at least they remain ambiguous.

Anyhow, I'll finish reading the details, but it's upholding my stances.

Davidski said...

Yamnaya and Afanasievo are the same group; those are the boundaries; that places it clearly as a Eurasian culture. But to each his own.

Yeah, and the British and Spanish are Euro-Americans not Europeans because their relatives colonized the Americas.

Awesome logic.

Atriðr said...

Yeah, and the British and Spanish are Euro-Americans not Europeans because their relatives colonized the Americas.

You obviously do not understand the Steppes if you think this is a question of colonization. Not even a comparable analogy. Read the paper you just posted. The first page will explain it to you.

And eventually, you'll need to come to grasp with what EHG actually is, even though you've stated in a past blog that you understood its breakdown. But you've yet to absorb the weight of that meaning.

Davidski said...

Yamnaya formed on the European steppe, therefore it's native to Europe. I'll post that quote again...

This is consistent with the idea that the blend of EHG and Caucasian elements in carriers of the Yamnaya culture was formed on the European steppe and exported into Central Asia and Siberia.

The reason Afanasievo is so similar to Yamnaya is because it's a European culture exported by European migrants to Asia.

So what's the issue? What are you trying to debate with me?

And that part of the Scythian paper that you quoted that you're so joyous about, well, they're just talking about the earliest Scythian Kurgans from the Iron Age.

You do know that Kurgans already existed on the European steppe during the Eneolithic, don't you?

Atriðr said...

that you're so joyous about
You're delusional. You take me for one of your other commentators. Maybe this is how you react, but in no way describes me.

And you are confused.

Witness you above at 10:34 PM: Yeah, and the British and Spanish are Euro-Americans not Europeans because their relatives colonized the Americas.

Now you at 11:06 PM: Yamnaya formed on the European steppe, therefore it's native to Europe. This was you response to EHG because you understood the implications of what EHG is. Pathetic.

And you also post this jewel: The reason Afanasievo is so similar to Yamnaya is because it's a European culture exported by European migrants to Asia. A European culture? What a jewel, what a jewel. I'm sure the Steppe folk back in the day greeted each other as such too: "Hello, my fellow Europeans! Good day, good day, fellow European!"

Enjoy your navel-gazing.

The smarter readers on your website will understand the subtexts of my comments.

Davidski said...

There aren't subtexts in your comments.

You're just rambling nonsense because you can't accept that Yamnaya is European and Afanasievo is just an Asian offshoot of Yamnaya.

Clearly you were hoping that Yamnaya was derived from Afanasievo, and you can't let go of that fantasy for some reason.

Atriðr said...

You're just rambling nonsense because you can't accept that Yamnaya is European and Afanasievo is just an Asian offshoot of Yamnaya.
EHG is just an offshoot of ANE. See how ridiculous this sounds?

Clearly you were hoping that Yamnaya was derived from Afanasievo, and you can't let go of that fantasy for some reason.
Really, why? But no, no fantasy here. More of your projection. I'm making the point that to find the source of PIE, one has to isolate between various variables, but that considers all variables. Once we can isolate, we can pinpoint the environmental insult that will point the arrow to the source. I'll repeat again, I was a huge supporter of the Yamnaya position until its strongest advocates started ignoring key pieces of information. Then I distanced myself to get a clearer picture. You're proof of this: confusing genetics, languages, with next to no historical sense of the unfolding of events.

You're fairly tolerant with your comments section, I'll give you that. That's a compliment. Scores enough points with me to tolerate your rude mouth.

Unknown said...

If you are saying Afanasievo aren't an offshoot of Yamnaya or at least from the Pontic Steppe where did they get their Caucuses' ancestry? Do you think it existed in the east?

Anonymous said...

Confirmation bias, the blog.

Daniel Brockert Actor said...

k/s variations happen within languages and the phenonomenon isn't limited to Indo-european languages. It's also found in semitic languages. Which seems to be a solid reason to drop the Centum-Satem distinction entire. ElectriC / ElectriCity