search this blog

Tuesday, August 22, 2017

Last South Asian qpAdm crapshoot


I had one last go at modeling South Asians with qpAdm using ancient samples from the Eurasian steppe and Iran before the (imminent?) publication of more proximate and relevant ancient data from Central and South Asia. Indeed, I used the updated qpAdm methods from Lazaridis et al. 2017 (ie. I packed the right pops with 16 outgroups), and, initially, the exercise was an utter failure.

The normally successful (p-value >0.05) model with Han, Neolithic farmers from Iran (Iran_N), Onge, and Early/Middle Bronze Age herders from the Eurasian steppe (Steppe_EMBA), turned out either infeasible or a really bad fit (for instance, see here). Clearly, something important was missing.

However, using Sarmatian_Pokrovka instead of Steppe_EMBA, as proxy for Bronze Age Eastern European ancestry, and the Lebbo (from Borneo) instead of Han and Onge, as a stand-in for Ancestral South Indian (ASI) ancestry, worked well enough, at least for the South Asian populations that I chose to run. Why? No idea. Feel free to speculate in the comments.

But importantly, unlike the Indo-Iranians, Dravidian speakers from India - Gond, Kapu and Malayan - could be modeled successfully without any ancient Eastern European admixture. No need to explain why this is so important.

Brahmin
Iran_N 0.277±0.042
Lebbo 0.304±0.037
Sarmatian_Pokrovka 0.418±0.045
P-value 0.293270281
chisq 15.229
Full output

Gond
Iran_N 0.095±0.061
Lebbo 0.905±0.055
Sarmatian_Pokrovka 0.001±0.063
P-value 0.785177372
chisq 8.836
Full output

Kalash
Iran_N 0.373±0.042
Lebbo 0.137±0.038
Sarmatian_Pokrovka 0.489±0.047
P-value 0.0661709327
chisq 21.356
Full output

Kapu
Iran_N 0.324±0.054
Lebbo 0.573±0.053
Sarmatian_Pokrovka 0.104±0.060
P-value 0.945176825
chisq 6.027
Full output

Kol
Iran_N 0.310±0.055
Lebbo 0.539±0.053
Sarmatian_Pokrovka 0.151±0.058
P-value 0.929121312
chisq 6.429
Full output

Malayan
Iran_N 0.192±0.062
Lebbo 0.738±0.056
Sarmatian_Pokrovka 0.071±0.066
P-value 0.774189887
chisq 8.983
Full output

Pathan
Iran_N 0.354±0.038
Lebbo 0.154±0.034
Sarmatian_Pokrovka 0.492±0.041
P-value 0.21064991
chisq 16.757
Full output

Tajik_Pomiri
Iran_ChL 0.319±0.043
Lebbo 0.143±0.035
Sarmatian_Pokrovka 0.538±0.052
P-value 0.835612167
chisq 8.121
Full output

Interestingly, using Caucasus Hunter-Gatherers (CHG) instead of Iran_N, as proxy for Neolithic Near Eastern input into South Asia, also worked, while, at the same time, potentially lowering the ancient Eastern European ancestry proportions (although not if we assume that the early Indo-Iranians moving into South Asia had a higher ratio of CHG-related admixture than the Pokrovka Sarmatians).

But it's impossible to say whether this is meaningful, because I had to remove CHG from the outgroups to add it to the reference populations. Also, this model is less parsimonious geographically, because South Asia is much closer to Iran than to the Caucasus, therefore there's no need to assume at this stage that the Neolithic farmers moving into South Asia were more like CHG than Iran_N.

Brahmin
CHG 0.399±0.060
Lebbo 0.366±0.036
Sarmatian_Pokrovka 0.236±0.071
P-value 0.18924389
chisq 17.224
Full output

Gond
CHG 0.120±0.045
Lebbo 0.880±0.045
P-value 0.735236332
chisq 10.362
Full output

Kalash
CHG 0.527±0.056
Lebbo 0.204±0.034
Sarmatian_Pokrovka 0.269±0.064
P-value 0.434945033
chisq 13.168
Full output

Kapu
CHG 0.404±0.044
Lebbo 0.596±0.044
P-value 0.815283592
chisq 9.241
Full output

Kol
CHG 0.383±0.072
Lebbo 0.602±0.053
Sarmatian_Pokrovka 0.015±0.080
P-value 0.780646206
chisq 8.897
Full output

Malayan
CHG 0.218±0.045
Lebbo 0.782±0.045
P-value 0.773142044
chisq 9.849
Full output

Pathan
CHG 0.517±0.046
Lebbo 0.219±0.031
Sarmatian_Pokrovka 0.264±0.053
P-value 0.954914864
chisq 5.746
Full output

Tajik_Pomiri
CHG 0.387±0.054
Lebbo 0.126±0.035
Sarmatian_Pokrovka 0.486±0.062
P-value 0.780117456
chisq 8.904
Full output

And now, we wait. I don't know when we'll see the first big ancient DNA paper on South Asia. Hopefully this year, and hopefully at bioRxiv first, so we can have a robust discussion about its conclusions before the final version appears in a journal. Apparently the Broad MIT/Harvard team is close to getting something out. If anyone has info about this paper, feel free to post it in the comments.

See also...

Ancient herders from the Pontic-Caspian steppe crashed into India: no ifs or buts

The Out-of-India Theory (OIT) challenge: can we hear a viable argument for once?

Late PIE ground zero now obvious; location of PIE homeland still uncertain, but...

120 comments:

Nirjhar007 said...

Without aDNA nothing can be seriously debated , we can perhaps get some indications at best . But since the Indo-Aryan issue is also a cultural issue , we can be sure that it had nothing to do with Central Asian Steppes .

Hard data of archaeology suggests a continuity from Harappan times , perhaps we will just get the re-affirmation of that fact from aDNA as well, with y-dna and others .

Davidski said...

Without aDNA nothing can be seriously debated, we can perhaps get some indications at best. But since the Indo-Aryan issue is also a cultural issue, we can be sure that it had nothing to do with Central Asian Steppes.

All indications are that what you're saying there is total BS.

Nirjhar007 said...

No its not BS . Study archaeology,ancient texts , genes don't tell much about culture.

Situation is so bad that your friend Mr.Khan , also seeks a possibility, that R1a will be found in Rakhigarhi around 2500 BC , he recognize an earlier 'Migration' to be taken as an option . Perhaps it was a reason you didn't cite that article .

But what is very clear is that a big revelation is underway .

Davidski said...

Can't see the possibility of R1a in Rakhigarhi around 2500BC. I'd say it would be a miracle if it's there even around 2000BC. Just after 2000BC: maybe.

You seem to be paying too much attention to rumors that are flying around. But unless the result is published and it's proven not to be contamination, then sit tight and wait.

Nirjhar007 said...

I would say at least from Early Harappan era we will see those R1a's, if not Neolithic already .

I don't listen much to rumors BTW.

Davidski said...

And where would this Neolithic R1a come from, considering that it's an Eastern European and Siberian forager paternal marker?

Which Neolithic group of farmers carried R1a? Did you honestly clarify to yourself this issue before formulating your awesome theory? Or are you just bullshitting to yourself and everyone else while you're at it?

Nirjhar007 said...

Relax. No one can simply see a massive migration that occurred during BA to S Asia. If they have done studies on the area .

A significant level of ''Steppe like'' ancestry , can only reflect an archaic ancestry and can not be related to any hypothetical event not backed by any hard data ! . Scientifically saying, its only a matter of time, that this suggestion will get directly proven .

I will also say that this CHG/Iran_Neo component, works better to explain the relations that IE has with Afrasian, Sumerian and others .


Davidski said...

The steppe-like ancestry can't be archaic ancestry because it's missing from Neolithic Iran and, as you'll soon see, from Chalcolithic Central Asia and India.

And it's linked to the steppe Y-haplogroup R1a-Z93, which is a very young haplogroup and not native to South Asia.

postneo said...

@David
"Also, this model is less parsimonious geographically, because South Asia is much closer to Iran than to the Caucasus"

No ...
The Iran neolithic Ganj Dareh samples bordering Iraq are almost the same distance to Afghanistan as the caucasus.There is a moist corridor connecting Southeast caucasus, south Caspian and South Asia.

In contrast the connection between the Iran Neolithic is with Baluchistan and Sindh following a southern coastal route.

Intervening areas are Arid and support much less population. This is also evident in modern population density maps.

Samuel Andrews said...

So are Gond people basically pure ASI?

Rob said...

What's the significance of "Sarmartians" in the fit ?

Davidski said...

So are Gond people basically pure ASI?

Clearly not. Max 90%.

What's the significance of "Sarmatians" in the fit?

That a population living on the Late Bronze Age/Iron Age steppe may have migrated to South Asia.

Jijnasu said...

@davidski An Iron age migration to south asia is highly unlikely.

Davidski said...

An Iron age migration to south asia is highly unlikely.

Let's say instead a population or populations living on the steppe well after Yamnaya.

blogspot said...


Nirjhar007 stop apamming. Give up. Its was clear already at start that the old outdated european theories about an Indian origin of the IE were falsificated long time ago.

Samuel Andrews said...

Arguing R1a Z93 existed in pre-BA India and Eastern Europe really is like saying humans originated in Africa and India. There's no ifs ands or buts about it R1a Z93 and R1a in general is "foreign" to India.

Seinundzeit said...

David,

"But importantly, unlike the Indo-Iranians, Dravidian speakers from India - Gond, Kapu and Malayan - could be modeled successfully without any ancient Eastern European admixture."

These qpAdm models make perfect sense, and are well aligned with output obtained using other methods (especially with regard to the distribution of ancient Eastern European ancestry in Central/South Asia. Not to mention the more sensible ENA levels).

I mean, the older models were quite generous with Steppe_EMBA admixture for South Indian "Tribal" and "Scheduled Caste" populations. They were also pretty generous with the ENA/ASI proportions in southern Central Asia.

At the end of the day, having "Scheduled Caste" South Indians at 20%-25% Steppe_EMBA, and the Kalash at 20%-25% ENA, simply didn't gel with any other method.

Based on other analyses, non-Brahmin South Indians are essentially 0%-5% Steppe_EMBA/MLBA-admixed (closer to 0%), and populations like the Kalasha + the "Pathan" samples are (at max) only 15% ASI/ENA (and probably closer to 10%).

Also, the fact that Sarmatian_Pokrovka works best (in the context of Central/South Asia) is a very good sign for this new qpAdm methodology.

As per Fst distances, Sarmatian_Pokrovka is the closest ancient steppe population to South Asians.

In my view, probably because Sarmatian_Pokrovka is essentially a genetic mixture between Srubnaya-related people and people related to the Srubnaya_outlier.

And I'm willing to bet that the Indo-Aryans were, in large part, genetically intermediate between Steppe_MLBA and ANE-rich/extremely eastern (genetically speaking) groups like the Srubnaya_outlier, which explains why the Sarmatian samples do well + are so close to South Central Asians in terms of simple Fst distance.

Anyway, the Lebbo angle is also very interesting, considering the fact that ASI is now understood to be genetically intermediate between the Onge and East Asians.

postneo,

"No ...
The Iran neolithic Ganj Dareh samples bordering Iraq are almost the same distance to Afghanistan as the caucasus.There is a moist corridor connecting Southeast caucasus, south Caspian and South Asia."

CHG show no haplotype linkages with South Asians, and their ADMIXTURE results anchor them near contemporary populations from the Caucasus.

By contrast, Iran_N show very intense haplotype linkages with South Asians, and their ADMIXTURE results anchor them near contemporary South Asians.

With formal methods, one is going real deep into the phylogeny, which explains why CHG works decently in these fits.

But again, haplotype-based analyses, PCA, ADMIXTURE, etc, show that only two streams of Near Eastern ancestry dominate Central/Southern Asia (with some minor exceptions): Iran_Chl-related in the "Stans", and Iran_N-related in Greater India.

Regardless, I do await aDNA from Central/South Asia with great excitement/interest.

Raj Pradip Chakraborty said...

@ David,OIT proponents mainly relies on R1a-M17 bearers among Saharia,Chenchu tribes.which are about 28%in them.It is not determined whether they are carriers of Z94 or not.Did you include tribal R1a data for your analysis?.If not ,then OIT proponents will perhaps not accept AMT/AIT.Unfortunately tribal R1a study heavily based on STRs.Tribal R1a need to be revisited to determine how close they are to Z94 bearers in India.

Nirjhar007 said...

David,

'' you'll soon see, from Chalcolithic Central Asia and India.''

We will see things .

Dave+Sam
''And it's linked to the steppe Y-haplogroup R1a-Z93, which is a very young haplogroup and not native to South Asia.''
''Arguing R1a Z93 existed in pre-BA India and Eastern Europe really is like saying humans originated in Africa and India. There's no ifs ands or buts about it R1a Z93 and R1a in general is "foreign" to India.''


I don't think so . Even if R1a didn't originate in S Asia , it is older than 2000 BC and certainly not limited to Z-93 , a proper study on S Asian population will clarify these things.

blogspot,

I am not an OIT proponent .



Acharya Agnimitra said...

Nirjhar
"Situation is so bad that your friend Mr.Khan , also seeks a possibility, that R1a will be found in Rakhigarhi around 2500 BC"

I read that hilarious article. It was chicken shit of him make such a drastic volte-face so close to the end game, considering how relentlessly vociferous he has been for classical AIT/AMT.

Be assured and forewarned, many more such spontaneous U turns, unilateral and arbritary date manipulations and genetic hair splitting are on their way. Mark my words, two months from now they will be pushing for a ludicrous Mature Harappan entry(though there isn't a snow ball's chance in hell for that)

Even David here will very soon begin to retract like an oyster. But thankfully, the data spanning millennia will not leave gaps in the R1a phylogeny to permit such obfuscations.

Davidski said...

@Acharya Agnimitra

You still seem to be unaware of the simple fact that R1a can't be native to both Eastern Europe and South Asia.

And since it's native to Eastern Europe, because this is proven with ancient DNA from Mesolithic Eastern Europe, then it's intrusive in South Asia.

Simple as that. The rest are potentially interesting details. Deal with it.

aniasi said...

Could there have been two migrations into early-Neolithic India? One would be the Iran-N people along the coast, and the other would be closer to CHG in the North?

The reason I say this is that proto-Dravidian seems to be a poor match for the Vedic Substratum. Yes, there are signs of retroflex consonants replacing secondary articulation and gerunds, but proto-Dravidian would have heavily simplified Old Indic's phonemic inventory, accents, and rhythms.

@Davidski

Could you see if Iran-N suits South Indian castes better, and CHG the Northwestern ones? Or even try both to see if there were two migrations that mixed?

postneo said...

@Sein
"CHG show no haplotype linkages with South Asians, and their ADMIXTURE results anchor them near contemporary populations from the Caucasus."

the role of haplotypes to link CHG is irrelevant if CHG is represented by 2 really old samples on one hand and you compare with modern south asians.

there are some shared haplotypes btw kura-araxes, maikop and south asia.

Davidski said...

South Asians do show a very strong genealogical relationship to Neolithic Iranians from the Zagros Mountains via genome-wide haplotypes.

http://eurogenes.blogspot.com.au/2016/07/early-neolithic-genomes-from-eastern.html

So if we're to pick either the CHG or Iran_N model as an exclusive one, then the Iran_N model is it.

By the way, no one has yet done such a haplotype analysis of Kura-Araxes, Maikop and South Asians, so posteno is referring to something else.

postneo said...

@aniasi
"The reason I say this is that proto-Dravidian seems to be a poor match for the Vedic Substratum. Yes, there are signs of retroflex consonants replacing secondary articulation and gerunds, but proto-Dravidian would have heavily simplified Old Indic's phonemic inventory, accents, and rhythms."


The metrical complexity in spoken dravidian dwarfs that of any other world language. It's innate, even more evident in percussion.


That said dravidian is not a great fit as a substratum for Vedic on lexical and other grounds.

Retroflexion in Vedic and Sanskrit is not from Dravidian either. simple words like pATha

Chetan_Vit said...

@Nirjhar007

An Outside India origin for the Indo European langauges is very obvious from the linguistic evidence alone. There is absolutely NO way that Sanskrit in India was the ancestor of all IE languages. So even if there was some R1a - Z93 in the samples before 2000 BC, I am sure most academics would consider it a very early isolated migration from the steppes

postneo said...

@Chetan_Vit said...
There were other IE languages in S Asia. Sanskrit cannot fully explain IE languages within South Asia.

@davidski
I was talking about shared uniparental markers

Chetan_Vit said...

@postneo

"There were other IE languages in S Asia"

Evidence?

Ryan said...

@Nirjhar007 - The only people who doubt there was a bronze age migration from Central Asia into India are people who have ideological reasons to doubt it. That in and of itself should tell you you are barking up the wrong tree. Only the impotent are pure, and your ancestors were not impotent. There isn't a single culture in the world that wasn't build on the bones of another.

@Chetan_Vit - Tocharians maybe?

@David - In terms of why the Lebbo are better proxies for ASI and ENA ancestry... maybe the Onge have some specific drift from being stuck on an island for 8-10k years? Kind of like why using Ashkenazi Jews or Finns isn't the greatest idea?

And I'd think the Han would probably have acreted some northern influences that wouldn't be reflective of any recent East Eurasian ancestry in South Asia?

I dunno just spitballing here.

Lee Albee said...

The Onge and Han may not be working because you have Ust_Ishim in your right population. Various analyses have shown that Onge and Han may have related heritage with Ust_Ishim.


Using Ust_Ishim as right population could be skewing your data as I think some data suggests a related population may have given rise to ASI.... I will see if I can substantiate that with a reference or two later

Just something to think about

Acharya Agnimitra said...

@Chetan Vit

There is a profusion of indications that there were other IE languages in South Asia. If you were familiar with the historical verses of the Rig, it would be clear to you. There is so much information in it if only you are curious enough to look.

1- The fact that Vedic Sanskrit was spoken ONLY by the Paurava tribe who dwelt in Haryana and Western Uttar Pradesh. THEY and ONLY THEY patronized and composed the Rig Veda. And they were NOT familiar with any territory outside this region in the oldest books.
2- The fact that the other four Arya tribes were located in the cardinal directions around the Sanskrit speaking Paurava. Together they were known as Panchajanya(Five peoples/tribes)
3-The fact that the one of the five Arya trbes, the Anava conglomeration of the Punjab region , is easily identified to be the proto-Iranians. Even Michael Witzel was forced to accept that the attested Parthava and Parsava enemies of the Paurava Indo-Aryans are none other than the Proto- Parthians and Persians. Obviously they did not speak Vedic but proto-Avestan. The other three Arya tribes, ostensibly, did NOT speak sanskrit.
4-The fact that Rig Veda mentions many hostile Indo European tribes with CLEAR IE NAMES but who are NOT called Arya or even Panchajanya such as Udavraja, Brishyas, Paravatas, Pani, Ajas etc

5-The fact that Bangani, a centum language was discovered in Uttarkhand. IT IS TOTALLY inexplicable by current models.
6- The fact that Sinhalese language of Sri Lanka is almost certainly NOT OF IA ORIGIN, as argued by linguists such as M.W.S. de Silva and alluded by Wilhelm Geiger. It has MANY words and features that are NOT SANSKRITIC and have no parallels with North Indian dialects. SHOCKINGLY,it shares features with such widely separated branches like Hittite and Germanic. For example, the Sinhalese word Watura bears nothing with the Sanskrit word Udaka for water.
7- The fact that Tocharian is textually attested even in Vedic(Brahmanic) times as the Uttarakuru located north of Kashmir.
8- The fact that Vedic tradition attests the existence of culturally and linguistically non Indo Aryan groups such as Vratyas and Mrdravaks('they of faltered speech')
9- The fact that the hostile non Arya Kikatas mentioned in the Rig Veda who lived in the Bihar region and who later became the Magadhans with their Magadhi dialect, is the first place where Prakritic literature was attested, showing they were not sanskrit speakers.
10- Meillet, M Deshpande etc have shown the inferred presence of dialects in Eastern India with many non-Indo Iranian features(such as initial l instead of r) which were absorbed into Classical liturgical sanskrit.

I could go on and on with this. But the crux of the matter is that the various prakrit languages which have for centuries been belittled by poor,stagnated scholarship as mere corrupted forms of Sanskrit is no longer valid. They are descendants of SPEECH FORMS WHICH EXISTED SIDE BY SIDE WITH RIG VEDIC SANSKRIT. These are long lost Non-Vedic IA/IE languages.

Infact, I dare say, perhaps none of the modern Indian languages are directly descended from Sanskrit. But all of their vocabulary has been profusely inundated by Sanskrit. They are all Prakrits by descent. Vedic Sanskrit(Not the classical) is a ELITE, FROZEN, PERFECTLY PRESERVED LINGUISTIC SPECIMEN FROM THE BRONZE AGE. It is the equivalent to a fossil unearthed by a geologist.

Ryan said...

@Acharya Agnimitra - I don't have the interest in checking all of your claims, but this claim stuck out as so obviously wrong that I felt it had to be addressed.

The fact that Sinhalese language of Sri Lanka is almost certainly NOT OF IA ORIGIN, as argued by linguists such as M.W.S.

I'm not even sure where to start with this. Sinhalese is an Indo-Aryan language and anyone arguing otherwise is just wrong. Find new sources.

You're also taking the term "Aryan" too literally here, and the random use of ALL CAPS does not help your credibility.

Acharya Agnimitra said...

@Ryan

"You're also taking the term "Aryan" too literally here"

For someone who cannot even tell the difference between the modern racist misnomer "Aryan" and the Rig Vedic word "Arya", for someone who cant even get the spelling right, you are quite demanding.

Of course you have no interest!I know. You are content like a frog in a well with the childish, simplistic story of a group of sanskrit speaking Kurgans invading and Indo-Europeanising all of North India in a couple of centuries. And lo! All Indian languages were born thus! Anything that shakes the convenience you feel in that will not be graced with your interest.

But unfortunately, that story does not hold in light of the above and a host of other varied kinds of evidences. Nevertheless, I had no intention of convincing anyone(I know it doesn't work that way with you folks). I only wanted to give the person who asked the question a glimpse of what he is overlooking. I believe I have done that.

Nighty nighty

Ryan said...

@Acharya Agnimitra - For someone who cannot even tell the difference between the modern racist misnomer "Aryan" and the Rig Vedic word "Arya", for someone who cant even get the spelling right, you are quite demanding.

The linguistic term is Indo-Aryan.

Kurgans invading

Kurgans are piles of rocks. People do the migrating, not rocks, and they would have spoken an ancestor to Sanskrit, not Sanskrit.

Get your facts straight man.

And lo! All Indian languages were born thus!

Not all of India speaks Indo-European languages, so maybe you should roll back your outrage a bit.

Anything that shakes the convenience you feel in that will not be graced with your interest.

No, half-assed arguments from someone who doesn't spend the time to even learn the basics of the fields they are arguing about is what doesn't keep my interests.

My ancestors were assholes on horseback too. They assimilated large swaths of Europe in only a few generations, just like yours did to India. Our pre-history is the same. Get over it.

aniasi said...

@Agnimitra

1) The Bhrgus definitively use the language of the Rg Veda, and the same can be said of the other groups mentioned in the battle of the Ten Kings. The only identification of them as separate peoples comes from attempts to use later names of areas and groups as cognates.

2) The Bangani thesis has been disproven. Even if there was a Bangani centum substrate, it could easily have come from Tocharian or Para-Tocharian.

3) Sinhalese has been heavily influenced by a substrate language, as well as archaic Tamil forms (to mainland Tamilians this is seen as very comical in both Sri Lankan Tamil and Sinhalese speakers)

4) Magadha was one of many states in the Kikata region, and we cannot assume that the Kikatas represent all the inhabitants of the area. Prakrits existed throughout India, as attested by Panini, and Magadhi stands out earlier only because Magadha was the first state to quickly organise as an Empire, probably since they could support larger populations with their control of the many branches of the Ganga river.

aniasi said...

@Davidski

I read that before, but got a bit confused about whether that was out of date, especially with the posts about Iran-Chalcolithic expansions.

Do you mind testing to see if there was a pulse of both Iran_N and CHG in a few Northern and Southern castes? It would be interesting.

Chetan_Vit said...

@Acharya Agnimitra

The fact that Bangani, a centum language was discovered in Uttarkhand. IT IS TOTALLY inexplicable by current models.

Bangani has been accepted to be a pure satem language. The theory of Bangani having a centum substrate (not being a centum language like you falsely claimed) were based on some spurious findings and have been discredited since then. Furthermore even if Bangani had four or five centum words, that much more likely indicates the influence of a western centum language rather than pointing to an out of India origin of IE.

The fact that Sinhalese language of Sri Lanka is almost certainly NOT OF IA ORIGIN

Sinhala has a lot of non Sanskritic words so what? Every IA language and Sanskrit itself has a significant percentage of non Indo European words and grammatical elements.That only serves to prove my point. The Old Indo Aryan language was foreign to India at first brought in by some IE migrants and later adopted by the population in the subcontinent. In the process the Old Indo Aryan speech was changed by the addition of indigenous features. I am going to have to ask you to give some reliable references that Sinhalese words have Germanic origins.

The fact that Tocharian is textually attested even in Vedic(Brahmanic) times as the Uttarakuru located north of Kashmir.

EVIDENCE?????

The fact that the one of the five Arya trbes, the Anava conglomeration of the Punjab region , is easily identified to be the proto-Iranians.

Yes the Vedic tribes may have interacted linguistically with some of the Iranian tribes at a very early stage. That is to be expected since the date of the Indo Iranian split is estimated as late as 2000 BCE.

the inferred presence of dialects in Eastern India with many non-Indo Iranian features(such as initial l instead of r) which were absorbed into Classical liturgical sanskrit.

L / r isogloss is a well known feature of eastern / western IA dialects. Does nothing so ever to support your cause.

I could go on and on with this.

You say you can go on and on but none of the "evidence" you give is remotely consequential enough to upturn the huge amount of evidence in favour of an Outside India origin

1)At least 3 IE subfamilies in Europe and the Caucasus / Middle East region.Only one subgroup in India

2)Presence of naive IE terms for flora fauna and environmental features that are foreign to the Indian subcontinent but present in the steppe environment. OTH a lack of native terminology for prominent Indian plants, animals and other features.

3) Presence of wheeled chariot and horse riding terminology in abundance when the horse is NOT an animal native to India and no horse remain found in India till after 2000 BC

4) The historical fact that almost every migration has been from the north west into India and almost never in the opposite direction.

5) Finally the huge amount of archaeological and genetic evidence in support of the Kurgan hypothesis.

FACE IT YOU ARE TRYING TO FIGHT A LOST BATTLE. THE ONLY PLACE WHERE THE OUT OF INDIA THEORY IS TAKEN SERIOUSLY IS WITHIN INDIAN NATIONALISTIC CIRCLES

andrew said...

I've collected a lot of the available genetic and anthropological information about the Lebbo' people in a blog post. for anyone who is interested.

Acharya Agnimitra said...

@aniasi

ONE branch of the Bhrgus shifted loyalties from Iranian to Indo Aryan and used Vedic to compose hymns which were included in book 10. It was the Jamadagni branch. The majority of Bhrgus led by the Atharvan branch served as allies and fire priests to the Anava (Proto-Iranans) It is these Bhrgus who appear as enemies of the Paurava in the Battle of the ten kings.

Even the Jamadagni branch show signs of their previous proximity with Iranians.In Vedic, the name should be "Gamadagni". The G-->J shift s visible. And is son is curiously called "Parasu Rama". Although later Puranas say it s because he wielded a battle axe, a more plausible interpretation would be that he was associated by the Indo Aryans with the Parsu(Proto-Persians)

Also note, that the Indo-Iranians have not split into separate branches during the time of Jamadagni, but conflicts had begun and phonetic changes were happening.

Bangani's origin from a remote Centum language has not been disproven. Ever since Claus Zoller discovered it, there was concerted effort to suppress its destructive implications. But his exhaustive work on it and also the counter rebuttals by those like Anvita Abbi to the detractors have more or less sealed the case. No, it could not have 'easily' come from Tocharian. There is nothing to show for it. Tocharian itself is a huge anomaly in the centum-satem split which was used to prove AIT/AMT. But having another centum tongue inside India destroys it.

How and when Sinhalese crossed an ocean and reached the island itself is far from clear. Added to this is a lot of words that are neither Sanskrit nor Tamil in origin. They are neither loan words for derived. Even more confounding are the words shared with branches like Hittite. Three only three notable scholars who have worked and stumbled on its uniqueness have said in unison that it is difficult to place it in the Indo-Aryan group. All three of them( de Silva, Geiger, S. Paranavitana) suggested that the proto-Simhala people arrived in the Island from North West India, specifically from the upper Indus Valley.

Your fourth para states a few facts, none of which seem to contradict anything I said. The Kikata were the easternmost people mentioned in the RV. They were not the whole of the IE population east of the Vedic territory, but they were the most powerful and they gave rise to Magadha of the Iron age.

Jijnasu said...

@acharya agnimitra
I find the idea that the anavas were 'proto-iranians' or that the druhyus were 'most other indo-european sub-famillies' problematic. It seems more meaningful that they represent different indo-aryan tribes along with the purus, yadus and ikshvakus. One would expect an iranian substrate across the punjab and northern sindh and possibly even some evidence for the influence of other indo-European groups in gandhara. However hydronymy and toponym suggests that Indo-Aryan was spoken even more widely further north-west than it is at present. Also the north-west was known to set the standard for Indo-Aryan speech being in many ways very conservative

Salden said...

Who knew Hindu Nationalists were also Fundies? Looked for Noah's Ark? Did the Ancient Indians and Isrealis find dinosaurs?

andrew said...

With regard to the Gonds it is useful to observe that this population is currently classified as a "Scheduled Tribe" rather than as a "Scheduled Caste" or another kind of Caste in India's socioeconomic system.

Ryan said...

Andrew - Your post on the Lebbo just raises further questions lol.

David - Which Pokrovka are the Sarmatian samples from by the way?

LPoropat said...

Is it possible that will be found here-and-there R1a samples in India dated before 2000 BCE, similar to recently found R1b samples from Serbia dated before 9000-8000 BCE, several thousand of years before the cultural expansion of IE people from the Eastern Europe (Kurgan Steppe) to Western Europe? I mean, before the main cultural expansion and conquest, like after 2000 BCE in India or 4500 BCE in Europe, there's always, especially in previous thousand years old periods, a long period of small intrusion&migration due to regional proximity. Perhaps the IE tribes were more widespread in Asia than previously thought? Is it possible that such much ancient samples could be manipulated by the Indian nationalists/leftists, intentionally misrepresenting their context?

Ryan said...

@LPoropat - There's a pretty big difference between "before 2000 BCE" and 9000 BCE. I think it's really damned unlikely you'll find R1a in India much before 4000 years ago though. There isn't regional proximity here and human populations seemed to be pretty isolated before the Holocene. Even an intrusion of ANE folks probably would have carried R2 instead. People can misuse whatever they want but I don't think that really matters.

Davidski said...

@Ryan

Put Sarmatian in the search box above.

@LPoropat

R1a and R1b are old North Eurasian paternal markers, but they were only widespread across North Eurasia until the Bronze Age.

There's no chance that there will be any R1a in Neolithic samples from South Asia or even Central Asia south of the steppes, unless the samples are contaminated and/or affected by other types of lab errors.

DDeden said...

Please correct: Malayan -> Malayalam. They are utterly different.

P Piranha said...

Truly great that this result obtains. We already knew for a long time that the population most assymetrically related to high caste and low caste members of the same linguistic group in India are Andronovo, Sintashta and following that modern Central and Northern Europeans, such as Icelandic, Orcadian, Lithuanian and so on, but not Yamnaya. In pca of D-stats results Yamnaya load equally between upper and lower caste persons, while Europe_LNBA, Steppe_LNBA and Northern Europeans load on upper caste samples and Iran_N and Middle Eastern populations load onto lower castes . This was always very difficult to square with the idea that the only input in from the Steppe into India was from Yamnaya.

Furthermore only one set of high quality ancient steppe samples shares high levels of IBD with both Indians and Europeans, all other steppe samples don't. That's Srubnaya.

SRUBNAYA
Ukrainian_Lviv Average 2.9312005
Tajik Average 3.004709757
Dharkar Average 3.010097
Polish Average 3.041164153
Kalash Average 3.04122627
Slovakian Average 3.094962107
Estonian Average 3.240149929
Croatian Average 3.338848475
Pathan Average 3.342796418
Lithuanian Average 3.4186915
Latvian Average 3.53712965
English Average 3.56383166
Danish Average 3.673768075
Kshatriya Average 3.71316304
Orcadian Average 3.964734
Irish Average 4.009592884
Swedish Average 4.041079083
German Average 4.153687709
Ukrainian_Belgorod Average 4.490012533
Ukrainian_Kharkov Average 6.252296
Ukrainian_Poltava Average 6.3027062

Hopefully if the authors of the new paper use 16-population qpAdm--there's no reason why they shouldn't--we will get an answer substantially similar to this one, and one of the last strikes against the Kurgan hypothesis--that upper caste Subcontinentals favour Yamnaya instead of Europe_LNBA, simply falls away. Maybe the evidence will shift further to confirm that Steppe LNBA descendants of Sintashta and Andronovo type populations (e.g. Samartian Pokrovka) are the best proxies of Steppic input into India.

Davidski said...

@DDeden

Please correct: Malayan -> Malayalam. They are utterly different.

The Dravidian samples I have are labeled Malayan. Not up to me to change such labels.

postneo said...

@chetan
"Bangani has been accepted to be a pure satem language."

that was a decade back. Its finding growing acceptance as Kentum. However its not just about kentum, the verb ordering is different from Sanskrit and IA. Its not isolated but is related to Kashmiri and Dardic languages.

http://himalaya.socanth.cam.ac.uk/collections/journals/ebhr/pdf/EBHR_31_06.pdf

aniasi said...

@Agnimitra

That does not make sense. You present no evidence, except for a cognate list. By that reckoning, Parasu-Rama might be Perseus, Parisian, or Parthian. Any association with Iranian locations is Puranic, distancing itself from the Battle of the Ten Kings by at least a thousand years, if not more.

The Rg Veda presents no evidence of a different language, or different identity. Jamadagni isn't mentioned, only the Bhargavas as a whole, and they are certainly IA. Stop mixing the Puranas, cognate-comparison, and the Rg Vedic account.

Also, concerning a centum substratum, its pretty clear that it arrived with Tocharian. Quite possible considering the connections between India and Central Asia and the Tarim Basin.

In fact, that is Witzel's theory: http://www.people.fas.harvard.edu/~witzel/Sinttashta_Qu.pdf

Also, we have a good idea how it arrived, by sea. Yes there will be non-Dravidian words, because not all of the island was Tamil speaking. The substratum appears to be related to Vedda.

@Jijnasu

Unlikely. Indo_Aryan was the first language in the region, and forms a substratum in Iranian. Hydronyms indicate a that the NW region was IA long before it was Ir, and Ir itself has a number of words that are IA 's' contributions instead or Ir 'h'

https://openaccess.leidenuniv.nl/bitstream/handle/1887/2691/299_054.pdf

aniasi said...

@DDeden

The Malayans are also a tribal people in Kerala. The name is from the same Tamil word for hill (Malai/Malay) that gives us the name Malayalam.

Jijnasu said...

@aniasi
I agree. By one would expect, I meant these substrates would be expected if talageri's hypothesis was correct

andrew said...

@Ryan Unfortunately, in this field, questions are cheap and answers are expensive.

Jijnasu said...

@Acharya Agnimitra
Also it is extremely unlikely that the sinhalese were from the extreme NW/ upper Indus. The one distinguishing feature of NW IA is that it preserves all three sibilants of OIA which either merged with the dental sibilant (central and southern dialects) or with the palatal sibilant (eastern dialects). Nor does it preserve many of the consonant clusters. links with the east and south west seem far more likely

postneo said...

@Chetan
There are many strange every day IE forms in ordinary NIA that cannot be explained by Sanskrit.

Take the word for bone. haDDI(Hindi), hARa(outer NIA) etc... the closest forms are is Kurdish and Hittite and not Sanskrit which is closer to greek osti and latin ossi. you cannot really derive these words from Sanskrit.

there are many

@Acharya Agnimitra
I would not say that Sinhala are non IA. However being an exotic isolated population who branched they will preserve traits not found in others.


Nirjhar007 said...

Ryan,

My ancestors were assholes on horseback too. They assimilated large swaths of Europe in only a few generations, just like yours did to India. Our pre-history is the same. Get over it.

Thanks for the zen wisdom Ryan :) . But I don't consider them as such and also, horse back riding is late 2nd millennium BC phenomenon with earliest depictions instead coming from 2000 BC Near East IIRC.

Acharya Agnimitra,

I quite agree with you but I don't support OIT now . I was a proponent of OIT few years back and before that I was a stark proponent of AIT also !. I now consider PIE to be from around Near East -West Asia and propose that India was Indo-European at least from Early Harappan period or 4th millennium BC at least.

About linguistics . for you and other soldiers here ;) . I am giving an interesting article , I do not fully agree with it but its more or less sound and robust :

Some unlikely tentacles of early Indo-European

(ch.3 in Angela Marcantonio& Girish NathJha, eds.: Perspectives on the Origins of Indian Civilization, Center for Indic Studies, Dartmouth MA, and DK Printworld, Delhi, 2013, from Dartmouth conference 2011)
http://www.academia.edu/20086861/_Some_unlikely_tentacles_of_early_Indo-European_published_as_ch.3_in_Angela_Marcantonio_and_Girish_Nath_Jha_eds._Perspectives_on_the_Origins_of_Indian_Civilization_Center_for_Indic_Studies_Dartmouth_MA_and_DK_Printworld_Delhi_2013_from_Dartmouth_conference_2011

Its quite rich . I hope many will find it interesting .

Chetan_Vit said...

@postneo There are many variant and non conforming forms in different IA dialects. The problem seems to be because the development of Indo Aryan took place in an environment that was distinct from any other IE language. The subcontinent had a significant (in numbers) population speaking languages belonging to many extant and extinct language families. People who adopted the Indo Aryan language brought their own mother tongue's characteristics from multiple languages into MIA and NIA. We should expect to see divergent forms that are not explained easily by the normal rules of word change.

But my question to you is this : Just because a few words stand out as anomalies, what makes you think that you can resort to an origin from languages spoken thousands of miles away? (such as Hittite you say)

We have no evidence to think that Hittite was spoken anywhere in the vicinity of the Indian subcontinent. This is mere grasping at straws. The mere anomaly of a few words do not defeat the overwhelming case that almost the whole of IA vocabulary can be derived from Sanskrit according to the regular rules of derivation. A language family is defined by its rules not by the exceptions

Rob said...

Well the cards are on the table, and it'll be great to finally see the results, either way
It'll help put some final details on the variations of models we have been hypothesising
As my last prediction (not that I'm deeply knowledgeable about South Asia), as an alternative to the parsimonious Steppe EMBA model, it might be possible that there wasn't any (major) migration from the western steppe to India, such that Gujarat and Punjabs, eg, are essentially local ANE, Iran Neolithic and "south Asian forager". The slight WHG they might pick up could be explained by presence of mtDNA lineages such as U2e, U4, U5a seen in south Siberia since 5000 BC (as Kristiina points out). This would create a sort of EHG as that seen further west, but a "hyper-EHG".
And against a more western origin is that it seems Indians do not have an affinity to Barcin Neolithic, only Iran Neolithic.

Iranic speakers like Tajik, on the other hand, definitely have more Western / ANF, via from BMAC and historical Scytho-Sarmatians.

Davidski said...

@aniasi

Do you mind testing to see if there was a pulse of both Iran_N and CHG in a few Northern and Southern castes? It would be interesting.

I can't do that with this method, or in fact any method that I can think of, because Iran_N and CHG are too similar. CHG is just less basal by a few per cent.

@andrew

Thanks for the Lebbo write up. I was wondering who they were.

Jijnasu said...

@Postneo
Extant Vedic and post-vedic Sanskrit literature offer only an incomplete picture of old-indic vocabulary , being particularly biased towards the usage of the upper classes of N.W and central India. There are several reconstructible O.IA forms that remain unattested in an text yet appear first in the middle indo-aryan stage and are universal in the modern languages. The word in question MIA haDDa (in NIA the dimmunitive with a feminine -i ending sometimes used) has a pan Indian distribution. On what basis do you connect it to the Hittite form though? it might be an entirely different (pre-IA word). The word 'watura' cited as an evidence of an anatolian substrate in sinhalese seems to have initially meant flood or storm and likely derived from OIA vAtUla or storm.

Nirjhar007 said...

@Rob

Well the cards are on the table, and it'll be great to finally see the results, either way
It'll help put some final details on the variations of models we have been hypothesising
As my last prediction (not that I'm deeply knowledgeable about South Asia), as an alternative to the parsimonious Steppe EMBA model, it might be possible that there wasn't any (major) migration from the western steppe to India, such that Gujarat and Punjabs, eg, are essentially local ANE, Iran Neolithic and "south Asian forager". The slight WHG they might pick up could be explained by presence of mtDNA lineages such as U2e, U4, U5a seen in south Siberia since 5000 BC (as Kristiina points out). This would create a sort of EHG as that seen further west, but a "hyper-EHG".
And against a more western origin is that it seems Indians do not have an affinity to Barcin Neolithic, only Iran Neolithic.

Iranic speakers like Tajik, on the other hand, definitely have more Western / ANF, via from BMAC and historical Scytho-Sarmatians.
.

Makes sense :) .

aniasi said...

@Davidski

I didn't realise they were that close.

ak2014b said...

"Situation is so bad that your friend Mr.Khan , also seeks a possibility, that R1a will be found in Rakhigarhi around 2500 BC , he recognize an earlier 'Migration' to be taken as an option . Perhaps it was a reason you didn't cite that article ."

This turns out to be a reference to Razib Khan. I searched for the article that's alluded to. I'm blown away. It's really the case that Razib has tried to go back on his own original predictions from not so far back, just January 2016. At that time, Razib Khan said of the anticipated Rakhigarhi results: "I predict that the Y chromosomal haplogroups will be H or J2" and strongly associated them with Dravidian, as indicated in the title https://www.unz.com/gnxp/the-dravidian-migration-theory-vindicated/

So now he's creating the impression that he's argued for R1a in South Asia at 3000-2500 BC all along, and that instead of Dravidian, he had envisioned steppe waves ending up in the Indus Valley civilisation at that time? That's not what he had originally been arguing, is it?

I figure he must have picked up some rumours that resulted in him totally overturning his prior claims (or he's just hedging his bets), so he doesn't come off looking as having been both clueless and prejudiced against South Asia, just in case his original prediction of an exclusively H and/or J2 civilisation wasn't vindicated by the results. Especially in the, by all accounts so far, unlikely event that R1a pops up somewhere in South Asia before 2000 BC.


@Salden
"Who knew Hindu Nationalists were also Fundies? Looked for Noah's Ark? Did the Ancient Indians and Isrealis find dinosaurs?"

Right? But that's not even all of it. There's christian fundies also proliferating here.

I was going to commend Chetan Vit for demanding evidence, but he turns out to be a self-professed Christian Theist, who's actually propagating dangerous nonsense, as well as the usual resorting to pseudo-science in order to dress his chosen fables up as history. Like others of that sort, he will retreat to the circular logic that his religious book is proof of the fables it contains, or to known forgeries and other arguments of faith when it comes to providing evidence for his own pet theories. So his asking for evidence only in specific situations could end up having to do with his religious point of view or interests in those topics. Can't be ruled out at all. He actually tries to sound clever by saying that evolution by natural selection and the rest is not parsimonious and unlikely. Bringing up Noah's Ark is very apt.

And earlier, velvetgunther was essentially propagating origin myths exclusive to a Middle-Eastern Christian denomination in South Asia, about one of the disciples, Thomas, and how he was to have arrived in South Asia in the 1st century. No other Christian denomination allows for this, with some going on record to deny it. The rest of humanity has had no evidence for disciple Thomas or any of it and doesn't subscribe. Absolutely no one who claims to adhere to facts would bring up such confabulations as possible history unless it were part of their specific fundy beliefs.

There's clearly multiple creepy religions and ideologies invested in such things and becoming more active here, and everyone must be more wary of all their spokesmen and the unapparent reasons that might actually be making them argue for one thing or another.

Acharya Agnimitra said...

Nirjhar,

Thank you for the reply.

"I now consider PIE to be from around Near East -West Asia and propose that India was Indo-European at least from Early Harappan period or 4th millennium BC"

I do not know your reasons, but I trust that they are grounded on facts. I suspect that perhaps discontent with just PIE, you are on a Nostratic trail.

We are agreed on the 4th mBCE part. As for me, there is recorded IE history and then there is prehistory. And that history does not go any further back into the past than the few precious hymns talking about the five tribes.

And as you agreed, there were many IE languages spoken in SA, apart from Sanskrit. The true implications of this cataclysmic fact struck me only recently. If the reconstructed PIE and the elaborately manufactured steppe hypothesis only took into consideration the 12 recorded IE languages, thinking that the number was near total and workable, almost burying the many unrecorded and lost 'Para-IE' dialects forever, the resultant conclusions are bound to be complete hogwash. Though churned and mixed beyond recognition, thankfully, we may luckily find their traces in their modern descendants.

As for Pre-PIE and your suggestion, I have not gone into such depths and cannot give an opinion. But I will say this, that if the coming results show continuity into the neolithic and Mesolithic, my opinion would solidify.

Jijnasu said...

@salden
Hindu support for an OIT isn't as much about fundamentalism as much as it is a reaction to the racism which colours many popular and until fairly recently even academic versions of the AIT. In addition the fact that it is often used by leftists and missionaries to widen societal rifts add to the problem

Jijnasu said...

The archaeologist Erdosy notes that certain central asian traits intrude upon the Indo-Iranian borderlands at the end of the mature harappan phase, as do certain cultural traits associated with Indo-Aryans. He says some evidence for small scale migration does exist in the late 3rd/early 2nd millenium BCE, thus archaeology too seems more support an early migration of 'yamnaya like' people into South Asia than a late Bronze age post-andronovo migration.

ak2014b said...

@Rob

In that case, how would we be able to detect the Indo-Iranian migrations and from where the languages were brought and at what time it entered South Central Asia and South Asia? What should be the genetic signatures we're expecting to find?

@Nirjhar007
"horse back riding is late 2nd millennium BC phenomenon with earliest depictions instead coming from 2000 BC Near East IIRC."

To be precise, and I'm aware I've already brought this up before: horse-riding was already present in Iran during its Jiroft civilisation (terminus ante quem), where there is definite evidence for horse-riding warriors. The phenomenon does not appear on the steppes, until the end of the 2nd millennium BC, according to Kuz'mina and others. A gap of around a thousand years, if not more.

Ryan said...

@David - Searching didn't help.

f we are to take these qpGraph models fairly literally, and I don't see why not, since they're very tight fits overall, then the early Sarmatians from what is now Pokrovka, Russia

There are several dozen towns named Pokrovka in Russia. Which one does this refer to?

Jijnasu said...

@ak2014b
I believe there is a group of archaeologists who trace beginnings horse riding to the botai culture while others see know evidence for horse riding until much later. Mounted warfare though was definitely a later phenomenon

Anthro Survey said...

@David and Rob

"Also, this model is less parsimonious geographically, because South Asia is much closer to Iran than to the Caucasus, therefore there's no need to assume at this stage that the Neolithic farmers moving into South Asia were more like CHG than Iran_N."

Iran_N people probably shared the plateau with ANE foragers at one time. Over the millenia, the two lineages began to mix and there was a pull towards the CHG direction.(After all, CHG is Iran_N+extra ANE) We see this when comparing early Iran_N to later samples.

It would be quite analogous to the increase of Villabruna-related ancestry in Neolithic Levantines relative to Natufians. Likewise analogous to the the "absorption" of extra WHG along Europe's Atlantic Facade regions
during the late Neolithic/Chalcolithic.

Some copper-age DNA from Hissar, Kelteminar, etc. would be nice to investigate this.


Chetan_Vit said...

ak2014's attempts to discredit the commentators here just reflects a common feature that is prevalent among supporters of OIT. Which is to cast racist and religious aspersions on the motivations of AIT supporters when they throw the evidence on their face. Why can't you face the facts head on for once OIT supporters, instead of resorting to personal and religious attacks. Why can't you bring yourself to accept that horse back riding nomadic pastoralists crashed into India and subjugated a large section of the population? Could it be because they feel that the final outing of the truth that has long been kept suppressed by the nationalists in India is going to seriously damage their political interests. It's easy to guess

Seinundzeit said...

David,

Iran_Hotu is also quite similar to Iran_N, but with extra ANE (by contrast, CHG differs from Iran_N due to some sort of European affinity; basically, a combination of extra WHG/EHG-like ancestry, and an Anatolian pull), and is the geographically closest ancient sample (which we currently have) to Central/South Asia.

So, I think one possible way to test the "local ANE" scenario might be to use Iran_Hotu, for Brahmin_UP.

With that in mind, do you think you can model UP Brahmins as a mixture between Iran_Hotu, Sarmatian_Pokrovka, and Lebbo?

Also, just to see what happens, could you eventually try Iran_Hotu, Sarmatian_Pokrovka, and Agta?

As always, only once you find the time to do this, and only when you have the inclination to do this.

No rush whatsoever.

Thanks in advance.

Salden said...

http://www.cell.com/ajhg/fulltext/S0002-9297(17)30291-4

The Zoroastrian study is now officially published.

Seinundzeit said...

aniasi,

When one applies scaling to the Global_10 PCA, the results become very consistent, and always turn out to be quite reasonable.

Based on that data, the West Eurasian ancestry in South India is Iran_N + extra ANE (so, rather like Iran_Hotu).

In North India, the extra ANE becomes less prominent, but is still perceptible, along with some admixture from the Eurasian steppes.

In northwestern India/eastern Pakistan, Iran_N + Steppe_EMBA/MLBA becomes sufficient.

In southern Central Asia (northern/northwestern Pakistan, Afghanistan, Tajikistan, etc), Iran_Chl becomes quite important, with a gradient forming from eastern Pashtuns (12%-25% Iran_Chl, along with substantial Iran_N) to the Tajiks (only Iran_Chl, no Iran_N).

Basically, CHG ancestry is not to be found in this part of the world, except via something Iran_Chl-related (which is already known to have some sort of CHG affinity).

This is also obvious from looking at genealogical relationships between Iran_N and South Asians vs CHG and South Asians (please refer to the original Broushaki et al. paper).

Also, ADMIXTURE fails to detect much of a South Asian affinity for CHG, while Iran_N do look rather South Asian in those kinds of analyses.

Although, CHG do cluster with the Brahui/Baloch, on the first two dimensions of a West Eurasian-only PCA (please refer to David's West Eurasian PCA).

This is probably due to the fact that CHG, like contemporary South Central Asians, deviate from Iran_N in a "European" direction.

I mean, CHG has some extra European-like ancestry (some WHG/EHG and Anatolian stuff), while South Central Asians have Steppe_EMBA/MLBA admixture, which creates an artificial similarity, and also probably explains why the levels of steppe admixture for South Central Asians take such a big hit when one uses the CHG samples.

Still, nothing is certain, until we have aDNA from India and the "Stans". Thankfully, the data is coming, very soon.

Rob said...

@ AK2014b

"In that case, how would we be able to detect the Indo-Iranian migrations and from where the languages were brought and at what time it entered South Central Asia and South Asia? What should be the genetic signatures we're expecting to find?"

I'm not quite understanding your question.
But I guess the key is the EHG/Iran Neol mix for IA
With Iranics like Tajiks having additional layer of something CHG/ ANF like. I've pointed this before and others noted also. Some suggested it was because the Indics "went around BMAC" thus not absorbing it, although that's a bit hard to grasp. Instead too me it seems like a later layer associated with the implied west Asian influences in BMAC formation.

Davidski said...

The pseudo-steppe theory, which posits that the Yamnaya-related ancestry in South Asians is not really from the steppe, but native to Central and South Asia, really doesn't work, and this should be obvious to everyone by now.

Indo-European speaking South Asians, and especially those from the upper castes, share a lot of genetic drift with Yamnaya that is specific to Bronze Age Eastern Europeans and their offshoots in Central Asia, like Afanasievo and Andronovo. Iran_Hotu, who is basically a Mesolithic or Neolithic Central Asian, doesn't share much of this drift.

There's simply no way to explain this, except to say that there was a big migration from the Eastern European steppe to Central and South Asia after Iran_Hotu was alive. That's because a big diffusion of Yamnaya-like ancestry into South Asia during the Mesolithic, from anywhere, could not have bypassed eastern Iran. This is easy to show, and I might do so this weekend.

Btw, Sein, Iran_Hotu doesn't have enough markers for robust qpAdm analyses with so many outgroups.

P Piranha said...

The application of scaling on nMonte dimensions produces extremely inconsistent results at the margin, e.g. no CHG in Estonians, only EHG, WHG and EEF, and there are so many issues with nMonte on PCA e.g. discovery of "African" and "Middle Eastern" contributions in Bell Beaker, that all results from it should be taken with a pinch of salt. Also, since this new qpAdm run it is simply no longer the case that the best models ("most parsimonious") we have for Indians remain as Yamnaya clones with Iran Neolithics and ASI. This shows that at higher levels of resolution, i.e. more outgroups to test the fit, Yamnaya + Iran Neolithic is conclusively rejected, while Iron Age Steppe and Iran Neolithic is accepted. Of course, better samples will improve the fit further but the old model is simply rejected. We should keep that in mind.

Honestly after the IBD segments with Srubnaya were found all indications were very strong that the question was already answered, because such strong signals in IBD, and from so recent a sample too--i.e. Indians have almost as prominent IBD with Srubnaya as Eastern Europeans--are just impossible to explain without recent gene flow from Steppe Late Bronze Age or Steppe Iron Age types with Indians. Of course the qpAdm indicated otherwise, but now this has changed so the picture is very coherent, and fits with the classic theory of Andronovo and Sintashta as Indo-Iranians and one of their descendant cultures as Aryan. We should always keep in mind that shared long chromosomal segments are always the strongest evidence of genetic connection, followed by qpAdm and D-stats and formal stats-reliant methods like treemix. nMonte and ADMIXTURE come in last. Chromosomal segments, given that your analysis is of sufficiently high quality, admit no alternative interpretations other than genealogical relationships; drift paths in formal stats admit several combinations as acceptable interpretations, and ADMIXTURE and nMonte just offer indications as to the affinities of our samples.

Likely the Samartian Pokrovka in this case, which are mostly Andronovo and Sintashta but contain minor additional inputs from EHG-ANE types in Siberia, some East Asian input, and contributions from the Caucasus and Iran, stand in for the admixed Andronovo descendants that entered India.

Davidski said...

I agree that this model with Sarmatians makes the most sense thus far, and IBD, or at least haplotype data, is the most conclusive way to demonstrate the close relationship between Indo-European South Asians and Bronze Age steppe groups.

But ADMIXTURE can be very useful in this context too, because it's able to pick up ethnic-specific drift. That's often one of its main weaknesses, because it skews ancestry proportions, but in this case it's a strength, because it's able to show a Bronze Age steppe-specific signal, and the fact that Indo-European South Asians carry this signal, while ancient Iranians mostly lack it, and the very little that they do show can be explained by CHG/Iran_N-like input in Bronze Age steppe groups.

I'll write something up about this, because any theories about South Asia have to take into account the reality that there was a population movement from the Bronze Age steppe to South Asia. This can't be explained away in any plausible way.

P Piranha said...

@ David

Minor quibble, but since I'm a sucker for accuracy: technically speaking there are many upper-caste and lower-caste populations from the same linguistic group who have the same amount of drift with Yamnaya. In fact IIRC UP Brahmin and the various low caste groups from UP are like that, in a double outgroup plot for them using D-stats Yamnaya sits squarely on the diagonal. The reason is probably because low caste groups have more ASI which is "Crown Eurasian" and increases their shared drift to Yamnaya itself. The populations that really differentiate upper caste groups from lower are modern Northern Europeans, Europe_LNBA, Steppe_LNBA and Steppe Iron Age populations.

P Piranha said...

About ADMIXTURE, no disagreement about this, just that we need to keep in mind the number of alternative interpretations that each method allows. Jaydeep and the whole "West Asian", "Teal" fiasco with Dienekes is such a situation.

An idea: You may want to focus on the occurrence of minor EEF in upper caste Indians in ADMIXTURE. This is a frequent and consistent pattern that appear in a fraction of analyses. Maybe the 16-population qpAdm for Indians can be redone with the "big six" as well, where the concern here is not the fit but whether or not EEF appears at different levels, if at all, among South Asians and thus whether or not independent confirmation of Samartian Pokrovka type material can be made.

Seinundzeit said...

P Piranha,

"The application of scaling on nMonte dimensions produces extremely inconsistent results at the margin, e.g. no CHG in Estonians... "

Not the kind of scaling that I have in mind.

In your case, I think you're referring to the clunky weighting of the PCs by their eigenvalues.

David,

"Btw, Sein, Iran_Hotu doesn't have enough markers for robust qpAdm analyses with so many outgroups."

Ah, I see.

Rather unfortunate, since as you noted, Iran_Hotu was basically a Mesolithic Central Asian.

Although, Iran_Hotu doesn't do much to the Steppe_EMBA/MLBA levels for South Central Asians; it mainly impacts the percentages seen in South Indians.

Seinundzeit said...

Also, just to clarify...

With regard to the Steppe_EMBA vs Steppe_MLBA difference, I've always kept myself in a state of epoché.

The evidence is just much too equivocal; so I'm waiting on those "Indo-Aryan" aDNA samples, before making the "Steppe_EMBA or Steppe_MLBA" determination.

Regardless, if genealogical linkages are the main clincher, we shouldn't forget this (please see below).

http://eurogenes.blogspot.com/2017/08/a-bronze-age-dominion-from-atlantic-to.html

As David noted in the aforementioned post:

"Interestingly, the graphs below, based on the cM values in my coancestry matrix, suggest that upper caste Indo-Aryan-speaking Brahmins from Northern India share relatively more ancestry with the Afanasievo genome than Iranic-speakers such as Pamir Tajiks, who generally share relatively more ancestry with the younger Andronovo and Sintashta samples."

Chad Rohlfsen said...

Here you go folks! David can continue with more models using these out-groups. Steppe_MLBA works well with extra ANE on top of the Iran_N ancestry.

left pops:
Pathan
Iran_N
AfontovaGora3
Steppe_MLBA
Onge

right pops:
Ust_Ishim
Mbuti
WHG
ElMiron
MA1
Levant_N
Karitiana
Kostenki14
Anatolia_N
CHG
Papuan

best coefficients: 0.412 0.140 0.307 0.140

std. errors: 0.085 0.039 0.074 0.047

fixed pat wt dof chisq tail prob
0000 0 7 6.670 0.464008 0.412 0.140 0.307 0.140

Chad Rohlfsen said...

Adding EHG and Europe_MNChL to the out groups really improves it.

left pops:
Pathan
Iran_N
AfontovaGora3
Steppe_MLBA
Onge

right pops:
Ust_Ishim
Mbuti
WHG
ElMiron
MA1
Levant_N
Karitiana
Kostenki14
Anatolia_N
CHG
Papuan
EHG
Europe_MNChL

best coefficients: 0.433 0.134 0.294 0.139

std. errors: 0.084 0.040 0.074 0.049

fixed pat wt dof chisq tail prob
0000 0 9 6.718 0.666437 0.433 0.134 0.294 0.139

Davidski said...

The Pokrovka Sarmatians are basically Steppe_MLBA, except with extra ANE, as well as some ENA and CHG/Iran_N admix, and less Europe_MNChL.

So I'd say it's basically the same thing.

Anthro Survey said...

@Seinundzeit

I realize CHG has a BIT of additional Villabruna-like ancestry, which is what confers it extra affinity to Anatolia_N and EHG(which is ANE+Villabruna_like) relative to Iran_N.

Nevertheless, the bulk of CHG is ultimately Iran_N+ANE, is it not?

Also, to tie in with what I said earlier in my comment regarding Kelteminar----do you believe there existed Iran_hotu-like populations with higher ANE content, sans Villabruna-like?

Anthro Survey said...

@Sein

This figure from a Laz paper illustrates my current paradigm when it comes to the relationship between CHG, Iran_N and Iran_hotu(note slight Villabruna shift in CHG):
http://oi65.tinypic.com/311pgrq.jpg

Seinundzeit said...

Anthro Survey,

"I realize CHG has a BIT of additional Villabruna-like ancestry, which is what confers it extra affinity to Anatolia_N and EHG(which is ANE+Villabruna_like) relative to Iran_N."

Honestly, there is some sort of looping dynamic going on between Anatolia_N and CHG (basically, it's a very complex relationship).

Anatolia_N often seems to be CHG-admixed, but in some methods one can get CHG to be around 5%-10% Anatolia_N, along with the extra Villabruna shift.

This is a case where we need further sampling of the Near East + the Caucasus, with a focus on acquiring samples dating to the Paleolithic and Mesolithic eras.

"Nevertheless, the bulk of CHG is ultimately Iran_N+ANE, is it not?"

Not literally. If my memory serves me right, the CHG samples predate the Iran_N samples.

But if in the sense of deep ancestral streams, then sure, CHG and Iran_N are quite closely related in terms of very deep genetic ancestry.

Although, ANE is a very complex construct (another case for which we need more samples, primarily from Paleolithic and Mesolithic Siberia + Central Asia).

"Also, to tie in with what I said earlier in my comment regarding Kelteminar----do you believe there existed Iran_hotu-like populations with higher ANE content, sans Villabruna-like?"

At the moment, I'm willing to bet that this is true.

That being said, I'm also willing to bet that the genetic legacy of such people is no longer to be found in southern Central Asia.

Rather, one sees the genetic signature of such people only in peninsular South Asia.

Of course, I could be wrong on both counts (perhaps such people never existed, and/or their genetic signature doesn't survive in peninsular India).

Without Kelteminar samples, one can only speculate.

Regardless, besides what I've seen in my own analyses, as supporting evidence I would note the presence of South Asian-specific U2 mtDNA lineages across the length and breadth of the Sub-Continent.

Anthro Survey said...

@Seinundzeit

"Not literally. If my memory serves me right, the CHG samples predate the Iran_N samples."

Right, I did not mean that Iran_N directly contributed to the ancestral composition of CHG. Essentially, I think of them sharing the same sort of deep relationship that Natufians, Levant_N and Anatolia_N do. Eastern analogs to that, essentially.

So, basically, they are quite interchangeable in nMonte, but less so when formal methods(concerned w/direct allele comparison) like qpAdm are used?

As for the dating:
Kotias predates them all. Hotu is the next oldest. Satsurblia is roughly contemporaneous with Abdol Hossein and Ganj Dareh. It is worth mentioning that Hotu and CHGs are the northernmost geographically and inhabited the same bio-climatic zone, which differs from the parched, arid environment of GD and AH. The same holds in the west, whereby the more basal-rich populations inhabited the more arid, UV-beaten environment.

"Rather, one sees the genetic signature of such people only in peninsular South Asia."

Peninsular---I take it you mean areas south of Vindhya river? Hmm, could it be that our imperfect population choices mask such an ancestral layer in northern areas like Punjab?

postneo said...

@jijnasu, @chetan

Chetans objections are thus. "How can you link with Hittite spoken thousands of miles away.?"

Actually the NIA word is much more becoming its asian and easterly confines. It is fact Sanskrit which is the geographic outlier. Since asthi is closer to greek and latin vs NIA. Latin was much farther away vs Kurdistan and Anatolia. Consider that Balochi is a western Iranian dialect spoken in Pakistan. similarly there are relic IA speakers alive today in Kurdistan. So such linkages are not that anomalous.

In fact the very goal of this blog is to prove an even more distant CW impact on South Asia. what about that?

apart from Kurdish, going further, even balto-slavic, Armenian and Albanian have an initial consonant like NIA and unlike Sanskrit and Avestan.

Other objection:
"all kinds of "mother tongues" brought all kinds of influences .to Sanskrit..its just a handful of words etc..

Its not a hand full, its a rather large set and a large subset are IE. Our goal should be to charactierize , identify patterns and sound shifts, not gloss over.

Seinundzeit said...

Anthro Survey,

"Essentially, I think of them sharing the same sort of deep relationship that Natufians, Levant_N and Anatolia_N do. Eastern analogs to that, essentially."

Exactly; no doubt about this.

"So, basically, they are quite interchangeable in nMonte, but less so when formal methods(concerned w/direct allele comparison) like qpAdm are used?"

I'd rather say that the opposite is the case; the higher-dimensional the data, the easier it is to distinguish between Iran_N and CHG.

"Hmm, could it be that our imperfect population choices mask such an ancestral layer in northern areas like Punjab?"

That's certainly a possibility.

Although, I think we'll be on much firmer ground, once we see the Central/South Asian aDNA paper(s).

Chetan_Vit said...

Thanks for the reply @post neo

I have a few questions. To start with, do we have enough reasons to be absolutely certain that the word in question haDDHi / haDDHa is a derivative of ásthi since the word shows a retroflex D that shouldn't be there by the rules. As someone already mentioned, the word seems to be pan India (does it mean that it occurs in other language families too?) Then we may have to think about a non IA origin for the word.

As shown in this map, the Afanasevo and Andronovo cultures had quite a degree of overlap in the late bronze age. In fact I believe that the Afanasievo culture's range extended far beyond the Tarim basin before the arrival of the second wave of Corded ware/ Sintashta migrants. The Andronovo then largely replaced the Afanasievo in Central Asia. In that case it is certainly not inconceivable that the earlier migration of centum languages could have left a substrate in relatively isolated pockets of speakers who shifed linguistic allegiance.

There is also the possibility that some early speakers of Greek who were present in the north west of the Indian subcontinent who changed their language to IA could have survived in the isolated Himalayan region (there is evidence of Greek influence in the India-Iran border going back even before Alexander's time) However I myself continue to believe that the Bangani substrate reports were spurious and there really is no centum substrate in the language

A third possibility that word initial "h" in certain words is not a relic of the PIE laryngeal at all rather unrelated purely coincidental innovations in each of the separate branches (IIR and Slavic).

The thing is I would consider any of these theories to be vastly more probable than the explanation that the ancestors of Hittite or Greek languages were present in India. That is really an unthinkably absurd proposition with zero evidence.

Anthro Survey said...

@Seinundzeit

"Exactly; no doubt about this."

Ok, great! This is my thinking, too. Just wanted to clear up and see whether we were on the same page. (thumbs up) It's just that, initially, I got the mistaken impression you overemphasizing their "Euro" affinity.

And, yes, I momentarily overlooked that nMonte utilizes the 10-dimensional Global10. Makes sense. Still, a bit surprised that 10 axes of variance give it an edge over qpAdm's sensitive, allele-by-allele nature unless I am missing something here.

Regarding Tajiks---is it really true that their non-steppe consistently comes out to being entirely Iran_Chl across various modeling schemes(or at least much higher Iran_Chl/Iran_N relative to Pashtuns)? A bit of a shocker if yes. If not, could be pseudo Chl affinity due to some ANF in late steppe groups.

Jijnasu said...

@postneo
haDDa seems to have a Pan-NIA distribution being attested from the west of the Indus to Bengal and marathi/sinhalese in the south. http://dsalsrv02.uchicago.edu/cgi-bin/philologic/getobject.pl?c.4:1:1158.soas
I don't think a connection to non-Indic IE languages is absurd purely based on the present location if a proper case is made out for it. The word looks suspicious due to the geminate -DD-. Middle Indic -DD- can either be traced either to Old Indic -rd- or -dr- or to OIA -ND- or DD-. An OIA form *harda or *hadra or *haNDa are clearly unconnected to any of the other IE forms for bone. Even *haDDa eould be hard to explain as a loan from any of the other Indo-European

I'm not convinced the CW has any direct connection with the IAs.
Who are the relic IA speakers in kurdistan??
Also Baloch is a recent migrant to the subcontinent, got there only around 1000AD

Oaie Porc said...

I am thinking of how all these data on ancient peoples relate to the ancient astronauts theories. I follow you with great interest. if more advanced civilizations intervened to our DNA at some point in the past would it be something today's genetic researchers would notice?

Davidski said...

I am thinking of how all these data on ancient peoples relate to the ancient astronauts theories. I follow you with great interest. if more advanced civilizations intervened to our DNA at some point in the past would it be something today's genetic researchers would notice?

Pretty sure they would at some point.

postneo said...

@chetan
"certain that the word in question haDDHi / haDDHa is a derivative of ásthi since the word shows a retroflex D that shouldn't be there by the rules."

Not Its not certain That would depend on further analysis. For that matter most of PIE is not cast in stone but a compendium of observed patterns.

It's haDDa / haDDI not haDDhI and no it's not pan Indian only north India ..even within that there is regional variation.

In Bengali it's hARa and I think marathi and Gujarati have similr forms where geminates are compensated by a long vowel. this shared isogloss between Marathi and bangla is evident in other words too maTTI vs mATI Etc . Today these languages are separated by atleast 300 million Hindi speakers who don't share this trait.

As for retroflexion it is common to see dental of Sanskrit as retroflexes in NIA

Also compounds replaced by geminates.

So this brings up another question was Marathi spoken in bengal or vice Versa. No ..Similarly noone is claiming greek or Hittite was spoken in India. But there are shared traits. Sometimes Sanskrit maybe more European shifted but other times NIA may have throwback forms not seen in Sanskrit.

Romulus said...

Pretty sure Davidski is going to have to fall back on ancient astronauts at sone point to keep steppe theory alive.

Acharya Agnimitra said...

@postneo

I think you are wasting your energy and time replying to those who heckle you to the minutest detail and to no end, no matter what you say. What you say really does not matter to them. When it comes to explaining themselves, they will either refuse to take a shred of the burden of proof or reply with utter crap whose refutations have been repeated ad nauseum over the years----" No horse bones in Indus valley"....."No chariot wheels"....."Steppe flora fauna in Vedas"......"dead bodies unearthed in Harappa".....

The moment you notice such brazen subscription to double standards, you should avoid unnecessary exertion.

Btw what do you think about Isidore Dyen's findings that suggest Proto-IndoEuropean and Proto-Austronesian during formative phases were close to each other as evidenced by sharing of basic words such as land and water, pronouns, first four numerals etc? He was by all means a Steppe proponent and found his own findings inexplicable. Steppes were far to the west and the Austronesian was in SE Asia. There was no way there could have been contact and he declared that the linguistic evidence was at odds with the Urheimat. Well, he may have overlooked an elephant in the room....
A Linguistic sum-up by Elst..
http://web.bookstruck.in/book/chapter?id=20205

postneo said...

@jijnasu
as you note haDDa is exclusively NIA which is significant. I have not checked Tibetan. I agree that the voicing in the medial position weakens its relationship with asthi. There are very few such transitions, I can think of one. But I know of no other neighboring languages that the word can be attributed to and certainly not Masica's small list. .. strapped for time.. will respond if I get a chance

Chetan_Vit said...

It's haDDa / haDDI not haDDhI and no it's not pan Indian only north India ..even within that there is regional variation.

Sorry my mistake. HaDDa it is. That makes things all the more clearer

The word HaDDa cannot be derived from the PIE **h₃ésth₁ by any rules of derivation in any known IE language family.

The word appears to be non IE

P Piranha said...

Attached below are some PCAs using David's fastIBD output, averaged by commenter Matt.

Here is a PCA of comparative IBD sharing between four Indo-Aryan populations, four West-Central Europeans, and alll the ancient Indo-European samples with higher fractions of Steppe ancestry. Yamnaya_Kalmykia sits dead center between Indians and Europeans. Andronovo, Afanasievo, Sintashta, Mezhovskaya, Karasuk, and the Hungarian Iron Age pre-Scythian are on the Indian side in descending order of sharing, while Ireland_BA, Unetice, Nordic_LN and IA, Roman Britain and Hungary_IA are on the European side, ordered as well.

Below are some PCAs comparing IBD between all Indo-Iranians and all modern Western Eurasians. Without labels : Dark blue is Uralic, Blue East is European, Light blue Central European, Cyan Mediterranean, Green is Anatolian and Balkans and South Italy, Purple Iranian plateau, Pink North Caucasians.

Here is the image without the biplot, which lets you see clearly that East European and Uralic populations are on the lower right, i.e. they dominate IBD sharing, followed by North Caucasians. Here is the image with population labels.

Indus Valley populations and Dravidians are distinguished by excess IBD with the Iranian plateau, Dravidians less so than Indus, which you can see from the shorter length of the biplot for Dravidian populations. The rest of the Indians and Indo-Iranians are distinguished by a very strong signal of IBD with East Europeans, both Slavs and Uralics. The populations with the strongest signals, that Kshatriya and Brahmin point straight towards, are Ukrainian Cossacks and Swedish. Sharing with North Caucasians is also significant.

I'm going to go out on a limb and propose that the steppic ancestry of Indo-Aryans derived from Sintastha and Andronovo clones, with EEF, that admixed into an Afanasievo and Karasuk substrate, without. Non-steppic contributions came along for the ride from the North Caucasus and Siberia. This stream is distinguishable by IBD, i.e. late, recent gene flow in Indians. Deep connections with Iran for North Indians would be less identifiable by shared segments or IBD, and will be as weak as the signal for Dravidians. In fact, IBD sharing with Indus Valley genomes will be quite low compared to the very long segments shared with the Aryan incomers, despite the Indus Valley genomes being more similar on the genome-wide level.

@ Sein

If we award the right level of evidential hardness to each of these methods, the situation would be clearer much longer ago. As I said, since now we have ancient samples of good quality to perform IBD, algorithms that detect long segments should be recognized to dominate even formal stats from now on.

@ Anthro Survey

nMonte is not better. The qpAdm model with CHG does not have a CHG-rich population other than Iran_N in the outgroups to judge the fit. In other words, either Iran_N or CHG are in the outgroups, never both. Since the rest of the outgroups are all very ancient with Palaeolithic divergences and do not have recent entanglements with CHG or Iran_N that will allow them to be differentiated, you get the same statistical outcome once you shift the proportions around. That's the only reason why that's happening. nMonte is not reliant on outgroups, but that does not make it better.

Chetan_Vit said...

@Jijnasu

Hindu support for an OIT isn't as much about fundamentalism as much as it is a reaction to the racism which colours many popular and until fairly recently even academic versions of the AIT.

Yes but facts don't care about anyone's hurt feelings nor can they be racist towards anyone. And I frankly do not understand what is so racist about the fact that there was a large scale migration of Indo Aryan speaking steppe populations which brought those languages to India. It is not like these Hindu nationalists who oppose this theory believe that all modern humans and languages originated in India (Sanskrit is the mother of all languages or some crap like that). Or perhaps they do?

You seem like a reasonable and well informed person but neither we nor the academics working on this should be obligated to suppress a true fact (which the Kurgan steppe theory is becoming more and more each day) just to spare some crybaby nationalists' hurt feelings. It is time these people learn to get over their "feelings" without standing in the way of academic research

Jijnasu said...

@post-neo
Apparently it is haDDa is attested as sanskrit word by certain lexicographers, not sure if it was ever used in any text though (it might be a MIA loan in sanskrit).

btw standard Hindi does simplify geminates with a compensatory lengthening of the previous vowel too, however it occurs only in monosyllabic words. (eg OIA. adya > MIA. ajja > H. Aj) One explanation for the bisyllabic nature of certain Hindi words is that they derive from an OIA. diminutive. eg. OIA.mRttikA > MIA.miTTiA > NIA. MiTTi.

Jijnasu said...

@chetanvit
I am not an OITist, I think that the steppe hypothesis is cnsiderably more probable than any of the other alternatives . However it has to be acknowledged that Indo-European studies had been closely associated with racism for much of the 19th and first half of the 20th century. Pop version of the kurgan hypothesis continues be linked to pseudo-scientific ideas of racial superiority. Read up on the ridiculous ideas being propagated at JNU culminating in the demonisation of the goddess durga and a creation of a neo-myth of mahishasura as an aborginal martyr (All this initiated by 'forward press' a magazine claiming to work for the upliftement of backward ethnicities - thinly disguising the fact that its editors are evangelical christian missionaries).

As for mainstream OIT, it doesn't suggest that sanskrit = PIE or anything of that sort). Its arguments are primarily based on archaeological and some of the early textual sources

Chetan_Vit said...

@Jijnasu I agree with you about the racist character of 19th and early 20th century indology and linguistic studies. But as we can see mainstream academic scholarship has moved on well past it. Now the idea that languages and cultures and languages are correlated with "race" has been discredited. The Indo-Europeans who entered India and Europe were genotypically a mixed bunch in all probability not a pure "Nordic Aryan race" or any such bullcrap

Only certain far right groups dare to bring in race to discussions on linguistics and genetics as it is generally accepted that all humans have had contributing ancestors from several ancestral populations. You may run into a few racists occasionally on the AIT side who use it to justify White supremacy or far right ideologies. But they are very few in number and frankly no better or worse than the Hindu supremacists in India

I guess there are crazies on both sides ;)

ak2014b said...

@Rob
Thanks for elaborating.

I'm trying to see if I can rightly follow your view of events based on your statements.

It seemed from a recent comment that you were contemplating a possible homeland for PIE "South of the Caucasus".

You mention BMAC and Tajiks are shifted more West Asian because of an additional layer of CHG/ANF like which the Indics don't have. Together with "Indians do not have an affinity to Barcin Neolithic, only Iran Neolithic", it seems you're thinking the EHG/Iran neolithic in IA is the migration signature in the Indics, while that the CHG/ANF West Asian came after with BMAC, and overlaid onto South Central Asia at a time forming the Iranics there.

However, from "it might be possible that there wasn't any (major) migration from the western steppe to India", "I guess the key is the EHG/Iran Neol mix for IA" and "This would create a sort of EHG as that seen further west, but a "hyper-EHG"" (for Gujarat and Punjabs), it seems like you're not placing an emphasis on EHG or the steppe as the route of entry into South Central and South Asia.

Further, it looks like you're thinking the hyper-EHG may either be old in South Asia or may have arrived earlier, possibly along with Iran Neolithic. And since you weren't particularly contemplating the steppe as the direct source of migration (bringing in the EHG), and possibly not from the Caucasus either, since you said something CHG/ANF like may only have been present from BMAC onwards and present in SC Asian Iranics, it looks like you're thinking the migration to South Asia came directly from Iran?

Do you think the source population of this migration carried both hyper-EHG and Iran Neolithic to South Asia, or just Iran Neolithic? From Lazaridis et al 2016, the Iran Chalcolithic samples had some EHG, as well as some more Levant (do South Asians have any of this?) and some WHG.

Are you considering that either in Chalcolithic or Bronze Age Iran, there may have been a distinct population with just Iran Neolithic and possibly also with hyper-EHG or some EHG, existing contemporaneously with Lazaridis' Iran Chalcolithic related population, but without mixing? Assuming both spoke some variants Indo-European, are you thinking the firstwould then have been part of the South Asia migration you seem to be describing, while the second could have been part of the later migration related to the BMAC, bringing in the "additional CHG/ANF like" in Tajiks and other South Central Asians?

Based on what I understand of your statements, the migration carrying Indo-European into South Asia would have been before BMAC, so before 2300 BC, but any idea how much before?

It's like a variant of one of the Near Eastern hypotheses, such as Sarianidi I think. I suppose further ancient sampling can tell us more, including the upcoming ones from South Central and South Asia.

However, did I actually understand you correctly at all, or have I totally misunderstood?

ak2014b said...

@Jijnasu
"I believe there is a group of archaeologists who trace beginnings horse riding to the botai culture"

Botai is supposed to attest to local horse domestication, though this view too is not uniformly supported, as it has been argued that it could merely point to early use of domesticated horses. Some remains indicated possible bridling. Assuming bridling, Outram thought this could therefore possibly suggest riding, but no certainty there either.

Unless there have been further archaeological developments, there appear too many uncertain factors in the chain going from horse milk to horse-riding in Botai.

Further, Anthony mentioned that archaeologists found in Chalcolithic Armenia a similar, but potentially earlier case of possible bridling, dated between 4000 and 3500 BC. Therefore, if the possible indications of bridling at 3700-3000 BC in Botai implies horse riding, Armenia more likely developed horse-riding earlier, based on archaeological evidence of the same nature.


@Chetan_Vit
Did you really just argue that anyone who concludes you're a religious fundy based on your flat earther scale statements must be an OIT? It's just that sort of unhinged reasoning which is also behind your write ups on evolution being a supposed "belief" that is "unparsimonious and incredibly unlikely". As well as your other attempt to deceive people that I pointed out, where you wrote that cosmology supposedly has proved your religious fables, despite cosmologists and theoretical physicists, like Krauss, Hawking and others, having frequently explained how cosmology has revealed there's absolutely no need for god. And physicists and biologists have done so specifically to prevent people like you from ongoing attempts to teach religious nonsense as "science" in schools.

Pointing out the religious fundies here, and what they're really here for, does not make me or Salden or anyone else OIT. OIT is something that's moreover restricted to South Asia.

Genetics is based on evolution and the processes driving it, like mutations and natural selection, which is exactly what your write-ups (that I already linked to further above) have been in religious denial about. So that means the only reason why religious fundies like yourself would choose to wade into genetics discussions now is for some ulterior purpose. And it will be the same crazy religious purpose for which you've been wading into discussions about cosmology and evolutionary biology.

Chetan_Vit said...

@ak2014 There is no need to give a response to your totally out of topic stalkerish ad hominem attacks nevertheless misinformation must be corrected for the readers' sake at least

"Genetics is based on evolution and the processes driving it, like mutations and natural selection, which is exactly what your write-ups"

I don't deny evolution and I believe that anyone who questions if evolution has occurred and is occurring in the face of all the evidence we have today is totally unhinged. What I would differ on is the explanation of random natural selection alone as the producer of complex biology from simpler systems. Anyway this is totally irrelevant to the archaeogenetic studies that are central to this blog since that requires an analysis of the changes that have taken place in the human genome and their chronology only.

My religious or philosophical position is completely irrelevant to my views on this matter and the arguments I give. Do you know that trying to discredit someone's arguments based on their religious affiliation alone would be a case of a logical fallacy.If you must, engage my arguments next time instead of launching silly ad hominem attacks. Good day to you

ak2014b said...

@Chetan Vit

"I don't deny evolution"

Yet, then:

"What I would differ on is the explanation of random natural selection alone as the producer of complex biology from simpler systems."

People like you are infuriating. You're absolutely denying evolution as it factually happens. So you're arguing that there was some non-random, hence deliberate "design" (and next, a "designer") to guide evolution along to some purposeful state.

Here's what you said, attempting to sound more reasonable than your more low-brow Intelligent Design (ID) fellows:

"Unlike what some ID- proponents say, evolution doesn’t violate any laws of thermodynamics. But it is just incomprehensibly improbable even over billions of years

Secondly, the production of genetic mutations is supposed to be totally random and blind to the consequences. In that case, it is difficult to imagine how a complex organ, say for example the eye, could evolve from scratch. Just imagine, the evolution of the eye or the heart would have required thousands, probably millions of changes in the preexisting genetic code. What is the probability that each step along the way was somehow beneficial to the organism and got selected? Compare this to what we have observed - the vast majority of mutations that occur naturally are inconsequential if not harmful for the organism!

So it is not irrational to believe in the theory of evolution by natural selection. I would just say it is an incredible unlikely and unparsimonious theory."


You should not be allowed to make statements on evolution (and therefore also genetics). Instead, you should force yourself to read Dawkins' "The Blind Watchmaker: Why the Evidence of Evolution Reveals a Universe without Design". It explains why the eye, heart and the rest, is made by the very evolutionary processes you dismiss as "unparsimonious" explanations for their development.

Wikipedia provides a description of the content where it addresses the very eye itself:

"To dispel the idea that complexity cannot arise without the intervention of a "creator", Dawkins uses the example of the eye. Beginning with a simple organism, capable only of distinguishing between light and dark, in only the crudest fashion, he takes the reader through a series of minor modifications, which build in sophistication until we arrive at the elegant and complex mammalian eye. In making this journey, he points to several creatures whose various seeing apparatus are, whilst still useful, living examples of intermediate levels of complexity."

Why don't you let biologists teach biology while you content yourself with Bible Studies, instead of tainting biology and physics with your gradual insinuations of ID and other religious fables.


"Anyway this is totally irrelevant to the archaeogenetic studies that are central to this blog since that requires an analysis of the changes that have taken place in the human genome and their chronology only."

It's absolutely not irrelevant. You can't pick and choose. This is merely a magnified view of a short-term window of evolution. It's part of what demonstrates how natural selection works. Like the drop in Neanderthal and everything.


"Do you know that trying to discredit someone's arguments based on their religious affiliation alone would be a case of a logical fallacy."

You discredited yourself with your dangerously loopy views. They're not so much a problem if you at least kept them to yourself, but you fundies choose to propagate them in order to force others into your delusions.

Seinundzeit said...

@ Anthro Survey

(I'm very sorry for taking so long; finally have some free time)

"Still, a bit surprised that 10 axes of variance give it an edge over qpAdm's sensitive, allele-by-allele nature unless I am missing something here."

Truth be told, I'm of the opinion that these methods can't be construed as having true epistemic edge in relation to each other.

Also, I was rather vague with my usage of "high-dimensional".

Honestly, I didn't have in mind the 10 PC dimensions; rather, what I meant by "high-dimensional" was the ability to examine things on the plane of (relatively) recent genetic differentiation (and with specificity), versus the ploughing of deeper phylogenetic ground with f-statistics, not to mention the reliance on outgroups with qpAdm.

"Regarding Tajiks---is it really true that their non-steppe consistently comes out to being entirely Iran_Chl across various modeling schemes(or at least much higher Iran_Chl/Iran_N relative to Pashtuns)? A bit of a shocker if yes. If not, could be pseudo Chl affinity due to some ANF in late steppe groups."

As is often the case, I can get loose with my ethnic labeling, for the purposes of easing/simplifying conversation.

To be more precise, Yaghnobi and Pamiri peoples (with the possible exception of the Ishkashimi speakers) show only Iran_Chl.

Although, I'm sure ethnic Tajiks will partake of the same pattern.

Among Pashtuns, the Ghilzai and Karlani are also strongly skewed towards Iran_Chl vs Iran_N, although the Iran_N skew is greater than what we see in Tajikistan.

Chetan_Vit said...

@ak2014b

This is a simple case of difference in philosophical worldview between you and me. You like Dawkins believe that "a simple organism, capable only of distinguishing between light and dark in the crudest fashion" can go through "a series of minor modifications which build in sophistication until we arrive at the elegant and complex mammalian eye". All this through the random work of the natural selection process acting on random changes in the genetic code produced during reproductive cell division.

I for one refuse to believe that such a process could ever result in the evolution of a complex animal from single celled organisms even over billions of years. It is just enormously unlikely and the onus is on the proponent of this theory to offer a mechanism by which this could happen. Does Dawkins realize how complex the minutest part of the eye is? (the eye lens designed for precise visual calibration, the light sensitive retina cells, the optic nerve and the visual cortex of the brain which somehow "evolved together" with the eye during this whole process)

This is a case of failure to appreciate the sheer magnitude of complexity found in living systems, down to the smallest components of the cells. Even comparing this sort of complexity to the most sophisticated inventions made by the hand of man would be a futile exercise. If you must accept that a spacecraft or a computer or a vehicle had to be intelligently designed, then I think it is at least rational to think the same about a living body which is many times more complex.

This is digressing too far from the subject matter of this blog. As I said before, archaeogenentic studies (as it is relevant to human prehistory) don't require a knowledge of the evolutionary mechanism. Only a chronology of the genetic changes is required, so there is no need to go into a debate on the mechanism of evolutionary processes here

Acharya Agnimitra said...

"the eye lens designed for precise visual calibration, the light sensitive retina cells, the optic nerve and the visual cortex of the brain which somehow "evolved together" with the eye during this whole process"

Dazzle them with your brilliance or baffle them with your bullshit. There is no 'precise calibration' at the lens. Of the entire refractive power of the eyes, the mostly acellular lens contributes only a third of the total refractive power, the multi layered cornea in front of it does the heavy lifting. The only thing this entire complex ultimately does is ensure light falls in the vicinity of the fovea depending on acuity needs. The real physiology of vision begins at the receptors. The occipital lobe turns the upside down image erect, meaningfully interprets it, and that is how you SEE. There is NOTHING unusual about the optic nerve, or the optic tract or the cortex evolving together. EVERY STEP of the visual pathway has been explained by science with NO gaps involved, so also the other cranial nerves or any part of the body.

The fact that you cannot understand it does not mean a magician created it. And trust me, no other organ in the body suffers as many infectons, inflammiatons and degenerations as the human eye. Nothing magical there.

Want more proof? HERE! There are fish species which evolved to loose their vision. HOW? Natural selection. The reverse of the process which confuses you...
https://en.wikipedia.org/wiki/Mexican_tetra


"archaeogenentic studies (as it is relevant to human prehistory) don't require a knowledge of the evolutionary mechanism."

You are a cherry picking creationist. You reject the entire book of human evolution of which population genetics is JUST a single page. Yet you dabble in archaeo-genetics for your pseudo-scientific/religious propaganda. Your ideas are very similar to some of the missionary pastors I have seen. I think you are one of them...

Chetan_Vit said...

@Acharya Agnimitra

And I am almost certain that you are one of those yoga fanatic Hindu nationalists who believe that ancient Indians flew airplanes in the stone age. Actually your name leaves little doubt about it!

Chetan_Vit said...

@Acharya Agnimitra

Here's a bit of data that will help you to put things in perspective and refrain from making supremely arrogant statements such as "EVERY STEP of the visual pathway has been explained by science with NO gaps involved". By the way, that statement of yours proves that you are not at all well acquainted with the developments in that field since even the best biologists working on this will not dare to make a claim such as this. It's funny when die hard Hindu fanatics who believe in stuff like stone age airplanes and test tube births in ancient India have no problem misusing and misappropriating scientific theories when it fits their agenda.

Read this:

"While today's digital hardware is extremely impressive, it is clear that the human retina's real-time performance goes unchallenged. Actually, to simulate 10 milliseconds (one hundredth of a second) of the complete processing of even a single nerve cell from the retina would require the solution of about 500 simultaneous nonlinear differential equations 100 times and would take at least several minutes of processing time on a Cray supercomputer. Keeping in mind that there are 10 million or more such cells interacting with each other in complex ways, it would take a minimum of 100 years of Cray time to simulate what takes place in your eye many times every second."

Evolutionists rarely attempt to calculate the probability of chance occurrence in their imagined evolutionary scenarios.

Souce : https://answersingenesis.org/human-body/eyes/can-evolution-produce-an-eye-not-a-chance/

(By the way don't be too quick to judge the content after reading the title of the website, since that is again a common mishap of people who are unable to evaluate arguments critically without resorting to ad hominem attacks)

Acharya Agnimitra said...

" you are not at all well acquainted with the developments in that field since even the best biologists working on this will not dare to make a claim such as this."

And yet here you are desperately looking for knowledge gaps in the same field to plant your missionary seeds of magic- because your nonsense can only hide and thrive in such darkness.
Your logic--If something's yet unexplained(which it's not) it's intelligently designed!

"Keeping in mind that there are 10 million or more such cells interacting with each other in complex ways, it would take a minimum of 100 years of Cray time"

You and the charlatan who authored this might have fooled a lot of gullible people with this sort of numerology crap. But this has no bearing on what you were originally arguing.
You said the eye could not have evolved and you are yet to show what part of the eye or what physiological process or what biochemical reaction is yet to to explained by evolution. Instead you praise the retina's efficiency (with an unverified quote) and somehow think that implies whatever you are arguing. Unless you are claiming that an extra-cranial intelligence helps the retina function, your quote implies nothing.

If you knew the ten layers of the retina and what each of them does , if you knew the basic function of the retina- phototransduction( a complex reaction to turn light into electrical impulses and send it to the brain), you wouldn't be claiming divine aid in its function. Go and read about it. With your mentality, you might not understand it. You dont EVEN understand the BASIC fact that image simulation, visualization and interpretation actually happen INSIDE THE BRAIN, not in the eye. You are SEEING with the back of your head, the eye is only a receiver and transponder. Science has broken down vision down to even the basic chemicals.

"it is clear that the human retina's real-time performance goes unchallenged."

Oh really? Many animals beat humans hands down based on the criterion considered(field, acuity, night vision etc) You were saying?

The only thing unexplained, and what the cunning author of your asinine creationist article tried to pass off as evidence of ID in the guise of numerology, is what happens inside the cortex. That is the mind-brain enigma-- It is part of cognitive neuroscience. It is a FRONTIER OF KNOWLEDGE and asks one of the existential questions of life-how a bunch of sodium and potassium in billions of neurons together create complex mental phenomenon- thoughts, memories etc.

Naturally, such intense frontiers are better left to scientists. It is a dangerous place to allow pastors and creationists to hang out and pollute.

" It's funny when die hard Hindu fanatics who believe in stuff like stone age airplane"

There are retards in every corner of the globe. Guaranteed. But stupidity is not a crime as long they keep it to themselves. Most of em are harmless. They don't actively try to push their nonsense on the gullible masses. They dont try to manipulate the education systems of entire countries. They don't tamper with history to suit their missionary, creationist agenda( as your sort does in the West and in India)

But you, pastor hiding under a Hindu name, you are quite the trouble... and I wont be replying to your abysmal nonsense anymore. You'll be talking to yourself.

Chetan_Vit said...

"But you, pastor hiding under a Hindu name, you are quite the trouble... and I wont be replying to your abysmal nonsense anymore. You'll be talking to yourself"

LOL so much for the much touted tolerance and diversity of Hinduism. This Agnimitra guy is a fanatic

Unknown said...



P Piranha what software do you use for PCA?

P Piranha said...

@ Unknown

I use Past3, same software that Matt uses. Its very convenient, you can find it here: https://folk.uio.no/ohammer/past/