search this blog

Friday, May 4, 2018

The protohistoric Swat Valley "Indo-Aryans" might not be exactly what we think they are


I need some help interpreting these linear models of ancient and present-day South Asian populations. Overall, the Iron Age groups from the Swat Valley, or SPGT, look like rather obvious outliers. The relevant datasheet is available here.


This might be because of significant bidirectional gene flow and/or continuity between Central Asia and the northern parts of South Asia before Sintashta-related steppe herders showed up in the region, and even before the Bactria Margiana Archaeological Complex (BMAC) got going. Note that Dzharkutan1_BA is an BMAC sub-population from near South Asia, but it doesn't quite have the same effect on those Swat Valley samples as the pre-BMAC Shahr_I_Sokhta_BA1 from the present-day Iranian/Afghan border.

If true, it probably means that most of the Iron Age peoples of the Swat Valley shouldn't be modeled as simply a mixture of Indus_Periphery and Steppe_MLBA. That's because they appear to be in part of the same or similar type of ancestry as Shahr_I_Sokhta_BA1. And indeed, qpAdm also suggests that they are.

SPGT
Indus_Periphery 0.692±0.042
Shahr_I_Sokhta_BA1 0.104±0.045
Sintashta_MLBA 0.204±0.015
taildiff: 0.659609
Full output

I'm trying to incorporate this new information into my Admixture graph models of the SPGT groups (see here). If I manage to come up with something useful I'll update this post with the results.

Update 08/05/2018: see here.

332 comments:

«Oldest   ‹Older   201 – 332 of 332
ryukendo kendow said...


Just face it Chetan. You South Indians never built the IVC and never were the indigenous group in Northern South Asia. You got loads of things in South India to feel proud of. Though I've been reading your comments above and they make little sense and to be honest you just come across as insecure and with some kind of South Indian inferiority complex. Trying to so hard to push theories that even a child could figure make no sense...

ryukendo kendow said...

@ Vara

The two are very difficult to separate in West Asia because EEF+Yamnaya combinations exist in local substrates by the time of Kassites even if we assume some Steppic input, which will confuse our attempts to tease out the steppic contribution. But Armenia_MLBA and Tepe Hissar BA tgt with Armenia_Chl/EBA and Tepe_Hissar_Chl are two pairs of populations that bracket the mid second mil date for arrival of Kassites, and I think the indications lean towards them having Sintashta-Andronovo ancestry when coanalysed with the preceding populations as the source.

Ultimately we need very high res data to perform haplotypes on to trace the source unambiguously; however another thing to bear in mind is the minority status of the putatively IAr elites. The Hajji_Firuz_BA is an extremely, extremely steppic individual and cannot be representative of the locals, and it is purely due to chance that we found her and dated her in an intrusive burial when the scientists were in fact focusing on populations from another era.

ryukendo kendow said...

^^ I mean Hajji_Firuz_BA and Hajji_Firuz_Chl, not Tepe Hissar! Apologies.

Chetan said...

"Just face it Chetan. You South Indians never built the IVC and never were the indigenous group in Northern South Asia. You got loads of things in South India to feel proud of. Though I've been reading your comments above and they make little sense and to be honest you just come across as insecure and with some kind of South Indian inferiority complex. Trying to so hard to push theories that even a child could figure make no sense..."

Whoa that went inappropriate pretty fast. Anyway, it's not as if I need to "push the steppe theory" on anyone. It's already mainstream and accepted by scholars, both Indian and non-Indian.

Chetan said...

Actually, I don't even know what this guy (ryukendow/mzp) is talking about.

ryukendo kendow said...

Lol I'm not the same person as mzp. I Just think what he says is super funny.

Grey said...

ryukendo

"the BB migration to Britain replacing 90% of the autosomes across wide areas is very, very difficult to reconcile with the archaeological signal, which shows widespread continuity"

patchwork settlement might explain it

say in the far NW of Europe only a limited number of sites were viable for early farming leaving the rest of the land still containing HGs (or the farmers cleared the HGs but then fell back to a more restricted range cos they exhausted the soil) but if most of that land was viable for herders then when the herders arrived maybe they simply settled all that empty land becoming 90% of the new combined population as a result.

Chetan said...

@ryukendow Oh, Apologies. So the original comment was deleted then?

mzp1 said...

Chetan you're crazy and have some serious psychological issues that are apparent with the way you are pushing the steppe theory.

Lets take a look..

You're talking nonesense about some kind of 'fire cult' that was supposedly introduced to South Asia via BMAC or something. TBH I dont know what your on about but you seem to be taking it very seriously.

The problem you have though, is that IndoIranian burial customs clearly reflect a South Central Asian origin. Because,

1.Zoroastrian Exposure practices can only be derived from Tibet. And that area is where Zoroastrians themselves place their origin.

2. Ashvamedha is only found in later Vedic literature but parallel practices are found in Celtic and are proposed for Andronovo.

But yet you choose to ignore all this clear evidence and instead talk about some seriously sketchy ideas about cremation.

You are so myopic in your fervour to push the migration theory that you are missing major parts of the puzzle, or for some psychological reason you are blocking it all out except for what you want to hear.

ryukendo kendow said...

@ Chetan

LMAO yep.

old europe said...



Grey

but if most of that land was viable for herders then when the herders arrived maybe they simply settled all that empty land becoming 90% of the new combined population as a result.

Yes it is a scenario pretty much likely but I think what you say must be combined with the factor of higher fertility rates too .

Chetan said...

@mzp I never made any definitive statements about the origin of cremation. Just proposed an explanation based on its existence in Andronovo in Xinjiang.

Cremation could have been practiced in a whole bunch of places, for all we know. As such, it's not the best signature to trace the movement of Indo-Aryans.

But anyway, this whole discussion is looking ridiculous now, since you have accused me of saying a lot of things I have never said here (like the Harappa = Dravidian claim)

You are free to not accept the steppe origin theory if you wish so. But to me and a lot of other people, it's already looking like the best explanation. But please do not accuse me of pushing a steppe agenda.

mzp1 said...

@Chetan

"But to me and a lot of other people, it's already looking like the best explanation"

Ofcourse, because you keep ignoring the vast amount of contradictory evidence.

And you just proved that right now, because I just gave you two pieces of data for you to look at, and you just ignored it, right now! It's obvious nothing's going to change your mind now, you are too invested in the migration theory.

Ric Hern said...

@ old europe

If you buy an imported German Vehicle and several other equipment for your home, do you suddenly start to speak German ? Or you decide to do yoga, meditation and like Indian food.... I guess no. So language can not always be connected to cultural goods or practices.

old europe said...

ric

Pardon, what are you referring to?

old europe said...

"I also disagree with his conclusion that an absence of steppe ancestry would imply a Near Eastern PIE homeland. Instead, it would imply that the Anatolian languages arrived in Anatolia largely via elite transmission which could also help explain why it is more divergent than Indo-European languages that arrived with large communities of native speakers via mass migration."

This is a quote from an article in Dispatches from turtle island.

The logic is this for people that push the steppe theory
1) If there's steppe ancestry......there's language change
2) if there's not steppe ancestry language change came via elite dominance and cultural influence

Maybe he's right maybe he's wrong but this is not the issue
Just tell these guys that IE on the steppe could have spread via elite dominance from caucasus or farmer europe or iran and they start immediately to pontificate ( sometimes to insult) and to bully.

Obviously not all the steppist are like that

Vara said...

@RK

That confused me but no worries lol.

Chl Iran and Armenia are simply too old to have proper Yamnaya ancestry. The context of how the Steppe ancestry made it is the real important part, which is why archaeology is still the best method for the PIE question. Mycenaean with high steppe ancestry is a female as is the Hajji Firuz individual. Do you see a trend here? Doesn't that have parallel in Macedonia where they were bringing in Scythian women?

Ric Hern said...

I begin to think that Indo-Europeans had the right products to trade in Crisis situations and the keen intuition to be at the right place at the right time and look like the saviours of those in Crisis...maybe this could at least be one of the major methods of spreading Language ?

ryukendo kendow said...

@ Vara

But the Steppe ancestry I'm talking about here is in the Armenia EBA-Armenia MLBA transition (which seems to come from Sintashta) and in the Hajji_Firuz_Chl-Hajji_Firuz_BA transition (which seems to come from Yamnaya), i.e. I'm not referring to Armenia Chl or Iran Chl, and both of these transitions are earlier than the Kassites, Guti or Mitanni but later than Yamnaya, showing that IE autosomal influences are in the region at about the right time.

We only have one high steppe sample from Greece and one high steppe sample from Hajji_Firuz, I think its very difficult to claim a pattern of "they were bringing in Steppe women" from sample sizes of one. In fact, I don't even think the archaeology supports this in any consistent way, I mean how would you even tell the "bringing in Steppe women" part, how would this manifest itself in the archaeology in a very visible manner esp if the women were low status foreigners?

For the genetics I think the only thing that can either prove or disprove sex bias in admixture in an unambiguous way is not even the Y or mtDNA haplogroups (especially when the sample size is small) but differences in ancestry when the autosome is compared to the X chromosomes.

For that, all the biases are in the correct direction in aDNA from Europe and modern DNA in South Asia at least. For CW and BB, and even the Balkans Chl which had very sparse representation of R1a or R1b Y haplogroups, the autosome is still heavily steppic compared to EEF-rich X, and for South Asian moderns the autosome is heavily steppic compared to AASI-rich X. It will be interesting to look at the patterns for the SPGT and West Asian aDNA samples, though for the SPGT there may be quality issues.

ryukendo kendow said...

^^ Balkans BA I mean.

Davidski said...

@Vara

The only high status Mycenaean sampled to date is that female. So why should the Mycenaean male show higher steppe input? Duh.

And the only C14-dated Bronze Age Hajji_Firuz sample is also a female.

Do you see how stupid and desperate you make yourself look with these sorts of idiotic arguments?

ryukendo kendow said...

Another mistake:

"But the Steppe ancestry I'm talking about here is in the Armenia EBA-Armenia MLBA transition (which seems to come from Sintashta) and in the Hajji_Firuz_Chl-Hajji_Firuz_BA transition (which seems to come from Yamnaya)

The two sources should be switched, Hajji_Firuz_BA with Sintashta and Armenia_MLBA with more Yamnaya, but really a small slice of Sintashta appears consistently for that as well.

Vara said...

@RK

Sorry I misunderstood your post.

I wasn't entirely serious about the steppe women I was comparing it to the simplistic scenario EHG boys kidnapping CHG girls. However, we have more than one male sample from Early Mycenaeans so there is a pattern here it as it should be the warriors who have high steppe. Archaeology does support trade between greater Mesopotamia and the steppes, this has been known for a while. What archaeology does not support is a movement from Sintashta to Iran. While new weaponry and tools show up in the Near East they are not of steppe origin. Yes it is possible that the steppe traders decided to settle down in the south but the Mitanni were not traders they were intrusive warrior elites.

Vara said...

@Davidski

I knew I would trigger someone. That Mycenaean is not labeled as elite. Keep repeating that maybe it would become a reality.

ryukendo kendow said...

I wanted to make sure the Sintashta signal was not some kind of fluke, so I threw all Steppe pops and West Asian pops from the Chalcolithic and EBA before at the Hajji_Firuz_BA individual to see what sticks.

Note that to resolve the temporal discrepancy the pre-Sintashta Poltavka may be interesting, its included below.

[1] "distance%=1.831 / distance=0.01831"


Hajji_Firuz_BA
"Armenia_EBA" 21.8
"Yamnaya_Samara" 21.25
"Armenia_ChL" 20.25
"Yamnaya_Bulgaria" 12.95
"Sintashta_MLBA" 10.75
"Poltavka" 10.4
"Ganj_Dareh_N" 2.2
"Seh_Gabi_LN" 0.4
"Anatolia_BA" 0
"Anatolia_ChL" 0
"AfontovaGora3" 0
"Sarazm_Eneolithic" 0
"West_Siberia_N" 0
"Onge" 0
"Shahr_I_Sokhta_BA3" 0
"Mala" 0
"Shahr_I_Sokhta_BA1" 0
"Yamnaya_Kalmykia" 0
"Yamnaya_Ukraine" 0
"Levant_BA" 0
"Tepe_Hissar_ChL" 0
"Geoksiur_Eneolithic" 0
"Seh_Gabi_ChL" 0

The signal remains for 10+10 = 20% Sintashta-Poltavka type Steppe MLBA.

ryukendo kendow said...

@ Vara

"Archaeology does support trade between greater Mesopotamia and the steppes, this has been known for a while. What archaeology does not support is a movement from Sintashta to Iran. "

What archaeology has been interpreted to support or not support should be malleable in light of new data, like for the Swat situation. There was supposed to be no evidence for Central Asian movements as well, Steppic or not, excepting very tentative connections drawn by AMT theorists.

Vara said...

@RK

"What archaeology has been interpreted to support or not support should be malleable in light of new data, like for the Swat situation. There was supposed to be no evidence for Central Asian movements as well, Steppic or not, excepting very tentative connections drawn by AMT theorists."

I'm not sure what Swat situation? Swat was always pushed due to the evidence of contacts with Andronovo and archaeology was right again. It wasn't a big movement nor did it have a very big impact.

ryukendo kendow said...

But even the AMT theorists themselves did not believe the evidence for migration was even necessarily present, they picked Swat as the least worst option. People like Anthony made the weaknesses of their claim amply clear from the start.

The migration wasn't necessarily small either, Steppe and BMAC each give ~10-20% but together this implies close to a third of the ancestors of the Swat population coming from Central Asia.

My point is that the interpretation of sites and material culture leads to a very wide confidence interval as to the amount of migration, or even the presence of migration. To add to this the Globular Amphora were identified as Indo-Europeanized from archaeology but in fact they were separated from EHG and steppe populations by 1000 years of genetic isolation. So I'm not sure our priors should be biased so strongly based on the conventional interpretations we have now.

Vara said...

@ryu

"The migration wasn't necessarily small either, Steppe and BMAC each give ~10-20% but together this implies close to a third of the ancestors of the Swat population coming from Central Asia."

Why are you lumping BMAC and Steppe together? They are two different populations and they both have a different relationship with South Asia. 2300-1700BCE BMAC had strong contacts with the Indus while the steppe folk that migrated to SA bypassed BMAC so BMAC ancestry most likely reached SA separately, and if that is true then the impact wasn't that significant. I hope I'm making sense here.

"So I'm not sure our priors should be biased so strongly based on the conventional interpretations we have now."

Sure. Wanna move on with PIE having a southern origin? Which is old news btw.

Sanuj said...

"Why are you lumping BMAC and Steppe together? They are two different populations and they both have a different relationship with South Asia. 2300-1700BCE BMAC had strong contacts with the Indus while the steppe folk that migrated to SA bypassed BMAC so BMAC ancestry most likely reached SA separately, and if that is true then the impact wasn't that significant."

Absolutely.

The argument seems to be:
1. Archaeology is mostly irrelevant.
2. SWAT is a dead end, died out along with it's language X.
3. It was never expected here, let's look for invasion elsewhere.

Funny.

ryukendo kendow said...

@ Sanuj

You definitely have problems with reading comprehension.

@ Vara

"2300-1700BCE BMAC had strong contacts with the Indus while the steppe folk that migrated to SA bypassed BMAC"

But given David's most recent qpAdm we have strong indications that this isn't true. There is probably some variation in the amounts of BMAC ancestry in different migrating groups, which is plausible because extremely large variations in Steppe MLBA groups w.r.t representation of local substrates is already present in the current data from the MLBA (compare Dali_EBA with Molaly or Dzhaylau).

Sanuj said...

"About archaeology, I really do not know how much certitude we actually get by trying to infer migration from interpretation of material artifacts."
"further sampling is needed from contemporaneous cultures close by to look for something closer to present-day South Asians (which by necessity will be closer to Sintashta)."
"Are there any "weird" linguistic substrates or historically extinct languages in the Swat area? The Butkara_IA samples are very dissimilar to present-day South Asians, resembling prior SPGT, despite being from ~0AD. These people seem to have went extinct in the historical period."

You definitely have problems with reading comprehension.

No, I don't. Just simplified things into easy English for everyone's benefit.

ryukendo kendow said...



Lol you definitely have comprehension issues. How does that represent an "argument"? What even am I arguing for?

You seem to have hyperactive facility for stringing together non-sequiturs.

Sanuj said...

You are arguing that archaeological continuity does not count for much, that we will get a heavily Sintashta shifted population lying buried in some hidden corner of South Asia, that Swat was not even meant to be the IAs as Anthony had already hinted.

You know why you are wrong? Because you know that evidence is hinting towards a late influx of Steppe ancestry, possibly via Indo-Scythians or Kushans or others, but you are not ready to delink IE in India with Steppe, you are running out of explanations so you're obfuscating the arguments with some great mental gymnastics.

Rob said...

@ RK

"To add to this the Globular Amphora were identified as Indo-Europeanized from archaeology but in fact they were separated from EHG "

Yes that was suggested by Gimbutas and other "Indo-Europeanists" who don't grasp the origins of pastoralism, the actual specialists rightly identified it as essentially a Megalithic group and the impact of agro-pastoralist revolution from the Carpathian basin



"The migration wasn't necessarily small either, Steppe and BMAC each give ~10-20% but together this implies close to a third of the ancestors of the Swat population coming from Central Asia"

Perhaps even more. But if BMAC is 30% and arriving hundreds of years earlier, whilst MBA steppe 10% in the Iron Age, then they probably represent different events.

ryukendo kendow said...

@ Sanuj

Lol as far as I am concerned, both Sintashta and BMAC ancestry are present in Swat for a total contribution of approximately 30%, and this is more or less completely in line with Parpola, Anthony, and co., so there is no need for me to argue against you. Not everything is some kind of plot that revolves around the OITers, especially when their scenarios are so weak.

I am interested in the far more interesting question of how the Butkara_IA type people came to be displaced by Kashmiri and Pashtun like people in their current range, and whether or not the immigrating groups contemporaneous with Swat had some mosaic of BMAC-Sintashta-Siberian ratios that impacted further migrations later, especially when we already know that almost all Indian caste groups were founded by small numbers of ancestors around 500BC-0AD or so, but of course you people are quick to treat any nuanced approach to the subject with screeches of "the Steppe theory is on its last legs!" and show no evidence of being able to hold more than two thoughts in your head at once.

Vara on the other hand raises very valid objections, pls try to reach at least his level.

ryukendo kendow said...

@ Rob

"Perhaps even more. But if BMAC is 30% and arriving hundreds of years earlier, whilst MBA steppe 10% in the Iron Age, then they probably represent different events."

This sounds quite possible to me, if that is the case maybe some of the BMAC features of the IAr culture in S Asia actually come from local post-Harappan groups that already had some BMAC cultural contributions, especially since, as Vara has quite rightly pointed out, that there is evidence for extensive material contacts and Narasimhan finds IVC contribution to late BMAC.

ryukendo kendow said...

@ Davidski
Hmm, if you're interested one way to check for BMAC-Sintashta single wave or two wave arrival is to check for ancestry covariance. If the qpAdm gives BMAC-Sintashta varying positively with each other it was probably a single wave, and if its negative then two waves (or, I'm guessing less parsimoniously, variation in BMAC-Sintashta ratios in a single wave).

^^@ Rob
This is of course just a poor-man's method given the little data we have, and the usual caveats for coverage & etc. apply.

ryukendo kendow said...

^^ In SPGT I mean.

Davidski said...

As far as I can see, I can't model most South Asians as Sintashta and part BMAC. They usually just get zero BMAC, and I often see a lot of "infeasible".

So I don't think there were any major migrations of BMAC people to South Asia in any form.

But there were probably economic and political contacts between BMAC and IVC, or at least parts of IVC, that resulted in limited bidirectional population movements between the two. Also, some steppe-derived groups with BMAC admixture probably settled in former IVC territory, and maybe even deep in India.

None of this is very controversial. But I'm sensing that some people are seeing in BMAC the only real chance to argue that Sintashta wasn't Proto-Indo-Iranian (because IVC sure as hell wasn't) so this type of discussion will probably keep popping in the comments here for a while.

Seinundzeit said...

RK,

"Are there any "weird" linguistic substrates or historically extinct languages in the Swat area? The Butkara_IA samples are very dissimilar to present-day South Asians, resembling prior SPGT, despite being from ~0AD."

I know that the question was directed at Sanuj, but I can't resist... (lol)

There is nothing unusual/unique about the linguistic situation of the Swat region, in relation to southern Central Asia and northern South Asia.

Prior to the Pashtun invasion by the Yusufzai, the area was divided between Dardic-languages in the upper valleys and Prakit-derivatives (Punjabi-like) in the lower reaches.

"These people seem to have went extinct in the historical period."

I don't think this is true.

I can model some of my northern Pakistani friends as being 60%-65% SPGT, but with the addition of intense Scythian, Near Eastern, and BMAC-related admixture.

The additional Scythian percentages can be chalked up to historical incursions from steppe-rich populations (Kushans, Hepthalites, etc). The Near Eastern percentages can be attributed to the intensification of West Asian/Central Asian/South Asian trade and interaction during the historical period. And the additional BMAC-related percentages are likely reflective of the fact that my samples are from deeper inside Central Asia than the Swat valley.

At the end of the day, people like the ancient Swat valley folk might never have existed in South Asia.

Contemporaneous people in South Asia proper may even have been nearly identical to contemporary Indians (perhaps "nearly identical" is unnecessarily strong language, but you get the idea).

Essentially, these ancient samples might not be quite relevant for Indians and Pakistanis with connections to Greater India, but they are certainly of importance in the modelling of Dardic people/Nuristani people/Northeastern Pashtuns.

ryukendo kendow said...

@ Davidski

Thanks for the heads up.

However, do you see any variation in BMAC-Sintashta ratios in the SPGT, who do have it? Are they positively or negatively correlated among the SPGT samples themselves?

@ Sein

"I can model some of my northern Pakistani friends as being 60%-65% SPGT, but with the addition of intense Scythian, Near Eastern, and BMAC-related admixture. "

So you confirmed Rob's finding of affinity to some subgroups only? Do you mind posting a table or a list of fits? Will be nice to see geographical structure in affinity to SPGT in S Asia.

The main reason I thought it would not be possible was I couldn't see how later waves could dilute the "excess similarity to BMAC and Siberia_N" signal in SPGT in D stats in the figs posted by David above unless the SPGT% was close to 0, so that the SPGT-containing pops hug the trendline formed by other S Asians. Maybe the Scythians and so on in your fits have some effect that actively reduces affinity to SPGT instead of just diluting it.

Davidski said...

@rk

There is some variation. The SPGT groups just barely above the trend lines in the graphs above behave more like modern-day South Asians in this respect.

Davidski said...

@Vara

That Mycenaean is not labeled as elite. Keep repeating that maybe it would become a reality.

Well, in the paper she's described as the elite Mycenaean individual from the ‘royal’ tomb at Peristeria.

Seinundzeit said...

RK,

Sure, I'll collate the results for you.

"Maybe the Scythians and so on in your fits have some effect that actively reduces affinity to SPGT instead of just diluting it."

I'm not too sure about that, since we're not really seeing SPGT percentages in actual South Asians.

As one would expect, the only populations with a strong affinity to SPGT are South Asians from the northern fringe, and South Central Asians.

Chetan said...

"Because you know that evidence is hinting towards a late influx of Steppe ancestry, possibly via Indo-Scythians or Kushans or others"

Because all the South Asian samples we have now are in the 1200 - 500 BCE period. Nevermind that's still too early for Sakas and Kushanas.

Alberto said...

@RK

Yes, I saw those graphs with the "traditionally priestly" groups highlighted. It seems to be based on the results from the qpAdm models in the supp. tables. I wish we could see the raw information from the other figures (S4.1 and S4.2), which I think would give a better picture regarding geographical origin, caste, language, recent history, etc... to see if there is any correlation. From the populations in G25 I don't see anything striking, rather a moderate variation (mostly around 20-30% of the West_Eurasian side) in a rather random distribution, with some correlation to some geographical areas (but when North Kannadi seems to be in the same ballpark as Kalash and higher than Kho, I don't think that modern populations are going to be informative of ancient events. I just think it's better to dispel dubious claims unless they're clearly not dubious).

Re: the variation in the rations of BMAC to Sintashta in the SPGT samples, it's much better to look at the individuals than at the groups. For that G25 is quite useful. Here a quick example:

https://docs.google.com/spreadsheets/d/17uUbaggAOWRbqMSqnMldGbvbnbnO_aEcavMs5WTGxws/edit?usp=sharing

I don't seem to find correlation, but just looked at it briefly before uploading it.

BTW, for these kind of tests you might find useful the script I wrote to do them, which is compatible with nMonte but easier for these tests (the README file contains info about it):

https://drive.google.com/open?id=1GnzWqr-hsa_GnwnWqx_as7ky0eHGKg-4

Davidski said...

@rk

In qpAdm, Sintashta is a relatively poor reference population for Hajji_Firuz_BA.

The reason it's showing up in your model, and also producing a good fit, is because you're balancing it out with more CHG-rich populations, like Armenia_EBA.

This is more difficult to do in qpAdm without seeing "infeasible" or negative coefficients, and generally messy results. So it's probably not something I can copy.

Chetan said...

@Vara Can you give some references to back up your certainty that the Indo-Aryan superstrate or adstrate in Kassite language dates to 1800 BCE? I have not looked into this much, but Wikipedia gives a date 1600 - 1155 BCE for Kassites


A much more parsimonious expanation would be that the Kassite and Mittani superstratum was due to the same Aryan movement into Mesopotamia around 1500 BC. And we have already seen how that is made more than possible by the early presence of Andronovo groups in Kyrgystan and Uzbekistan.

Sanuj said...

@RK

"we already know that almost all Indian caste groups were founded by small numbers of ancestors around 500BC-0AD or so"

http://www.pnas.org/content/113/16/E2215
"Emergence of sociocultural norms restricting intermarriage in large social strata (endogamy) coincides with foreign invasions of India"

Try to be more up to date with facts next time. Also, stop correlating caste with purported coming of IE in India, caste as we know it today are a result of very late social dynamics.

mzp1 said...

@Chetan,

At the moment you are trying to derive two different language groups from one cultural horizon. It is a non-starter as we do not have any differentiation between Andronovo groups that would allow us to identify it as consisting of two different groups.

YOU CANNOT JUST SAY ANDRONOVO IS INDO-ARYAN WHEN IT SUITS YOU, AND THEN SAY IT IS IRANIAN AT OTHER TIMES. YOU NEED TO PRESENT A CONSISTENT AND COHERENT HYPOTHESIS.

Maybe they even spoke Tamil and German? Lets just throw all the languages into this undifferentiated mass until we can derive every language from the Steppe.

mzp1 said...

YOU GUYS REALLY NEED TO WORK ON WHICH PART OF ANDRONOVO IS INDO-ARYAN AND WHICH PART IRANIAN.

I dont like to keep going on about it. But this steppe theory has been around for a long time and it is not a minor point. Now you have both the Genetic and Archeological data.

Because almost every other aspect of the theory depends on this early differentiation.

It is not acceptable to gloss over such a major event, one which would define the structure of the Andronovo horizon more than anything else. Without this the theory cannot be taken seriously.

Vara said...

@RK

I wanted to add another point. The actual percentages of Steppe and BMAC are entirely dependent on the actual IVC samples we will get.

@Chetan

Just because they took over 1500BCE doesn't mean that's the invitation date. In one of the texts of Tell Leilan dated to 1760BCE we get the first mention of the 'maryanni' warriors. They are believed to be a warrior-noble class intrusive to the Hurrians and Semetic people. The word is thought to be a Hurrianized Indo-Aryan word. The Mitanni were also called 'maryannu'. Since you're familiar with Parpola you can check his thoughts on that because he goes over it. Mind you the Old Babylonian period is when we get the first mention of Baghdadu as well.
PS. Witzel thinks there is an II influence on Gutians and Lullubi but doesn't go over it much.

mzp1 said...

Guti would be derived from Dasa->Daha->Dahae->Daghae->Getae->Guti

And they would be more Iranian than Indo-Aryan

old europe said...

mzp1

Andronovo is a very big and large archeological complex spanning thousand kilometers . I think there's no doubt it could contain both Indo Aryan and Iranian

mzp1 said...

There is a lot of doubt, because Indo Aryan is not really attested in the Steppe, unlike Iranian, which is very dominant.

How do you explain that?

This is why we need evidence of it being there, or something pointing to language differentiation.



ryukendo kendow said...

@ Sanuj

Please at least read your article before posting it.

I went to read the author's disagreement with Basu et al, coming along fine, yep, until the part when he used "historically more accurate" human generation times of 18, 15 and 13(!) years, and did a double take.

I mean, human fertility rises and falls in a curve, and it has long been known that the generation times used in most papers (25 years) is an underestimate of the mean generated by the curve, with 27 years used by Poznik et al, for example, and 27.5 years proposed by ISOGG, even for those societies where early marriages were thought to bias the average closer to 22.5. The author's shrinking of the generation times is the only change which the author uses to support his timeframe, of Islamic invasions.

You telling me that on average in India, people were born on average when their parents were 18 years old and younger, and even 13(!) years is plausible? In fact the author does not even provide any historical data to back up his estimates of human generation times, he resorts, quite openly, to circular reasoning, saying that the shorter times cause the onset of endogamy to fall with the Islamic invasions and "we know" the Islamic invasions should coincide with the onset of Caste, so we should use these lengths.

Despite elsewhere aknowledging that Basu et al themselves allude to historical data showing the Caste endogamy in the Pala Empire, i.e. before the Islamic invasions.

I was wondering how such an extremely shoddy piece could end up in PNAS, then checked and realised it was not a peer-reviewed article, just a letter to the authors.

If he had subjected it to greater scrutiny by geneticists or historians, perhaps in a peer-reviewed article, it would probably be summarily rejected, especially an article proposing the middle estimate of human generation times in India to be 15 years old.

A finding that South Asians have generation times a full decade shorter than all other human populations should be one of the most significant findings in anthropology, medical genetics, and population genetics of South Asia, if it is true. It will affect the calculation of many things, even economic history and demography (the population would grow much, much faster than elsewhere and hit the malthusian limit much quicker, rates of death will be constrained to be much higher, lifespan much shorter and standards of living lower than the estimates obtained so far, for example). It is such an utterly implausible thing to suggest that I wonder at the author's motives.

old europe said...

mzp1

here's a quote from Kuzmina I posted last month. I do not pretend it can settle the matter.

"Indo-Iranian toponyms are also found on the Middle Volga and the Urals (Popov and Loyfman 1962). N. L. Chlenova (1983a; 1984) plotted the toponyms and demonstrated that Iranian toponyms were spread over Timber-grave and Andronovo territories, but part of the Andronovo toponyms can only be interpreted as Indo-Aryan; here also was included the Altai by A. M. Maloletko (1986: 70-75). These data are extremely important for the final resolution of the problem of the cultural and ethnic attribution of the Fedorovo complex. Due to the fact that Indo-Iranian toponyms of the pre-Scythian period have been found on the territory populated only by Fedorovo tribes, the hypothesis identifying the Fedorovo population as Ugrian that has been proposed by V. N. Chernetsov must be rejected, and the hypothesis of the Indo-Aryan attribution of Andronovans can be supported".

This does not rule out that you can be right and that it could have been gone the other way round.

ryukendo kendow said...

@ Alberto

I just think it's better to dispel dubious claims unless they're clearly not dubious

... I think it is extremely difficult to see that scatterplot and conclude that the distribution is due to chance, a simple T-test is all that is necessary and even to have everything fall in the top 60% of the distribution is already very nonrandom, much less have the Brahmins come in at the top few ordinals for all horizontal slices of the distr.

Not sure this is something you can dispel by saying, "I don't see anything", (even then you clearly can) there has to be some statistical thinking.

Chetan said...

@MZP Andronovo was a large horizon of cultures and they almost certainly didn't all speak the exact same language.

https://erenow.com/ancient/the-horse-the-wheel-and-language/the-horse-the-wheel-and-language.files/image147.jpg

And like I said and you chose to ignore, Gathic Avestan and Rigvedic Sanskrit are so close linguistically that they can, in parts, be treated as two dialects of the same language.

Really, there is nothing complicated in this. It's just your reluctance to be objective.

Sanuj said...

@RK

India has one of the highest instances of child marriages in the world. Even today a lot of child births happen at 18 years of age.

But leaving that, even going by the proposed dates of Basu et al, taking the age as 22.5 years, the timeline falls after 300 AD.

Even if I take the maximum age that you propose(which can't be true for the time and location), we still fall to around 100 AD.

None of this anywhere near your arbitrary spitting of dates like "500 BC to 0 AD", which you loosely throw around just to fit into some mythical narrative that you have in your head.

mzp1 said...

@Chetan,

"And like I said and you chose to ignore, Gathic Avestan and Rigvedic Sanskrit are so close linguistically that they can, in parts, be treated as two dialects of the same language."

Interesting random fact, but what does it have to do with it? They are different languages is my point, so there is linguistic (and possibly religious) 2-way differentiation, where is it found in the Andronovo?


"Really, there is nothing complicated in this. It's just your reluctance to be objective."

Ok buddy, glad it is not complicated, so can you answer my question then..

"YOU GUYS REALLY NEED TO WORK ON WHICH PART OF ANDRONOVO IS INDO-ARYAN AND WHICH PART IRANIAN."

Still waiting for your response?

And what happened to the Indo-Aryan river names?
And the Indo-Aryan speakers, did they ALL migrate to South Asia?

If you dont know, just say "I dont know", atleast Eastpole is honest about it, and I have no issues with him, you on the other hand...

ryukendo kendow said...

Lol so ~50% of all births happen before 18 years and ~50% happen after? When was the average age at first birth, 13?

There is a long tail of older births and conceptions, stretching to 40s for females and even later than that for males, such that the most recent measurement for avg age of male conception, among small-scale societies and historic societies, is 31. There is a strong mental tendency to conflate age at first birth with average age of all births--this is certainly the case for me, and has to be avoided.

The 500BC to 0AD figure for onset of endogamy is taken from a series of estimates for different Caste groups in Reich et al's original paper. Basu et al are not the only ones who have come up with estimates.

Either way, I am interested in the populations that were implicated in being the founders of various Indian caste groups today, which lie away from the SPGT and belong to a cline of their own, low in Siberia_N and BMAC ancestry and higher in Sintashta ancestry. If you want, we can widen the age to 1000BC-400AD, it is immaterial.

Much more arbitrary is the range of 18-13 years (why not 22.5 years? 20 years?), which were probably specifically chosen so as to fit all the dates right at the Islamic invasion.

mzp1 said...

Chetan you are so deluded you are attempting to imply an area with only

Iranian River Names
Iranian borrowings into other languages
Iranian Tribal Names
Iranian literature
Iranian speakers as far back as we can see

And NO CONCLUSIVE INDO-ARYAN items (we even have this in Mittani for gods sake)

...as the source of INDO-ARYAN languages into India.

I mean, dont you even see how absurd this is. I mean we even have Indo-Aryan attested in the Middle East, and not one conclusive example in the whole of the Steppe, itself a massive region, and supposedly it's homeland.

You guys are losing the plot.

Chetan said...

@mzp Indo-Aryan and Iranian developed from two closely related dialects spoken by two (or perhaps more) Andronovo groups.

If you are failing to appreciate the arguments objectively, I don't think there is any point in carrying on a discussion.

Have it your way. Early PIE (ie Sanskrit) was spoken in the Neolithic in the Indus Valley. From there, it magically morphed into Hittite, Tocharian, Celtic, Italic, Germanic and Balto-Slavic which all migrated their way from South Asia into their respective positions by 2000 BCE. (it can't be earlier than 1900 BCE when the first words of Hittite are attested)

Chetan said...

correction-later than 1900 BCE

mzp1 said...

@Chetan,

I know you're being sarcastic but to be honest it feels pretty good to atleast see that written on here. Now, that is what I call I nice elegant theory.

Celtic, Germanic Balto-Slavic would be derived directly from Iranian.
Not sure about the others.

I need to do some research on linguistics to be able to fit it in.

Iranian is already somewhat in between those languages and IA in terms of Centumization and vowel diversification.

Chetan said...

@mzp Also, I notice that you have been doing nothing other than trying to falsify the hard though-out models of others on flimsy grounds without offering any credible alternative yourself.

Easy job huh?

mzp1 said...

Indo-Aryan->Iranian-> Most Euro Languages (Diversification West of Pontic Steppe as Steppe Iranians merge into Euro populations in different waves)

mzp1 said...

@Chetan, What do you mean, I have been clear about my theory here, and have posted it a few times.

mzp1 said...

You can ask me any question, you will probably get a lot of "I dont knows" for now but atleast I can be honest. But you wont find many problems with the theory I dont think. Unlike the Steppe 'theory'

ryukendo kendow said...

@ mzp

Celtic, Germanic Balto-Slavic would be derived directly from Iranian.
Not sure about the others.

I need to do some research on linguistics to be able to fit it in.

mzp1 said...

@RK,

Nice comment brah, very enlightening.

mzp1 said...

You guys are obviously so clueless on the topic that you cant even pick holes in an ill-thought-out theory of an amatuer like myself, yet you have such strong opinions on the urheimat question, strange combination to be honest...

old europe said...


mzp1

I'm open-minded as I told before. I was referring to IA and Iranic place names in the steppe. I do not know to be honest how massive are IE place names in the indian subcontinent maybe you know better.
To play the devil's advocate two things
1) AFAIK the processes is from Centum to Satem
2) If the homeland of IE is India I would wonder why they managed to influence the language of ireland and portugal ( for example) but failed to make it to southern india.

mzp1 said...

@Old_Europe,

I appreciate your comments and you seem genuine and engaging, it is nice when we can atleast talk about our theories.

I didnt see any examples, in your comment, of IA places names in the Steppe, so couldnt verify anything.

As far as I know it is a consensus that (Northern) South Asia is very much IA in terms of place names. I just assume they are all IA.

1. Yes, but I dont know why. They say K->S is more common so it must have started with K, but I am not sure of that. In IA->Iranian it goes S->H/K with S being the original.

2. The initial movements were to the North, BMAC being an early result of Iranian civilization. Populations in the Punjab were more interested in trade routes and resources in the North and West, as it is until now. Not much reason to go South I dont think. From South Central Asia they expanded into the Steppe and went into Europe, with possibly the Greeks and Anatolian branches taking routes from South Caucuses.

Mr. Kulkarni said...

@chetan

"Also, I notice that you have been doing nothing other than trying to falsify the hard though-out models of others on flimsy grounds without offering any credible alternative yourself."

Falsifying a theory by providing contrary evidence while giving no solid alternative is a perfectly acceptable form of science.

mzp1 said...

Chetan thinks that

"No IA attested in Steppe"
is "flimsy grounds"
for the theory that
"IA originated in the Steppe"

mzp1 said...

@Old_Europe,

To be honest, I dont really want to push any particular theory, as I rather just say we dont know where they originated, but I really dont like the Invasion Theory so I am just going to make their lives very difficult.

I am going to set up a youtube channel to talk about IE stuff, and woulnt want to push people away by pushing for a particular origin.

I am very critical of 'science' of late, it is one thing to determine if you have a particular gene, quite another to determine when, for example, caste endogamy started in South Asia. I have seen 'science' proved wrong too many times to give these guys credit without understanding what is going on myself.

RyuKendo, for instance, is confident he knows when caste endogamy started in India. I would like for him to explain it to me.

Sanuj said...

@RK

Your estimates were wildly wrong based on the most recent study on the subject, and which was conducted on top of the Reich et al study. Unless you have something substantive to refute it, your quoted period was wrong.

Coming to "populations implicated in being the founders of various Indian caste groups today". They were already the elites in society, and have been historically mollycoddled throughout Indian history by the ruling classes as they were/are the priestly class and held sway over the masses. This included intermarriage.

The same would also apply on Indo-Scythians and Kushans who also had their Eastern capitals in Mathura in UP, which is a major religious center of North India. All of this has nothing to do with IE or "Aryans".

Chetan said...

@Mr Kulkarni @Mr MZP Both of you need to gain some perspective on the issue.

This is an archaeological/anthropological model. So there are still gaps which will hopefully get addressed by future development. So don't expect it to answer every tiny (irrelevant) objection you have in mind.

Like how mzp thinks that there aren't any Indo-Aryan toponyms from the steppe. I don't even see how this "falsifies" the steppe theory at all. Do you expect every place on earth to always bear the toponyms of all the languages that were ever present there, especially one that belonged to pastoralists That's such a ridiculous argument.

Overall, the steppe theory explains best, how IE languages spread over such a huge region during the Bronze Age. If you have a better model, feel free to share it with us. Real models, not flowcharts like mzp posted a while ago.

Chetan said...

I think it's best if you guys take some time out to gain some perspective on the issue. Because as it is evident, you are clearly wasting everyone's time without adding anything new. The same decades-old worn out objections.

People like Anthony have invested a lot of their work in coming up with a coherent model for how late PIE spread through the Eurasian steppes. You can start reading them for starters, maybe even post a detailed counter-model for the rest of us to evaluate.

Anyway I'm done with this discussion.

mzp1 said...

I'm glad that you're done.

People like you shouldnt be talking about this topic.

mzp1 said...

" Real models, not flowcharts like mzp posted a while ago."

Dude, my model is real, and not filled with made-up bullshit like PIE and Indo-Iranian.

Lol, this guy wants to tell me my model isnt real, but has a fit when asked a simple question about where his model places IA in the Steppe.

ryukendo kendow said...

@ Sanuj

Lol what do you mean by "based on the Reich study?" They didn't even sample the same populations as Reich, or use his samples. In fact they sampled much fewer populations than he did.

Reich gives 100 generations for onset of endogamy for Vysya, which translates to ~500BC given 25 yrs generation time.

mzp1 said...

What happened in ~500BC to cause the onset of endogamy?

I'm asking RyuKendo

old europe said...


Chetan
Like how mzp thinks that there aren't any Indo-Aryan toponyms from the steppe. I don't even see how this "falsifies" the steppe theory at all.

If there weren't ( but I believe there are) this is a factor that at least weakens the steppe theory... absolutely yes.

Do you expect every place on earth to always bear the toponyms of all the languages that were ever present there, especially one that belonged to pastoralists.

If you push for a big demic diffusion of a language family absolutely yes....absence of toponyms give grounds to a model of elite dominance or cultural influence ( but then all the movement of R1b ...R1a ecc.)
For example i'm often told that as an italian my ancestors ( the neolithic farmers) should have spoken Vasconic. They say that ancient europe from portugal to poland and from britain to bulgaria was Vasconic......but then were are all the vasconic place names gone since in Europe you cannot find a pattern of toponyms that are not IE.

old europe said...

mzp1

best wishes for your channel!

mzp1 said...

There are 'castes' in Early Vedic literature. So caste endogamy must have developed after that.

'Scientific' dating of the onset of Caste endogamy needs to be expressed in the context of South Asian history/literature, and needs to fit in with everything else we know.

Otherwise it is meaningless. This is the problem with people like you, you think one data point, even that a highly dubious one, dependant on the reputation of a particular scientist, is enough to overturn many other pieces of information. Highly arrogant and misguided.

Obviously the onset of caste endogamy would be related to historical societal changes and combining the two would give a much more useful narrative than just "oh this scientist guy put something into a computer and it says X"

mzp1 said...

@Old europe

"best wishes for your channel!"

Thanks a lot

I am going to put up a video about the Migration Theory to get things started (already have a video up but it is not IE-related). I will post a link here once it is up.

Mr. Kulkarni said...

@chetan

Kalibangan 2500bc-1800bc fire altars have been found. many of them in public places, with the sacrificial stone (pali peedam).

One of these is very close to a public bath, correlating with vedic practice of bath before rituals.

However, some western experts reject this because it doesnt fit into their 'models'. so pardon me for not buying into their theories wholesale.

Lets recap previous hypothesis of male aryan invasion:
1. No chariots or horses of 'invaders' have been found
2. In fact, no chariots found till 200BC in fact - there goes the chariot as an IE marker.
3. No correct R1a adna found yet
4. No male R1a adna found yet

Yet we should stick to the old theories?

2nd half of yajurveda has vedic people moving east to ganga plains. Iron work is mentioned in this. 1800BC iron work has been found in Chandauli district on ganga plains.

Therefore, I shall stick to my hypothesis of IA speakers entering India from Iran prior to 2000BC, closer to 2500-3000BC.

Sanuj said...

@RK

You are 'lol'ing too much. ON top of Reich study means, using the broad outline of Reich's study and refining it further. Learn to accept defeat.
From Basu et al:
"We surmise that the number of ancestral components in the populations of India may have been underestimated by Reich et al. (9) because of (i) lack of inclusion of tribal populations, who are considered by anthropologists to be the autochthones of India, and (ii) inadequate representation of the geocultural diversity of India in the set of sampled populations, and (iii) selective removal of some populations based on deviance of their genomic profiles. Our study has corrected this deficiency and has provided a more robust explanation of the genomic diversities and affinities among extant populations of the Indian subcontinent, elucidating in finer detail the peopling of the region."

ryukendo kendow said...

@ Sanuj

Sanuj, the fact of the matter is that you have no idea how to read scientific literature, and no idea how to understand what the scientists are talking about. You are obsessed with 'defeat' and 'victory' to a degree that you have embarassed yourself.

Reich et al chose to look at populations on the ANI-ASI cline, of course they would remove the TB and AA populations and will not discover their components. How on earth does that account for Caste Endogamy for populations speaking Dravidian or IAr languages, which are all on the Dravidian IAr cline? Did Basu et al even sample the Vysya? Reich et al focused on endogamy in groups like Vysya and Naidu and etc, but Basu explicitly said that endogamy set in at around 70 gen ago in "upper caste IAr populations" which the Vysya and Naidu and so on are not. How then could they contradict each other's findings?

Reich sampled more populations than Basu et al, this is just a fact. Go and read the supplementary material of both papers and the population list. The innovation of Basu is to include TB and AA populations in their ADMIXTURE and other analyses, and to talk openly about the AA and TB admixture components and their dating.

Just stop posting selective readings already. If you wanna engage, do so honestly.

Matt said...

@RK: To add to this the Globular Amphora were identified as Indo-Europeanized from archaeology but in fact they were separated from EHG and steppe populations by 1000 years of genetic isolation.

On that tangent, I do wonder though; just from messing around with Fsts, GAC as a whole are set a little "east" on the cline, while the Ukrainian Eneolithic samples, putatively Indo-European in some models, are relatively "west" in keeping with substantial Ukraine_N ancestry, and variable within that (apparently according to cultural layers?), and further there are some differences between the Polish and Ukrainian GAC which seem apparent in some direct D-stat comparisons.

Perhaps something will square the circle here between what Marija Gimbutas I believe saw as signs of interaction with the nascent steppe cultures of Ukraine and the generally fairly similar pattern of overall GAC's ancestry to Britain / Germany / Sweden middle to late Neolithic. Not a mass migration of the Corded Ware type but some kind of contact between a Funnelbeaker like group and groups which were a bit different from the EHG / Steppe_EMBA. (Contacts which could become over-emphasised in degree if we were specifically had a hammer of kurganisation and saw suitable nails everywhere).

(Or perhaps (probably) I'm looking at the Fst scores too much, and trying to be too clever).

Chetan said...

@kulkarni

"1. No chariots or horses of 'invaders' have been found"

Domesticated horses are known. Not just in PGW, but also a lot of post-Harappan cultures

"In fact, no chariots found till 200BC in fact - there goes the chariot as an IE marker."

About chariots, I'm not aware of the findings. But if there are no chariots in this period (1200-800 BCE), that would falsify our own scriptures which states chariots were used! So if you want to take that leap, go ahead

"3. No correct R1a adna found yet"

Your previous two objections were quite reasonable, but this. What are the chances that R1a-L657 alone is from somewhere other than steppe/ forest steppe.

"4. No male R1a adna found yet"

For now

Chetan said...

@kulkarni Why should we even expect chariots to turn up in the archaeological record ? Unless the Indo-Aryans practiced chariot burials like the residents of Sintashta. I think there is no textual evidence for that. Not even many human burials, since cremation became the norm. Unfortunately for us!

mzp1 said...

"Your previous two objections were quite reasonable, but this. What are the chances that R1a-L657 alone is from somewhere other than steppe/ forest steppe."

I dont know. How about using this formula : R1A found / Steppe Pop sample size
So that would give us: 0/x = 0

mzp1 said...

@Chetan

For once I agree with you

"@kulkarni Why should we even expect chariots to turn up in the archaeological record ? Unless the Indo-Aryans practiced chariot burials like the residents of Sintashta. I think there is no textual evidence for that. Not even many human burials, since cremation became the norm. Unfortunately for us!"

Same goes for IVC

Sanuj said...

@RK

You have been shifting goalposts throughout this conversation. Your exact words were "especially when we already know that almost all Indian caste groups were founded by small numbers of ancestors around 500BC-0AD or so"

Starting with grand reasoning about what age is the best statistical age, to Basu is talking about just "IAr upper castes". So that does not make it "almost all", does it? Who is selective here?

Other than that, Moorjani et al have also come to similar conclusions of which Reich too is a co-author.
"Our analysis documents major mixture between populations in India that occurred 1,900–4,200 years BP, well after the establishment of agriculture in the subcontinent. We have further shown that groups with unmixed ANI and ASI ancestry were plausibly living in India until this time."
https://www.sciencedirect.com/science/article/pii/S0002929713003248

Just stop posting selective readings already. If you wanna engage, do so honestly.

Davidski said...

R1a-M417 is from the Pontic-Caspian steppe.

That's obvious now, and the fact that it's a huge marker in India, especially in Indo-Aryan speakers and upper castes, means that the game's over.

There's really nothing left to debate now except the details of how it all happened.

Santosh said...

@ ryukendo kendow
"@ Santosh

If you are reading this, I just wanna say thanks for the posts previously on the Dravidians, well appreciated!

Are there any "weird" linguistic substrates or historically extinct languages in the Swat area? The Butkara_IA samples are very dissimilar to present-day South Asians, resembling prior SPGT, despite being from ~0AD. These people seem to have went extinct in the historical period."

Yeah, the stupid idiotic ancient Dravidians have somewhat become an obsession of mine- so I just grabbed that opportunity there like anything to write all I knew and thought till the point. I also felt a bit silly about concentrating perhaps a bit too much on the etymological side, though I also somewhat found that it took me in the right direction- the best example to illustrate this would be the history of the word paNTakApu. Wikipedia which I checked later said that paNTakApu is the name of cultivators specific to a particular region. It probably had nothing to do with the paNTa, 'crop' part of it- at least exclusively. For some reason, the farmers who got regionally separated in a particular case styled themselves paNTakApu, 'crop farmers'- it may not be because they were suddenly beginning to grow crops and all their predecessors did not. But that said, etymology still led me in the right direction because it appears each and every Telugu agriculturist caste now separated were called by the name Kapu in the olden days. Without the Dravidian Etymological Dictionary, I would never have been able to know this as I don't typically read Wikipedia a lot and also not bother to remember all the minute historical details even if I read. The kApu, kApuram connection blew my mind when I encountered it in the DEDR. It so happened that the current kApu at a previous stage in Telugu descended from a form with a preconsonantal nasal in it- *kAmpu. Now this kApuram word which means 'married life', 'household', etc. was also apparently *kAmpuram at some olden point. The etymological connection between these two words is I think, lost from the mental grammars of most Telugu people, including myself. One reason may be that the new form kApu without the nasal got identified more and more with a genuine Old Telugu kApu which descends from the root *kA, 'to protect', and thus means 'protector', 'protection'. It is not uncommon to hear Telugu people popular-etymologically associate this 'protector' meaning in the senses of 'protector of fields', etc. with the caste name kApu! Somewhere between these two times, kApuram also apparently got ossified in the mental grammars, in fact like many pure Telugu words really.

I'm mostly clueless about Indo-Aryan linguistics, except perhaps some very minor Classical Sanskrit- so I would have done you zero help with the Swat area languages. I see Seinundzeit did a wonderfl job about that.

Thank you finally. Hopefully, we'll find the Indus language someday- but there really is no other way than to decipher that script (if that is that) conclusively; people cannot begin to associate Dravidian languages with Harappan civilisation also (please note that I'm not making any comment whatsoever about Indo-Aryan languages and Harappan civilisation)- it may be that just some Harappan cultural and genetic influence got into native cultures of Gujarat or Maharashtra or some such place (showing the genetic correlation) and that Dravidian was one of the native languages of those regions that have survived and need not be the Harappan language.

ryukendo kendow said...

@ Sanuj

That accusation, coming from you, is pretty f*cking incredible.

The reason why I stated that almost all Indian Caste groups became endogamous around 500BC-0AD or so was because I wanted to provide some context for my question of which populations were most important in the genesis of caste groups (because its not SPGT, which instead seems Dardic), and where they could have been at 1000BC-0AD. Why is this is so triggering to you, such that you obsessively focus on pushing the dates later and later? It does not matter to me whether it happened 1000BC or 400AD.

The range produced by 100-70 (from Reich and Basu respectively) generations assuming 25 yrs, i.e. >500BC and <250AD, is not even that that different from my claim ("500BC-0AD or so"). You are not gonna get the data to support "very late appearance of Caste in its present form due to recent processes! " or "Islamic invasions changed Hindu culture, they introduced caste!" no matter how you misrepresent it.

The variable dates that Basu stated they obtained (and Basu did see that all groups in their dataset became endogamous at some point, not just upper Castes! on this pt you are just wrong, pls read the paper), including very late dates, were all for Tribal AA and ST populations, (Gond, Ho, Paniya, Irula, Kadar, Santal, Korwa, Birhor) which are not caste populations. For the Caste populations they are indeed at 70 generations (~250 AD, probably <250AD since they use the 25 yrs genlength estimate), in fact very precisely at 70 generations for all Brahmins which is striking and kinda surprising.

If you wanna focus on "but 0AD" then you're pretty focused on proving me wrong in every circumstance, than in engaging in any fruitful engagement with my question.

I actually find it very interesting that the data seem to indicate an onset of Caste endogamy first in the lower-status IAr-Dravidian populations and then ending in the highest-status groups, with tribals entering the system much later. This may be in keeping with the sequence of events in texts (i.e., first without denigration, then Outcasteism appears and the lowest status nontribal groups are ostracised ever since the Manusmriti, then more groups get ostracised, finally by Pala times we see historic evidence of social sanction of endogamous cleavages throughout society, and lastly territorial states control tribals in late historic times). The different dates of occurrence of endogamy form an interesting pattern that may bear investigation by historians.

Proud farmer said...

I can't understand why there is South Indian bashing in the very first post on this page. We are coming off as an insecure people or something. The research people are doing are learning to make us more proud of our heritage not less. We always thought the indo Aryana conquered us imposed upon us. Now we know there is hardly any steppe admixture in us. The religion we must have accepted voluntarily. Our Kings were our own. The Brahmins in south came here by invitation. They didn't invade. Our contribution to religion is also considerable. The Bhakti movement started in the south which changed the face of Hinduism all over India. Some of the greatest interpreters of Vedas in the medieval period were shankaracharya and madhavacharya and ramanujacharya who were from the south. Our part of the country is more advanced we have less incidence of rape or other criimes in spite of having more than fifty percent hunter gatherer ancestry. What is not to be proud of. Don't judge by one person who may have his own axe to grind. You geneticists have made us proud. And our politicians have one less pet peeve to take advantage of.

ryukendo kendow said...

You also completely misinterpret the paper by Priya Moorjani, in fact it directly contradicts your claims.

If the majority of mixing happened between ASI and ANI at 4200-1900BP then that means by 1900BP the mixing basically ended. Moorjani says:

" In a subset of groups, 100% of the mixture is consistent with having occurred during this period. "

This means after 1900 BP (i.e. 100AD) a subset of populations basically stopped outbreeding 100% with those with higher or lower ANI-ASI ancestry, while for many others the endogamy was still a bit leaky, but was still there.

Moorjani concludes:

"These results show that India experienced a demographic transformation several thousand years ago, from a region in which major population mixture was common to one in which mixture even between closely related groups became rare because of a shift to endogamy."

So they conclude that the endogamy for sure began when the ANI and ASI stopped mixing in the various populations in India they sampled, i.e ~100AD, but the termination of gene flow was probably gradual and had an onset before that, as is usually the case for such phenomena.

ryukendo kendow said...

Its nice that she established that for some populations in India statistically speaking 100% of all ANI-ASI admixture happened before 1900BP, meaning more or less that for that subset caste endogamy can be firmly established to date to before that time and was very tightly enforced after.

Hmm, previously many ppl were extremely skeptical of Moorjani's results (they expected dates, especially in S India, to be earlier). However given the new aDNA it appears ASI is redefined to consist of a separate admixture event between IVC_Periphery pops and extra AASI, which could plausibly have happened at a very late time, even after the late Harappan. This makes Moorjani's results look much more sensible.

mzp1 said...

@Proud Farmer,

I have nothing against South Indians like yourself.

But I have seen South Indians on the internet trying very desperately to push for the Aryan Migration Theory, which, obviously, comes off as desperate and insecure.

I agree, and I said, South Indians have much to be proud of. The IVC is not one of those things though, just to be clear.

Sanuj said...

@RK

Your whole context around this question is wrong. Because you are starting with an assumption that some migrating group is the cause of caste in India, which is not the case as the data is showing. Studies are consistently showing that the onset of endogamy has happened in the historical era. 1900 BP or 1700 BP is the onset of endogamy not 500 BC as you claimed earlier, to tie it to some migrating group.
It has nothing to do with any migration which should be tied to the question IE groups or languages. I am not saying that caste is unreal, Islamic invasions caused it, it might very well have happened in the Gupta era or earlier due to strict priestly strictures.

And what do you mean "because its not SPGT, which instead seems Dardic". Dardic is a ethno-linguistic grouping, which would be as much a part of the caste setup as others back in time. Kashmiris are a Dardic people, and we have Kashmiri Pandits, probably the oldest Brahmins we have, who knows.

Just a fun trivia(not claiming anything): The story of RigVedic Rishis starts with Kashyap, on whose name Kashmir is. Many people feel that IE story has to have a beginning somewhere in the mountains. :-)

ryukendo kendow said...

@ Sanuj

Lets lay out the data:
Basu et al get only data from upper caste groups where gene flow stops at 70 gens (plus tribal groups with much more variable dates).

Reich et al get stronger signals from lower caste and Dravidian groups starting at 100 gens and later, i.e 500BC and later (some are later, but the earliest is at 100 ~500BC or so).

Moorjani find latest termination of admixture at 1900BP, i.e. all endogamy began to set in before 100AD, and admixture mostly stopped by 100AD, or even completely stopped by 100AD for a subset of populations sampled.

This is what the facts show.

On the question, here is what I stated:
The fact is that the migration from Central Asia already impacted Indian groups by the time of the SPGT, i.e. the prehistoric Iron Age, far before any issue with Caste. The problem is the SPGT do not represent the direct ancestors of modern-day caste IAr groups, because they were not impacted by the migration enough (too little Sintashta) and because they were impacted by some extra ancestry from C Asia that is not now present in caste groups (but seems to be present in some Dardic groups). The question then becomes, if we have some samples that look pretty Dardic (and which are the results of C Asian migration), where are the samples that look like mainline caste IArs in South Asia (which will also be affected by this migration, in fact more affected because IAr have even more Sintashta ancestry compared to Dardics, but they must themselves not look Dardic)?

This is the issue. Notice I can talk about this without even talking about linguistics, just "populations" (Dardic with SPGT contributions, and IArs without), though linguistics will help in answering the question (e.g., if there was a mysterious substrate in Swat, or if Swat was Dardic throughout its entire history). You either have something to contribute or you don't.

Sanuj said...

@RK

Yes, and that is the situation. Now two things follow from here,
The Central Asian contribution to the Dardic groups is a very natural process, because they are right next door to C Asia, there would be obvious movements for trade etc. as the archaeologists have noted. This would also be a natural spread out populations after the formation of Andronovo like groups, who would venture out and mix with surrounding people, ultimately reaching the fringes of South Asia.

Now, what about Steppe ancestry in IAr? Some might have come through this ongoing trickle, and some more in the later phases of known Central Asian groups ruling over large parts of North India for centuries, wouldn't you attribute any contribution to them, they just came, ruled for centuries and vanished? What you want to do is to get one big group of Steppe, moving in as a regiment to account for all the admix in India. This is not a meaningful approach to take.

Apart from that we still need a proper IVC sample to really know, how much of a Steppe mix has really happened. But, even with the situation right now, things are clear. IE started south of Caucuses(where exactly- we'll figure out), and didn't come to India via Steppe.

Seinundzeit said...

Sanuj,

"Dardic is a ethno-linguistic grouping, which would be as much a part of the caste setup as others back in time. Kashmiris are a Dardic people, and we have Kashmiri Pandits, probably the oldest Brahmins we have, who knows."

The Kalasha, Kho, Shina, Kohistani, Pashayi, etc, are unfamiliar with the Indian concept of caste.

That being said, the caste system must have existed in Swat, since the Pashtuns of Swat are unique among people of their ethnicity for having a very "Indian"-like system of social hierarchy, with themselves on top (as landowners), and with the Pashtunized autochtones working their fields, and performing various occupations considered "unworthy" of the "Pukhtun zamindar".

No doubt, this whole system is a cultural holdover from the pre-Pashtun inhabitants.

Proud farmer said...

i never talked about the ivc. No mention of it in fact. You are the one coming off as insecure now. Who was the ivc do you think? You sound very knowledgeable. Kindly educate me.

Sanuj said...

@Sein

Yes, i meant in the historical context, it must have existed in some form in those areas as well, as it did in the rest of the subcontinent.

Seinundzeit said...

Sanuj,

For what it's worth, the caste system never seems to have been known to the Dardic peoples of what are now northeastern Afghanistan and northern Pakistan.

Sure, it must have been a part of Swati culture, but only among the Hindkowan-related Swati, not among the Kohistani (a Dardic people) of the upper reaches.

Jijnasu said...

@sein
I guess the problem is in identifying dardic as a distinct clade within indo-aryan, when it may really be the retention of archaicisms and shared innovations that distinguish dardic. The pre-islamic Kashmiris, Shina and swatis (who shared many cultural/religious traits with adjacent regions of medieval India) were culturally very different from the kalash or the kho who seem to have been isolated from a fairly early period. The strange thing for me though is that the conservative North Western (dardic like) prakrit seems to have been used even east of the Indus. Is it possible that the people of iron age swat shared their language with territories to its south and east, yet were genetically distinct?

ryukendo kendow said...

Sorry, factual errors again, I meant to say that IArs have higher Sintashta ancestry than SPGT, not Dardics. Just in case anyone is confused, Dardics>IAr>SPGT in Sintashta ancestry, but Dardics take SPGT and caste IArs don't.

@ Sanuj

Well, if Sintashta ancestry was increased by Scythians around 0AD we should look at steppe genetics at around 0AD.

This ranged from ~75% West Eurasian and still quite Sintashta-like (e.g. Hunnic Tien Shan) to ~75% East Asian and very unlike any putative source for South Asians (e.g. Pazyryk Scythian).

It is possible some Steppe groups contributed to IAr populations in the Scythian period, but as Sein already alluded to, Scythian and Sarmatian percentages only appear for Dardic and Iranic groups in SC Asia and not for IAr groups, which consistently prefer Sintashta.

Haplogroups like N and specific subclades of Q appear among Scythian-Sarmatian samples, as well as much East Asian mtDNA, which are not represented in upper caste South Asians.

But this is nMonte and uniparentals and a more secure way to rule this out (or not) would be to try qpAdm with Sarmatian/Scythian instead of Sintashta groups, which David could try if he's interested. If all later groups have too much East Asian ancestry to serve as plausible sources, then the hypothesis of 'Scythian upper castes' is not very believable.

ryukendo kendow said...

@ Mzp

The question is, do you actually want to be helped?

This article provides an overview of linguistic trees for IE languages. Try to understand its approach, read Ringe and Atkinson if you can, and tell me if its possible that Anatolian and Celtic and Tocharian and Italic and Germanic and etc... can all originate from Iranic languages.

PIE is not bullshit, flowcharts are not enough, and you'll see why.

Jijnasu said...

@mzp
Neither North Indians nor South Indians can take 'credit' for the IVC. Of course they were our ancestors in part, but we know very little about them. The cultural ancestors of modern Indians are the mainly people of the PGW and BRW as well as the regional cultures of the early historical period.

old europe said...



RK

here's a quote from the article you posted:

Conversely, Germanic, Celtic and Italic were adjacent to each other at the beginning of the Iron Age and yet Germanic is about as far, morphologically, from Celtic and Italic as it’s possible to get.

I was speaking about the stunning difference between these two family languages despite the fact that they arrived with the R1b invasion ( allegedly) of the post 2500 BC scenario. I really do not understand how IC and germanic could diverge so much in slightly more than 1000 years above all bearing in mind that nordic regions ( the proposed urheimat of Germanic) and central europe france and italy were in close cultural contact during the bronze age with polada, unetice, nordic bronze age and so on.

old europe said...



schrijvers wrote about the fact of protogermanic being the product of balto finnic switching to centum indoeuropean but for this to happen we need a clear majority of balto finnic people around northern europe and west baltic regions. There's no trace of massive invasion of uralic in these regions.... there's only corded ware..........

Chetan said...

Regarding toponyms (since a lot of people appear to making arguments on this basis)

While the presence of toponyms belonging to a particular language is evidence that its speakers were present there at one time, the reverse is not true.

You can't take a place and point to the lack of toponyms there to establish that a language was not spoken there at some time in the past.

In the case of Indo-Aryans, they started out as one among many groups of Andronovo pastoralists in the trans-Urals region. They then migrated through Siberia and Central Asia in a couple of centuries. They hardly had any lasting presence that would justify IA toponyms being found in these areas, which were occupied later by Scythians, Turks, Mongols and Russians. Many places in Kazakhstan bear Russian names inspite of the very recent Russian hegemony over the region.

Toponyms, at least on their own, are not a reliable source to look for evidence.

old europe said...


chetan

"While the presence of toponyms belonging to a particular language is evidence that its speakers were present there at one time, the reverse is not true".

If you are referring to me well if you read my posts I'm saying the same thing.

"They hardly had any lasting presence that would justify IA toponyms being found in these areas"

We DO find IA and Iranic toponyms all along the eurasian steppe from the Don to the Ural and beyond.

The toponyms per se are not the last word obviously they must match with other criteria.
But as you pointed out a IE toponym is much more reliable than a pot or other archeological findings ( even my much beloved cremation) to know if IE people were around in particular places.

Chetan said...

@old europe Yes, agree with you. Can you provide the reference for specifically IA toponyms in these areas? I know there are Iranian toponyms all over, but are there specifically IA ones?

IA is close to reconstructed to Proto Indo-Iranian in many ways. So there have been debates on the nature of early loanwords into Uralic for example - were they specifically Indo-Aryan or from an undifferentiated Indo-Iranian language.

There can be no doubt however that IIr (maybe even IA itself) gave many loanwords to Proto-Uralic.

Jijnasu said...

@chetan
There was a case for specifically IA toponymya in the crimean peninsula. Apparently the indus (sindhu) and kabul (kubha) shared their names with rivers there. I think the original paper is only available in russian though

Alberto said...

@RK

The statistical significance of that scattered plot is what I'm questioning, asking for more clear data (for example from the Figure S4.1 and S4.2 plots).

Those ones that you are posting are a bit more cryptic. But also look at those 5 populations that they highlight:

Brahmin Nepal Z=3
Bhumihar Bihar Z=3.5
Brahmin Tiwari z=4.3
Brahmin_UP Z=4.1
Bhumihar_UP Z=2.6

Looking at the models, it seems that these 5 groups are autosomally almost identical and seem to come from one source population. The Brahmins from Nepal, Bihar and Tiwari are probably not locals. We could be looking at 5 populations that are essentially the same one, only recently split and migrated to other areas. Or maybe not. But do you know the history of those 5 groups highlighted? More importantly, does the person who wrote the text on the pre-print has a deep knowledge of the history of these groups or not even Wikipedia level knowledge of it?

So you see, the statistical significance depends on the correctness of test and the knowledge about who they are testing. And whether that text accompanying it is an accurate account belonging to a scientific paper or a sensationalist headline belonging somewhere else, in line with other parts of the text, is what you should be looking at.

It's good that you apply critical thinking to the opinions of people here, including mine. But it's much more important that you apply it to the data in the pre-print (and to the data in general). You used to do that some years ago.

old europe said...



chetan and Jijnasu

Here's a more complete quote:

The geographic zone of the ancient seat of the Indo-Iranians is indicated by the absence in the Indo-Iranian languages of the common Indo-European words for ‘spruce’ and ‘bog’. This makes it possible to localize them in the steppe (Schrader 1913). The suggested identification of the legendary river Ra (Sanskrit Rasa, Iranian Raŋha) with the Volga and the Ripa mountains with the Urals (Marquart 1938; Grantovsky and Bongard-Levin 1970; 1998; Grantovsky 1976 supported by A. I. Dovatur, D. P. Kallistova et al. (1982: 248-249) and N. L. Chlenova (1983: 56-60;1989)) is extremely important for locating the Indo- Iranian homeland. V. Miller (1887) recognized the Iranian etymology of a number of geographical names in the North Pontic, later augmented by V. I. Abaev (1949). A large number of Iranian hydronyms with the root don (Iranian ‘water’) has been revealed for the drainages of the Dnieper, Desna, Seim, Severny Donets and the Poltava region (Toporov and Trubachev 1962; Strizhak 1965). The Indo-Iranian treatment of toponyms in the North Pontic has also been supported in a series of works by O. N. Trubachev (1975; 1976; 1999) and L. A. Lelekov (1980) while E. A. Grantovsky and D. S. Raevsky (1984: 47-62) agreed that the Iranian element is “the only undisputed (element) in the North Pontic.” S. S. Berezanskaya (1982: 206-209) in charting the hydronyms of the Upper and Middle Dnieper area on an archaeological map showed that the area of the Iranian names does not coincide with the distribution of the Scythian culture but correlates completely with the territory of the Timber-grave culture, which is a strong argument in favor of assigning a Proto-Iranian identity to the Timber- grave people.
Indo-Iranian toponyms are also found on the Middle Volga and the Urals (Popov and Loyfman 1962). N. L. Chlenova (1983a; 1984) plotted the toponyms and demonstrated that Iranian toponyms were spread over Timber-grave and Andronovo territories, but part of the Andronovo toponyms can only be interpreted as Indo-Aryan; here also was included the Altai by A. M. Maloletko (1986: 70-75). These data are extremely important for the final resolution of the problem of the cultural and ethnic attribution of the Fedorovo complex. Due to the fact that Indo-Iranian toponyms of the pre-Scythian period have been found on the territory populated only by Fedorovo tribes, the hypothesis identifying the Fedorovo population as Ugrian that has been proposed by V. N. Chernetsov must be rejected, and the hypothesis of the Indo-Aryan attribution of Andronovans can be supported.

mzp1 said...

Old_Europe,

I don't agree with that except.

They seem to be using the word Indo Iranian where they should say Iranian. There are no clear examples which is all I was asking for.

Also. Timber Grave Culture is related to one of Dacians or Cimmerians. I identify Cimmerians with Sarmatians that would make timber grave Iranian.

mzp1 said...

@old Europe

"area of the Iranian names does not coincide with the distribution of the Scythian culture but correlates completely with the territory of the Timber-grave culture, which is a strong argument in favor of assigning a Proto-Iranian identity to the Timber- grave people."

Timber Grave does not get us to the home of Iranians in South Central Asia.

Timber Grave is identified with Cimmerians who moved Westward.

So that language should not be classed as "proto-iranian" but Iranian. And these Iranians would have moved West and their language became European. This is precisely how Balto-Slavic, Celtic and Germanic would have formed.

Ric Hern said...

@ old europe

We see the differences also in Goidelic because of its relative isolation compared to Brittonic and Gaulish. Scandinavia was also relatively isolated and when we look at the Y-DNA found there we can identify at least three different contributing populations of equal proportions contributing to the formation of Germanic.

This is not the same case for Celtic and Italic.

old europe said...

mzp1

we should read the books quoted in the post to determine the method they used to investigate the toponyms. I remeber one of the most famous: DANUBE-DNIEPER-DNIESTER-DON ecc.
Also you should point out the archeological trail at the base of this westward movement of the iranian language together with a time reference.

Ric
Scandinavia surely had a life of its own but ( even today it is quite a world apart from mainland europe)but in the timing of birth of protogermanic it was deep involved in exchanges of idea with central and even east mediterranean regions. We can find an example in the rock art of Boushlan in many ways so similar to that of Valcamonica ( North Italy) a deep relations involving also religious ideas.

mzp1 said...

@Old Europe,

Yes, I know it's just an idea I have, I need to take a closer look at the those early Steppe cultures and their languages along with their timeframes.

I just need finish of some things before I can dive into it and get you more information.

Gill said...

David, have you run any tests to see whether modern Baloch/Brahui have more affinity to BMAC or perhaps a population like Sarazm_EN?

direita said...

No, the rigveda tells us another story:

Aryan men carrying a superior culture dominated and enslaved natives. the various degrees of miscegenation created the castes: the higher the percentage of Aryan blood = the higher the caste.
4,000 years of miscegenation has extinguished the Aryans, but not their traits that still resist
https://farm5.staticflickr.com/4331/37318996996_dd7dfcb39b_o.jpg

«Oldest ‹Older   201 – 332 of 332   Newer› Newest»