Friday, August 26, 2022

The Southern Arc paper and it's data upends the Steppe Theory in multiple ways


Hasanlu Indra chariot



Lazaridis, Iosif, et al. “The Genetic History of the Southern Arc: A Bridge between West Asia and Europe.” Science, vol. 377, no. 6609, 2022, https://doi.org/10.1126/science.abm4247.


This is a recently published paper ('the paper', 'this paper') with many new samples from what the authors call 'The Southern Arc'. The authors define it as 'a region centred on the large Anatolian peninsula (Turkey), including in the west (in Europe) the Balkans and the Aegean, and in the south and east, Cyprus, Mesopotamia, the Levant, Armenia, Azerbaijan, and Iran.'

The main conclusion of this paper which upends one part of the Steppe theory is that West Asia is now considered to be the homeland of the first Indo-European speakers, and the steppe to be only a secondary homeland. From the paper:

However, the link connecting the Proto-Indo-European–speaking Yamnaya with the speakers of Anatolian languages was in the highlands of West Asia, the ancestral region shared by both.
This is because the Anatolian region (Anatolian languages split from Proto Indo-European first as per consensus) has negligible ancestry from the steppe, but rather shows a significant ancestry from its east. I had already proposed an Armenian origin for Anatolian languages in my article published a few months ago, so I agree with this conclusion for now.

However, the paper also notes:

Contributions of Indo-European to Uralic languages (spoken in the forest zone of Eastern Europe and Siberia) appear to have involved only Indo-Iranian speakers ~4200 years ago . This is important because it constrains the migratory history of Proto-Indo-Iranian, consistent with genetic evidence that it spread through the steppe to South Asia and ruling out the possibility that it spread from West Asia to South Asia over the Iranian plateau.
This is a very old argument in the linguistic community and it is not a settled issue by any long shot. Edwin Bryant, a noted Indologist who has taught at Harvard U and now teaches at Rutgers University, in his book (12) dispassionately notes all the viewpoints regarding this and other arguments surrounding the Indo-Iranian debate. For example, Bryant cites Misra (1992) who notes that the borrowing only occurs one way from Indo-Iranian into Uralic/ Finno-Ugric. Why should it be so if Indo-Iranians and Finno-Ugric interacted on the steppe before Indo-Iranians came south to India? Misra asks. The borrowing should happen both ways if proto FU and Proto I-Ir were nearby, according to Misra. Misra argues that the scenario which makes the most sense is Indo-Iranians leaving to the north and giving loanwords to Finno-Ugric so that I-Ir could give loans to FU unidirectionally. Edwin Bryant notes that this argument is 'Ingenious'. To date, to my knowledge, there are zero widely accepted Finno-Ugric borrowings in Old Indic or Old Iranian. 

Bryant adds
While most linguists seem to agree that the loans are Indo-Iranian and not Proto-Indo-European, there is disagreement over whether they are specifically Indo-Iranian, Iranian, or Indo-Aryan. Misra (1992), in addition to reversing the direction of language flow, is of the opinion that most of the words can be accounted for as Old Indo-Aryan forms and not Iranian. Shevoroskin also considers them to be Indo-Aryan (and even Middle Indo-Aryan). Most recently, Lubotsky (forthcoming) concurs that the oldest layer of borrowings are often of Sankrit and not of Iranian. D'iakonov (1985) and Dolgopolsky (1993) consider them Indo Iranian. Gamkrelidze and Ivanov (1983b) in contrast, are quite specific that they should be interpreted as early Iranian and not as Indo-Iranian, “or even less as Old Indic”. Joki (1973, 364-365) also considers them to be mostly Iranian or Middle Iranian. As Mallory (1997) quips: “Will the 'real linguist please stand up? It should be obvious that linguists have as much difficulty in establishing the chronological relationships between loanwords as any other 'historical science".
If steppe_Mlba was Proto-Indo-Iranian, how can it give Sanskrit/Early Iranian/ middle Iranian/ Middle Indo-Aryan, etc loanwords to Finno-Ugric? 

Bryant finally concludes
Ultimately, there is little in the history of loanwords that can eliminate a variety of historical possibilities.


There was increased contact from SC Asia into the Eurasian steppes starting 3rd millennium BCE. Spengler et al 2014 (26) state:

The presence of free-threshing wheat (Triticum aestivum/turgidum) and broomcorn millet (Panicum miliaceum) at Central Eurasian highland campsites also illustrates the first transmission of domesticated grains from both southern Central Asia (south of the Syr Darya River) and China into Central Eurasia during the third millennium BC.

These contacts are viable sources of loanwords from the south into Uralic, and they continued well into the 2nd Millenium BCE.

Therefore, I strongly disagree with the authors of the 'Southern Arc' paper in that they are extremely narrow in their line of thinking in constraining Indo Iranians from the steppe, based on ambiguous evidence. The only constraining factor in locating the Indo-Iranians should be the Mitanni evidence (discussed later). Post-1600 BCE various loanwords, Gods & proper names of Indo-Iranian origin appear in the Mitanni context in Syria.

This narrow perspective is very visible in their paper when they choose to not even entertain the possibility of earlier CHG/IranN ancestry common between West Asia and the Indian subcontinent as the carriers of IE language to the region. After all, if CHG/IranN is the carrier of IE languages into Anatolia and the steppe, what language did it bring into the Indian subcontinent? IranN ancestry is the largest source of ancestry in IVC people (13)(5). What language did this largest source of ancestry bring to the region? Do the authors have an answer for that? In comparison, the steppe ancestry in Swat valley was minuscule at an average of 15%.


Johannes Krause of Max Planck Institute of Evolutionary Anthropology, and his linguist colleagues at the University of Jena are also open to the possibility of IE spreading to the Indian subcontinent via the Iranian Plateau (17). Paul Heggarty's A2 Hypothesis (20) also posits the homeland of Indo-Iranians to be south of the Caspian, and not on the steppe. Here, it is important to note that the rejection of Heggarty's trans-Iranian-plateau route for Indo-Iranian by Shinde et al 2019 (21), has been overturned by Maier et al 2022 (preprint)(22) as well as my own recent work due to presence of minor but unmistakeable Anatolian /Levantine ancestry in the Indus Periphery samples published in Narasimhan et al 2019 (5).



The second disagreement I have with the authors is their refusal to test for IVC ancestry in the new Southern Arc samples. From 'Southern Arc' paper supplement page 267, emphasis is mine:
Bronze Age samples from Shahr-I Sokhta(34) from eastern Iran at the edge of the region we are considering also derive almost all their ancestry from a CHG-related source. These individuals were modeled as having South Asian ancestry, which is not represented in the 5- source model used to model samples from the Southern Arc in our paper.(34) Samples from Tepe Hissar in northeastern Iran ((34) and this study) also have predominantly CHG ancestry but with 7±2% Levantine Neolithic ancestry; the latter could be derived from the CHG/Levant cline of which the Mesopotamian Neolithic was a part.


So the authors did not even try to use an IVC source in their modelling. They didn't even test it. I am extremely displeased by this. Especially when the contacts between the Indian subcontinent and the near east are so well known and attested in archaeology. Many archaeologists have noted the presence of Mitanni Aryan elements at Hasanlu. Very disappointing!!

In the next section, I will analyze some of the new samples. The section will get technical, so lay readers could skip it and go to the next sections.


ANALYZING SOME OF THE NEW DATA


In this article, I will analyze the samples from the NW Iranian site of Hasanlu Tepe and Dinkha Tepe, and discuss the results and implications of the analysis.

I will be analyzing these samples dispassionately, using all relevant sources from surrounding regions that I can in a rotating qpAdm model as described by Harney et al 2021(1), with slight modifications. The advantage provided by this strategy is that it allows us to directly compare competing models against each other. For example, for a target label (whose ancestry we want to find) if we have 20 sources we want to test, then for a 2 source model we will test 20C2 = 190 models one by one. The other 18 populations will go into the reference 'right pops' of qpAdm. Only models which cross the threshold p-value of 0.05 AND with non-negative admixture coefficients will be accepted, except if I can reject them due to some other specified reason. If no 2-source models work, I move on to 3-source models and so on.

The modification I make will be that of the 20 populations (as an example), I will be removing the Outgroup (African label) and a few other (ex 6 labels) Neolithic labels from contention as sources while keeping them as reference populations. So I will only test 20-1-6 = 13 Labels as sources while utilizing the other 7 purely as references. The number of models tested then will be 13C2 in this example. This helps me cut down on computing time and I don't even want to utilize distal labels as sources (eg 8000BCE Ganj_Dareh, 9000BCE PPN, etc.) and only want to check proximal sources which are closer in time to the targets. To this end, I have developed a Python script that helps me run all the models sequentially and record the output in a CSV file systematically.

Unless otherwise specified, this is the list of all the labels I will use for the models.

RotatingPoplist: Mbuti.DG, ONG.SG, PPN, EHG, CHG, Turkey_N, Azerbaijan_Caucasus_lowlands_LN, Iran_HajjiFiruz_N, Iran_GanjDareh_N. Serbia_IronGates_Mesolithic, Tarim_EMBA1, Iran_BA_HajjiFiruz, Russia_Samara_EBA_Yamnaya, 
Russia_MLBA_Sintashta, Uzbekistan_Bustan_BA, Russia_Steppe_Catacomb, IVCp, Maikop_Novosvobodnaya, IRQ_PPN, ARM_Aknashen_N, ARM_Masis_Blur_N, Tajikistan_C_Sarazm, Russia_Steppe_Maikop, Russia_Caucasus_Eneolithic, Russia_Steppe_Eneolithic, Iran_C_SehGabi, Israel_C, Hasanlu_MBA, Hasanlu_LBA_B, Hasanlu_LBA_A, DinkhaTepe_A

Of this list, these are the labels I will exclude from sources:

ONG.SG, PPN, EHG, Turkey_N, Iran_GanjDareh_N, CHG, Tarim_EMBA1, Serbia_IronGates_Mesolithic

Some other labels will be removed completely if they are younger in time than the target to be tested. Eg. Hasanlu_LBA_B, Hasanlu_LBA_A, and DinkhaTepe_A will be removed while testing Hasanlu_MBA.

I will explain the labels. The usual nomenclature is Country_SiteName_Timeperiod. N = Neolithic, C = Chalcolithic, BA = Bronze Age, MBA = Middle Bronze age, LBA = Late Bronze age, IA = Iron age.


Working Rotating models for Hasanlu_MBA


Location: Hasanlu Tepe
Date: 2100BCE. 
Y hg J1a2b1b2~

These are the only acceptable 2-source models.

Hasanlu_MBA 2 source
*S%1 means Admixture coefficient of Source 1, Src1SE means 1 Standard Error for Source1 admixture coefficient. S%1 of 0.125 and Src1SE of 0.057 can be read as Iran_BA_HajjiFiruz having a coefficient of 12.5% +- 5.7%


Given the presence of R1b at Hasanlu, and the presence of Yamnaya (but not Sintashta) ancestry at the nearby site (within 100km) of Hajji_Firuz in 2400BCE, I prefer the Aknashen_N + Hajji_Firuz_BA/ Yamnaya_Samara models for Hasanlu_MBA. Yamnaya's presence in chalcolithic Armenia is also noted in the Southern Arc paper.


Working Rotating Models for Hasanlu_LBA_B


Location: Hasanlu Tepe
Date: ~1400BCE. 
Female: mtDNA W6. 

Can be modelled simply as Hasanlu_MBA with a p-value of 0.73.



Working Rotating Models for Hasanlu_LBA_A


Location: Hasanlu Tepe
Date: ~1350BCE. 
Male. Y hg J1a2a1b1a~. mtDNA R2.

Hasanlu_LBA 2 source

6/172 models pass. Looks like a Sarazm-type population had a big genetic impact (55-75%) on this sample. Likely a 2nd gen migrant.



Working Rotating Models for Hasanlu_IA


Location: Hasanlu Tepe
Date: 18 samples dated from 1300BCE to 700BCE. 
Out of 13 males, 10 are R1b-M269+ (of the Yamnaya kind). 1 is J2a2a1a1a2, 1 is J2a1a1b1a1a and 1 L1a (specifically L-Y28524). L1a is ISOGG notation and L-Y28524 is YFull notation which is more precise.

This L1a subclade is from the Indian subcontinent, and the immediate parent L-Y6288 (Yfull) is found in the Swat valley Iron age (I6555, 900BCE). The only other sample on YFull in L-Y28524 is a Konkani Indian.

Hasanlu_IA rotating
DinkhaA and DinkhaB are excluded from the source and reference.

Out of 210 models, only the above ones pass. The Sintashta model can be rejected because the base population Aknashen_N is 5000 years older than Hasanlu_IA and is unlikely to have existed in the pure form given that we know the composition of Hasanlu_MBA and both Hasanlu_LBA samples.

I prefer the Hasanlu_LBA_A + Hasanlu_LBA_B model as it requires no external ancestry.

Given the presence of  L1a, and the non-marginal p-value of the model with IVCp as the source, I decided to check further, but found that for the whole label of 18 samples, Hasanlu_LBA_A + Hasanlu_LBA_B still fits the best.

However, using G25 there seem to be a few samples that show elevated South Asian Ancestry. The highest is seen in I6430. So next, I analyze that sample alone to prove the existence of South Asian ancestry. 

I added sources like Udegram_IA and Aligrama_IA from Swat valley. Udegram has steppe admixture but Aligrama does not and likely represents the pre-steppe ancestry of Swat valley. Also added another BMAC label - SappaliTepe and another steppe label Srubnaya.


RotatingPoplist: Mbuti.DG, ONG.SG, PPN, EHG, CHG, Turkey_N, Iran_GanjDareh_N, Serbia_IronGates_Mesolithic, Tarim_EMBA1, Russia_Samara_EBA_Yamnaya, Russia_MLBA_Sintashta, Uzbekistan_Bustan_BA, Russia_Steppe_Catacomb, IVCp, IRQ_PPN, ARM_Masis_Blur_N, Iran_C_SehGabi, Israel_C, IRN_Hasanlu_MBA, Russia_Srubnaya, Pakistan_Udegram_IA, Pakistan_Aligrama_IA, IRN_DinkhaTepe_BIA_AUzbekistan_SappaliTepe_BA


Out of 364 models, 4 work. We can see that the steppe source in all these models shows a ~0% coefficient. Essentially, the following 2 source models work for I6430.

1. Hasanlu_MBA + Aligrama_IA
2. Hasanlu_MBA + IVCp

The IVC/Swat coefficient is 20-22%. Important to note here that the Udegram_IA has steppe ancestry but is rejected in favour of IVCp and Aligrama_IA which don't have steppe ancestry.

With this, in addition to the presence of South Asian Y haplogroup at Hasanlu, we have proof of IVC autosomal ancestry at Hasanlu.



Archaeological connections between Swat and Hasanlu are also seen. The below post by Nezih provides some excerpts.


Working Rotating Models for DinkhaTepe_A


Location: Dinkha Tepe, 25km from Hasanlu
Date: 6 samples, of which 4 are carbon dated and calibrated between 2000-1700BCE and 1 is from around 1000BCE. The last one is not carbon-dated. 
The 2 males have Y hg J-FGC16079 and J-ZS1711.

Out of 171 2-source models, only 4 pass the p-value threshold. 
DinkhaA models
The ancestry for Dinkha_A is either local, or there is minor BMAC/IVCp ancestry (up to 13%).


Working Rotating Models for DinkhaTepe_B


Location: Dinkha Tepe
Date: 8 samples, of which 1 is dated to ~1800BCE, 5 are dated to between 1400-1000BCE, 1 is dated 1000-850BCE, and 1 is only archaeologically dated. 
Out of 3 males, 2 belong to Y hg G1a and 1 to J2a1a1a2a.

None of the 2 or 3-Source rotating models passed, so below I present the results of the 4-Source models. 

Only 4/220 models pass, and all 4 contain Azerbaijan_LN (46-51%) + BMAC(29-35%) + IVCp (9-11%) + a steppe source (~10%). All 4 steppe sources work - Sintashta, Yamnaya, Catacomb as well as Hajji_BA. 

Dinkha Tepe B models

RotatingPoplist: Mbuti.DG, ONG.SG, PPN, EHG, CHG, Turkey_N, Azerbaijan_Caucasus_lowlands_LN, Iran_HajjiFiruz_N, Iran_GanjDareh_N, Serbia_IronGates_Mesolithic, Tarim_EMBA1, Iran_BA_HajjiFiruz, Russia_Samara_EBA_Yamnaya, Russia_MLBA_Sintashta, Uzbekistan_Bustan_BA, Russia_Steppe_Catacomb, IVCp, Maikop_Novosvobodnaya


I worked on Dinkha_B further, reducing the available pops in the rotating model which do not work and adding Swat_IA and Swat_Aligrama_IA (no steppe_mlba in this label ) to the models. Importantly, I added Dinkha_A to the list to test for Dinkha_B = DinkhaA + X + Y. Because from the PCA and previous results it is clear that Dinkha_A is a native population very similar to 5000BCE Masis_Blur_Neolithic population with minor additional BMAC/IVCp like inflow.

Dinkha B alternate

RotatingPoplist: Mbuti.DG, ONG.SG, Turkey_N, Iran_GanjDareh_N, CHG, EHG, Serbia_IronGates_Mesolithic, Tarim_EMBA1, Russia_Samara_EBA_Yamnaya, Russia_MLBA_Sintashta, Uzbekistan_Bustan_BA, IVCp, IRQ_PPN, ARM_Masis_Blur_N, IRN_DinkhaTepe_BIA_A, IRN_DinkhaTepe_BIA_B, Pakistan_Aligrama_IA, Pakistan_Udegram_IA

Among 105 models, the only feasible model (p-value: 0.52) for Dinkha_B is 

Dinkha_A + BMAC + Yamnaya. 

Note that Swat sources are rejected, which means there wasn't any further inflow after DinkhaA.
Note that Sintashta is part of the poplist, and as no model with Sintashta_MLBA works, Steppe_MLBA as a source can be rejected.  Indeed, DinkhaA + Sintashta + BMAC gets a failing p-value of only 0.009. Therefore, a mixed BMAC + SteppeMLBA population like TKM_IA (850BCE, Turkmenistan) can be rejected as the source. This sample was posited by Davidski of Eurogenes Blog to be an early Iranian due to steppe admixed BMAC ancestry.

G25 runs with the above sources give great fits, with avg admixture coefficients matching exactly.


G25 dinkha B



Working Rotating Models for IRQ_Nemrik9_LBA


Location: Nemrik9, Iraq. 300km west of Hasanlu. (middle Assyrian period)
Date: No carbon dating, archaeological date 1500-1200BCE
Male, Y hg J-Y6094

This sample can be modelled as IRN_DinkhaTepe_BIA_A with a p-value of 0.19.

So, the BMAC ancestry at DinkhaTepe in the B cluster did not extend further west till Assyria.



DISCUSSION

Let me recap the results obtained so far.

1. First, we see Yamnaya-like inputs of around 8% in Hasanlu MBA dated to 2100BCE. This is in concurrence with the findings of Lazaridis et al and also with earlier findings of Yamnaya ancestry at Hajji_Firuz_BA dated to 2400BCE. 

2. In the LBA period, we see a huge amount of Sarazm-like ancestry (55-75% depending on model) around 1350BCE at Hasanlu. The other sample is simply like Hasanlu_MBA, indicating that the external admixture probably did not happen too long before the sample date.

3. Hasanlu_IA (1300-700BCE) can be modeled as 

     2-source: Hasanlu_LBA_B + Hasanlu_LBA_A

The presence of L-Y28524 at Hasanlu strongly indicates an IVC/Swat connection. Sample I6430 has 20% ancestry from IVC.

4. DinkhaTepe is 25km from the Hasanlu site. 
DinkhaTepe_A from 2000-1700BCE can be modelled as a 2-source model of local Arm_Masis_Blur_N + IVCp/BMAC/SehGabi. It can also be modelled as 2-source: Masis_N + Hajji_Firuz_N, both local much older populations.

5. DinkhaTepeB mostly has samples from 1400-1000BCE, which means this cluster is later in time than DinkaA. DinkhaB sees a large inflow of BMAC-like ancestry which can also be seen in the oldest sample dated to around 1800BCE (I3912, check Vahaduo output above. I3912 has very little Yamnaya ancestry, maybe it entered the site later? Or is the sample misdated?).

Putting 1-5 together, we see a post 2000BCE entry of Sarazm/BMAC or IVCp-like ancestry entering NW Iranian sites of Hasanlu and DinkhaTepe. The signal is most clear in Hasanlu_LBA_A and DinkhaTepe_B with 35-75% external ancestry.

If the dating of sample I3912 is correct, then this entry happened between 2000BCE and 1800BCE. Instead, if the dating is wrong, then that would mean the entry of this ancestry between 1700BCE and 1400BCE.

WHAT DOES THIS EXTERNAL ANCESTRY MEAN?


Viktor Sarianidi, the excavator of BMAC considers BMAC to be an Iranian culture (2). BMAC has many sites with temples, fire worship, horse burials, and horse-headed sceptres (10)(11)  - indicating an IE culture. Importantly, steppe ancestry in main BMAC populations is nil, only a couple of outliers with some steppe ancestry have been found.

Mallory (in reply to Lamber-Karlovsky) (32) notes the following with respect to Indo-Iranian origins:

Andronovo and the Bactrian Margiana complex really do appear to be the only game(s) in town, and Lamberg-Karlovsky has indicated why they appear to be mutually exclusive solutions, neither capable of resolving the problem by itself. There is no single culture (using the word in the widest—even most unjustified—sense) that can link the Indus, Iran, Central Asia, and the steppe lands together.
Elsewhere, Parpola (31), citing Ghirshman (1977) notes:

Ghirshman (1977) had already advanced numerous arguments in favour of an Aryan identification of the Gurgan Grey Ware complex; these arguments must now be understood as covering the BMAC as well. Thus, Ghirshman traced the horsemanship, for which the Mitannians were so famous, back to the Gurgan plain, where a locally made cylinder seal from Tepe Hissar IIIB bears the representation of a horse-drawn chariot. Miniature models of trumpets, made of gold and silver, have been found in the Gurgan area (Tepe Hissar IIIC; Treasure of Asterabad) as well as in Bactria. As Ghirshman pointed out, trumpets were needed for directing chariots in battle, and an Egyptian bas-relief of Ramses ΙΠ (early 12th century B.C., soon after intensive Egyptian contacts with Mitanni) shows that trumpets were used in training horses. Horse bones have been found in Gurgan (Shah Tepe), but not, so far, in Bactria or Margiana.* The horse is represented, however, on several of the ceremonial weapons said to have come from the looted graves of Bactria; here one bronze statuette even shows a horse with a naked ityphallic rider, who is pressing his bent legs backwards beneath the horse, without stirrups (Bothmer 1990: 43: no. 29). The abundance of weapons itself suggests that the ruling elite of the BMAC was actively engaged in warfare.

* Dubova (33) asserts the presence of a domesticated horse at Gonur necropolis

Some words need to be said about the discovery of horse remains. It has become clear now that we can talk only about one whole horse burial at Gonur: this was a young animal, leaning against the east wall of the courtyard of royal tomb 3200 and fixed against the side of a four-wheel cart.

It can be said with confidence that there was a  domestic horse in Margiana.

Hiebert (31) suggests that the movement of people from BMAC introduced Indo-Iranian to the Iranian plateau.

The identification of the intrusive BMAC burials has suggested a movement of population from Central Asia into an area óf towns and cites on the Iranian plateau (Hiebert and Lamberg-Karlovsky 1992). The lack of imported objects, in addition, into the oases of Bactria and Margiana at this time strongly indicates the unidirectionally of this movement. However, the presence of these burials does not indicate massive migration, since these tombs and cenotaphs outside Central Asia are intermittently dispersed and continuity of occupation in the oases of Central Asia is confirmed. As previously suggested (Hiebert and Lamberg-Karlovsky 1992) and discussed in this volume (by Erdosy in Chapter 1), the appearance of the Central Asian peoples on the Iranian plateau for this brief period at the beginning of the second millennium B.C. is the best candidate for an archaeological correlate of the introduction of Indo-Iranian speakers to Iran and South Asia.

 

THE HASANLU GOLD BOWL


How and where was the bowl found? Very interesting context from Iranica Online
The ‘gold bowl of Ḥasanlu’ was found in the debris of Burned Building I West on the Citadel Mound at Ḥasanlu in 1958. It had fallen into room 9 in the southeastern corner of the building where, at the end of the 9th century B.C.E., it was buried under the collapsed mud brick walls of the second story along with the bodies of three men. One of these men was carrying the bowl. Although badly crushed, it could be determined that the man had fallen on his right shoulder, with his head facing the north wall. His lower right arm lay upright against the wall face, his hand bent at the wrist. This arm, from wrist to shoulder, was protected by a gauntlet made perhaps of leather (represented by a thin, half-centimeter thick film of powdery yellow and plum-red material). It was backed by six or seven rows of round, flat-topped bronze buttons attached by loops at the back. Two larger flanged buttons formed attachments at either end. The left arm was unadorned, the upper part lying in front of, and parallel to, his chest, with the lower part lying against the wall. Between his hands, lying at forty-five degree angle, was the flattened gold bowl. The bowl was being carried in an upright position and contained three other valued artifacts: a mottled red and white uncut stone cylinder with gold caps, a shattered figurine (perhaps a bird) of laminated ivory (?), and a sword-hilt with a U-shaped bronze guard set over the shoulder of a broken iron blade with round tang. The grip was composed of three red sandstone and three white stone discs fitted over the tang. The pommel was missing. This object is of considerable interest as it matches one found in the former Russian Armenia at Mouci-yeri, dating to between 1200 and 900 B.C.E. (de Morgan, p. 91, figs. 34, 35; Schaeffer, p. 501, nn. 6 and 7).

Images from the Penn Museum site 

Hasanlu bowl part 1
Detail, upper zone of the Hasanlu gold bowl: Storm god in chariot and priest. Only the doubled horns of the front bull indicate that there are indeed two overlapping animals. (Photo courtesy of the Hasanlu Project)



Hasanlu bowl part 2






 Detail, middle zone of the gold bowl: Goddess on a lion. (Courtesy: Penn Museum, Hasanlu Project)


Hasanlu gold bowl 3
Detail, lower zone of the gold bowl: combat scene. 3 headed animal was also seen (Photo courtesy of the Hasanlu Project)


A detailed description of the scenes on the bowl can be found at Iranica Online, do give it a read.

So we see a Goddess who rides a lion (just like Goddess Durga), who may be the Iranian Goddess anAhita. From wiki,
Aredvi Sura Anahita is principally addressed in Yasht 5 (Yasna 65), also known as the Aban Yasht, a hymn to the waters in Avestan and one of the longer and better preserved of the devotional hymns.

GN Kurochkin(3) also considers the imagery on the bowl as Avestan

The illustrations on this Hasanlu bowl conform to integral passages of the Avesta, found in the 21st chapter of Vendidat and the 5th last devoted to Anâhita. 

For example, a hero dressed in  a peculiar  skirt  is shown  on  the bowl  twice: once associated with a bird of prey and once with a three-headed monster. In the Avestan hymn, the hero Thraétaona is also mentioned twice: for the first time in connection with his battle with the three-headed Azi-Dahäka (Zeit 5.33—34) and a second time when he helps the  hero Päurva  by  making him fly  up  to heaven  in the form of  a bird  of  prey  ( fest 5.61).  After  three days and  nights,  Piurva  asks the goddess Anahita to return him to the earth, which she does in the shape of a beautiful girl ( Ya:st 5.62—65). Päurva then performs a sacrifice with the sacred Haoma drink ( Yokt 5.66) (the Hasanlu  bowl  shows a man with a vessel) (Kuroch- kin 1974).

Some of the scenes on the Hasanlu bowl reappear later in Sasanian toreutics.

The  Gandhara Grave Culture of Swat in Pakistan, dating to the second half  of the second millennium BC, has been identified by some researchers with the Vedic Aryans. It is significant that also  the  material  culture of  Swat  offers many  parallels to the finds of Tepe Hissar III, Marlik and Hasanlu V.


HP Francfort notes the following in his 2008 paper (4). I apologize for quoting large parts of his paper, but I thought they were too good to paraphrase.

A part of the iconography seems to be related to the Oxus Civilization (labeled also BMAC: Bactro-Margiana Archaeological Complex). The focus or core of the Hasanlu bowl imagery is none other than the basic ancient water cycle, as observed by various students. Water flows down from the muzzles of the bulls of a chariot in the upper register to a composite being in the shape of a mountain monster (“dragon“) that is fought by a hero in a kilt while, as fertility is expressed by various images such as rams and female deities. Links seem clear with an ancient Middle/Inner Asian mythological/naturalistic cycle of water represented earlier, in the first half of the third millennium, on the carved dark stone vases from Jiroft (Kerman, Iran).

On the Hasanlu bowl we then see a hero vanquishing a lion (dragon) AND a goddess taming a lion: this a fundamental core scheme present in the Oxus Civilization iconography. Incidentally, we may observe that the two rams behind the dragon are by means of graphic contact walking on the recumbent lion’s tail: this mean that the naked goddess standing on the rams dominates the lion (an old Oxus Civilization scheme).

Hero in a kilt, as we have seen, fighting against the dragon (hero ON the lion): as observed by various students of the Hasanlu bowl, it is rep-resented twice, once as a fighter with his hands protected by shields and a second time as victorious, larger in size and wearing a diadem (with snake? as trophy?) and bow, quiver and arrows. He wears a kilt ornamented with what apparently are animal tails. This is the costume of the Oxus Civilization warriors-hunters  (figs.11, 12) and of the Oxus Civilization hero (eagle anthropomorphic) fighting a dragon or snakes.  We can notice in passing the  graphic contact of the walking hero with the naked goddess’ hand behind and with a ram’s leg above: in the Oxus Civilization the eagle-hero acts for the goddess and he is also the provider of caprids (see below).

The upper register of the bowl, in the present hypothesis, could represent the Mitanni-Arya deities, apparently in the order of some of the texts: Indra, Mitra, Varuna and the Nāsatya. Indra is the warrior on a chariot with a headdress in the shape of horns. Mitra is the anthropomorphic being with the headdress in a winged (solar) disk. Indra and Mitra drive horse- or equid-drawn chariots. The winged being, the first deity on the right, is Varuna on a chariot drawn by two oxen pouring water (and snow) on the dragon mountain. Introduced by a “priest,” the Nāsatya, two twin bearded men, face these three deities; they lead twin rams, which seems to be a “promotion” of the Nāsatya.” Indra, Mitra and Varuna belong to the sovereign gods of the Indo-Europeans, according to Dumézil, who devoted some pages to the “Indian” gods of Mitanni. Regarding the Vedic Nāsatya, or Aśvin, their status oscillates between human and divine(“promotion”), sacrificing and beneficiaries of the soma sacrifice. The Hasanlu bowl iconography depicts the proposed Aśvin introducing a sacrifice, in front of the sovereign deities, but on the same footing by means of the artistic rules of isocephalic representation and in the upper register. We can recognize here not an exact reflection of the Vedic texts, but a Mitanni-Arya variant expressing the ambivalence of their status.

In this perspective we might see the graphic representation of the Aśvin, twins with twin rams in our interpretation in the upper register, as precisely positioned above the naked goddess on the twin rams(again), purposely to present an artistic vision of the Mitanni-Arya version of the Vedic myth of the Aśvin desirous of enjoying Sukanyā, the spouse of the old rejuvenated Cyavana.


 His conclusions are as follows:

The main conclusion is the proposal that the Hasanlu bowl ornamentation is composed of two blocks of beings linked together:- The Mitanni-Arya in the upper register, the heaven and the hierarchical domain of the ruling elites; it seems that it is the exaltation of a priestly cast and of male divinities in chariots;- The Oxus-Hurrian-Mesopotamian in the lower register, which is an Ancient Oriental-Eurasian complex scheme in the process of being minimized.

The present analysis of the Hasanlu bowl might give more substance to the hypothesis of the religious in-fluence of Arya groups in Western Iran at some point. However, we do not pretend that the Hurrian “reading” that proposes to identify images as Kumarbi, Tešub, Ullikummi, Ištar, Šaušga, a “reading” favored by Winter in her comprehensive study of 1989, has been invalidated. Simply, our “reading” gives more consistency to the upper register deities (not simply as Tešub, sun and moon facing sacrificers), the key element being the identification of the Aśvin.


Parpola also agrees about the Indo-Aryan elements of the Hasanlu Bowl (31)

 

Hasanlu, near Lake Urmia, is an important site associated with Early West Iranian Grey Ware in the second half of the 2nd millennium B.C.. Kurochkin (1990, 1994) has convincingly argued for an Aryan affinity of the famous golden bowl from Hasanlu V, dated to the Early Iron Age (c. 1500-1000 B.C.). He notes that two pictorial scenes on the Hasanlu bowl show a hero dressed in a peculiar skirt: in one the hero is associated with a bird of prey and in the other he is combatting a three-headed monster. In the Avestan hymn to the Goddess Anähitä, preserved in the 5th YaSt, the hero Θraetaona is also mentioned twice: first in connection with his battle with the three-headed AZi-Dahäka (YaSt 5.33-34) and secondly when he helps the hero Päurva by making him fly up to heaven in the form of a bird of prey (YaSt 5.61).

Kurochkin's parallels are from Iranian sources, but I should like to point out that at least the former myth is pre-Iranian, for 0raëtaona the son of ÀΘwya has an exact counterpart in the Rgveda, namely Indra's associate Trita Àptya, who slew VisvarOpa, the three-headed son of Tvastr. The Avestan name of the beast, AZi, corresponds to Vedic ahi "snake, dragon", used of Indra's demonic enemy; it also reminds one of the snake or dragon motif so frequently found on the Bronze Age seals of Bactria and Margiana (Pottier 1984; Sarianidi 1986). Moreover, the epithet dahâka, which the dragon has in the Avesta, connects this adversary with the Däsas, against whom the Àryas of the Rgveda were fighting.

The Hasanlu bowl thus seems to reflect Indo-Aryan mythology.



The Evidence from Archaeology and Assyrian texts


That these newcomers to NW Iran are proto-Iranians is completely borne out by archaeological and Assyrian textual evidence. From Iranica Online:
The Medes are for the first time attested in 836 B.C. in an Assyrian inscription of King Salmanasar III (who fought against the Mataí). 

The Persians are first attested some years earlier (843 B.C., which is the date of the first definite evidence for Iranian settlement in Iran proper), under the name Parsuaš (Iranian *Pārsva-). They apparently have to be located to the south and west of Lake Urmia.



TC Young,1967 (18) posited that 1300-1000 BCE Iron Age 1 sites at Hasanlu V, and Sialk V are the best fits for the first Iranians in Zagros.

The written evidence points to the east as the direction from which the Medes and the Persians came, and the Iron I ceramic tradition has roots in Hissar III and related sites in the northeast which were occupied long before Iron I grey ware appeared in western Iran. That tradition and those who bore it spread fairly rapidly over much of western Iran in numbers which perhaps match what the written sources tell us of the earliest Iranians in the Zagros. Furthermore, the areas where this pottery is found match in considerable detail those areas where we have postulated that early Iranians were found by the Assyrians. 


Douglas Adams, 2020 (19) also states it is consensus that "Proto-Iranians to the north (arriving on the edge of the Fertile Crescent [Assyrian records] by 1100 BC)"


Even noted archaeologists who favoured a steppe origin of India-Iranians proposed that Iranians had already reached NW Iran by 1000 BCE. Kuz'mina notes (24),

Evidence for the emergence of horse riding in the 12th century BC is important for resolving the problem of the migration of the Iranian-speaking peoples. E. A. Grantovsky (1970) maintained that analysis of Iranian names indicated that they appeared on the Iranian plateau at the end of the 2nd millennium BC and became numerous only in the 8th century BC. In Iranian culture of this period there appear a number of important innovations. At Hasanlu, Dinka- Tepe, Marlik, and Babadzhan archaeologists have discovered a horse burial rite that is alien to early Iranian cultures.

These ritual burials and images of horse and riders probably mark the route of Iranian-speaking tribes moving from the Eurasian steppes to the Iranian plateau who brought with them horse raising, riding, mounted combat and the cult of the horse.

Thus the archaeological materials of Iran analyzed by M. N. Pogrebova and myself correspond well with the historical data on the migration of the Iranian speaking people to Iran. The contradiction between the hypothesis of E. A. Grantovsky about the Caucasian route and M. D’yakonov’s opinion on the migration through Central Asia is thus eliminated: it is obvious that both routes were used. The beginning of the migration may be believed to have happened in the last quarter of the 2nd millennium BC. 


These dates line up perfectly with what we see in the genetic data of NW Iran: BMAC-like ancestry entering the region post 1500BCE. The question is if this approximate date is the consensus, and we have Hasanlu samples dated till 600bce at the latest, then why did Lazaridis et al reject the Hasanlu iron age people as non-Iranian speakers purely based on their idea of what Iranian genetic profile should be like? 


 Putting it Together


So, the Gods on the Hasanlu bowl may be the Mitanni Vedic deities. Other deities may be linked to the Avesta. And we see the presence of IVCp ancestry (including the L1a Y-hg) and BMAC ancestry in this region. 

Notably, there is no evidence of Sintashta-related steppe_mlba ancestry at Hasanlu or Dinkha Tepe which according to Narasimhan et al 2019 (5) and this paper was the carrier of Indo-Iranian languages and culture to SC Asia. Rather, it seems that NW Iran in the bronze age only had contact with BMAC and IVC regions. There is additional evidence of IVC people migrating to the near east post 2000BCE. IVC people/traders mediated movements of elephants (6) and Zebu cattle (7) to the near east post 2000BCE. Do note that since the Aryans did not change the language of the Mitanni kingdom subjects but rather only provided loanwords and some cultural elements (8), a large-scale migration (defined by me as >30% admixture visible in autosomal ancestry) is neither required nor proposed by me. 

BMAC ancestry is also the main ancestry component in modern Iranians compared to steppe_mlba ancestry. If one follows Viktor Sarianidi's school of thought and considers BMAC Iranian, then it elegantly explains the spread of the Iranian language family. The Assyrian texts and archaeological data also support the entry of proto-Iranians into Zagros by 1300BCE.

Now that the CHG/Iran component has been proposed to be the first carrier of IE languages, geneticists, archaeologists, anthropologists and linguists must examine in detail what this ancestry brought to regions where it dominated, and shed existing biases that stop them from exploring neglected ideas. Perhaps linguists should also focus more on finding out what loanwords and cultural practices came into Indo-Iranian as a result of the genetic contact with the steppe, rather than apriori assuming that the minor steppe ancestry caused language change over most of the Indian subcontinent and Iran.

After all, 40% steppe ancestry and almost 100% steppe male markers (14) could not change the language of the Basque people in Iberia, how can one be sure an average 15-20% steppe ancestry can do that job without leaving major traces (through large substrate effect)  of the native languages in a vast region like the Indian subcontinent?


The R1a and R1b quandary, Rules for thee but not for me



One of the reasons given by geneticists to prove that steppe bronze age populations brought Indo-Aryan languages to the Indian subcontinent is the presence of R1a Y haplogroups in a significant % (~25-35%) of Indian males (5), which according to them is of Steppe origin*. However, this paper itself provides evidence that this may not be correlated with linguistic change. Consider the following statements from Supplement pg 267, the context is that 11/14 Hasanlu males have R1b Y hg related to Yamnaya which is supposed to be an IE-speaking population (which I agree with).

* This is debatable, Indian R1a-L657 is born around 2200BCE much earlier than steppe migration to India and has not been found anywhere on the steppe yet or in modern Europeans/Russians. Rather, the steppe in the BA has the brother clade R1a-Z2124+ in overwhelming proportion.
There are some reasons to think that the spread of the steppe ancestry into NW Iran is not associated with the historical Iranian languages: first, it's very minor impact; second, its association with the movement of steppe ancestry into the South Caucasus that could plausibly be associated with Proto-Armenian speakers and would thus predict an (Iranian, Armenian) linguistic relationship rather than the universally accepted relationship of Iranian languages with the Indo-Aryan (or Indic) languages of South Asia. Third, South Asian populations have virtually none of the Yamnaya-derived Y-chromosome lineages, and conversely, the samples from Iran included in our study do not have the R-Z93 lineage tracking the spread of steppe ancestry into Central/South Asia. Finally, our samples are from the periphery of the Iranian world (NW Iran) and may also predate the establishment of the Median and Achaemenid kingdoms of Iran during the 1st millennium BCE.
The authors say that despite 80% Y haplogroup R1b+ at Hasanlu_IA, given the minor autosomal impact chances are that they did not speak an IE language at all. If at all they did speak an IE language, it must be related to Armenian since the minor steppe ancestry came to the region in the same pulse which according to them also gave birth to proto-Armenian.

 Lazaridis et al 2022 (25) also say this:

Present-day Iranians do have R-Z93 Y chromosomes or the more general upstream R1a-M17 ones [observed in every one of 19 diverse populations from Iran , as well as in present-day Indians, and modern Iranians almost completely lack R1b Y chromosomes (<1% frequency)]. Thus, it appears that R1a haplogroup Y chromosomes represent a common link between ancient and modern Indo-Iranians, whereas R1b haplogroup Y chromosomes (to which many of the Hasanlu males belonged) do not. The absence of any R1a examples among 16 males at Hasanlu, who are instead patrilineally related to individuals from Armenia, suggests that a non–Indo-Iranian (either related to Armenian or belonging to the non–Indo-European local population) language may have been spoken there and that Iranian languages may have been introduced to the Iranian plateau from Central Asia only in the first millennium. 

Firstly, the authors are wrong about R1b in modern Iranians being absent (<1%). Grugni et al 2012 (15) found >10% R1b+ males among Iranians. Secondly, what is this obsession with R1a and Indo-Iranians? Surely a few R1a markers cannot signify a sudden change in language, it has to be associated with a significant influx of external autosomal ancestry (9)(23) which is nowhere to be seen in Iron age NW Iran or later. Kivisild et al 2003 (27), show the presence of R1a in 26% of Chenchu males, a Dravidian tribe found in South India. R1a1 is found in all castes of Dravidian-speaking Tamil Nadu and Andhra Pradesh states in South India (28), irrespective of  'caste hierarchy. So how can the authors simply and lazily assume the presence of R1a = language change to Indo-Iranian?

R1a proportion in Iranians is also low, as compared to 80% R1b at Hasanlu. Grugni et al 2012 (15) found 14.4% R1a+ among 938 Iranian subjects from 15 ethnic groups, ranging from 0% - 25% across groups. Lopez et al, 2017 found 10.5% R1a in non-Zoroastrian Iranians, 5.7% in Indian lay Parsis, 0% in Indian Parsi priests, 5.3% in Iranian lay Zoroastrians, and 12.5% in Iranian Zoroastrian priests.

Ok, so 80% R1b did not bring about Iranian languages till 600BCE (the youngest Hasanlu sample date) as per the authors. Why did the authors throw the available archaeological and literary data when some of it clearly suggests that Hasanlu and Dinkha Tepe were already Iranian by this time, to then just rely on R1a as the marker of Indo-Iranian languages? Is this really scientific? It sure seems extremely dogmatic and unscientific.

So did a lot of Sintashta-related steppe ancestry enter NW Iranians post 1000BCE Hasanlu?


If Hasanlu people didn't speak Iranian, how do we explain this?


Modern Iranians seem to be mostly descended from the Hasanlu_IA population with very minor later steppe ancestry.

Modern Iranian
Modern Iranian groups (avg) are descended from Hasanlu_IA


The fits of the above G25 model are <0.01 so they are fantastic. So is a measly 3% Sintashta steppe-related ancestry since 1000BCE Hasanlu supposed to explain NW Iranian languages like Kurdish when 40% steppe couldn't change the language of Basques?

In fact, Kurdish people who speak a Western Iranian dialect can be modelled simply as Hasanlu_IA with great fits, no extra ancestry needed. This is evidence that Hasanlu_IA was likely Iranian speaking and Yamnaya or Sintashta or Armenians had nothing to do with it. 

Target: Kurdish_Average
Distance: 0.9144% / 0.00914426
97.4 IRN_Hasanlu_IA
2.6 Levant_Baqah_BA

Kurds PCA
3D PCA: Kurdish cluster inside the Hasanlu_IA cluster


However, to provide fair context, G25 and PCA are not formal tools to prove admixture. More formal analysis needs to be done for Kurdish genomes to prove the absence of a significant amount of excess steppe ancestry.

Another line of evidence to prove that Sintashta_MLBA ancestry has no special relation to Modern Iranians over Hasanlu_IA is through F4 stats. 

Iranian F4
F4(Chimp, Steppe, Hasanlu_IA, Modern Iranians)


As control, we can check F4 stats for modern Indians

F4(Chimp, Steppe, IVCp, Modern North Indian) >> 0



The highly significant + F4 tells us that Sintashta is more closely related to Modern North Indians than with IVC and is proof that modern North Indians (Ror, Gujjar, Kamboj) received additional Sintashta-related ancestry after the pre 2000BCE IVCp period. This same evidence is not seen in modern Iranians.

If Iranians had excess Sintashta-related ancestry over Hasanlu_IA (to explain their Iranian language as per steppe theory), the F4 should have been positive with a significant Z-Score. But we see a negative F4 with a non-significant Z-Score, indicating similar Sintashta-related ancestry (actually Yamnaya related) in both Hasanlu_IA and Modern Iranians.


All of this matches perfectly with the entry of BMAC ancestry into Hasanlu and Dinkha Tepe by the late bronze age (1400BCE). We have solved the issue of Proto-Iranians, they were people with BMAC-like ancestry and carried little to no Steppe_MLBA ancestry, and certainly didn't carry the R1a 'supposedly Indo-Iranian' male marker.

I propose that the BMAC-related autosomal ancestry in Iranians is responsible for the spread of Iranian languages, not the bronze age steppe. I propose that Hasanlu_IA people were already Iranian speakers which has nothing to do with their R1b Y hg but rather the large-scale BMAC-like admixture from the east. This is the second way that this new paper upends the Steppe theory.

The question becomes - how did BMAC become Iranian speaking without steppe contact before 2000BCE? That should be the future direction of research.


THE MITANNI EVIDENCE

A question separate (but maybe not) from the first Iranians reaching western and north-western Iran concerns the unambiguous presence of Indo-Iranian words and proper names related to the Mitanni Kingdom around Syria. The latest date by which Indo-Iranian speakers should be present in the proximity of Syria is the 16th century BCE (Witzel 2019, Diakonoff 1996)(29, 30). 

While Witzel 2019 (29) describes these words as early Indo-Aryan:

However, a major milestone is the appearance of early IA speakers in the near east, on the northern rims of Mesopotamia. They have left many loanwords in a non-related Caucasus language, Mitanni Hurrite, from ca. 1600 BCE onwards. The Mitanni kingdom covered northern Iraq and Syria; IA speakers will have reached it from its eastern border in the Zagros mountains. The loanwords extend from horse culture (colors, number of rounds in chariot races), to the throne names of Mitanni kings, and a brief list of four Vedic deities (Mitra, Varuṇa, Aryaman, Nāsatya) attached to a treaty with the Hittites (1380 BCE).
Diakonoff connects these loanwords to the Dardo-Nuristani branch of Indo-Iranian.
We have tried to show that linguistic data seem to point to the possibility that ancient Iran could have been invaded by earlier Indo-Iranian Dardo-Nuristani tribes before the arrival of Iranian-speaking tribes, that is, the Medes, about the middle of the second millennium BCE.
Lazaridis et al pay absolutely zero heed to this Mitanni evidence. After all, this is the hard evidence of the earliest unambiguous presence of Indo-Iranians. Constraining the origin of I-Ir to Eastern Europe because of loanwords in proto-Uralic is not hard evidence, just a hypothesis which has other plausible explanations.

So, where is the genetic evidence of Sintashta steppe-related people moving towards Syria in the mid-2nd Millennium BCE? Where is the genetic evidence of Sintashta steppe-related people populating Western Iran and spreading Median or pre-Median Iranic languages? All genetic evidence from NW Iran points to Oxus-related ancestry (along with minor IVC-related ancestry) moving west during both these time periods in question. Moreover, the sample of the 'Well Lady' from Alalakh, Turkey dated to ~1550BCE is also of Central Asian /Oxus origin (34).



CONCLUSION


I propose that the high percentage of BMAC-related ancestry along with minor Indus Valley ancestry at Hasanlu_IA, Hasanlu_LBA, and DinkhaTepe_LBA can explain the spread of Iranian languages in this region as language spread (and later persistence) almost always requires a sizeable migration (9)(23) and also explain the Mitanni elements found in this region. The exact process through which this occurred needs more archaeological research.

The second area of further research is the explanation of how the Oxus & NW Indian regions became Indo-European speakers.

The third area of further research is to examine the linguistic and cultural impact of the steppe ancestry which is present in modern Indians, through the lens of cultural exchange rather than that of imposition and dominance.




REFERENCES:

(1) Éadaoin Harney, Nick Patterson, David Reich, John Wakeley, Assessing the performance of qpAdm: a statistical tool for studying population admixture, Genetics, Volume 217, Issue 4, April 2021, iyaa045, https://doi.org/10.1093/genetics/iyaa045

(2) Sarianidi, V. (2014). The Indo-Iranian Problem in the Light of the Latest Excavations in Margiana. Studia Orientalia Electronica, 94, 417–442. Retrieved from https://journal.fi/store/article/view/43986

(3) G. N. Kurochkin, “Archaeological Search for the Near Eastern Aryans and the Royal Cemetery of Mārlik in Northern Iran,” in South Asian Archaeology 1993, ed. A. Parpola and P. Koskikallio, vol. 1, Helsinki, 1994, pp. 389-95.

(4) FRANCFORT, HENRI-PAUL. “A Note on the Hasanlu Bowl as Structural Network: Mitanni-Arya and Hurrian?” Bulletin of the Asia Institute, vol. 22, 2008, pp. 171–88. JSTOR, http://www.jstor.org/stable/24049242. Accessed 28 Aug. 2022.

(5) Narasimhan VM, Patterson N, Moorjani P, et al. The formation of human populations in South and Central Asia. Science. 2019;365(6457):eaat7487. doi:10.1126/science.aat7487

(6) Canan Çakırlar & Salima Ikram (2016) ‘When elephants battle, the grass suffers.’ Power, ivory and the Syrian elephant, Levant, 48:2, 167-183, DOI: 10.1080/00758914.2016.1198068

(7) Verdugo, Marta Pereira, et al. “Ancient Cattle Genomics, Origins, and Rapid Turnover in the Fertile Crescent.” Science, vol. 365, no. 6449, 2019, pp. 173–176., https://doi.org/10.1126/science.aav1002.

(8) Aram Kosyan. “An Aryan in Išuwa.” Iran & the Caucasus, vol. 10, no. 1, 2006, pp. 1–6. JSTOR, http://www.jstor.org/stable/4030938. Accessed 28 Aug. 2022.

(9) Bellwood, Peter. “Some Observations on the Transeurasian Language Family, from the Perspective of the Farming/Language Dispersal Hypothesis.” Evolutionary Human Sciences, vol. 2, 2020, p. e37., doi:10.1017/ehs.2020.34.

(10) Dubova N. A., 2015, Animal burial and their cults in Margiana, in N. Laneri (Ed.), Defining the Sacred. Approaches to the Archaeology of Religion in the Near East, 13-23, Oxbow Books.

(11) Julio, BENDEZU-SARMIENTO. “2021, J.BENDEZU-SARMIENTO, Horse Domestication History in Turkmenistan and Other Regions of Asia, MIRAS, 1, p. 17-29.” MIRAS (2021): n. pag. Print.

(12) Bryant, E. (2001). The quest for the origins of Vedic culture. Oxford University Press.

(13) Shinde, Vasant et al. “An Ancient Harappan Genome Lacks Ancestry from Steppe Pastoralists or Iranian Farmers.” Cell vol. 179,3 (2019): 729-735.e10. doi:10.1016/j.cell.2019.08.048

(14) Olalde I, Mallick S, Patterson N, et al. The genomic history of the Iberian Peninsula over the past 8000 years. Science. 2019;363(6432):1230-1234. doi:10.1126/science.aav4040

(15) Grugni V, Battaglia V, Hooshiar Kashani B, et al. Ancient migratory events in the Middle East: new clues from the Y-chromosome variation of modern Iranians. PLoS One. 2012;7(7):e41252. doi:10.1371/journal.pone.0041252

(16) López S, Thomas MG, van Dorp L, et al. The Genetic Legacy of Zoroastrianism in Iran and India: Insights into Population Structure, Gene Flow, and Selection. Am J Hum Genet. 2017;101(3):353-368. doi:10.1016/j.ajhg.2017.07.013

(17) KRAUSE, JOHANNES. “Chapter 6: Europeans Find a Language.” Essay. In Short History of Humanity. S.l.: NIELSEN BOOKDATA, 2022. 

(18) Young, T. Cuyler. “The Iranian Migration into the Zagros.” Iran, vol. 5, 1967, pp. 11–34. JSTOR, https://doi.org/10.2307/4299585. Accessed 31 Aug. 2022.

(19) Adams, Douglas. “Language Log.” Language Log " The Geographical, Archeological, Genetic, and Linguistic Origins of Tocharian, 14 July 2020, https://languagelog.ldc.upenn.edu/nll/?p=47645.

 (20) Heggarty, P. (2018). Why Indo-European? Clarifying cross-disciplinary misconceptions on farming vs. pastoralism. In G. Kroonen, J. P. Mallory, & B. Comrie (Eds.), Talking Neolithic: Proceedings of the workshop on Indo-European origins held at the Max Planck Institute for Evolutionary Anthropology, Leipzig, December 2-3, 2013. Washington, D.C.: Institute for the Study of Man.

(21) Shinde, Vasant et al. “An Ancient Harappan Genome Lacks Ancestry from Steppe Pastoralists or Iranian Farmers.” Cell vol. 179,3 (2019): 729-735.e10. doi:10.1016/j.cell.2019.08.048

(22) Maier, Robert et al. "On the limits of fitting complex models of population history to genetic data"
bioRxiv 2022.05.08.491072; doi:10.1101/2022.05.08.491072

(23) Heggarty, Paul. (2014). Prehistory through language and archaeology [Routledge Handbook of Historical Linguistics]. 

(24) Kuz'mina, E. E., Pitina, S., Prudovsky, P., &amp; Mallory, J. P. (2007). Chapters 8, 26. In The origin of the indo-Iranians (pp. 140–380). essay, Brill. 

(25) Lazaridis, Iosif et al. “A genetic probe into the ancient and medieval history of Southern Europe and West Asia.” Science (New York, N.Y.) vol. 377,6609 (2022): 940-951. doi:10.1126/science.abq0755

(26) Spengler, Robert et al. “Early agriculture and crop transmission among Bronze Age mobile pastoralists of Central Eurasia” Proc. R. Soc. B.2812013338220133382 (2014). http://doi.org/10.1098/rspb.2013.3382

(27) Kivisild T, Rootsi S, Metspalu M, et al. The genetic heritage of the earliest settlers persists both in Indian tribal and caste populations. Am J Hum Genet. 2003;72(2):313-332. doi:10.1086/346068

(28) Watkins WS, Thara R, Mowry BJ, et al. Genetic variation in South Indian castes: evidence from Y-chromosome, mitochondrial, and autosomal polymorphisms. BMC Genet. 2008;9:86. Published 2008 Dec 12. doi:10.1186/1471-2156-9-86

(29) Witzel, Michael. (2019). Early ‘Aryans’ and their neighbours outside and inside India. Journal of Biosciences. 44. 10.1007/s12038-019-9881-7. 

(30) DIAKONOFF, I. M. (1996). Pre-Median Indo-Iranian Tribes in Northern Iran? Bulletin of the Asia Institute, 10, 11–13. http://www.jstor.org/stable/24048881

(31) Erdosy, G. (2012). The Indo-Aryans of Ancient South Asia: Language, Material Culture and Ethnicity. Berlin, Boston: De Gruyter. https://doi.org/10.1515/9783110816433

(32) C. C. Lamberg‐Karlovsky , "Archaeology and Language: The Indo‐Iranians," Current Anthropology 43, no. 1 (February 2002). https://doi.org/10.1086/324130

(33) Dubova, Nadezhda. (2012). Animal Burials and Cult of them in Margiana. 

(34) Skourtanioti E, Erdal YS, Frangipane M, et al. Genomic History of Neolithic to Bronze Age Anatolia, Northern Levant, and Southern Caucasus. Cell. 2020;181(5):1158-1175.e28. doi:10.1016/j.cell.2020.04.044







61 comments:

  1. Do the authors of this paper, reject IranN/IndiaN geneflow to the Steppe/Yamnaya?

    ReplyDelete
  2. They haven't even modeled with IranN as far as I can see. They label everything as CHG, very shoddy. Also they haven't modeled steppe pops, only the new ones. But I'll look at the papers with fresh eyes tomorrow.

    ReplyDelete
  3. All these southern arc populations are admixed between chg and IranN. Nothing like that seems to be explored. There's also SC asian inflow into Armenia which hasn't been explored, overall it's underwhelming.

    ReplyDelete
  4. Plus, they seem to have found 'Indo Anatolian' language family from somehwere. It was in the dustbin, and suddenly these people have brought it out and they keep using it - Lazridis's twitter thread for example.

    ReplyDelete
  5. I was about to send you these links to you, but I see that you've seen them. I am really curious to know the origin of the other branches of PIE, specially Greek, as it seems to have "Anatolian like words"

    ReplyDelete
  6. @Daniel

    Greek, Armenian and IIr form a grouping at some point in time and had to migrate from a common source after being together for a while. However the current consensus that these three groups spread from steppe is still held. I think Greek came from Anatolia, not Yamnaya.

    ReplyDelete
  7. Here is a good primer on some of the research on evidence for or against Indo Anatolian

    http://www.worldcat.org/search?q=Greater+Anatolia+and+the+Indo-Hittite+Language+Family

    ReplyDelete
  8. @3rdacc If you search for people north of the Caucasus in the 3rd millenium you will see that there was a lot of mining, whose products were exported to the Balkans and East Europe. So, perhaps, you had some sort of serfdom/slavery going on, with IE being carried by trade of goods and slaves.

    But I don't see people consideinrg the Y chromossomes as also due slavery of males.

    ReplyDelete
  9. The 'BMAC impact' in Hasanlu LBA is interesting, but have you tried with Tepe Hissar instead of Bustan? It is closer, and there are well known archaeological connections through gray ware: https://www.iranicaonline.org/articles/ceramics-ix

    About L1a, it is found also in Areni cave in Chalcolithic Armenia, that here is analyzed as 91/92% Armenia Neolithic + 7/8% EHG.

    About Swat, the archaeological connection points rather to an Iranian origin: for instance, the gray ware in Swat and Hasanlu comes from NE Iran as you can see in the link above.

    ReplyDelete
  10. The Armenian L1a is different, rather go by the yfull nomenclature. I have included yfull link in blogpost. You will immediately see that the Hasanlu L1a is right under swat IA, and next to a konkani.

    Wrt tepe hissar, tepe hissar is almost same as bmac. So it should work too. However i can't include it coz I have too many populations to compute already.

    Yes, swat has bmac/tepehissar ancestry. But ivc ancestry in that region looks unmistakeable so far.

    ReplyDelete
  11. Go to https://www.yfull.com/tree/L-Y6288/

    See the sample under it. I6555 from Pakistan_loebanr_IA (red means ancient sample). carbon date ~900bce.

    Hasanlu L-Y28524 falls directly under L-Y6288. There is also a konkani at L-Y28524*.

    The clade above this, is L-l1307, that too is full of Indians, pakistanis and Lankans.

    Armenian L1a is L-Y31961 https://www.yfull.com/tree/L-Y31961*/

    ReplyDelete
  12. @Ashish,
    The L-Y6288 is downstream of L-1307. This L-1307 is also called L-M357 or L1a2. This is mostly present in North-west India , pakistani and afghans(its the one of the major haplogroups of jats and rors) though a few occurrences are present in south too. Just because it's present in a few modern konkanis, it doesn't mean it originated in india. Also, there are a lot of middle easterners in this clade(apart from chechens for whom it seems to be due to very recent founder founder effect).

    The upcoming 4600 BCE sample from turkemenistan (Monjukli Depe) is from this L-1307 or L-M357(L1a2) clade. High probability that its origin lies in a region spanning from caucasus to NE iran to SC Asia. Might've entered southvasia when that eastern iranic like ancestry cane !

    ReplyDelete
  13. I make the connection because of

    1. I found IVC ancestry in Hasanlu (not because of trying, but because it worked)
    2. Archaeological connections
    3. That exact subclade is only found in India and Pak on Yfull.

    I did not claim anything about origin, I only said this L-Y28524 in Hasanlu seems to be from IVC / swat.

    ReplyDelete
    Replies
    1. 2. Archaeological connection ? You mean that petroglyph ?

      3. The L-Y28524 seems to have UAE and Qatar samples too,
      https://www.yfull.com/tree/L-Y28524/ so how this subclade found only among indian and pak on YFull ?

      And going by the formation and TMRCA date of L-Y6288(>7000 years), the assignment of the swat ancient is not as downstream/terminal as expected but that seems normal in case of ancient DNAs where downstrean SNPs are often not found or are damaged.

      Delete
  14. @vASISTHA, have you seen this?
    https://eurogenes.blogspot.com/2022/08/dear-iosif.html

    ReplyDelete
  15. Yes daniel, I left some comments.

    ReplyDelete
  16. "2. Archaeological connection ? You mean that petroglyph ?"
    Did you not read Parpolas excerpt. Same gray vase ware with button/disk bases at swat and hasanlu?

    "The L-Y28524 seems to have UAE and Qatar samples too,
    https://www.yfull.com/tree/L-Y28524/ so how this subclade found only among indian and pak on YFull"

    The UAE and Qatar samples have 60 additional SNPs than L-Y28524, are downstream, and are L-Y60258.

    "And going by the formation and TMRCA date of L-Y6288(>7000 years), the assignment of the swat ancient is not as downstream/terminal as expected but that seems normal in case of ancient DNAs where downstrean SNPs are often not found or are damaged."

    YFull only accepts the highest quality samples and it does not put up the samples if there are missing markers downstream. You can download the dataset and manually check I6555.

    ReplyDelete
    Replies
    1. "YFull only accepts the highest quality samples and it does not put up the samples if there are missing markers downstream."

      lnteresting, then the 3500 year old uzbekistan sample here at https://www.yfull.com/tree/R1b/ is really basal R1b as of now & it's assignment is not due to further missing SNPs downstream ?

      Delete
  17. Many smples from those countries on yfull are from recent South Asian migrants living there. Remember Coldmountians on Anthrogenica mentioning it before.

    ReplyDelete
  18. Nope, no missing information. That sample is R1b* or R1b-prePh155. Same clade found at tarim.If you click on the i next to it, its written R1b*. It has some of Ph-155 mutations but not all, hence prePH155.

    ReplyDelete
  19. Thank you for clarifying about L1a and Tepe Hissar.
    What about L1a in Shahr-i Sokhta and Bustan? Is it related?

    About the archaeological connection, "same gray vase ware with button/disk bases at swat and hasanlu" does not suggest a movement from Swat to Hasanlu, considering the Iranian origin of Hasanlu gray ware. The period of iron age Hasanlu and Swat seems the same 1000-800 BCE, but gray ware arrived earlier in Hasanlu and also in Swat: "The polished black-gray pottery of Gandhara grave culture during the Ghalegay IV period, earlier considered to run from 1700-1400 BCE, has been associated with that of other BMAC sites like Dashly in Afghanistan, Tepe Hissar and Tureng Teppe... However, these datings between 1700 and 1000 BCE are currently considered outdated, and Gandhara grave culture has been re-dated to c. 1400 to 800 BCE." https://en.wikipedia.org/wiki/Gandhara_grave_culture#Indo-Aryan_migrations
    Ghalegay IV gray ware clearly belongs to the bronze age and derives from Iranian or BMAC gray ware. Bronze age Swat has also many cases of West Asian E1b1b1b2.

    ReplyDelete
  20. You are correct. But the predominant view is that BMAC had contact with Swat rather than iran, although both BMAC and iran had similar ancestry/culture. E1b is present in BMAC too.

    ReplyDelete
  21. "What about L1a in Shahr-i Sokhta and Bustan? Is it related?" Havent checked their Y-Full terminal clades.

    ReplyDelete
  22. " the predominant view is that BMAC had contact with Swat rather than iran" in which sense? From Iran to BMAC, we have the movement of Gray Ware: "Carinated gray wares found in small quantities in Namazga IV are considered by Soviet scholars to be forerunners of Namazga V unpainted wares with similar shapes, reflecting a possible eastward migration of the “burnished gray-ware people.” Isolated burnished gray-ware vessels occur sporadically still farther east, for instance at Dashly I and Gonur Depe in the area around Mary (old Margiana) and even at Sarazm, east of Samarkand" https://www.iranicaonline.org/articles/ceramics-ix
    In the other direction, Tepe Hissar IIIC has typical BMAC objects, but I don't know of other BMAC movements more to the west.

    About IVC ancestry, it was frequent in BMAC and Eastern Iran, so migrations from there could spread that ancestry to western Iran.

    ReplyDelete
  23. Sorry I meant to say that Swat had contact with BMAC rather than Tepe Hissar or N/NE iran directly

    ReplyDelete
  24. From Sarianidi
    "It is worth mentioning that, apart from common funeral ceremonies, the graves at the Gonur necropolis and the Swat graveyard manifested similar ceramic complexes which represent the late variation of the BMAC. The similarity between Swat and Margiana was also supported by another detail. Some anthfopomorphic statuettes had a specific common feature: a cup-like hollow on their heads (Sarianidi 1998a, fig. | 1,l). This detail was earlier unknown in Central Asia. Statuettes with high "crowns" widening upwards were found in Swar graves at the same time (Antonini & Stacul 1972; Dani 1967, pls. li & liii). These "crowns" find their analogies in Turkmenistan of the Namazga V period' There is only one iconographic difference between them: all statuettes of this type from Margiana and Swat are represented standing (Sarianidi 1998a, fig. 16.2) while the Namazga V statuettes are always seated."

    ReplyDelete
  25. They commented in the supplementary info on the methodology, Iran_N didn't produce results with robust enough p-values like CHG did, or something in the likes.

    ReplyDelete
  26. Have you checked the afinity of the Basques with these samples?

    ReplyDelete
  27. Nope. Thats the last thing on my mind. Im busy with finishing these iran samples. Then might move on to anatolians.

    ReplyDelete
  28. @Vasistha that's to put doubt that Yamana was only composed by IE speakers.

    ReplyDelete
  29. Have edited and finished this post. Excited!

    ReplyDelete
  30. Well I always said the BMAC was the origin of the Iranian civilization ie Zoroastrianism, Iranian languages, shahname etc so this bowl definitely is consistent with that.

    I dont have these new samples merged yet, but if you want to know if there is additional South Asian ancestry in any samples you can just run the calc

    f3(A=test, B=ONG.SG, C=Chimp.REF) and set outgroupmode= TRUE/YES whatever

    then just look to find which samples have greater ONG affinity.

    I have done this for Steppe samples and others

    1 ONG.SG Kazakhstan_Eneolithic_Botai Chimp.REF 217. 1.45 149. 62958
    2 ONG.SG Russia_Late_Sarmatian.SG Chimp.REF 216. 1.34 162. 73724
    3 ONG.SG Russia_HG_Karelia.SG Chimp.REF 215. 1.44 150. 63238
    4 ONG.SG Tajikistan_C_Sarazm Chimp.REF 215. 1.46 147. 57720
    5 ONG.SG Russia_Andronovo.SG Chimp.REF 214. 1.28 168. 77585
    6 ONG.SG Luxembourg_Loschbour Chimp.REF 214. 1.58 135. 52164
    7 ONG.SG Russia_Sidelkino_HG.SG Chimp.REF 214. 1.44 148. 60555
    8 ONG.SG Russia_Afanasievo Chimp.REF 214. 1.19 179. 86733
    9 ONG.SG Russia_MLBA_Sintashta Chimp.REF 214. 1.18 181. 87459
    10 ONG.SG Turkmenistan_IA.SG Chimp.REF 213. 1.51 141. 60093
    11 ONG.SG Georgia_Kotias.SG Chimp.REF 212. 1.42 149. 72021
    12 ONG.SG Turkmenistan_Gonur_BA_1 Chimp.REF 212. 1.29 164. 68232
    13 ONG.SG Iran_GanjDareh_N Chimp.REF 212. 1.22 174. 74888
    14 ONG.SG Anatolia_N Chimp.REF 212. 1.17 180. 88294
    15 ONG.SG Iran_C_TepeHissar Chimp.REF 211. 1.23 171. 73247

    ReplyDelete
  31. Could you please clarify some notations used? For example, what is meant by "Hasanlu_MBA + IVCp/ Hasanlu_LBA_A"? What do the / and + signify, are they different? Is this meant to represent a 3 source model of Hasanlu_MBA, IVCp, and Hasanlu_LBA_A... or something else? Thanks.

    ReplyDelete
  32. Hasanlu_MBA + IVCp/ Hasanlu_LBA_A

    means

    Hasanlu_MBA + IVCp as well as Hasanlu_MBA + Hasanlu_LBA_A.

    So 2 different 2 source models, not a 3 source model.

    ReplyDelete
  33. Some borrowings might be due Tocharian. Tocharian borrowed words to chinese and other languages of Siberia. But people don't want to acknoledge these same borrowings to Uralic, which is a circumpolar family of languages.

    ReplyDelete
  34. @Vasistha: "After all, 40% steppe ancestry and almost 100% steppe male markers (14) could not change the language of the Basque people in Iberia", maybe "pre-basque" was the dominating Steppe language, which would also justify its similarities to caucasian languages?

    ReplyDelete
  35. @daniel, im not aware of that research.

    ReplyDelete
  36. I have added a section about R1b and Iranians at the end. Do read that part again.

    ReplyDelete
  37. @Vasistha , I quoted you in the paper, if that's what you mean. If it is the tocharian thing, I can provide it later.

    ReplyDelete
  38. Ashish, do BMAC and IVC have shared ancestry?

    ReplyDelete
  39. Yes Hasanlu deals a great blow to the Sintashta Indo-Iranian nonsense.

    However, a few comments on common misconceptions:
    1." BMAC " is a terrible designation for a network that spanned from Hissar to Quetta. A lot of the characteristics of BMAC don't originate in Bactria nor Margiana, eg. Grey Ware from Hissar, chlorite and alabaster from Sistan, the pilastered buildings with fire altars from Mundigak..etc.

    2. While the Indo-Aryans did not have a significant impact in Mesopotamia they did settle in Northern Iran. The Mazena(Mazandaran) Daevas of the Younger Avesta even left behind toponyms; Mount Bikni(Vighna). Actually, even Urmiya is considered IA.

    3. Zoroastrianism did not spread with Zarathustra and Vishtaspa but with a community that came later. Wherever the original homeland was doesn't matter but the group that spread it was centered in western Sistan and Herat(Zamyad Yasht). The younger Avesta itself went through multiple redactions for example, the Anahita Yasht during the time of Artaxerxes II.

    4. Francfort's interpretation of the Hasanlu bowl is correct but he switched Indra and Varuna. However, his claim of the East Iranian eagle hero being non-IE is dubious. Every IE dragonslayer has been associated with eagles it's even the favorite form of various thunder gods. Also, somehow Francfort forgot about the original IE river goddess Danu and the various water goddesses of the Scythians. To be blunt, BMAC is one of the best examples of IE cultures from rituals to iconography.

    I've written about this in detail to Alberto but I haven't finished the last part. Though, It looks like the Mitanni did not come from Hissar IIIC. North Iranian and Mundigak samples will pretty much solve the Proto-Indo-Aryan debate.

    ReplyDelete
  40. "Ashish, do BMAC and IVC have shared ancestry?"

    Yes, while BMAC ancestry is largely like Tepe Hissar, IVC sees a good chunk of tepe hissar like ancestry (carrying anatolian component) coming in at an unknown time.

    On top of that we already see IVC migrants at Gonur (2100bce), and also bmac ancestry in swat iron age. So, intense contacts between the two.

    ReplyDelete
  41. @vara Thank you for the comments.

    "To be blunt, BMAC is one of the best examples of IE cultures from rituals to iconography."

    I agree.

    As for Mitanni, 2 Hasanlu samples I6430 and I6428 are showing a high amount of swat_udegram or swat_aligrama_nosteppe ancestry of upto 30%.

    I will add the results as soon as my models finish running. My idea is that Mitanni Aryan elements do not need a large migration (as large scale language change didn't occur) which is consistent with the ivc ancestry connections being seen at hasanlu

    However, bmac like ancestry definitely does overpower this NW iranian region post 1700bce, highly consistent with language changes in the region.

    ReplyDelete
  42. "BMAC " is a terrible designation for a network that spanned from Hissar to Quetta. A lot of the characteristics of BMAC don't originate in Bactria nor Margiana, eg. Grey Ware from Hissar, chlorite and alabaster from Sistan, the pilastered buildings with fire altars from Mundigak..etc."

    Right. Because my approach is from the pov of genetics primarily rather than archaeology, and since BMAC sites have lots of aDna samples available to use unlike northern /eastern iran, i tend to prefer that nomenclature.

    But in general, bmac samples are very similar to shahr-sokhta siestan samples and can be used interchangeably in most models. I wish we get LBA and IA samples from siestan soon.

    ReplyDelete
  43. @Vasistha "IVC sees a good chunk of tepe hissar like ancestry (carrying anatolian component) coming in at an unknown time" Interesting, where can we see that?

    About BMAC/Swat-like ancestry in Hasanlu, at least initially it could be mostly the Mitanni Aryan component. It was thought that Hasanlu with gray ware and other changes reveal around the middle of 2nd mill. BCE the arrival of Mitanni Aryans or Iranians. See https://www.iranicaonline.org/articles/hasanlu-teppe-i: "In the second half of the second millennium B.C.E., a major change took place in burial customs, ceramics, and architecture with the appearance of burnished gray, black, red, and brown pottery made in a variety of shapes, including stemmed goblets, spouted jars, and vessels reproducing metal techniques such as gadrooning. At the end of the second millennium B.C.E., the top of the Citadel mound was occupied by a series of monumental buildings, of which at least one was a large temple with a columned hall measuring 18 by 24 meters with four rows of six columns each, a forerunner of later columned halls in Media and Achaemenid Anshan (q.v.)." Recently, Danti has questioned the arrival of new people (see https://www.academia.edu/41532652/The_Iron_Age_at_Hasanlu_Iran_New_Perspectives_by_Megan_Cifarelli), but the genetic data can confirm the migration.

    Now, Parpola saw this as a migration of Sintashta Indo-Aryans from BMAC. Now we know that BMAC was not conquered by steppe people at the time Parpola thought (1900 BCE). To me, BMAC and near areas like Tepe Hissar were proto-Iranian, and Mitanni Aryan too: http://new-indology.blogspot.com/2017/05/were-mitanni-really-indo-aryans.html
    The IVCp component should not be separated, but it can be simply part of this BMAC-like population, that shared the pre-Zoroastrian pantheon with Vedic Indo-Aryans.

    ReplyDelete
    Replies
    1. Thanks for the perspective.

      Wrt, tepe hissar like ancestry in IVCp samples from SiS and Gonur, Maier et al 2022 preprint has confirmed ~7% anatolian ancestry in those samples, overturning conclusions of Narasimhan et al and shinde et al.

      I also confirm the same in my 2previous blog posts on Rakhigarhi and IVCp, except I note that Levant PPN also works instead of Anatolian . The population which mediated this ancestry into India is likely something like Anau chalcolithic. The inflow could have happened between 5000-4000bce.

      Delete
  44. Added another section

    The Evidence from Archaeology and Assyrian texts

    That these newcomers to NW Iran are proto-Iranians is completely borne out by archaeological and Assyrian textual evidence. From Iranica Online:
    The Medes are for the first time attested in 836 B.C. in an Assyrian inscription of King Salmanasar III (who fought against the Mataí).

    The Persians are first attested some years earlier (843 B.C., which is the date of the first definite evidence for Iranian settlement in Iran proper), under the name Parsuaš (Iranian *Pārsva-). They apparently have to be located to the south and west of Lake Urmia.

    TC Young,1967 (18) posited that 1300-1000 BCE Iron Age 1 sites at Hasanlu V, and Sialk V are the best fits for the first Iranians in Zagros.

    The written evidence points to the east as the direction from which the Medes and the Persians came, and the Iron I ceramic tradition has roots in Hissar III and related sites in the northeast which were occupied long before Iron I grey ware appeared in western Iran. That tradition and those who bore it spread fairly rapidly over much of western Iran in numbers which perhaps match what the written sources tell us of the earliest Iranians in the Zagros. Furthermore, the areas where this pottery is found match in considerable detail those areas where we have postulated that early Iranians were found by the Assyrians.
    Douglas Adams, 2020 (19) also states it is consensus that "Proto-Iranians to the north (arriving on the edge of the Fertile Crescent [Assyrian records] by 1100 BC)"

    ReplyDelete
  45. How can one do qpAdm rotation on Linux version of Admixtools?

    ReplyDelete
  46. Great insights! We know Iranian languages and Indo-Aryan languages are sister families. Since BMAC is the Iranian vector according to you, what ancestry is the propagator of Proto-Indo-Iranian and Indo-Aryan? And why don't we see authoritative archeologists classifying Indus Valley sites as Indo-Aryan and BMAC sites as Iranian if MLBA Central Asian Steppe sites are really irrelevant to Indo-Iranian culture?

    ReplyDelete
    Replies
    1. @Legend. Thanks.

      QpAdm in Linux gives single results for single models. For rotation with less samples, you may use admixr on R by Bodkan, but the problem there is that it runs a fresh qpAdm each time rather than using qpfstats output, so time taken is in multiple hours.

      Personally, I have written a python script which runs various rotations from qpfstats output and writes it to file.

      Wrt Indo Iranian, if BMAC was Iranian, proto Aryan should be somewhere in NW india or Afghanistan.

      Wrt, 'authoratitive archaeologists', Sarianidi was the excavator of Gonur and he already proposed BMAC as Iranian based on his findings. Recently, parpola published a paper on sanauli and claimed that Sanauli was indo Aryan.

      Delete
    2. @vAsiSTha The problem with admixr is that I've tried it and honestly it's a hassle to run. I personally run Admixtools in Ubuntu in WSL in windows which is very convenient. I've tried to manually rotate the models too but that again is almost impossible to do by myself as too many unique combinations. I've seen your models are from Admixtools' linux version so if you don't mind can you help me in setting up qpAdm rotation in linux?

      I think Proto-Aryan in Afghanistan makes sense because Rig Veda 3.33
      shows Vishvamitra and Bharatas crossing Beas and Sutlej to reach Kurukshetra. So thanks for the sources for Gonur and Sinauli, I'll read them.

      Delete
  47. The material found at BMAC, including wagons, Horse burials (refer Sarmiento, Dubova, Sarianidi) (before Sintashta culture was even born), Fire rituals, Fire temples, soma equipment (refer Sarianidi), poppy and ephedra seeds (disputed by others, but refer sarianidi) and motifs of serpents and eagles, etc. all point to BMAC culture being Iranian prior to 2000bce.

    Edwin Bryant notes the views of most noted archaeologists in his book 'Quest for the Origins of Vedic culture'. There is complete absence of evidence regarding any 'Aryan migration/invasion'. All indian archaeolgists, including the legend BB Lal, and outsiders like Kenoyer, Lamberg-Karlovsky, Jim Schaffer, R Dyson etc. agreed (at least at some point, don't know if they changed views) that evidence of cultural change in north Indian archaeological record is sparse to non existent.

    The opponents to this view always has been the Anthony, Mallory, Kuz'mina camp and their friends at Harvard like Witzel. They claim that lack of cultural change in archaeological record doesn't mean anything. It's a shame that only this second set of people are given voice now, mostly due to their collaboration with Harvard in their recent papers. Furthermore, the primary marker of Vedic culture, fire rituals and fire altars (7 altars together near a bathing tank, similar to community ritual spaces seen today for post death Vedic rituals) have been found at Kalibangan and Lothal, long before 2000bce.

    For instance in Śatapatha Brāhmaṇa 12.9.2.1 describes about ritualistic bathing [8].

    "Having performed the sacrifice they betake themselves to the purificatory bath; for after a Soma-sacrifice they do betake themselves to the purificatory bath, and the Sautrâmanî is the same as the Soma (sacrifice)."

    Good article on the Kalibangan Fire altars
    https://www.indica.today/long-reads/roots-vedic-rituals-harappan-fire-worship-vedic-parallels/

    ReplyDelete
  48. @vAsiSTha Yeah actually, this is what I meant. Scholars like Anthony, Mallory, Kuz'mina and Witzel have so much clout that I've seen only them being cited. You're the only blogger which I've seen, to represent the other side. In my honest opinion, archeology and genetics hold a lot of weight. Considering that we don't see any archeological influence from MLBA Central Asian Steppe cultures into Indian subcontinent and very minor ancestry influx (which we aren't sure to happen before 1500 BCE in mainland India), it always amazed me how this proposition was emerging victorious in academia.

    ReplyDelete
  49. Also, I think it's necessary to point out that Parpola also has very weird theories in mind like how he shows 2 waves of Indo-aryans arriving into India. Jamison's review of Parpola's book was really great. I think I should share it here.

    https://lockwoodonlinejournals.com/index.php/jaos/article/view/776

    ReplyDelete
  50. @Legend

    Shoot me a mail at vikramraj12344321 (at) gmail dot com

    ReplyDelete
  51. It's likely their is confusion of Paleolithic pre-Iranian-S.Asian vestigial genes as IVC. That properly reflects east-shiftedness of early Iranians, but not necessarily, authenticate, IVC admixture. Hassanlu is doubtful to have IVC influence.

    ReplyDelete
  52. @blogmaster

    Hasanlu_IA_I6430 - This particular sample has definite IVC related admixture.

    ReplyDelete
  53. This recent paper in the Nature from 10 days back would be interesting, it seems to prove my hunch that CHG originated in the Indian subcontinent. Reading the Lazardis paper in light of this new paper hints towards AIT

    ReplyDelete
  54. https://www.nature.com/articles/s41586-023-06705-1#Fig2

    ReplyDelete
  55. https://www.researchgate.net/figure/GD1hu3a-appears-to-be-related-to-Caucasus-Hunter-Gatherers-and-to-modern-South-Asian_fig1_306017875

    Even though part of the Western
    Eurasian component found in India can be linked to Bronze Age migrations by dating the last contact using
    Linkage Disequilibrium (thus coming from the Kotias lineage), our results highlight the possibility of an older
    contribution from a source genetically close to GD13a (which would be hard to disentangle from the later gene
    flow from the Steppe). Eventually, ancient DNA from the Indus Valley will be needed to detect conclusively
    whether any genetic traces were left by the eastward Neolithic expansion from the Near East, or whether this
    process was mostly cultural.


    Also, found the above study which postulates CHG (Neolithic female from Ganj Dareh dated to caliberated 10000- 9700BCE to be the source of steppe ancestry in S Asia, much earlier back in time. Or was the direction of diffusion in reverse direction?

    ReplyDelete

No censorship unless spam