(Last Update: Dec 21)
Over at the Dodecad blog, I have been tracking a mysterious "Dagestan" genetic component in South Asia and Europe. This component is modal in populations of Dagestan: Dargins from Urkarah, Lezgins, and Kumyks from Stalskoe.
Dargins and Lezgins are Northeast Caucasian speakers, and while Kumyks are Turkic, this is probably due to a small East Eurasian component in their ancestry, and it's a fair guess that they too are natives to the region who underwent language shift.
Surprisingly, this component occurs at a high frequency in some South Asian populations, including Telugu and Tamil Brahmins from South India. These are believed to be descended from Indo-Aryan speakers from North India and to have maintained a genetic distinctiveness vis a vis the native inhabitants of South India.
There are two reasons why I think I am picking a really meaningful signal:
- Differentiation between Tamil and Andhra Pradesh Brahmins and non-Brahmin populations from the same states
- The levels of the component in India are as high as any other Eurasian populations save for the Dagestanis immediate neighbors in the Caucasus
Indeed, if you inspect the project members' results on this component, you will see that the "Dagestan" component is found at many South Asians who belong to other populations than those available in the public references. Without giving anything away, it's clear that the "Dagestan" component is found frequently in many parts of India.
Indo-Aryans outside India
The question of the origin of the Indo-Aryans is obviously very difficult, and attempts at tracing their origin use either a type of "geolinguistic trigonometry" to make them fit in the broader context of Indo-European dispersals, or rely on archaeological interpretation of the material culture of sites such as Arkaim or the BMAC.
I'll leave these speculations to those competent to make them, and I will turn to the concrete: the Mitanni of Syro-Anatolia. Here we have what seems to be Indo-Aryan personal/deity names, as well as numerals in what appears to have been a largerly Hurrian-speaking population.
The Kingdom of the Mitanni flourished in the 2nd millennium BC, a time also generally considered to be that of the earliest Indo-Aryan linguistic monument, the Rigveda.
What is most interesting, however, is that Hurrian, like Urartian are part of the Hurro-Urartian language family, which has been linked by some linguists to Northeast Caucasian.
I won't comment on whether such a link exists, but a relationship between Hurro-Urartian and Northeast Caucasian coupled with the known proximity of Indo-Aryans with Hurrians would immediately supply an explanation for the "Dagestan" component: it might be the legacy of an absorption of Hurrian elements by the ancestors of the Indo-Aryans while the latter were still in the Near East.
By the 2nd millennium BC, the Indo-Aryan element seems to have been well on its way to disappearance in the region, and we find no trace of it after the demise of the Mitanni. A millennium later came the disappearance of Urartian, replaced by various types of Indo-European (such as Armenian and Iranian), and Semitic. We are fortunate that the Indo-Aryans of Syro-Anatolia left traces of their existence before their demise.
The Bactria-Margiana Archaeological Complex was excavated by Viktor Sarianidi (here is a trailer for a documentary about it, and a Discover magazine article). According to Sarianidi, it is at least Indo-Iranian (if not specifically Indo-Aryan) and can be traced to the Syro-Anatolian region. The formation of the BMAC begins in the 3rd millennium BC, and it ends its existence in the 2nd, at around the time when Indo-Aryans are said to have made their appearance in the subcontinent. The geographical region of the BMAC (aka Oxus River civilization) is pictured on the left.
Whether we trace it to Anatolia, or, more modestly, to Iran, the likely western origin of the BMAC would almost certainly mean that it could have served as a conduit for the dispersal of the mystery "Dagestan" component to India.
Non-Indo-Europeans of Pakistan
The scenario I have outlined above derives the Indo-Aryans from the Transcaucasus where they pick up the "Dagestan" component, south of the Caspian, to the Oxus River civilization (BMAC) of Turkmenistan, to Afghanistan, and then via the Khyber Pass to Pakistan. The Mitanni are seen as Indo-Aryans who "stayed behind" and became thoroughly Hurrianized in the 2nd millennium BC.
We have one piece of evidence of the association of this component with Indo-Aryans: its presence in Brahmins and absence from low caste and tribal groups.
But, there is a different source of evidence that can potentially complete the argument: the non-Indo-European speakers of Pakistan: the Dravidian Brahui and the Burushaski speakers.
My theory predicts that they should have less of this component than Indo-Aryan and Iranian speakers from Pakistan such as Pathans, Sindhi, and Balochi. To test this, I repeated my South Asian experiment, but this time I added these two populations as well.
Below are the results:
In comparison to my previous experiment, the East Asian and Southeast Asian components have "folded back" into one, and the Burusho isolate has formed its own cluster.
But, here are the numbers that clinch the argument: the proportions of the Dagestan component #10 in South Asia:
As you can see, the component is minimized in tribals, low castes, and non-Indo-European groups of Pakistan (Burusho and Brahui).
I am not going to bet that all the details presented in this scenario are correct. But, this theory seems to make sense of many different pieces of evidence (such as the presence of Indo-Aryans in the Near East) and to harmonize with the genetic evidence.
UPDATE (Dec 21)
Since I've posted analysis of Europeans and South Asians separately, I decided to do a joint analysis.
Admixture proportions can be found in the spreadsheet. The "Dagestan" component is #1 (black) in this figure.
Here is the table of Fst distances between components:
The presence of component #1 in Uygurs and in Pathans may mean that it may have been carried by Indo-Iranians rather than specifically Indo-Aryans.
It would be wonderful to have samples from Kazakhstan, Turkmenistan, Afghanistan, Uzbekistan, or Tajikistan to fill the Central Asian gap between the Caucasus and China/South Asia. So, if you've chanced upon any people from those countries who happen to have tested with 23andMe, feel free to let them know about the Dodecad submission opportunity.