The USA has been in Afghanistan for over 10 years now. Like many Americans my own personal preference is that we get out as soon as possible. Because of American involvement we see terms like “Pashtun” bandied about in the media, but there is little further exploration. But politics and international relations are the not focus of this post, at least not politics and international relations in our time. A new paper in PLoS ONE examines the Y-chromosomal patterns as they partition across ethnic groups in Afghanistan. By this, we mean the direct paternal lineage of Afghan men. Additionally, the authors place the results in a broader Eurasian context. The results are not surprising, though they add greater precision and power to our picture because of their sample size. The main downside is that they did not include mtDNA (maternal lineage) or autosomal analysis (the total ancestry, not just the paternal or maternal line).
At this point most Americans should in theory have a general sense of Afghan ethnography. But let’s go over it again. First and foremost you have Pashtuns, who are a broad coalition of tribes who are Sunni Muslims, and speak East Iranian languages. The Tajiks are nominally non-tribal Sunni Muslims who speak a variant of Persian (Dari). The Hazara are Shia Muslims who also speak a variant of Persian (Dari). Finally you have Uzbeks, who are Turkic Sunni Muslims. It is visibly clear that the Uzbeks and Hazara are admixtures between West Eurasian and East Eurasian populations, though the Uzbek language should also make that an obvious likelihood. The Hazara claim an origin as descendants of Mongol refugees who fled Iran after the fall of the Il-Khan regime; the genetics does support his. The Uzbek identity is somewhat confused insofar as the ethnonym “Uzbek” is actually relatively new as a term which covers a range of Turkic populations in southern Central Asia (see “Sart”). In regards to the Pashtuns and Tajiks, despite their common religion and Iranian language, the two are distinguished strongly due to a very divergent history. A cut-out would be that the Pashtuns are part of greater South Asia and its cultural sphere; the Kabul valley were dominated by Hindu-Buddhist dynasties before the Muslim conquest. In contrast, the Tajiks are heirs to a long standing Persian cultural presence in Central Asia, what was once termed Turan. The fact that they are Sunni Muslims rather than Shia is a quirk of history. In the 16th and 17th centuries the Safavid dynasty of Iran (which was culturally Turkic) converted Persia and Persians from a predominantly Sunni domain and population to an exclusively Shia one (the main exceptions in Iran today are ethnic minorities such as Kurds and Baloch). But the Persians of Central Asia were under Sunni Turkic hegemony, and so maintained their ancestral religion (there seem to have been no continuous Zoroastrian communities in Central Asia, in contrast to Iran). It is also notable that Dari exhibits some archaic features.
The main results of the paper are illustrated in this figure:
What you see here is that an isolation-by-distance model does not predict the Y-chromosomal variation in Afghanistan. Hazara and Uzbeks do not cluster with Tajiks or Pashtuns, their neighbors, presumably because they have recent East Eurasian ancestry. This is not so surprising. The Uyghurs are a similar population, in the center of Eurasia, and geographically midway between East and West Eurasians. But a close examination of patterns of genomic variation indicates that the Uyghurs are the products of recent admixture (~2,000 years). To my knowledge no such analysis has been performed on Uzbeks or Hazara, but I am willing to bet $400 against $40 for someone taking the other side that they too are recent admixtures. The history here is is clear. Central Asia was dominated by Iranian populations up to ~2,000 years ago. Then pulses of nomadic populations began to issue out of the Altai region; the Turks. Though today there remain a residual non-Turkic population in Central Asia, the Tajiks being the most numerous, it is primarily a Turkic domain. But the physical features of Central Asian Turks indicate clear non-East Eurasian ancestry, almost certain the Iranian substrate of Turan (apparently the Turkic dialects of Central Asia have specifically Iranian features as well in terms of lexicon).
The same dynamics obviously apply in Afghanistan. Only a massive folk wandering can explain why the Hazaras, in the middle of Afghanistan, exhibit a large dollop of the Genghis Khan haplotype. The Uzbeks are the bleeding edge of a wave of demographic advance which has been inexorably sweeping out of northeast Asia for nearly 2,000 years. This is important in the larger scale, because it is illustrative of a tendency where continuous clines can crash and burn due to the power of human culture to mix & match, and, transplant and translocate. As one moves from the Kabul Valley into North or North-Central India the changes genetically are relatively mild (at least on the Y-chromosome) in comparison to that which occurs as one pushes into the highlands of central Afghanistan, or to the norther marches which have been populated by Uzbeks. That is because for thousands of years the null isolation-by-distance dynamic had been operative across the expansive of greater South Asia. Before the arrival of the Turks one might suppose, with some qualifications, that Iran, Turan, and Hind, exhibited a cultural and genetic wholeness in continuity (Puranic Hinduism and Zoroastrianism are both arguably derived forms of one strain of Aryan religion). But the intrusion of a Turkic population, alien linguistically and genetically, disrupted this continuous gradient. An isolation-by-distance model becomes useless without the information of anthropology and history.
When attempting to construct a taxonomy of human relationships I think it is important to distinguish between the alternative dynamics which have been operative in generating the palimpsest of human genetic variation. Isolation-by-distance and clinal gradation is highly informative in many cases (e.g., North European plain, the North Indian plain, much of China). But there are also many specific instances when historical and geographically contingencies are such that one is confronted by genetic chasms (e.g., across the Pamirs, or across the Bab-el-Mandeb). Both cases are true, and part of the broader picture. But they are not the total picture alone.
Related: Dienekes has some related comments. The finding that Afghan R1a1a is of the South Asia, and not East European, clade suggests to me that R1a1a arrived with West Asians who brought the dominant package of “Ancestral North Indian” to South Asia.