The Tarim Basin and the European Presence in Ancient China

There is a seductive quality to a single, well-publicized genetic paper. It arrives with the authority of a Nature headline, gets compressed into a Reuters paragraph, and suddenly a century of archaeology and linguistics is supposed to fold itself quietly into a footnote. The 2021 Nature study of the Tarim Basin mummies has been wielded precisely that way, as if one dataset, drawn from thirteen individuals, had closed the case on Indo-European presence in western China. It has not. What it has done, if read carefully, is relocate the question. And in doing so, it has inadvertently strengthened the very thesis its popularisers claim to have buried.

Let me be precise about what the 2021 paper actually showed. Genome-wide data from the earliest securely dated Tarim individuals, roughly 2100 to 1700 BCE, revealed a highly isolated, autochthonous ancestry: approximately 72 percent Ancient North Eurasian and 28 percent Ancient Northeast Asian, with no detectable contribution from the Afanasievo or Yamnaya steppe populations, nor from the Bactria–Margiana oasis cultures. The conclusion drawn was that these particular individuals were not recent migrants from the Pontic steppe.

That is a finding about thirteen people buried in particular boats at particular sites. It is not a finding about the entire linguistic and demographic history of the Tarim Basin across three millennia. The leap from 'the earliest mummies were not steppe migrants" to "Indo-European peoples never reached this region" is not a scientific inference. It is a rhetorical one.

Here is the uncomfortable fact the paper's own authors acknowledge: the Dzungarian Basin populations, a few centuries earlier, carried substantial Afanasievo-related ancestry. The Afanasievo culture is the easternmost expression of the early Indo-European steppe horizon, the same cultural-genetic complex that, on the western arc, fed into the Yamnaya expansion into Europe. When you find Afanasievo ancestry in a neighbouring basin, you have found Indo-European-associated peoples sitting at the doorstep of the Tarim. The 2021 data do not remove them. They merely say the specific individuals in the Xiaohe-era boats were not the ones carrying that signal.

This is the point that the popularised version of the study consistently mangles. Human geneticists measure ancestry. Linguists measure language. These are not the same quantity, and they do not move at the same speed or in the same packages.

A population can adopt a language: wholesale, generation after generation, without leaving a detectable genetic trace, if the demographic pressure is cultural rather than reproductive. This is not a marginal phenomenon; it is the historical norm. The Celts of Britain, the Turkic peoples of Anatolia, the Spanish speakers of Latin America, all are cases where language and culture spread in ways that outran, or only partially overlapped with, genetic replacement. The Tarim Basin is no different.

The Tocharian language is the decisive witness here. Tocharian is an extinct branch of Indo-European, attested in manuscripts from the fifth to tenth centuries CE in the northern Tarim oases. It is not a dialect of some neighbouring language; it is its own branch, with distinctive phonological and grammatical features. Linguists have long noted that its divergence from other Indo-European branches is consistent with an early split, meaning the speakers of Proto-Tocharian must have separated from the rest of the Indo-European-speaking world at a very early date, before the major western branches had fully differentiated. That is precisely the signature you would expect from a group that moved east early and spent millennia in relative isolation.

The Afanasievo presence in the Dzungarian Basin provides the entry route. The Tocharian language provides the linguistic witness. The fact that the earliest mummies themselves were not steppe migrants does not contradict either. A millennium separates the earliest mummies from the written Tocharian record, a millennium in which a language could easily have moved, spread, and been adopted by populations whose ancestry remained locally dominant.

The 2021 study is at its most persuasive when it describes the Tarim Basin as a crossroads, a place where people, languages, and technologies circulated continuously. And that description, far from undermining the Indo-European thesis, is precisely what it requires.

Consider the material record the paper itself cites: cattle, sheep, wheat, barley, millet, and dairy residues. These are not the possessions of a hermetically sealed population. They are the signature of the Inner Asian Mountain Corridor, a network of exchange that linked the steppe, Central Asia, and the oases of the Tarim for millennia. Wheat and barley came from the west. The dairy pastoralism that left its mark in the dental calculus of the Xiaohe people is a steppe tradition. The weaving techniques and burial customs echo broader Eurasian pastoralist practices.

A population that is genetically isolated can still be culturally porous. The Tarim people adopted the technologies and practices of their neighbours, including, almost certainly, their languages. The 2021 data show that the founding population was not a simple transplant from the West. They do not show that the West never touched them. A crossroads, by definition, is a place where things pass through.

The story becomes even clearer as you move forward in time. The Iron Age and historical populations of Xinjiang show substantial admixture, including steppe-related ancestry associated with Iranian-speaking groups such as the Saka. The Saka were horse-riding pastoralists of the Central Asian steppe, speakers of an Eastern Iranian language, and they left their mark across the region. By the first millennium BCE, the Tarim Basin was demonstrably inhabited by, and mixed with, Indo-European-speaking peoples. The Saka presence is not contested; it is documented.

So the full arc is this: an early, genetically isolated Bronze Age population in the Tarim Basin proper; Afanasievo-related steppe peoples in the neighbouring Dzungarian Basin; a millennium of cultural exchange along the Inner Asian Mountain Corridor; the appearance of Tocharian, an early-diverging Indo-European language, in the northern oases; and by the Iron Age, substantial Iranian-speaking admixture. The Indo-European footprint on the region is not a hypothesis awaiting confirmation. It is an established, multi-layered fact. The 2021 study refines the timing and the vectors. It does not refute the presence.

The stakes here are not merely academic. They are political, and they are identical on both sides of the argument. Chinese official historiography has long emphasized continuous settlement and cultural continuity within the borders of the modern state; the mummies' non-East-Asian appearance and the existence of an Indo-European language in the basin have been downplayed or reframed in state-aligned accounts. On the other side, some Uyghur nationalist narratives have claimed the mummies as ancestral, despite the fact that modern Uyghurs do not model as direct descendants of the Bronze Age Tarim genetic cluster. Both projects require a pristine, static origin story. Both are contradicted by the actual record.

The truth is messier, and it is more interesting. Populations moved. Languages travelled. Technologies crossed ecological barriers. Isolation was the exception, not the rule. The Tarim Basin was not a sealed chamber where an autochthonous people sat untouched for millennia; it was a crossroads where a genetically distinctive founding population met, absorbed, and transmitted the currents of the wider Eurasian world. That is what the archaeology says. That is what the linguistics says. And, read correctly, that is what the genetics says too.

A single study, no matter how prestigious, does not overturn a century of convergent evidence. The 2021 Nature paper is a valuable contribution, it correctly identifies the earliest Tarim individuals as locally derived rather than recent steppe migrants. But it does not, and cannot, erase the Indo-European presence in the region. That presence is attested by the Afanasievo ancestry in the neighbouring basin, by the early-diverging Tocharian language in the oases, by the Iranian-speaking Saka of the Iron Age, and by the continuous cultural exchange that the paper itself documents.

The Tarim mummies remain what they always were: a striking, unsettling window into a world far more mobile and interconnected than our founding myths allow. The people who want to bury that evidence, and the people who want to distort it, are both wrong. The truth is that Indo-European peoples and languages did reach this region, and the genetics, properly read, do not refute it. They refine it.

Zhang, F., Ning, C., Scott, A., et al. (2021). The genomic origins of the Bronze Age Tarim Basin mummies. Nature, 599, 256–261.