Aryan migration: what does genetics say?

12 October 2026 · By Himanshu Pandey

Why this question still matters

Ask ten people in India who the Aryans were and you will probably get ten different answers. For some the word means foreign conquerors who pushed older peoples to the south. For others it means the authors of the Vedas, who never came from anywhere because they were always here. For most, it is a word they met in a school textbook and then in a heated argument on social media.

Over the last fifteen years the debate has changed shape, because a new kind of evidence has entered it: DNA recovered from ancient skeletons. This article sets out, as plainly as possible, what that evidence shows, what it does not show, and how it fits with the older evidence from language and texts. It also explains why the ancestry figures that appear in personal DNA reports should not be read as an "Aryan" score or a caste marker.

Throughout, this article keeps three things apart that are often mixed up. The first is a language family: Indo-Aryan, a branch of the larger Indo-European family. The second is a body of culture and ritual, preserved in the Vedic texts. The third is biological ancestry, the DNA people carry. Much of the confusion in this debate, from the nineteenth century to today, comes from treating these three as if they were one.

Before genetics: a word and its baggage

The modern career of the word "Aryan" began with European philology. When William Jones and others noticed in the late eighteenth century that Sanskrit was related to Greek, Latin and the Germanic languages, they opened the comparative study of what came to be called the Indo-European family. The Sanskrit word ārya, a term of self-description in the Rigveda, was borrowed as a label for the people who first spoke these languages.

Max Müller did more than anyone to spread the term, and he is often blamed for turning it into a racial category. The record is more complicated. In 1888 he wrote that an ethnologist who spoke of an Aryan race or Aryan blood was as great a sinner as a linguist who spoke of a long-headed dictionary [1]. For Müller, Aryan was a language word. His warning was ignored. Thomas Trautmann has shown how colonial administrators, philologists and race scientists together turned a language family into an imagined race of fair-skinned conquerors, and how this idea was then used to explain caste [2].

Indian writers answered in very different ways. Bal Gangadhar Tilak accepted that the ancestors of the Vedic poets had come from outside, but argued in The Arctic Home in the Vedas (1903) that their first home lay near the North Pole [3]. B. R. Ambedkar rejected the whole theory. In Who Were the Shudras? (1946) he called the theory of invasion "an invention" and argued that the idea of an Aryan race collapsed on every point [4]. Later, a separate indigenist school argued that Indo-European languages began inside India and spread outward, a position now usually called Out of India.
— the late twentieth century, historians were working to separate the strands. Romila Thapar argued that the nineteenth-century equation of language, race and nation was a colonial construction, and that the texts point to migration and gradual change of culture, not one military conquest [5]. Michael Witzel set out the linguistic and textual case that Old Indo-Aryan came into the subcontinent from the north-west [6]. Edwin Bryant gave the indigenist arguments a fuller hearing than they usually got, while still finding the migration reading stronger [7, 8]. All of this work rested on language, texts and archaeology. Until about 2009 nobody could bring genetic evidence from ancient people into the discussion.

How geneticists read ancestry

A chromosome does not record which language a person spoke. What genetics can do is reconstruct how populations mixed and moved, and then ask whether any of those movements match, in time and place, the spread of a language that we know about from other evidence.

The first large picture of Indian population history came from David Reich, Kumarasamy Thangaraj and colleagues in 2009 [9]. Studying 25 Indian groups, they modelled most mainland populations as a mix of two inferred ancestral groups. They called these Ancestral North Indians (ANI), related to people of West Asia, Central Asia and Europe, and Ancestral South Indians (ASI), who are not closely related to any population outside the subcontinent. They estimated ANI ancestry at between 39 and 71 percent across the groups they studied, higher on average in traditionally upper-caste groups and among Indo-European speakers.

Priya Moorjani and colleagues then dated this mixing. Using data from 73 groups, they placed ANI and ASI admixture between about 1,900 and 4,200 years ago, and showed that after this period many groups turned to endogamy, so that mixing slowed sharply [10]. Analabha Basu, Neeta Sarkar-Roy and Partha Majumder later found that a wider sample needed five ancestral components to describe it, adding Austro-Asiatic, Tibeto-Burman and Andamanese strands to ANI and ASI [11].

Two cautions apply to all of this. First, ANI and ASI are statistical reconstructions. Nobody has found an ANI or ASI skeleton; they are the two ends of a mixing line worked out from living people. Second, ancestry percentages depend on the model. The main tool now used, qpAdm, tests whether a group can be described as a mix of chosen source populations, and if so, in what proportions. Change the sources and the percentages change. A figure such as "20 percent Steppe" describes how well a model fits, not a fixed quantity of anything.
— the late 2010s the two-way ANI and ASI model had been replaced by a three-way one: an indigenous South Asian hunter-gatherer ancestry, usually called AASI (Ancient Ancestral South Indian) and approximated by the present-day Andaman Islanders; an Iranian-related ancestry linked to early farmers and herders; and a Steppe ancestry from Middle to Late Bronze Age pastoralists of Central Asia.

What the ancient bones say

The real change came when DNA could be taken from ancient skeletons. The most important study so far is Vagheesh Narasimhan, Nick Patterson, Priya Moorjani and colleagues, published in Science in September 2019 [12]. It reported genome-wide data from 523 ancient individuals from Iran, Central Asia and the far north of the subcontinent. Four of its findings bear directly on the Aryan question.

First, it found eleven individuals at two sites with known trade contacts with the Indus Valley Civilisation, Shahr-i-Sokhta in eastern Iran and Gonur in Turkmenistan, whose ancestry stood out from their neighbours. They carried about 45 to 82 percent Iranian farmer-related ancestry and about 11 to 50 percent AASI-related ancestry, with no detectable Anatolian farmer ancestry. The authors called this pattern the Indus Periphery Cline and argued that it was probably typical of people in the Indus Valley Civilisation itself. These individuals had no Steppe ancestry.

Second, it dated the arrival of Steppe ancestry. The main population of the Bactria-Margiana Archaeological Complex in Central Asia had none, and the study rejected that population as a major source for South Asians. Steppe-derived ancestry first appears in outlier individuals at these sites between about 2100 and 1700 BCE. In the Swat Valley of northern Pakistan, where the study had more than a hundred individuals from about 1200 BCE onward, the authors estimated that Steppe ancestry mixed in between about 1900 and 1500 BCE. They concluded that Steppe ancestry entered South Asia in the first half of the second millennium BCE and today makes up as much as about 30 percent of the ancestry of some South Asian groups.

Third, it found that the sex balance of this mixing differed by place. In the Bronze Age and Iron Age Swat individuals, Steppe ancestry came mainly through women: only about 5 percent of their 44 Y chromosomes belonged to the R1a-Z93 lineage typical of Steppe men, against about 20 percent Steppe ancestry in the rest of the genome. In present-day South Asians as a whole the pattern is reversed, and the Steppe ancestry has come mainly through men.

Fourth, groups that see themselves as traditionally priestly, especially Brahmin and Bhumihar groups, tended to carry more Steppe ancestry than the simple ANI and ASI model would predict. The authors read this as a second line of evidence linking Steppe ancestry with the custodians of Sanskrit texts, while warning that it was not a formal test and that there are exceptions.

The authors also noted a difficulty that historians should take seriously: there is very little shared material culture between the Central Asian steppe and South Asia in this period. They pointed to the Beaker culture in Europe as an example where people with Steppe ancestry moved in large numbers while adopting much of the local material culture.

The Rakhigarhi episode

Published the same week was the first genome from a burial inside the Indus Valley Civilisation. Vasant Shinde, Vagheesh Narasimhan and colleagues reported DNA from one woman buried at Rakhigarhi in Haryana [13]. Her ancestry matched the Indus Periphery pattern: Iranian-related and AASI-related, with no Steppe ancestry. The paper also argued that her Iranian-related ancestry separated from that of the early farmers and herders of the Iranian plateau before farming began there. If that is right, Indian agriculture did not arrive with a large movement of farmers from the west.

What followed is worth recording. At a press conference in Delhi in September 2019, the lead author presented the findings as setting aside the Aryan migration theory, and much of the Indian press reported them that way. Some of his co-authors publicly disagreed, and Shoaib Daniyal of Scroll.in pointed out that the companion Science paper argued the opposite [14]. In the same article, Vagheesh Narasimhan, Niraj Rai and Nick Patterson are reported as rejecting parts of the press statement. In fact, a Harappan genome without Steppe ancestry is exactly what the migration model predicts, since the Steppe ancestry arrived after the mature Harappan period. The lesson for readers is simple: read the paper, not the press conference.

The wider Eurasian picture

The South Asian evidence fits a pattern seen across Eurasia. Wolfgang Haak and colleagues showed in 2015 that the Corded Ware people of Neolithic Germany drew about three-quarters of their ancestry from Yamnaya-related pastoralists of the Pontic-Caspian steppe, which points to a large movement of people into Europe in the third millennium BCE [15]. Morten Allentoft and colleagues reached similar conclusions from Bronze Age genomes across Eurasia [16]. Peter de Barros Damgaard and colleagues showed that the earliest known horse herders, at Botai in Kazakhstan, were not related to the Yamnaya, and that the Anatolian branch of Indo-European, which includes Hittite, cannot be explained by a simple steppe migration [17].

Most recently, Iosif Lazaridis and colleagues studied 435 ancient individuals and traced the Yamnaya and Anatolian-speaking peoples to a common ancestral population in the Caucasus and Lower Volga region in the fifth millennium BCE [18]. In their model, later Yamnaya expansions carried the non-Anatolian branches of Indo-European, including Indo-Iranian, outward. Narasimhan's team also noted that the Steppe ancestry in South Asia has the same profile as that in Bronze Age Eastern Europe, which may explain some features shared by the Indo-Iranian and Balto-Slavic languages [12]. For the archaeology behind the steppe model, David Anthony's book remains the fullest account [19]; for an accessible summary from the genetics side, see David Reich [20].

What we still do not know

The genetic evidence is stronger than it was ten years ago, but it is narrower than the headlines suggest. Several questions remain open.

The origin and date of the Iranian-related ancestry in South Asia are not settled. The Rakhigarhi paper argued for a very old lineage, present before farming. The 2025 whole-genome study by Elise Kerdoncuff and colleagues, based on 2,762 people in the LASI-DAD cohort, describes the source as Neolithic farmers related to Iranian and Central Asian cultures [25]. Both are statements about which sampled ancient group is the closest match, and new samples could change them.

Genetics can date the arrival of Steppe ancestry. It cannot date the arrival of a language. The link between the two is an argument built from timing, geography and linguistics, and it should be presented as such.

The largest gap is geographic. There is still no published ancient DNA from the Ganga plain, and none from skeletons firmly dated to the Rigvedic period. The Swat Valley lies at the north-western edge of the subcontinent, and Rakhigarhi has given us one genome. Claims about the heartland of early Vedic culture rest on samples from its margins and on reconstructions from living people, whose ancestry has been shaped by more than two thousand years of endogamy.

The other side of the argument

The migration reading is now the majority view, but the case against it deserves to be stated fairly.

The strongest scientific challenge came from two Y-chromosome studies published in 2006, before ancient DNA from the region existed. Sanghamitra Sengupta and colleagues analysed Indian Y-chromosome lineages in detail and concluded that the influence of Central Asian pastoralists on the existing gene pool was minor [21]. Sanghamitra Sahoo and colleagues reached a similar conclusion, arguing that the data did not support any major movement of people from the north-west tied to farming or to Indo-Aryan languages [22]. These were serious papers in good journals, and indigenist writers relied on them heavily. In 2015 Peter Underhill and colleagues mapped the R1a lineage in detail, dividing it into a mainly European branch (Z282) and a Central and South Asian branch (Z93), and suggested that its earliest diversification took place near present-day Iran [23].

The field changed its view for two reasons. The 2006 studies estimated history from the diversity of lineages in living men, a method that can mistake an old lineage for an old local presence. Once DNA from dated ancient skeletons became available, R1a-Z93 and the wider Steppe signal appeared in the region late and from the north. At the same time, Marina Silva and colleagues combined maternal, paternal and genome-wide data and found that the maternal ancestry of South Asians is largely old and local while part of the paternal ancestry reflects Bronze Age input from Central Asia [24]. Narasimhan's study then found R1a-Z93 in Bronze Age and Iron Age Swat, and none in the Indus Periphery individuals [12].

The Out of India position still has supporters among Indian scholars. To be convincing, it has to explain why the oldest branch of Indo-European, Anatolian, is found far from India; why early Indo-Aryan words appear in the Mitanni records of northern Syria in the second millennium BCE; and why ancient DNA shows Steppe ancestry moving into South Asia rather than out of it. No peer-reviewed ancient DNA study so far supports an outward movement of this ancestry from India. On present evidence the migration reading is much better supported, with the open questions noted above.

What this means for your own DNA result

A whole-genome report may show a figure for Steppe ancestry, or a breakdown into AASI-related, Iranian-related and Steppe components. Here is how such numbers should be read.

A Steppe percentage is not an Aryan score. Aryan is a word for a language and a cultural world, not a genetic group. The people who brought Indo-Aryan speech carried a mix of ancestries, the people they met carried others, and over time the language spread much further than the genes. How much Steppe ancestry you carry says nothing about whether your ancestors spoke an Indo-Aryan language.

It is not a caste marker either. It is true that, on average, some traditionally priestly groups carry more Steppe ancestry, and that ANI ancestry is higher on average among upper-caste groups [9, 12]. These are averages across groups, with wide overlap and many exceptions. They do not allow anyone to read caste from a genome or rank individuals. The dating work of Moorjani and Basu also shows that the endogamy which preserved these differences is younger than the mixing it froze in place [10, 11].

The percentage depends on the model. The same person can be described with different proportions depending on which reference populations are used, as the move from the ANI and ASI model to the three-source model shows.

The broadest finding is one of shared descent. Kerdoncuff and colleagues found that most genetic variation in India goes back to a single migration out of Africa about 50,000 years ago, and that a later shift towards endogamy produced strong founder events, which shape much of the genetic difference between Indian communities today [25]. Nearly everyone in the subcontinent is a mix of the same few ancient sources in different proportions.

References

[1]
Müller, F. Max. Biographies of Words and the Home of the Aryas. London: Longmans, Green & Co., 1888.
[2]
Trautmann, Thomas R. Aryans and British India. Berkeley: University of California Press, 1997.
[3]
Tilak, Bal Gangadhar. The Arctic Home in the Vedas. Poona: Tilak Bros., 1903.
[4]
Ambedkar, B. R. Who Were the Shudras? Bombay: Thacker & Co., 1946.
[5]
Thapar, Romila. "The Theory of Aryan Race and India: History and Politics." Social Scientist 24, no. 1/3 (1996): 3–29.
[6]
Witzel, Michael. "Autochthonous Aryans? The Evidence from Old Indian and Iranian Texts." Electronic Journal of Vedic Studies 7, no. 3 (2001): 1–115.
[7]
Bryant, Edwin. The Quest for the Origins of Vedic Culture: The Indo-Aryan Migration Debate. New York: Oxford University Press, 2001.
[8]
Bryant, Edwin, and Laurie L. Patton, eds. The Indo-Aryan Controversy: Evidence and Inference in Indian History. London: Routledge, 2005.
[9]
Reich, David, Kumarasamy Thangaraj, Nick Patterson, Alkes L. Price, and Lalji Singh. "Reconstructing Indian Population History." Nature 461 (2009): 489–494.
[10]
Moorjani, Priya, et al. "Genetic Evidence for Recent Population Mixture in India." American Journal of Human Genetics 93, no. 3 (2013): 422–438.
[11]
Basu, Analabha, Neeta Sarkar-Roy, and Partha P. Majumder. "Genomic Reconstruction of the History of Extant Populations of India Reveals Five Distinct Ancestral Components and a Complex Structure." Proceedings of the National Academy of Sciences 113, no. 6 (2016): 1594–1599.
[12]
Narasimhan, Vagheesh M., et al. "The Formation of Human Populations in South and Central Asia." Science 365, no. 6457 (2019): eaat7487.
[13]
Shinde, Vasant, et al. "An Ancient Harappan Genome Lacks Ancestry from Steppe Pastoralists or Iranian Farmers." Cell 179, no. 3 (2019): 729–735.
[14]
Daniyal, Shoaib. "Aryan Migration Theory Upheld by Two New Genetic Studies. Why Did Indian Media Report the Opposite?" Scroll.in, 12 September 2019.
[15]
Haak, Wolfgang, et al. "Massive Migration from the Steppe Was a Source for Indo-European Languages in Europe." Nature 522 (2015): 207–211.
[16]
Allentoft, Morten E., et al. "Population Genomics of Bronze Age Eurasia." Nature 522 (2015): 167–172.
[17]
de Barros Damgaard, Peter, et al. "The First Horse Herders and the Impact of Early Bronze Age Steppe Expansions into Asia." Science 360, no. 6396 (2018): eaar7711.
[18]
Lazaridis, Iosif, et al. "The Genetic Origin of the Indo-Europeans." Nature 639 (2025): 132–142.
[19]
Anthony, David W. The Horse, the Wheel, and Language: How Bronze-Age Riders from the Eurasian Steppes Shaped the Modern World. Princeton: Princeton University Press, 2007.
[20]
Reich, David. Who We Are and How We Got Here: Ancient DNA and the New Science of the Human Past. New York: Pantheon, 2018.
[21]
Sengupta, Sanghamitra, et al. "Polarity and Temporality of High-Resolution Y-Chromosome Distributions in India Identify Both Indigenous and Exogenous Expansions and Reveal Minor Genetic Influence of Central Asian Pastoralists." American Journal of Human Genetics 78, no. 2 (2006): 202–221.
[22]
Sahoo, Sanghamitra, et al. "A Prehistory of Indian Y Chromosomes: Evaluating Demic Diffusion Scenarios." Proceedings of the National Academy of Sciences 103, no. 4 (2006): 843–848.
[23]
Underhill, Peter A., et al. "The Phylogenetic and Geographic Structure of Y-Chromosome Haplogroup R1a." European Journal of Human Genetics 23 (2015): 124–131.
[24]
Silva, Marina, et al. "A Genetic Chronology for the Indian Subcontinent Points to Heavily Sex-Biased Dispersals." BMC Evolutionary Biology 17 (2017): 88.
[25]
Kerdoncuff, Elise, et al. "50,000 Years of Evolutionary History of India: Impact on Health and Disease Variation." Cell 188, no. 13 (2025): 3389–3404.