Two studies published in 2019 changed this old argument. Together they show that a new ancestry component, derived from pastoralists of the Eurasian steppe, entered South Asia in the second millennium BCE — after the mature Harappan cities had declined — and that a Harappan individual from Rakhigarhi did not carry it.[1][2] That much is agreed among geneticists. What the genomes cannot tell you is what language anyone spoke, or whether anything like an “invasion” happened. Both limits matter as much as the finding.
The evidence
Narasimhan et al., Science (2019). Over 500 ancient genomes from Central and South Asia. The paper finds steppe-derived ancestry entering South Asia in the second millennium BCE, after the Harappan decline. In living South Asians that component is distributed on a gradient that correlates both with geography and with traditional priestly status.[1] Consensus among geneticists
Shinde et al., Cell (2019): the Rakhigarhi genome. One individual from Rakhigarhi in Haryana, the largest known Harappan site. The genome shows no steppe ancestry. Its ancestry mixes South Asian hunter-gatherer ancestry with an Iranian-related component from a lineage that split off before farming existed on the Iranian plateau. The paper’s title — “An Ancient Harappan Genome Lacks Ancestry from Steppe Pastoralists or Iranian Farmers” — carries its main point: farming in South Asia was not brought by migrating Iranian farmers.[2]
The sample is the paper’s weakness, and it should be stated plainly: 61 skeletons were screened and exactly one yielded usable DNA, at low coverage.[2] One genome is one person, and conclusions about “the Harappans” from it should be held with that in mind.
The deep population. Most modern South Asian ancestry comes from the people who were already there — the hunter-gatherer and Iranian-related ancestry of the farming and Harappan populations — not from later arrivals.[1][2] Kerdoncuff et al. (Cell, 2025), with about 2,700 whole genomes from living Indians, now tests that three-way ancestry model at scale.[3] Consensus
The independent lines it fits. The genetic result was not the first evidence for movement of Indo-Aryan speakers; it is the latest line to agree with it:
- Language. Sanskrit sits inside the Indo-Iranian branch of Indo-European, and Vedic deva / asura inverts exactly into Avestan daēva “demon” / ahura “lord” — the trace of two communities that once shared a pantheon.[4]
- Inscriptions. The Mitanni treaty of c. 1350–1320 BCE names Mitra, Varuṇa, Indra and the Nāsatyas in Syria, two centuries before anything written in India.[5][6]
- The horse. No securely identified horse, spoked wheel or chariot comes from mature Harappan levels, and the domestic horse lineage only began spreading across Eurasia c. 2200 BCE; the Rigveda is saturated with horses and chariots.[7][8]
- The text’s own map. The Rigveda knows the Punjab well and the Ganga barely.[9][10]
What the DNA does not show
It carries no language. Steppe ancestry does not by itself prove that Indo-Aryan speech arrived with it. Combined with the linguistic evidence, that is the strongly favoured inference — but it is an inference.[1] Majority
It shows no invasion. The genetic signal is demographic, spread over centuries. It says nothing about violence, conquest or displacement. The nineteenth-century “Aryan invasion” — chariot armies destroying Harappan cities — has been rejected by essentially all scholars for decades. The Mohenjo-daro “massacre” skeletons show no battle trauma and come from different periods; there is no destruction layer at any Harappan city.[11][7] The cities emptied gradually from c. 1900 BCE amid aridification, several centuries before the migration.[12][13] Consensus
It is not a claim about belonging. Ancestry proportions say nothing about anyone’s authenticity, rank or right to their tradition. The racial “Aryan” of colonial and European theory was a misuse of what began as a linguistic term.[14]
A short timeline
| Date | What happens | Confidence |
|---|---|---|
| Pre-farming | Iranian-related lineage in South Asia splits from later Iranian farmers[2] | Majority |
| c. 2600–1900 BCE | Mature Harappan cities; Rakhigarhi individual, no steppe ancestry[2] | Consensus |
| c. 2200 BCE | Domestic horse lineage begins spreading across Eurasia[8] | Majority |
| c. 1900 BCE | Harappan de-urbanization begins; no destruction layer[7] | Consensus |
| c. 1900–1400 BCE | Mainstream window for Indo-Aryan speakers entering the northwest[10] | Majority |
| 2nd millennium BCE | Steppe ancestry enters South Asia[1] | Consensus |
| c. 1500–1200 BCE | Rigveda’s Family Books composed in the Punjab[15] | Majority |
| c. 1350–1320 BCE | Mitanni treaty names Vedic gods in Syria[5] | Consensus |
| 2019 | Narasimhan et al.; Shinde et al. published[1][2] | Consensus |
What scholars disagree about
Political — this argument is shaped by modern politics on every side, and it deserves the strongest version of each position.
The mainstream migration model. Indo-Aryan speakers entered northwest South Asia c. 1900–1400 BCE as one branch of a wider Indo-Iranian dispersal, mixing with existing populations. Its case rests on the Mitanni evidence, the nesting of Indo-Iranian inside Indo-European (which requires a homeland outside India unless every other branch migrated out), the horse and chariot, and now the steppe signal present later and absent at Rakhigarhi.[10][16][1]
The indigenist position (Out of India). Indo-Aryan is native to South Asia and the dispersal ran outward. Its strongest points are real: the Rigveda preserves no memory of a foreign homeland or arrival; if the Sarasvatī is a mighty Ghaggar-Hakra, composition may be earlier than the standard chronology allows; the migration lacks a clear archaeological signature; and colonial Aryan theory was entangled with racial ideology.[17][4] On the genetics specifically, indigenists stress what the mainstream also concedes: DNA cannot identify a language, and a minority ancestry component entering a large population does not by itself show which way a language spread. The mainstream reply is that the genetic, linguistic and epigraphic lines point the same way, and that this convergence is the argument.[18] Contested
A third position that is neither. Since 2023 the simple steppe model has a serious mainstream challenger. Heggarty et al. (Science, 2023) date Proto-Indo-European to roughly 8,100 years ago and propose a hybrid model with a southern route via the Iranian plateau alongside the steppe route. Lazaridis et al. (Nature, 2025) trace the Yamnaya themselves largely to a Caucasus–Lower Volga population. Neither is indigenist; both are peer-reviewed; both are contested.[19][20] Contested
Where this series lands. The migration model, because the linguistic and epigraphic evidence is the strongest kind available and the genetics is consistent with it — while treating the scale, mechanism and exact timing of the movement as open.[1][10]
Common misconceptions
- “Rakhigarhi disproved the Aryan migration.” The Rakhigarhi genome lacks steppe ancestry, which is what the migration model predicts for a Harappan who lived before the steppe component arrived.[2]
- “The DNA proves an Aryan invasion.” It shows movement and mixing over time, not conquest.[1]
- “Modern South Asians descend mainly from the migrants.” The great majority of South Asian ancestry comes from earlier populations; the steppe component is one part of a mix.[1]
Why it matters
The genetics settles less than either side sometimes claims, but it does settle something: people and ancestry moved into South Asia after the Harappan peak, and the Harappans themselves did not carry that ancestry. That fits a Rigveda composed in the Punjab after the cities had emptied (§3) — while leaving Hinduism’s deeper roots, in the populations and practices that were already there (§1, §2), firmly South Asian. The full account of the methods and their limits is in §0.
References
- ^ a b c d e f g h i j Vagheesh M. Narasimhan et al. "The formation of human populations in South and Central Asia". Science 365, 2019. doi Over 500 ancient genomes. The steppe-ancestry signal, and its limits.
- ^ a b c d e f g h Vasant Shinde et al. "An Ancient Harappan Genome Lacks Ancestry from Steppe Pastoralists or Iranian Farmers". Cell 179, 2019. doi The single sequenced Rakhigarhi individual.
- ^ Elise Kerdoncuff, Laurits Skov, Nick Patterson et al. "50,000 years of evolutionary history of India". Cell 188, 2025. doi ~2,700 whole genomes; where the three-way ancestry model now gets tested at scale.
- ^ a b Edwin Bryant The Quest for the Origins of Vedic Culture. Oxford University Press, 2001. open library The fairest book-length survey of the origins debate; takes the indigenist arguments seriously without endorsing them.
- ^ a b Mitanni–Hittite treaty — Boğazköy archive. c. 1350–1320 BCE. Akkadian, with Indo-Aryan divine names. Invokes Mitra, Varuṇa, Indra and the Nāsatyas as witnesses — Indo-Aryan gods attested in writing in Syria two centuries before anything written in India. about
- ^ Kikkuli horse-training text — Hittite archive. c. 1400 BCE. Hittite with Indo-Aryan numerals. aika-, tera-, panza-, satta-, na-vartana — Indo-Aryan numerals in an Anatolian technical manual. about
- ^ a b c Jonathan Mark Kenoyer Ancient Cities of the Indus Valley Civilization. Oxford University Press, 1998. full text The standard accessible account by an excavator.
- ^ a b Pablo Librado et al. "Widespread horse-based mobility arose around 2200 BCE in Eurasia". Nature 631, 2024. doi Turns the Harappan "no horse" point from an argument from absence into a positive one: the domestic lineage had not yet spread anywhere.
- ^ RV 10.75.5 — Rigveda. c. 1100 BCE. Vedic Sanskrit. The Nadīstuti river hymn. Fixes the Rigvedic world in the Punjab; the Ganga appears once, at the far edge. read the text
- ^ a b c d Michael Witzel "Substrate Languages in Old Indo-Aryan". Electronic Journal of Vedic Studies 5, 1999. The case for a lost pre-Indo-Aryan language family in the northwest.
- ^ Kenneth A. R. Kennedy "Have Aryans been identified in the prehistoric skeletal record from South Asia?". in The Indo-Aryans of Ancient South Asia, 1995. The reanalysis that ended the Mohenjo-daro "massacre".
- ^ Liviu Giosan et al. "Fluvial landscapes of the Harappan civilization". PNAS 109, 2012. doi The geomorphology behind the Ghaggar-Hakra argument.
- ^ Rita P. Wright The Ancient Indus: Urbanism, Economy, and Society. Cambridge University Press, 2010. full text Strong on economy and on what the absence of palaces does and does not imply.
- ^ Thomas R. Trautmann Aryans and British India. University of California Press, 1997. open library How the Aryan idea was constructed, and how it went wrong.
- ^ Michael Witzel "Early Sanskritization: Origins and Development of the Kuru State". Electronic Journal of Vedic Studies 1, 1995. doi The Kuru state thesis — canonization as a political project.
- ^ David W. Anthony The Horse, the Wheel, and Language. Princeton University Press, 2007. full text The standard statement of the steppe hypothesis.
- ^ Michel Danino The Lost River: On the Trail of the Sarasvatī. Penguin India, 2010. The best-argued case on the other side of the Sarasvatī question. Cite it if the debate is to be stated fairly.
- ^ Edwin Bryant & Laurie Patton, eds. The Indo-Aryan Controversy: Evidence and Inference in Indian History. Routledge, 2005. Indigenist and mainstream scholars in the same volume, arguing directly.
- ^ Paul Heggarty et al. "Language trees with sampled ancestors support a hybrid model for the origin of Indo-European languages". Science 381, 2023. doi Dates PIE to ~8,100 BP and proposes a southern route alongside the steppe one. A mainstream, non-indigenist challenge to the simple steppe model — and contested.
- ^ Iosif Lazaridis, Nick Patterson, David Anthony et al. "The Genetic Origin of the Indo-Europeans". Nature 639, 2025. doi Derives the Yamnaya largely from a Caucasus–Lower Volga population, revising the older EHG–CHG picture.