Skip to content
Mela Keela
Veḷi

Exhibit · the pattern

It was not three expansions. It is one pattern, five thousand years long, and it has not stopped.

The usual count is three: Yamnaya into Europe, Indo-Iranian into South Asia, European colonialism across the planet. The record shows at least nine — and the people underneath each wave were, more often than not, the descendants of an earlier one. Nearly half of humanity now speaks a language descended from a community that lived north of the Black Sea five thousand years ago. This page maps every expansion, what it did to the people already there, and the two that most people have never heard named.

~46%

of humanity speaks an Indo-European language natively

400+

languages in the family

~80%

of Yamnaya ancestry from one Caucasus–Volga source

9+

distinct expansions, 4000 BCE to 1950 CE

Before the steppe.

the 2025 finding

The Yamnaya were not the beginning. A 2025 study in Nature of 435 ancient genomes identifies a Caucasus–Lower Volga (CLV) population, c. 4500–3500 BCE, from which the Yamnaya derived about 80% of their ancestry — and which also contributed roughly a tenth of the ancestry of Bronze Age Hittite-speakers in Anatolia. This resolves the oldest problem in the field: Anatolian, the earliest Indo-European branch, showed no steppe ancestry because it split off before the steppe phase, from the CLV source directly. The homeland is the North Caucasus–Lower Volga region. The Yamnaya were its first great amplification.

Every expansion.

what moved, where, and who was underneath

c. 4500–3500 BCE · CLV → AnatoliaHittite, Luwian. Beneath: Hattic, Hurrian speakers. c. 3300–2500 BCE · Yamnaya → Corded Ware → EuropeLargest genetic event in Europe in 5,000 years. Beneath: Anatolian-descended farmers; near-total Y replacement in Britain. c. 3300–2500 BCE · Afanasievo → Altai, MongoliaThe eastward mirror. Tocharian. Beneath: Ancient North Eurasian populations. c. 2200–1600 BCE · into GreeceMycenaeans 13–18% steppe; Minoans none. Beneath: pre-Greek "Pelasgian" substrate, Linear A. c. 1900–1200 BCE · Sintashta → Indo-Iranian → South AsiaMale-biased. Beneath: the Indus population — CIHG + AASI — and every forager group behind them. c. 1500–800 BCE · Iranian → the plateau; Scythian → the steppeMedes, Persians, Saka. Beneath: Elamite, and the earlier plateau population. c. 300–800 CE · Germanic, then Slavic → across EuropeSlavic: the largest linguistic expansion in Europe — Baltic, Finnic, Turkic speakers beneath. Mostly assimilation, not replacement. 1500–1950 CE · European colonialism → five continentsEnglish, Spanish, Portuguese, French. Beneath: the Americas, Australia, Africa, and South Asia — already ranked. 1580–1900 CE · Russian → Siberia to the PacificThe one nobody counts. Beneath: 40+ indigenous Siberian peoples.
Red: demographically replacive or strongly male-biased. Orange: significant admixture, partial replacement. Blue: primarily linguistic and cultural assimilation. The categorisation follows the ancient-DNA evidence where it exists; for the historical expansions it follows the documentary record.

The two nobody names.

and why they matter to the argument

Slavic — the counter-example

Between roughly 500 and 800 CE, Slavic spread across more of Europe than any language family before or since — from the Elbe to the Volga, from the Baltic to the Adriatic — over territory held by Baltic, Finnic, Germanic, Turkic and Illyrian speakers. And the genetics show it was mostly not replacement. In the Volga–Oka region, about half of modern Russian R1a carriers descend from the pre-Slavic Finnic population — the Slavs assimilated the Meshchyora and others rather than displacing them. Slavic speakers today are genetically close to their non-Slavic neighbours.

This matters because it proves the pattern is not inevitable. An Indo-European expansion can be linguistic without being demographic, and can absorb without ranking. What the Slavic case lacks is what South Asia has: a hereditary legal order built on top of the arrival.

Russian Siberia — the one that mirrors India

From 1580 the Russian state expanded east across Siberia to the Pacific — a distance greater than the Atlantic crossing — over the lands of more than forty indigenous peoples: Nenets, Khanty, Evenki, Yakut, Buryat, Chukchi, Nivkh. Fur tribute, forced conversion, hostage-taking of clan leaders, and later the Gulag. Most Siberian languages are now endangered or extinct.

Structurally, it is British India with the polarity reversed: a steppe-descended Indo-European state colonising indigenous northern Asia by land, while a sea-borne Indo-European state colonised indigenous South Asia by water. Both found tribal peoples on marginal land and administered them into categories. Neither is counted as an "Indo-European expansion" because the word is reserved for prehistory. It should not be.

Each wave found the survivors of the last one on the worst land, and pushed them further.

What travelled with the language.

the package

Inspect

sort by date, by direction, by what travelled

ElementWhere it is documented
Patrilineal descent groupsYamnaya kurgans are male-dominated; Corded Ware households show patrilocality and female exogamy; the Sintashta package includes the chariot warrior burial; gotra in South Asia; the Roman gens; the Germanic sippe.
The male-lineage bottleneckR1b's near-total replacement of Neolithic Y lineages in Britain and Iberia; R1a's concentration in South Asian upper castes; the global 5–7 kya collapse in male effective population size that the expansions ride on.
Hierarchy coded into lawThe tripartite ideology — priests, warriors, producers — reconstructed across Vedic, Iranian, Roman and Celtic material by Dumézil. Only in South Asia did it become hereditary, endogamous, and enforced for two thousand years.
The displaced into refugiaBasque in the Pyrenees; the Vedda in the Sri Lankan highlands; the Irula in the Nilgiris; the Nenets on the tundra. Every surviving isolate in Eurasia sits on land the expansion did not want.

The page's claim is that these are one pattern rather than three events. Sorting by date is the test.

Three layers on one people.

the South Asian case, stated once

Where the pattern closed on itself

First layer, c. 1900–1200 BCE: Indo-Iranian speakers arrive, male-biased, and over the following millennium a hereditary ranking is written into law with the arrivals' descendants at its head and the earlier population's descendants beneath — the foragers pushed to the hills and made a category.

Second layer, 1600–1947 CE: Indo-European speakers arrive again, from the sea — descendants of the branch that went west while the first went south. They find the ranking already in place, measure it with nasal indices, enumerate it in censuses, and codify the forest-dwellers as criminal by birth. They administer what the first layer built.

Third layer, now: the science that reconstructs all of this is written in English, held in Boston, and read back to the subcontinent through institutions controlled by the first layer's heirs. The Irula are beneath all three, and the people at the top of each are cousins.

The honest limits.

Mela Keela is an independent digital museum. Sources, evidentiary limits and review status are identified on individual pages.