The map of human speech is a lie.
It shows about 7,500 living languages. We take this number for granted. It feels static. But new research suggests we are looking at a shadow.
A team of researchers from Spain, the United States, and Germany used complex modeling to reconstruct the last 12,00 years. Their findings, published in Science, rewrite the timeline of linguistic diversity. The world did not start with thousands of languages and steadily lose them.
It started with fewer.
Then it exploded.
For thousands of years, language diversity climbed alongside the human population. It didn’t peak in the distant, mysterious Paleolithic. The peak was surprisingly recent. The “golden age” of human language occurred roughly 1,000 to3,000 years ago—long before Columbus sailed. At its height, the number of languages may have been ten times higher than it is now. Tens of thousands of tongues may have been spoken.
This challenges the standard narrative. We blame European colonialism for language loss. Colonialism accelerated the death of languages, yes. It is undeniable. But the decline started earlier. It began with the rise of empires.
The hidden peak of linguistic diversity
How did we get here?
The problem is simple: language leaves no bones. No skeletons. No pottery shards labeled “Proto-Indo-European dialect 4.” Before writing emerged roughly 6,000 years ago the historical record goes blank. We are guessing.
Most people assume that hunter-gatherer societies were wildly diverse. Small groups, isolated by mountains and forests. One tribe per valley. This leads to the idea that linguistic diversity peaked early in human history and has only fallen since.
The data says otherwise.
The researchers looked at 171 modern hunter-gatherer and fishing groups. These groups had minimal contact with agricultural communities. Their movement and subsistence patterns remained traditional. These groups provide a baseline for population size.
They estimated how large ethnolinguistic groups were at the start of the Holocene, roughly 12,000 yearsago. At that time, the global population was between 4.4 million and7 million people.
If we assume typical group sizes for foragers, the math is startling.
The early Holocene likely had between 3,300 and 7,800 languages. That is fewer than today. It was not an era of hyper-diversity.
Then agriculture arrived.
Farming fuelled the explosion
Agriculture changed everything. It created surplus food. Surplus food created population growth. But it also changed how we organized ourselves.
Early human communities grew larger. This allowed more people to share a single language. It created a drag on diversity. More mouths, same tongue.
Yet, the models show that linguistic diversity rose.
Why?
The researchers generated thousands of possible historical pathways. They combined population estimates with ethnographic data. The result was consistent.
Language diversity grew alongside the human population for millennia. The number of languages increased even as average population size rose. The peak came between 1,000 and3,000 years ago.
“This period of greatest linguistic diversity may have not occurred in the remote Paleolithic but during the era of expanding states andearly empires,” says Claire Bowern of Yale University.
Think about that.
The era of expanding states. The rise of empires. The Iron Age. The height of the Han Dynasty. The Roman Republic. This is when the world was most linguistically diverse.
This is the lost golden age.
It is overlooked. Historical records from that time were written by expanding societies. Their languages became dominant. Their accounts erased the smaller groups. The “golden age” vanished because the survivors didn’t bother to count the dead.
Empires killed language diversity
After 1,000 years ago?
The curve crashes.
Linguistic diversity falls rapidly. The drop is steep. It continues over the last two thousand years.
European colonialism is often cited as the sole culprit. It accelerated the process. It swept across continents, replacing local tongues with imperial ones. But the models indicate the decline began earlier.
Multinational states expanded. They brought formerly separated populations into contact. They spread politically dominant languages. They brought new technologies. Cultural traditions. Pathogens.
The analysis does not pinpoint which empires drove this. It is a global pattern. A massive contraction.
The languages we see today are not representative of the past. They are the survivors.
Languages spoken by expanding populations survived. Languages spoken by absorbed, displaced, or decimated groups disappeared. This creates a selective bottleneck.
This matters more than we think.
Common linguistic features—grammatical patterns, sounds, structures—are often viewed as cognitively superior. Or efficient. Or easy to learn. The study suggests these traits are common not because they are “better.”
They are common because they rode the waves of expanding populations. Extinction shaped linguistic diversity more than innovation did.
What else died with the languages?
Languages are not just words.
They are vessels.
Religious beliefs. Kinship practices. Social institutions. Ecological knowledge. All of this travels with the language.
When a language goes extinct, the vessel cracks. The contents spill.
The researchers argue that the same bottleneck that killed tens of thousands of languages also erased vast swathes of cultural knowledge. Direct historical traces are rare. We only see the survivors.
The study is a broad reconstruction. It relies on assumptions. It cannot recreate short-term collapses caused by plague or war in specific regions. Regional patterns likely varied.
But the main pattern is stable.
Language diversity did not fall steadily from a Paleolithic peak. It rose. It hit a brief, unusually high peak around 2,000 years ago. Then it contracted. Severe contraction.
We are left with a fragment. A sliver of what was.
The map of human speech is incomplete. The silence in the data is loud.
