Study finds language diversity peaked before empires, not colonialism

A new study coauthored by Yale linguist Claire Bowern, lead author Damián Blasi (ICREA research professor at the Center for Brain and Cognition, Pompeu Fabra University, Barcelona), Russell Gray (director of the Department of Linguistic and Cultural Evolution at the Max Planck Institute for Evolutionary Anthropology, Leipzig) and Marcus Hamilton (University of Texas at San Antonio and the Santa Fe Institute) traces the trajectory of global language diversity over the past 12,000 years. It was published July 23 in the journal Science.

Because direct evidence of languages does not exist before writing appeared about 6,000 years ago, the researchers combined ethnographic data from 171 hunter-gatherer and fisher societies whose traditional way of life had not been reshaped by contact with food-producing populations, with independent estimates that the global human population 12,000 years ago was between about 4.4 and 7 million people. From this they modelled thousands of possible trajectories connecting language counts at the start of the Holocene with the present, estimating that between 4,500 and 6,200 languages were spoken 12,000 years ago.

The models consistently show linguistic diversity growing for thousands of years after the start of agriculture, then peaking between 1,000 and 3,000 years ago, when upwards of tens of thousands of languages were spoken worldwide. This is the period the researchers call the world's linguistic 'golden age.'

That golden age ended as large states and multinational empires, such as the Roman Empire, spread their languages, cultures and pathogens to previously independent populations, causing rapid language loss. The finding challenges the commonly held view that widespread language extinction began about 500 years ago with European colonial expansion: according to the study, colonialism intensified the decline in linguistic diversity but did not cause it. Languages of expanding states and empires were disproportionately likely to survive, while the languages of populations that were absorbed, displaced or in decline disappeared.

The researchers say this means the roughly 7,600 languages spoken or signed today are a small and historically biased sample of the languages that once existed, since they are disproportionately drawn from expansionist states rather than being a representative cross-section of past human linguistic diversity. Bowern said the work shows that language patterns can be used to reconstruct large demographic events over millennia, and that language has never been isolated from political currents. She added that the finding matters for today's endangered languages because political currents do not have to move in one direction, and extinction is not inevitable. Blasi said the pre-agricultural world was probably not exceptionally rich in languages, with diversity growing alongside population over thousands of years. Gray said today's languages are the survivors of a massive, highly selective historical bottleneck, and that a linguistic feature can be common not because it works better for communication but because it happened to be carried by populations that expanded.

Key facts

The study, led by Damián Blasi with Yale's Claire Bowern, Russell Gray and Marcus Hamilton, was published July 23 in Science and models global language diversity across the past 12,000 years.
Linguistic diversity peaked in a 'golden age' 1,000 to 3,000 years ago, when upwards of tens of thousands of languages were spoken worldwide, versus an estimated 4,500 to 6,200 languages 12,000 years ago.
The decline was driven by the rise of large states and empires, such as Rome, not by European colonialism, which began about 500 years ago and intensified but did not cause the drop, per the study.
The model draws on ethnographic data from 171 hunter-gatherer and fisher societies plus population estimates of about 4.4 to 7 million people 12,000 years ago.
Today's roughly 7,600 spoken or signed languages are a small, historically biased sample: languages of expanding, dominant groups survived disproportionately while those of absorbed or declining populations vanished.

Why it matters

The study pushes back the timeline for mass language extinction by up to two millennia, from the start of European colonialism about 500 years ago to the rise of large states and empires 1,000 to 3,000 years ago. It reframes today's roughly 7,600 languages not as the normal, stable baseline of human linguistic diversity but as a small, politically skewed remnant of a far larger set that has been shrinking for millennia.

Who it affects

The finding speaks to linguists, historians, demographers and evolutionary biologists studying how human diversity has changed over time, and to researchers and advocates working on endangered languages today, since it argues the drivers of language loss are political and demographic rather than fixed to any one historical episode.

How to use it

As a research finding rather than a tool, its main use is reframing: it suggests language survival tracks the expansion of dominant political groups rather than any inherent communicative advantage of a language, and Claire Bowern argues this means the loss of endangered languages today is not inevitable, since political currents can move in more than one direction.

How solid is it

The paper was peer reviewed and published in Science, and is the product of an international collaboration spanning linguistics, demography and evolutionary biology, backed by ethnographic data from 171 societies and independent population estimates. Because no direct evidence of prehistoric languages exists before writing appeared about 6,000 years ago, the key figures, such as the peak number of languages and the 12,000-year-ago baseline, are modelled estimates given as ranges (for example 4,500 to 6,200 languages) or as an order of magnitude ('upwards of tens of thousands'), not precise counts.

Risks and caveats

The study's earliest figures rest on assumptions connecting population size to the number of distinct language-speaking groups, applied to societies whose way of life had not been transformed by contact with agricultural populations; this is a model of plausible trajectories, not a direct count. The source text does not detail the statistical and social-computational methods beyond naming them, and gives no exact figure for the peak number of languages during the golden age.

“The languages we see today are the survivors of a massive and highly selective historical bottleneck.”

— Russell Gray, study coauthor, Max Planck Institute for Evolutionary Anthropology