Study estimates humans spoke up to 10 times more languages 2,000 years ago
A new study suggests humanity may have lost nine out of every ten languages it once spoke over the past two millennia.
Tens of thousands of human languages may have vanished over the past two thousand years, leaving today's roughly 7,000 surviving tongues as a fraction of what once existed. A new study proposes that at a historical peak — a kind of 'golden age' of linguistic diversity — humans worldwide spoke or signed up to ten times as many distinct languages as are documented today. The research spans communities across the globe and includes signed as well as spoken languages. The findings suggest that linguistic diversity has declined sharply as populations expanded, consolidated, and dominant languages spread at the expense of smaller ones. While today approximately 7,000 languages are counted by researchers, the study's model implies the historical total could have reached 70,000 or more at its height a few millennia ago. The work adds a quantitative dimension to longstanding concerns among linguists that language extinction is not a modern phenomenon alone, but a process with deep historical roots predating colonialism and globalization.
Why it matters
Language extinction erases unique ways of encoding knowledge, culture, and human experience; understanding the historical scale of loss helps researchers and communities assess what has already disappeared and what remains at risk.
Key facts
- The study estimates humans may have spoken up to 10 times as many languages as exist today
- Approximately 7,000 languages are documented in the world today
- The proposed historical peak occurred roughly 2,000 years ago
- The estimate includes both spoken and signed languages
- The findings imply a historical total potentially reaching around 70,000 distinct languages
Bias & framing notes
Both sources align on the core claim and the 10-times figure, but Science News frames the story around a poetic 'golden age' concept while Scientific American leads with the quantitative estimate. Neither source provided significant methodological detail in the available reporting, making it difficult to evaluate the strength of the underlying study's model.
NewsClear — neutral news & congressional tracking · Bill of the Week