(© sommart – stock.adobe.com)
In a nutshell
- Human language capacity existed at least 135,000 years ago, much earlier than previously thought, according to genomic evidence from population divergence studies.
- The 35,000-year gap between language capacity (135,000 years ago) and widespread symbolic behaviors (100,000 years ago) suggests language gradually transformed how humans thought and interacted with their world.
- All modern human populations share language capabilities with similar structures, indicating that language emerged before the first major population split occurred in our ancestral line.
CAMBRIDGE, Mass. — It’s long been wondered exactly when our ancestors developed the ability to speak. Of course, there’s no surefire way to know specifically when our unique “superpower” first began, but scientists at the Massachusetts Institute of Technology offer compelling evidence that humans possessed the capacity for language at least 135,000 years ago—much earlier than many previous estimates suggested.
Led by MIT linguist Shigeru Miyagawa, the research team tackled this age-old question with a surprisingly fresh approach: they examined genetic evidence from population divergence studies to pinpoint when language must have first appeared in our evolutionary story.
Their core finding? Language likely fueled the explosion of symbolic behaviors and cultural innovations that define modern humans, which began appearing in the archaeological record around 100,000 years ago.
How Scientists Solved the Language Puzzle
The research team’s approach hinges on a clever observation. All modern human populations today have fully developed language capabilities. The roughly 7,000 languages spoken worldwide share remarkable structural similarities in how they’re built—their sounds, grammar, and meaning systems follow similar patterns regardless of geography or cultural isolation.
This universality reveals something profound: the cognitive machinery for language must have been present in our ancestral population before any major splits occurred. If language had developed after populations began diverging, we would expect to find some human groups today without language or with fundamentally different communication systems. Neither scenario exists.
To identify when this crucial split happened, Miyagawa’s team analyzed 15 different genomic studies published between 2007 and 2023. They categorized these studies by the genetic markers used: some examined DNA passed down through mothers, others looked at DNA passed down through fathers, and the most comprehensive analyzed whole-genome comparisons.
Most studies employed statistical methods to determine when populations diverged, with the most advanced using techniques that examine patterns of genetic variation to reconstruct population history. By calculating average values from these various studies, genomic evidence consistently points to the Khoisan peoples of southern Africa as representing the earliest division from our original stem population
Whole-genome comparisons pointed to 136,000 years ago (give or take 23,000 years), while other genetic markers showed similar timeframes. Combining these lines of evidence, the study concludes the original split occurred approximately 135,000 years ago.
The 35,000-Year Gap: From Language to Culture
If language capacity existed 135,000 years ago and symbolic behaviors began appearing widely in the archaeological record around 100,000 years ago, what happened in the gap in between? This 35,000-year interval might represent a critical period when early humans were developing their linguistic abilities, eventually triggering an explosion of innovation.
These innovations include ochre pieces with symbolic engravings found at Blombos Cave in South Africa (dated to around 77,000 years ago), marine shells used as personal ornaments, and geometric patterns engraved on ostrich eggshells found at sites like Diepkloof and Klipdrift Shelters.
While Neanderthals and other extinct human relatives occasionally demonstrated similar behaviors, only Homo sapiens made such symbolic activities routine across populations. This pattern hints that language, with its unique capacity for quite complex symbolic representation, transformed early humans into the symbolically-oriented beings we recognize today.
Debunking Prior Research
Miyagawa and colleagues propose that language worked as a catalyst that accelerated various gradual processes taking place throughout the Middle Stone Age. Language had “a direct and enormous impact on all facets of human life,” helping our ancestors create new connections between existing symbols and develop new behavioral patterns.
The research, published in Frontiers in Psychology, challenges some prominent linguists, including Noam Chomsky, who previously suggested linguistic capacity emerged much later—possibly around 50,000 years ago. Based on the genetic evidence, the authors consider such late emergence “highly implausible.”
While establishing that linguistic capacity existed in Homo sapiens at least 135,000 years ago, the researchers can’t yet pinpoint exactly when or how language itself first emerged. Future studies combining genetic evidence with archaeological findings may further narrow this window.
Nevertheless, this work pushes back the timeline for one of humanity’s defining features and helps explain the subsequent flowering of symbolic behaviors that have characterized our species ever since.
Paper Notes
Limitations
The researchers acknowledge several limitations to their approach. First, while they can establish when linguistic capacity must have been present, they cannot determine exactly when language itself first emerged—only that it had to be present by around 135,000 years ago. Second, despite improvements in genetic analysis techniques, some imprecision in molecular clock data remains unavoidable with current technology. The estimates from different studies show considerable variation, with upper estimates ranging from 110,000 to 210,000 years ago and lower estimates from 53,000 to 178,000 years ago. The study also doesn’t resolve ongoing debates about whether language evolved gradually or emerged suddenly as a result of a cognitive revolution. Additionally, the researchers note that their findings don’t address the specific evolutionary changes that led to language emergence, remaining neutral on competing theories about language evolution. Finally, the study relies on genomic evidence rather than direct archaeological evidence of language use, which remains elusive given that language doesn’t fossilize.
Funding and Disclosures
The research was financially supported by the São Paulo Research Foundation (FAPESP) through grants awarded to Shigeru Miyagawa (grant number 2018/18900-1) and Vitor Augusto Nóbrega (grant number 2023/03196-5). The authors declared no conflicts of interest, stating that the research was conducted in the absence of any commercial or financial relationships that could be construed as potential conflicts. The researchers also explicitly declared that no generative AI was used in the creation of their manuscript, an increasingly common disclosure in academic publications as AI tools become more prevalent in research and writing processes.
Publication Information
The paper, titled “Linguistic capacity was present in the Homo sapiens population 135 thousand years ago,” was published in Frontiers in Psychology on March 11, 2025. It was authored by an international team including Shigeru Miyagawa (Massachusetts Institute of Technology, University of São Paulo, and Seikei University), Rob DeSalle (American Museum of Natural History), Vitor Augusto Nóbrega (University of São Paulo), Remo Nitschke (University of Zurich and University of Arizona), Mercedes Okumura (University of São Paulo), and Ian Tattersall (American Museum of Natural History). The article was submitted on September 30, 2024, accepted on February 6, 2025, and published the following month. It appeared in volume 16 of the journal with the DOI identifier 10.3389/fpsyg.2025.1503900. The paper is classified as a “Perspective” article, which typically presents a viewpoint on a research area supported by evidence but is not necessarily a full original research study.
