Language borrowing often accompanies cultural exchange, yet tracing its history can be difficult without reliable records of population contact. A team led by researchers at the University of Zurich has now turned to genetics to fill these gaps, using population mixing as a marker for linguistic exchange.
The study, published in Science Advances, integrates genomic data from more than 4,700 individuals across 558 populations with two large-scale linguistic databases cataloging grammatical, phonological, and lexical features. “By using genetic data as a proxy for past human contact, we were able to get around the problem of missing historical records and we could detect over 125 comparable instances of contact across the globe,” said Anna Graff, lead author of the study.
The team found that genetic contact between populations increased the probability of linguistic sharing in unrelated languages by 4–9 percent. “What surprised us most is that no matter where in the world populations come into contact, their languages become more similar to remarkably consistent extents,” added Chiara Barbieri, senior author and population geneticist at the University of Cagliari.
Not all features were equally likely to spread. Word order and consonant sounds tended to transfer more readily than other structural elements, yet the study found no universal rules for borrowability.
“This challenges long-standing assumptions about what makes a linguistic feature more or less borrowable,” explained Balthasar Bickel, director of the NCCR Evolving Language. The findings also highlighted instances where languages diverged after contact, reflecting efforts to assert group identity.
According to Graff, both convergence and divergence are integral to understanding global language evolution. The results demonstrate how genetic and linguistic data, studied together, can provide a clearer record of how human societies interact and change.
Publication Details
Anna Graff et al., Patterns of genetic admixture reveal similar rates of borrowing across diverse scenarios of language contact. Sci. Adv. 11, eadv7521(2025).DOI:10.1126/sciadv.adv7521
The study, published in Science Advances, integrates genomic data from more than 4,700 individuals across 558 populations with two large-scale linguistic databases cataloging grammatical, phonological, and lexical features. “By using genetic data as a proxy for past human contact, we were able to get around the problem of missing historical records and we could detect over 125 comparable instances of contact across the globe,” said Anna Graff, lead author of the study.
The team found that genetic contact between populations increased the probability of linguistic sharing in unrelated languages by 4–9 percent. “What surprised us most is that no matter where in the world populations come into contact, their languages become more similar to remarkably consistent extents,” added Chiara Barbieri, senior author and population geneticist at the University of Cagliari.
Not all features were equally likely to spread. Word order and consonant sounds tended to transfer more readily than other structural elements, yet the study found no universal rules for borrowability.
“This challenges long-standing assumptions about what makes a linguistic feature more or less borrowable,” explained Balthasar Bickel, director of the NCCR Evolving Language. The findings also highlighted instances where languages diverged after contact, reflecting efforts to assert group identity.
According to Graff, both convergence and divergence are integral to understanding global language evolution. The results demonstrate how genetic and linguistic data, studied together, can provide a clearer record of how human societies interact and change.
Publication Details
Anna Graff et al., Patterns of genetic admixture reveal similar rates of borrowing across diverse scenarios of language contact. Sci. Adv. 11, eadv7521(2025).DOI:10.1126/sciadv.adv7521