There are sophisticated quantitative methods to reconstruct the expansion of language families in space and time, but no comparable methods to trace language contact, the spread of features between unrelated languages. In this research project we aim to close this research gap.
We developed a Bayesian model to quantify evidence for language contact. The model looks at the change in linguistic similarity over time: regions where languages have become more similar are considered as evidence for contact. When applied to languages in north-western Europe, the model reveals a substantial gain in linguistic similarity in the British Isles, cross-cutting lines of linguistic ancestry.
Currently, we are working on a Baysian model to automatically find pathways of language diffusion and language contact in space.