Towards a Complete Mapping of Kurdish Dialectology

An ongoing effort to create a comprehensive interactive map of Kurdish dialect geography, continua, and mutual intelligibility.

Kurdish is often treated as a single language in NLP resources, but in reality it is a dialect continuum - or more accurately, several overlapping continua - spanning a vast geographic area across Turkey, Iraq, Iran, Syria, and diaspora communities worldwide. This project works towards a complete, systematic mapping of this variation: where are the boundaries, how sharp or gradual are they, and what linguistic features mark them?

One of the most interesting puzzles in Kurdish dialectology is the asymmetry of mutual intelligibility: speakers of Central Kurdish (Sorani) report understanding Northern Kurdish (Kurmanji) relatively well, while the reverse is far less true. This counter-intuitive finding can be partly explained by the dialect continuum - 'Central Kurdish' is not a uniform block but a gradient of sub-varieties, some of which are geographically and linguistically close to Northern Kurdish. Mapping these sub-varieties is essential to understanding both the linguistics and the NLP challenges they pose.

The project is ongoing. Initial results were presented at ICKL-6 in 2023. Work continues on expanding the variety coverage and refining the geographic representations. [TODO: add links to updated maps and publications as they become available]