Introducing Micrela: Predicting mutual intelligibility between related languages in Europe

The Wikipedia lists 91 indigenous languages that are spoken in Europe today. Forty-two are currently being used by more than one million speakers. Only 24 are recognized as official and working languages of the European Union. Most of these belong to one of three major families within the Indo-European phylum, i.e. Germanic, Romance and Slavic languages. Languages and language varieties (dialects) within one family have descended from a common ancestor language which has become more diversified over the past centuries through innovations. Generally, the greater the geographic distance and historical depth (how long ago did language A undergo an innovation that language B was not part of), the less the two languages resemble one another, and - it is commonly held - the more difficult it will be for speakers of language A to be understood by listeners of language B and vice versa. A working hypothesis would then be that the longer ago two related languages split apart, the less they resemble one another and the smaller their mutual intelligibility. We are involved in a fairly large research project that was set up to test these hypotheses. The basic idea was to first measure the level of mutual intelligibility between all pairs of languages within a family, them compute the degree of structural linguistic similarity between the members, and try to predict the observed level of intelligibility from the linguistic distances measured. The historic component is seen as a subsidiary issue: we will not try to predict mutual intelligibility between languages from the distance between them in the traditional linguistic family tree (cladistic distance) but simply check to what extent the linguistic relatedness between two languages corresponds with their mutual intelligibility and/or measured linguistic distance.