There are reasons for assuming that different types of communities provide different social conditions for linguistic changes with consequences at least for the speed and for the type of grammatical changes. In order to explore this question, one needs both a typology of communities and a model for measuring the extent of linguistic change. This article suggests and discusses a way of measuring by which it should be possible to systematically compare the degree of linguistic change. The tentative calculations will primarily be illustrated by data from dialects in Norway in the 20th century.