Data‐Driven Dialectology

Abstract Most studies of language variation proceed from the geographic or social distribution of single elements (features), and find it difficult to proceed further. Data‐driven dialectology, and more generally, data‐driven variationist studies, begin instead from an aggregate view of language variation and reap immediate benefits in dealing with well‐known exceptions in the distributions of single features and in avoiding the need to select which features to use as the basis of characterizations. But the major advance is the opportunity to characterize general tendencies in linguistic variation.