Dialect distances based on orthographic and phonetic transcriptions

Abstract

This paper describes ongoing work in which the main objective is to quantitatively determine the linguistic distances between languages and dialects. The Levenshtein distance measure is applied to orthographic and phonetic transcriptions of words from 15 Norwegian dialects. Clustering of the distances between the different dialects shows the relationships between the dialects in terms of regional groupings and closeness. Although orthographic transcriptions generate distinctive north and south groupings, the more detailed phonetic transcriptions group the dialects more decisively into their regional groups. When the phonetic transcriptions are employed, the dendrogram of distances between regions is very similar to that computed from perceptual assessment of dialect distances