Linking Linguistic and Geographic Distance in Four Semantic Domains: Computational Geo-Analyses of Internal and External Factors in a Dialect Continuum

Abstract

Dialectometry studies patterns of linguistic variation through correlations between geographic and aggregate measures of linguistic distance. However, aggregating smooths out the role of semantic characteristics, which have been shown to affect the distribution of lexical variants across dialects. Furthermore, although dialectologists have always been well-aware of other variables like population size, isolation and socio-demographic features, these characteristics are generally only included in dialectometric analyses afterwards for further interpretation of the results rather than as explanatory variables. This study showcases linear mixed-effects modelling as a method that is able to incorporate both language-external and language-internal factors as explanatory variables of linguistic variation in the Limburgish dialect continuum in Belgium and the Netherlands. Covering four semantic domains that vary in their degree of basic vs. cultural vocabulary and their degree of standardization, the study models linguistic distances using a combination of external (e.g., geographic distance, separation by water, population size) and internal (semantic density, salience) sources of variation. The results show that both external and internal factors contribute to variation, but that the exact role of each individual factor differs across semantic domains. These findings highlight the need to incorporate language-internal factors in studies on variation, as well as a need for more comprehensive analysis tools to help better understand its patterns.