Measuring socially motivated pronunciation differences

Abstract

This paper applies a measure of linguistic distance to differences in pronunciation which have been observed as a consequence of modern speakers orienting themselves to standard languages and larger regions rather than local towns and villages, resulting in what we shall call regional speech. We examine regional speech, other local "varieties" in the Dutch of the Netherlands and Flanders, and also standard Netherlandic Dutch and Belgian Dutch.2 Because regional speech is difficult to study, as it may not constitute a linguistic variety in the usual sense of the word, we focus on the speech of professional announcers employed by regional radio stations. We examine their speech in light of Auer and Hinskens' (1996) cone-shaped model of the speech continuum, which includes REGIOLECTS, which they define as a sort of comprise between standard languages and local dialects (more below). In this examination we use a measure of pronunciation difference which has been successful in dialectology (see Nerbonne & Heeringa 2009 for an overview) and which has been demonstrated to be valid both for measuring dialect differences and also for measuring speech differences due to limited auditory acuity (cochlear implants). We thereby introduce a technique into sociolinguistics to measure the difference between regional speech and standard Dutch as well as the difference between regional speech and the local speech of towns and villages, providing a perspective on the issue of whether regional speech functions as "standard" within more restricted areas or whether it serves rather to mark regional identity.