Rundfunknachrichten als linguistisch-phonetische Ressource für die Sprachwandelforschung

Ein halbautomatisiertes Verfahren und eine Demosprachdatenbank

Abstract

Spoken language is subject to synchronous variation at the segmental and suprasegmental levels, which in turn can lead to diachronic change. This study presents a semi-automatically processed radio news corpus as a useful source for diachronic trend studies and provides an initial exploratory assessment of prosodic change in terms of speech rate. The corpus comprises radio news from 1956 to 2017, obtained from the public service broadcaster Saarländischer Rundfunk. An initial analysis revealed that during the 60-year period under investigation, the proportion of core news was reduced in favor of other news elements. Further signal-based and signal+text-based analyses of the core news items showed that speech rate increased over time, which was accompanied by fewer and shorter pauses. The articulation speed, on the other hand, did not change. Advantages and disadvantages of news corpora for prosodic change research are discussed.