Clear Speech — Mere Speech? How Segmental and Prosodic Speech Reduction Shape the Impression That Speakers Create on Listeners

Abstract

Research on speech reduction is primarily concerned with<br/>analyzing, modeling, explaining, and, ultimately, predicting<br/>phonetic variation. That is, the focus is on the speech signal<br/>itself. The present paper adds a little side note to this fundamental<br/>line of research by addressing the question whether<br/>variation in the degree of reduction also has a systematic<br/>effect on the attributes we ascribe to the speaker who produces<br/>the speech signal. A perception experiment was carried out for<br/>German in which 46 listeners judged whether or not speakers<br/>showing 3 different combinations of segmental and prosodic<br/>reduction levels (unreduced, moderately reduced, strongly reduced)<br/>are appropriately described by 13 physical, social, and<br/>cognitive attributes. The experiment shows that clear speech is<br/>not mere speech, and less clear speech is not just reduced either.<br/>Rather, results revealed a complex interplay of reduction<br/>levels and perceived speaker attributes in which moderate<br/>reduction can make a better impression on listeners than no<br/>reduction. In addition to its relevance in reduction models and<br/>theories, this interplay is instructive for various fields of<br/>speech application from social robotics to charisma coaching.