Paralanguage and human-computer interaction. Part 1: identification of recorded vocal segregates

Publisher Website
Google Scholar
Abstract
Vocal segregates are short non-lexical utterances such as ‘mm-hmm’. They are frequently observed in natural dialogue, which they help to regulate and maintain. Twenty-five naïve subjects were asked to identify the meanings of isolated vocal segregates recorded by unfamiliar speakers. The segregates were recorded both with natural articulation, which enables the differentiation of vowel sounds, and with the speaker's mouth closed. Responses were made using a seven-alternative forced choice procedure. Mean identification accuracy was 69.4% significantly above chance level, and performance was higher for vocal segregates which contained vowel sounds. The possible role of vocal segregates in human-computer interaction is discussed.

This publication has 9 references indexed in Scilit: