Dispreferred speech acts in virtual reality: Analysis of tone choices and hesitations

Abstract

This study compared participants' dispreferred speech acts between two oral discourse completion tasks (oral DCT): computer-based and virtual reality (VR)-based. Both tasks involved a series of scenarios eliciting speech acts of refusals and complaints (i.e., dispreferred acts). While the computer-based oral DCT elicited participants' speech directed to their imaginary interlocutor, in the VR-based task participants produced their speech to an actual interlocutor appearing in 360° visualizations. Sixty-two native and nonnative speakers of English completed both tasks. Using mixed effects models, the impact of task conditions (computer-based or VR-based) and language background (native or nonnative speakers) was assessed on two linguistic features: tone choices (rising, falling, and level tone) and hesitations (e.g., occurrences of filled and unfilled pauses). Both participant groups produced more level tones and hesitations in the VR-based than the computer-based DCT. Although native speakers used more falling tones, nonnative speakers used more rising tones, and their speech involved more hesitations.