Prosody in spontaneous and read speech: Some preliminary findings from a tone language

Abstract

Intuitively it appears easy to identify whether someone is speaking spontaneously or reading out loud from a text. Previous research has shown that prosody plays a role for this perceptual distinction. The current study examines a number of acoustic prosodic parameters in the production of the two speaking styles in Mandarin Chinese. We ask whether the use of contrastive pitch in this tone language is implicated in the prosodic parameters related to the two speech modes. Spontaneous speech from three female speakers of Mandarin was recorded during interviews with the experimenter. For each speaker, ten declarative sentences were selected and transcribed. The speakers read the transcribed sentences in a second recording session 2 weeks later. Thus, ten pairs of lexically and syntactically identical sentences were obtained. Acoustic analyses were conducted of a number of intonational and temporal parameters (i.e., F0, F0 range, F0 standard deviation, F0 declination, and speaking rate). Preliminary analyses show that among the acoustic parameters, only the speaking rate consistently distinguishes between the two speaking styles. This result contrasts with previous research showing the contribution of F0-related parameters to the spontaneous-read distinction.