An investigation of functional relations between speech rate and phonetic variables. (July 2022)

• Record Type: Journal Article
• Authors: Kim, Seung-Eun; Tilsen, Sam
• Is Part Of: Journal of phonetics. Volume 93(2022)
• Publication Date: 2022-07
• Subjects: Speech rate -- Prosody -- Functional relations -- Non-linearity -- Attenuation effects -- Bayesian analysis -- Syntactic context; Phonetics -- Periodicals; Phonetics -- Periodicals; Phonétique -- Périodiques; Phonetics; Periodicals; Electronic journals; 414.05
• DOI: 10.1016/j.wocn.2022.101152
• Languages: English
• ISSNs: 0095-4470

Abstract

Highlights: Functional relations between speech rate and phonetic variables are investigated. Non-restrictive and restrictive relative clauses (RC) are elicited at various rates. Post-RC boundary is more susceptible to variation in speech rate. Pre-RC boundary better reflects the distinction between RC types. Some phonetic variables exhibit ceiling/floor attenuation effects at extreme rates. Abstract: It is well known that speech rate is correlated with many phonetic variables. The current study aims to obtain a more precise characterization of how phonetic measures covary with speech rate. Specifically, we assess whether there is evidence for linear and/or non-linear relations with rate, and how those relations may differ between phrase boundaries. Productions of English non-restrictive (NRRCs) and restrictive relative clauses (RRCs) were collected using a method in which variation in speech rate is cued by the speed of motion of a visual stimulus. Articulatory and acoustic variables associated with phrase boundaries were analyzed; for each variable, Bayesian regression was used to obtain posterior parameter distributions for a set of generalized linear models. Analyses of posterior predictions showed that phonetic variables associated with a phrase boundary that follows the relative clause (post-RC boundary) were more susceptible to rate variation than those at a boundary that precedes the relative clause (pre-RC boundary). Phonetic variables at the post-RC boundary also showed evidence for non-linear relations with rate, which suggest floor or ceiling attenuation effects at extreme rates. On the other hand, substantial differences between syntactic contexts were found primarily at the pre-RC boundary. A high degree of participant-specificity was observed in F0-related variables.