Social meaning in Scottish voices: A sociophonetic investigation of voice quality combining auditory-perceptual, acoustic, and qualitative approaches

Pearce, Joe (2024) Social meaning in Scottish voices: A sociophonetic investigation of voice quality combining auditory-perceptual, acoustic, and qualitative approaches. PhD thesis, University of Glasgow.

Full text available as:
[thumbnail of 2024PearcePhD.pdf] PDF
Download (31MB)


Voice quality – the characteristic features of a person’s voice – varies according to macro-social categories like social class, age and gender (Esling 1978b, Stuart-Smith 1999b, Beck & Schaeffler 2015), and within speakers as they present different personae to others (Podesva 2007). Here, I focus on laryngeal aspects of voice quality, also called phonation, and consider how combining auditory-perceptual, acoustic, and qualitative methods can enrich the study of voice quality and its social meaning. I take the context of Scottish voice quality, and ask how this varies by age, gender and region, and then use Interpretative Phenomenological Analysis (IPA) to consider how these meanings are constructed and understood by an individual speaker. I first considered the relationships between acoustic measures and descriptive voice quality labels, with a view to evaluating the feasibility of conducting an acoustic analysis at scale in spontaneous speech. 90 seconds of speech from 24 speakers of Glasgow, Lothian (in and around Edinburgh) and Insular Scots from the Scots Syntax Atlas (SCOSYA) (Smith et al. 2019), stratified by age (18-25 and 65+) and gender (male and female), were investigated in a linked auditory-perceptual and acoustic study. Phonation Profile Analysis (PPA), a novel method that takes the descriptive phonation labels from Vocal Profile Analysis (Laver et al. 1991[1981]) and applies them to short stretches of sonorants, revealed a prevalence of whispery and tense-whispery voice. PPA results for 2170 stretches were compared to automatic f0-based categorization of creak (Dallaston & Docherty 2019) and acoustic analysis of H1*-H2*, H2*-H4* and H4*-2kHz* and Cepstral Peak Prominence (CPP) taken using VoiceSauce (Shue et al. 2011) informed by the psychoacoustic model (Kreiman et al. 2021). I concluded that it would be possible to use these acoustic methods at scale and maintain interpretability of results: f0-based categorisation of creak and PPA coding of creak showed high agreement, and whispery, breathy, tense, modal, and tense-whispery voice showed different acoustic profiles each on at least one acoustic measure, aiding in the interpretation of later corpus results. I then used these methods with 180 seconds of speech from 95 male and female, older and younger speakers from SCOSYA (Glasgow n=48, Lothian n=28, Insular n=19). In contrast to previous findings by Stuart-Smith (1999b) and Beck & Schaeffler (2015), I found no gender difference in the use of creak, instead finding that Insular Scots speakers were particularly creaky. Analysis of spectral tilt and CPP revealed that female speakers used more tense phonation than male speakers, while younger speakers used a tenser and more near-modal quality than older speakers. I then combined this acoustic approach with Interpretative Phenomenological Analysis (IPA), a qualitative approach interested in how people understand major life experiences, in a case study of how Carrie, a Scottish transgender woman, uses her voice and understands her experience with it. Experiential themes highlight the importance of situational control in how she uses her voice: Carrie understands others to perceive a disconnect between her appearance and her voice, and while she expresses feeling self-conscious about her voice and feeling a need to change it for others in everyday situations, she takes joy in using her voice professionally and exploiting the perceived disconnect her voice creates in situations where she is in control. She uses harsh voice to voice how she believes others see this disconnect, drawing on iconised links between harsh voice and wider societal attitudes towards trans women as monstruous. Through combining auditory-perceptual analysis, automatic categorisation of creak, multiple measures of voice quality, and in-depth qualitative data, this research gives a fuller picture of view of social meaning in voice quality, by not only tracing how multiple aspects of voice quality vary according to macro-level social factors for Scottish speakers, and what one Scottish speaker’s own use of voice quality means to her, and where her understanding of her voice fits into the wider landscape of social meaning in Scottish voices.

Item Type: Thesis (PhD)
Qualification Level: Doctoral
Additional Information: This work was supported by the Economic and Social Research Council (Grant number: 2178789).
Subjects: P Language and Literature > P Philology. Linguistics
Colleges/Schools: College of Arts & Humanities > School of Critical Studies > English Language and Linguistics
Funder's Name: Economic and Social Research Council (ESRC)
Supervisor's Name: Stuart-Smith, Professor Jane and Cohen, Dr. Clara
Date of Award: 2024
Depositing User: Theses Team
Unique ID: glathesis:2024-84331
Copyright: Copyright of this thesis is held by the author.
Date Deposited: 22 May 2024 14:53
Last Modified: 22 May 2024 14:58
Thesis DOI: 10.5525/gla.thesis.84331

Actions (login required)

View Item View Item


Downloads per month over past year