TY - BOOK

T1 - Phonetic and dialectal variation in phonologically contrastive laryngealisation

T2 - A case study of the Danish stød

AU - Siem, Andrea

PY - 2024/2/16

Y1 - 2024/2/16

N2 - A growing body of research is documenting the variability of the phonetic manifestation of phonological contrasts in voice quality. This thesis is an empirical addition to the typology of cross-linguistic phonetic variation in the production of phonological laryngealisation. Investigating this variation is done through a comparative lens of dialectal differences within a language and with the Laryngeal Articulator Model (Esling et al 2019) as a guiding framework for investigating laryngealisation. Dialectal differences are analysed using the Danish stød as a case study, comparing speakers from Copenhagen and Aarhus. The stød has previously been reported to be highly variable in its production, making it an excellent source of phonetic variation. Further, the dialect of Aarhus has been speculated to contain a tonal stød variant under certain conditions (Kyst 2004) which has not yet been studied phonetically, making the inclusion of this dialect novel both in a typological perspective and in the context of the phonetics of the stød specifically. Data from 10 speakers of Modern Standard Copenhagen and 11 speakers from Aarhus is analysed using both acoustic and articulatory measurements. Acoustic measures include fundamental frequency, intensity, the amplitude difference between the first and second harmonic (H1-H2), Cepstral Peak Prominence, Harmonics-to-Noise Ratio and Subharmonic-to-Harmonic Ratio. The articulatory measures are obtained via Electroglottography, a novel method in the context of investigating the stød, which enables the vocal fold contact patterns to be modelled for analysis. Three main research questions are explored: (i) how much gradient phonetic variation in voice quality occurs during the Danish stød (ii) which acoustic and articulatory measurements correlate with the subtypes of the stød and (iii) how the voice quality changes are timed. The different phonetic types of laryngealisation are categorised in reference to five subtypes of creaky voice described in Keating et al (2015). To accommodate the expected variation in the Aarhus dialect, the stød in divided into two types for analysis, regular and tonal, both elicited in a contrastive minimal pair. The differences are analysed dynamically using two different statistical methods, Generalised Additive Mixed Models and random forest models. The findings generally confirm that investigating dialectal differences when exploring types of contrastive non-modal phonation is a rich resource to draw upon in widening our empirical understanding of phonetic variation in phonological voice quality across different languages. They demonstrate that stød in Danish is not just one type of stød, and even the standard Copenhagen variant exhibits differences in acoustics and articulation based on its stød basis. The study of timing finds high variability but uncovers some general patterns according to stød type and dialects, solidifying timing as a rich source of phonetic differences. Lastly, the findings from this study support the notion that the larynx is an active articulator and that various structures within it can affect voice quality independent of activity in the glottis. This encourages more research into how these different laryngeal structures interplay and in what ways this affects commonly used acoustic and articulatory correlates of these mechanisms.

AB - A growing body of research is documenting the variability of the phonetic manifestation of phonological contrasts in voice quality. This thesis is an empirical addition to the typology of cross-linguistic phonetic variation in the production of phonological laryngealisation. Investigating this variation is done through a comparative lens of dialectal differences within a language and with the Laryngeal Articulator Model (Esling et al 2019) as a guiding framework for investigating laryngealisation. Dialectal differences are analysed using the Danish stød as a case study, comparing speakers from Copenhagen and Aarhus. The stød has previously been reported to be highly variable in its production, making it an excellent source of phonetic variation. Further, the dialect of Aarhus has been speculated to contain a tonal stød variant under certain conditions (Kyst 2004) which has not yet been studied phonetically, making the inclusion of this dialect novel both in a typological perspective and in the context of the phonetics of the stød specifically. Data from 10 speakers of Modern Standard Copenhagen and 11 speakers from Aarhus is analysed using both acoustic and articulatory measurements. Acoustic measures include fundamental frequency, intensity, the amplitude difference between the first and second harmonic (H1-H2), Cepstral Peak Prominence, Harmonics-to-Noise Ratio and Subharmonic-to-Harmonic Ratio. The articulatory measures are obtained via Electroglottography, a novel method in the context of investigating the stød, which enables the vocal fold contact patterns to be modelled for analysis. Three main research questions are explored: (i) how much gradient phonetic variation in voice quality occurs during the Danish stød (ii) which acoustic and articulatory measurements correlate with the subtypes of the stød and (iii) how the voice quality changes are timed. The different phonetic types of laryngealisation are categorised in reference to five subtypes of creaky voice described in Keating et al (2015). To accommodate the expected variation in the Aarhus dialect, the stød in divided into two types for analysis, regular and tonal, both elicited in a contrastive minimal pair. The differences are analysed dynamically using two different statistical methods, Generalised Additive Mixed Models and random forest models. The findings generally confirm that investigating dialectal differences when exploring types of contrastive non-modal phonation is a rich resource to draw upon in widening our empirical understanding of phonetic variation in phonological voice quality across different languages. They demonstrate that stød in Danish is not just one type of stød, and even the standard Copenhagen variant exhibits differences in acoustics and articulation based on its stød basis. The study of timing finds high variability but uncovers some general patterns according to stød type and dialects, solidifying timing as a rich source of phonetic differences. Lastly, the findings from this study support the notion that the larynx is an active articulator and that various structures within it can affect voice quality independent of activity in the glottis. This encourages more research into how these different laryngeal structures interplay and in what ways this affects commonly used acoustic and articulatory correlates of these mechanisms.

U2 - 10.17635/lancaster/thesis/2274

DO - 10.17635/lancaster/thesis/2274

M3 - Doctoral Thesis

PB - Lancaster University

ER -