Proceedings of the 19th Conference on Laboratory Phonology - Long-term phonetic convergence vs. speaker-specificity - creaky voice in L2 English

Sep 11, 2026

Upload by Himbo

Long-term phonetic convergence in L2 English is examined through the creaky voice feature (CV) in a longitudinal speech corpus from University College Utrecht. The study targets whether Dutch L1 speakers acquire CV in English, use it in their L1, and show convergence over their three-year stay. Automatic low-f0 detection using the antimode classifier estimates creak prevalence. Results reveal high inter-speaker variation and limited intra-speaker change, with recording effects in spontaneous and L2 data and language-by-style interactions.

LanguageEnglish

CategoryPaper Templates

FormatPDF

Pages3

Reading Time1 min

AccessFree to download

Proceedings of the 19th Conference on Laboratory Phonology

Long-term phonetic convergence vs. speaker-specificity: creaky voice in L2 English

Method: automatic detection and antimode classification

Results: creak prevalence across time, language, and style

Utrecht University Repository | Title | Long-term phonetic convergence vs. speaker-specificity: creaky voice in L2 English | | --- | --- | | Authors | Sebregts , K. D. C.J. ; Vriesendorp , Hielke; Quené , Hugo; White , Yosiane | | Published in | Proceedings of the 19th Conference on Laboratory Phonology | | Publication Date | 2024-06 | | Link | [https://dspace.library. uu. nl/handle/1874/482601](https://dspace.library. uu. nl/handle/1874/482601) | | Citation | Sebregts , K D C J , Vriesendorp , H , Quené , H & White , Y 2024 , Long-term phonetic convergence vs. speaker-specificity: creaky voice in L2 English. in T Cho , S Kim , J Holliday & S Lee-Kim (eds) , Proceedings of the 19th Conference on Laboratory Phonology. Association for Laboratory Phonology, pp. 387-388. | | Versions / License | Publisher version | | Rights | [https://www. uu. nl/en/university-library/license-and-reuse-conditions](https://www. uu. nl/en/university-library/license-and-reuse-conditions) | Proceedings of the 19th Conference on Laboratory Phonology, June 27-29, 2024, HIPCS, Hanyang University, Seoul, Korea,(Eds. T. Cho, S. Kim, J. Holliday and S. Lee-Kim) Long-term phonetic convergence vs. speaker-specificity: creaky voice in L2 English Koen Sebregts, Hielke Vriesendorp, Hugo Quené, and Yosiane White Institute for Language Sciences, Utrecht University Phonetic convergence (also imitation, accommodation, alignment, entrainment) between interlocutors within conversations and in experimental tasks is by now a well-documented phenomenon, both between L1 speakers of the same language [1, 2, 3], and in the L2 speech of speakers with various L1 backgrounds [4] . While some researchers have made explicit claims that such short-term accommodation may provide the seeds for sound change in the longer term, this isnot uncontroversial [5, 6] . Longitudinal studies have shown accent change and phonetic convergence over the medium to long term, but to a modest extent and modulated by various factors [7, 8] . Our study uses a longitudinal speech corpus of a relatively tight-knit community to shed light on potential change and phonetic convergence over a number of years. An additional factor here is that the variety converged upon is L2 English. The D-LUCEA corpus [9] consists of recordings of students at University College Utrecht, an international campus college in the Netherlands where English is the lingua franca. Participants were recorded up to five times over their three-year stay in L2 English as well as their L1 . The subcorpus (n=29) used for the current study consists of the first and last recordings of all female speakers from the first cohort with Dutch as their L1 . Speakers in the corpus have previously been shown to converge over time on phonetic features, both segmental (the realisation of /s/ in their L1 and L2 [10]) and prosodic (L2 speech rhythm [11]), though not on their use of filled pauses [12], where they maintained speaker-specific patterns. The current study looks at potential convergence by these speakers on their use of creak/creaky voice (CV) . CV is a well-known social marker in L1 (especially American) English, where it is associated with upwardly mobile, urban, young female speech. It is also a feature that speakers have been shown to converge on in conversation [13, for Australian English], and that L2 speakers are able to learn through exposure alone [14] . CV has been much less studied in other languages; in Dutch, it has been described as rare and idiosyncratic [15, 16], though there have been recent mentions on social and in popular media, drawing explicit parallels with the sociolinguistic profile of CV in English [17] . We examine whether the students in our corpus acquire CV in their English, use it in their L1 as well, and whether they converge over time. Our method involved the automatic detection off0 [18] across students’ (read and spontaneous) speech, as many (though not all) types of CV involve low f0 [19]

FAQs

What is the document’s research focus?

It investigates whether speakers converge on creaky voice (CV) over long-term interactions and whether CV relates to speaker specificity in L2 English versus L1 Dutch.

How is creaky voice identified in the study?

CV is detected automatically using low f0 measurements, with the antimode (AM) serving as a cutoff to classify creaky (<AM) versus modal (>AM) voice and compute creak prevalence per speaker.

What do the results indicate about convergence over time?

Creak prevalence increases over time, but the effect is paired with limited intra-speaker variability and strong inter-speaker differences, suggesting CV behaves more like an idiosyncratic feature than a gradually acquired one.

Sep 11, 2026

Upload by Himbo