Devoicing is a process whereby a voiced consonant such as /bdg/ is realized as voiceless [ptk]. Some theorists [1,2] propose that this phenomenon is an instance of fortition, or consonant strengthening, especially when it occurs word-initially. This study proposes an in-depth exploration of voicing alternations in word-initial position in five Romance languages (Portuguese, Spanish, French, Italian, Romanian) using large corpora (ca. 1000h of speech) and automatic alignment. Our results show that (i) there is initial devoicing in all languages, and (ii) this devoicing is conditioned by the preceding context. This allows the languages to be divided into those displaying (a) only phrase-initial fortition (Spanish), (b) phrase-initial and post-obstruent fortition (French, Romanian and possibly Italian) and (c) generalized word-initial fortition (Portuguese).
Cite as: Hutin, M., Wu, Y., Jatteau, A., Vasilescu, I., Lamel, L., Adda-Decker, M. (2021) Synchronic Fortition in Five Romance Languages? A Large Corpus-Based Study of Word-Initial Devoicing. Proc. Interspeech 2021, 996-1000, doi: 10.21437/Interspeech.2021-939
@inproceedings{hutin21_interspeech,
author={Mathilde Hutin and Yaru Wu and Adèle Jatteau and Ioana Vasilescu and Lori Lamel and Martine Adda-Decker},
title={{Synchronic Fortition in Five Romance Languages? A Large Corpus-Based Study of Word-Initial Devoicing}},
year=2021,
booktitle={Proc. Interspeech 2021},
pages={996--1000},
doi={10.21437/Interspeech.2021-939}
}