Harsh voice and its interaction with vowel quality in Fuzhou Min Chinese

Fuzhou Chinese (Fuzhounese), a variety of Eastern Min, is known for its tone-vowel interaction. The tones /21, 241, 24/ are associated with lower, diphthongal vowels and non-modal phonation, while tones /44, 51, 32, 5/ are associated with higher, monophthongal vowels and modal phonation. This study tests the hypothesis, based on the laryngeal articulator model (Esling et al., 2019), that epilaryngeal constriction responsible for non-modal phonation mediates vowel quality alternation due to articulatory synergy with lingual retraction. The most common phonation type observed among 18 speakers is harsh voice (with occasional aryepiglottic trilling). Creaky and modal voice, and less often ventricular or whispery voice, are also observed. Spectral tilt and noise measures reveal that noise (cepstral peak prominence and harmonic-to-noise ratio) is most reliable in distinguishing modal from non-modal tones. The only spectral tilt measure that does so is H2-H4, while H1-H2 differentiates subtypes of constricted phonation. The relationship between phonation and vowel quality is also examined. Noisier phonation is associated with higher F1 for all vowels, lower F2 for [ei], and higher F2 for [ou], in accordance with cross-dialectal variation (Chen, 1998), supporting the hypothesis that tone-vowel interaction is mediated by epilaryngeally constricted phonation.