Devoicing of Historically Voiced Obstruents in Xiangxiang Chinese––an Acoustic-Phonetic Perspective

In this paper the present-day pronunciations of the historically voiced obstruents in Xiangxiang Chinese were examined from an acoustic-phonetic perspective. Results strongly supported an ongoing process of devoicing for the voiced stops, fricatives and affricates, which is conditioned by a couple of factors including sex of speaker, place of articulation, manner of articulation and historically tonal type. Specifically, historically voiced obstruents in syllables corresponding to the historical tone category Ru have all become voiceless unaspirated or voiceless aspirated, meaning the process of devoicing has already completed; in contrast, for those that occur in syllables corresponding to the other three tonal categories, i.e., Ping, Shang , and Qu , i) the historical devoicing is still in progress, as evidenced by a considerable amount of intra- and inter-speaker free variation in the pattern of voicing among “fully voiced”, “partially devoiced” and “fully devoiced” obstruents; ii) the degree of devoicing is the highest in voiced fricatives, followed by the voiced affricates, and is the lowest in voiced stops; and iii) the devoicing also tends to be more extensive for obstruents with a more posterior place of articulation, and for male than female speakers. Candidate explanations for these patterns of devoicing were also presented.