This dissertation examines the phonological rule of High Vowel Devoicing (HVD) in Standard Modern Japanese, the dialect of the Kanto plane region. The data herein support the analysis of loss of high vowel voicing as both phonological and phonetic processes (Maekawa 1990; Kondo 1997; Tsuchida 1997). Phonological devoicing involves spread of the feature [+spread glottis] from the preceding obstruent, although it was noted in this dissertation that oral closure may also be playing a role. Phonetically, gestural overlap (as per Browman & Goldstein 1990; 1992) of the vowel's voicing gesture by the preceding glottal spread is also supported by the data (Tsuchida 1997 and references therein). However, while the data of Tsuchida (1997) limited gestural loss of voicing of high vowels to the environment of between voiceless fricatives, the data here suggests that this overlap occurs with high vowels in all devoicing environments.
Characterization of the various stages of vowel reduction within in a Feature Geometry framework are presented. It is also noted that the typical devoiced vowel is actually fricativized (that is, becomes a vowel-colored fricative), with the frication produced by airflow through a narrowed oral tract providing the energy for vocal resonance. The presence of fricativized voiced vowels in the current data set suggests independent manipulation of the features controlling glottal spread and oral closure; i.e. [spread glottis] and [continuant].
|
E-mail this message to a friend
