2021/06/09 by Kikuo Maekawa · 1 citation
Psychology · Computer Science · Mathematics · Medicine · #Phonetics and Phonology Research #Speech Recognition and Synthesis #Speech and Audio Processing #Utterance #Articulation (sociology) #Vowel #Realization (probability) #Speech production #Place of articulation #Nasal vowel #Phonology #Linguistics #Audiology #Psychology #Speech recognition #Mathematics #Computer science #Medicine #Philosophy #Statistics
paper · pdf · doi:10.1017/s0025100321000050
openalex publication_date 2021/06/09 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/04
ABSTRACT Japanese moraic nasal /N/ is a nasal segment having the status of an independent mora. In utterance-medial position, it is realized as a nasal segment sharing the same place of articulation as the immediately following segment, but in utterance-final position, it is believed to be realized as a uvular nasal. This final-/N/-as-uvular view, which is wide-spread in the literature on Japanese phonetics and phonology, was examined objectively by use of a real-time MRI movie of the articulatory movement of eleven Tokyo Japanese speakers. It turned out that the utterance-final /N/ is realized in a wide range of locations on the palate from the hard palate to the uvula. GLMM modeling showed that the closure locations of the utterance-final /N/ can be predicted accurately from the identity of the preceding vowel. In addition, leave-one-out cross validation showed that the model can be generalized to new data. We conclude that the realization of utterance-final /N/ is not fixed to uvular; its place of articulation is determined largely by the property of the preceding vowel.