Tongue action in silence: Japanese voiceless vowels

Abstract

The study presents acoustic and articulatory analyses of the Japanese voiceless vowels [(i)] and [(u)] arguing Japanese listeners have internal representations for the vowels and perceive them from articulatory information while physical properties of vowels are absent. Ogasawara (2006) reports that there is no formant band in Japanese voiceless vowels. It is an open question as to whether there is an articulatory gesture associated with the voiceless vowels, and whether this gesture is the same or different as that of voiced vowels. To answer the question, we investigated production of voiceless vowels by a native Japanese speaker. We created 11 pairs of nonsense words such as [hogito] vs [hok(i)to] and [hoguto] vs [hok(u)to]. We recorded the speakers production of the stimuli with the voiceless vowels and their voiced counterparts using the ultrasound device to capture tongue images. The same recording was examined for acoustic analysis. An investigation of the ultrasound images reveals there are tongue gestures for the voiceless vowels, movements differ depending on which voiceless vowel it is, and the movements parallel those of the voiced counterparts. These results argue in favor of the view that Japanese speakers internal representations of words include a representation of distinct voiceless vowels.