The Contribution of Voice Quality to Politeness in Japanese

This study investigates the contribution of voice quality to the expression of “positive politeness” under varying conditions of relative social status between male speakers of Japanese. In previous studies, suprasegmental features have been found to express politeness. Bandwidth of the first formant (BW1) was also found to be relevant to perceived politeness. However, the association between BW1 and voice quality is complex. BW1 could be a parameter indicating breathiness, whispering, or nasality. In order to clarify the contribution of voice quality, in this study, spectral tilt and aspiration noise in high frequency region were observed as the focus of interest. Vowel data (/a/) extracted from natural utterances produced by two male native speakers of Tokyo dialect were observed. For measurement of breathiness, the method suggested by Hanson (1995) was employed in order to avoid known disadvantages of inverse filtering. Acoustic parameters were observed separately between two speakers, one whose overall BW1 was wide and whose speech were perceived as polite, and the other whose BW1 showed negative correlation with perceived politeness. The result indicates two tendencies. Firstly, when a speaker had overall wide BW1 and was rated polite overall, the ratings with aspiration noise in high frequency region showed positive correlation with politeness. Secondly, when a speaker showed negative correlation between BW1 and politeness, the speaker also showed negative correlation between spectral tilt and politeness, but showed no correlation with aspiration noise rated. Overall, it seems that when a speaker is rated polite, the speaker’s politeness is associated more with aspiration noise, and therefore related to breathiness.