Coarticulation of non-modal phonation.

Abstract

Although there are many studies on the acoustics of phonation, few deal specifically with the coarticulation of non-modal phonations. Using data from three languages (English, White Hmong, and Korean), this project describes the acoustics of breathy-to-creaky phonation contours in vowels. In English, such contours are found in vowels preceded by [h] or an aspirated stop, and followed by a preglottalized coda-stop, e.g., “cat.” In Hmong, the contours are found in vowels with a phonemic creaky tone that are preceded by an aspirated stop or [h]. In Korean, the contours are derived from vowels preceded by an aspirated or lenis stop and followed by a coda-stop, which is found to show preglottalization. Twelve speakers of each language were recorded saying monosyllables with expected phonation contours in carrier sentences. The phonation was analyzed using VOICESAUCE [Shue et al. (2009)], which outputs a value at each millisecond for several acoustic measures. Time courses for measures like H1*-H2* and harmonics-to-noise ratio show support for the presence of breathy-creaky contours in the three languages. I analyze the cross-linguistic differences in the timing of the contours in light of previous findings on the timing of non-modal phonation [e.g., Blankenship (2002)] and current theories of coarticulation.