Master of Arts, The Ohio State University, 1982, Linguistics

Experimental investigations of tone 3 sandhi in Mandarin Chinese have failed to find well-defined connections between sandhi and speech rates. In this study, empirical evidence is provided that indicates that the traditional claims of tone 3 sandhi are too arbitrary. Tone 3 does not necessarily change into tone 2 by all means. Furthermore, empirical, historical and theorectical arguments are raised concerning the validity of the duration of tone 3 in sandhi situation. Spectrograms and graphs are used to study durations and fundamental frequency of tone 3 in tone sandhi at slow, normal and fast speech by three females and three males. The degree to which the values of tone 3 are influenced appears to depend on various speech speeds. In addition, it is possible that greater variance in tone 3 duration and fundamental frequency is concomitant with the variance of speech speeds. The present study may be viewed as raising some interesting questions and careful cautions in future research on tone 3 sandhi.

Committee: Robert Fox (Advisor) Subjects: