Abstract
Diphthongs in Mexian Hakka Chinese have been divided into two categories (Cheung 2007): Endpoints of the falling (in vowel space) diphthongs have formant values equivalent to those of the monophthongal vowels, while the endpoints of the rising diphthongs generally have more centralized formants. This categorization provides a chance to test the hypothesis of Iskarous etal. (2010) that in speech production the location of constriction (CL) changes discretely in transitions between successive targets, while the degree of constriction (CD) changes continuously. We therefore hypothesize that the centralization in these diphthongs results from a reduction in CD (possibly due to undershoot) rather than CL, thus making it parallel to common consonant reduction changes in languages, such as spirantization of stops. The current study tests this hypothesis using TADA (TAsk-Dynamic) modeling, in which CD and CL in speech production can be manipulated. Diphthongs [ia], [ua], [ai], and [au] were modeled, and the preliminary results indicate that a change in target CD alone yields formant patterns that are more similar to those reported compared with a change in CL. Results of additional studies will be presented that analytically derive CD and CL in these diphthongs from new data collected from speakers of Hakka.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have