Consonant types can influence F0 values of the adjacent vowels (this F0 perturbation effect is henceforth referred as C-F0). C-F0 may be enlarged to maximize perceptual distinctiveness and thus reinforce aspiration contrast. In tonal languages, the effect may also be inhibited to cue tone contrast by constraining F0 variability by the demands of the lexical tone system. This dissertation investigates how C-F0 can be related to tone contrast and aspiration contrast by asking the following questions:(1) Is C-F0 conditioned by lexical tones? (2) Is C-F0 conditioned by F0 difference between tones? (3) Is there a cue trading relation between F0 and VOT for aspiration contrast?Fifteen Mandarin speakers and fifteen Cantonese speakers participated in the production experiments. All stimuli followed a CV template. The Mandarin stimuli had four Mandarin tones: M-T1(55), M-T2(35), M-T3(214) and M-T4(51). The Cantonese stimuli covered six Cantonese tones: C-T1(55), C-T2(25), C-T3(33), C-T4(21), C-T5(23) and C-T6(22). Initial consonants were aspirated obstruents, unaspirated obstruents and sonorants. F0 values following sonorants were the baseline for evaluating C-F0. The major results are as follow: for Question 1, the major findings are that tones can influence the vowel duration that C-F0 can extend and the difference between F0 following different consonants in Mandarin and Cantonese. A consistent direction was found: F0 following aspirated stops was higher than F0 following unaspirated stops and following sonorants. The trajectory of F0 following aspirated stops started as the second highest and converged with the lowest baseline F0 following sonorants. The results indicate a robust aspiration raising and a weaker voiceless unaspirated raising effect in both Mandarin and Cantonese.For Question 2, the pattern of the vowel duration that C-F0 can extend was found to follow the pattern of the F0 difference between the target tone and its closest tone in the tone inventory. However, the difference between F0 following different consonants within the first 10ms did not follow the patterns of F0 difference between tones in Mandarin or Cantonese. Finally, the cross-linguistic comparison provides support to the hypothesis that higher degree of tone competition may restrict C-F0.For Question 3, the results show that VOT is a strong cue for the aspiration contrast while onset F0 is a weak cue. However, the findings do not provide evidence for a cue trading relation between VOT and onset F0 in Mandarin or Cantonese. This dissertation has offered a new angle that few previous studies have dealt with before: traditionally, the enhancement of voicing contrast or the physical properties of producing voicing were proposed as the major trigger that give rise to C-F0. This dissertation has introduced a new contrast, i.e. the contrast of lexical tones, to explore the question. Furthermore, the major findings for all three questions confirm a hybrid account for explaining C-F0 in tonal languages: the tone enhancement account and the consonant automatic account.
Read
