| This dissertation covers three areas of phonetic study of Chinese tones: (1) the interaction between tone and vowel, (2) the interaction between tone and consonant, and (3) acoustical analysis of tone sandhi. Five major Chinese dialects, Standard Mandarin, Southwestern Mandarin, Shanghai, Taiwanese and Cantonese, were used for investigation. The material is organized in seven sections.;Section 2 investigates the effect of tone on vowel quality in Taiwanese. Results show that (a) vowels {i}, {e}, {o} and {u}, not however {a}, associated with the high tone occupy areas in the acoustical vowel space that are distinct from those occupied by the same vowels associated with the low tone, and (b) the average fundamental frequency of the high tone correlates with vowel height. No such correlation is found in vowels produced with the low tone.;Section 3 investigates the difference between the effects of {p('h)} and {p} on the fundamental frequency onset of the following diphthong {ei} in Cantonese. Results show that the F(,0) onset for {ei} following {ei} is higher, however, the intensity onset is higher for {ei} following {p}, which implies that a higher F(,0) may be produced with a decreased subglottal pressure.;Section 4 investigates whether the duration of the prevocalic {s} is influenced by the tone on the following vowel and to examine whether temporal compensation exists between the duration of the fricative and the duration of the vowel within a certain tone class. Results show that (a) the duration of the fricative is influenced by the tone on the following vowel, and it varies according to the onset value of the tone, and (b) within each tone class, temporal compensation does not exist between the durations of the segments {s} and {i} in the syllable {si}.;Section 1 investigates whether tone correlates with both duration and intensity in Taiwanese. Results show that (a) the high tone has an overall higher intensity than the mid tone and the mid tone has an overall higher intensity than the low tone, (b) the high-falling tone has an overall higher intensity than the low-rising tone, and (c) the rising tone is longer than the level tone and the level tone is longer than falling tone.;Section 5 investigates whether the peak intraoral air pressure in a voiceless unaspirated stop {p} is affected by the tone on the following vowel or diphthong in Mandarin, Chungking, Shanghai and Cantonese. Results show that (a) there is a good correlation between the peak intraoral air pressure in {p} and the F(,0) onset of the tone on the following vowel or diphthong, and (b) the amount of change in subglottal pressure cannot account for the large variation of F(,0).;Section 6 investigates whether the pitch contours on monosyllabic words and on compounds are actually similar. The results show that tone spreading will partially explain the patterns of tone contours, on the compounds, but that there are also some more arbitrary processes involved in tone sandhi in Shanghai. A phonological analysis is presented which regards contours as sequences of level tones, and tones are treated as having an 'autosegmental' association with syllabic units.;Section 7 is a spectrographic investigation of the tone sandhi at the phrase and sentence levels in Mandarin. Results show that none of the three claims proposed by Chao, 1948, 1968 and Cheng, 1968 applies to tone sandhi at the productive or acoustic level in the speech of today's young generation of Peking Mandarin speakers. |
