In the currently dominant autosegmental-metrical (AM) theory of intonational phonology, intonational forms are derived from observed intonational contours without reference to their associated functions. Consequently, not only the categorical status of the resulting intonational components needs subsequent proof, but also the meaning of the intonational contours requires explanations outside the definition of the components. In order to counteract these problems and to better understand speech intonation, this dissertation investigates intonation systems of Mandarin Chinese and General American English through a functional approach---surface forms being analyzed through underlying linguistic functions. Specifically, the following theoretical issues are explored on the intonation of the two languages: (1) the functional domains of lexical tone/stress, focus, and sentence type, (2) the role that focus plays in distinguishing sentence types, and (3) the interaction between lexical tone/stress, focus, and sentence type.;Five experiments were conducted to address these issues. Experiments 1 and 2 investigated whether focus and sentence type could be produced and perceived simultaneously in Mandarin, and if yes, how they would interfere with each other. Experiment 3 aimed to identify feature vectors that are most effective in characterizing statements and yes/no questions in Mandarin, where decision trees were implemented in the classification of intonational contours. Experiments 4 and 5 examined whether focus and sentence type are realized differently through lexical items (tone vs. word stress) in Mandarin and English, and how the results are explained by the Parallel Encoding and Target Approximation (PENTA) model and the AM theory of English intonation.;The main findings include: (1) statement/question intonation is realized in parallel with focus and lexical items that also use pitch for their encoding in both languages, and (2) the similarities and differences between Mandarin and English intonation are essentially caused by the way sentence type interacts with focus and lexical tone/stress in the two languages. These findings are in support of the functional view of intonation, according to which components of intonation are defined and organized by individual communicative functions that are independent of each other but are encoded in parallel.
展开▼