Abstract:
This research study investigates the prosodic constituents in Thai. It first identifies prosodic boundaries using experimental data. It then analyzes the prosodic boundary strength based on Speech-Pause Score and classifies the prosodic categories according to the strength. Lastly, it analyzes the mechanism involved in the mapping between syntax and prosody. The data comprised 107 sentences including 1) simple, compound and complex sentences; 2) thii3 or sɨŋ3 relative clause constructions, thii3 or waa3 verb-complement clause constructions and thii3 noun-complement clause constructions; 3) independent phrases that have no grammatical relation with other words in the sentence; and 4) thuuk2 or doon1 passive constructions and haj3 or tham1haj3 causative constructions. Two experiments were conducted in order to locate the prosodic boundary of each sentence. Subjects were asked to judge which positions between words can potentially be used as pauses. Prosodic boundaries can be categorized into three groups according to the syntactic constituents that they are aligned with. In order to find the levels of prosodic boundary strength, one-way ANOVA and post-hoc Scheffes test were performed. The results show that there are two levels of prosodic boundary strength: the intonational phrase and the phonological phrase. To investigate the boundary strength of vowel durations, the recorded sounds of 4 informants (NC, NK, NM, and NN) were measured. To control the effects of differences in segmental composition of rime and differences in duration affected by stress associated with content words, vowel duration was classified by rime structure and then by word class before it was normalized to a standard score. In order to test whether prosodic boundary strengths corresponded to pauses or to the syntactic status suggested by the analysis under the experiment, two methods of grouping the standard scores were used. In the first method, the standard scores were classified into three groups--the standard scores of syllables occurred before pauses (P), the standard scores of final syllables of the sentences (F) and the standard scores of the syllables that are in other positions (O). In the second methods the standard scores were classified into three groups corresponding to the types of the syntactic constituents that they are aligned with. The results show that the prosodic boundary strength varies with the position of the pauses, not the syntactic status. The three factors influencing prosodic phrasing in Thai were syntactic factor, information structure factor, and prosodic factor. The syntactic factor played a vital role in phrasing both the intonational phrases and the phonological phrases. An intonational phrase refers to the right or left edges of a clause that are inputs in the intonational phrasing mechanism. The information structure factor played a critical role in cases where the information unit functioned as a topic which is an intonational phrase. In spite of that, the topic necessarily referred to the right or left edges of the syntactic constituents in demarcating the intonational phrase. In phonological phrasing, the syntactic factor and the prosodic factor competed with each other. In order to balance the size of the phonological phrases in accordance with the Uniformity Principle and Symmetry Principle, some syntactic edges were disregarded and, instead, the syntactic edge within the verb phrase was chosen in phonological phrasing. As choosing the syntactic edge within the verb phrase cannot satisfy those principles, the syntactic edges were referred to again and the prosodic constituents and the syntactic constituents were isomorphic as a result.