The Reexamination Of The Third Tone Sandhi In Beijing Chinese

Third tone sandhi is a process in Beijing Chinese which states that the first T3 is changed into T2 when T3 comes before another T3 in the same domain. However, the rule itself says nothing about the domain or scope that circumscribes the application of tone sandhi. The question that arises is whether the application of third tone sandhi rule is conditioned by syntactic factor or prosodic factor. This thesis attempts to explore the actual tonal behaviors in T3 sequences at a normal speaking rate, to investigate the evidence further, and to try to understand the conditioning factors and their interactions.Phonetic experiment and OT analysis are both used in this thesis to have a comprehensive study on third tone sandhi in Beijing Chinese. In the phonetic experiment part, the recording data (of a male informant and a female informant) of citation forms, disyllabic T3 sequences, trisyllabic T3 sequences and quadrisyllabic T3 sequences are measured by the phonetic software Praat (Version 4.2.29) and we get the FO value of each syllable in all the sequences, as FO value is the most important parameter to determine the tone category. Then the FO values are converted to Chao's tone letters which are convenient for us to see the tonal behaviors of the sequences. The sequences are subclassified according to their syntactic structures. Whether the application of third tone sandhi rule is isomorphic to syntactic structure can be concluded by observing whether we can get the same tonal behavior in the sequences belonging to the same syntactic pattern. In OT analysis, those T3 sequences are organized into prosodic units called Minimal Rhythmic Units (MRUs) by OT evaluation. Then the resulting prosodic structures serve as input to third tone sandhi rule and finally the predicting readings are presented. Whether the application of third tone sandhi rule is isomorphic to prosodic structure can be concluded by observing whether the final readings we get through OT formation of prosodic structures keep with the experimental results.Through the phonetic experiment, we get the actual tonal behaviors of all the T3 sequences at a normal speaking rate. Through the whole analytical process, we conclude that third tone sandhi in Beijing Chinese operates on prosodic structures which are sensitive to, but by no means isomorphic to syntactic structures. A central proposal is that the domain of third tone sandhi is a new prosodic category called the Minimal Rhythmic Unit (MRU). The MRUs are isolated as the domain of obligatory tone sandhi through the OT evaluation process, and then the resulting prosodic structures serve as input to the third tone sandhi rule. MRU construction is stratum-ordered:it applies (cyclically) at the lexical level (or we can say "at the word or compound level"), then iteratively across the board at the phrasal level. Likewise, third tone sandhi applies first within the MRU, then (optionally) across MRUs.