Previous studies argue that there are two types of passives, namely long and short passives for Mandarin bei passive sentences, and it is suggested that they have different structures, namely that long passives have the Agent present at the syntactic structure, with the Agent completely missing from the syntactic structure in short passives (Huang, 1999; Tang, 2004, 2006; and Xiong, 2003), which is called the structure separation approach in this dissertation. The other view on the relation between the two types of passives is the ellipsis/deletion approach, namely that short passives are derived from the result of omitting the Agent from long passives in syntax (Lü, 1980; Li, 1994). Which approach is more appropriate and correct is one of the questions that this dissertation tries to answer.
Unlike previous studies which are conducted at the sentential level, this dissertation investigates the bei passive sentences from a discourse perspective. Specifically, this dissertation takes a Centering perspective which examines the transition states from one utterance to next and investigates the roles of Mandarin bei passives in discourse. A total of more than 100 three-utterance mini-discourse segments with the bei utterance in the middle is examined on the basis of 11 novels and the data analysis suggests that the function of both types of bei passives is to preserve the backward-looking center (Cb) from its preceding utterance and pass it to its following utterance, though the roles of the bei-object in these two types of passives are different: while the overt bei-object helps prepare the introduction of a new Cb for the following utterance, the covert bei-object can only help retain the original Cb from the previous utterance. The data analysis also reveals that both types of bei-objects can function as the Cb for the bei passive sentence, linking the bei utterance with its previous utterance, and they can also be the antecedents of the pronouns in the following utterance, which suggests that the bei-object must be realized in syntax, namely occupying a syntactic position, unlike what is suggested by the structure separation approach. This finding argues for the ellipsis approach to the relation between the two types of bei passives in Mandarin Chinese. To further support this proposal, Mandarin ba-construction is incorporated into this study, and a similar inter-sentential function is also observed.
Furthermore, this dissertation makes a series of additional comparisons. It compares the division of labor between the pre-bei/ba topic and the bei/ba-object, the different behaviors of bei/ba-sentences in Narrative and Report modes, the different properties of the Centering framework and Givón's tradition, and so forth. These comparisons have led to the following conclusions:
a. From a Centering perspective, different center transition states witness distinct constructional patterns with a strong homogeneous nature within each pattern. Considering these patterns, an inter-sentential function for ba/bei-utterances is thus proposed, as pointed out earlier. Similar to the bei case, the pre-ba topic also preserves the Cb from its preceding utterance; while the ba-object, competing with the post-verbal Recipient, functions to introduce a new Cb for the following utterance or retain the original Cb from the previous utterance;
b. From a discourse mode viewpoint, it is shown that the bei passives in the Narrative mode perform differently from those in the Report mode in aspects such as referential choice, transition states, and possible violations of the Zero-anaphora Rule. These differences are attributed to their distinct interpretation patterns of tense, i.e. continuity vs. deixis in this dissertation, and it is pointed out that the bei passives in different modes bear distinct discourse functions;
c. By comparing the covert bei-object with the semantically inferable entity as well as the logical subject in the Middle Construction, it is shown that only the covert bei-object could act as a center. This result further vindicates that a covert bei-object also occupies a syntactic position, as argued earlier. Hence a uniform syntactic structure is proposed for both long and short passives, though the long and short passives differ in their discourse functions - the former is adopted to introduce a new center so as to prepare for a possible Cb Shift, whereas the latter is to avoid the introduction of unnecessary centers so as to keep the information flow in a more fluent manner, as pointed out earlier; and
d. Last by not least, by comparing the Centering approach with Givón's tradition towards discourse analysis, we think that Centering Theory is the true analytical tool of calculating the degree of discourse coherence, while Givón's tradition is better treated as a measurement for discourse relatedness, not directly the discourse coherence.
Unlike the previous studies, which follow Givón's (1983) tradition, such as Xing (1990), Myhill & Xing (1994), and Liu (2011), which scrutinize the contribution of Mandarin bei passives related to topic continuity, the theoretical perspective of Centering Theory, as adopted in this dissertation, explores in-depth the unnoticed linguistic facts about Mandarin ba and bei-constructions, and thus helps provide a more comprehensive portrait concerning the discourse functions and syntactic properties of these two constructions in Mandarin Chinese.
| Date of Award | 2 Oct 2013 |
|---|
| Original language | English |
|---|
| Awarding Institution | - City University of Hong Kong
|
|---|
| Supervisor | Haihua PAN (Supervisor) |
|---|