The history of Chinese shows inherent variability: the dominant sentential position of adverbial PPs gradually moved from postverbal to preverbal position, from Early Old Chinese to Early Middle Chinese, and has resulted in stable variation from that time until the present day. This work aims to better understand the motivation and mechanism behind the historical positional shift of PPs and the non-categorical morpho-syntactic behavior of PPs in Modern Chinese.;The diachronic analysis of my study identifies a correlation between historical changes. The split of PPs is correlated with the emergence of the serial verb construction, a hypotatic structure which is capable of denoting delimitedness. I argue that the postverbal structure came to gradually attract PPs, which denote delimitedness, as speakers acquired a hypotatic method to express delimitedness using a serial verb construction in Late Old Chinese. Furthermore, I argue that, for the time after Late Middle Chinese up through Modern Chinese, a postverbal PP that immediately follows V, as in [V PP], must be in the gradual process of reanalysis as V-P NP, in analogy to the emergence of V-R NP in Late Middle Chinese.;I also present a logistic regression model that accounts for the dynamic variation in the sentential positions of PPs. The model confirms that economy within VP and delimitedness constraints are the main predictors for the gradual preverbal positioning of PPs in the dynamic variation. A monosyllabic V, a non-delimiting PP and a simple postverbal structure co-occur with preverbal PPs much more than with a disyllabic V, a delimiting PP, or a complex postverbal structure. The odds of preverbal positioning of PP for a monosyllabic verb and a non-delimiting PP increase consistently towards the later historical stages. The odds of a PP being in preverbal positioning for a simple postverbal structure increase, especially from Early Old Chinese until Late Middle Chinese.;The synchronic part of my study accounts for the morpho-syntactic variation of PPs in Modern Chinese. While the majority of PPs appear only in preverbal position (i.e. no variation), some subclasses of PPs may appear in a total of four different syntactic patterns. This work models the modern variants in Lexical Mapping Theory and proposes a language specific feature assignment for Goal and goallike roles in Chinese. The argument structure of a given lexical verb predicts the potential morpho-syntactic behavior of its argument as a theory-internal motivation for variation.;This work argues that the rise of the pressure of economy within the lowest VP (economy within VP), and the rise of a syntactic-semantic interdependent constraint for postverbal elements (delimitedness constraint), both helped to motivate the positional shift of PPs. This work proposes universal but violable optimality theoretic constraints and their ranking, as the mechanism of variation in the constituent order of PPs throughout the diachrony and the synchrony of Chinese.;In this work, I also propose a relative ranking of constraints as a mechanism for variation. Among high-ranking constraints, the delimitedness constraint prefers an alignment with non-delimiting adverbial phrases in the preverbal position, and delimiting adverbial phrases in the postverbal position. The high ranking of the economy within VP constraint prefers an NP argument over a PP argument within the lowest VP. The thematic hierarchy constraint ensures that a constituent order among arguments that obeys the UG thematic hierarchy is preferred over others within the lowest VP. The relative ranking of low-ranking constraints determines which variant surfaces as the optimal expression.;Exploring both dynamic and stable variation in the morpho-syntactic behavior of PPs provides us with a more comprehensive picture of the motivation and mechanism behind that variation. Variation emerges from competition among potential expressions in the evaluation system, by means of the ranking of constraints, both universal and language-specific.
展开▼
