Constructional Schemas of Chinese Separable Words (LHCs)

The present thesis focuses on Chinese LHCs (líhé cí – separable words) within the framework of constructionist approaches to grammar. The constructionality of linguistic strings is defined here in terms of their degree of entrenchment, which can be measured by the association strength of components for decomposable linguistic strings like LHCs. In addition, the semantic integrity of a linguistic string also reflects its degree of entrenchment. Combining the association strength of component morphemes of LHCs and their semantic integrity, we examined the degree of entrenchment of LHCs. The data consisted of the Chinese Web corpus 2017 Simplified, i.e., a sub-corpus of zhTenTen17, comprising 13.5 billion words collected from the internet. This was analyzed using the online corpus query system in Sketch Engine. The results show that most LHCs are entrenched units, which can either be reflected by their semantic integrity or strong association strength. Such results explain why LHCs have long been seen as constructions. Moreover, the typicality of LHCs was measured by combining token frequency and type frequency of their separated uses, and compared with the degree of entrenchment of LHCs. The results show that there is a positive correlation between the typicality and entrenchment of LHCs. That is, more typical instances of LHCs are also more likely to be entrenched than less typical LHCs. We further examined the semantic and pragmatic reasons behind the separated uses of LHCs by applying collostructional analysis to separated constructions. In the present thesis, we investigated separated constructions used with aspect markers, which are also the most typical separated structures according to existing studies. The results revealed nuanced difference between formally similar separated structures, such as those with the aspect marker le, suggesting that these structures should be viewed as constructions with constructional meanings that can hardly be deduced just by the semantics of their components. In addition, we also analyzed the hierarchy of constructional schemas of LHCs and separated structures at different levels. By comparing low-level schemas and high-level schemas with instances of separated uses of non-LHCs, we designated low-level schemas as “basic schemas”, and conclude that these should be given a central position in the teaching and learning of Chinese as a second language.