Compound And Transitivity

A transitive resultative construction (e.g.,他打跑了小偷/Ta da-pao-le xiaotou) is featured by a verb compound (V1+V2, e.g.,打跑/da-pao) in a transitive sentence form. This dissertation addresses two issues with regard to the transitive resultative construction. The first is the formation of the verb compound it contains, and the second is the choice of its subject and object (e.g.,三瓶酒/sanpinjiu and老王llaowang in三瓶酒喝倒了老王/Sanpinjiu he-dao-le Laowang).As to the first issue, the compounds are characterized by constrained productivity, (e.g.,打哭/da-ku is acceptable but*打笑/da-xiao is not). Existing research has three weaknesses in describing and explaining this charateristic. First, it stipulates that the compound codes a causative meaning via the syntactic structure of V2 functioning as the complement of V1. This syntactic representation of causation is not adequately justified. Second, the syntactic and aspectual constraints it proposes fail to constrain the compounding. The constraints make the compounding process both overgenerative (i.e., by licensing unacceptable cases such as*打笑/da-xiao) and undergenerative (i.e., by ruling out acceptable cases such as吓病/xia-bing). Third, it ignores the role of background information in licensing the otherwise unacceptable compounds. For example, semantic information associated with a certain syntactic or discourse context may increase the acceptability of some novel compounds such as留学生吻瘫机场/Liuxuesheng wen-tan jichang. With regard to the first weakness, this study establishes the compound's representation of causation as its constructional semantics and argues that various kinds of the compounds (e.g.,打死/da-pao,吃饱/chi-ba,学会/xue-hui,跑丢/pao-diu, and饿死/e-si) instantiate the constructional semantics in different ways. The constructional semantics explains the shared and individual features of each type of compounds such as the foregrounded status of V2. With regard to the second weakness, this study proposes that the compound codes a prototypical causation with an intended or expected result, which constraints the productivity of the compound construction. For instance, in打死/da-si ("beat to death"), V1 ("beat") leads to an intended or expected result ("death"), while in打笑/da-xiao ("beat to laugh"), the result is not normally intended or expected. With regard to the third weakness, this study argues that background information helps to license the elsewhere unacceptable compounds. It can be achieved by either using a certain syntactic structure associated with a strong causative meaning or supplying contextual information to help establish the causative relationship in the compound. For example, in reporting the event of the airport closedown as a result of the alarm triggered by students kissing in a restricted area, the newspaper uses the headline of留学生吻瘫机场/Liuxuesheng wen-tan jichang.The second issue of this study focuses on the choice of the subject and object in the transitive resultative construction. This issue has been tackled from three approaches in existing literature. In the compound-verb approach, the subject and the object are determined by the event structure coded in the compound. In the light-verb approach, a light verb is posited to select the causer subject. In the constructional approach, a construction determines the subject and object. All three approaches mix up two kinds of causations, that is, the causative relationship between two events and that between two participants. Moreover, they fail to constrain the choice of the causer subject (e.g., they fail to rule out the sentence*领导醉倒了老王/Lindao zui-dao-le Laowang).This study proposes that the transitive resultative construction (e.g.,老王打跑了小偷/Laowang da-pao-le xiaotou) results from the fusion between the transitive construction (e.g., the relationship between老王/Laowang and小偷/xiaotou) and the compound construction (e.g.,打跑/da-pao). Such fusion is motivated by the categorization between the transitive and compound constructions, and hinges on their semantic overlap. The fusion reconciles the two types of causation (e.g., the causative relationship between老王/Laowang and小偷/xiaotou and that between打/da and跑/pao) and accounts for all kinds of transitive resultative constructions in a unified manner, including the non-canonical types (e.g.,三瓶酒喝倒了老王/Sanping jiu he-dao-le Laowang and三瓶酒醉倒了老王/Sanpingjiu zui-dao-le Laowang). In particular, following the fusion operation, the research further constrains the choice of the subject by arguing that the subject must be licensed by the compound verb, thus ruling out sentences with an unlicensed subject (e.g.,*领导醉倒了老王/Lindao zui-dao-le Laowang), because the subject is not semantically related to the compound verb.Moreover, the fusion operation can account for the argument alternation of the compound (e.g.,喝醉/zuidao as in老王喝醉了酒/Laowang he-zui-le jiu and三瓶酒喝醉了老王/Sanping jiu he-zui-le Laowang). The alternation is attributed to the constructional semantics of the compound, which can be categorized by, and thus fused with, two kinds of transitive structures (e.g., from老王/Laowang to酒/jiu and from三瓶酒/sanpingjiu to老王/Laowang).This study also explains the three-way ambiguity of some transitive resultative constructions (e.g.,老王追累了老李/Laowang zhui-lei-le Laoli). In this example,累/lei can refer to the state of either the patient or the agent. In the former case, it produces the interpretation of Laowang chasing Laoli and Laoli getting tired. In the latter case, there are two possibilities累/lei can refer to the change of state of the agent, be it either老王/Laowang or老李/Laoli, thus giving rise to the two interpretations that Laowang chased Laoli and Laowang got tired, and that Laoli chased Laowang and Laoli got tired. Moreover, the present account rules out the fourth reading that Laoli chased Laowang and Laowang got tired, because in this case a categorization cannot be established between the compound and the transitive construction, without which the fusion is not licensed.This study has theoretical and empirical implications. Theoretically, a cognitive constructional analysis of both issues lends support to the constructional view of grammar and demonstrates the importance of categorization and encyclopedic semantics in sanctioning grammatical constructions. In particular, it sheds light on the sanctioning of the argument when the construction is fused with the verb. Empirically, the study has implications for the analysis of the intransitive resultative constructions and the de-resultative construction. The construconal analysis can be extended to them.