The unaccusative hypothesis and the syntax of archaic Chinese


Student thesis: Doctoral Thesis

View graph of relations


  • Mengbin LIU


Awarding Institution
Award date15 Jul 2015


Perlmutter (1978) puts forth the Unaccuative Hypothesis that one-place verbs are divided into unaccusative and unergative types. Based on that, Burzio (1986) claims that two-place verbs are divided into transitive and causative verbs. The unaccusative hypothesis has been applied to Chinese by many linguists (cf. Li,Y.-H 1985, Lü 1987, Huang 1989, etc.). Based on previous studies, Huang (2007) claims that all verbs in Mandarin Chinese can be divided into unergative type and unaccusative type. According to Huang, transitive verbs (such as da 打 "beat", ma 罵 "scold" and xie 寫 "write") and three-place rob-type verbs (such as bo 剝"peel", ti 踢 "kick" and qiang 搶 "rob") are unergative verbs. Causative verbs (such as kai 開 "open", guan 關 "close", chen 沉 "sink" and xia 嚇 "frighten") and give-type verbs (such as song 送 "send" and gei 給 "give") are unaccusative verbs. The studies of Mandarin Chinese under the unaccusative hypothesis focus on the object in-situ constructions (cf. Xu 1999, Han 2000, Han and Pan 2005, 2008, etc.), while the studies of classical Chinese under the unaccusative hypothesis focus on causative verbs (cf. Song 2005, Sun 2006, etc.) However, there is a lack of studies about Archaic Chinese from the perspective of formal syntax and the newly developed theory postulated by Huang (2007). Therefore, in this thesis we try to study verbs and the related constructions of Archaic Chinese from the perspective of formal syntax and under the framework of Huang (2007). First, we discuss the "NP theme+V unacc" construction in detail by comparing it with the object-preposing (after subject) construction, the topic construction and the middle construction. We argue that it is different from object-preposing construction and typical topic construction in Archaic Chinese. Also, it is different from the middle construction. We find that the ways of differentiating English unaccusative verbs from middle verbs cannot be applied to Archaic Chinese. This kind of comparison is significant as it hasn't been done before. Second, we discuss the formation of "NP1 exp+V+NP2" ("王冕死了父親") construction from the perspective of historical development and formal syntax. We argue that the "NP1 exp+V+NP2" construction comes from a morphological or lexical causative construction and the "V+NP theme" ("來了一個人") construction are actually two-place unaccusative sentences with omitted causer or experiencer. Only the NP theme+V" ("一個人來了") construction is the real one-place unaccusative construction in which the NPtheme generates in the object position in deep structure and moves to the subject position in surface structure because of case requirement. Several pieces of evidence have been provided in this thesis. The historical relationship between "NP1 exp+V+NP2" construction and causative construction was not considered in detail in the previous studies. We take it into consideration and provide more reasonable explanations for the formation of "NP1 exp+V+NP2 construction. Third, we investigate the one-place unergative verbs and two-place unergative verbs (transitive verbs) in Archaic Chinese under the framework of Huang (2007). In Huang (2007), it is argued that a patient may be inserted to a one-place unergative verb and form a two-place unergative verb. A two-place unergative verb may become a one-place unergative when it doesn't take a patient. The result of our investigation shows that in Archaic Chinese only some of the one-place unergative verbs may have an additional patient argument and form two-place unergative verbs. When a patient doesn't appear after a two-place unergative verb, it is hard to tell whether the patient is omitted or it is a one-place unergative verb. Fourth, we discuss whether the rob-type double object construction exists in Archaic Chinese. We mainly investigate the syntactic properties of "duo 奪 +zhi 之 +noun" construction by applying movement test to the nouns after the main verb duo 奪, and we also consider the syntactic properties of "duo 奪 +noun+noun". We argue that it is more reasonable to analyze "duo 奪 +zhi 之+noun" as double object construction but the "duo 奪 +noun+noun" construction can either be analyzed as a double object construction or a transitive construction with a genitive phrase. Fifth, we discuss the relationship among the three constructions: "V+DO+IO", V+IO+DO" and "V+DO+yu 于 +IO". We argue that none of them can be at the same time the deep structure of other two constructions. It is possible that the "V+DO+IO construction is derived from "V+DO+yu 于 +IO" by omitting the preposition yu 于, and the "V+IO+DO" is not directly relevant to the other two constructions. However, we cannot deny the possibility that the "V+IO+DO" construction behaves as the basic structure of "V+DO+IO" structure. Sixth, we investigate whether the syntactic criteria used in Huang (2007) to distinguish between rob-type construction and give-type construction can be applied to Archaic Chinese. Results show that passivization and yi 以 phrase in Archaic Chinese cannot tell us the difference of syntactic structures between these two types of constructions. Therefore, we conclude that syntactic evidence is not enough for dividing the three-place verbs into unergative type and unaccusative type. The only thing we can be sure is that the yi 以 phrase, rather than yu 于 phrase, makes the difference between rob-type verbs and give-type verbs in Archaic Chinese. Seventh, we discuss what functions unaccusative verbs and unergative verbs have in general. We argue that when the verb stands in the subject or object position, it behaves as a noun because of the existence of zero nominalizer. We also argue that when the verb stands in the position of prenominal modifier, it actually behaves as the predicate of a prenominal relative clause. Eighth, we investigate how yu 于 phrase acts in both active sentences and passive sentences. We find that the theta-role of the NP following yu 于 is almost decided by the properties of the main verb. As yu 于 phrase can be used in all kinds of sentences, yu 于 is obviously not the marker of passive sentences. In passive sentences, the NP after yu 于 which used to be analyzed as agent can be also analyzed to other roles. Ninth, we discuss whether the passive sentences with jian 見/wei 為/bei 被 are syntactically true passives. For the unmarked passive constructions, we compare them with the object-preposing construction, topic construction and ergative (ie. unaccusative) sentences. The result shows that the unmarked passive construction is different from the object-preposing construction and the typical topic construction. It is also different from ergative sentences. Lastly, we discuss the causative constructions from perspective of the unaccusative hypothesis and formal syntax. For the morphological and lexical causatives, we discuss the properties of the causative verbs, argument structure and syntactic representations. We put more emphasis on the three-place verbs which are used as causative verbs. For the syntactic causatives, we investigate the properties of the causative verb V1 and V2 . What's more, we discuss in detail the syntactic structures of causatives with overt causative verbs shi 使 and ling 令.

    Research areas

  • Syntax, Verb, To 600, Chinese language