A Comparative Study Of Chinese And English Middle Constructions

Middle construction, as a special syntactic and semantic structure, has aroused numerous discussions and become a hot controversial issue in linguistic field over the recent decades. Especially, the definition and derivation process of middles have exerted all kinds of ideas.Based on the previous study, this paper is attempting to give a unified answer about Chinese and English middles under the theory of generative grammar. Active form with a passive meaning, this nature of middles is generally acknowledged by the field. It also shows other properties. Semantically, they are non-eventive, with a generic meaning, which means the middles have an implicit argument. Syntactically, they always contain a adverbial adjunct and the internal argument is realized as the surface subject. Verbs also have something to do with the acceptable form. The unique features of middles differs it from other similar structures.Linguists bring forth different criteria about the defining of the middles. Two of them are dominant: implicit agent argument criterion and modification criterion. Representatives as Iwata and Keyser & Roeper think the implicit agent argument should be the defining criterion of middles. However, Massam advocates the modification approach. This paper proposes a combination approach of syntactic and semantic in defining middles. Syntactically, the surface subject of the sentence must be theme or patient, excluding the experiencer. Roberts points out that the logical subject of middles must be the agentive argument. Verbs of psych and senses are also excluded. Semantically, middles always with an implicit agent, refers to situations without any specific time reference.On the basis of affectedness effect and aspectual classes of verbs brought by some linguists, this paper states that only the volitional transitive verbs can form grammatical middles. Also, the responsibility of the subject offers some constraints on it. About the selection of the adjuncts, we think the semantic features of the verb and the properties of the subject work together to decide which adjunct they are going to pick.Semantic approach and syntactic approach are the two methods about the derivation of middle constructions. Semantic approach argues that middle formation is strictly a pre-syntactic operation, mainly because the external argument doesn't project in syntax and the verb becomes one-place verb, then according to externalization, the internal argument turns to be the subject of the sentence. While the syntactic analysis proposes that middles form in syntactic representation, and the external argument of the middle construction is syntactically active. Only because middle verbs (like passive verbs) cannot assign case to its internal argument, so the internal argument has to move to the subject position to get a case. This paper agrees with the latter, for the lexicon approach cannot hold water when facing many realistic language problems. For example, in many cases, the implict agent can be recovered. Also, we know that in passives, we can insert a preposition by or bei to lead the cover agent, but it is not allowed in middles. Thus, we wonder this is because a implicit agent has already existed.This paper borrows ideas from Levin & Rappaport and Gu Yang, who think besides lexical-syntactic representation, another interface exists between lexicon and syntactic representation, which is called lexical-semantic representation. A verb in this interface can only have thematic role. Only when the verb entering into lexical-syntactic representation interface, its thematic role can be an argument. If a thematic role is suppressed in lexical-semantic representation, it cannot display in lexical-syntactic representation, let alone syntactic representation. So, on middle formation, we claim that in lexical-semantic representation, the verb is a transitive with external and internal thematic role. But after it entering into the lexical-syntactic representation, its external thematic role is suppressed, so we cannot find its external argument in syntactic representation, but can still feel the agentive meaning. At last, the internal argument is raised up to the subject position to fulfill Extended Projection Principle. This paper also assumes that the requirement of the subject has something to do with the motivation to suppress the external argument, and it causes a transitive verb turns into a stative one and cannot assign external case.This thesis also displays some demerits and problems. For example, it only gives semantic understanding about why middle verbs lack of the ability to assign case, but mentions little about its syntactic motivation.