Table of Contents:
The topic of this study concerns the incremental sentence processing of transitive constructions in Mandarin Chinese. In such constructions, nominal elements (as verb arguments) and their properties, e.g. morphological marking, verb type and word order, play an important role. The processing of these constructions is observed empirically by recording and analyzing the electrical brain activity (EEG) of native Mandarin speakers during the presentation of auditory stimuli. The electrical potentials are averaged over the relevant words (word-onset).
The main focus of this investigation deals with the role of animacy and the linear order of arguments in Mandarin Chinese transitive constructions. The linear order is related to the thematic roles these arguments take. Specific semantic roles are determined by the verb; however, it is also possible to establish the thematic character and linear order of each argument´s role independently of the verb. In Mandarin Chinese, this can be done with the coverbs „bă“ and „bèi“. Both coverbs fix the linear order of the verb and its arguments as NNV (N = noun, V = verb) while the coverb itself stands between the arguments (N-bă-NV or N-bèi-NV). Additionally, the coverbs encode different thematic role hierarchies for the nominal arguments: the bă-construction calls for an agent-before-patient role assignment, while the bèi-construction (also known as the „Chinese passive voice“) calls for a reversed, patient-first assignment. The unmarked syntactic equivalent, NVN/SVO, is possible for both types of constructions, but this word order does not convey the same meaning as its bă- and bèi-counterparts. A further important difference between the bă- and bèi-constructions is that only transitive verbs of a certain character (known as „disposal“ or „causative“, Li & Thompson 1981) may occur with the coverb „bă“. This constraint does not hold for bèi-constructions, which are often associated with an „adversative reading“ (adversative passive, Huang 1999, adversative character = psychological affectedness of the undergoer by the event described by the verb).
Chapters 1-4 first discuss some theoretical aspects of this investigation - why, for example, the traditional grammatical terms „subject“ and „object“ or „agent“ and „patient“ are not always optimal to more closely describe these two types of constructions in Mandarin Chinese (chapter 1). It will be suggested, along the lines of Primus (1999), Van Valin (2005) and Van Valin & LaPolla (1997), to use the macrorole or protorole terminology „actor“ and „undergoer“ (Van Valin & LaPolla 1997) or „proto-agent“ and „proto-patient“ (Primus 1999), instead of „subject“ and „object“, respectively. Following this theoretical discussion, the issue will be approached from a psycholinguistic perspective. Chapter 4 deals with how the animacy of the nominal arguments might influence the incremental processing of transitive constructions. The argumentation is as follows: due to its implementation of prominence hierarchies and generalized semantic roles (GRs), the „extended Argument Dependency Model“ (eADM; Bornkessel & Schlesewsky 2006, Bornkessel-Schlesewsky & Schlesewsky 2009) is best suited to describe the effects that the coverbs „bă“ and „bei“ have on the processing of verb-final constructions in Mandarin Chinese (chapter 3).
Experiment I investigates the influence of the sentence-final verb in a bă-construction, as well as the animacy of the second argument (NP2). It involves the following manipulations/conditions: bă-compatibility of the verb, lexical-semantic plausibility of the sentence, animacy of NP2. NP1 is always animate, and the same lexical material was also tested with the neutral NVN-order. Experiment II systematically examines the influence of the linear orders of the thematic roles, as well as each argument´s animacy, using both bă- and bèi-constructions. Here the focus is only on the preverbal arguments. The results of Experiment II (see below) led to a third experiment, which further varies the order of thematic roles using the coverbs „bă“ and „bèi“ in relative clause constructions, which are head-final in Mandarin Chinese. That way it is possible to establish an undergoer-before-actor order for the bă-constructions, and an actor-before-undergor order for the bèi-constructions.
The main results from Experiment I are a) an N400 effect for bă-incompatible verbs in verb-final constructions (cf. Ye et al. 2007) and b) an N400 at the verb for the animate NP2 condition. The results from the second and third experiments are a) an N400 effect for an inanimate NP2 when NP1 is the undergoer (relational prominence effect), as well as b) a position-independent N400 effect for an inanimate undergoer when occuring with an animate actor (adversative passive effect).
The following conclusions were drawn from these results: bă-compatibility of the verb depends on a telicity condition, which is imposed on the event described by the sentence. The coverb „bă“ limits the interpretation of the events to one with a telic character so that only those verbs which allow such an interpretation are permitted. The prominence-based effect of animacy results from the raised cognitive (semantic) interference of two activated concepts during the establishment of the prominence hierarchy (NP2) or during the verb-argument-linking (verb position). The adversative effect can be explained by the „bei“-specific constraint on the interpretation of the undergoer argument (NP1). More precisely, „bei“ reduces the undergoer to an experiencer role only, and therefore renders it an animate entity. As a whole, the data show that, in a single language, cross-linguistically observable processes like argument hierarchization via prominence features and language-specific characteristics (interpretative constraints and pragmatic functions) both equally have a direct influence on incremental language processing.