| The present study investigates the role of lexical semantics in sentence processing within the framework of the Competition Model (Bates & MacWhinney, 1982, 1989). A close look at the Competition Model literature reveals that lexical semantics is not well represented in the model. First, among the semantic cues, only animacy has been explored. However, studies investigating the animacy effect1 generally employ verbs that require animate nouns as their subject (simple verbs2). A question raised is whether the animacy effect is a unitary effect caused by animacy itself or it is some effect caused by the use of simple verbs. Second, although nominal semantics is identified as a category of cues, only animacy is represented as a cue; nothing is said about the role of other nominal semantic features. Third, verb semantics has no representing cues in the form-function mapping process. In face of increasing evidence for the role of lexical semantics in sentence processing, it would be unreasonable for a sentence processing theory like the Competition Model to ignore the role of lexical semantics.As an attempt to explore the role of lexical semantics within the framework of the Competition Model, the present study proposes a semantic cue to represent lexical semantics of nouns and verbs: argument specification satisfaction (ASS). ASS argues that form-function mapping can be interpreted as many-to-many noun-argument mapping. In sentence processing, nominal lexical forms compete to be mapped onto arguments in the functional level. Arguments carry semantic requirements that have to be satisfied. A noun meeting more of these requirements of an argument is higher in ASS for that argument and has a better chance of being mapped onto it.To test the proposed cue of ASS, the present study conducted 3 experiments. The experiments adopted a task of simple agent identification, where participants wereasked to identify the subject/agent of a simple transitive sentence. Thirty Chinese monolinguals and thirty English monolinguals participated in the experiments. Experiment 1 was intended to test whether the animacy effect is a product of the use of simple verbs. Three types of verbs were employed: simple verbs, Stimulus-Experiencer (SE) verbs3 and nonword verbs. Experiment 2 was to explore the role of ASS in sentence processing. Three levels of ASS were manipulated: ASS1 where the first noun was favored by the verb as the subject/agent, ASS2 where the second noun was preferred, and ASS3 where no preference was given to either noun. Experiment 3 aimed to explore the role of nominal semantic features other than animacy in sentence processing and to test the hypothesis that noun-argument mapping is many-to-many. Nominal animacy was controlled. Four ASS conditions were manipulated. In the first ASS condition, the two nouns were equal in both the internal ASS and the external ASS4. In the second condition, one noun was higher than the other in the internal ASS, though they were equal in the external ASS. In the third condition, one noun was higher in the external ASS though they were equal in the internal ASS. In the fourth condition, one noun was higher in the external ASS while the other was higher in the internal ASS.The results were consistent with the hypotheses laid down by the ASS hypothesis. First, we found that the animacy effect varied across verb types and could be accounted for by the ASS effect and the animacy per se effect. Second, ASS was not only a reliable but also strong strategy for sentence processing. Third, nominal semantic features other than animacy affected sentence processing through ASS. Fourth, when a noun was mapped onto an argument, the other noun was chosen to be mapped onto the argument left, strongly suggesting that noun-argument mapping works in a many-to-many fashion.It is believed that the conclusions have important implications for the...
|
PDF Full Text Request