An ERPs Approach To The Mental Processing Mechanism Of The Structurally Ambiguous Sentence

As a special sentence type, ambiguous sentence has been widely studied inlinguistics. Among different kinds of studies on ambiguous sentence, the mechanismof ambiguous sentence processing is of great importance. Mental mechanism refers tothe mentally operational processes such as syntactic analysis, syntactic analysis,semantic retrieval, integration and selection, etc., which are conducted to construct themeaning of a sentence during sentence comprehension. There are mainly four modelsdealing with ambiguous sentence processing: the Constraint-based Model provided byMcClelland (1987), the One-by-one Retrieval Model provided by Jiang Tianfang(1999), the Referential Support Model provided by Crain and Steedman (1985), andthe Concurrent Retrieval Model provided by Boland and Blodgett (2001). The formertwo models support the idea of “one-phase manner” of ambiguous sentenceprocessing. They claim that when people comprehend an structurally ambiguoussentence without context, probably the most common meaning will be activated,because it contains some high-frequency-collocated words or phrases; when providedwith context, people will only activate the context-appropriate meaning with thecontext inappropriate meaning eliminated, because the context-appropriate meaningwill perfectly overwhelms the other meanings. However, the latter two modelssupport the idea of “two-phase manner” of ambiguous sentence processing. Theyclaim that when people comprehend a structurally ambiguous sentence, they will firstactivate all the possible meanings and then choose the appropriate meaning based onthe pragmatic or contextual information, no matter whether it is provided with contextor not. So they think all the meanings of an ambiguous sentence will be activated andthese meanings will compete with each other. Though a number of previous studies onambiguous sentence processing had been conducted, there are some contradictions intheir investigations owing to different experimental design, subject selection, materialcollection, and research methods in their studies.Based on the achievements and limitations of the previous studies, the present study investigates the mechanism of structurally ambiguous sentence processing again,using ERP technique to conduct the experiment. The research question of the presentstudy is as follows: What is the mental processing mechanism of the structurallyambiguous sentence under on-line processing? Specifically, under on-line processing,do the participants only activate the common meaning of an ambiguous sentence, asstated by the Constraint-based Model and the One-by-one Retrieval Model; or do theyactivate both the common and the uncommon meanings of an ambiguous sentence, asclaimed by the Referential Support Model and the Concurrent Retrieval Model?The present study was completed in the Key Lab of Cognitive Neuroscience&Foreign Language Learning in Sichuan International Studies University. Totally,30students were selected in the present study; all of them are right-handed and agedfrom22to28. There were240structurally ambiguous sentences (experimentalsentence) and20filter sentences in the experiment. The experimental sentences weredivided into the following four types. The first type was the ambiguous sentence,which contains two meanings, the common meaning and the uncommon meaning(“ambiguous sentence” for short). The second one was the ambiguous sentence withthe common meaning, in which an underline was used to eliminate the uncommonmeaning (“disambiguated sentence1” for short). The third one was the ambiguoussentence with the uncommon meaning, in which an underline was used to eliminatethe common meaning (“disambiguated sentence2” for short). The last one was theambiguous sentence with the common meaning, in which an anterior context directingto the common meaning was provided to eliminate the uncommon meaning(“context-embedded ambiguous sentence” for short). During the experiment, everysentence in the material were divided into two parts (the unambiguous part and theambiguous part) to be successively presented, and the participants were required tojudge whether the sentence has one meaning or two meanings after it was presented.The ERPs waves were analyzed by Neuroscan System. After the data beinginputted into the software SPSS16.0and analyzed, the following results were gained:(1) The ERPs waves between the ambiguous sentence and the disambiguatedsentence1are similar to each other, from which we can infer that when the participants comprehend a context-absent structurally ambiguous sentence underon-line processing, they only activate the common meaning. It is in accordance withthe one-phase manner of ambiguous sentence processing and supports theConstraint-based Model and the One-by-one Retrieve Model.(2) The ERPs waves between the context-embedded ambiguous sentence andthe disambiguated sentence “1+2”(the overlying wave of disambiguated sentence1and2, referring to the activation of both the common and the uncommon meaning)are similar to each other, from which we can infer that when the participantscomprehend a context-embedded structurally ambiguous sentence under on-lineprocessing, they activate both the common meaning and the uncommon meaning.This result is identical with the perspective of the Referential Support Model and theConcurrent Retrieval Model.To sum up, only the common meaning is activated in the context-absentstructurally ambiguous sentence under on-line processing, but both the common andthe uncommon meanings are activated in context-embedded structurally ambiguoussentence under on-line processing. Therefore, the structurally ambiguous sentenceprocessing does not show a uniform mental mechanism, and has to be discussedseparately.