基于ERPs的结构性歧义句的心理加工机制研究

发布时间：2018-09-04 17:06

【摘要】：作为一种特殊的句子类型，歧义句在语言研究领域中一直备受关注。在歧义句的各类研究中，对歧义句加工机制的研究是一个重点。所谓句子加工机制是指人们在理解一个句子时为最终达致句子意义所进行的诸如句法分析、意义提取、意义整合、意义选择等心理操作过程。目前提出的有关歧义句加工机制的模型主要有四个：McClelland (1987)提出的“基于限制的模型”(Constraint-based Model)，江天方(1999)提出的“逐一提取模型”(One-by-oneRetrieval Model)，Crain和Steedman (1985)提出的“参照支持模型”(Referential Support Model)，以及Boland和Blodgett (2001)提出的“同时提取模型”(Concurrent Retrieval Model)。前两个模型支持歧义句“一步加工完成”的观点，认为人们在理解无语境的歧义句时，倾向于激活其常用意义，原因在于体现常用意义的句子成分之间的搭配关系牢固；当歧义句被赋予语境时，人们倾向于激活符合语境的意义，忽略不符合语境的意义，原因在于前者处于绝对优势。后两个模型支持歧义句加工“两步完成”的观点，认为人们在理解歧义句时，无论其是否带有语境，均首先激活其所有潜在的意义，然后根据语用或语境信息选择出恰当的意义。前人对这些歧义句的加工机制模型也开展了大量实证研究，但由于不同研究在研究设计、被试选择、实验材料编写、研究手段等方面多不一致，可比性有限，导致研究结论彼此多有抵牾。本研究针对现有研究存在的不足，采用ERPs新技术，对结构性歧义句的心理加工机制，进行再探讨，旨在回答如下问题：结构性歧义句的加工机制是什么，即，被试在在线加工结构性歧义句时是激活一个意义还是两个意义？本研究在四川外语学院外语学习认知神经实验室完成，共选择被试30人，均为右利手，年龄在22-28岁之间。实验材料为240个结构性歧义句（实验句）和20个填充句。实验句分为四类：含有“常用义”和“生僻义”两种意义的歧义句（简称“歧义句”），以“下划线”排除“生僻义”的歧义句（简称“消歧句1”），以“下划线”排除“常用义”的歧义句（简称“消歧句2”）以及在歧义句句首附加了具有“常用义指向性”的语境句的歧义句（简称“指向性语境歧义句”）。实验时，每个句子均分成无歧义和有歧义部分先后呈现，要求被试在每个句子呈现完毕时判断该句含有一种还是两种意义。本实验利用Neuroscan软件采集脑电，并对ERP数据进行离线分析，结果输入SPSS软件进行统计分析。通过对比各类句子加工的ERPs脑电数据，最终得出了以下结果： (1)歧义句和其消歧句1的脑电波幅相似。由此推测被试在线加工不带有语境的结构性歧义句时，只激活了常用的意义，没有激活生僻意义。在线研究结果支持“基于限制的模型”和“逐一提取模型”的观点。 (2)指向性语境歧义句的脑电波幅和消歧句1、2的叠加脑电波幅相似。由此推测被试在线加工指向性语境歧义句时，常用意义和生僻意义均被激活。在线研究结果基本支持“参照支持模型”和“同时提取模型”的观点。综上所述，被试在在线加工不带有语境的结构性歧义句时只激活常用意义，而在在线加工指向性语境歧义句时既激活常用意义，也激活生僻意义。这表明，，结构性歧义句的加工并没有表现出统一的加工机制，需进行分类讨论。
[Abstract]:As a special sentence type, ambiguous sentences have attracted much attention in the field of linguistic research. Among all kinds of ambiguous sentences, the study on the processing mechanism of ambiguous sentences is a key point. There are four main models of ambiguity processing: Constraint-based model proposed by McClelland (1987), One-by-one Retrieval model proposed by Jiang Tianfang (1999), Crain and Stedman (1985). "Referential Support Model" and "Concurrent Retrieval Model" proposed by Boland and Blodgett (2001). The first two models support the view that ambiguous sentences are "processed in one step" and that people tend to activate their common meanings when they understand contextual ambiguity sentences. When ambiguous sentences are given context, people tend to activate contextual meaning and ignore non-contextual meaning because the former is absolutely superior. The latter two models support the view that ambiguous sentences are processed in two steps and people understand ambiguity. Previous studies on the processing mechanism of these ambiguous sentences have also carried out a large number of empirical studies, but due to different research designs, subjects'choice, experimental materials compilation, research methods and so on. Many inconsistencies and limited comparability lead to conflicted conclusions.
Aiming at the shortcomings of the existing research, this study uses ERPs technology to re-explore the psychological processing mechanism of structured ambiguity sentences, aiming to answer the following questions: What is the processing mechanism of structured ambiguity sentences, that is, whether the subjects activate one meaning or two meanings when processing structured ambiguity sentences online?
This study was conducted in the Cognitive Neurology Laboratory of Foreign Language Learning, Sichuan University of Foreign Languages. Thirty subjects, all right-handed, aged 22-28, were selected. The experimental materials consisted of 240 structural ambiguity sentences (experimental sentences) and 20 filler sentences. The experimental sentences were divided into four categories: ambiguous sentences with two meanings of "common meaning" and "unfamiliar meaning" (abbreviated as "ambiguity"). In the experiment, the ambiguous sentences with "common sense directivity" were added at the beginning of the ambiguous sentences. Each sentence was divided into unambiguous and ambiguous parts, and the subjects were asked to judge whether the sentence contained one or two meanings at the end of each sentence.
In this experiment, EEG was collected by Neuroscan software, and ERP data were analyzed offline. The results were input into SPSS software for statistical analysis.
(1) Ambiguity sentence is similar to its disambiguation sentence 1 in EEG amplitude. It is inferred that the subjects only activate the common meanings and do not activate the remote meanings when they process structured ambiguity sentence without context on-line.
(2) The amplitudes of EEG of directional context ambiguity sentences are similar to those of disambiguation sentences 1 and 2. It is inferred that the common meaning and remote meaning are activated when the subjects process directional context ambiguity sentences online.
To sum up, the subjects only activate common meanings when they process structured ambiguity sentences without context on-line, and activate both common and remote meanings when they process directional ambiguity sentences on-line.
【学位授予单位】：四川外语学院
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：H04

【相似文献】