PSYCH OpenIR  > 认知与发展心理学研究室
中文组合歧义词的加工和切分机制
其他题名The mechanism of word segmentation and recognition for Chinese combinational ambiguous word
周钧毅
学位类型博士
导师李兴珊
2017-05
学位授予单位中国科学院研究生院
学位授予地点北京
学位专业基础心理学
关键词中文阅读 中文组合歧义词 表征激活 时间进程 词切分
摘要

相较于大多数的拼音文字系统的文本而言,中文文本的一个最为显著的特性是没有明确的词间空格来标记词边界。鉴于词汇在阅读过程中起到重要的作用,对于文本的理解需要建立在词汇加工识别的基础上,而需要对词汇进行加工又需要将词汇从文本中切分出来,因此中文读者在阅读中的词切分和词汇识别过程是语言心理学中的一个重要问题。在以往研究的基础上,本研究对这个问题展开了系统的探讨,通过四个实验,对中文阅读中组合歧义词的词汇识别和词切分进行考察。
本研究包含两部分研究。研究一包含两个实验,使用正常阅读任务探讨了中文组合歧义词加工过程中的表征激活。组合歧义词指的是由多个字组成,其中一部分字能够组成另外一个能够独立使用的词汇,例如“酒精灯”,其中头两个汉字能够组成“酒精”(在本研究中我们称其为嵌套词)。实验1使用正常阅读任务考察了包含前嵌套词的组合歧义词的整词和嵌套词表征的激活。我们分别操纵了整词词频和嵌套词词频。结果发现,在加工组合歧义词的过程中整词和嵌套词的表征都会得到激活,对首次注视时间的生存分析结果表明,嵌套词表征的激活要早于整词的表征。在实验2使用正常阅读任务对包含后嵌套词的组合歧义词的嵌套词表征的激活进行考察。我们操纵了后嵌套词词频。结果发现,后嵌套词表征得到了激活。研究一的实验结果说明,在组合歧义词的加工过程中,整词表征和嵌套词表征都会得到激活。
我们通过研究一揭示了组合歧义词加工过程中的表征激活情况,然而,得到激活的表征是否会影响读者的词切分?我们在研究二中同样使用正常阅读任务探讨读者对组合歧义词的切分和读者的词切分策略。研究二包含两个实验,实验3中使用正常阅读任务考察了读者对于包含前嵌套词的组合歧义词的切分。通过变化目标词之前的动词,在整词合理的条件下,操纵了组合歧义词中嵌套词的合理性(合理性指的是词汇或者文本相对于背景文本信息的合理程度)。结果没有观察到嵌套词的合理性效应,这说明嵌套词的激活并不会使得嵌套词被切分出来,读者倾向于将组合歧义词整词作为一个词汇单位进行切分。在实验4中,我们进一步探讨嵌套词的合理性和整词的合理性之间的关系,以此考察中文读者的词切分策略。我们使用正常阅读任务。通过变化组合歧义词(目标词)前面的动词来对组合歧义词的合理性和其中的嵌套词的合理性同时进行操纵。结果发现,在整词合理的条件下,没有观察到嵌套词的合理性效应,这重复了实验3的结果。然而,在整词不合理的条件下,我们观察到了嵌套词的合理性效应。这个结果支持了我们对于读者词切分策略的假设,即,在整词不合理的情况下,读者发现这种切分方式无法构建一个合理的文本结构,因此读者会迅速调整词切分策略,转向更小的词汇单位进行切分,试图通过切分方式的改变来重新构建一个合理的文本结构。研究二在研究一的基础上,进一步说明了读者对于组合歧义词的切分方式和在阅读过程中的词切分策略。
在最后的讨论中,我们探讨了当前研究中所揭示的组合歧义词的加工机制、词切分策略以及当前结果对于中文词切分与词识别建模的启示。同时对于未来的研究进行了展望。

其他摘要

One of the most obvious characteristics of Chinese text, compared with most alphabetic languages, is that there is no interword spaces to mark word boundaries. As words play very important roles in Chinese reading, the understanding of text should be based on word recognition. Word recognition is found on the premise that text is segmented properly. Therefore, the word recognition and word segmentation of Chinese readers is a key issue of linguistic psychology. On the basic of previous studies, present dissertation examine the word recognition and segmentation of Chinese combinational ambiguous word.
Present dissertation includes two main studies. In the first study, we explored the activations of word representations during the processing of Chinese combinational ambiguous word. These are words where some of the component characters constitute another word (i.e. embedded word). In Experiment 1, participants read sentences with combinational ambiguous wordswhose firs two characters can constitute an embedded words while their eye movements were monitored. Reading times on these words were longer when the frequencies of the embedded words were lower, and were longer when the frequencies of the whole words were lower. Survival analyses showed that the divergence pointfor the embedded-word frequency set was earlier than that for the whole-word frequency set. In Experiment 2, participants read sentences with combinational ambiguous words whose last two characters can constitute an embedded words while their eye movements were monitored. The frequency effect of embedded word were observed as well. These results suggest that embedded words are activated when Chinese readers processcombinational ambiguous word, and this activation is earlier than for the whole word.
The first study only revealed that the representation of embedded word was activated during reading combinational ambiguous word. However, whether this activation influences the segmentation of combinational ambiguous word remains an open question. In the second study, we examined the mechanism of word segmentation during reading Chinese combinational ambiguous word and explored Chinese readers strategy of word segmentation. In Experiment 3, participants read sentences with combinational ambiguous words whose firs two characters can constitute an embedded words while their eye movements were monitored. We varied verbs prior to target words such that the plausibility of target word were manipulated. Four conditions of sentences were constructed: (a) a plausible combinational ambiguous target word with its embedded word plausible at the point it appeared, (b) a plausible combinational ambiguous target word with its embedded word implausible at the point it appeared, (c) a plausible two-character target word, (d) an implausible two-character target word. No plausibility effect of embedded word were observed which indicated that the activation of embedded word would not lead to its segmentation. Chinese readers tend to segment the combinational ambiguous word as a whole. In Experiment 4, we explore the relationship between the plausibility of embedded word and the plausibility of the whole word. Also, We varied verbs prior to target words such that the plausibility of target word were manipulated. Four conditions of sentences were constructed: (a) a plausible combinational ambiguous target word with its embedded word plausible at the point it appeared, (b) a plausible combinational ambiguous target word with its embedded word implausible at the point it appeared, (c) a implausibility combinational ambiguous word with its embedded word plausible at the point it appeared, (d) an implausible combinational ambiguous word with its embedded word implausibility at the point it appeared. No plausibility effect of embedded word were observed when the combinational ambiguous target words were plausible. However, we do observed the plausibility effect of embedded word when the combinational ambiguous target words were implausible. The results showed that when the combinational ambiguous target words are implausible, readers found that segmenting combinational ambiguous word as a whole cannot lead to a coherent mental representation. Readers then adjusted their strategy of segmentation to segment embedded word so that they can rebuild a coherent mental representationwhich showed that Chinese reader’s strategy of word segmentation is flexible and immediate.
Finally, we discussed some inspiration of present dissertation on Chinese word segmentation and recognition as well as building models of word segmentation and recognition in Chinese reading. We also raise some questions for proceeding potential further studies.

语种中文
文献类型学位论文
条目标识符http://ir.psych.ac.cn/handle/311026/21486
专题认知与发展心理学研究室
作者单位中国科学院心理研究所
推荐引用方式
GB/T 7714
周钧毅. 中文组合歧义词的加工和切分机制[D]. 北京. 中国科学院研究生院,2017.
条目包含的文件
文件名称/大小 文献类型 版本类型 开放类型 使用许可
周钧毅-博士学位论文.pdf(3158KB)学位论文 限制开放CC BY-NC-SA浏览 请求全文
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[周钧毅]的文章
百度学术
百度学术中相似的文章
[周钧毅]的文章
必应学术
必应学术中相似的文章
[周钧毅]的文章
相关权益政策
暂无数据
收藏/分享
文件名: 周钧毅-博士学位论文.pdf
格式: Adobe PDF
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。