Abstract: The neural dynamics of prosodic processing in spoken language comprehension has drew a lot of attention. However, little is known to what extent prosodic information constrains neuro-cognitive processes of written language processing. Exploration into this issue provides a better understanding of the role of prosodic information. The present research is designed to investigate whether and how a comma and the lexical deflection influence the reading of Chinese sentences, by using eye movements recording. Eye-tracking is one of the ideal methodologies for exploring this problem because it allows for recording the fixation of various observers on various parts of the sentence in real time with relatively high temporal and spatial resolution. Three types of eye movement measures such as first fixation duration, gaze duration, and total fixation duration were examined to determine how lexical deviation and comma placement was activated and the interactions of lexical deviation and comma placement during silent reading of Chinese text. A self-paced, line-by-line, reading paradigm was used, in which participants were asked to read sentences silently and press a key each time when they want a new one to appear. Some of the sentences followed by a question which required participant to judge whether the two sentences having the same meaning. The Chinese sentences had or not had a comma insert at the position of the ambiguous structure. The experiments were conducted by manipulating the meaning biased by commas and lexical deviation. Analysis indicators were first fixation duration, gaze duration and total fixation duration. According to the analysis of fixation time on the lookzones，it was found that: when there was different lexical deflection, the ambiguous sentences disambiguated by the comma had a different paten on first fixation duration, gaze duration and total fixation duration. The main effects of the comma and the lexical deflection were significant (p﹤0.05). The interaction between the comma and the lexical deflection duration was also significant(p﹤0.05). Further simple effect inspection found that the times were shorter for the target region with a comma than that of without a comma no matter what the lexical deflection was(p﹤0.05). The lexical deflection had an affect on sentences processing only when the comma had biased the adjective modify the entire phrase(p﹤0.05). The results showed that the lexical and comma deflection both had an impact on sentences processing at the early stage. The insertion of commas shortened the fixation times in region when the lexical deflection in accordance with the direction the comma disambiguated to. The interaction between the comma and the lexical deflection took place at the early stage of processing. The comma had an affect on sentences processing no matter what the lexical deflection was. The lexical deflection had an affect on sentences processing only when the comma had biased the adjective modify the entire phrase. In conclusion, our data were not only compatible with the implicit prosody hypothesis, but also suggested the interaction between the suprasegmental cues and the lexical deflection occurred at the early processing during sentence reading.

Key words: the comma, lexical deviation, ambiguity resolution, eye movement

摘要： 本研究采用眼动法，探讨了被试在理解层次分隔歧义句时，逗号作为书面韵律边界所起到的作用，以及书面韵律边界与词汇信息如何共同影响句子加工。研究选取了裸视视力或矫正视力均正常的母语为汉语的成人，完成一个在自然情境下阅读层次分隔歧义句的句子理解实验，同时记录他们句子阅读时的眼动数据。结果发现：有逗号解歧的句子的阅读时间明显少于没有逗号解歧句子的阅读时间；逗号偏向类型和词汇偏向类型之间存在交互作用，即两者在三种眼动指标上都达到了显著性水平，研究所选的三种眼动指标为首次注视时间、第一遍阅读时间和总注视时间；此外，当逗号偏向将句子结构确定为形容词修饰整个短语结构时，词汇偏向会对语义理解进行调整，表现在三种眼动指标的显著性上。实验结果支持内隐韵律假设，书面韵律信息对歧义句歧义消解的作用从句子加工早期开始，书面韵律信息与词汇偏向的交互作用也发生于句子加工早期。

关键词: 书面韵律边界, 词汇偏向, 歧义消解, 眼动