自然阅读中的词跳读：视觉和语言信息的交互作用

It is well known that our eyes move to different locations in text reading to obtain the useful information for comprehension. Eye movements in reading is not a sequence of word-to-word movements, many words are not fixated but skipped. While in reading English texts, about one third of words are skipped. Our recent findings suggest that there are approximately 47% Chinese single character words skipped during Chinese reading. The reasons why words are sometimes skipped, whether these skipped words are processed, and if so, when they are processed, are not well understood. The underlying mechanisms of word skipping have been an important research topic in eye movement control models. The previous researchers using eye tracking technique have investigated readers’ word skipping during reading of alphabetic languages, they have been puzzled by this phenomenon and suggested various theoretical explanations. However, unlike English, written Chinese is unspaced, densely packed and word boundaries are less clear in the parafovea than in spaced languages like English. The cause of word skipping during Chinese reading is far from understood. In this project, we will employ a co-registration of eye tracking and ERPs to record readers’ eye movements and EEG simultaneously during natural Chinese sentence reading, to investigate the cognitive and neural mechanisms of word skipping during Chinese reading. The specific research questions are: (1) Is Chinese word skipping based on complete recognition or partial “coarse” recognition of a word in the parafovea? (2) Does foveal processing load influence the skipping of the upcoming word? (3) Will both parafoveal and foveal processing efficiency influence word skipping in Chinese reading? And (4) the visual and linguistic information: which is more important in explaining word skipping in Chinese reading. Our findings will provide important theoretical implications for our understanding of the cognitive and neural mechanisms of Chinese word skipping and constructing of eye movement control models during Chinese reading. Furthermore, they will provide evidence for improvement of individual’s reading performance.

在阅读过程中，眼球的运动不是逐词进行的，而是不断地跳过许多词。在英文阅读中至少有30%的词被跳读。本课题组近期的研究表明，中文阅读中约47%的字词被跳读。那么被跳读的字词是否得到加工，何时得到加工，具体来说（1）词跳读是基于副中央凹处词汇的完全识别还是部分识别后发生；（2）中央凹加工负荷是否影响对下一个词的跳读；（3）副中央凹和中央凹加工效率与词跳读的关系是什么；（4）视觉和语言信息如何交互作用决定一个词的跳读。这是当前阅读中眼动控制模型关注的核心问题。中文阅读中词跳读的研究非常有限。本项目基于中文书写系统的特点，运用眼动和ERP同步技术，实时地记录读者在中文自然阅读时词跳读过程中的眼动和脑电活动特点，将对以上四个关键科学问题进行回答，从而系统地揭示中文词跳读产生的机制及影响因素。本项目为建构中文阅读的眼动控制模型提供启示，还对提高不同群体读者的阅读效率提供参考依据。

在阅读过程中，眼球的运动不是逐词进行的，而是不断地跳过许多词。在英文阅读中至少有30%的词被跳读，中文阅读中至少有40%的字词被跳读。那么被跳读的字词是否得到加工，何时得到加工，是当前阅读中眼动控制模型关注的核心问题。中文阅读中词跳读的研究非常有限。与拼音文字书写系统不同，中文书面排列更紧密，单位空间内的信息密度更大，读者从副中央凹获得更多信息，可能导致更多跳读；然而，中文没有词间空格，可能使读者在利用副中央凹视觉区进行词汇切分时存在困难，进而影响词的跳读。此外，中文词长变异小，可能使读者采用不同于拼音文字阅读的跳读策略。基于中文书写系统这些特点，项目运用眼动研究中经典的边界范式，实时地记录读者在中文自然阅读时词跳读过程中的眼动特点，系统地揭示中文词跳读产生的机制及影响因素。主要研究发现：（1）当副中央凹视觉区呈现一个高频简单的预视词时，即使该词与先前的语境不符，读者也倾向于跳过随后呈现的目标词，表明读者更多地依赖于副中央凹处词汇的熟悉性做出跳读的决定。该效应不受读者对呈现发生变化意识的影响；（2）中央凹加工负荷影响目标词的跳读，但该效应独立于副中央凹预视的效应；（3）在拼音文字阅读中，老年人通常采用冒险的阅读策略，为补偿较慢的阅读速度，总是不断地进行跳读，但这种猜测可能是无效的，所以不断进行回视。在中文阅读过程中，发现老年人的跳读率比较低，向前的眼跳距离比较小，更喜欢采取谨慎的阅读策略，这一结果可能与中文书写系统的特性有关；（4）采用眼动与脑电同步记录技术，初步验证了副中央凹预视效应以及视觉和正字法的副中央凹-中央凹效应，即副中央凹呈现的视觉和正字法信息影响当前中央凹词汇的加工，为进一步探讨自然阅读过程中副中央凹和中央凹加工的神经机制奠定基础。项目结果为建构中文阅读的眼动控制模型提供启示，对提高不同群体读者的阅读效率提供参考依据。在项目资助下，我们在《Journal of Experimental Psychology: General》等国际知名的实验心理学期刊上共发表论文十九篇，获得全国教育科学研究和天津市社科研究优秀成果奖等。