心理科学 ›› 2023, Vol. 46 ›› Issue (4): 770-778.DOI: 10.16719/j.cnki.1671-6981.20230401

• 基础、实验与工效 •    下一篇

预视对阅读中语境预测性效应的影响*

赵赛男1,2, 李琳1, 张俐娟1, 王敬欣**1   

  1. 1天津师范大学心理学部,天津,300387;
    2临沂大学教育学院,临沂,276005
  • 出版日期:2023-07-20 发布日期:2023-08-14
  • 通讯作者: **王敬欣,E-mail:wjxpsy@126.com
  • 基金资助:
    *本研究得到国家自然科学基金项目(32271119)和国家自然科学基金项目(32000790)的资助

The Influence of Preview on Contextual Predictability Effects during Reading

Zhao Sainan1,2, Li Lin1, Zhang Lijuan1, Wang Jingxin1   

  1. 1Faculty of Psychology, Tianjin Normal University, Tianjin, 30038;
    2School of Education, Linyi University, Linyi, 276005
  • Online:2023-07-20 Published:2023-08-14

摘要: 研究采用增量呈现文本范式操纵预视类型,并同时操纵目标词语境预测性,考察了预视对词汇加工中早期语境预测性效应的影响。实验1发现预视类型和语境预测性不存在交互作用。无论有没有预视信息,语境预测性均对早期词汇加工产生了影响。实验2采用※预视创设了更接近自然阅读的文本呈现模式,再次发现语境预测性效应不受预视类型的影响。研究结果表明,语境预测性对阅读中词汇加工的影响并不依赖于与目标词相同的预视信息,在没有预视信息情况下同样会表现出语境预测性对词汇加工的影响。

关键词: 语境预测性, 预视, 增量呈现文本范式, 眼动

Abstract: Most previous research have found contextual predictability effects eliminated in invalid preview which indicates contextual predictability effects depend on valid preview (like normal reading). The researchers manipulated invalid preview conditions by presenting various kinds of words or nonwords that were different from target words in parafoveal vision. It has been well established that invalid preview can cause various costs that may overwrite the contextual predictability effects. However, it is still hard to tell which is the cause of the elimination of contextual predictability effects in invalid preview conditions: the cost caused by invalid preview, or the absence of valid preview. Solving this problem is crucial to understand how the top-bottom predictability is influenced by bottom-top preview information. The present research investigated this effect with incremental paradigm by manipulating parafovea without preview information.
EyeLink 1000 Plus eye-tracker recorded participants' (40 participants in experiment 1 and 44 participants in experiment 2) right-eye gaze when they read the sentences that contained target words. Sentences were displayed in Song font in black-on-gray text on a 24-inch ASUS LCD monitor (1920×1080 pixels) with each character subtended approximately 0.9 degrees of visual angle. Experiment 1 was a 2(contextual predictability: high, low)×2(preview type: normal, none)within subjects design. Stimuli were 164 sets of Chinese sentences containing two interchangeable target words that were of either high or low contextual predictability. There was no preview information before directly fixed word in none preview condition, which was different from normal reading pattern and may influence the results. The aim of experiment 2 was to further verify and extend the findings from experiment 1 in a more normal reading form. In order to create a normal reading pattern with minimal interference for vocabulary processing, experiment 2 used meaningless and simple ※ as invalid preview. It was a 2 (contextual predictability: high, low) × 2(preview type: normal, ※) within subjects design.
The results showed clear effects of preview type in both experiments with shorter reading times and word skipping rate for normal preview condition, in line with findings from previous studies. It also replicated robust and reliable contextual predictability effects on eye movement time measures (first fixation duration, gaze duration, total reading time) in both experiment 1 and experiment 2, which were contributed to longer fixation durations for high predictability words than low predictability words. More importantly, the current results showed no interaction between contextual predictability and preview types on any measures in both experiment 1 and experiment 2. It suggested that the contextual predictability effects with none preview and ※ preview were similar to normal preview. The results of Bayes analyses also provided strong evidence for the additive models.
The key point of present study is the interaction between contextual predictability and preview type. The robust addictive effects suggest the elimination of contextual predictability effects in invalid previews is not due to the lack of valid preview but the overwrite of the invalid preview costs. Therefore, this research indicates contextual predictability influences word processing independently rather than depending on the valid preview information.

Key words: contextual predictability, preview, incremental revealing text paradigm, eye movement