Abstract: Preview effects have been found in Chinese and alphabetic languages. However, previous studies showed that the preview effect in Chinese was greater and larger than alphabetic languages (Vasile & Angele, 2017). In Chinese, there is no word boundary information between words. A Chinese character can be formed by a word independently or combined with other Chinese characters into a word. While processing the current fixed character(s), Chinese readers may preview more character(s) to determine the position of word boundary in advance, thus increasing the preview effects (Yen et al., 2009). However, Sheridan et al. (2016) compared preview processing in a normal text condition contained spaces with an unsegmented text condition contained random numbers instead of spaces and larger preview effects were found for normal condition than unsegmented text condition. Present study further explored the influence of word demarcation information on preview processing. In the present study, we recorded readers’ eye movements to explore the influence of word demarcation information on preview processing for one-character words and two-character compound words. The one-character words were also the first character of two-character compound words. We adopted the boundary paradigm (Rayner, 1975) to manipulate the target words presented normally or replaced by a pseudocharacter for previews. Instead of word space, we adopted word highlighting to demarcate word boundary. As Cui et al. (2014) stated, inter-word spaces provide word boundary information while also reducing word boundary masking. Moreover, according to Bai et al. (2008), word spacing condition will increase the length of the sentence, which will affect the sensitivity of vision and then influence the fixation position and fixation duration. This research used the Eyelink 2000 eye tracker which has 1000Hz sampling rate. The experimental material was presented on a 19-inch flat screen display with a refresh rate of 150 Hz and a resolution of 1024 pixels×768 pixels. There were six practice sentences to help the participants familiarize with the experimental procedure. In addition, in the formal experiment, every three sentences were followed by reading questions to ensure that the participants could read the sentences carefully by asking the participants to judge "yes" or "no". Firstly, consistent with the findings of Bai et al. (2008), the results indicated that word highlighting could facilitate the processing of one-character words and two-character compound words, participant read word highlighting condition faster than normal condition. Secondly the results indicated that the preview effects of two-character compound words were greater than one-character words. Thirdly the results showed that word highlighting did not affect the preview effects for neither one-character words nor two-character compound words. These results are consistent with the model of word segmentation and word recognition (Li et al., 2009), which proposed that the character in the visual field be processed in parallel. When the target words are recognized, the word boundary information is also recognized; consequently, word highlighting did not affect the preview effects for neither one-character words nor two-character compound words. Totally, word boundary information is not the most fundamental reason for the preview processing differences between Chinese and alphabetic languages. We still need to further explore the reasons from more deeply cognitive processing to explain preview processing differences between Chinese and alphabetic languages in following studies.

Key words: word highlighting, one-character word, two-character compound word, preview processing, eye movement control

摘要： 本实验采用眼动轨迹记录法并结合边界范式（Rayner, 1975），操纵单字词和双字复合词的首字为一致预视或假字预视，考察单字词和复合词预视加工的差异性。另外，通过在中文句子中加入词边界信息——词间阴影，以考察词间阴影对单字词和复合词预视加工的影响。实验结果表明，复合词的预视效应大于单字词，并且词间阴影对单字词和复合词的预视加工均没有显著影响。因此，与印欧语系相比，中文阅读中更大更强的预视效应需从更深层次的语言和认知因素进行探讨。该结果更倾向于支持词切分和词汇识别模型。

关键词: 词间阴影, 单字词, 双字复合词, 预视加工, 眼动控制