Abstract:

In the present study, readers’ eye movements were recorded to investigate the influence of word highlighting information on the preview processing of between-words and within-words.

Most studies on preview effects have found that the size of the preview effects is 30~50 ms. Hy?n? et al. (2004) examined parafoveal processing of the end lexeme of a long Finnish compound while the beginning lexeme of the compound was fixated. And the results found 80ms preview effect which was more pronounced than previous literatures. The larger preview benefit may have been due to the fact that the preview word is part of one larger linguistic unit (within-words); however, in previous experiments, the preview word and the current fixated word belong to different words (between-words). Consequently, researchers speculated that within-words might induce larger preview effect than that of between-words. Some researchers used compound words (within-words) and phrases (between-words) to further explore this issue and they found that there were no differences between within-words and between-words in preview effects. The results cannot exclude the possible explanation that larger preview effect for within-words is caused by the higher syntactic expectations of nouns comes from adjectives in phrases inducing larger preview effect and then counterbalances the possible differences between the two kinds of words. The present study adopted the boundary paradigm to probe the preview processing differences between within-words and between-words. In the present study, the first character of a two-character compounds (between-words) and the second character of a two-character compounds (within-words) was manipulated to be presented normally or replaced by a pseudo-character for previews. Moreover, word highlighting sentence and non-word highlighting sentence were introduced to examine whether the word boundary information could exert different influences on the preview processing of between-words and within-words. Marking word boundary by word highlighting has its unique advantages. Most of all, compared with word spaces, word highlighting can not only keep the same sentence length meanwhile providing the word boundary information but also control the same word lateral masking on different conditions.

Firstly, the results indicated that the preview effect for between-word was smaller than that of within-word. The results were consistent with the results of Hy?n? et al. (2004), which showed that the morphological information of target word could impact on preview processing. Secondly, we found that there were no differences among normal condition, highlighting condition and non-word highlighting condition. Even so, we did not found the significant influences of word boundary demarcation for preview processing, the possible benefit effect of word boundary still could not be ruled out thoroughly. As Bai et al. (2008) pointed out that readers are familiar with the text without any word boundary signals in normal reading; consequently, the null effect between normal and word boundary text may show a priming effect on word boundary condition, which is the one readers are not familiar with. Thirdly, the results showed that word boundary information had similar effects on within-words and between-words.

Results of the present study indicated that word morphological information could affect its preview processing; however, word boundary information do not necessarily facilitate preview processing for both between-words and within-words. The possible explanation may be that word segmentation and word recognition occur simultaneously. These results are consistent with the model of word segmentation and word recognition.

Key words: word boundary, between words, within words, preview processing, word segmentation