关键词:  中文阅读, 视觉编码, 消失文本范式, 眼动

Abstract:  Encoding the visual information of text into a more permanent representation is necessary for text processing and comprehension. It was found that young English readers are able to encode all the visual information of a fixated word within 55 ms, but cannot encode word n+1 within this time in English reading. Evidence has shown that young Chinese adult readers could not only encapsulate a fixated word as soon as young adults English readers, but also encode the text of word n+1 within this time. It has also been found that the process of encoding a fixated word in Chinese reading is more susceptible to the effects of childhood development and age than English; thus, it may be language-related. With two disappearing text experiments, we explore the characteristics of text and visual encoding of words n and n+1 during Chinese reading. By doing so, we want to clarify whether the process of encoding visual information during Chinese reading is word-or character-based. If the encoding process is character-based, the impairment in reading performance from disappearing manipulations should be mediated by the number and locations of disappeared characters. On the contrary, a word-based hypothesis denies these results. The sentences used in both experiments consisted of 7 or 8 two-character words. The first experiment was conducted to explore the visual encoding in preview frames. The experiment was a 4 (disappearing manipulations: no disappearing vs word n+1 disappearing vs character n+1 disappearing vs character n+2 disappearing) ×2 (disappearing onset: 0 ms vs 60 ms) mixed design. After word n was fixated for 0 or 60 ms in each condition, word n+1, the left character of word n+1, and the right character of word n+1 disappeared respectively. An immediate refixation on word n did not result in the reappearance of disappeared text. Reading time was the most important reference in the present study. The results showed that the disruption caused by disappearing manipulations was mediated by the number and locations of disappeared characters; thus, indicating that the encoding process in preview frames is character-based during Chinese reading. The second experiment also had four disappearing manipulations and two disappearing onset times. This experiment was a 4 (disappearing manipulations: no disappearing vs word n disappearing vs fixated character disappearing vs no fixated character disappearing) × 2 (disappearing onset: 0 ms vs 60 ms) mixed design. In the word n disappearing condition, word n disappeared after it was fixated for 0 or 60 ms. In the fixated character disappearing condition, the fixated character disappeared after word n was fixated for 0 or 60 ms. In the no fixated character disappearing condition, the nonfixated character, which also belonged to word n, disappeared after word n was fixated for 0 or 60 ms. An immediate refixation on word n did not result in the reappearance of disappeared text. It was found that all of the disappearing manipulations disrupted reading performance equally when the disappearing onset time was 0 ms, and that not all of them prolonged reading time in the conditions with a 60 ms onset time. These results indicated that the encoding process in fixation frames is word-based during Chinese reading. By integrating the two results mentioned above, it may be concluded that Chinese readers encode the visual information of word n and characters in word n+1 parallel. Further, the encoding process is character-based in preview frames and word-based in fixation frames.

Key words: Chinese reading, visual encoding, disappearing text, eye movements