SIZE OF SACCADE AND FIXATION DURATION OF EYE-MOVEMENTS DURING READING - PSYCHOPHYSICS OF JAPANESE TEXT-PROCESSING

The size of saccadic eye movements and eye fixations during Japanese text reading (written in both hirakana phonograms and kanji ideograms) were analyzed. Hirakana is sound based, i.e., it consists of symbols for syllables, but such text was processed differently from English text in terms of saccadic eye movements and fixations. In experiment 1 the reader was asked to read three types of text (hirakana only, kanji-hirakana mixed, and English) in a natural reading manner. The results show that kanji-hirakana-mixed text, which has picturelike symbols, requires shorter eye fixations and longer saccades than those required by hirakana text. The findings show that kanji and hirakana are processed differently. Logographlike kanji text greatly facilitated text processing and understanding during reading. In experiment 2 we found that kanji-based text had a wider perceptual span of reading than did hirakana text. The wider span is consistent with the longer saccade length for kanji-based text found in experiment 1. In experiment 3 we studied the convenient viewing position, i.e., the position within a word where the eye should fixate first for the isolated word to be recognized most quickly. A convenient viewing position was confirmed to exist both for hirakana- and kanji-based words, but for hirakana words the position changed as word length changed, as was expected from the results of experiment 2. In experiment 4 we measured lexical access time for hirakana- and kanji-based words and found that the reaction time to kanji-based words was faster than the reaction time to words written with hirakana. In experiment 5 we studied the detection of kanji and hirakana, using spatial-filtering techniques, i.e., Gaussian, Laplacian-Gaussian, and low pass, and found that the space constant required for detection was larger for hirakana than that for kanji. Moreover, the critical spatial-frequency bandwidth required for hirakana was found to be lower (2 cycles per character) than that for kanji (4-8 cycles per character). The results show that kanji and hirakana are detected in different ways in the early stages of visual processing and are consistent with the results of experiment 1. A dual-strategy model for the computation of the saccade length and fixation duration to the next word in the parafovea was proposed, based on a linguistic-control hypothesis of eye movement during Japanese text reading.