Acquisition of linguistic information to the left of fixation during the reading of Chinese text

Abstract

The linguistic properties of the first (critical) character of a two-character Chinese word were manipulated when the eyes moved to the right of the critical character during reading to determine whether character processing is strictly unidirectional. In Experiment 1, the critical character was replaced with a congruent or incongruent character or left unchanged. Critical character changes did not influence the fixation duration, but incongruent changes led to more regressions than congruent changes. In Experiment 2, the critical character was replaced with either a homophonic or a non-homophonic character when it was to the left of fixation. The fixation following the change was now longer when the replaced character and the critical character were homophones than when they were phonologically dissimilar. These results indicate that readers obtain phonological and semantic information to the left of a fixated character and that the recognition of consecutive Chinese characters is not strictly unidirectional.

Acknowledgements

This research was supported by grants from the Taiwan National Science Council (95-2413-H-010-002), the theme project on brain, cognition and behavioral science from Academia Sinica (AS-93-TP-C05) and by Grant HD043405 from the National Institutes of Health (United States). We are grateful to Curtis Hardyck and Mallory Bersamira for their help with the paper. We also thank Sarah White and Barbara Juhasz for their comments on an earlier version of the article.

Notes

¹In Experiment 1, contextually congruent and incongruent critical (replacement) characters could form a word with the 2nd character in almost all cases (94% in the congruent and 91% in the incongruent condition). The word frequency of these words was 7.7, 9.5, and 3.3 per million in the identical, related, and unrelated conditions, respectively, and the difference between the congruent and incongruent condition was not significant, t(89) = 1.532, p = .129. Furthermore, the frequency of congruent and incongruent characters themselves was also similar in the two congruent and incongruent condition, I(89) = 1.293, p = .199.

²The semantic congruity of critical character replacement had no effect on R-AOI viewing even when we analysed only those cases in which the eyes landed on the second character of the target word. The numeric effect pattern matched that of the overall data. We considered these data less compelling due to the large variability in this relatively small set of data

³The probability of fixating at the second character of the target word ranged between 32%–36% in the four experimental conditions. Relative fixation frequency was increased to 61%–64% when the next character was included in the R-AOI.

⁴The by-item analysis was not possible here due to the limited number of observations per condition.