The effects of graphemic and phonetic cues in text reading: Evidence from eye movements

碩士 === 國立中央大學 === 學習與教學研究所 === 93 === This thesis is an eye movement study of Chinese character reading. The central question of the study is to investigate the prevalent but unclear phenomenon: Do homophonic or graphemic control characters interrupt readers when they are reading? In the pilot study...

Full description

Bibliographic Details
Main Authors: Ching-Wei Chang, 張靜維
Other Authors: Hwa-Wei Ko
Format: Others
Language:zh-TW
Published: 2005
Online Access:http://ndltd.ncl.edu.tw/handle/27217731217141895084
Description
Summary:碩士 === 國立中央大學 === 學習與教學研究所 === 93 === This thesis is an eye movement study of Chinese character reading. The central question of the study is to investigate the prevalent but unclear phenomenon: Do homophonic or graphemic control characters interrupt readers when they are reading? In the pilot study, articles were presented in EYELINK eyetracker, a large amount of general data from adult readers was collected, such as fixation duration, saccade length, and reading speed to establish the instrument reliability. Afterward, there are two experiments in the study. First of all, four types of characters were manipulated in the text - homophonic and graphemic similar characters, phonetic dissimilar but graphemic similar characters, homophonic but graphemic dissimilar characters, and phonetic and graphemic dissimilar characters. The results of the first experiment show that the phonetic and graphemic dissimilar characters slow down the readers the most. A second experiment was conducted to validate the previous results. In the second experiment, the texts were modified to produce semantic anomaly. Two types of characters, phonetic dissimilar but graphemic similar characters, homophonic but graphemic dissimilar characters, were embedded in the anomalous sentences. The results are consistent with those of the first experiment. It is indicated that only phonetic and graphemic dissimilar characters would interrupt reading, and the other three types of characters would not. The experiments support the fact that readers can access correct word meanings in a text even with similar but incorrect phonetic and graphemic clues.