A Developmental Study On The Role Of Different Linguistic Codes In Parafoveal Processing

In recent years, the study on parafoveal processing is one of the most flourishing and debating fields of eye movement research in reading. There has been a growing interest in whether the word information can be extracted from the parafovea, and if so, what kind of information can be obtained. Many studies have been carded out to explore these questions in alphabetic languages. In the present study, the boundary paradigm was used to investigate the role of different linguistic codes in parafoveal processing and its developmental characteristics in Chinese sentence reading.Three experiments were conducted in this study:Experiment 1 aimed at exploring the linguistic information extraction parafoveally at the character level. There were five preview types: graphemically similar homophonic previews; graphemically dissimilar homophonic previews; graphemically similar nonhomophonic previews; synonyms; and controls. The target characters frequency were manipulated to determine its modulation of parafoveal preview benefit.Experiment 2 manipulated the regularity of the phonetic radical with the whole character to determine the phonological and graphemic preview benefit at the subcharacter level. There were four preview types: homophones sharing phonetic radicals, nonhomopbones sharing phonetic radicals, visually dissimilar homophones and controls.Experiment 3 manipulated the semantic relatedness between the semantic radicals and the whole characters to determine the graphemic and semantic preview benefit at the subcharacter level. Four preview types were set, which were synonyms sharing transparent radicals, nonsynonyms sharing semantic radicals, visually dissimilar synonyms and controls.The results showed as follows:Firstly, the main effect of grades was significant on the viewing times. The fixation duration on different preview types was longer for junior middle school students than senior middle school students and college students.Secondly, the phonological codes in the parafovea did contribute to Chinese character identification at both the character and subcharacter level. The graphemic information also had an early facilitatory influence; however, the semantic preview benefit was difficult to be obtained parafoveally.Thirdly, the target characters frequency and the regularity influenced on Chinese character identification. The fixation durations on the low-frequency, phonetically irregular characters were longer than the high-frequency, phonetically regular characters. And the preview benefit was influenced by the characters frequency, regularity and the age.Finally, our findings were generally compatible with the E-Z reader model. Meanwhile, evidence for the linguistic control hypothesis was given. The linguistic information could beobtained parafoveally to guide the eye movements of the readers.