Revealing The Time Requirements Of Information Extraction From Words In Foveal And Parafoveal For Reading In Chinese

There is a very close relationship between this eye movement behavior and the mental activity that underlies reading.The new visual information necessary for reading is acquired by the processing system during eye fixations that typically last about200-250ms. The information abstraction (encoding) is not only the beginning of lexical identification, the primary approcah to post-lexical integration or high linguistical processing, but also plays an essential role during coordineted process of visual system with cognitive systems. Thus it is an important and fundmental reading skill for readers if they can efficiently acquire enough information for normal reading during one fixation. However, while numerous studies have examined the temporal characteristics of the information acquisition during fixational pauses when reading aphpabetic language, little is known about how rapidly information is acquired from different gaze locations when reading non-alphabetic languages such as Chinese. The main purpose of current research is to investigate the efficientcy of extraction (encoding) of useful information both from words at the point of fixation and from words at different eccentricities to the right and left of this location when reading in Chinese, and also to investigate whether there are developmental changes in the time needed to capture enough information for normal Chinese reading. In particulay,the present study used the disappearing text paradigm in three sub-studies (eight experiments) to manipulate the display duration for words in foveal area (word n) and for words in parafoveal area (word n+1and word n-1) separately.Study1involving three experiments explored the time needed to acquire information from fixated words (word n) in the foveal area for reading to be normal. In experiment1, participants read sentences composed uniformly of two-character words, including a target word that was of high or low lexical frequency.Sentences were displayed normally or the fixated word was visible for only a short period (20,40,60, or80ms) before disappearing. Overall sentence reading times were normal when words were displayed for as little as40ms, and the effects of word frequency on fixation times for target words were normal even when words were displayed for only20ms following the onset of fixation. However, readers made fewer and longer fixations and slightly more regressions for disappearing text. The results indicated that, fixated words need only to be visible for very briefly during fixation for normal processes of word identification to take place. In experiment2, the longer displayed durations (80ms,110ms,140ms,170ms,200ms) were manipulated to assess the reasons for regressions increasing in experiment1. The results showed that the number of regresstions were no longer increased when words were displayed for80ms and more, supporting the assumption that the regression increasing caused by unsufficiency time displayed for sublexical processing to take place, instead of low level of visual changes. In experiment3, to examine whether the regularity of using sentences composed of sequences of2-character words have contributed to this effect in experiment1, the same100experimental sentences were presented but intermixed with a large set of filler sentences(100special fillers and200normal fillers) that were composed of words of varying length. The results replicated the experiment1and also denied the possibility of getting contribution from regularity of two-character words sentences.Study2explored the time needed to acquire information from words at different locations in parafovea for reading performance to be normal. In experiment4and5, participants read sentences presented normally or under’ disappearing text conditions in which the word n+1(experiment4) or word n-1(experiment5) will disappear a fixed period after the beginning of fixation (the same duration as in experiment1) while the fixated word n will remains available throughout the fixation.There was surprisingly little disruption to overall reading time when word n+1is visible for40ms or more following fixation on word n, which contrasts with findings for alphabetic languages, although readers made longer fixations and re-fixated words more often.Similarly,there was surprisingly little influence of the disappearance of word n+1on the lexical processing of words, even when word n+1was visible for as little as20ms. Moreover, the effect of word frequency on word-skipping in particular shows that lexical identification was initiated even for parafoveal words displayed for as little as20ms following fixation on word n. The results indicated that Chinese readers could acquire information of word n+1at the beginning period of fixation on word n (about20ms or less), even the amount of information could not reach the highest level of requirements for normal reading. The results of experiment5showed that, there were considerable disruptions on overall reading times when reading all disappearing texts (20ms-80ms, even more), but little influence on he lexical processiong of words. It indicated that the important of word n-1for Chinese reading, but the pattern of time needed for word n-1is quite different from that for word n+1, at least beyone the first80ms and possibly throughout a whole fixation.In experiment6, we sought to further examine the difference of time needed between word n+1and word n-1, which located at different directions of fixated word but similary area of visual acuity, under more extreme disappearing text conditions. Here we manipulated simultaneously both the location (word n+1and word n-1) and display time (0ms and40ms) in the disappearance of parpfoveal words. There were most serious disruption on overall reading times caused by word n+1for0ms displays,and more important, for word n+1disappearing, the the disruptions casused by0ms and40ms displays were equal. Similarly, The difference caused by0ms displays were larger than that caused by40ms displays; but for word n-1disappearince of influences on lexical processiong of words between0ms and40ms conditions for word n+1disappearing was larger than that for word n-1disappearing. Thus, it again supported the findings from experiment4and5, and the information of word n+1and n-1were acquired at different period of fixation on word n, although they were both very important for reading.Study3explored whether there was a developmental change in the speed of information capturing during fixations in reading. Two experiments recorded the eye movements of Chinese developing child readers (10-toll-yrs)and skilled adult readers(18-to21-yrs)(experiment7:20ms,40ms,60ms,80ms or control; experiment8:20ms,60ms,100ms,140ms or control).Both experiments showed that there were no overall cost to reading times for either age group except in the20ms display condition, which was longer compared to when sentences were shown normally. However, the disappearing text manipulation has more effect on children’s eye movements than adults’. Children make more regressions back to the target words and more changes on refixation than adults, when reading disappearing text. More important, when displayed time (20ms) was insufficient for both adult and children to obtain the necessary informmtion from fixated words, chidren showed more overall reading time cost than adults for normal reading processing to be initiated. To conclude,(1) Fixated word in foveal need only to be presented very briefly (about50ms) during fixation for lexical processing and reading comprehension to be relatively normal, which replicated previous findings for alphabetic languages.(2)The extraction of usful information from word n+1in parafoveal is initiated in the early period of fixation on word n(within50ms, even less20ms), readers could obtain most or all of nececcary information from word n+1for norma lexical and reading processing during this brief period of time.(3) The necessary information in both foveal (word n) and parafoveal (word n+1) for normal reading to take place might occur simultaneously during the beginning brief period of a fixation. These findings reveal the remarkable efficiency of the visual and lexical processing of text when reading Chinese, and which can be better accounted for by the parallel procseeing models of eye movements control in reading.(4) The information of word n-1is very important for reading which also located in parefoveal area, unlike word n+1, the information extraction from word n-1occurred at relatively late period of fixation on word n (later than80ms).(5) There were some developmental changes in the time needed for information capture during a fixation in Chinese reading, indicated developing child readers’ encoding of words is slower than that of skilled adult readers.