| In natural reading,readers can not only process the currently fixated word in foveal vision,but also preprocess the upcoming word in parafoveal vision.Parafoveal preprocessing,which plays a critical role in skilled reading,involves the acquisition of lexical information in word recognition and post-lexical information in word integration.As to the question how foveal load modulates parafoveal preprocessing,the E-Z reader and SWIFT model,which are the two most prominent models,have different hypotheses in lexical level processing and are poor interpretative in post-lexical level processing.These two models diverge in their predictions about the interaction between foveal and parafoveal processing in lexical level.According to the E-Z reader model,words are processed serially and the depth and span of parafoveal information acquisition decreases with increasing foveal load.In contrast,the SWIFT model suggests that words are processed in parallel and the span of parafoveal information acquisition decreases with the increase of foveal load while the depth of parafoveal information acquisition increases.Previous researches in alphabetic reading investigating the modulation on the depth of parafoveal processing are based on varied preview span and thus induce inconsistent results,which means that hypotheses of these two models still need to be verified.Only the E-Z Reader model can partially explain the parafoveal processing of post-lexical integration,such as the effect of preview plausibility that has been demonstrated by large numbers of researches,while the SWIFT model lacks a theoretical explanation of post-lexical preprocessing.The current research focused on how foveal load modulated parafoveal processing of lexical and post-lexical information,which was composed of three studies containing a series of six experiments totally.Foveal load was manipulated by the word frequency of foveal words.The first study investigated the modulation of foveal load on the depth of parafoveal lexical processing,containing a series of three experiments.In experiment 1,we manipulated the fovea load(low,high)and the preview type(identical,orthographically related,unrelated)to testify how parafoveal orthographic processing was modulated by foveal load.The results showed that preview words that were orthographically related to the targets did not cause increasing skipping of the target words or decreasing first-pass fixation duration relative to preview words that were unrelated to the targets,indicating no orthographic preview effect or the modulation of foveal load.In experiment 2,the fovea load(low,high)and the preview type(identical,phonologically related,unrelated)were manipulated to investigate how foveal load modulated parafoveal phonological processing.The results showed that previews words phonologically related to the targets induced increasing skipping of target related regions and decreasing gaze duration only when foveal load was high,which indicates that readers can not obtain phonological information parafoveally under low foveal load.Experiment 3 investigated the interaction between foveal load and parafoveal semantic processing by manipulating the fovea load(low,high)and the preview type(identical,semantically related,unrelated).We found semantic preview effects whenever foveal load was low or high,indicated by decreasing first fixation duration and single fixation duration on target related regions when preview words were semantically related to the target.This result suggests that readers can obtain semantic information parafoveally while such processing is not affected by the foveal load.The second study examined how foveal load modulated the span of parafoveal lexical processing,which contained two experiments.In experiment 4,a three-character word was used as the parafoveal word and the fovea load(low,high)and the preview span(0-character,1-character,2-character,3-character,full sentence)was manipulated to investigate how foveal load and the span of parafoveal lexical processing interacted.It was found that the preview span was larger than three characters whenever foveal load was low or high,indicated by longer first fixation duration,single fixation duration and gaze duration on targets when previewed with three characters of the target word than when with full sentence.This result implies that readers can obtain lexical information from more than three characters parafoveally while such processing is not affected by the foveal load.Three one-character words were adopted as the parafoveal words in experimental 5 in order to enlarge the difficulty of parafoveal processing and observe the potential modulation of foveal load.The fovea load(low,high)and the preview span(0-word,1-word,2-word,3-word,full sentence)was manipulated similarly.Dissimilar results were found in this experiment: when foveal load was low,readers provided with three words parafoveally showed longer single fixation duration on target words than with full sentence parafoveally,indicating the preview span larger than three words(characters);when foveal load was high,fixation durations on targets for 3-word preview were not significantly different from those for full sentence preview,indicating a three-word(character)preview span.Such results suggest that the number of words(characters)that readers can preprocess parafoveally decreases with the foveal load.The third study was conducted to explore how foveal load modulate parafoveal post-lexical processing.In experiment 6,the fovea load(low,high)and the preview type(identical,unrelated but semantically plausible,unrelated and semantically implausible)were manipulated,with plausibility referring to a post-lexical integration.An interaction was found between foveal load and the parafoveal-on-foveal effect of plausibility: when foveal load was low,preview words that were unrelated to the targets and implausible with the preceeding context caused no longer first fixation duration and daze duration on pretargets than those that were unrelated to the targets but plausible with the preceeding context,with the latter causing longer regression path duration on pretargets;when foveal load was high,first fixation duration,gaze duration and regression path duration on pretargets for unrelated and semantically implausible preview were significantly longer than those for unrelated but semantically plausible preview.These results suggest that post-lexical integration of plausibility information decreases with the foveal load.Under the experimental conditions,the following conclusions can be drawn.Firstly,the foveal load modulates both the depth and the span of parafoveal lexical processing.In addition,the direction of the foveal load modulation on parafoveal lexical processing is not in accordance with the E-Z Reader model,but is in line with the parallel processing hypothesis of the SWIFT model.Furthermore,the foveal load modulates parafoveal post-lexical processing and readers discontinue the integration of parafoveal words when the foveal load was low.Finally,the direction of the foveal load modulation on parafoveal post-lexical processing suggests that parafoveal integration processing occurs in parallel with foveal processing. |
PDF Full Text Request