Phrases And Sentences Of The Semantic, Syntactic And Contextual Effect Research

An ERP experiment was conducted to investigate semantic and syntactic processes,as well as explore semantic contextual effect in Chinese phrases and sentencescomprehension. Participants were presented with Chinese words, which wereconstructed to induce either phrase or sentence task. Participants were asked to judgethe rationality of all phrases or sentences. Phrase task included match, classifier-nounmismatch, and classifier-verb mismatch, while sentence task included match, localmatch, classifier-noun mismatch, and classifier-verb mismatch. The materials of thephrase task (e.g., a (zhang: classifying thin and complanate things)+sheet/heart/insist)were contained in the sentence task to constitute the local context of the sentence task,and the subjects (e.g., mother) and predicates (e.g., sets) constituted the global contextof the sentence task.Behavior data found that semantic violations accompanied longer reaction timesand lower accuracy rates in both tasks, but there were no differences between syntacticviolations and match. Behavioral performance was better for the sentence task than forthe phrase task while local context in both tasks was the same.EEG data showed that classifier-noun mismatch presented an apparent N400effectin midline sites and right fronto-temporal for phrase and sentence tasks, however,classifier-verb mismatch only showed the N400effect in parieto-central for the sentencetask. In parieto-central area, P600effects were found in the sentence task but not in thephrase task. These indicated that Chinese semantic and syntactic processes couldn’t becompletely separated, and syntactic integrations couldn’t be easily realized until in acertain context. In addition, there were more negative N400s for phrase task than forsentence task in parieto-central, suggesting it was more difficult to perform the phrasetask. Activation of N400was more robust in left fronto-temporal than rightfronto-temporal in the phrase task, but the activation was the opposite in the sentencetask. Brain-posterior P300was an index of semantic violation. There was nohemispherical difference for the P300effects in the phrase task, but there was much larger P300in left than right hemisphere in the sentence task. The hemisphericaldifferences of N400and P300indicated the left hemisphere was the dominanthemisphere for language processing, and more right-hemispherical areas were involvedin the sentence processing.sLORETA analysis found that there were no activation differences during allconditions in the phrase task for the N400and P600time windows. But in the sentencetask, the N400effects were mainly located in fronto-temporal network, and the P600effects were mostly located in fronto-parietal network. In N400time window,activations of the parietal, frontal and temporal lobes were much stronger in the phrasetask than in the sentence task. This might be due to a shorter context in the phrase taskresulting in more difficult in retrieving, accessing and integrating words. In P600timewindow, activations of inferior frontal gyrus, superior temporal gyrus, postcentral gyrusand superior parietal lobule et al. were stronger in the sentence task than in the phrasetask. This might be due to a longer context in the sentence task resulting in moreterminal integrations of semantic and syntactic involved.Therefore，Chinese can not be understood preferably when it is much concise. Thisstudy supports the facilitation effect of sentence context.