Effect Of Working Memory Capacity On The Processing Of Pronouns In Weak Crossover Structure In L2 High School Students

According to the literature on pronouns,the cue-based retrieval mechanism affects pronoun resolution(Badecker & Straub,2002;Choy & Thompsom,2010),working memory capacity affects the language resolution model of the subjects,and inaccessible antecedents affect the pronoun resolution of the subjects with low working memory capacity(Cunnings &Felser,2013),but it is not clear whether the cue-based retrieval mechanism plays an important role in the early or late stage of pronoun processing(Lv Jun,2020).Syntactically inaccessible antecedents have no effect on pronoun resolution,but gender congruence affects pronoun anaphora(Sturt,2003),and antecedents in relative clauses do not affect the early processing of pronouns(Kush et al.2015).However,some studies have found that syntactically inaccessible antecedents can also constrain variable pronouns(Felser & Drummer,2017;Kush et al.2017),and the clue retrieval has better influence on pronoun processing than weak crossover syntactic principle(Kush et al.2017).In the early stage of pronoun processing,language processor only considers legal antecedents.Clue retrieval affects pronoun resolution at the late stage of processing(Sturt,2003;Chow et al.2014;Kush et al.2015;Qian Jingjing& Wu Fuyun,2016).Therefore,the author takes 104 students from two classes of a senior high school in Zhangzhou City as the research subjects,and studies whether working memory capacity affects pronoun processing of senior high school students under different syntactic principles through offline Oxford cloze test,online operation span working memory capacity test and self-paced reading task test.The data are analyzed by spss21.0 and this paper attempts to answer the following two questions: 1)Is the cue-based retrieval mechanism superior to the syntactic principle of weak crossover? 2)In the online test,do the high and low working memory groups show superiority in the cue-based retrieval mechanism when they respond to pronoun processing under the noun type effect in the weak crossover /complement clause structure?The results are as follows:(1)In the online test,the author took reading judgment score as dependent variable,noun type,syntactic structure,gender matching and working memory capacity as independent variables,and conducted a 2 × 2 × 2 four factor ANOVA.The results show that the main effect of pronouns and names is significant,with F(1,383)= 41.180,P = 0.000;the main effect of gender consistency is significant,with F(1,383)= 5.278,P = 0.022;the main effect of working memory span is significant,with F(1,383)= 4.649,P = 0.032;the main effect of syntactic structure is significant,with F(1,383)= 4.060,P = 0.045.However,the interaction effect between working memory capacity and different syntactic structures was not significant,with F(1,383)= 0.103,P = 0.749,indicating that the interaction between working memory and different syntactic structures had no significant effect on the accuracy of online reading judgment.(2)In the self-paced reading online test,the response time of pronoun in critical region was taken as the dependent variable,and the working memory capacity,noun type,syntactic structure and gender matching were taken as the independent variables.It is found that the main effect of working memory is significant,with F(1,1535)= 945.658,P = 0.000;the main effect of noun type is significant,with F(1,1535)= 28.486,P = 0.000;the main effect of gender consistency is significant,with F(1,1535)= 36.496,P = 0.000;the main effect of syntactic structure is significant,with F(1,1535)= 38.661,P = 0.000.The interaction between working memory capacity and syntactic structure is significant with F(1,1535)=14.565,P = 0.000,the significant value is less than 0.05,which indicates that working memory capacity can affect the response time of pronoun processing under different syntactic structures.(3)In the online test,the author takes the reaction time of the critical region of high and low working memory groups as dependent variables,and takes pronoun name,gender matching and syntactic structure as independent variables to carry out a 2 × 2 × 2 three factor ANOVA.The main effects of noun types,gender matching,and syntactic structure in the high working memory group all reached a significant level(F values were 10.248,12.203,and3.098,respectively,and the significance levels were 0.000,0.001,and 0.001,respectively);The main effects of type,gender matching,and syntactic structure in the low working memory group also reached a significant level(F values were 18.697,25.162,and 47.325,respectively,with a significant level of 0.000).The paired sample t test results showed that in the online processing response of the low working memory group,there was only a significant difference between the gender inconsistency of the pronoun complement clause and the gender of the name complement clause,with a significant probability of 0.003;the significant probability of the other three groups were 0.055,0.60 and 0.159;while the high working memory group has no significant difference in the four sets of paired variables,and the significant probabilities are 0.096,0.365,0.60,and 0.080,respectively.The response time of the two groups of subjects under the gender-unmatched stimulus conditional sentence was higher than that of the gender-matched stimulus sentence.Research shows:(1)Regardless of the syntactic structure,the accuracy rate of the subjects in the low working memory group was lower than that of the subjects in the high working memory group.Moreover,under the gender matching effect,the accuracy rate of the subjects under the gender-consistent syntactic structure was higher than that of the gender-inconsistent syntactic structure.This research result is the same as that of Lv Jun(2020),Kush et al.(2017),Felser& Drummer(2017)and others.It shows that the cue-based retrieval mechanism has priority in sentence processing.(2)The response time of online processing in the low working memory group under any syntactic structure was higher than that in the high working memory group.Moreover,no matter what the syntactic structure is,the participants’ response time to conditional stimulus sentences containing gender-mismatched pronouns and antecedents was longer than that of gender-matched conditional stimulus sentences between pronouns and antecedents,which indicates that the weak crossover syntactic principle has priority in sentence processing.The research result is the same as the research results of Badecker & Straub(2002),Sturt(2013),and Cunnings & Felser(2013).(3)The low working memory group only had differences in the gender inconsistency of pronoun complement clause and person name complement,while the high working memory group had no differences in the four variables.In other words,the high working memory capacity group can process sentences online without relying on semantic clues.This is consistent with the results of Lewis & Vasishth 2005;Lewis et al.2006,but inconsistent with the results of Felser & Drummer(2017)German study and Kush et al.(2017)English self-paced reading study,which indicates that the low working memory group,due to its low working memory capacity,will establish a dependency relationship based on semantic feature matching,thus reducing the working memory capacity load.Therefore,in teaching,teachers should pay attention to the cultivation and improvement of learners’ working memory ability,properly teach the differences of syntactic principles and syntactic morphology,strengthen the processing of complex sentences,and reduce the consumption of cognitive resources.