How Does Non-linguistic Gender Information Affect Linguistic Gender Information Processing In Chinese Learners Of English

The misuse of third person singular pronouns is very common in English speech of Chinese learners of English.In Chen(2004),the gender error rate was 17.65%.Even among advanced Chinese learners of English,the average gender error rate was6.47%(Dong & Jia,2011).Studies(Dong & Li,2011;Dong,Wen,Zeng,& Ji,2014)assert the cause of gender errors lies in the absence of gender information encoding in the Conceptualizer.In English,third person singular pronouns referring to female and male entities are distinct in both oral and written forms(e.g.,she/he),while in Chinese,the counterparts only differ in the written(e.g.,她/他)but not the oral forms(e.g.,ta/ta).Such difference might be one of the reasons accounting for Chinese learners’ failure to encode gender information in producing third person singular pronouns in English.Nonetheless,few studies explore how enhancement of gender information affects gender encoding.The present study thus aims to find the influences of non-linguistic gender information on linguistic gender information processing.If non-linguisitc gender information helps gender encoding,its enhancement would affect linguistic gender information processing,otherwise,it would not.Both talker voices and pictures of people are strong carriers of gender information.And we can extract non-linguistic gender infromation directly and quickly from these two carriers.In our study,we employed self-paced reading paradigm and recorded the reading time for pronouns either matching or mismatching its antecedent in gender.The reading time for two chunks of words following the pronouns was also recorded in case of a spillover effect.To ensure that any difference in the reading time was the result of experimental manipulation,a baseline experiment without additional gender information was conducted.Results showed that in both Chinese and English reading,there was no significant difference in the reading time when antecedents and pronouns either matched or mismatched in gender.Experiment One was to explore how auditory gender information affects linguistic gender information processing in Chinese learners of English.Experiment One was identical to the baseline experiment except that participants heard talker voices lasting almost 1000 ms before the reading.The gender of talkers was in line with the antecedent in the reading.The results showed that in Chinese reading,when antecedents mismatched pronouns in gender,the reading time for the second chunk after the pronoun was significantly longer than the time in the matched condition.However,in English reading,there was no significant difference in the reading time when antecedents and pronouns either matched or mismatched in gender.This was probably due to English listening anxiety.Altogether,these results indicate that talker voices affect linguistic gender information processing.Experiment Two was to investigate how visual gender information affects linguistic gender information processing in Chinese learners of English.Two groups of pictures were involved: pictures of Chinese people and pictures of American white people,with an aim to split gender information and race information.The experiment was conducted in English.Participants saw pictures before reading and the gender of the people was in line with the antecedent in the reading.The results showed that in the sub-experiment with pictures of Chinese people,when antecedents mismatched pronouns in gender,the reading time for the pronoun was significantly longer than the time in the matched condition.However,in the sub-experiment with pictures of American white people,there was no significant difference in the reading time between the matched and unmatched conditions.This was probably due to the other-race effect which delayed gender information extracting.Altogether,these results indicate that gender information in pictures of people affects linguistic gender processing.In conclusion,results of our study verified that non-linguistic gender information affects linguistic gender processing.The enhanced gender information promotes gender encoding.Our study also supports the account that the cause of gender errors lies in the absence of gender information encoding.