An Eye-movement Study On English Pronominal Anaphora Resolution Of Native And Chinese Speakers Of English

Pronouns contain only limited information about their antecedents such as gender,number and case,which are frequently ambiguous.Pronominal anaphora resolution has always been one of the major concerns of psycholinguistic research.A number of studies have been carried out to investigate the mechanisms of pronoun processing by using experimental methods like the naming task,the probe recognition task,the self-paced reading task etc.The effect of gender ambiguity on English pronominal anaphora resolution has been investigated in a number of studies(e.g.,Caramazza et al.,1977;Stevenson & Vitkovitch,1986;MacDonald & MacWhinney,1990;Rigalleau & Caplan,2000)but the results were not consistent.The effect of processing depth on English pronominal anaphora resolution were less frequently discussed(e.g.,Rigalleau et al.,2004;Stewart et al.,2007;Wu et al.,2018)and there are still some limitations due to their experimental methods and design.Eye-movement experiment can provide resourceful information of the online sentence processing in reading.Only a few existing studies investigated the influence of gender ambiguity and processing depth on English pronominal anaphora resolution of Chinese speakers of English and eye-movement experiments are not often used.The current study uses four eye-movement experiments to investigate the influence of gender ambiguity and processing depth on pronominal anaphora resolution of native and Chinese speakers of English with the focus of the following four research questions:(1)How do gender ambiguity and processing depth affect the pronominal anaphora resolution of native English speakers?(2)How do gender ambiguity and processing depth affect the pronominal anaphora resolution of Chinese speakers of English?(3)What are the differences between native and Chinese speakers of English in pronominal anaphora resolution under deep processing?(4)What are the differences between native and Chinese speakers of English in pronominal anaphora resolution under shallow processing?In the eye-movement experiments,participants read sentence items with gender-ambiguous/ gender-unambiguous pronouns and answer a question after each sentence by pressing computer keys.There are three conditions(ambiguous vs.subject-antecedent vs.object-antecedent)of each sentence items and two types of questions(comprehension question vs.Yes/No question)after the sentence items.Experiment 1 and 2 investigate the pronominal anaphora resolution under deep and shallow processing of native speakers of English and Experiment 3 and 4 investigate the same issue of Chinese speakers of English.The main findings are as follow:Firstly,English pronominal anaphora resolution of native speakers of English process under deep processing differ from that under shallow processing.Under deep processing,sentence items with a gender ambiguous pronoun were processed easier in early processing but more difficult in later processing,suggesting that early disengagement of the unambiguous pronoun with a salient non-referent is effortful and later competition between interpretations of the ambiguous pronoun increased the processing difficulty.Under shallow processing,the ambiguous condition enjoyed an advantage in processing over the subject-antecedent condition and the object condition,suggesting an underspecified representation of the ambiguous pronoun and a disengagement of the unambiguous pronoun with its non-referent.The shallow processing results reconcile with the “good-enough” representation accounts.The finding of disengagement under shallow processing is new to the research of English pronominal anaphora resolution and it provides evidence that disengagement is a non-strategic mechanism under different depths of processing.Secondly,English pronominal anaphora resolution of Chinese speakers of English process under deep processing differ from that under shallow processing.Under deep processing,the ambiguous condition has lower fixation probability than the unambiguous conditions in early processing but longer reading times in late processing.It provides evidence for an early facilitation and late suppression of gender ambiguity,similar to the pattern of that of native English speakers.In shallow processing,reading times of the ambiguous condition was no difference with the unambiguous conditions but reading time of the object-antecedent condition was longer than the subject-antecedent condition,which suggests that Chinese speakers constructed underspecified representation of ambiguous pronouns which were “good-enough” for the task at hand and there was still a non-strategic disengagement process.The difference of ambiguity effect on pronominal anaphora resolution in early and late processing and the existing of disengagement of unambiguous pronoun under shallow processing are new to the research of English pronominal anaphora resolution of Chinese speakers of English.Thirdly,the pronominal anaphora resolution under deep processing for native speaker of English is similar to that of Chinese speakers of English in that they both have an early facilitation and late suppression of gender ambiguity.However,the exact time and region that the ambiguity effect takes place are different for two groups of English speakers.In early processing,the facilitation of gender ambiguity occurred before the pronoun region of Chinese speakers but at the pronoun region of native speakers.During late processing,gender ambiguity had a suppression on reading times of both native and Chinese speakers of English,showing that two antecedents of the ambiguous pronoun compete which resulted in processing difficulties.Fourthly,under shallow processing,both native and Chinese speakers of English constructed “good-enough” underspecified representation of the ambiguous pronoun,but the degree of underspecification and difficulties of disengagement of two unambiguous pronouns are different.The underspecified representations constructed by native speakers were less complete and hence easier to process.Disengagement of unambiguous pronouns are comparatively more effort-consuming but the two unambiguous conditions did not differ from each other.For Chinese speakers of English,underspecified representations of the ambiguous pronouns made the pronominal anaphora resolution fairly easy,but not any easier than the unambiguous pronouns.Reading time of the object-antecedent condition was longer than the subject-antecedent condition of Chinese speakers,while the reading time of the ambiguous condition was lower than both the unambiguous conditions.These differences indicate that the pronominal anaphora resolution of native speakers is not affected by gender ambiguity,salient of antecedent and other factors.Underspecified representations are constructed just sufficient for the task at hand.However,of Chinese speakers of English,they are still sensitive to such factors and are influenced to a larger extent in such second language comprehension.The underspecified representations constructed by Chinese speakers of English are more specified that that are constructed by native speakers.The current eye-movement study investigated the pronominal anaphora resolution under deep and shallow processing for native and Chinese speakers of English and shows its significance in three aspects.Firstly,it provides evidence for some existing arguments about English pronominal anaphora resolution and helps to clarify some unresolved questions.Secondly,it supplements and brings fresh ideas for the existing studies about English pronominal anaphora resolution of Chinese speakers of English by discussing the new factor of processing depth.Thirdly,it verifies the original theoretical by providing evidence for a nonstrategic disengagement process for both groups of speakers.Lastly,the comparison of native and Chinese English speakers enlightens the teaching and learning of English reading and also the practice of English-Chinese translation and interpretation.