The Role Of Semantic Transparency In English Compound Processing In Chinese EFL Learners

The representation and processing of poly-morphemic words(i.e.,derived,inflected and compound words)has been one of the central focuses in psycholinguistic studies over the past four decades.Recently,compound words that consist of at least two individual morphemes have been gaining popularity among scholars at home and abroad,for its unique word formation characteristics provide an ideal condition for researchers to examine the morphological and semantic properties of morphemes in mental lexicon.There has been a fierce debate about whether individual morphemes have independent representation units in mental lexicon.However,the majority of the current empirical evidence came from L1 compound processing,while the number of relevant literature in L2 field was very limited and the conclusions were rather mixed.Moreover,the current L2 researches mainly focused on exploring the mental route of compound processing,but not the potential factors that may influence the compound processing.To bridge such a gap,this study aimed to investigate the role of semantic transparency of the constituent morphemes in English compounds in highly proficient Chinese EFL learners via a masked constituent priming experiment and a masked semantic priming experiment with visual lexical decision tasks.There were two independent variables in this study,namely compound type and prime type.The target bimorphemic compounds were divided into four categories on the basis of the semantic transparency of their constituents,namely TT(e.g.,haircut),TO(e.g.,background),OT(e.g.,cowboy),and OO(e.g.,honeymoon),in which “T” stands for“transparent” while “O” refers to “opaque”.Prime types for each experiment included morphologically(experiment 1)or semantically(experiment 2)related condition and unrelated condition.A total of 44 postgraduate English majors(2 males and 42females)from a national key university in Jiangsu Province participated in the study.There were two experiments.Experiment 1 was the masked constituent priming experiment,in which “constituent-compound” pair was used as the related condition(e.g.,hair-haircut),while “unrelated simple word-compound” pair as the baseline condition(e.g.,son-haircut),and the Stimulus Onset Asynchrony(SOA)was set as 50 ms.Experiment 2 was the masked semantic priming experiment,in which“semantically related prime-compound” pair was adopted as the related condition(e.g.,head-brainstorm),while “unrelated simple word-compound” pair as the baseline condition(e.g.,word-brainstorm),and the SOA was 200 ms.Each experiment contained 40 key compound stimuli,10 for each condition,and 40 compound non-words to counterbalance the YES/NO responses.Participants’ response accuracy and RTs were tested by two-way repeated measures ANOVA on SPSS 26.From the statistical analysis of the behavioral data in the two experiments,several findings were obtained as follows:1.As for the role of compound type in compound processing,a main effect of compound type was observed not only in the masked constituent priming experiment but also in the masked semantic priming experiment,indicating that the semantic transparency of the individual constituent morphemes played a vital role in compound processing in the advanced Chinese-English bilinguals.Besides,participants in this study seemed to be more sensitive to the semantic transparency of the first constituent,which is different from the results in L1 field.What’s more,the RTs of the TT and TO compounds were very close across all priming conditions in the two masked experiments,implying a similar mental processing pattern.2.In regard to prime type in compound processing,both of the experiments showed main effects for prime types.In experiment 1,significant constituent priming effect was found in all the four categories of compounds,but the priming effect size was not equivalent and compounds with a transparent first constituent showed stable priming effects while those with an opaque initial constituent displayed elusive priming effects.In experiment 2,even though the main effect of prime types appeared in the statistical analysis of RTs,significant semantic priming effect was found only in the TO compounds by participant analysis,while other groups showed no such priming effect.Besides,the TT compounds surprisingly yielded the highest error rate and the longest response times while the OO compounds the highest response accuracy rate and the shortest response times.3.No interaction between the compound type and prime type emerged throughout the response accuracy or response time analysis in either of the two experiments.As the results of the two experiments revealed,compound type and prime type played a significant role in modulating the compound processing in Chinese EFL learners.The semantic transparency of the constituent morphemes was a crucial factor,especially the semantic transparency of the first constituent.The observed constituent priming effect validated that morphological decomposition occurs in the early stage of compound processing,but the unbalanced priming effects size among the four categories of compounds implied that morphological information may not be the sole source for such a processing advantage.Salient semantic priming effect only emerged in the TO compounds,showing that prior presentation of semantically related primes could not greatly facilitate the compound processing.Meanwhile,the unexpected accuracy and RTs distinction between the TT and OO compounds suggested that the semantic information of the transparent constituent morphemes might be passively activated and the semantic integration could cost extra processing burden,while the connection between the semantically associated primes and the meaning of the opaque constituents failed and the fully opaque compounds were more likely to be processed by whole word form.All these findings in this study provided supports for the hybrid model.