Metaphoric Competence And Processing Model Of Metaphor Under Different Second Language Proficiency:An ERP Study

The pervasiveness of metaphors in everyday language has made it drawn more intensive attention than other linguistic phenomena. Metaphor is not only a linguistic phenomenon, but the way of thinking. Western scholars have conducted a variety of experiments to examine native speakers’ metaphoric processing models. However, the study on English as foreign language(EFL) learners’ processing model of metaphor and their metaphoric competence is at the initial stage.The present study adopted event-related potentials(ERPs) technique to examine EFL learners’ metaphoric competence and processing model of metaphor. Specifically, 36 Southwest University graduates were divided into 2 groups: high language proficiency group with 18 English major graduates who have achieved Certificate of Test English Major Eight(5 male, 13 female) and low language proficiency group with 18 non-English major graduates who have achieved Certificate of College English Test Six(5 male, 13 female). Participants were required to perform a semantic judgement task which asked them to decide whether the randomly displayed sentences were meaningful or not. The research stimuli were 224 sentences of 4 types with the same target word: conventional metaphors, novel metaphors, anomalous sentences and literal sentences. The major findings of the present study were as follows:First, EFL learners’ metaphoric competence is significantly influenced by their language proficiency, that is, high language proficiency EFL learners’ accuracy rate was significantly higher than low language proficiency EFL learners’. However, even if EFL learners’ language proficiency is advanced, their metaphoric competence is not satisfied. Compared with native speakers’ accurate responses, we suggested that(1) though high language proficiency EFL learners’ accurate responses approximated to native speakers ’responses, while the differences were found between conventional metaphors and literal sentences. Therefore, we assumed that high language proficiency EFL learners’ metaphoric competence was at a moderate level.(2) Low language proficiency EFL learners’ accurate responses were much lower than the native speakers’ and especially the accuracy for novel metaphors. Consequently, we assumed that low language proficiency EFL learner’ metaphoric competence level was low.Second, the reaction time for the semantic judgement task was not significantly influenced by different language proficiency, but high language proficiency group learners generally spent less time processing each sentence types than the low language proficiency group. The reaction time for each sentence types in the high language proficiency group was in a graded manner: novel metaphors/anomalous sentences >conventional metaphors>literal sentences. In this case, we presumed that the high language proficiency EFL learners favored the structural mapping/ the career of metaphor model. The reaction time for conventional metaphors, novel metaphor and literal sentences was not varied in the low language proficiency group and the time for anomalous sentences was also equal. Therefore, low language proficiency EFL learners initially processed all types of linguistic expressions literally and they preferred to the three-stage or indirect access model.According to the previous studies and results of the present study, we observed the brain waveforms on the early N400(320-440ms) and the late N400(440-560ms). The brain waveforms were recorded at 9 electrode sites(F3, FZ, F4, C3, CZ, C4, P3, PZ, P4) and there was no interaction effects among language proficiency, sentence types and electrode sites in both time windows. However, in high language proficiency group, we found a main effect between sentence types in the early window 320-440ms. Novel metaphors elicited more negative waveforms than conventional metaphors and literal sentences. Surprisingly, the amplitudes of anomalous, conventional metaphors and literal sentences were similar. The reason for this is because N400 was a negative waveform elicited by incongruent meaning or violations of semantic expectancies and EFL learners’ foreign language sensitivity was lower than native speakers, the amplitude of conventional and anomalous metaphors was reduced. Besides, the behavior data also proved the differences between literal sentences and conventional metaphors. We suggested that high language proficiency EFL learners preferred to the structural mapping model rather than the graded salient model. No interaction effect or main effect was found in low language proficiency group, so we proposed that low language proficiency EFL learners favored the indirect access model or the three-stage model.Last but not the least, the present study did not find the relationship between EFL learners’ metaphoric competence and their processing model of metaphor. Though language proficiency significantly influenced EFL learners’ metaphoric competence, EFL learners’ processing model of metaphor did not affected by their language proficiency. Therefore, we suggested that EFL leaners’ metaphoric competence was not correlated with their processing model of metaphor at least from the evidence of learners’ language proficiency. On one hand, this result was influenced by the limitations of the present study; on the other hand, the relationship was affected by other factors, such as learners’ interlanguage. This will be the direction for our future research.