Contextual Effects On The Processing Mechanism Of Chinese3VO Metaphor:an ERP Study

As a common linguistic phenomenon, metaphor has always been the research focus of linguists. With the rapid development of cognitive science and scientific technology, techniques such as fMRI, ERP and eye-tracking, etc. are adopted to investigate the cognitive mechanism of metaphor. Researchers abroad have achieved fruitful results in the field of metaphor comprehension. There are three models widely accepted about the processing mechanism of metaphor, including the standard pragmatic model, the direct access model and the graded salience hypothesis. The first two models are simply and roughly concerned with the role of context ignoring the effect of salience, while the third one puts much stress on the effect of salience underestimating the important role of context. Compared with the fruitful achievements abroad, domestic scholars have carried out few studies on Chinese metaphor by employing scientific techniques mentioned above. Besides, most of the empirical studies have focused on the "A is B" metaphorical sentences and overlooked metaphors with other structures.In order to fill up the gap of researches home and abroad, the present study selected Chinese three-character verb-object (hereafter referred to as3VO) metaphors as materials by employing the ERP technique and Prime-Target-Probe paradigm. Both context and salience were taken into consideration. Manipulating salience of all the Chinese3VO metaphors, the present study aims to explore the contextual effects on the processing mechanism of metaphor and examine the relevant theories of the graded salience hypothesis. By means of a questionnaire survey, Chinese3VO metaphors whose metaphorical meaning is salient were chosen as Target. Then metaphorical, literal and irrelevant contexts were created respectively for Target and the three types of context were referred as Prime. Moreover, the author set two-character phrases as Probe which represented the metaphorical meaning of Target. Another questionnaire survey using a5-point scale was conducted to select Prime which were supposed to highly activate the corresponding meaning and to select Probe which was highly semantically related to Target. At last, the ERP experiment was carried out recording the behavioral data and ERP data. After analyzing all the collected data, the major findings were obtained as follows: 1. The behavioral results show that participants took longer time and made more errors to do the judgmental task when Target was primed by the literal context (LC). ERP results for Target exhibit no difference in the N400amplitudes under different types of context. However, ERP results for Probe reveal a significant main effect of context type. LC elicited a larger N400amplitude significantly different from metaphorical context (MC). Combining behavioral results with ERP results, a conclusion is drawn that context exerts a profound effect on the processing mechanism of Chinese3VO metaphor. Although the metaphorical meaning is salient, LC can activate the non-salient but congruous literal meaning first. The results indicate that the processing mechanism of Chinese3VO metaphor is not consistent with the graded salience hypothesis but with the direct access model when the appropriate context is provided.2. As for the the hemispheric involvement in language processing, ERP results for Target show a whole-hemispheric involvement in processing metaphor. Both LH and RH were engaged in Chinese3VO metaphor comprehension. ERP results for Probe show that in the right hemisphere (RH), LC evoked a significantly larger N400amplitude than MC while no such significant difference was found in the left hemisphere (LH). LC activated the corresponding literal meaning which made it difficult to relate Probe to Target. This implies RH is not sensitive to metaphoricity or literality but to task difficulty, semantic distance, novelty, etc. The more difficult the task is, the more RH is activated.3. An unexpected new ERP component referred to as P2was found in the present study. ERP results for Probe yield a significant difference in the P2amplitude for context type. When Target was primed by LC, the corresponding literal meaning of Target was derived which lengthened the semantic distance between Probe and Target and in turn made it more difficult to the semantic judgment. On the contrary, when Target was primed by MC, the semantic distance was shortened and the judgmental task became easier. The more difficult the task is, the larger the P2amplitude becomes. It is inferred that P2may be another index to investigate semantic processing.