Abstract
This study conducts an ERP experiment to explore the online processing mechanism of Chinese xiehouyu, a subcategory of Chinese idiomatic expressions with a metaphorical two-part allegorical saying, regarded as a non-literal language construct. Using a 2 × 2 design, (high familiarity (HF)/low familiarity (LF)) × (literally-biasing context (LC)/metaphorically-biasing context (MC)), the researchers have obtained the following findings: (1) familiarity plays an important role in Chinese xiehouyu processing, i.e. the metaphorical meaning of a HF Chinese xiehouyu can be directly activated while that of a LF one has to be derived from its literal meaning first; (2) contextual information also weighs in the process, i.e. the metaphorical meaning of a Chinese xiehouyu can be promoted in MC condition but suppressed in LC condition; (3) the interactive effect of familiarity and contextual information can be explained by the career of metaphor hypothesis; and (4) the Standard Pragmatic Model (SPM) of non-literal languages can explain the processing of LF xiehouyu, and the Direct Access Model (DAM) may to some extent account for the mechanism of HF one but fails to explain the case of LF one, while the Graded Salience Hypothesis (GSH) can provide an acceptable explanation for the processing mechanism of Chinese xiehouyus of varied familiarity.