An ERP Study On The Way Construction

Using event-related potentials (ERPs), the present study attempted to investigate the processing of way constructions with different levels of acceptability both for English native speakers (L1) and advanced Chinese English learners (L2 advanced). Way construction is represented as the common structure "subject verb possessive way oblique" in which the verb stipulates the means or manner of the motion by subject and possessive way takes the object position. According to Goldberg (1995), way construction can be divided into three types:"means type", "manner type" and "vanilla type". Among these three types of way construction, "means type" implying the creation of the way, such as" Tom pushed his way to house", is most acceptable; "manner type" describing the concomitant actions or states, such as "The boy whistled his way to school", is less acceptable; and "vanilla type" expressing the pure motion such as "The boy went his way to bedroom", is unacceptable. Based on the above classification, this paper distinguished three levels of acceptability of way constructions:acceptable "means type", marginal "manner type" and unacceptable "vanilla type". ERPs were measured at the critical word "way" across all three conditions. In this experiment, we found that ERPs responses to way were significantly more positive in unacceptable "vanilla type" condition for English native speakers. This pattern appeared on both P3a and P600 (P3b). However, no ERPs component response was evoked in both marginal "manner type" and unacceptable "vanilla type" conditions for L2 advanced. Based on the results, P3a might be associated with the top-down attention-driven process, since "vanilla type" is more effective than "means type" in engaging participants' attention; and p600 (P3b) might be related to the bottom-up memory-driven process, since incoming influent, odd "vanilla type" was mismatched with the "orthodox", fluent "means type" stored in native speakers' memories. The P3a and P600 (P3b) pattern suggests the process of language follows general human cognitions such as perception, attention and memory process. Moreover, the result also indicates that only English native speakers are sensitive to the distinction between acceptable and unacceptable types of way constructions. We suggest the reason why Chinese English learners fail to detect the difference between way constructions is that the input of way constructions is not frequent enough to form language chunks in their memories.The results of this experiment also had some implications for both construction theory and pedagogy:firstly, the teaching of collocation between verbs and their typical constructions should be placed front and center; secondly, the coercions of idiosyncratic constructions on verbs and the restrictions of specific verbs on constructions should also be taken into consideration in our teaching practices; thirdly, since the acquisition of constructions is input-driven and usage-based, we should increase the frequency of inputs of various kinds of constructions so that they can be entrenched as grammatical knowledge in the learners' minds.