Native and non-native processing of English wh- questions: Parsing strategies and plausibility constraints

Extract

The two experiments reported here investigated the processing of English wh- questions by native speakers of English and advanced Chinese, German, and Korean learners of English as a second language. Performance was evaluated in relation to parsing strategies and sensitivity to plausibility constraints. In an on-line plausibility judgment task, both native and non-native speakers behaved in similar ways. All groups postulated a gap at the first position consistent with the grammar, as predicted by the filler-driven strategy and as shown by garden path or filled-gap effects that were induced when the hypothesized gap location turned out to be incorrect. In addition, all subjects interpreted the plausibility of the filler-gap dependency, as shown by a reduction in the garden path effect when the initial analysis was implausible. However, the native speakers' reading profiles showed evidence of a more immediate effect of plausibility than those of the non-native speakers, suggesting that they initiated reanalysis earlier when the first analysis was implausible. Experiment 2 showed that the non-native speakers had difficulty canceling a plausible gap hypothesis even in an off-line (pencil and paper) task, whereas for the native speakers there was no evidence that the sentences caused difficulty in this situation. The results suggest that native and non-native speakers employ similar strategies in immediate on-line processing and hence are garden-pathed in similar ways, but they differ in their ability to recover from misanalysis.