We could have loved and lost, or we never could have love at all: Syntactic Misanalysis in L2 Sentence Processing

Abstract

This study investigated sentence-processing strategies adopted by advanced nonnative speakers (NNSs) and native speakers (NSs) of English in the context of an English structure with which NNSs reportedly have an acquisition difficulty (e.g., Swan & Smith, 2001) - namely, modal perfect (MP). Participants read MP sentences such as 'He could have worked at the shoe factory' and closely related analogous sentences (e.g., 'He could have work at the shoe factory'), and reading times and errors were measured in an online grammaticality-judgment task. It was hypothesized that NSs would have a processing preference for MP sentences compared to the analogues, reflecting the primacy of syntactic information in NS processing and a preference for late closure, whereas NNSs would show no such preference because they rely less on syntactic information when processing sentences. The results revealed, however, that both NSs and NNSs read MP sentences more quickly and with fewer errors than the closely related analogues, consistent with a processing preference for MP sentences. Both groups were also influenced by word-category frequency information, which moderated, but did not fundamentally alter, their syntactic preference for MP. The significance of these findings is discussed in terms of models of second-language sentence processing and NNSs' reported MP acquisition difficulty.