Processing and production of unique and non-unique-to-L2 syntactic structures: The case of English articles and tense-aspect

The L2 acquisition of English articles and tense-aspect (TA) have been popular research areas over the last two decades. Compared with the numerous applications of metalinguistic knowledge and oral production tasks, the use of online (real time) processing methods to investigate these morphosyntactic structures has been far less common. In perhaps the only eye-tracking study on L2 English article processing, Trenkic et al. (2014) showed that L1 Mandarin/L2 English learners are able to use articles in real time to resolve referent ambiguity in a similar manner to L1 English speakers. In one of the rare self-paced reading (SPR) studies on L2 English TA processing, Roberts and Liszka (2013) found that while L1 English and L1 French/L2 English speakers were sensitive to English present perfect violations, L1 German/L2 English speakers, whose first language grammaticalises tense but not aspect, were not. However, beyond these important findings, our understanding of the L2 online processing of these morphosyntactic structures remains limited.
To address these gaps, the present thesis tested 24 L1 Mandarin/L2 English, 22 L1 Croatian/L2 English and 24 L1 English participants on an SPR task and in oral production. The SPR task used novel article stimuli and TA items adapted from Roberts and Liszka (2013) to test (implicit) sensitivity to violations, while a grammaticality judgement task (GJT) on the same stimuli was used to ascertain participants’ explicit knowledge. The comprehension data were triangulated with oral productions of English articles and tense elicited via an animated film retelling task.
A linear mixed-effects model analysis revealed that the participants’ performance on both the SPR and oral production tasks was highly influenced by their L1. The findings lend support to the morphological congruency hypothesis (Jiang, Novokshanova, Masuda, & Wang, 2011) which posits that late L2 learners cannot fully acquire morphosyntactic features that are incongruent (realised differently or absent in the L1 and L2), which suggest several implications for further research.