Roberts, L. 1 & Liszka, S. 2
1 University of York
2 University of Greenwich
Research suggests L1-L2 parsing differences may be attributable to learners' being less able to predict upcoming input based on previously processed material (Gruter, et al., 2014; Kaan, 2014). Previous research on L2 processing expectations has centred on morphosyntactic (determiner-noun/subject-verb) agreement rather than inter-clausal relations. Results are mixed (Teal & Slabakova, 2016), and cross-linguistic influences remain under-researched.
Using self-paced reading, we examined 100 L2 learners? processing of subject-object ambiguities [1] with subordinate verbs in the past simple or progressive. In L1 English, grammatical aspect influences readers' online expectations that a direct object will follow (Frazier et al. 2006).
1. While the man hunted/was hunting the rabbit escaped/it escaped.
Both RTs and off-line judgments showed a clear L1 influence. Although all were matched on English tense/aspect knowledge offline, the [+Aspect] learners (60 French/Spanish/Chinese) showed a processing asymmetry like the English controls: with strong expectations for a direct object, but only following the past simple where an immediate and lasting garden-path effect emerged. The [-Aspect] learners (40 German/Dutch) expected a direct object to follow in both conditions equally.
In sum, learners are not less able to anticipate upcoming structure in general, but rather their L2 parsing expectations are mediated by L1 (grammatical/semantic) knowledge.