Edited by Megan M. Brown-Bousfield, Suzanne Flynn and Éva Fernández-Berkes
[Studies in Bilingualism 65] 2023
► pp. 205–235
Research on transfer in third language acquisition mostly centers around early stages of development, and little is known on processing after the initial state. This study explores processing of prepositions in L3 English to examine sensitivity to violations of grammaticality as a function of cross-linguistic similarity in the L3 development process. The study aims to investigate whether two groups of subjects (L1-Kurdish/L2-Turkish L3 learners of English and L1-Turkish L2 learners of English) are sensitive to violations of grammaticality in the processing of English prepositions at an intermediate level of proficiency and whether they can access implicit knowledge of prepositions during real-time processing in a self-paced reading task. Bayesian multilevel analyses were carried out to estimate potential sensitivity to violations. Results suggest that L3 learners are sensitive to violations of grammaticality in conditions in which ungrammatical items were presented as their L1 shares structural similarities with the adpositional system of English. Such sensitivity was not found in the L2 learner group whose L1 does not have any structural overlaps with the adpositional system of English. For L3 learners, reading times in the critical region were slower when the sentence cued a violation rather than when it did not. Slower reading times to violations spilled over onto the post-critical region. In addition to online sensitivity, post-sentence grammaticality judgements were also indicative of sensitivity to grammaticality violations by L3 learners. Overall results suggest that performance differences between the two groups can be explained by structural similarities and typological proximity between the L3 and the L1 of the trilingual group. Only L3 participants whose L1 has prepositions as part of its adpositional system seemed to be sensitive to the grammaticality violations online, and thus, could be argued to benefit from the facilitation of cross-linguistic similarity.