Structure-dependent sensitivity to morphosyntactic violations in L3 Norwegian: A self-paced reading investigation

Aims and Objectives: The study aims to investigate online sensitivity to morphosyntactic violations in Norwegian as a third language (L3).

Design/methodology/approach: L1 Polish–L2 English–L3 Norwegian trilinguals (N = 56) and a control group of native Norwegian speakers (N = 25) performed a self-paced reading task in Norwegian, with sentences including adjectival gender agreement and pronominal possessor agreement violations.

Data and analysis: Reading times were analysed on the critical (adjective in the predicative position and possessive pronoun, respectively) and spill-over (following preposition) words using mixed-effects regression modelling. Findings/conclusions: The results revealed significant between-group differences in reading times as a function of sentence grammaticality for gender agreement, but not for possessor agreement: while native Norwegian speakers showed sensitivity to both types of violations, L3 learners slowed down while reading incorrect sentences testing possessor agreement. This differential performance might be related to the greater salience of possessor agreement, or to transfer from English, as advocated by the Typological Primacy Model (Rothman, 2011) and the L2 Status Factor model (Bardel & Falk, 2007). Originality: In contrast to the growing body of literature on transfer effects in L3 acquisition, tested mainly with untimed judgement tasks, the present study offers insight into incremental processing of L3 morphosyntax, reflecting implicit knowledge.

Significance/implications: The results indicate that native-like processing in the L3 is achievable, but is influenced by such factors as structure salience and cross-linguistic similarity.