Relative clause attachment preferences of late bilinguals

Aims and Objectives/Purpose/Research Questions: This study examines how late bilinguals parse temporarily and globally ambiguous relative clauses, and whether this preference is modulated by their reading habits. Research questions are as follows: (1) Do late bilinguals' attachment preferences differ as the task varies, for example, translating versus identifying the NP modified by the given RC? (2) Are attachment preferences influenced by locally or globally ambiguous subject relative clauses? (3) Do late bilinguals' reading habits have any influence on their attachment preferences in the two tasks?Design/Methodology/Approach: We used a self-paced reading task and translation task with a group of late bilinguals whose native language (L1) was Turkish and who learned English (L2) as a foreign language in formal settings (N = 43).Data and Analysis: We performed a repeated-measures analysis of variance (ANOVA) on the data.Findings/Conclusions: We found that task type influenced participants' attachment preferences. Greater use of the second noun phrase (NP2) as opposed to the first noun phrase (NP1) was found in the translation task, regardless of ambiguity type, which shows that proficient late bilinguals were able to perform like native speakers. In the question-answer task, however, the reverse pattern observed can be interpreted either as the L1 influence, or alternatively, the underuse of syntactic information in L2, supporting the Shallow Structure Hypothesis (SSH). We found no correlation between the attachment preferences and reading habits of the participants.Originality: This study was the first to examine parsing ambiguity in RCs through a translation task.Significance/Implications: The results may be taken into consideration in a psycholinguistic approach to translation studies. Further studies may investigate ambiguity parsing by using online tasks backed by the eye-tracking method.