No abstract

Syntactic cues help individuals to assign thematic roles (agent/patient) during sentence processing. In the present study, we examined whether the use of syntactic cues changed in bilinguals depending on the task they performed. Spanish (L1)/English (L2) bilinguals read sentences in Spanish either to produce them in the same language or to translate them in English. Three syntactic cues were evaluated: animacy (Experiment 1), subject-verb agreement (Experiment 2), and word order (Experiment 3). In Experiments 1 and 2, word order was stronger than animacy and subject-verb agreement when bilinguals read for translation. Moreover, when the same syntactic cue favoured a different agent for each language (word order in Experiment 3), the bilinguals performed as native speakers of Spanish in the within-language task, whereas they performed as native speakers of English in the translation task. Together, this research indicates that translation changes the way bilinguals use syntactic cues in sentence processing. Public Significance StatementThis study suggests that the way bilinguals understand sentences depends on the goal of the reading. Spanish-English bilinguals behave like Spanish monolinguals when they read and produce sentences in Spanish whilst they behave like English speakers when reading the same sentences for translation in English. This finding has important practical implications for training in translation tasks.

The goal of our research was to explore the possible online co-activation of both the target language (TL) syntactic structure representation and TL attachment strategies in translation, and to look over a possible interaction between both syntactic properties. To this purpose, Spanish (L1) – English (L2) bilinguals were instructed to read complex noun phrases with an ambiguous relative clause in Spanish to either repeat them in Spanish or translate them into English. The final word of the sentences and the syntactic congruency between the source language (SL) and TL syntactic structure were manipulated. The results revealed co-activation of both TL syntactic properties: participants interpreted sentences more accordingly to the TL preferred strategy (low attachment) in the reading for translation task, read congruent sentences faster, and used the TL preferred interpretation strategy in the congruent condition of the sentences more. These results indicated TL activation at different syntactic levels during comprehension of the SL in translation.