| Our study is focused on cross-linguistic influences in Spanish L3 acquisition by learners with English L1 and French L2. We based our observations in a corpus of written compositions from which we extracted the errors and classified them according to the possible influence from French and English that might have caused them. In addition, we did an error analysis database with the lexical, morphological and structural features that these errors exhibited, which we crosstabulated with information obtained in a linguistic profile that our participants responded.Our study incorporates the interpretation of the mental representation of the language mode of trilinguals as proposed by Grosjean (2001 2004), applied to third language learners. We offer an extended perspective for the application of the model in which the activation of previously acquired languages is independent of external factors such as the perceived linguistic background of the receptor of the message, as originally proposed by Grosjean.In sum, our holistic approach integrates qualitative and quantitative elements of the error analysis and participant profile data. We discuss how this approach can shed light on the dynamics that lie behind the production of crosslinguistic influences in the written production of third language learners.Keywords. Third language acquisition (L3 acquisition), Multilingualism, Crosslinguistic influences, Language mode, Transfer, Error analysis, Multiple language acquisitionThis mixed approach of the analysis of this written corpus combined with inferential statistics of the qualitative descriptions of error analysis data and the linguistic profiles information has allowed us to support the argument that L3 acquisition is not an identical process to L2 acquisition (Hall, 2001 Leung, 2005). The study also provides quantitative evidence to confirm the notion that the principles of recency, typological-lexical similarity, and proficiency are activated in Spanish L3 learners (Cenoz, 2001 Dewaele, 1998 Hammarberg, 2001). In addition, we have shown that the social and academic contexts of the L3 learners are related factors to the production of CLIs in the written performance of Spanish L3 individuals. |
