This project proposes a reflection on interference in translations from a foreign language into the mother tongue, by means of an empirical study of comparable and parallel corpus together with cognitive data. The subjects will be Brazilian undergraduates in Spanish as a foreign language. The definition adopted is that of interference as the incorporation of elements from one language in a production in another language (Mackey, 1970 apud PRESAS, 2000, p.25). We also propose that when someone chooses to translate a frequent structure from the source language into a less usual structure of the target language, that could be seen as a kind of interference, especially when "naturalness" is considered to be a desirable translation feature. The research will be based on empirical data to verify, first, whether interference in translation really happens in the opposite direction of interference in foreign language learning, that is, if it occurs from L2 on L1, as Presas (2000) pointed out. This author also highlights that this interference direction is surprising and perhaps could not be completely understood by hypotheses offered to explain interference of L1 on L2, in L2 learning contexts. Secondly, we depart from the hypothesis that, due to contrastive peculiarities between Portuguese and Spanish, certain structures would favor the interference of L2 on L1 in this linguistic pair. If our data confirmed that, we will try to offer tentative explanations using cognitive process data.The entire corpus will be comprised of three tasks. The subjects are supposed to (1) write a recipe in Portuguese following an image sequence; (2) complete a recipe written in Portuguese with temporal an purpose subordinate clauses; (3) translate into Portuguese a recipe in Spanish. The third one will be the main corpus, in which we will mainly observe the translations proposed for: 1) temporal subordinate clauses with cuando and hasta; 2) purpose subordinate clauses with para. To characterize the use of certain syntactic constructions as interference, it will be necessary to check whether the structures found in the translations are not the most frequent in Portuguese. To do that, we will also analyse a corpus of recipes originally written in Brazilian Portuguese.That should allow to observe: a) which structures are more frequent for temporal and purpose subordinate clauses in Spanish and in Portuguese, in recipes; b) if the subjects use the same or different structures to write a recipe directly in Portuguese, to fill the blanks of a recipe in Portuguese, to translate a recipe into Portuguese; c) if there are differences between the structures in the subjects' productions and those found in recipes collected on the Internet, originally written in Portuguese. This corpus organization may allow to verify the hypothesis of a special probability of interference in the translation of such subordinated clauses in this text type.Finally, all tasks will be made using the program Translog (JAKOBSEN, 1999), which provides data to study translation process (eg. pauses measurements) and may be useful to formulate explanatory hypotheses for the findings.Our main theoretical framework will be the concepts of naturalness and of interference in L2 acquisition and in translation. We will also draw on a contrastive study of temporal and purpose subordinate clauses between Spanish and Portuguese. Methodologically, principles of Corpus Linguistics and techniques for translation process studies will be adopted.
News published in Agência FAPESP Newsletter about the scholarship: