Constrained communication and learner translated language: A corpus-based pilot study of L1 and L2 student translations

Aguiar De Souza Penha Marion, Laura [UCL] Gilquin, Gaëtanelle [UCL] Lefer, Marie-Aude [UCL]

This presentation, situated at the interface of Corpus-Based Translation Studies (CBTS) and Learner Corpus Research (LCR), aims at comparing translated language produced by native speakers and foreign language learners of the target language (TL). As repeatedly shown in the CBTS literature (e.g. Laviosa 2002), translated language is characterized by a range of typical features that set it apart from non-translated language, such as simplification, increased explicitness, normalization/conventionalization and source language (SL) influence. As rightly pointed out by Lanstyák & Heltai (2012), however, these features have been attributed to other varieties of constrained communication as well, such as L2 writing, which suggests that they may be more general characteristics of language contact and constrained language production (cf. also Kruger & Van Rooy 2016, Kruger 2017). With this in mind, we address an issue that has not received much attention in CBTS to date, namely translation directionality, i.e. whether translation is done into the translator’s L1 or L2 (Campbell 1998, Beeby 2009, Pokorn 2010). We hypothesize that L2 translation, being both translated and learner language, displays more visible features of constrainedness than L1 translation. In order to investigate this empirically, we conduct a corpus-based pilot study devoted to the impact of directionality on the linguistic traits of student translations, focusing on simplification, explicitation and SL interference. The corpus data consist of two French source texts (300-word newspaper articles) and their translations into English by 13 university students majoring in modern languages or translation (6 native speakers of English and 7 French-speaking learners of English). The analysis relies on the automatic extraction of simplification and explicitation indicators traditionally studied in CBTS (e.g. lexical variety, core vocabulary coverage, mean sentence length for simplification; see Laviosa 1998, Bernardini et al. 2016). The automatic stage is complemented with a thorough manual annotation of simplification and explicitation shifts as well as traces of SL interference, as, arguably, these can take on different forms, many of which cannot be fully captured by automatic analyses. The differences between L1 and L2 translation are further analyzed in the light of the keylogging data and screen recordings collected while the participants were translating the two source texts, so as to illustrate the added value of methodological plurality for learner translation research.