Writing in a second Language with Machine translation (WiLMa)

Writing in a Second Language with Machine Translation (WiLMa) Margot Fonteyne1,2, Maribel Montero Perez2, Joke Daems1, and Lieve Macken1 1 LT3, Language and Translation Technology Team 2 MULTIPLES, Research Centre for Multilingual Practices and Language Learning in Society Ghent University, Belgium margot.fonteyne@ugent.be Abstract The WiLMa project aims to assess the ef- fects of using machine translation (MT) tools on the writing processes of second language (L2) learners of varying profi- ciency. Particular attention is given to in- dividual variation in learners’ tool use. 1 Introduction WiLMa (2021–2024) is a predoctoral research project funded by Ghent University’s Special Re- search Fund (Grant No. BOF.DOC.2021.0001.01). The objectives for this project are: • To compare the L2 MT-assisted writing pro- cess across proficiency levels with L2 writing processes not assisted by MT • To map the individual variation in MT con- sultation behaviour (i.e., how L2 learners use MT during writing), investigate its correlation with learners’ L2 proficiency level, and study its effects on the L2 writing product The learners studied in this project are Dutch (L1) learners of Swedish (L2). In this project de- scription, we report on the pilot study we carried out with these learners and present the next steps of this project. Over the past decade, using MT has become a widespread practice among L2 learners, with writ- ing tasks being one of the technology’s most popu- lar use cases (Jolley and Maimone, 2022). A num- ber of publications has already investigated the ef- fects of MT use on L2 writing using a product- oriented approach. These studies have shown that © 2022 The authors. This article is licensed under a Creative Commons 3.0 licence, no derivative works, attribution, CC- BY-ND. writing products for which MT use was allowed differ from products for which it was not allowed. However, our knowledge of the effects of MT use on the L2 writing process is still limited. By boosting learners’ linguistic skills, MT may help learners to handle the competing demands on the different writing subprocesses better. These effects might also be larger for learners with lower profi- ciency levels (R´ev´esz, 2021). So far, two studies have investigated whether there are any differences to be found between learners’ online writing be- haviours (speed fluency, pausing, reading, and re- vising) in MT and non-MT conditions (Garcia and Pena, 2011; Ra´ıdo and Torr´on, 2020). However, it is difficult to draw conclusions from these studies, as they were based on a very small sample size and cover only a limited range of proficiency levels. Moreover, despite consulting the tools being a major component of the MT-assisted writing pro- cess, few studies have investigated how learners use MT during writing. Cancino and Panes (2021) report for example that (untrained) learners look up 95 words per 100 words written. Fred- holm (2015) notes that, on average, 44% of learn- ers’ texts is MT. However, these studies do not tell us whether learners’ consultation behaviour when having access to MT is any different from when learners use …

Content truncated to 500 words