Margot Fonteyne

This paper discusses the selection of a machine translation system for use in asylum reception centres, to translate messages written in English, French, and Dutch into 14 languages, including low-resource ones like Pashto and Somali. Translation quality is assessed through automatic metrics and expert human evaluations to support effective multilingual communication.

Most existing empirical work on the effects of Machine Translation (MT) use on second language (L2) writing has concentrated on its impact on writing products, with much less research addressing its effects on L2 learners’ behaviours during writing. We therefore investigate whether the L2 writing process varies depending on whether learners are provided access to MT or to an online bilingual dictionary. Twenty-seven L1 Dutch learners of Swedish were assigned four writing tasks, with two tasks completed in each condition (MT or dictionary). While writing, participants’ keystrokes were registered. Descriptive measures were used to summarize and compare participants’ writing speed, tool engagement, and pausing behaviour across the two conditions. Results indicated that participants wrote more in less time, consulted the tool more frequently but for shorter periods of time, and paused longer between higher textual units when provided access to MT.

Due to the growing success of neural machine translation (NMT), many have started to question its applicability within the field of literary translation. In order to grasp the possibilities of NMT, we studied the output of the neural machine system of Google Translate (GNMT) and DeepL when applied to four classic novels translated from English into Dutch. The quality of the NMT systems is discussed by focusing on manual annotations, and we also employed various metrics in order to get an insight into lexical richness, local cohesion, syntactic, and stylistic difference. Firstly, we discovered that a large proportion of the translated sentences contained errors. We also observed a lower level of lexical richness and local cohesion in the NMTs compared to the human translations. In addition, NMTs are more likely to follow the syntactic structure of a source sentence, whereas human translations can differ. Lastly, the human translations deviate from the machine translations in style.