The Ethics of Machine Translation Post-editing in the Translation Ecosystem

Abstract

The metaphor of the translation ecosystem originates from situational models of translation that conceptualise the translation process as a complex system. This includes not only the translator, but also other people—cooperation partners such as clients, project managers, proof-readers or co-translators—their specific social and physical environments as well as their cultural artefacts (Risku, Translationsmanagement. Interkulturelle fachkommunikation im kommunikationszeitalter. Narr, Tübingen, p 19, 2004). These artefacts, understood as objects made or used by humans for a particular purpose, have a high relevance for the translation process and for the translator’s cognition. The artefact group of translation technology includes, among others, tools for terminology and project management, translation memory (TM) systems, alignment software and machine translation (MT) systems (Krüger, Lebende Sprachen 61(2):297–332, 2016a). From the perspective of ecosystemic theories of translation, we are able to include situational factors which are external and internal to the translator and provide a holistic means for the analysis of translation performance. In this respect, the ethics of machine translation post-editing (MTPE) poses a question of central importance, a question that can be addressed from the stance of the ecosystem metaphor.

MTPE as an object of study is directly linked to the different developments in MT over time. During the first years of MT, it was largely empirical and focused on MT usability and comprehensibility, with a view to further developing the technology. Eventually, when MT reached a maturity, research interests concentrated on the practicalities of MTPE, with case studies and best practice examples (Garcia, Anglo Saxonica 3(3):291–308, 2012). With the latest developments in neural MT, MTPE is in a “state of terminological flux” (Vieira, The Routledge handbook of translation and technology. Routledge, London, p 320, 2019), comprising different, yet complementary, tasks and procedures: as a separate service with its own international standard, a dynamic activity that goes beyond the static cleaning of MT outputs, and a task associated by default with lower quality expectations. The instability of MTPE as a concept leads to the discussion of human agency in the MTPE process, and the exploration of the extent to which translators are able to intervene in the use of MT in MTPE. Furthermore, the analysis of the different degrees of human control triggers diverse issues in the ethics of MTPE. This chapter explores such issues in the light of the translation ecosystem, analysing three specific ethical dilemmas: (a) Dilemma #1: the post-editor’s status; (b) Dilemma #2: the post-editor’s commitment to quality; and (c) Dilemma #3: digital ethics and the post-editor’s responsibility. Rather than offering a set of closed conclusions, the chapter should be read as an invitation to the reader to think about key ethical elements and the way MTPE is affecting the translator’s work.