Non-Binary genders in (machine) translation

Machine translation (MT) is known to suffer from gender bias, with non-binary individuals being often neglected in research on the topic. While gender-fair language (GFL) could foster their visibility and increase gender sensitivity, its implementation in language technology can be regarded as highly controversial, ideologically loaded, and technically difficult. This chapter provides a critical discussion of GFL, its impact on society as a form of language planning, and its implications for translation and MT. We discuss influencing factors on the success of introducing new gender forms and acknowledge that the strategy to be preferred is still an open debate in research and the non-binary community, with the current consensus that different contexts will require different strategies. This complicates finding technically feasible strategies for debiasing MT beyond the binary and raises issues for translators to apply GFL in their daily work. We critically discuss the possibilities of introducing GFL to MT and possible pitfalls of this endeavour for the technology and the community developing it. We argue that a change in the MT paradigm, which currently exclusively relies on the source text to produce exactly one target text, might be necessary to achieve gender-fair MT.