Abstract: Multilingual text versions pose challenges for digital scholarly editions (DSE), particularly in the encoding and representation of textual variation. Hannah Arendt for example authored most of her works in English and German, carefully rewriting and editing the texts herself. The resulting versions should not be considered as mere translations or revisions. Arendt used the plurality of languages as a creative tool. In comparing the versions, seeming textual shifts create meaningful tensions. For a DSE, this imposes the task of enabling readers of Arendt’s works to explore “these tensions as a significant and signifying Between.” (Wild, 2024, p. 29) The COMUTE project [1] was initiated to analyse such “tensions” in multilingual writing. Funded by the German Research Foundation, it aims to algorithmically identify variant passages [2] and, in collaboration with the Hannah Arendt Critical Edition [3] as one use case, explores how such findings can be evaluated, visualised and integrated into editorial workflows. In order to support data reuse, especially in DSE workflows, the collation results should be encoded according to the TEI guidelines. [4] While it is feasible to use TEI markup for critical apparatuses for simpler cases, this is not straightforward when it comes to more complex cases, as phenomena of multilingual textual variation manifest themselves not only at the character level, but also in linguistic nuances and in content and context. This poster presents a TEI-based approach for encoding multilingual and non-parallel textual variation and integrating it into DSEs. [1] https://www.comute-project.de/ [2] The COMUTE-Project is building on the existing monolingual collation tool LERA (Pöckelmann et al., 2022) and is extending it by adding a multilingual layer. [3] https://hannah-arendt-edition.net/home [4] As discussed for monolingual variation, for example, by Bleeker et al. (2018) and Cayless et al. (2019) and for translation scenarios in the TEI mailing list (Various Authors, 2013)