EPIC English and German original and interpreted subcorpora annotated for modal verb hedging

This dataset consists of sentences containing modal verbs from the German original, English original, German to English interpreted and English to German interpreted subcorpora of the EPIC corpus (Przybyl et al. 2022), which were then annotated as hedges or non-hedges.

As many sentences have several modal verbs, some sentences repeat. The subcorpora are thus available with and without repeats. The subcorpora with repeats include a field for the modal verb in question and are tagged according to if that modal verb acts as a hedge. The subcorpora without repeats are tagged as hedges if one or more of the modal verbs in the sentence act as hedges.