Cross-linguistic annotation of modality: a data-driven hierarchical model

We present an annotation model of modal- ity which is (i) cross-linguistic, relying on a wide, strongly typologically motivated ap- proach, and (ii) hierarchical and layered, ac- counting for both factuality and speaker’s attitude, while modelling these two aspects through separate annotation schemes. Modal- ity is defined through cross-linguistic cate- gories, but the classification of actual lin- guistic expressions is language-specific. This makes our annotation model a powerful tool for investigating linguistic diversity in the field of modality on the basis of real language data, being thus also useful from the perspective of machine translation systems.