Utterance unit annotation for the Japanese Sign Language Dialogue Corpus Towards a method for detecting interactional boundaries in spontaneous sign language dialogue

This chapter defines 'utterance units' and describes their annotation in the Japanese Sign Language (JSL) dialogue corpus, first focusing on how human annotators - native signers of JSL - identify and annotate utterance units, before reporting on part of speech (POS) tagging for JSL and semi-automatic annotation of utterance units. The utterance unit is an original concept for segmenting and annotating movement features in sign language dialogue, based on signers' native sense. We postulate a fundamental interaction-specific unit for understanding interactional mechanisms (such as turn-taking) in sign language social interactions from the perspectives of conversation analysis and multimodal interaction studies. We explain differences between sentence and utterance units, the corpus construction and composition, and the annotation scheme, before analyzing how JSL native annotators annotated the units. Finally, we show the application potential of this research by presenting two case studies, the first exploring POS annotations, and the second a first attempt at automatic annotation using OpenPose software.