DE PROFUNDIS: THE PROBLEMS OF THE DEEP ANNOTATON OF THE MULTIMODAL RUSSIAN CORPUS AND THE WAYS OF RESOLUTION


2019. № 3 (21), 319-325

 National Research University Higher School of Economics

Abstract:

The paper focuses on the manual gesture annotation in the Multimodal Russian Corpus (MURCO), which was started by E. A. Grishina and is being developed by the authors of this paper. The important aspect of the annotation process is the attempt to provide 

“the uniformity and commonality of the markup” [Grishina 2010] to the maximum degree possible. To do so, the annotator should carefully study the MURCO data which was marked by E.A. Grishina (since sadly it is no longer possible to ask the questions directly) and define the rules that govern the gesture annotation and that were probably implied by E. A. Grishina herself.

The paper describes three of such rules: 1) the choice between the gestures to open one’s eyes wide and to raise one’s brows (both meaning ‘fear’) — we state that the main factor here is the distinctness of gesture performance, 2) the choice between the meanings ‘confirmation’ and ‘emphasis’ of the gesture to close one’s eyes, and 3) the choice between the same two meanings of the gesture to nod. In both cases the meaning ‘confirmation’ is preferable if the utterance accompanied by the gesture is the reply to somebody’s remark. The other factor is the cohesion — if the utterance accompanied by the gesture conveys the same meaning as the previous utterance of the speaker, the meaning ‘confirmation’ should be preferred.