Abstract This paper presents a model of size and shape depiction in the manual modality, based on a micro-level analysis of the lexicon of Adamorobe Sign Language (Ghana). The model combines strong elements of previous models, such as a dyadic labeling system separating form and meaning as suggested in