The semantic content of gestures varies with information status, definiteness and clause structure

When speakers mention referents for the first time in discourse, they can also accompany them with gestures. This study explores whether the semantic content expressed by such gestures – whether gestures focus on the entity itself or on an action that the referent is involved in – is related to discourse properties. We consider the potential effects of a referent’s information status (brand-new/inferable), definiteness of the referring expression (indefinite/definite), and the clause structure in which it occurs (more/less specialized for referent introduction). Results suggest that ‘entity’ gestures specifically accompany brand-new referents expressed by indefinite nominals (e.g., indicating the shape of ‘a broom’), and in clause structures specialized for the introduction of referents (e.g., ‘there was a broom’). In contrast, ‘action’ gestures are more likely to occur with inferable referents expressed by definite nominals (e.g., pretending to hold the broom), and in less specialized clause structures focusing on events (e.g., ‘a spoon came flying in). The study highlights the importance of considering variations on the word, clause and discourse levels for understanding the semantic content of gestures.