Multimodal-first or pantomime-first?

A persistent controversy in language evolution research has been whether language emerged in the gestural-visual or in the vocal-auditory modality. A “dialectic” solution to this age-old debate has now been gaining ground: language was fully multimodal from the start and remains so to this day. In this paper, we show this solution to be too simplistic and outline a more specific theoretical proposal, which we designate as pantomime-first. To decide between the multimodal-first and pantomime-first alternatives, we review several lines of interdisciplinary evidence and complement it with a cognitive-semiotic experiment. In the study, the participants saw – and then matched to hand-drawn images – recordings of short transitive events enacted by 4 actors in two conditions: visual (only body movement) and multimodal (body movement accompanied by nonlinguistic vocalization). Significantly, the matching accuracy was greater in the visual than the multimodal condition, though a follow-up experiment revealed that the emotional profiles of the events enacted in the multimodal condition could be reliably detected from the sound alone. We see these results as supporting the proposed pantomime-first scenario.