Gestures are often categorized into types: iconics, metaphorics, and pantomimes (having representational relationships with spoken semantics), deictics (i.e., pointing), emblems (having their own conventional meaning), and beats (temporally coinciding with spoken content for emphasis). These originate from research often involving unnaturalistic paradigms where participants' gestures during responses (e.g., retelling a narrative) are recorded. Approaching these types implicitly requires a stance on why we gesture; a conscious aim to communicate or an unconscious effort to orchestrate speech. Focus on them has led to the understudying of the interactive role gestures can play, where intersubjective acknowledgment and information transfer are central. This paper has two main aims: to profile the interactive role of gesturing as a proportion of all gesturing and to investigate its relevance for why humans gesture. We report data from 48 28-min dyadic conversations with a naïve participant and a confederate, varying interlocutors' gesture visibility and conversation complexity. Our first, preregistered, analysis coded for the six traditional gesture categories, which resulted in ∼28% being uncodeable. Our second analysis asked whether these were interactive, which accounted for nearly 90% of uncoded gestures and a quarter of the entire database. Occluding gesture visibility significantly decreased the amount of interactive gestures participants made, resulting from a drop in interactive gestures made during simple conversations; complex topic interactive gesture frequency is stable between visually available and occluded conditions. Our data support both philosophies, but advocate for a subconscious account: that we gesture for the intrinsic motivations to express ourselves and to be understood.
Williamson et al. (2026) studied this question.