We make a case for using gestures and actions to understand and convey spatial and dynamic properties of systems. Problems in learning elementary astronomy are analysed in the context of demands of spatial thinking, in a system which is not amenable to direct perception, namely, the sun–earth–moon (SEM) system. We describe a pedagogy which uses gestures (most often in combination with concrete models and diagrams) to facilitate the visualisation and simulation required in elementary astronomy. These gestures are presented in terms of their purpose in pedagogy: to internalise a natural phenomenon, or an astronomical model, or general properties of space. In terms of design these pedagogical gestures mediate between concrete models of the SEM system and related spatial configurations on the one hand, and their corresponding abstract diagrammatic representations on the other: called here the model–gesture–diagram pedagogical link. Next we present some video data on students’ gestures observed during collaborative problem‐solving which took place in the course of our pedagogic intervention. Implications of these results are drawn for embodiment and multimodality of thought.