When children encounter museum exhibits, they find rich opportunities for action, perception, learning, and other forms of cognition. Can we see systematic organization in the children’s behavior, and by extension, their cognition? Based on an account of cognition as embodied, situated, and culturally mediated, this research illustrates how some cognition can be directly observed, manifested through interactions among modalities, people and objects in a distributed cognitive system. This study uses micro- and macro-analyses of behavior recorded on video to discover organizing structure in children’s behavior and cognition, evidenced in allocation of visual and haptic attention, manipulation of objects, and use of written, spoken, and gestured language. At micro- and macro-scales, perceptual engagement with concrete objects precedes engagement with abstract concepts, as evidenced in motor behavior and content of speech and gesture. In this context, we describe learning as a process of adaptive coordination, rather than a product to be measured.