Drawing is recognized as a powerful tool to learn science. Although current research has enriched our understanding of the potential of learning through drawing, scarce attention has been given to the social-cognitive interactions that occur when students jointly create drawings to understand and explain phenomena in science. This article is based on the distributed and embodied cognition theories and it adopted the notion of we-space, defined as a complex social-cognitive space, dynamically established and managed during the ongoing interactions of the individuals, when they manipulate and exploit a shared space. The goal of the study was to explore the role that collaborative drawing plays in shaping the social-cognitive interaction among students. We examine this by a fine-grain multimodal analysis of a pair of middle school students, who jointly attempted to understand and explain a chemical phenomenon by creating drawings and thinking with them. Our findings suggest that collaborative drawing played a key role in (i) establishing a genuine shared-action space, a we-space, and that within this we-space it had two major functions: (ii) enabling collective thinking-in-action and (iii) simplifying communication. We argue that drawing, as a joint activity, has a potential for learning, not restricted to the cognitive process related to the activity of creating external visual representations on paper; instead, the benefits of drawing lie in action in space. Creating these representations is more than a process of externalization of thought: it is part of a process of collective thinking-in-action.