How do ideas come into being? Our contribution takes its starting point in an observation we made in empirical data from a prior study. The data center around an instant of an academic writer’s thinking during the revision of a scientific paper. Through a detailed discourse-oriented micro-analysis, we zoom in on the writer’s thinking activity and uncover the genesis of a complex idea through a sequence of interrelated moments. These moments feature different degrees of “crystallization” of the idea; from gestures, a sketch, a short written note, oral explanations to a final spelled-out written argument. For this contribution, we re-analyze the material, asking how the idea gets formed during the thinking process and how it reaches a tangible form, which is understandable both for the thinker and for other persons. We root our analysis in a notion of language as social, embodied, and dialogical activity, drawing on concepts from Humboldt, Jakubinskij, and Vygotsky. We focus our analysis on three conceptual nodes. The first node is the ebbing and advancing of language in idea formation – observable as a trajectory through linguistically more condensed or more expanded utterance forms. The second node is the degree of objectification that the idea reaches when it is performed differently in a variety of addressivity constellations, i.e., whether and how it becomes understandable to the thinker and to others in the social sphere. Finally, the third node is the saturation of the idea through what we call intrapersonal intertextuality, i.e., its complex and dialogically related re-articulations in a sequence of formative moments. With these considerations, we articulate a clear consequence for theorizing thinking. We hold that thinking is social, embodied, and dialogically organized because it is entangled with language. Ideas come into being and become understandable and communicable to other persons only by and within their different, yet, intertextually related formations.