Some descriptions of learning represent the process as the development of organisations of elements. Various organisations have been proposed, for example, schemata and conceptual structures. Such representations assume that mental entities, such as concepts, are sufficiently stable and differentiated to be treated as units. We discuss these assumptions and propose a new term, assemblages, to refer to a person's activation of two or more conceptual resources in a context. Methodological challenges have resulted in a lack of research that examines how assemblages are formed. This study presents data from a microgenetic, multiple case study of four 16-17-year-old students. The participants were interviewed weekly, using various probes related to forces and motion over six months. We focus on two aspects of the assembly process in our analysis. First, we report data that indicate that participants perceived the units they assembled differently from expert conceptualisations and reflect on the stability of their assemblages. Second, we discuss how participants' expectations about the coherence of knowledge impact their assembly. We propose that future research investigates the stability and boundaries of conceptual resources and suggest teachers and researchers are cautious in assuming that data indicate a conceptual resource is stable or unitary.