Recent research has revealed the existence of intuitive representations strongly rooted in diverse knowledge domains and the difficulties of modifying those representations through instruction by means of conceptual change processes (Carey, 1995; Gopnik & Meltzoff, 1997; Vosniadou, 1994). According to some interpretations, these representations are organized in the form of intuitive or implicit theories, among whose features would be representational consistency across different contexts or tasks. Other authors, however, have argued in favor of the partial or situated nature of these representations so that we cannot talk about authentic theories (Caravita & Hallden, 1994; diSessa, 1993, 1994). In this research, we analyzed the consistency of the representations on the nature of matter held by people with diverse levels of instruction from adolescents aged 12 to students with advanced studies in chemistry. Our objective was to analyze the consistency of these representations with respect to the core concepts of chemistry, the intrinsic movement of particles, and the discontinuous nature of matter, which would be constituted by particles separated by empty space. However, these two core concepts would be in contradiction to our embodied representation of matter based on its macroscopic or sensorial properties according to which matter is static and continuous. The results confirm a majority and consistent use of intuitive representations as against scientific ones, confirming that at least from the point of view of consistency, these representations constitute authentic implicit theories. Moreover, it is those representations that correspond to the macroscopic properties of matter that are used in the most consistent way, which also supports the hypothesis of the embodied nature of these theories. Although the corpuscular theory of matter, when used appropriately, provides more consistent representations, our data show a quite limited use of this theory, even among students with advanced studies in science, which once more shows the difficulties of achieving, through instruction, a change in these implicit, embodied theories. In the article, we discuss the implications of these data for science teaching.