In response to theories of learning that attempt to get rid of contents, two arguments are provided which suggest that it is theoretically disputable that knowledge develops along the same kind of process for biology and history, physics and mathematics, or the geometry of the triangle, and the geometry of space. The first is that empirical studies show that the biggest difficulties met by students depend heavily on the contents of situations to be mastered. The second is that the search for general theories misses a very important epistemological point, namely, that concepts and competencies are solutions to specific problems. Two conceptual fields (additive and multiplicative structures) and the nature of a content-oriented approach concerning natural language are discussed and illustrated, which leads to a comprehensive definition of the nature of a concept. It is a triplet of three sets: the set of situations that make the concept meaningful; the set of invariants (theorems-in-action) that characterize the variety of competencies of students (these invariants are properties of the concept); and the set of symbolic representations that can be used to represent these properties and the situations. The importance of these three sets, the nature of the conceptual field, and the importance of a careful and profound analysis of the contents of knowledge are addressed. (JN)