The definitions and theoretical cores of 'heterodox' economics remain subject to debate, as do those of 'mainstream' economics. The theoretical elements that could unambiguously differentiate both perspectives are therefore debated. Common features are difficult to highlight within 'heterodox' economics. Symmetrically, 'mainstream' economics is uneasy to describe via necessary and sufficient features, as numerous authors within its field claim to have enriched neoclassical economics, and to use concepts from other social sciences. On the other hand, 'heterodox' policies and their differences with mainstream ones (e.g., 'austerity') appear to be more identifiable. Despite these uncertainties, the paper shows that some core concepts and theoretical assumptions may represent a distinction between heterodox and mainstream economics. The contribution of the paper is to argue that many conventional distinctions may in fact not be unambiguous criteria of distinguishing the two epistemic frameworks (e.g., self-interested agents, the methodology of modelling, use of concepts of social sciences), and to suggest a few theoretical issues that could constitute lines of difference-notably related to the concepts of causality, time and society.