The aim of this thesis is to argue in favour of the embodied music cognition paradigm (hereafter: EMC), as opposed to traditional (computational) theories of musical mind. The thesis consists of three chapters. The goal of the first chapter is to (1) examine computational (disembodied) music cognition, focusing on the main problem of this approach (namely: the symbol–grounding problem). In the second chapter, (2) I will present and discuss Marc Leman’s EMC, that may serve as a response to the problems of the computational view of the musical mind. Although this framework is interesting, it is unclear in several places. For that reason, I intend to enrich it with the references to the recent works on human mirror neuron system (hereafter: MNS) and enactivist views on music cognition. Given that, in the last chapter, I will (3) answer the question, whether mental representations are necessary in music cognition.