Abstract The topic addressed here is the role of emotions in music creation. This work presents research that is expected to become a precursor for the development of a software that will have the capacity to generate musical compositions at a human aesthetic level, and will serve as an intelligent creative assistant to a composer. The goal in this study is to develop a general cognitive model of music creation based on a semantic cognitive map of musical sounds, such as chords and their combinations. To achieve this goal, the basic theory of music needs to be analyzed first, and models of emotional perception of music need to be connected to it. Then, an empirical approach is undertaken to construct a semantic map, containing combinations of chords, allocated based on their emotional assessments. This semantic map will be further used as the central element in the creative assistant. Several options were explored in this study as a means of construction of the semantic maps. Preliminary results reported here provide an insight into an understanding of the principles of the future composer assistant. This emergent technology is scalable and potentially expandable to other domains of digital arts, including fine art and virtual dance.