In this paper, I challenge the idea that music is language-like, in the sense it has a semantic-like dimension, as apparently implied in Peter Kivy's view on the relationship between music and language. Kivy suggests that music is semantic-like because it expresses something at the level of meaning which appeals to "musical expressivity". Musical expressivity is captured by the emotive properties constituting the musical content and recognized by a competent listener. I discuss two positions on musical expressivity, cognitivism and emotivism, which characterize the two aspects of musical expressivity differently - the emotional experience of the listener, and the musical objects and their features - the connection between them, and how they shape musical content. I conclude that since none of them provides an exhaustive explanation of musical expressivity, we should abandon the idea that music is semantic-like and, a fortiori, that music is language-like, at least within a framework which considers the semantic dimension of music to be related to emotive properties and musical expressivity.