This paper outlines a theoretical framework for interaction with sound in music mixing. Using cognitive linguistic theory and studies exploring the spatiality of recorded music, it is argued that the logic of music mixing builds on three master metaphors—the signal flow metaphor, the sound stage metaphor and the container metaphor. I show how the metaphorical basis for interacting with sound in music mixing has changed with the development of recording technology, new aesthetic ideals and changing terminology. These changes are studied as expressions of underlying thought patterns that govern how music producers and engineers make sense of their actions. In conclusion, this leads to suggestions for a theoretical framework through which more intuitive music mixing interfaces may be developed in the future.