This chapter pursues the notion that, quintessentially, music enables the prediction of the unpredictable. Our focus is on the perception of music using ideas from theoretical biology and neuroscience to explain the nature of musical stimuli and their perceptual synthesis. In brief, we will consider music as a perceptual construct that supports (unconscious) inference on the causal structure of auditory input, in the sense of Helmholtz (1860). We examine the motivation for this particular perspective on music and consider the neuronal architectures that underlie its perception. The basic premises and supposed neuronal implementation—in terms of embodied inference—are then illustrated using simulations of (bird) song generation and perception; with a special focus on reproducing perceptual and neurophysiological responses that are seen in empirical neuroscience studies.