The origins of music remain obscure. Here it is pointed out that the outer hair cells in the cochlea lie approximately in a regular, hexagonal array, so it appears significant that important musical ratios – in particular the semitone, octave, perfect fifth, and major third – appear in the relative distances between adjacent cells. The speculation made here is that if the intercell distances are inherently tuned in this way, then incoming sound could initiate reverberating activity between the cells, and a musical ratio could be signaled by simultaneous standing waves in one cell–cell distance and in another which adjoins it. Essentially, the spacings between the cells might correspond to the lengths of miniature, musically tuned cavities. This speculative model of cell–cell interaction can explain recent observations that the human cochlea spontaneously emits low-level sound at intervals close to a semitone, and that the hearing thresholds of some subjects exhibit a whole “keyboard” of semitone-like intervals. These recent findings are striking, and appear more than coincidence. They prompt the key question, why? A possible answer may lie, it is suggested, with the distinct 2-dimensional geometry of the outer hair cells in the plane of the basilar membrane, which commonly exhibits a 19° alignment. This angle corresponds to a relative distance of 1.06, which is close to a semitone. It is pointed out that the same geometry which generates a natural semitone also produces other musically significant ratios. Perhaps, then, music might be an innate property of the human auditory system – and hence that there might be a natural basis for preferred musical intervals. Natural theories have often been criticised, with some saying that music is a learned faculty and depends only on culture. However, evidence has accumulated that there do seem to be musical universals, and therefore that music might indeed have a natural basis, most commonly thought to be via some neural processing in the brain. The explanatory model put forward here as the basis for further exploration suggests that musical analysis might actually begin in the periphery – in the cochlea itself.