Renormalization-group approach to ordered phases in music

Abstract

The organization of disordered sounds into the ordered structures of music can be understood through an analogy to the emergent ordering of physical systems undergoing phase transitions. This work builds off a prior mean-field model for pitch in music [J. Berezovsky, Sci. Adv. 5, eaav8490 (2019)] by using renormalization-group techniques to study the effects of dimensionality and local correlations. We corroborate the results of the mean-field model by showing convergence of the phase diagram as lattice dimension is increased, while also uncovering new phases which the mean-field model does not reveal. We also compute the nearest-neighbor correlations and provide comparisons to the mean-field model, as well as historical tuning systems used by different groups of musicians. The new phases and resulting correlations revealed in this work suggest a number of possible avenues for further exploration, including generating new music using the pitch distributions suggested by the model.