Hierarchical structures are rapidly and flexibly built up in the domains of human language and music. These domains require a tree-building capacity – “dendrophilia” – to dynamically infer hierarchical structures from sensory input, based on subunits stored in a lexicon. This dynamic process involves a crucial class of abstracta overlooked in the target article.

Savage et al. make a compelling case, Mehr et al. less so, for social bonding and credible signalling, respectively, as the main adaptive function of human musicality. We express general advocacy for the former thesis, highlighting: overlap between the two; direct versus derived biological functions, and aspects of music embedded in cultural evolution, for example, departures from tonality.