Our commentary addresses how two neurodevelopmental disorders, Williams syndrome and autism spectrum disorder, provide novel insights into the credible signaling and music and social bonding hypotheses presented in the two target articles. We suggest that these neurodevelopmental disorders, characterized by atypical social communication, allow us to test hypotheses about music, social bonding, and their underlying neurobiology.

The hypotheses in both target articles rely implicitly on much the same logic. For a “social-bonding” device to make sense, there must be an underlying reason why an otherwise-arbitrary behaviour sustains alliances – namely, credible signals of one's value to partners. To illustrate our points, we draw on the parallels with supposed bonding behaviours in nonhuman animals.

Mehr et al. seek to explain music's evolution in terms of a unitary proper function – signalling cooperative intent – which they cash out in two guises, coalition signalling and (allo)parental attention signalling. Although we recognize the role signalling almost certainly played in the evolution of music, we reject “ultimate” causal explanations which focus on a unidirectional, narrow range of causal factors.

The view of music as a byproduct of other cognitive functions has been deemed incomplete or incorrect. Revisiting the six lines of evidence that support this conclusion, it is argued that it is unclear how the hypothesis that music has its origins in (musi)language is discarded. Two additional promising research lines that can support or discard the byproduct hypothesis are presented.

I propose an adjunct to the two models presented in the target articles, a function of music that is ubiquitous and would have solved a clear adaptive problem, that of transmitting important survival information among pre-literate humans. This class of knowledge songs uniquely preserved cultural, botanical, medical, safety, and practical information that increased the adaptive fitness of societies.

Discerning adaptations from by-products is a defining feature of evolutionary science. Mehr, Krasnow, Bryant, and Hagen posit that music is an adaptation that evolved to function as a credible signal. We counter this claim, as we are not convinced they have dispelled the possibility that music is an elaboration of extant features of language.

The argument by Mehr et al. that music emerged and evolved culturally as a credible signal is convincing, but it lacks one essential ingredient: a model of signaling behavior that supports the main hypothesis theoretically and empirically. We argue that signaling games can help us explain how musical structures emerge as population-level phenomena, through sender–receiver signaling interactions.

We suggest that the accounts offered by the target articles could be strengthened by acknowledging the role of group selection and cultural niche construction in shaping the evolutionary trajectory of human music. We argue that group level traits and highly variable cultural niches can explain the diversity of human song, but the target articles' accounts are insufficient to explain such diversity.

This article is extraordinarily rigorous and rich, although there are reasons to be skeptical of its theory that music originated to signal group quality and infant solicitude. These include the lack of any signature of the centrality of these functions in the distribution or experience of music; of a role for the pleasure taken in music; and of its connections with language.

The two target articles agree that processes of cultural evolution generate richness and diversity in music, but neither address this question in a focused way. We sketch one way to proceed – and hence suggest how the target articles differ not only in empirical claims, but also in their tacit, prior assumptions about the relationship between cognition and culture.

Each target article contributes important proto-musical building blocks that constrain music as-we-know-it. However, neither the credible signaling nor social bonding accounts elucidate the central mystery of why music sounds the way it does. Getting there requires working out how proto-musical building blocks combine and interact to create the complex, rich, and affecting music humans create and enjoy.

Here, we compare birdsong and human musicality using insights from songbird neuroethology and evolution. For example, neural recordings during songbird duetting and other coordinated vocal behaviors could inform mechanistic hypotheses regarding human brain function during music-making. Furthermore, considering songbird evolution as a model system suggests that selection favoring certain culturally transmitted behaviors can indirectly select for associated underlying neural functions.

It is premature to conclude that music is an adaptation. Given the danger of overextending the adaptationist mode of explanation, the default position should be the byproduct hypothesis, and it should take very strong evidence to drag us into the adaptationist camp. As yet, the evidence isn't strong enough – and the proposed adaptationist explanations have a number of unresolved difficulties.

Human language and human music are both unique communication systems that evolved in the human lineage. Here, I propose that they share the same root, they evolved from an ancestral communication system yet to be described in detail. I suggest that pre-hunt charade was this shared root, which helped organize and coordinate the hunt of early hominins.

In their article, Mehr et al. conclude that the design features of music are consistent with adaptations for credible signaling. Although appealing to design may seem like a plausible basis for identifying adaptations, probing adaptive theories of music must be done at the genomic level and will require a functional understanding of the genomic, phenotypic, and fitness properties of music.

The evolutionary origins of complex capacities such as musicality are not simple, and likely involved many interacting steps of musicality-specific adaptations, exaptations, and cultural creation. A full account of the origins of musicality needs to consider the role of ancient adaptations such as credible singing, auditory scene analysis, and prediction-reward circuits in constraining the emergence of musicality.

I challenge Mehr et al.'s contention that ancestral mothers were reluctant to provide all the attention demanded by their infants. The societies in which music emerged likely involved foraging mothers who engaged in extensive infant carrying, feeding, and soothing. Accordingly, their singing was multimodal, its rhythms aligned with maternal movements, with arousal regulatory consequences for singers and listeners.

The credible signaling theory underexplains the evolutionary added value of less-credible affective musical signals compared to vocal signals. The theory might be extended to account for the motivation for, and consequences of, culturally decontextualizing a biologically contextualized signal. Musical signals are twofold, communicating “emotional fiction” alongside biological meaning, and could have filled an adaptive need for affect induction during storytelling.

Both Mehr et al.'s credible signaling hypothesis and Savage et al.'s music and social bonding hypothesis emphasize the role of multilevel social structures in the evolution of music. Although empirical evidence preferentially supports the social bonding hypothesis, rhythmic music may enable bonding in a way uniquely fitted to the normative and language-based character of multilevel human societies.

Mehr et al. propose a theory of the evolution music that can potentially account for most animal vocalizations as precursors to human music. Therein lies its appeal but also its Achilles' heel, for the wider the range of animal vocalizations treated as premusical expressions, the wider the gap to human music. Here, I offer a few critical observations and constructive suggestions that I hope will help the authors strengthen their case.