We're delighted to see two target articles about the evolution of music appear together in BBS. Each tackles music very differently: One developing an adaptationist scenario emphasizing music's role in credibly signalling cooperative intent (Mehr et al.), and the other focusing on gene-culture coevolution (Savage et al.).

Savage et al. approach music as a multifaceted, complex coevolutionary phenomena demanding a treatment which interweaves cognitive and morphological innovation with cultural and biological evolution, emphasizing feedback between these processes. Regardless of the success of Savage et al.'s specific proposal, we think something along those lines is required to explain music's evolution. In our view, music's coevolutionary origins render Mehr et al.'s approach to music via the adaptation/byproduct dichotomy misguided. Let us explain via analogy.

Distinctively hominin hand morphology, the hand's executive control, and lithic technologies all arose through processes of niche construction and coevolutionary dynamics and feedback. As Downes (Reference Downes, Ayala and Arp2010, p. 249) has put it: “The hand did not evolve in response to a particular environmental stimulus at any particular time. Rather, various selection pressures, including bipedalism, the occupation of niches with widely varying food resources, and our own niche construction, led to the musculature and bone structure that supports the range of activities for which human hands can be used.” In light of these coevolutionary dynamics, it is not productive to ask whether hominin hands and cognition were “adapted” for tool production and use, “exapted” for tool production and use, or whether tool production and use are “byproducts” of morphological and cognitive adaptations. We think the evolution of the dynamic, complex, coevolved mosaic “musicality” – the suite of morphological and cognitive capacities supporting the production and perception of musical displays – is roughly analogous. It doesn't follow from this that explanations appealing to adaptations and byproducts are never effective (say, the heart might be an adaptation for pumping blood, whereas blood's colour is a byproduct of its biochemistry). Rather, that vocabulary implies a causal simplicity which overlooks music's likely complex, niche-constructed, coevolutionary path (Killin, Reference Killin and Joyce2018a).

When dealing with a complex biocultural phenomenon, any account focused on a narrow range of unidirectional evolutionary pressures will be highly incomplete at best. But, that is exactly what Mehr et al. offer. They target psychological mechanisms underwriting signalling of cooperative intent, purportedly “adaptations for credible signalling, which give rise to a universal human psychology of music” (sect. 5.3, para. 2). Of course, it doesn't follow from music's multifaceted nature and complex evolution that signalling isn't an important part of its story (indeed, we suspect signalling roles were probably quite important and suggest that Savage et al. do more to acknowledge this). But, it does undermine appeals to a small number of narrow causal factors. Many other selective processes driving music's evolution are left by the wayside.

That said, we think Mehr et al.'s specific signalling hypotheses need further interrogation on their own terms. First, they see music as “a means for groups to credibly show off their qualities to other groups” (sect. 4.2.1, para. 10). As the authors point out, senders and receivers must be appropriately incentivized for signalling to evolve. The senders and receivers in this scenario are groups, with the incentive initially being territorial, “to deter intruders and avoid a fight” (sect. 4.2.1, para. 1), coopted later in human evolution for advertising/assessing alliance opportunities. But, are these sufficient, given the costs of music-acquisition? It is hard to say. This is complicated by the need for individual-level decomposition, as groups don't perceive music, individuals do, and they also pay its costs. Chimpanzees and bonobos can be vocal and rowdy but they do not sing or rhythmically entrain with conspecifics: Ancestral hominins would have needed to acquire these abilities even if they were rudimentary by modern human standards. This would have been cognitively demanding and imposed opportunity costs. We question whether the posited payoffs here could have provided a consistent, deep-time selective environment to explain coordinated music's origins, as Mehr et al. claim, or whether this is better seen as a form of stabilizing selection (which would also explain the ethnographic evidence discussed).

Consider synchrony. As the authors say, “a high level of synchronous coordination among signalers requires considerable effort to achieve” (sect. 4.2.1, para. 4). But, is it sufficiently well-correlated with a valuable trait (cooperative ability generally?) for an index signal mechanism to evolve? And, were such highly-synchronous displays really so ubiquitous as to be a valuable source of such information, incentivizing adaptive response? We find it equally plausible that (1) these hypothesized synchronous displays were actually quite loosely correlated with the actual fitness-relevant virtues supposedly being signalled for (as natural variation in musical aptitude partially suggests), and (2) that they instead worked (insofar as they did) by manipulating pre-existing responses to audible cues. Therefore, they may well be part of music's evolutionary story but the authors' case for ultimate causation here is weak. Indeed, unlike their example of bitter taste (for which their dichotomous ultimate/proximate causal framework is arguably adequate), coevolutionary feedback across evolution and development renders “ultimate” causation for music somewhat artificial.

Second, without a chronology, Mehr et al.'s hypothesis that infant-directed maternal contact calls evolved to signal (allo)parental attention is difficult to assess, partly because it requires the shaky assumption that our long-ancient ancestors' interbirth intervals resembled those of modern foragers. Whether we are talking about erectines or heidelbergensians or archaic sapiens matters here. For if interbirth intervals at the relevant time were longer – say, between those of modern foragers and chimpanzees (almost certainly true of erectines) – those ancient mothers would have had fewer fully-dependent young children at any given time, weakening selection for that signalling strategy.

Again, it doesn't follow that infant-directed vocalizations played no role in music's evolution. We think they would have. The point is that hypotheses focusing on sharply limited ranges of causal factors are bound to be inadequate for explaining complex traits (Currie, Reference Currie2014, Reference Currie2019) and musicality is a complex trait: We must instead point to the multifaceted and interwoven nature of the evolutionary dynamics behind music's evolution, as Savage et al. do (see also Killin, Reference Killin2016, Reference Killin2017, Reference Killin2018b).