One of Mehr, Krasnow, Bryant, and Hagen's aims is to show that the hypothesis that music is a byproduct of other faculties fails in light of six lines of evidence. The evaluation of various adaptation hypotheses led to the conclusion that the byproduct view is incomplete or incorrect. These six arguments are revisited below from a biolinguistic point of view in order to evaluate their individual validity in relation to one premise: the possible origins of music in (musi)language.

The first argument of Mehr et al. is that complex, song-like vocalizations have evolved convergently across distant species, and they are socially learned, like music is. However, if these elaborate, song-like vocalizations feature semantically meaningful and functionally referential information (Arnold & Zuberbühler, Reference Arnold and Zuberbühler2012; Elie & Theunissen, Reference Elie and Theunissen2016), they are analogous to human language, not human music. If so, this first argument does not show why the hypothesis that music is the byproduct of language is incomplete or incorrect.

The second argument is that music is a human universal. This claim, although correct, does not really discard the byproduct hypothesis specifically in relation to language, because language is universal too. The spontaneous development of home-signs in deaf populations (Senghas, Kita, & Özyürek, Reference Senghas, Kita and Özyürek2004) and the ability to lexicalize even in cases of severe absence of language input (Gleitman & Newport, Reference Gleitman, Newport, Gleitman and Liberman1995) are unparalleled indicators of the universality of language. For this second argument to successfully discard the byproduct hypothesis, cases of societies that have music but not language should be presented.

The next argument put forth by Mehr et al. is that music is complex and it features grammar-like structures analogous to those of language. The presentation of this argument makes clear that the identified feature is shared with language. This argument has not identified any flaws in the byproduct hypothesis.

The fourth argument is that music perception appears early in ontogeny. This argument has an ontogenetic basis, not a phylogenetic one, which is what would be needed for discarding the byproduct hypothesis about the origins of music. Although it is true that neonates detect the beat of music, this could well be the outcome of central auditory processing that is used when deciphering the rhythm of language in utero (Minai, Gustafson, Fiorentino, Jongman, & Sereno, Reference Minai, Gustafson, Fiorentino, Jongman and Sereno2017).

The fifth line of evidence that aims to show that the byproduct hypothesis is incorrect is that music perception displays evidence for neural specialization and is impaired in specific deficits (e.g., congenital amusia). The first part of this argument does not disprove that music originates in language: Identifying a neural basis does not entail a unique one-to-one mapping between a specific brain region and a specific cognitive function, because brain regions form part of networks that underlie the workings of many cognitive processes. Also, developmental disorders do not necessarily represent selective impairment to a specific locus, but may rather result from early perturbations of the interconnected neonate brain that prevent specialization of function from developing (D'Souza & Karmiloff-Smith, Reference D'Souza and Karmiloff-Smith2011). It is important that most studies describe language deficits in amusics (e.g., Liu, Jiang, Wang, Xu, & Patel, Reference Liu, Jiang, Wang, Xu and Patel2015; Sihvonen et al., Reference Sihvonen, Ripolles, Leo, Rodriguez-Fornells, Soinila and Sarkamo2016), leaving open the possibility of observing overlapping brain regions and shared loci of neural disruption in the two domains, language and music. The neural overlap hypothesis does not entail shared neural circuitries between language and music (Peretz, Vuvan, Lagrois, & Armony, Reference Peretz, Vuvan, Lagrois and Armony2015), but it does not discard the possibility of neural recycling either, in the sense that language may have been recycled during evolution for musicality (Honing, ten Cate, Peretz, & Trehub, Reference Honing, ten Cate, Peretz and Trehub2015).

The last argument is that music is ancient. The claim that the byproduct hypothesis is incomplete/incorrect does not follow from this argument, especially if one entertains this hypothesis in relation to language. Mehr et al. (2020) mention that flutes are at least 40k years old, but language has been argued to play a role in our species' evolution for the last 200k years (Pagel, Reference Pagel2017).

None of these replies to the six arguments put forth by Mehr et al. (2020) shows that the main hypothesis they advance is incorrect. However, these replies suggest that Mehr et al. (2020) have not established why the byproduct hypothesis that links the origins of music to language is flawed or why the hypothesis of shared origin via a common ancestral precursor that could scaffold both music and language (e.g., musilanguage; Brown, Reference Brown, Wallin, Merker and Brown2001) is unconvincing.

Suggesting that the main claim of Mehr et al. (2020) has not been established through the six pursued lines of argument does not entail that there are no experiments that can actually support it. Arguments both in favor or against it can be adduced through two promising lines of evidence. The first one is to examine species that show signs of musical perception, although their communication system does not rely primarily on acoustic signals. One promising example is Cyprinus carpio, which was found to be successfully trained to discriminate music from different genres (Chase, Reference Chase2001). The second approach is to determine whether music is innate as language is. The innateness of language is well-known: A typically developing infant does not have the option to suppress her innate linguistic predisposition and not pick up the language of her environment. The question is whether the same can be claimed for music. Another line of investigation of this argument boils down to examining the incidence of patients that perceive spoken language although they have never received linguistic input in the auditory modality. Congenitally deaf people may develop both auditory and tactile hallucinations such as hearing voices or being fingerspelled to (Atkinson, Reference Atkinson2006). A PubMed search of the terms “congenital” AND “deaf” AND “music” suggests that this is not the case for musical hallucinations. If music is innate as language is, and given that musical hallucinations in people with acquired deafness have been amply reported (e.g., Fisman, Reference Fisman1991), the extremely low – if existent at all – occurrence of musical hallucinations in congenitally deaf people is a fact that awaits explanation. Providing one possibly goes through talking about the innateness and centrality of music in human cognition.