As someone who accidentally kicked off two decades of theorizing about the evolution of music with a few pages of discussion in one book (Pinker, Reference Pinker1997), I've long been nonplussed at the fervor with which theoreticians have striven to show that music is a Darwinian adaptation. I had included that discussion partly to sharpen readers’ conceptions of the criteria for an adaptation. After 500 pages that had argued for the adaptive basis of many psychological faculties – stereo vision, the recency effect, disgust, jealousy, and revenge – I wanted to show that not everything is an adaptation. Any rigorous adaptationist hypothesis had to go beyond a trait merely being commonplace and lay out independent signs of engineering design for attaining some goal that was a subgoal of inclusive fitness. Whereas, it's easy to reverse-engineer, say, language, stereo vision, or fear, the function of music is far from obvious. If the concept of adaptation is not to apply to everything, and hence to nothing, we must entertain the possibility that music instead is a pleasure technology: an application of human ingenuity to the evolutionarily dubious but proximately compelling goal of activating our pleasure circuits. In the case of music, these circuits might belong to language, auditory scene analysis, habitat selection, emotional calls, motor control, and perhaps some non-adaptive features of the auditory brain, such as proximity to other systems and their entrainment by the periodicity in auditory signals.
And yet while many scholars despise the idea that psychological traits are adaptations, blowing it off as a bunch of after-the-fact just-so stories, they are equally offended by the idea that music is not an adaptation, and so have offered hypotheses that are dead on arrival, such as that music evolved to bond the group or attract females. The reason for the discrepancy, I suspect, is that adaptation is not conceived of as a testable hypothesis from evolutionary biology but as an affirmation of how we value, deplore, or frame features of human nature. To say that music is an adaptation is to exalt its value; to say it is a by-product is a philistine denigration.
For these reasons, it's a pleasure to see Mehr et al. transcend all this wooliness in their superb article. After performing masterful necropsies on the bond-the-group and woo-the-ladies hypotheses, and raising reasonable criticisms of the by-product possibility, they propose a two-part hypothesis – credible signaling of coalition quality and of attention to infants – that satisfies the criteria for an adaptation and has impressive support from phylogenetic, ethnographic, genetic, and behavioral evidence. Maybe the theory is even true, although I think that it has some shortcomings.
First, it's not easy to see how these two very specific functions can be reconciled with the broad range of forms and contexts in which music is produced and enjoyed. If coalition quality and infant care are the two pillars, and everything else a set of cultural embellishments and extensions, we should see signs that those two functions are particularly robust, universal, archetypal, pervasive, and salient in the panoply of musical experience. But, that is exactly what was not found in Mehr et al.'s (Reference Mehr, Singh, Knox, Ketter, Pickens-Jones, Atwood and Glowacki2019) mammoth cross-cultural survey. It was not the case that music exemplifying the two proposed cores, such as war songs and lullabies, were universal, whereas the supposed extensions, such as love songs, healing songs, dance music, and other genres, were distributed more patchily, followed paths of historical influence rather than species-wide universality, or had less reliable acoustic signatures. Our major conclusion was that the four kinds of music spotlighted in the paper, together with 16 other genres were pretty much equally robust, distinctive, and universal: “Music is not a fixed biological response with a single prototypical adaptive function: It is produced worldwide in diverse behavioral contexts ….”
In a similar vein, the contemporary phenomenology of music shows no signs of the core-plus-periphery structure their theory implies. I see no evidence that group-advertising genres such as anthems and team songs, together with lullabies, are the most popular or accessible musical genres, that listeners backslide to pondering formidable cliques or calm babies when they experience other kinds of music, or any other sign of centrality. Both the ethnography and the psychology imply that music involves a broad mapping between acoustic structures and human experience, with no obvious common reaction or instrumental benefit. We enjoy a diversity of musical forms equally, and with no characteristic outcome other than the pleasure we get as we listen.
This leads to my second reservation about the theory. The most blazingly obvious feature of music – people enjoy it – plays no role in the theory. Although the authors criticize the by-product hypothesis in generic terms, they don't focus on the specific version in which humans apply their know-how to tap sources of acoustic and motoric pleasure. Indeed, I find it hard to see how that hypothesis could not be true. If we're smart and our brains motivate us with pleasure, what could stop us from deploying our intelligence to gratify ourselves, as we do so flagrantly and maladaptively with non-procreative sex and recreational drugs (two examples for those of you, who, like me, don't even like cheesecake)?
My final reservation concerns something else that is conspicuous by its absence: language. Music, like language, lacks close homologs in other apes; has a hierarchical structure of phrases within phrases; has a complex rhythmic structure that matches that of language so uncannily that we can put words to music; feels like it's communicating something even when it isn't; and is composed of harmonically related frequencies found in no commonly experienced natural acoustic stimulus other than the human voice (and, less frequently, animal vocalization). And, as we showed in Mehr et al. (Reference Mehr, Singh, Knox, Ketter, Pickens-Jones, Atwood and Glowacki2019), all musical genres include lyrics. In the current theory, these are all coincidences.
Despite these zones of skepticism, I commend the authors on this tour de force, which elevates the topic to new levels of theoretical rigor and empirical richness.
As someone who accidentally kicked off two decades of theorizing about the evolution of music with a few pages of discussion in one book (Pinker, Reference Pinker1997), I've long been nonplussed at the fervor with which theoreticians have striven to show that music is a Darwinian adaptation. I had included that discussion partly to sharpen readers’ conceptions of the criteria for an adaptation. After 500 pages that had argued for the adaptive basis of many psychological faculties – stereo vision, the recency effect, disgust, jealousy, and revenge – I wanted to show that not everything is an adaptation. Any rigorous adaptationist hypothesis had to go beyond a trait merely being commonplace and lay out independent signs of engineering design for attaining some goal that was a subgoal of inclusive fitness. Whereas, it's easy to reverse-engineer, say, language, stereo vision, or fear, the function of music is far from obvious. If the concept of adaptation is not to apply to everything, and hence to nothing, we must entertain the possibility that music instead is a pleasure technology: an application of human ingenuity to the evolutionarily dubious but proximately compelling goal of activating our pleasure circuits. In the case of music, these circuits might belong to language, auditory scene analysis, habitat selection, emotional calls, motor control, and perhaps some non-adaptive features of the auditory brain, such as proximity to other systems and their entrainment by the periodicity in auditory signals.
And yet while many scholars despise the idea that psychological traits are adaptations, blowing it off as a bunch of after-the-fact just-so stories, they are equally offended by the idea that music is not an adaptation, and so have offered hypotheses that are dead on arrival, such as that music evolved to bond the group or attract females. The reason for the discrepancy, I suspect, is that adaptation is not conceived of as a testable hypothesis from evolutionary biology but as an affirmation of how we value, deplore, or frame features of human nature. To say that music is an adaptation is to exalt its value; to say it is a by-product is a philistine denigration.
For these reasons, it's a pleasure to see Mehr et al. transcend all this wooliness in their superb article. After performing masterful necropsies on the bond-the-group and woo-the-ladies hypotheses, and raising reasonable criticisms of the by-product possibility, they propose a two-part hypothesis – credible signaling of coalition quality and of attention to infants – that satisfies the criteria for an adaptation and has impressive support from phylogenetic, ethnographic, genetic, and behavioral evidence. Maybe the theory is even true, although I think that it has some shortcomings.
First, it's not easy to see how these two very specific functions can be reconciled with the broad range of forms and contexts in which music is produced and enjoyed. If coalition quality and infant care are the two pillars, and everything else a set of cultural embellishments and extensions, we should see signs that those two functions are particularly robust, universal, archetypal, pervasive, and salient in the panoply of musical experience. But, that is exactly what was not found in Mehr et al.'s (Reference Mehr, Singh, Knox, Ketter, Pickens-Jones, Atwood and Glowacki2019) mammoth cross-cultural survey. It was not the case that music exemplifying the two proposed cores, such as war songs and lullabies, were universal, whereas the supposed extensions, such as love songs, healing songs, dance music, and other genres, were distributed more patchily, followed paths of historical influence rather than species-wide universality, or had less reliable acoustic signatures. Our major conclusion was that the four kinds of music spotlighted in the paper, together with 16 other genres were pretty much equally robust, distinctive, and universal: “Music is not a fixed biological response with a single prototypical adaptive function: It is produced worldwide in diverse behavioral contexts ….”
In a similar vein, the contemporary phenomenology of music shows no signs of the core-plus-periphery structure their theory implies. I see no evidence that group-advertising genres such as anthems and team songs, together with lullabies, are the most popular or accessible musical genres, that listeners backslide to pondering formidable cliques or calm babies when they experience other kinds of music, or any other sign of centrality. Both the ethnography and the psychology imply that music involves a broad mapping between acoustic structures and human experience, with no obvious common reaction or instrumental benefit. We enjoy a diversity of musical forms equally, and with no characteristic outcome other than the pleasure we get as we listen.
This leads to my second reservation about the theory. The most blazingly obvious feature of music – people enjoy it – plays no role in the theory. Although the authors criticize the by-product hypothesis in generic terms, they don't focus on the specific version in which humans apply their know-how to tap sources of acoustic and motoric pleasure. Indeed, I find it hard to see how that hypothesis could not be true. If we're smart and our brains motivate us with pleasure, what could stop us from deploying our intelligence to gratify ourselves, as we do so flagrantly and maladaptively with non-procreative sex and recreational drugs (two examples for those of you, who, like me, don't even like cheesecake)?
My final reservation concerns something else that is conspicuous by its absence: language. Music, like language, lacks close homologs in other apes; has a hierarchical structure of phrases within phrases; has a complex rhythmic structure that matches that of language so uncannily that we can put words to music; feels like it's communicating something even when it isn't; and is composed of harmonically related frequencies found in no commonly experienced natural acoustic stimulus other than the human voice (and, less frequently, animal vocalization). And, as we showed in Mehr et al. (Reference Mehr, Singh, Knox, Ketter, Pickens-Jones, Atwood and Glowacki2019), all musical genres include lyrics. In the current theory, these are all coincidences.
Despite these zones of skepticism, I commend the authors on this tour de force, which elevates the topic to new levels of theoretical rigor and empirical richness.
Conflict of interest
None.