“Music As a Coevolved System for Social Bonding” (MSB) is a unifying hypothesis making a broad set of interwoven claims and predictions. However, too many predictions are vaguely defined, such as: “social bonding design features operate at multiple levels simultaneously, in the same way that a couple dancing can intensify their own relationship, and their relationship with the broader social group.” How do these different types of dance interactions influence social bonding? Specifically, would MSB predict greater social bonding after solo dancing at a rock concert versus couples ballroom dancing versus partner rotation in swing dancing or Greek Kalamatianos folkdance in a circle holding hands? Or predict higher levels of social bonding for experienced, habitual dancers versus occasional dancers? What's the time course of these effects? More importantly, how would MSB measure this and disentangle social bonding arising from music, or dancing, or simply being physically active in a congenial social setting (e.g., doing group calisthenics) with like-minded people? The authors conjecture “singing in large choirs should produce greater bonding than singing in small choirs” and mention studies suggesting oxytocin levels are elevated after singing, but admit “evidence linking oxytocin with music remains limited.” What are the best quantitative psychological measures of social bonding and what are its best measurable and accurate neurochemical and neurophysiological markers? Oxytocin, dopamine or EOS levels? Cortisol levels to measure reduced stress? Or the neuronal coherence of dancers or choir singers? The authors sidestep detailed predictions and experimental measures, crucial to test and validate MSB. An example of one possible approach measures prosociality following interpersonal synchrony (Cirelli et al., Reference Cirelli, Einarson and Trainor2014) or monitors performers and audience for multiple behavioral and neural measures during musical performance (Chang et al., Reference Chang, Livingstone, Bosnyak and Trainor2017; Swarbrick et al., Reference Swarbrick, Bosnyak, Livingstone, Bansal, Marsh-Rollo, Woolhouse and Trainor2019). Critical tests of MSB require measurements and analysis of psychological measures of social bonding and associated neuromarkers before, during, following live performance or participatory musical or dance events and in longitudinal studies.
Despite “ubiquitous use of music in communal ceremonies and rituals” and the assertion that “dance is a core part of music-making and not a separate domain,” there is simply no obligatory nor causal link between social bonding, dance and musicality at an individual level. Anecdotally, Beethoven was asocial and “never learned to dance in time with music” (Cooper, Reference Cooper2010). His greatest pleasure was solitary walks in the countryside where he was inspired to compose. Dancers in Merce Cunningham's company didn't dance to a beat, rather Merce created dances based on chance juxtaposition of movement and music. Would MSB predict less social bonding at Cunningham concerts or among dancers in his company? One of the most popular radio programs, Desert Island Disks, asked what music you would bring if you were a solitary castaway on a desert isle. Such solitary enjoyment of music doesn't jibe with the claim that musicality goes hand-in-hand with social interactions. In fact, much music making is solitary (professional musicians may spend 50–100 times more time alone practicing than performing in a group). However, MSB asserts solitary music making or listening is irrelevant from an evolutionary perspective, emphasizing participatory music. Although MSB's suggested testbeds (drumming circles, campfire singalongs, and folk dances) are certainly valid and important, this restricted realm of explanatory and predictive power limits the generality and applicability of MSB. One avenue for future neuroimaging studies might be to explore “neural homophily” (a functional similarity in brain activity in friends – Parkinson et al., Reference Parkinson, Kleinbaum and Wheatley2018) in relation to shared music preferences.
A key aspect of MSB is a neurobiological model for interactions between music perception, production, and social bonding, which gives a central role for dopamine in music prediction and reward (Ferreri et al., Reference Ferreri, Mas-Herrerro, Zatorre, Ripolles, Gomez-Andres, Alicart and Rodriguez-Fornelis2019; but see Goupil & Aucouturier, Reference Goupil and Aucouturier2019 and Castro et al., Reference Castro, L'heritier, Plailly, Saive, Corneyllie, Tillmann and Perrin2020 for other music-evoked emotions). However, although a good beginning, the MSB neural model (Fig. 3 of the Savage et al. target article) is overly simplified – and impoverished because it does not incorporate brain structures known to play important roles in music perception and production: (a) cerebellum contributes to timing, rhythm perception, and generation (Teki & Griffiths, Reference Teki and Griffiths2016; Teki et al., Reference Teki, Grube, Kumar and Griffiths2011), (b) other frontal areas such as frontopolar cortex (Medalla & Barbas, Reference Medalla and Barbas2014), and dorsolateral frontal cortex (Mas-Herrero et al., Reference Mas-Herrero, Dagher and Zatorre2018), (c) higher auditory association cortex, in anterior STG (Hackett, Reference Hackett2015; Moerel et al., Reference Moerel, De Martino and Formisano2014), and (d) insula. Their model doesn't mention evolution of hemispheric lateralization for musical processing (Albouy et al., Reference Albouy, Benjamin, Morillon and Zatorre2020; Zatorre, Reference Zatorre2001). Although simplification can be useful, it limits precise and well-formulated predictions.
Nor does MSB draw a compelling link between neural music and reward pathways with social cognition networks, focusing on the arcuate fasciculus and research indicating emotionally empathic people have higher arcuate microstructural integrity compared to autistic individuals with impaired social bonding, and less arcuate connectivity. However, music is a strength of autistic individuals, who often have excellent musical memory, pitch perception, and music-evoked emotions (Wenhart et al., Reference Wenhart, Bethlehem, Baron-Cohen and Altenmuller2019). If people with autism have good musical perceptual abilities, but poor integrity of the arcuate and poor social bonding, this is an argument against MSB's assumption of a strong link between musicality and social bonding.
MSB raises fascinating evolutionary questions, and opportunities to use animal models to help understand the evolution of “proto-musical” abilities and mechanisms for social bonding – and underlying neural networks and connections. Duetting, in which pair partners combine their songs into temporally coordinated joint displays, helps strengthen pair bonds (Baldassarre et al., Reference Baldassarre, Grieg and Webster2016). Hoffmann et al. (Reference Hoffmann, Trost, Voigt, Leitner, Lemazina, Sagunsky and Gahr2019) demonstrated synchronization of premotor brain activity in duetting pairs of sparrow weavers in the wild, consistent with MSB view that rhythmic cooperative behavior leads to “neural resonance” (synchronous brain activity across individuals) that facilitates social bonding. Is this resonance stronger in longer established pairs?
Does duetting in gibbons and turn-taking (Pika et al., Reference Pika, Wilkinson, Kendrick and Vernes2018) in other NHPs, and coordinated pant-hooting in chimpanzees promote group bonding? Chimpanzees engage in rhythmic group “rain dances” in the wild. Hattori and Tomonaga (Reference Hattori and Tomonaga2020) showed chimp rhythmic swaying could be sound-induced, providing a model for “proto-dance” studies to measure whether rhythmic group movement enhances social bonding in chimpanzees. Although monkeys don't entrain to rhythm, they're sensitive to harmonic structure. However, macaques lack relative pitch ability to recognize non-octave transpositions of melodies (Wright et al., Reference Wright, Rivera, Hulse, Shyan and Neiworth2000), perhaps related to the lack of differentiation of a “cortical music pathway” (Norman-Haignere et al., Reference Norman-Haignere, Kanwisher, McDermott and Conway2019). It would be valuable to confirm whether this result generalizes to other monkeys, such as more vocal and “musical” duetting species (marmosets, geladas, and gibbons; Geissmann, Reference Geissmann, Wallin, Merker and Brown2000), or to great apes. Research investigating the neural connections between auditory, vocalization, reward, and social cognition networks (Freiwald, Reference Freiwald2020; Shepherd et al., Reference Shepherd and Freiwald2018; Sliwa & Freiwald, Reference Sliwa and Freiwald2017) will illuminate neural mechanisms and the role of vocalizations, reward pathways, and social bonding in primates.
“Music As a Coevolved System for Social Bonding” (MSB) is a unifying hypothesis making a broad set of interwoven claims and predictions. However, too many predictions are vaguely defined, such as: “social bonding design features operate at multiple levels simultaneously, in the same way that a couple dancing can intensify their own relationship, and their relationship with the broader social group.” How do these different types of dance interactions influence social bonding? Specifically, would MSB predict greater social bonding after solo dancing at a rock concert versus couples ballroom dancing versus partner rotation in swing dancing or Greek Kalamatianos folkdance in a circle holding hands? Or predict higher levels of social bonding for experienced, habitual dancers versus occasional dancers? What's the time course of these effects? More importantly, how would MSB measure this and disentangle social bonding arising from music, or dancing, or simply being physically active in a congenial social setting (e.g., doing group calisthenics) with like-minded people? The authors conjecture “singing in large choirs should produce greater bonding than singing in small choirs” and mention studies suggesting oxytocin levels are elevated after singing, but admit “evidence linking oxytocin with music remains limited.” What are the best quantitative psychological measures of social bonding and what are its best measurable and accurate neurochemical and neurophysiological markers? Oxytocin, dopamine or EOS levels? Cortisol levels to measure reduced stress? Or the neuronal coherence of dancers or choir singers? The authors sidestep detailed predictions and experimental measures, crucial to test and validate MSB. An example of one possible approach measures prosociality following interpersonal synchrony (Cirelli et al., Reference Cirelli, Einarson and Trainor2014) or monitors performers and audience for multiple behavioral and neural measures during musical performance (Chang et al., Reference Chang, Livingstone, Bosnyak and Trainor2017; Swarbrick et al., Reference Swarbrick, Bosnyak, Livingstone, Bansal, Marsh-Rollo, Woolhouse and Trainor2019). Critical tests of MSB require measurements and analysis of psychological measures of social bonding and associated neuromarkers before, during, following live performance or participatory musical or dance events and in longitudinal studies.
Despite “ubiquitous use of music in communal ceremonies and rituals” and the assertion that “dance is a core part of music-making and not a separate domain,” there is simply no obligatory nor causal link between social bonding, dance and musicality at an individual level. Anecdotally, Beethoven was asocial and “never learned to dance in time with music” (Cooper, Reference Cooper2010). His greatest pleasure was solitary walks in the countryside where he was inspired to compose. Dancers in Merce Cunningham's company didn't dance to a beat, rather Merce created dances based on chance juxtaposition of movement and music. Would MSB predict less social bonding at Cunningham concerts or among dancers in his company? One of the most popular radio programs, Desert Island Disks, asked what music you would bring if you were a solitary castaway on a desert isle. Such solitary enjoyment of music doesn't jibe with the claim that musicality goes hand-in-hand with social interactions. In fact, much music making is solitary (professional musicians may spend 50–100 times more time alone practicing than performing in a group). However, MSB asserts solitary music making or listening is irrelevant from an evolutionary perspective, emphasizing participatory music. Although MSB's suggested testbeds (drumming circles, campfire singalongs, and folk dances) are certainly valid and important, this restricted realm of explanatory and predictive power limits the generality and applicability of MSB. One avenue for future neuroimaging studies might be to explore “neural homophily” (a functional similarity in brain activity in friends – Parkinson et al., Reference Parkinson, Kleinbaum and Wheatley2018) in relation to shared music preferences.
A key aspect of MSB is a neurobiological model for interactions between music perception, production, and social bonding, which gives a central role for dopamine in music prediction and reward (Ferreri et al., Reference Ferreri, Mas-Herrerro, Zatorre, Ripolles, Gomez-Andres, Alicart and Rodriguez-Fornelis2019; but see Goupil & Aucouturier, Reference Goupil and Aucouturier2019 and Castro et al., Reference Castro, L'heritier, Plailly, Saive, Corneyllie, Tillmann and Perrin2020 for other music-evoked emotions). However, although a good beginning, the MSB neural model (Fig. 3 of the Savage et al. target article) is overly simplified – and impoverished because it does not incorporate brain structures known to play important roles in music perception and production: (a) cerebellum contributes to timing, rhythm perception, and generation (Teki & Griffiths, Reference Teki and Griffiths2016; Teki et al., Reference Teki, Grube, Kumar and Griffiths2011), (b) other frontal areas such as frontopolar cortex (Medalla & Barbas, Reference Medalla and Barbas2014), and dorsolateral frontal cortex (Mas-Herrero et al., Reference Mas-Herrero, Dagher and Zatorre2018), (c) higher auditory association cortex, in anterior STG (Hackett, Reference Hackett2015; Moerel et al., Reference Moerel, De Martino and Formisano2014), and (d) insula. Their model doesn't mention evolution of hemispheric lateralization for musical processing (Albouy et al., Reference Albouy, Benjamin, Morillon and Zatorre2020; Zatorre, Reference Zatorre2001). Although simplification can be useful, it limits precise and well-formulated predictions.
Nor does MSB draw a compelling link between neural music and reward pathways with social cognition networks, focusing on the arcuate fasciculus and research indicating emotionally empathic people have higher arcuate microstructural integrity compared to autistic individuals with impaired social bonding, and less arcuate connectivity. However, music is a strength of autistic individuals, who often have excellent musical memory, pitch perception, and music-evoked emotions (Wenhart et al., Reference Wenhart, Bethlehem, Baron-Cohen and Altenmuller2019). If people with autism have good musical perceptual abilities, but poor integrity of the arcuate and poor social bonding, this is an argument against MSB's assumption of a strong link between musicality and social bonding.
MSB raises fascinating evolutionary questions, and opportunities to use animal models to help understand the evolution of “proto-musical” abilities and mechanisms for social bonding – and underlying neural networks and connections. Duetting, in which pair partners combine their songs into temporally coordinated joint displays, helps strengthen pair bonds (Baldassarre et al., Reference Baldassarre, Grieg and Webster2016). Hoffmann et al. (Reference Hoffmann, Trost, Voigt, Leitner, Lemazina, Sagunsky and Gahr2019) demonstrated synchronization of premotor brain activity in duetting pairs of sparrow weavers in the wild, consistent with MSB view that rhythmic cooperative behavior leads to “neural resonance” (synchronous brain activity across individuals) that facilitates social bonding. Is this resonance stronger in longer established pairs?
Does duetting in gibbons and turn-taking (Pika et al., Reference Pika, Wilkinson, Kendrick and Vernes2018) in other NHPs, and coordinated pant-hooting in chimpanzees promote group bonding? Chimpanzees engage in rhythmic group “rain dances” in the wild. Hattori and Tomonaga (Reference Hattori and Tomonaga2020) showed chimp rhythmic swaying could be sound-induced, providing a model for “proto-dance” studies to measure whether rhythmic group movement enhances social bonding in chimpanzees. Although monkeys don't entrain to rhythm, they're sensitive to harmonic structure. However, macaques lack relative pitch ability to recognize non-octave transpositions of melodies (Wright et al., Reference Wright, Rivera, Hulse, Shyan and Neiworth2000), perhaps related to the lack of differentiation of a “cortical music pathway” (Norman-Haignere et al., Reference Norman-Haignere, Kanwisher, McDermott and Conway2019). It would be valuable to confirm whether this result generalizes to other monkeys, such as more vocal and “musical” duetting species (marmosets, geladas, and gibbons; Geissmann, Reference Geissmann, Wallin, Merker and Brown2000), or to great apes. Research investigating the neural connections between auditory, vocalization, reward, and social cognition networks (Freiwald, Reference Freiwald2020; Shepherd et al., Reference Shepherd and Freiwald2018; Sliwa & Freiwald, Reference Sliwa and Freiwald2017) will illuminate neural mechanisms and the role of vocalizations, reward pathways, and social bonding in primates.
Conflict of interest
None.