The stressors of social life within large groups might have never been more apparent than in current society, but these stressors are not new. Dunbar argues that to maintain social cohesion despite these stressors, primates use social-structural, behavioral, and cognitive strategies, including vocalization, chorusing, and conversation. In this response, we spell out the links between Dunbar’s insights on group cohesion and theories of the evolution and social functions of music.

Extending Dunbar’s case leads to a reconceptualization of the social bonding hypothesis for the evolution of musicality, an account co-authored in BBS by one of us with other colleagues (Savage et al., Reference Savage, Loui, Tarr, Schachner, Glowacki, Mithen and Fitch2021). The social bonding hypothesis posits that rhythm-based and pitch-based processes are design features of music that bring larger groups together. We had previously proposed that prediction and reward, as implemented by the dopaminergic reward and endogenous opioid systems (EOS), serve as the underlying mechanisms for this bonding. We now agree with Dunbar that meta-cognition encompasses an additional suite of cognitive mechanisms – including imagination, episodic simulation, causal inference, and inhibition – through which musicality brings large groups together.

Evidence for EOS-mediated musical effects on social bonding includes, but is not limited to, temporally synchronized processes. In addition to people moving together being more cooperative together (Cirelli, Einarson & Trainor, Reference Cirelli, Einarson and Trainor2014; Hove & Risen, Reference Hove and Risen2009; Rabinowitch & Meltzoff, Reference Rabinowitch and Meltzoff2017), sound elicits rhythmic movement in male more than female chimpanzees (Hattori & Tomonaga, Reference Hattori and Tomonaga2020). Rhythmic swaying was sometimes accompanied by vocalization, reflecting a behavioral mechanism that allows music to trigger the EOS. Vocalizations of species like indris (Indri indri) and gelada baboons (Theropithecus gelada) constitute structured songs with distinct timing and pitch variations, suggesting a level of complexity that parallels aspects of human music (Gamba et al., Reference Gamba, Torti, Estienne, Randrianarison, Valente, Rovara, Bonadonna, Friard and Giacoma2016).

How meta-cognitive processes shaped the evolution of musicality is less clear in the animal world, but new work from music cognition shows that meta-cognition is entwined with musicality in humans. Trevor and Frühholz (Reference Trevor and Frühholz2024) argue that music, beyond its effects on social bonding, evolved as a tool for socio-affective fiction (Trevor & Frühholz, Reference Trevor and Frühholz2024). According to this theory, music added a dimension of immersiveness to storytelling that enabled the elicitation of strong emotions, allowing listeners to practice the emotion regulation skills that promote harmonious living within large groups.

The vivid imaginings that music can elicit make it a powerful tool for intergroup understanding. Music elicits vivid episodes of scene construction in spontaneous thought, including autobiographical memories, future simulations, and fictional imaginings (Belfi et al., Reference Belfi, Chen, Schneider and Tranel2016; Margulis et al., Reference Margulis, Wong, Turnbull, Kubit and McAuley2022; Margulis & Jakubowski, Reference Margulis and Jakubowski2024). These episodes engage the default mode network, including but not limited to the medial prefrontal cortex, which is also engaged during acts of musical creativity such as improvisation tasks (Belden et al., Reference Belden, Zeng, Przysinda, Anteraper, Whitfield-Gabrieli and Loui2019; Vergara et al., Reference Vergara, Norgaard, Miller, Beaty, Dhakal, Dhamala and Calhoun2021). Converging evidence from neuroimaging studies shows that regions of the default mode network also represent high-level musical event structure, thus providing a shared neural substrate for meta-cognition and socio-affective processing during music listening (Williams et al., Reference Williams, Margulis, Nastase, Chen, Hasson, Norman and Baldassano2022).

Part of meta-cognition entails causal inference, based on the available evidence for the perceptual system. In that regard, a recent review article argues that source sensitivity, defined as the capacity to infer the causal and contextual sources of music, plays an even stronger role in music appreciation than previously acknowledged (Thompson et al., Reference Thompson, Bullot and Margulis2023). Presented with musical sound, listeners automatically generate a representation of who created it, and the music’s typical social contexts and functions. This ability emerges early in life, with five-month-olds responding preferentially to a melody previously sung in person by their caregiver in comparison to another highly similar melody (Mehr, Song, & Spelke, Reference Mehr, Song and Spelke2016). No such preference arose if the infant had previously encountered the melody on a screen or emitted by a toy. The contextual circumstances within which they were exposed to the melody shaped how they subsequently received it. Adults interpret music “as the sonic trace of social relationships between a group” (Aucouturier & Canonne, Reference Aucouturier and Canonne2017) such that even when presented with novel musical excerpts, adults engage in mentalizing and infer their associational and contextual meaning, resulting in widespread intersubjectivity in the personal thoughts and imaginings the music evoked (Margulis et al., Reference Margulis, Wong, Turnbull, Kubit and McAuley2022; Wu et al., submitted). Individuals who report stronger musical skills (Müllensiefen et al., Reference Müllensiefen, Gingras, Musil and Stewart2014) were also more accurate at categorizing Chinese and Western melodies by their cultural origin. This suggests that alongside more robust internal models of musical structure comes an enhanced capacity to infer social context.

Finally, inhibition is a fundamental meta-cognitive process in music-making, playing a crucial role in both individual and group coordination. Dunbar describes inhibition as a meta-cognitive skill for exerting self-control in social situations, and cases of inhibition are pervasive in musical behavior. Self-control is essential for learning to make music, as executive functions such as response inhibition enable musicians to synchronize their actions in ensemble settings, ensuring that they start, stop, and adjust their playing in unison. Multiple studies in music cognition have shown that children with musical training perform better on executive function tasks, especially response inhibition (Hennessy et al., Reference Hennessy, Sachs, Ilari and Habibi2019; Sachs et al., Reference Sachs, Kaplan, Der Sarkissian and Habibi2017; Zuk et al., Reference Zuk, Benjamin, Kenyon and Gaab2014). Neurologically, inhibition is associated with the anterior cingulate cortex (ACC). This region is activated in musical tasks, for example, when singers adjust their vocal pitch in response to perturbed auditory feedback (Zarate, Wood, & Zatorre, Reference Zarate, Wood and Zatorre2010). This role of the ACC extends beyond human music-making – in primates, the same region governs explore/exploit behaviors, suppresses maladaptive responses, and encodes prediction errors as well as simulated reward signals (Hayden, Pearson, & Platt, Reference Hayden, Pearson and Platt2009, Reference Hayden, Heilbronner, Pearson and Platt2011). The shared reliance on inhibitory control in both human and primate cognition suggests that musical engagement not only refines executive function but may also contribute to fostering social cohesion by reinforcing the cognitive flexibility necessary for cooperative behavior.