Dunbar posits that grooming and social cognition are key mechanisms for maintaining group cohesion, particularly in primates. However, these mechanisms alone are insufficient to explain how stability is maintained in large groups where conflicts inevitably arise. Here, we highlight tolerance as an additional mechanism crucial for large-scale social cohesion. We argue that tolerance underpins flexible social skills and supports cooperative ties beyond close kin or small cliques, thus playing a foundational role in the evolution of complex primate societies, including our own.
Tolerance, defined as the ability to accommodate social stress and conflicts, strengthens cohesion by reducing social friction and enabling more stable and frequent coalitions. Without tolerance, even sophisticated cognitive skills, such as reasoning about others’ mental states, would have limited utility in social interactions, as individuals would struggle to share the benefits of joint effort (Hare & Tomasello, 2005). For instance, chimpanzees can understand false beliefs and strategically manipulate others’ visual access to maximize their own payoffs (Karg et al., Reference Karg, Schmelz, Call and Tomasello2015; Krupenye et al., Reference Krupenye, Kano, Hirata, Call and Tomasello2016). Yet, their cooperation is often constrained by low tolerance, even when they recognize that they cannot solve a problem alone. When rewards for joint effort can be easily monopolized or when sharing becomes necessary, tensions rise and cooperation collapses (MacLean & Hare, Reference MacLean and Hare2013; Melis, Hare, & Tomasello, Reference Melis, Hare and Tomasello2006). This limitation may help explain why chimpanzees, even in populations exhibiting relatively high levels of cooperation and ecological pressures favoring coordinated hunting, still tend to hunt in small parties of only three to four individuals (Boesch, Reference Boesch2002).
In contrast, bonobos, who share a common ancestor with chimpanzees dating back approximately 1.5 to 2.1 million years (Fontsere et al., Reference Fontsere, Kuhlwilm, Morcillo-Suarez, Alvarez-Estape, Lester, Gratton, Schmidt, Dieguez, Aebischer and Álvarez-Varona2022), exhibit significantly higher levels of social tolerance, particularly in feeding context, and collaborate more effectively (Hare et al., Reference Hare, Melis, Woods, Hastings and Wrangham2007; Jaeggi, Stevens, & Van Schaik, Reference Jaeggi, Stevens and Van Schaik2010; Nolte, Sterck, & Van Leeuwen, Reference Nolte, Sterck and Van Leeuwen2023). Associated with enhanced tolerance, bonobo foraging parties tend to be larger than those of chimpanzees, both in absolute size and as a proportion of their unit-group (Furuichi, Reference Furuichi2009; Mulavwa et al., Reference Mulavwa, Furuichi, Yangozene, Yamba-Yamba, Motema-Salo, Idani, Ihobe, Hashimoto, Tashiro, Mwanza, Furuichi and Thompson2008; Surbeck & Hohmann, Reference Surbeck and Hohmann2008). Moreover, when neighboring groups meet, bonobos frequently exhibit joint foraging behaviors and forms of intergroup cooperation (Cheng et al., Reference Cheng, Samuni, Lucchesi, Deschner and Surbeck2022; Lucchesi et al., Reference Lucchesi, Cheng, Janmaat, Mundry, Pisor and Surbeck2020; Samuni & Surbeck, 2023). They voluntarily share food even with unfamiliar individuals and display flexible cooperative-communicative skills based on their understanding of others’ mental states (Tan & Hare, Reference Tan and Hare2013; Townrow & Krupenye, Reference Townrow and Krupenye2025). Such extensions of sociality beyond immediate group boundaries underscore tolerance as a foundational element for larger-scale social comity.
While endorphin release may indirectly reduce aggression by promoting positive social experience, tolerance involves direct regulation of emotionally defensive responses. For instance, chimpanzees exhibit elevated testosterone reactivity in anticipation of social interactions, which prepares them for competition and lowers the potential for tolerance (Sobolewski, Brown, & Mitani, Reference Sobolewski, Brown and Mitani2012; Wobber, Hare, et al., Reference Wobber, Hare, Maboto, Lipson, Wrangham and Ellison2010). In contrast, bonobos show increased cortisol release in similar social contexts, indicating stress responses that facilitate passive coping strategies (Wobber, Hare, et al., Reference Wobber, Wrangham and Hare2010). These strategies include mitigating anxiety through sociosexual contact, which stimulates oxytocin release and is more effective than grooming alone (Clay & de Waal, Reference Clay, de Waal, Hare and Yamamoto2015; Moscovice et al., Reference Moscovice, Surbeck, Fruth, Hohmann, Jaeggi and Deschner2019). The aggression-buffering effects of stress responses are not unique to bonobos. Macaques, for instance, experience heightened social stress when closely related individuals are in proximity during foraging, as indicated by increased rates of scratching, a behavior associated with elevated cortisol levels (Sekizawa & Kutsukake, Reference Sekizawa and Kutsukake2025). However, following scratching, macaque interactions are less likely to escalate into aggression and more likely to become affiliative (Whitehouse, Micheletta, & Waller, Reference Whitehouse, Micheletta and Waller2017). Additionally, cooperative behaviors with socially close partners further reduce cortisol levels, reinforcing stress-alleviating mechanisms that preserve social bonds (Stocker et al., Reference Stocker, Loretto, Sterck, Bugnyar and Massen2020).
These tolerant responses to social stress likely result from selection against aggression. In animals artificially selected for tolerance toward humans, increased serotonin levels in the brain are among the earliest physiological markers of reduced emotional reactivity (Agnvall et al., Reference Agnvall, Katajamaa, Altimiras and Jensen2015; Popova et al., Reference Popova, Voitenko, Kulikov and Avgustinovich1991; Wang et al., Reference Wang, Pipes, Trut, Herbeck, Vladimirova, Gulevich, Kharlamova, Johnson, Acland and Kukekova2018). Correspondingly, compared to chimpanzees, bonobos have twice the density of serotonergic axons in the amygdala (Stimpson et al., Reference Stimpson, Barger, Taglialatela, Gendron-Fitzpatrick, Hof, Hopkins and Sherwood2016). In macaques, variation in serotonergic gene profiles is associated with sociality as well. Individuals carrying two copies of the long allele of the 5-HTTLPR gene, which encodes the serotonin transporter, show lower aggression and greater social engagement (Brent et al., Reference Brent, Heilbronner, Horvath, Gonzalez-Martinez, Ruiz-Lambides, Robinson, Skene and Platt2013; Watson et al., Reference Watson, Li, Brent, Horvath, Gonzalez-Martinez, Ruíz-Lambides, Robinson, Skene and Platt2015). More broadly, across primates, serotonin concentrations correlate with impulse control and play a central role in regulating vigilance, balancing the detection of social threats with affiliative behaviors (Higley & Linnoila, Reference Higley and Linnoila1997; Mikheenko et al., Reference Mikheenko, Shiba, Sawiak, Braesicke, Cockcroft, Clarke and Roberts2015; Weinberg-Wolf & Chang, Reference Weinberg-Wolf and Chang2019; Westergaard et al., Reference Westergaard, Suomi, Chavanne, Houser, Hurley, Cleveland, Snoy and Higley2003). These patterns suggest that shifts in social cognition are closely linked to hormonal adaptations that facilitate tolerance.
Although Dunbar’s framework gives limited attention to developmental processes, selection acting on an extended developmental window may also play a crucial role in shaping tolerance (Hare, 2017). Under pressures of social complexity, primates have evolved prolonged juvenile periods that allow for more opportunities for social learning, including the acquisition of aggression-regulating skills (Joffe, Reference Joffe1997; LaFreniere, Reference LaFreniere2011). This extended developmental phase may provide more affordance for sustained tolerance, as juveniles across primate species tend to be more tolerant than mature individuals (Pereira & Fairbanks, Reference Pereira and Fairbanks2002). Delays in the maturation of adult-level inhibition over social responses may help sustain juvenile-like tolerance into later developmental stages, as observed in bonobos and Barbary macaques (Rosati & Santos, Reference Rosati and Santos2017; Wobber, Wrangham, & Hare, Reference Wobber, Hare, Maboto, Lipson, Wrangham and Ellison2010). Humans exhibit the most pronounced form of this pattern. Our uniquely prolonged dependency, marked by early-developing cooperative-communicative capacities and an extended period for socialization, likely facilitates emotional and behavioral regulation essential for maintaining cohesion within large, diverse, and dynamic social groups (Hare, Reference Hare2017; Hawkes, Reference Hawkes2014; Vollhardt, Migacheva, & Tropp, Reference Vollhardt, Migacheva, Tropp and Rivera2009).
In summary, we broaden the discussion of primate social evolution by proposing tolerance as a crucial factor enabling species, including humans, to overcome the “glass ceilings” on group size by sustaining social cohesion. Moreover, since tolerance likely scaffolds advanced social skills, fostering affective regulation (e.g., through enriched rearing environments; Honess & Marin, Reference Honess and Marin2006) rather than relying solely on cognitive solutions may be key to managing complex social networks.
Dunbar posits that grooming and social cognition are key mechanisms for maintaining group cohesion, particularly in primates. However, these mechanisms alone are insufficient to explain how stability is maintained in large groups where conflicts inevitably arise. Here, we highlight tolerance as an additional mechanism crucial for large-scale social cohesion. We argue that tolerance underpins flexible social skills and supports cooperative ties beyond close kin or small cliques, thus playing a foundational role in the evolution of complex primate societies, including our own.
Tolerance, defined as the ability to accommodate social stress and conflicts, strengthens cohesion by reducing social friction and enabling more stable and frequent coalitions. Without tolerance, even sophisticated cognitive skills, such as reasoning about others’ mental states, would have limited utility in social interactions, as individuals would struggle to share the benefits of joint effort (Hare & Tomasello, 2005). For instance, chimpanzees can understand false beliefs and strategically manipulate others’ visual access to maximize their own payoffs (Karg et al., Reference Karg, Schmelz, Call and Tomasello2015; Krupenye et al., Reference Krupenye, Kano, Hirata, Call and Tomasello2016). Yet, their cooperation is often constrained by low tolerance, even when they recognize that they cannot solve a problem alone. When rewards for joint effort can be easily monopolized or when sharing becomes necessary, tensions rise and cooperation collapses (MacLean & Hare, Reference MacLean and Hare2013; Melis, Hare, & Tomasello, Reference Melis, Hare and Tomasello2006). This limitation may help explain why chimpanzees, even in populations exhibiting relatively high levels of cooperation and ecological pressures favoring coordinated hunting, still tend to hunt in small parties of only three to four individuals (Boesch, Reference Boesch2002).
In contrast, bonobos, who share a common ancestor with chimpanzees dating back approximately 1.5 to 2.1 million years (Fontsere et al., Reference Fontsere, Kuhlwilm, Morcillo-Suarez, Alvarez-Estape, Lester, Gratton, Schmidt, Dieguez, Aebischer and Álvarez-Varona2022), exhibit significantly higher levels of social tolerance, particularly in feeding context, and collaborate more effectively (Hare et al., Reference Hare, Melis, Woods, Hastings and Wrangham2007; Jaeggi, Stevens, & Van Schaik, Reference Jaeggi, Stevens and Van Schaik2010; Nolte, Sterck, & Van Leeuwen, Reference Nolte, Sterck and Van Leeuwen2023). Associated with enhanced tolerance, bonobo foraging parties tend to be larger than those of chimpanzees, both in absolute size and as a proportion of their unit-group (Furuichi, Reference Furuichi2009; Mulavwa et al., Reference Mulavwa, Furuichi, Yangozene, Yamba-Yamba, Motema-Salo, Idani, Ihobe, Hashimoto, Tashiro, Mwanza, Furuichi and Thompson2008; Surbeck & Hohmann, Reference Surbeck and Hohmann2008). Moreover, when neighboring groups meet, bonobos frequently exhibit joint foraging behaviors and forms of intergroup cooperation (Cheng et al., Reference Cheng, Samuni, Lucchesi, Deschner and Surbeck2022; Lucchesi et al., Reference Lucchesi, Cheng, Janmaat, Mundry, Pisor and Surbeck2020; Samuni & Surbeck, 2023). They voluntarily share food even with unfamiliar individuals and display flexible cooperative-communicative skills based on their understanding of others’ mental states (Tan & Hare, Reference Tan and Hare2013; Townrow & Krupenye, Reference Townrow and Krupenye2025). Such extensions of sociality beyond immediate group boundaries underscore tolerance as a foundational element for larger-scale social comity.
While endorphin release may indirectly reduce aggression by promoting positive social experience, tolerance involves direct regulation of emotionally defensive responses. For instance, chimpanzees exhibit elevated testosterone reactivity in anticipation of social interactions, which prepares them for competition and lowers the potential for tolerance (Sobolewski, Brown, & Mitani, Reference Sobolewski, Brown and Mitani2012; Wobber, Hare, et al., Reference Wobber, Hare, Maboto, Lipson, Wrangham and Ellison2010). In contrast, bonobos show increased cortisol release in similar social contexts, indicating stress responses that facilitate passive coping strategies (Wobber, Hare, et al., Reference Wobber, Wrangham and Hare2010). These strategies include mitigating anxiety through sociosexual contact, which stimulates oxytocin release and is more effective than grooming alone (Clay & de Waal, Reference Clay, de Waal, Hare and Yamamoto2015; Moscovice et al., Reference Moscovice, Surbeck, Fruth, Hohmann, Jaeggi and Deschner2019). The aggression-buffering effects of stress responses are not unique to bonobos. Macaques, for instance, experience heightened social stress when closely related individuals are in proximity during foraging, as indicated by increased rates of scratching, a behavior associated with elevated cortisol levels (Sekizawa & Kutsukake, Reference Sekizawa and Kutsukake2025). However, following scratching, macaque interactions are less likely to escalate into aggression and more likely to become affiliative (Whitehouse, Micheletta, & Waller, Reference Whitehouse, Micheletta and Waller2017). Additionally, cooperative behaviors with socially close partners further reduce cortisol levels, reinforcing stress-alleviating mechanisms that preserve social bonds (Stocker et al., Reference Stocker, Loretto, Sterck, Bugnyar and Massen2020).
These tolerant responses to social stress likely result from selection against aggression. In animals artificially selected for tolerance toward humans, increased serotonin levels in the brain are among the earliest physiological markers of reduced emotional reactivity (Agnvall et al., Reference Agnvall, Katajamaa, Altimiras and Jensen2015; Popova et al., Reference Popova, Voitenko, Kulikov and Avgustinovich1991; Wang et al., Reference Wang, Pipes, Trut, Herbeck, Vladimirova, Gulevich, Kharlamova, Johnson, Acland and Kukekova2018). Correspondingly, compared to chimpanzees, bonobos have twice the density of serotonergic axons in the amygdala (Stimpson et al., Reference Stimpson, Barger, Taglialatela, Gendron-Fitzpatrick, Hof, Hopkins and Sherwood2016). In macaques, variation in serotonergic gene profiles is associated with sociality as well. Individuals carrying two copies of the long allele of the 5-HTTLPR gene, which encodes the serotonin transporter, show lower aggression and greater social engagement (Brent et al., Reference Brent, Heilbronner, Horvath, Gonzalez-Martinez, Ruiz-Lambides, Robinson, Skene and Platt2013; Watson et al., Reference Watson, Li, Brent, Horvath, Gonzalez-Martinez, Ruíz-Lambides, Robinson, Skene and Platt2015). More broadly, across primates, serotonin concentrations correlate with impulse control and play a central role in regulating vigilance, balancing the detection of social threats with affiliative behaviors (Higley & Linnoila, Reference Higley and Linnoila1997; Mikheenko et al., Reference Mikheenko, Shiba, Sawiak, Braesicke, Cockcroft, Clarke and Roberts2015; Weinberg-Wolf & Chang, Reference Weinberg-Wolf and Chang2019; Westergaard et al., Reference Westergaard, Suomi, Chavanne, Houser, Hurley, Cleveland, Snoy and Higley2003). These patterns suggest that shifts in social cognition are closely linked to hormonal adaptations that facilitate tolerance.
Although Dunbar’s framework gives limited attention to developmental processes, selection acting on an extended developmental window may also play a crucial role in shaping tolerance (Hare, 2017). Under pressures of social complexity, primates have evolved prolonged juvenile periods that allow for more opportunities for social learning, including the acquisition of aggression-regulating skills (Joffe, Reference Joffe1997; LaFreniere, Reference LaFreniere2011). This extended developmental phase may provide more affordance for sustained tolerance, as juveniles across primate species tend to be more tolerant than mature individuals (Pereira & Fairbanks, Reference Pereira and Fairbanks2002). Delays in the maturation of adult-level inhibition over social responses may help sustain juvenile-like tolerance into later developmental stages, as observed in bonobos and Barbary macaques (Rosati & Santos, Reference Rosati and Santos2017; Wobber, Wrangham, & Hare, Reference Wobber, Hare, Maboto, Lipson, Wrangham and Ellison2010). Humans exhibit the most pronounced form of this pattern. Our uniquely prolonged dependency, marked by early-developing cooperative-communicative capacities and an extended period for socialization, likely facilitates emotional and behavioral regulation essential for maintaining cohesion within large, diverse, and dynamic social groups (Hare, Reference Hare2017; Hawkes, Reference Hawkes2014; Vollhardt, Migacheva, & Tropp, Reference Vollhardt, Migacheva, Tropp and Rivera2009).
In summary, we broaden the discussion of primate social evolution by proposing tolerance as a crucial factor enabling species, including humans, to overcome the “glass ceilings” on group size by sustaining social cohesion. Moreover, since tolerance likely scaffolds advanced social skills, fostering affective regulation (e.g., through enriched rearing environments; Honess & Marin, Reference Honess and Marin2006) rather than relying solely on cognitive solutions may be key to managing complex social networks.
Financial support
Funding for this commentary was provided by Duke Kunshan University. The work of BH is supported in part by grant 62280 from the John Templeton Foundation.
Competing interests
The authors declare that there are no competing interests.