In his target paper, Dunbar provides a compelling account (and a comprehensive testing) of the structural, behavioral, and cognitive mechanisms that make possible the creation and maintenance of large animal groups (with a focus on primates). According to his view, these mechanisms allow to circumvent the two main problems resulting from group life: stress and dispersion. Dunbar’s view is that stress can be attenuated through the potentiation of the endorphin system via direct grooming or grooming-like activities, like vocal chorusing (or, in the case of humans, a panoply of language-related activities). Furthermore, the development of advanced social cognition skills would allow to maintain weak ties with individuals one cannot groom with.

Since stress management is a crucial component of Dunbar’s model, other mechanisms contributing to reduce socially induced stress can be expected to have favored the creation and maintenance of larger primate groups. In this commentary, I wish to argue that Dunbar’s model would benefit from the consideration of the evolutionary strategies aimed, specifically, at reducing reactive aggression responses. The activation of the endorphin system that results from grooming and grooming-like activities can be seen as a reward mechanism that compensates for social stress. By contrast, the attenuation of reactive aggression responses could be seen as a desensitization mechanism that reduces the impact of social stressors. This latter mechanism entails the hypoactivation of the hypothalamic-pituitary-adrenal axis, which is typically observed in many domesticated animals (e.g., Ericsson et al., Reference Ericsson, Fallahsharoudi, Bergquist, Kushnir and Jensen2014), but also in humans (Chrousos et al., Reference Chrousos, Renquist, Brandon, Eil, Pugeat, Vigersky, Cutler, Loriaux and Lipsett1982), who have been claimed to have gone through a self-domestication process (Hare, Reference Hare2017). Because lower levels of reactive aggression favor closer and more frequent contacts between individuals (this in turn facilitating the creation of larger groups that are less prone to fragmentation), this mechanism could be seen as part of the “behavioral solution” advocated for by Dunbar. Nonetheless, it can be also regarded as part of his “cognitive solution”. One main reason is that, neurobiologically, the attenuation of reactive aggression responses is achieved by increasing the control of selected cortical areas, particularly the prefrontal cortex, over the subcortical components of the circuit of aggression, including the striatum (Lischinsky & Lin, Reference Lischinsky and Lin2020). This mechanism thus partially overlaps with the brain circuits supporting advanced social cognitions, since, as noted by Dunbar, the prefrontal cortex is important for mentalizing abilities as well, which are a key aspect of our social brain. Interestingly, the mechanism controlling reactive aggression also partially overlaps with the circuits that regulate abilities that are relevant for bonding, like vocalization or language. This is because the striatum also contributes to such (see Murphy, Hoshi, & Benítez-Burraco, Reference Murphy, Hoshi and Benítez-Burraco2022 for a detailed discussion). A piece of evidence supporting this view is that domesticated animals, in which stress signalling in the striatum is attenuated, exhibit more variable vocalizations than their wild conspecifics (O’Rourke et al., Reference O’Rourke, Martins, Asano, Tachibana, Okanoya and Boeckx2021). Likewise, moving now to the human self-domestication hypothesis, it has been proposed that selected changes in the brain circuits highlighted above might account for our sophisticated syntactic abilities, and more generally, for our cross-modal thinking, which are both at the core of our advanced linguistic, pragmatic, and narrative capacities (see Benítez-Burraco & Progovac, Reference Benítez-Burraco and Progovac2021; Benítez-Burraco, Ferretti, & Progovac, Reference Benítez-Burraco, Ferretti and Progovac2021 for details). The consideration of human cognitive disorders in particular, conditions impacting on our social brain, provides additional support to the view above. In autism spectrum disorders (ASD), high reactive aggression co-occurs with high stress levels, but also with reduced socialization abilities, specifically, an impairment of some of the high order cognitive skills highlighted by Dunbar as part of his Social Brain Hypothesis, most notably the ability to understand others’ intentions (mentalizing), but also linguistic and communicative capacities (see Benítez-Burraco, Lattanzi, & Murphy, Reference Benítez-Burraco, Lattanzi and Murphy2016 for an overview).

To finish, I wish to argue that the dopamine system could serve as a potential linker of the two mechanisms contributing to increase group cohesion in primates, namely, the grooming-dependent mechanism and the reactive aggression-dependent mechanism. Accordingly, as noted by Dunbar, the endorphin system acts in tandem with the dopamine system. But it also happens that attenuated stress responses, as found in species showing low levels of reactive aggression, also impact on dopamine signaling (through glutamatergic activity), further promoting some of the behavioral and cognitive modifications highlighted above, particularly, more variable vocalizations (see O’Rourke et al., Reference O’Rourke, Martins, Asano, Tachibana, Okanoya and Boeckx2021 for details). Compared to other primates, humans show, specifically, increased dopaminergic innervation of fronto-striatal circuits involved in episodic memory retrieval (Clos, Bunzeck, & Sommer, Reference Clos, Bunzeck and Sommer2019), and speech and language learning (Raghanti et al., Reference Raghanti, Edler, Stephenson, Wilson, Hopkins, Ely, Erwin, Jacobs, Hof and Sherwood2016). These dopamine-dependent changes can be thus related to our advanced mental travel abilities and our linguistic capabilities, respectively, and should be of interest to Dunbar’s model, in view of the fact that storytelling is one key grooming-like activity promoting group complexity.

In conclusion, Dunbar’s model certainly captures and explains in sound and novel ways how larger and more cohesive primate groups are formed and maintained, but it could be improved with the consideration of other mechanisms also promoting cohesion, specifically, the reduction in reactive aggression responses, considering the links and overlaps that exist between the neurobiology of grooming and the neurobiology of aggression.