The claim that humans possess a “core” system of social knowledge raises questions about its evolution. Either human core knowledge arose de novo – neither Thomas nor Fiske (Reference Fiske1992) supports this view – or it should be found in other species and linked to fitness. Data from nonhuman primates support this view. The data are not essential to Thomas’ argument, but they strengthen it in two respects: they suggest why knowledge of social relationships is “core,” and they raise questions about the evolution of social knowledge in different societies.

Nonhuman primates behave differently with different individuals, often forming close social bonds for months or even years (Seyfarth & Cheney, Reference Seyfarth and Cheney2012). Primates not only have differentiated relationships, but also they recognize them in others. In species with linear dominance ranks, individuals recognize both their own rank and the ranks of others (Cheney, Seyfarth, & Silk, Reference Cheney, Seyfarth and Silk1995a). Recognizing one’s own rank could arise simply because individuals distinguish, from personal experience, “dominant to me” from “subordinate to me.” Recognizing other animals’ ranks, however, is cognitively more interesting because it requires an individual to organize her knowledge from the observation of interactions in which she is not involved.

Primates have expectations about how others will behave and, as with humans, these expectations depend on properties of relationships. When female vervet monkeys and baboons hear the distress cry of an immature, they look toward the immature’s mother (Cheney & Seyfarth, Reference Cheney and Seyfarth1980, Reference Cheney and Seyfarth1999). When two capuchin monkeys exchange aggression and solicit support from an ally, they preferentially solicit allies who rank higher than their opponent and have a social relationship with the solicitor that is closer than their relationship with the opponent (Perry, Barrett, & Manson, Reference Perry, Barrett and Manson2004, Reference Perry, Manson, Muniz, Gros-Louis and Vigilant2008). After aggression, many primates “reconcile” with a victim by behaving in a friendly way toward her (Cords, Reference Cords1992). Baboons reconcile by grunting (Silk, Cheney, & Seyfarth, Reference Silk, Cheney and Seyfarth1996). Grunts increase the likelihood that the victim will interact in a friendly way toward her former aggressor (Cheney, Seyfarth, & Silk, Reference Cheney, Seyfarth and Silk1995b; Cheney & Seyfarth, Reference Cheney and Seyfarth1997). Moreover, grunts from a close relative of the aggressor can also function as “reconciliation-by-proxy:” victims treat the grunt as reconciliation if it comes from a close relative of the aggressor, but not if it comes from an individual belonging to a different matriline (Wittig et al., Reference Wittig, Crockford, Seyfarth and Cheney2007a). The perception of other animals’ alliances follows a similar pattern (Wittig et al., Reference Wittig, Crockford, Wikberg, Seyfarth and Cheney2007b, Reference Wittig, Crockford, Langergraber and Zuberbuhler2014).

In baboon society, males leave the group where they were born, whereas females remain in this group throughout their lives, forming close bonds with their mothers and matrilineal sisters. Adult females can be ranked in a linear dominance hierarchy that determines priority of access to resources. From birth, daughters acquire ranks adjacent to those of their mothers. Rank relations among females are generally stable over time, with few reversals occurring either within or between families. Ranks among males change often.

Baboons’ knowledge of other animals’ relationships is embedded in their knowledge of society. In addition to recognizing other animals’ ranks and kinship, individuals respond more strongly to evidence of a between-family rank reversal than to evidence of a within-family reversal. They act as if they classify females simultaneously according to both kinship and rank (Bergman et al., Reference Bergman, Beehner, Cheney and Seyfarth2003; Schino, Tiddi, & Polizzi di Sorrentino, Reference Schino, Tiddi and Polizzi di Sorrentino2006). Similarly, although the relative ranks of males change often, individuals act as if they “update” the rank order when listening to competitive displays (Kitchen et al., Reference Kitchen, Cheney and Seyfarth2005), recruiting allies (Silk, Reference Silk1999), or monitoring sexual partners (Crockford et al., Reference Crockford, Wittig, Seyfarth and Cheney2007).

These results are difficult to explain without assuming that baboons perceive others not just as a collection of individuals, or even a collection of dyads, but as an assembly of relationships organized according to an intuitive theory – an interconnected system of knowledge (Thomas, 2025). This knowledge is structured using implicit, abstract categories. The categories are implicit because they are not fully conscious, well-developed theories that the baboons can describe; instead, through processes we do not yet understand, baboons observe other animals and generate expectations about how each will behave in different contexts (Cheney & Seyfarth, Reference Cheney and Seyfarth2007:118).

Social categories are abstract because they cannot be reduced to any one or even a few perceptual attributes. High-ranking baboon females are not older or larger than low-ranking females, nor do they live in larger kin groups. Family members do not always look alike, sound alike, or share distinctive physical features. Yet none of this variation affects category membership (Seyfarth & Cheney, Reference Seyfarth, Cheney, Margolis and Lawrence2015). Moreover, the classification of individuals into categories persists even when the individuals within them change. From the moment an infant baboon is born, she is treated as a member of a high- or low-ranking matriline; from the moment a new male arrives in the group, he acquires a place in the male dominance hierarchy, with all the transitive implications that this implies. For baboons, social categories organize knowledge regardless of the individuals involved.