Jones argues that kinship and language are similar in their discrete, combinatorial structure, and that these similarities point toward “general principles of cognition or communication.” Although he does not discuss evolution explicitly, he suggests that the shared mechanisms underlying kinship classifications and language may have evolved from the same cognitive precursor. Recent studies of nonhuman primates support this hypothesis. Specifically, they demonstrate that both the recognition and classification of individuals and the recognition and classification of vocalizations constitute discrete combinatorial systems that interact closely. One system maps onto the other, suggesting that they share underlying cognitive mechanisms. The close link between recognition and classification of individuals and vocalizations supports the view that, during human evolution, the cognitive mechanisms underlying kinship classifications and those underlying language shared a common precursor.

Although the physical features and vocalizations of different animals vary continuously, free-ranging monkeys and apes recognize each other as individuals, both by appearance and by voice. This in itself constitutes a very simple, discretely coded system of knowledge that is probably pervasive among group-living animals. Moreover, from this relatively simple starting point, monkeys, apes, and perhaps many other species make further, more elaborate classifications of individuals. Many monkeys, for example, recognize other animals' dominance ranks. They also distinguish pairs of individuals who have close behavioral associations from those whose interactions are less frequent and predictable (reviewed in Cheney & Seyfarth Reference Cheney and Seyfarth2007). In monkeys, close bonds typically manifest in frequent grooming, support in aggressive coalitions, tolerance at feeding sites, and other behaviors. They may persist for many years. In baboons and macaques, for example, males disperse but females remain in their natal group throughout their lives, and close behavioral associations are most common among matrilineal kin – mothers, daughters, and sisters – who almost always form such bonds; but strong bonds may also occur among less closely related females, particularly when these individuals have no close kin present in the group (Schino Reference Schino2001; Silk et al. Reference Silk, Altmann and Alberts2006; Reference Silk, Beehner, Bergman, Crockford, Engh, Moscovice, Wittig, Seyfarth and Cheney2010). In chimpanzees, females disperse but males remain in their natal group throughout their lives, and close bonds are formed between males and, more rarely, between females. Such bonds manifest in grooming, coalitions, meat sharing, and participation in hunts (Langergraber et al. Reference Langergraber, Mitani and Vigilant2009; Mitani Reference Mitani, Kappeler and van Schaik2006; Muller & Mitani Reference Muller, Mitani, Slater, Rosenblatt, Snowdon, Roper and Naguib2005; Nishida Reference Nishida1983; Watts Reference Watts1998; Wittig Reference Wittig, Lonsdorf and Ross2010; Wittig & Boesch Reference Wittig and Boesch2003).

In baboons, experiments suggest that individuals not only recognize other animals' dominance ranks and kin relations but also combine this information to create a representation of social relations in which ranked individuals are grouped into families (Bergman et al. Reference Bergman, Beehner, Cheney and Seyfarth2003; Seyfarth et al. Reference Seyfarth, Cheney and Bergman2005; see also Penn et al. Reference Penn, Holyoak and Povinelli2008). Observational data on Japanese macaques yield similar results (Schino et al. Reference Schino, Tiddi and Polizzi di Sorrentino2006). The classification of individuals on the simultaneous basis of kinship and rank constitutes a conceptual structure in primates because, while the individual elements may change over time as animals are born, die, join or leave the group, the overall structure remains constant. Moreover, animals treat one another differently not just because of their own behavior, but also because of their membership in a particular kin group. For example, from the moment of its birth, baboons treat an infant that belongs to a high-ranking matriline differently from the way they treat an infant that belongs to a low-ranking matriline. This difference in behavior arises not because infants behave differently but because they belong to different kin groups. In this respect, nonhuman primates' recognition of kinship and rank can be thought of as a conventionalized system, based on discrete elements – individual identity, kinship, and rank – combined in a rule-governed manner (Cheney & Seyfarth Reference Cheney and Seyfarth2007).

Nonhuman primate vocalizations also constitute a very simple, discrete, combinatorial system of recognition and classification. Most calls are individually distinctive, and individual recognition by voice is common. Many call types are also used only in particular circumstances (reviewed in Seyfarth & Cheney Reference Seyfarth and Cheney2010). Baboons, for example, give acoustically different alarm calls only to particular predators, territorial calls only during encounters with other groups, screams only by lower-ranking to higher-ranking individuals, threat vocalizations only by higher-ranking individuals to lower-ranking individuals, and so on. Listeners, moreover, recognize these predictable relations. In the playback experiments cited above, where A was dominant to B, listeners showed little response to playback of the sequence “A threatens B and B screams” but responded strongly to the sequence “B threatens A and A screams” (Bergman et al. Reference Bergman, Beehner, Cheney and Seyfarth2003). In sum, just as nonhuman primates recognize individuals and classify them according to rank and matrilineal kinship, the same animals recognize and classify vocalizations according to caller identity, context, and the “rules” of call delivery.

Jones notes that in humans, we sometimes view kinship and language as separate cognitive domains. By contrast, among nonhuman primates the recognition of individuals, kin groups, and rank relations (on the one hand) and the meaning of vocalizations (on the other) are inextricably bound together. Both are discrete, combinatorial systems, and each influences the other. When young animals learn to recognize their matrilineal kin, they do so in part by learning to recognize these animals' vocalizations; conversely, a caller's identity, rank, and kin group membership help to determine the meaning of a call or a sequence of calls to a listener. When a baboon hears Sylvia giving threatening vocalizations to Hannah and Hannah screaming, the listener's interpretation of this sequence draws not only on her identification of the call types and the contexts in which they are given but also on her recognition of the callers, their ranks, and their matrilineal kin groups. In this respect, the cognitive processes involved in the recognition and classification of individuals overlap significantly with the cognitive processes involved in the recognition of call meaning. Such data offer indirect support for the hypothesis that, during the course of human evolution, the cognitive mechanisms underlying kinship classifications and those underlying language evolved from a common source.