That language is stable despite the incentive to deceive proves that humans have solved the “central problem” for the evolution of communication (Maynard Smith & Harper, Reference Maynard Smith and Harper2003). We humans indeed show a cooperative attitude, which includes the a priori mutual expectation of cooperative intent between communicators (Grice, Reference Grice1989), and, as Heintz & Scott-Phillips (H&S-P) rightly emphasize, extends well beyond communication: It is also reflected in teaching, and in fact permeates almost all our social behavior and cognition (see also Burkart, Hrdy, & van Schaik, Reference Burkart, Hrdy and van Schaik2009).

According to H&S-P, our partner choice ecology can explain these developments (Fig. 1, top row) because the strong ecological reliance on cooperation makes it vital to be chosen by others as a cooperation partner. The latter can be maximized by engaging in reputation management, that is, advertising one's own cooperativeness or prosociality to potential partners, which according to H&S-P is the result of a highly complex cognitive mechanism, that is, Gricean communicative pragmatics. Once evolved in the context of partner choice, this mechanism then would have become available in all the contexts where it is conspicuous today, including language.

Figure 1.

Sequence of events proposed by H&S-P (top row) and in the current commentary (bottom row). The gray shading represents the socio-ecological background of our ancestors. Both proposals agree that language could only emerge once fundamental prosociality was established, but H&S-P argue it is the result of cognitively demanding Gricean communicative pragmatics (red arrow) whereas the bottom row points at comparative evidence (green arrow) that shows high allomaternal care and cooperative breeding directly facilitate the emergence of prosociality without Gricean communicative pragmatics as a precondition.

Among nonhuman primates, partner choice is likewise widespread but tellingly, it is not based on cognitive book-keeping mechanisms (Schino & Aureli, Reference Schino and Aureli2009) or reputation management (e.g., in chimpanzees: Engelmann, Herrmann, & Tomasello, Reference Engelmann, Herrmann and Tomasello2012), most likely because they lack language (van Schaik, Reference van Schaik2016). In fact, reputation management is strikingly absent even in marmoset monkeys, who are renowned for their high levels of cooperation and intentional prosociality (Burkart & van Schaik, Reference Burkart and van Schaik2020; sensu Townsend et al., Reference Townsend, Koski, Byrne, Slocombe, Bickel, Böckle and Manser2017) and do engage in partner choice (Brügger, Willems, & Burkart, Reference Brügger, Willems and Burkart2021). When adult marmoset helpers were alone with immatures from their group who were not their own, they would not stop helping but even increased their food sharing tendency (Brügger, Kappeler-Schmalzriedt, & Burkart, Reference Brügger, Kappeler-Schmalzriedt and Burkart2018). This clearly confirmed that their motives were genuine and strong proactively prosocial rather than instrumental and serving to manage their reputation.

The high prosociality of marmoset monkeys is thus clearly not a consequence of partner choice, pace H&S-P. Comparative data directly comparing prosociality in a group-service paradigm among a large number of primate species (Burkart et al., Reference Burkart, Allon, Amici, Fichtel, Finkenwirth, Heschl and van Schaik2014; Verspeek, van Leeuwen, Laméris, Staes, & Stevens, Reference Verspeek, van Leeuwen, Laméris, Staes and Stevens2022) show that prosociality is best explained by reliance on allomaternal care, or cooperative breeding (when individuals other than the parents significantly help rearing offspring). Humans are cooperative breeders too (Hrdy, Reference Hrdy2009), and when directly comparing them with the same group-service paradigm, they fall right on the nonhuman primate regression line (Burkart et al., Reference Burkart, Allon, Amici, Fichtel, Finkenwirth, Heschl and van Schaik2014). Prosociality in humans thus does not require a uniquely human mechanism, and the claim of a cognitively demanding explanation as in H&S-P appears unwarranted (which is a more widespread problem; Adriaense, Koski, Huber, & Lamm, Reference Adriaense, Koski, Huber and Lamm2020).

Based on this primate background, we therefore propose a more parsimonious sequence than advocated by H&S-P (Fig. 1, bottom row). When our ancestors started to engage in cooperative breeding, this was accompanied by an increase in proactive prosociality, consistent with the general primate pattern (Burkart et al., Reference Burkart, Allon, Amici, Fichtel, Finkenwirth, Heschl and van Schaik2014). This general prosocial attitude allowed the emergence of low-cost, honest cooperative signaling and thus paved the way for language evolution in our great apelike ancestors (Burkart et al., Reference Burkart, Hrdy and van Schaik2009, Reference Burkart, Guerreiro Martins, Miss and Zürcher2018, Reference Burkart, Adriaense, Brügger, Miss, Wierucka and van Schaik2022). Only once language was in place, however, did reputation management become necessary on a grand scale because without, only those directly involved will know that a specific partner was non-cooperative in the past. Crucially, reputation management is only needed when one's reputation of being a good or bad collaborator can actually spread widely across a large and loose social network. Language can provide exactly this, and only when there exists a risk that deceptive behavior can be broadly advertised through gossip to everyone will it reinforce a strong concern for reputation in all group members (see also van Schaik, Reference van Schaik2016, p. 331). In sum, only with language could a concern for reputation become strong enough to install a general prosocial attitude as suggested by H&S-P. However, because language per se requires such an attitude, which moreover can easily emerge in the context of cooperative breeding, this is unlikely.

Among cooperatively breeding monkeys (marmosets and tamarins), prosociality is most evident in their propensity to provide and offer food to others. However, and particularly important for the proposal above and language evolution in general, this propensity also extends toward sharing information (reviewed in Burkart, Guerreiro Martins, Miss, & Zürcher, Reference Burkart, Guerreiro Martins, Miss and Zürcher2018, Burkart et al., Reference Burkart, Adriaense, Brügger, Miss, Wierucka and van Schaik2022). For instance, frequent vocalizations function to provide information useful to others, in particular about predators and food (Brown, Almond, & van Bergen, Reference Brown, Almond and van Bergen2004), thus satisfying the definition of “expression” by H&S-P (target article, sect. 2, para. 1). Food-offering calls in tamarins are given for young immatures to offer food to them, but to older immatures only to indicate the location where the immatures can learn to extract food, thus engaging in teaching-like behavior (Rapaport, Reference Rapaport2011; Troisi, Hoppitt, Ruiz-Miranda, & Laland, Reference Troisi, Hoppitt, Ruiz-Miranda and Laland2018). In fact, several marmoset and tamarin species show remarkable sensitivity to the skill level of immatures when deciding whether to advertise the presence of food and share it with them (Dell'Mour, Range, & Huber, Reference Dell'Mour, Range and Huber2009; Humle & Snowdon, Reference Humle and Snowdon2008; Moura, Nunes, & Langguth, Reference Moura, Nunes and Langguth2010; Sehner, van Schaik, & Burkart, Reference Sehner, Van Schaik and Burkart2022; Snowdon & Roskos, Reference Snowdon and Roskos2017). Critically, such teaching-like scaffolding was also reported during the vocal development of immatures (Chow, Mitchell, & Miller, Reference Chow, Mitchell and Miller2015; Takahashi et al., Reference Takahashi, Fenley, Teramoto, Narayanan, Borjon, Holmes and Ghazanfar2015, Reference Takahashi, Liao and Ghazanfar2017).

Among adults, the cooperatively breeding marmosets are more likely than independently breeding primates such as capuchin monkeys or macaques to use gaze as coordination smoothers when engaged in joint action with a partner (Miss & Burkart, Reference Miss and Burkart2018; Miss, Meunier, & Burkart, Reference Miss, Meunier and Burkart2022). Although cooperatively breeding groups are closely related on average, immigration and emigration events are frequent and group membership, rather than mere relatedness, determines helping (De Oliveira Terciero, Willems, Arruda, Burkart, & Araujo, Reference De Oliveira Terciero, Willems, Arruda, Burkart and Araujo2022). Overall, it thus very much seems that to thrive in their socio-ecological niche of cooperative breeding, marmosets and tamarins have evolved a convergent interaction engine (Levinson, Reference Levinson, Enfield and Levinson2006) that resembles the one of humans in many relevant respects (Burkart et al., Reference Burkart, Adriaense, Brügger, Miss, Wierucka and van Schaik2022).

Our complementary account does not argue that partner choice and reputation management are not crucial in the human social ecology. However, the cooperative breeding of our ancestors offers a more plausible point of departure: Great ape-like organisms who were more prosocial than extant great apes and had a greater inclination toward information donation. Immatures growing up in such an interdependent ecology would have acquired much of the cognitive infrastructure described by H&S-P (Hrdy & Burkart, Reference Hrdy and Burkart2020) before reputation management evolved.