Richerson et al. “do not see how any of the alternatives to CGS can easily account for the institutionalized cooperation that characterizes human societies” (target article, sect. 7, para. 2). While they make a solid case for the role of cultural group selection (CGS) in human evolution and history, we think it is the evolution of human cooperation that made the tribal social instincts and CGS possible in the first place, and that this occurred much earlier than the authors think. To explain, we refer to Weingarten and Chisholm's (Reference Weingarten and Chisholm2009) use of Bowlby's (Reference Bowlby1946) insight that “the psychological problem of ensuring persistent co-operative behaviour” (p. 62) in groups was solved by the evolution of the capacity of infants to develop, first, an attachment to (a powerful emotional valuation of) mothers and others, then, as they mature, to groups of others, group leaders, and group “policy” (i.e., belief/value system; Bowlby made specific reference to the cohesion, cooperation, and altruism of combat units in WWII).

From the perspective of attachment theory – and infants – everyone's first group, tribe, society, institution, ethnolinguistic unit, and culture is the mother–infant group. (By mother we mean “mother and/or others who act like mothers.”) Everyone's first experience of cooperation is in the resolution (or not) of parent–offspring conflict (Chisholm Reference Chisholm, Sterelny and Fitness2003). Because selection operates only on phenotypes, and phenotypes must develop before they are exposed to selection, we are skeptical of models of the evolution of cooperation that attribute fitness benefits to adults only. How did our adult ancestors learn to cooperate? What motivated them to cooperate?

Richerson et al. have little use for what they call “trigger hypotheses” for the evolution of cooperation because they are difficult to test and “even the most innatist theories [of language, for example] imagine a culture-led gene-culture coevolutionary process.” We have little use for them because they are all about the behavior of adults – but do imagine a mother–infant culture-led coevolutionary process that set the stage for CGS. Like Richerson et al., we do not think that kin selection and reciprocity are sufficient to explain large-scale cooperation, but again, think they can explain where our cooperative motivations came from and how they were exposed to selection – because many of the “mechanisms that maintain intergroup variation” are already part of the attachment process. To wit, infants are capable of accurate, rapid, social learning; infants conform their behavior to their mothers' behavior; the infant–mother group is a moral community, based on trust (security) and norms of interaction, and infants and mothers punish each other for “deviant” behavior with resistance, displays of anger, withdrawal, and so forth; infants learn disproportionately from their “prestigious,” all-powerful, all-resources-controlling mothers; and infants engage in “dances and similar motor rituals” with their mothers (Beebe & Lachmann Reference Beebe and Lachmann2014; Trevarthen Reference Trevarthen2007; Tronick Reference Tronick2007). Moreover, oxytocin and subcortical brain systems subserve both infant and adult attachment behavior (Lane et al. Reference Lane, Luminet, Rimé, Gross, de Timary and Mikolajczak2013; Nelson & Panksepp Reference Nelson and Panksepp1998; Schore Reference Schore, Narvaez, Panksepp, Schore and Gleason2013; Strathearn Reference Strathearn2011) and trust in experimental games played by adults (Baumgartner et al. Reference Baumgartner, Heinrichs, Vonlanthen, Fischbacher and Fehr2008; Kosfeld et al. Reference Kosfeld, Heinrichs, Zak, Fischbacher and Fehr2005). The basic mammalian social emotions (Panksepp Reference Panksepp2011) provide the utility function – emotional costs and benefits – guiding human choice, by infants in the attachment process and adults in cooperation with each other.

Because there is no point in thinking unless one already has something of value to think about, the basic mammalian, “bottom-up,” subcortical emotions have also been implicated in the development of “top-down,” neocortical thinking (Damasio Reference Damasio1994; Reference Damasio1999), of which theory mind is an early example. Theory of mind – and related constructs like intersubjectivity (Trevarthen Reference Trevarthen2011) and “time travel” (Suddendorf Reference Suddendorf2013) – enables infants to learn through (not just from) mothers by inferring (thinking about) their mothers' beliefs, desires, and intentions/actions in order to “choose” their own next action on the basis of their innate value system, the basic mammalian social emotions. Theory of mind is the beginning of culture – shared or “we” understanding/intentionality (Tomasello et al. Reference Tomasello, Carpenter, Call, Behne and Moll2005). It also varies as a function of attachment security (Fonagy & Target Reference Fonagy and Target1997; Fonagy et al. Reference Fonagy, Gergely and Target2007; Meins et al. Reference Meins, Fernyhough, Wainwright, Das Gupta, Fradley and Tuckey2002). As Hobson (Reference Hobson2004, p. 94) put it, “symbolizing, language and thought are possible only because of the nature of the emotional connection between one person and another, and because of each person's involvement with a shared world.” The cultural origin of human cognition is in the co-construction of meaning by mother and infant. We therefore partly disagree with Richerson et al.'s statement that “Perhaps the earliest cultural norms merely solidified the bonds of kinship and reciprocity that were evolving through participation in systems of cooperative breeding” (sect. 2.2, para. 2). (Again, how did kin and cooperative breeders learn to cooperate? What made them want to?) We think the earliest cultural norms evolved through mother–infant cooperation.

This brings us to “docility,” one of Richerson et al.'s “trigger hypotheses”: “Docile individuals more inclined to conform to norms would find it easier to enter larger, more norm-bound groups” (sect. 2.2, para. 3). If “docile” means less fearful and aggressive, we entirely agree, for two reasons. First, secure attachment is associated with less fear and aggression (Cassidy Reference Cassidy, Cassidy and Shaver2008; Fearon et al. Reference Fearon, Bakermans-Kranenburg, van IJzendoorn, Lapsley and Roisman2010). Second, the Russian farm fox studies (Belyaev et al. Reference Belyaev, Plyusnina and Trut1985; Trut Reference Trut1999) show that artificial selection for “tameness” (reduced fear/aggression) succeeded, not because “genes for” tameness were selected, but because of the delayed development of fearful/aggressive behavior in tame strains compared to non-selected, wild strains. As a correlated by-product, the “docile” strains also showed increased social-cognitive skills (Hare et al. Reference Hare, Plyusnina, Ignacio, Schepina, Stepika, Wrangham and Trut2005). Weingarten and Chisholm (Reference Weingarten and Chisholm2009) proposed that the evolution of delayed development in humans might have had the same effect, potentially explaining how human adults came to be (relatively) docile.

In conclusion, we believe that the tribal social instincts and CGS were important very early in human evolution as our infant ancestors evolved the enhanced social-cognitive skills (theory of mind) that enabled them to “bargain and negotiate” their way through parent–offspring conflict. The outcome, everyone's first culture – mother–infant culture – repeated in every mother–infant group for hundreds of thousands of years, resulted in culture-led, gene-culture coevolution, and CGS. We are thus inclined to agree with Ayala (Reference Ayala2010): Complex cumulative culture may well be an exaptation.