Central to Pietraszewski's computational theory of groups is a set of group-constitutive roles within four triadic primitives. To be considered part of a group, he argues that individuals are obligated to occupy such roles during conflict. Such building blocks for representing groups, if part of humans' evolved psychology, could be present ontogenetically and perhaps phylogenetically. Although some data from the developmental and biological sciences support Pietraszewski's theory, other data raises questions about whether “group membership applies to all of the triadic primitives” (sect. 7, sect. 2) and whether similar behavioral expectations hold across various ecological conditions and interactions. In reviewing this work, we open the door to discuss a broader set of conceptual primitives that support reasoning about groups.

Several studies have examined infants' capacity to reason about social groups, revealing that they are capable of making inferences about multi-agent conflict, and the types of behaviors that should be directed toward ingroup versus outgroup members (Pun, Birch, & Baron, Reference Pun, Birch and Baron2016, Reference Pun, Birch and Baron2021; Rhodes, Hetherington, Brink, & Wellman, Reference Rhodes, Hetherington, Brink and Wellman2015). For example, when 16-month-old infants witnessed a conflict between two agents from opposing groups, they were more surprised when these agents' social partners cooperated (instead of conflicted) with one another (Rhodes et al., Reference Rhodes, Hetherington, Brink and Wellman2015). This result is consistent with Pietraszewski's triadic primitive generalization, in which a conflict between two agents from opposing groups can be extended to another, uninvolved member of a group.

Research with infants as young as 9 months of age supports the early emergence of the primitives defense and alliance (Pun, Birch, & Baron, Reference Pun, Birch and Baron2021). Specifically, after watching two agents from opposing groups come into conflict, infants expected an ingroup member (that was not part of the initial conflict) to harm an outgroup member (by pushing them off the platform). Similar findings have been observed with nonhuman primates, children, and adults (e.g., De Dreu et al., Reference De Dreu, Gross, Méder, Giffin, Prochazkova, Krikeb and Columbus2016; Rhodes & Brickman, Reference Rhodes and Brickman2011; Rusch, Reference Rusch2013). This suggests that when defending the interests of the group, individuals may be obligated to help ingroup members (i.e., be allies), even if it requires harming outgroup members

Relatedly, by 6 months of age infants understand that there is “strength in numbers.” After observing two agents compete with one another, infants expected the agent with more group members to prevail over an agent with fewer group members (Pun et al., Reference Pun, Birch and Baron2016). To be able to make this inference, infants likely inferred that group members acted as allies, banding together to support one another against an opposing group. Consistent with this argument, recall that infants were more surprised when a group member did not provide aid to an ingroup member (Pun et al., Reference Pun, Birch and Baron2021).

Interestingly, even though infants did not witness “the initially uninvolved onlooker…behave” (sect. 5, para. 3) in Pun et al. (Reference Pun, Birch and Baron2016), they were still able to predict the outcome of a competition based on group size. Indeed, similar to many social species, from insects, to lions and nonhuman primates (Batchelor & Briffa, Reference Batchelor and Briffa2011; McComb, Packer, & Pusey, Reference McComb, Packer and Pusey1994; Wilson, Hauser, & Wrangham, Reference Wilson, Hauser and Wrangham2001), the capacity to predict the outcome of a competition based on the presence of group members is critical for survival and may reflect part of an evolved psychology that supports reasoning about social groups. Therefore, even if cues to group membership (e.g., spatial temporal cues and moving in synchrony) are considered ancillary according to Pietraszewski, they may be sufficient to activate group-constitutive roles of novel groups, even prior to extensive socialization.

Other studies with infants and nonhuman species suggest that variation in ecological conditions (e.g., competition, cooperation, resource acquisition, and predation) may activate different computations for ingroup and outgroup members, that ultimately influences behavior (e.g., Avilés, Reference Avilés2002; Bian, Sloane, & Baillargeon, Reference Bian, Sloane and Baillargeon2018; Bonner, Reference Bonner1982; Chapman & Teichroeb, Reference Chapman and Teichroeb2012; Krause, Ruxton, Ruxton, & Ruxton, Reference Krause, Ruxton, Ruxton and Ruxton2002; Lindstedt et al., Reference Lindstedt, Miettinen, Freitak, Ketola, López-Sepulcre, Mäntylä and Pakkanen2018). This raises the possibility that any computational theory of groups needs to account for the influence of these conditions. For example, 18-month old infants expect ingroup and outgroup members to be treated equally when resources are abundant. In contrast, ingroup members are expected to be prioritized when resources are limited (Bian et al., Reference Bian, Sloane and Baillargeon2018), suggesting that competition promotes ingroup favoritism. Relatedly, in the absence of conflict, 17–19-month-olds expect ingroup members to provide instrumental help to another ingroup member (but not an outgroup member) (Jin & Baillargeon, Reference Jin and Baillargeon2017).

Furthermore, ingroup loyalty is expected to be maintained such that behaviors that harm the group are not permissible (Rhodes et al., Reference Rhodes, Hetherington, Brink and Wellman2015, Rullo, Presaghi, & Livi, Reference Rullo, Presaghi and Livi2015; Ting, He, & Baillargeon, Reference Ting, He and Baillargeon2019). For example, when ingroup loyalty is violated (e.g., an ingroup member harms another ingroup member), infants in their first year of life expect ingroup members to refrain from helping the aggressor (as a form of punishment) (Ting et al., Reference Ting, He and Baillargeon2019). However, it is not clear whether committing this transgression leads infants to infer that this disloyal individual should be ostracized from the group. Future research may want to consider the possibility that additional primitives independently shape expectations about ingroups and outgroups.

Finally, research from the biological sciences has proposed that variation in ecological conditions effect the formation and maintenance of groups (Avilés, Fletcher, & Cutter, Reference Avilés, Fletcher and Cutter2004; Bonner, Reference Bonner1982; Chapman & Teichroeb, Reference Chapman and Teichroeb2012; Frank, Reference Frank2003). For example, in the species Dictyostelium discoideum (cellar slime mold), facing resource scarcity and starvation leads to the formation of groups. Ultimately, some of the aggregated cells sacrifice their own reproductive fitness for the survival of the group (Bonner, Reference Bonner1982). This demonstrates that the formation of the group and the roles individuals occupy can vary as a function of the environment (e.g., Avilés, Reference Avilés2002; Krause et al., Reference Krause, Ruxton, Ruxton and Ruxton2002). Given that these behaviors occur even at the cellular level, it may provide insight into the ways in which group-constitutive roles could be represented psychologically.

Together, research from the developmental and biological sciences provide some initial support for Pietraszewski's theory of groups. This study suggests that individuals may only be obligated to occupy the roles within the triadic primitives during a conflict, potentially because conflict may constrain the behaviors that are permissible for ingroup members to engage in. However, such expectations may not emerge simultaneously in development, even if all of them are part of an evolved psychology. Furthermore, variation in ecological conditions can influence how individuals within a group should behave toward ingroups and outgroups. For example, when no conflict is present, expectations of behavior may be less rigid; ingroup members may be permitted to behave more benevolently toward outgroup members, and outgroup harm may be less acceptable. Finally, because ingroup support is expected to be maintained, exploring how violating group-constitutive roles affects perceptions of group membership may help inform a more robust computational theory of groups.