“The question of how it is that humans can understand and coordinate infinitely varying and novel instances of meaningful social interactions is roughly parallel to the classical learnability question in theoretical linguistics – how it is that all children within a language community converge in learning the same grammar of their mother tongue, although the verbal stimuli they meet are impoverished, noisy, and highly variable. Ultimately, it amounts to asking how human social life is understandable, and hence possible.

The problem does not reduce to understanding that individuals are agents who act intentionally to fulfil their goals. Although by definition a social relationship takes place between intentional agents, knowing that several intentional agents are present does not specify what kind of social relationship they will have with one another. Pursuing the analogy with language acquisition, it is possible there is something analogous to universal grammar in the domain of social relations – a universal set of abstract kinds of social relations, and innate skeletal knowledge of these that guides learning, inference, and indeed experience in the social domain.”

The basic point of the paragraphs above (quoted from, p. 18) is remarkably similar to Jara-Ettinger & Dunham’s (hereafter, J-E&D) proposal that “a generative grammar that builds structured models of social collectives,” not just reasoning about the mental states of other individuals, is necessary for understanding the social world (J-E&D, p. 1). We whole-heartedly agree: Just as you know which sentences are correctly formed in your mother-tongue even without any explicit knowledge of formal syntax, so this shared social grammar enables you to intuit that whenever you are hungry you can eat food from the fridge in your family home (regardless who bought it) but not at the office, or that a boss can order her PA to fetch her coffee at work, but not Saturday night at the movies (J-E&D, p. 1). It also licenses inferences of the moral consequences and reasons for violating these social rules. To our minds, the basic, critical questions of any social grammar proposal are 1) what are its conceptual primitives and evolved fundamental rules, and 2) how are these recognized, instantiated, and functioning in coordinating social interaction – making social life understandable, guiding socialization and adaptive social strategies, and sustaining and regulating different kinds of social institutions. To these questions, the answers of J-E&D are unfortunately vague.

Seeking to understand the same interface between social mind and society that concerns the target article, relational models (Fiske, Reference Fiske1991, Reference Fiske and Haslam2004, Reference Fiske2025) and social dominance (Ho et al., Reference Ho, Kteily, Sheehy-Skeffington, Thomsen and Gawronski2025; Sidanius & Pratto, Reference Sidanius and Pratto1999) theories each build on classic social theory and decades of empirical social-scientific evidence. Fiske (Reference Fiske1991, Reference Fiske and Haslam2004, Reference Fiske2025) draws on vast records of ethnographic data to outline the evolved cues that universally signal and instantiate the fundamental relationships of communal unity, hierarchy, equality, and proportionality, complementing it with experimental evidence that people are cognitively encoded in terms of the basic kind of relationships one has with them (e.g., Fiske, Haslam, & Fiske, Reference Fiske, Haslam and Fiske1991). Sidanius and Pratto (Reference Sidanius and Pratto1999) argued that other groups are predominantly seen and treated not in terms of their specific identities but as a function of their position in the societal hierarchy and the basic, abstract hierarchical motives of the perceiver, a claim supported by thousands of research papers (Ho et al., Reference Ho, Kteily, Sheehy-Skeffington, Thomsen and Gawronski2025).

The Thomsen-Carey proposal synthesized these insights with the theory of core cognition (Carey, Reference Carey2009), arguing that structural relational models (e.g., communal/kin sharing and caring; hierarchy; equality) form part of an evolved and finite set of conceptual primitives for a coordinating, generative grammar that manifest even in the preverbal mind to make socialization possible, taking the perceptual input specified by Fiske’s ethnographic review (e.g., sharing bodily fluids/synchrony/close touch for communal clique structures; vertical/size/numerical transitive differentiation for hierarchy) (see also Ho et al., Reference Ho, Kteily, Sheehy-Skeffington, Thomsen and Gawronski2025; Sheehy-Skeffington & Thomsen, Reference Sheehy-Skeffington and Thomsen2020, Reference Sheehy-Skeffington and Thomsen2023; Thomsen et al, Reference Thomsen, Frankenhuis, Ingold-Smith and Carey2011).

The very point of these prior proposals is that the way in which infants and all other people understand any specific relationships and social interaction is precisely as concrete instantiations of basic and general, abstract, latent, and compositional relational forms and the coordination rules they entail: The structures of social life (as Fiske titled his Reference Fiske1991 book) scale and compose to societal institutions, not only in the way they are represented, but in the strategic and adaptive ways they are endorsed and navigated, thereby forming political psychology (Ho et al., Reference Ho, Kteily, Sheehy-Skeffington, Thomsen and Gawronski2025; Sidanius & Pratto, Reference Sidanius and Pratto1999; Thomsen, Reference Thomsen2010).

Surprisingly, J-E&D refer only briefly to this work with the claim that “rather than proposing a small number of innate relationship templates, we propose that even infants are able to flexibly construct all of these institutional models, and likely many more, via a particular type of generative grammar operating over roles.” (sect 5.3, para.3). Yet, the very point of a compositional, generative grammar is that it operates over some finite set of (likely innate, evolved) primitives. What J-E&D think these fundamental relations or roles are and how they function remains unclear. That they repeatedly turn to examples of communal care and hierarchy suggests that the notion of innate relational models (Fiske, Reference Fiske1991), as part of a compositional generative social grammar that scales up to society (Sheehy-Skeffington & Thomsen, Reference Sheehy-Skeffington and Thomsen2020, Reference Sheehy-Skeffington and Thomsen2023; Thomsen, Reference Thomsen2010, Reference Thomsen2020, Reference Thomsen2025), is more central to institutional understanding than the authors will allow.

Arguing that reasoning about mental beliefs/intentions and social structures/constraints is separate stance neglects the centrality of intersubjective rules for understanding self, others, and social institutions, in a manner reminiscent of methodological individualism, ignoring seminal work in the social sciences (e.g., Berger & Luckman, Reference Berger and Luckmann1966; Cooley, Reference Cooley1902; Garfinkel, Rawls, & Lemert, Reference Garfinkel, Rawls and Lemert2015; Goffman, Reference Goffman1959; Mead, Reference Mead1934; Vygotsky, Reference Vygotsky1978). Institutions and, we’d argue, any concrete social relationships are “structurated” (Giddens, Reference Giddens1984) through the repeated enactment of different relational roles they encompass. Arguing that a generative grammar builds structured models that constrain behavior begs the question of what its conceptually primitive relational/structural forms are that allow us to generate infinite rule- and role-based social coordination in the first place.