The target article proposes an elegant account of social cognition as reasoning about the roles people occupy, rather than only their mental states. The authors view this “institutional stance” as an evolutionarily endowed system that operates from infancy. Two features of the developmental trajectory of norm-based reasoning, however, raise the possibility of alternate origins and greater developmental change than the current proposal predicts. Rather than reflecting an early-emerging, domain-specific system, early norm-based reasoning may originate in domain-general representations of kinds with normatively connected features, with institutional reasoning emerging later as children develop richer representations of individuals and their mental states.

First, the weighting of normative and mental-state information shifts across development – young children show a surprising bias toward norms. For example, young children (e.g., ages 4–5) predict that agents will follow norms instead of their personal preferences (Berndt & Heller, Reference Berndt and Heller1986; Chalik, Rivera, & Rhodes, Reference Chalik, Rivera and Rhodes2014; Kalish & Shiverick, Reference Kalish and Shiverick2004), and expect norms more than preferences to generalize across groups and be stable over time (Kalish, Reference Kalish2002, Reference Kalish2012; Kalish & Lawson, Reference Kalish and Lawson2008; Kalish & Shiverick, Reference Kalish and Shiverick2004). They also misremember behavior driven by mental states as instead reflecting norms (Kalish, Reference Kalish2012). By around age 7, children instead prioritize mental states in each of these cases. These systematic developmental changes are not clearly specified within the proposed framework, which instead emphasizes early emergence and stability of both institutional and mental-state reasoning.

Second, children’s early norm-based reasoning is not restricted to the social domain. Young children extend normative reasoning to animal categories, responding that animals should display their characteristic features (Foster-Hanson et al., Reference Foster-Hanson, Roberts, Gelman and Rhodes2021; Haward et al., Reference Haward, Wagner, Carey and Prasada2018). They also view animals that most clearly exemplify these normative features as the best examples of their kinds (Foster-Hanson & Rhodes, Reference Foster-Hanson and Rhodes2019a). As in the social domain, these tendencies weaken with age, as children increasingly incorporate individuating information.

Domain-generality, along with developmental changes in the weighting of normative information, are both consistent with a general, early-emerging bias to represent kinds with characteristic, normatively connected features (Prasada & Dillingham, Reference Prasada and Dillingham2006; Reference Prasada and Dillingham2009), with a developmental shift toward more robust representations of individuals. Such a general shift is well-documented. Young children preferentially remember, choose to learn about, and generalize based on kinds more than individuals, with a shift across development toward more consideration of within-category variation and of individuals as distinct from the categories to which they belong (Brandone, Reference Brandone2017; Cimpian & Erickson, Reference Cimpian and Erickson2012; Cimpian & Park, Reference Cimpian and Park2014; Gelman, Sánchez Tapia, & Leslie, Reference Gelman, Sánchez Tapia and Leslie2016; Gülgöz & Gelman, Reference Gülgöz and Gelman2015; Leslie & Gelman, Reference Leslie and Gelman2012; Rhodes, Brickman, & Gelman, Reference Rhodes, Brickman and Gelman2008; Rhodes, Gelman, & Brickman, Reference Rhodes, Gelman and Brickman2008; Riggs, Kalish, & Alibali, Reference Riggs, Kalish and Alibali2014).

From this perspective, early norm-based reasoning may not yet reflect institutional reasoning in the sense described by the authors, but instead broader features of how children reason about kinds. A key distinction, as the authors note, is that in kind-based reasoning, normatively connected features are properties of the category members themselves, whereas in institutional reasoning, such features arise from occupying a role. A critical test, then, is whether children treat norms as intrinsic properties of people or as contingent on rolesFootnote ¹ .

Although direct evidence on this question is limited, existing findings are suggestive. Young children often treat role-based obligations as identity-defining, viewing individuals who occupy certain roles as being category members “deep down” (Foster-Hanson & Rhodes, Reference Foster-Hanson and Rhodes2019b). In addition, children tend to conflate norms and desires, assuming that people want to fulfill their obligations (Costanzo, Grumet, & Brehm, Reference Costanzo, Grumet and Brehm1974; Reference Costanzo, Grumet and BrehmKalish, Reference Kalish1998; Kalish & Shiverick, Reference Kalish and Shiverick2004; Riggs & Fast, Reference Riggs and Fast2025). Only later do children understand that norm compliance can diverge from individual preference, and respond that only those who genuinely care about fulfilling a role are members of the category “deep down.” Together, these findings suggest that young children may initially represent norms as intrinsic to category membership rather than as contingent on roles.

Thus, perhaps young children begin with domain-general kind representations that support normative expectations. Over time, they develop richer representations of individuals and come to understand norms as arising from roles rather than as an intrinsic part of category membership. One way to reconcile this developmental account with the target article is to interpret early norm-based reasoning as proto-institutional cognition. On this view, young children represent categories with normative expectations but lack the role-filler independence and structured coordination that characterize mature institutional reasoning. Ontogeny may therefore parallel the phylogenetic progression proposed, with fully flexible and generative institutional reasoning emerging later in development.

A central question is how this transition occurs. One possibility, which could present a challenge to the proposed framework, is that it depends in part on the development of mental-state reasoning. Understanding that people sometimes act counter to their own preferences, and that different people, with different mental states, can occupy the same role and behave similarly, may help children represent roles as abstract positions that constrain behavior independently of the people who occupy them. Without thinking about mental states, role and person may remain fused. Consistent with this possibility, the ability to predict behavior that conflicts with category norms is related to theory-of-mind development (Chalik et al., Reference Chalik, Rivera and Rhodes2014). When children explicitly represent mental states, deviations from norms become informative: the misalignment between a person’s preferences and the category demands makes the category visible as something operating outside of individual membership. In this way, representing mental states might enable a shift from a kind of representation where category membership reflects something deep and internal to a role representation where the role is an external position that people can occupy, resist, or leave. We view this type of interplay between mental state and norm-based reasoning in the development of institutional thought as consistent with current developmental evidence, but a departure from the authors’ proposal, as we understand it.

Taken together, these considerations suggest that institutional reasoning may emerge through the interaction of domain-general kind representations and developing capacities for representing individuals and their mental states. This view retains the authors’ emphasis on norm-based reasoning as central to social cognition, but suggests that full-fledged institutional reasoning may characterize a mature endpoint of social cognition rather than its starting point.