Understanding and improving institutions may be the most important task of social science. They determine much about every society’s economic outcomes and the life prospects of every human being (Acemoglu, Johnson, & Robinson, Reference Acemoglu, Johnson, Robinson, Aghion and Durlauf2005). Some economists recognize that neoclassical economics lacks the tools needed to fully understand institutions, pointing toward cognitive science to fill explanatory gaps (North, Reference North2005). Developing a cognitive economics of institutions is thus a crucial, yet largely neglected, project (Johnson, Reference Johnson and Vialein press), of which the evolutionary and developmental underpinnings would be a cornerstone.

It was with great interest, therefore, that I read about Jara-Ettinger and Dunham (J&D)’s institutional stance. J&D define “institutions” as a “network of interconnected roles” or “representation of a graph, where each node represents a role and each edge represents a relation.” As examples, J&D mention universities, governments, committees, charities, homeowner associations, businesses (e.g., corporate offices, rental-car agencies, cafés), and clubs (e.g., sports teams, knitting groups, book clubs, sororities, study groups). I learned much from J&D’s theorizing about role-based social inference.

I am less sure I have learned where “institutions” come from, in the sense that economists and political scientists use the term. They usually define “institutions” as the “rules of the game” within which economic and political activity occurs (Crawford & Ostrom, Reference Crawford and Ostrom1995; Hogdson, Reference Hodgson2006; North, Reference North1990). Prototypical examples include property law, insurance markets, democratic accountability, separation of powers, rule of law, accounting standards, schedules of taxation, and judicial rules of evidence. (J&D’s sense of “institution” maps more closely to what institutional economists call “organizations” [North, Reference North1990] – a subset of “institutions” in the broad sense. “Organization” is probably closer to the vernacular use of “institution.”) The intricate combination of formal legislation and informal norms constituting these rules exerts pervasive but largely invisible effects on our lives.

Rules and roles are, of course, related. The rights and responsibilities entailed by roles follow from the rules associated with those roles (e.g., the role of property owner is defined by the rules governing what a person can do with their property), and formal rules gain their force in part from enforcement by individuals in roles imbued with authority (e.g., rules of evidence are enforced by judges). If rules and roles are two sides of the same coin, then perhaps the institutional stance gives us all we need to understand institutions in the economists’ sense.

For several reasons, however, I believe that rules are more fundamental than roles, both logically and (I suspect) cognitively.

First, roles greatly underdetermine the rules. For example, an individual filling the role of “judge” has the power to issue rulings about intellectual property. But this tells us little about the content of those laws, and it is those details that matter for economic outcomes (Friedman, Reference Friedman2001). Copyrights provide weak but lengthy protections to authors against copying specific text (literally construed), whereas patents provide strong but short protections to inventors against imitating an idea (broadly construed). The stronger protections on patents are economically important because they raise the returns to innovation for the inventor; their shorter terms are equally important because they (eventually) reduce the cost of the innovation for the user. Although the judge has the power to enforce these rules, she has only modest discretion. Both the cognitive content and economic rationale of patent law focus on the rules, not the roles.

Second, some rules are abstracted so far from those who “enforce” them that it is unnatural to analyze them in terms of roles. A key “meta-institution” is the rule of law – the principle that abstract and consistent rules, rather than the particulars of who occupies positions of power, determine how institutions work (Fukuyama, Reference Fukuyama2011). For societies with rule of law, the enforcer of the rules is secondary to the rules themselves, enhancing the primacy of rules over roles. Just as importantly, the meta-institution of the rule of law is so abstract that it is hard to pinpoint any individual role as its “enforcer.”

Third, because formal enforcement is costly, institutions can effectively guide society only when their enforcement is rarely required; this is why effective formal institutions are almost always built on top of informal norms that are largely self-enforcing. In this sense, the law has an existence independent of – and more fundamental than – the roles that enforce it (North, Reference North1990).

None of this undermines the empirical claims made by J&D, which are both plausible and interesting. I also agree with the broader point that institutions rely on sociocognitive adaptations that are psychologically prior to institutions themselves. Institutions are cultural technologies built on top of evolved instincts, yet they come to transcend them by taking on a life of their own via cultural evolution. Where I differ from J&D is that I doubt that role-based construal is the primary sociocognitive adaptation needed for this job.

Consider an analogy. Spoken language is a set of instincts that biologically evolved at least 100,000 years ago, whereas written language is a technology that culturally evolved around 5,000 years ago. Reading and writing co-opt not only our language instinct but also other innate capacities such as object recognition and motor control. We likely wouldn’t have written language if humans had not evolved bipedalism to free our hands – but we would not thereby identify bipedalism as the primary instinctual underpinning of writing. Interpreting behavior in terms of roles, similarly, is at best one of many instinctual underpinnings of institutions.

What would a more comprehensive catalog of such instinctual precursors to institutional technologies include? As J&D would likely agree, norm psychology, theory-of-mind, and indeed, social-role construal surely all contribute. So too do many of our sociocognitive adaptations for cooperation (e.g., reciprocity), intuitive theories (e.g., lay economics; Boyer & Petersen, Reference Boyer and Petersen2018), the domain-general capacity for abstract reasoning, and the domain-specific capacity for reasoning about rational actors (as in Jara-Ettinger et al., Reference Jara-Ettinger, Gweon, Schulz and Tenenbaum2016). To the extent that it makes sense to speak of an “institutional stance” that has rules, rather than roles, as its domain, it is more likely a patching together of distinct cognitive systems rather than a single adaptation.