Levine et al. argue that cognitive strategies such as following rules, using cached welfare standards, and virtual bargaining are lower-cognitive-cost strategies for achieving the same coordination of social actions, and feelings of moral obligation, that could otherwise be achieved at much higher cognitive cost by performing actual bargaining in the world. We are sympathetic to the authors’ proposal. In our commentary, we consider another resource rational strategy at the heart of human sociality, beyond those considered in the target article: delegation of bargaining power to legitimate representatives and institutions. By doing so, we extend resource rational contractualism and tie it to other apparently distinct literatures.

In place of individually bargaining for social contracts, people sometimes delegate bargaining responsibility to another person to bargain on their behalf. Delegation is a lower-cognitive-cost strategy than bargaining on one’s own behalf (Bar-Gill & Sunstein, Reference Bar-Gill and Sunstein2015). It allows people to replace the cognitive demands of full negotiation with a much simpler choice of the delegate. Moreover, this choice of delegate can be made once and then generalize to new social coordination problems (for example, with elected representatives who bargain for laws that govern all domains of social life). In choosing a delegate to give bargaining and decision power on one’s behalf, people intuitively prefer individuals perceived to be competent, unselfish, impartial, and benevolent (Levi et al., Reference Levi, Sacks and Tyler2009; Mayer et al., Reference Mayer, Davis and Schoorman1995; Saxe, Reference Saxe2022; Tsai et al., Reference Tsai, Trinh and Liu2022; Tyler et al., Reference Tyler, Goff and MacCoun2015). A competent delegate with benevolent and impartial motives can be trusted to agree to bargains that adequately reflect the principal’s interests.

Once a delegate has been given bargaining power, they can reach negotiated agreements or make decisions that generate the same coordination of social actions and feelings of moral obligation as other examples of resource rational moral cognition (André et al., Reference André, Fitouchi, Debove and Baumard2022). When a leader or authority has been delegated power by a legitimate process, people feel obligated to follow their rulings even against their immediate self-interest (Baldassarri & Grossman, Reference Baldassarri and Grossman2011; Dickson, Gordon & Huber, Reference Dickson, Gordon and Huber2022; Kogler et al., Reference Kogler, Olsen, Kirchler, Batrancea and Nichita2023; Murphy, Reference Murphy2017) and expect others to comply with these rulings even if there is no threat of punishment for disobedience (Margoni, Baillargeon & Surian, Reference Margoni, Baillargeon and Surian2018). These normative expectations facilitate the effectiveness of contracts formed by delegates, for instance in public goods contexts.

Conceptualizing delegation and legitimacy as a form of resource rational moral reasoning highlights open questions and challenges for future theoretical and empirical work. First, how and under what circumstances do people choose to individually negotiate solutions themselves versus to delegate that responsibility to a representative? This choice introduces an accuracy-effort tradeoff comparable to the one discussed in the target article. Following rules and using cached welfare standards typically involves sacrificing accuracy in order to minimize effort. Likewise, delegation decisions involve a trade-off and can be prone to error. People may err in delegating, when they would be better off negotiating for themselves (e.g., if the delegate is insufficiently competent or impartial); but also, in negotiating for themselves when they would be better off delegating (e.g., when a truly expert and benevolent delegate is available; Bar-Gill & Sunstein, Reference Bar-Gill and Sunstein2015).

Second, even when individuals delegate negotiation power, the delegation remains at least partially contingent and open to re-evaluation. Delegates earn trust by showing that they are continuously responsive to the voice and interests of the principals (Hirschman, Reference Hirschman1970; Sunshine & Tyler, Reference Sunshine and Tyler2003). People may reject the rules and agreements arrived at by delegates, especially if the outcomes are at odds with their individual moral convictions (Skitka et al., Reference Skitka, Bauman and Lytle2009). In other words, there is a continual process of joint inference: beliefs about the delegate’s competence and impartiality influence attitudes toward their negotiated agreements, and attitudes toward those agreements influence the perceived legitimacy of the delegate. Therefore, cognitive scientists need to investigate how these inferences interact in shaping people’s interpretation and acceptance of the delegated contracts. We suggest that formal models of moral cognition should be extended to incorporate inferences about the negotiated contracts and motivations of those with delegated power. (In a related context, we found that observers can use Bayesian inverse planning to simultaneously update beliefs about an authority’s legitimacy and their punitive policies, Radkani et al. (Reference Radkani, Tenenbaum and Saxe2025).

Finally, delegation to an authority or institution (as a real or a legal person) is inevitably tied to, if not sometimes equivalent to, delegation to a process. Self- or delegated negotiations are often embedded in a set of collectively created and agreed-upon processes that constrain the choice sets and regulate decision-making, and people have strong intuitions about which processes are fair, impartial, and cognitively feasible (Thomas et al., Reference Thomas, Mitchell, Sumner, Terrizzi, Piff and Sarnecka2022; Tyran & Feld, Reference Tyran and Feld2006). For example, within the justice system, negotiating a fair solution on the defendant’s behalf by a lawyer amounts to following the final decision made by a jury process. Critically, these processes cannot be reduced to action standards because they often lead to specific unique solutions in a new context (most trivially random lottery, but also democratic discourse and voting). With the advent and integration of AI agents into social institutions at various scales, these processes and algorithms are becoming more widespread and simultaneously more complex, implicit, and opaque (Farrell et al., Reference Farrell, Gopnik, Shalizi and Evans2025).

Thus, it is important for a resource rational theory of morality to characterize the cognitive representations related to delegation, and how those interact with the other cognitive processes described in the target article.