Levine et al. present a novel framework for understanding moral cognition as an approximation of rational bargaining solutions. Their model integrates contractualist, consequentialist, and deontological approaches into a unified system, arguing that these modes of moral reasoning serve as heuristics that approximate ideal moral agreements under resource constraints. However, one of the central limitations of their account is its failure to engage with the problem of meta-uncertainty: the uncertainty individuals face not just about specific moral rules but about which moral framework itself is correct or should be applied in a given situation.

Meta-uncertainty refers to the problem of not knowing which moral principles should guide decision-making. In many real-world moral dilemmas, individuals are not merely deciding between different actions but between fundamentally different moral theories. Should one prioritize contractualist bargaining solutions, utilitarian welfare maximization, or deontological constraints? How should one decide between these frameworks when they yield conflicting prescriptions? One approach that addresses these questions is maximizing expected choice-worthiness (MEC), which suggests that individuals assign credence to different moral theories and choose actions based on their expected moral value across all plausible frameworks (MacAskill & Ord, Reference MacAskill and Ord2020; Kernohan, Reference Kernohan2021). In a case where one must decide whether to lie to save a life, a deontologist may prohibit lying under any circumstance, while a consequentialist might consider it not only permissible but morally required. MEC provides a structured way to resolve such conflicts. An individual might assign 60% credence to consequentialism and 40% to deontology and then choose the action with the highest expected moral value weighted across both theories. This does not resolve the disagreement by appealing to a single moral principle but rather rationalizes the decision under uncertainty. Such an approach enriches Levine et al.’s framework by offering a principled account of how people navigate situations in which heuristic approximations based on a single bargaining ideal break down. While Levine et al. interpret moral cognition as an efficiency-driven approximation of idealized rational agreement, incorporating MEC would address a broader epistemic challenge: how agents make decisions when the very basis of agreement is uncertain.

Beyond probabilistic moral reasoning, it is important to account for the meta-rules people use to select among moral frameworks in context. Individuals often employ different moral heuristics depending on the social or situational circumstances. Some may follow a principle of conservatism, deferring to historically entrenched moral norms. Others adopt a pluralist approach, selecting frameworks based on the domain of the dilemma (using, for instance, consequentialist reasoning in policy contexts and deontological principles in interpersonal ethics). Another strategy involves avoiding catastrophic moral errors by choosing actions that are unlikely to be deeply wrong under any reasonable moral theory.

While Levine et al. assume contractualist reasoning dominates moral cognition, empirical evidence shows that people flexibly shift between different heuristics rather than adhering to a single framework. Dual-process theories in moral psychology (Greene et al., Reference Greene, Brian Sommerville, Nystrom, Darley and Cohen2001; Cushman, Young, & Hauser, Reference Cushman, Young and Hauser2006; Brand, Reference Brand2016) show that deontological judgments are often fast, intuitive, and emotionally driven, whereas consequentialist judgments emerge through slower, more reflective reasoning. This cognitive architecture suggests that people switch between moral systems based on the type of reasoning they engage in, not because they are computing a single ideal bargain. Levine et al.’s model accommodates multiple heuristics but treats them as approximations of an underlying agreement-based logic. The empirical data instead imply that people use different systems depending on the type of dilemma, not merely to economize effort but to express distinct moral values.

Cross-cultural research further challenges the universality of a contractualist bargaining ideal. Some studies show that fairness norms and moral judgments vary significantly across societies (Bentahila, Fontaine, & Pennequin, Reference Bentahila, Fontaine and Pennequin2021; Graham et al., Reference Graham, Peter Meindl, Johnson and Zhang2016; Henrich, Heine, & Norenzayan, Reference Henrich, Heine and Norenzayan2010). In some cultures, people emphasize equality, while others prioritize need, hierarchy, or reciprocity. Other theories suggest that societies with stricter norms respond more punitively to violations, while looser cultures tolerate greater deviance, reinforcing the idea that default moral heuristics are shaped by local social ecology (Gelfand, Reference Gelfand2018). These patterns suggest that people fall back on locally shared norms when faced with uncertainty, a behavior more consistent with conservatism than with an internalized simulation of Nash-style bargaining. This variation shows that people draw on different moral systems depending on the situation, which sets the stage for moral pluralism as a central feature of cognition.

Moral pluralism illustrates how individuals and groups draw from multiple moral systems depending on context. Moral foundations theory identifies several distinct moral domains (such as care, loyalty, authority, and sanctity) that guide moral judgment, suggesting that people often invoke different foundations in different situations, adjusting their reasoning dynamically (Graham et al., Reference Graham, Haidt, Koleva, Motyl, Iyer, Wojcik and Ditto2013; Haidt, Reference Haidt2012). Relational models theory supports this pluralism by proposing that people use distinct relational schemas, such as communal sharing, authority ranking, and market pricing, to regulate behavior in different social settings (Fiske, Reference Fiske1992). Such theories suggest that moral cognition is modular and relationally grounded, rather than structured by an overarching agreement-based ideal.

If Levine et al. aim to naturally understand our moral minds, they must account for the lived complexity of moral decision making, where individuals constantly shift between normative frameworks – not because they miscompute a single ideal, but because moral life requires navigating many partially incompatible goods. What moral pluralism offers, and what Levine et al.’s model could absorb more explicitly, is a recognition that moral cognition goes beyond bargaining logic and often involves adjudicating between fundamentally different kinds of reasons, grounded in social roles, affective commitments, or culturally embedded values. Incorporating this heterogeneity into Levine et al.’s model would not mean abandoning the insight that people approximate cooperative solutions under constraints. Rather, it would refine that insight by embedding it within a richer ecology of competing moral aims, different cognitive strategies, and flexible framework changes.