People imagine agreements with others to decide whether an action is morally acceptable, rather than negotiating every decision in real time. This matters because contractualist moral theories—which hold that right actions are those people would agree to—face a practical problem: we cannot literally negotiate every moral choice we make. The research proposes that moral cognition works around this limitation by using three resource-rational shortcuts: imagining what others would agree to, recycling past agreements for new situations, and switching between effortful reasoning and quick heuristics depending on the stakes. The project tests these ideas through vignette experiments and multiplayer games, then builds AI-based decision aids grounded in contractualist principles. If successful, this research could reshape how we understand moral judgment—not as a set of fixed rules or emotional reactions, but as a flexible, resource-efficient negotiation process. The AI decision aids might eventually help people navigate complex ethical trade-offs in settings like healthcare allocation, policy design, or organisational governance. However, the work is primarily fundamental science: it aims to clarify the cognitive foundations of contractualist morality, with practical applications as a longer-term possibility.
View original technical description
Agreements are a fundamental part of human morality: people make promises, ask for consent, draw up contracts, and "talk it out" when disagreements arise. In philosophy, contractualist moral theories—which view actions as morally acceptable when they would be agreed upon—are deeply concerned with mutually beneficial arrangements. In moral psychology there is now growing interest in the role that agreement plays in moral judgment and decision making. For complex problems, however, practical constraints on individual cognition and social coordination pose an immediate challenge for psychological contractualist theories—we obviously cannot come to literal agreement, interpersonally, on everything. This proposal addresses this challenge by developing a resource-rational account of contractualist morality. It proposes that moral cognition can be understood as approximating the outcome of actual negotiation by making rational use of limited resources. First, people often substitute anticipated agreements (i.e., imagined ones) for actual ones. Second, people often recycle past agreements, generalizing them to novel circumstances, even if their fit to those circumstances may be less than ideal. Third, people can switch between cognitively effortful agreement-based reasoning, and cognitively efficient heuristics, depending on the importance of accurate moral judgments and decisions. This proposal develops these ideas theoretically and tests them empirically using manipulations of resource-rational factors (e.g., stakes, complexity, time pressure) in vignette-based experiments and multiplayer interactive games. It then builds on cutting-edge AI techniques to develop contractualist decision aids. Non-NSF funded collaborators: Nick Chater, Hossam Zeitoun, Behavioural Science Group, Warwick Business School, University of Warwick.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know