Moral Unknowns: An Exploration of Intellectual Humility and Categorizations of Normative Uncertainty

Abstract

Understanding how people make decisions under uncertainty is a vast field of study, whereas investigating how we make decisions under fundamental moral uncertainty—to decide how to decide what we ought to do—has not received a comparable level of attention. In other fields such as psychology and statistics, uncertainty has been divided into two types: epistemic uncertainty and aleatory uncertainty. Aleatory uncertainty arises from the inherently unpredictable nature of the system under study, whereas epistemic uncertainty is uncertainty due to a lack of knowledge or information. In this study, we applied this bifurcation of uncertainty to the realm of moral uncertainty and investigated how people understand moral uncertainty from an epistemic and aleatory perspective. Additionally, we explored the relationship between intellectual humility—a key individual difference measure that influences how people make judgements—and people’s understanding of moral uncertainty. Our findings indicate that people generally interpret moral uncertainty from a more epistemic lens (i.e., they see moral uncertainty as uncertainty that can be reduced with more information or knowledge). In addition, we found evidence that higher levels of intellectual humility, specifically an openness to belief revision and ego-intellect separation, predict a more epistemic understanding of moral uncertainty. Such findings have valuable implications not only when considering the relationship between an aleatory-epistemic lens and various metaethical positions regarding moral uncertainty, but also for the applications of this division of uncertainty in machine learning on ethical decision-making in artificial moral agents. Keywords: moral uncertainty, decision-making, epistemic and aleatory uncertainty, intellectual humility, belief revision