How to weigh lives. A computational model of moral judgment in multiple-outcome structures

When is it allowed to carry out an action that saves lives, but leads to the loss of others? While a minority of people may deny the permissibility of such actions categorically, most will probably say that the answer depends, among other factors, on the number of lives saved versus lives lost. Theories of moral reasoning acknowledge the importance of outcome trade-offs for moral judgments, but remain silent on the precise functional form of the psychological mechanism that determines their moral permissibility. An exception is Cohen and Ahn's (2016) subjective-utilitarian theory of moral judgment, but their model is currently limited to decisions in two-option life-and-death dilemmas. Our goal is to study other types of moral judgments in a larger set of cases. We propose a computational model based on sampling and integrating subjective utilities. Our model captures moral permissibility judgments about actions with multiple effects across a range of scenarios involving humans, animals, and plants, and is able to account for some response patterns that might otherwise be associated with deontological ethics. While our model can be embedded in a number of competing contemporary theories of moral reasoning, we argue that it would most fruitfully be combined with a causal model theory.