Plural Causes

Abstract Causal selection is the process underlying our intuition that an outcome happened because of a given event, or that an event is the cause of an outcome. When a forest catches fire after a lightning strike, for example, people tend to say that the lightning bolt was the cause of the fire, not mentioning the presence of oxygen in the air, although they are well aware that the latter was no less indispensable for the fire to occur. The extant literature on causal selection has so far operated on the implicit premise that the only relevant variables for causal selection are individual variables, corresponding to distinct nodes in the relevant network of causes. Ours is the first systematic study of plural causes in the context of causal selection. First, we establish by means of two behavioral experiments the psychological reality and non-triviality of plural causes, ruling out potential deflationary explanations. Second, we show that state-of-the-art models of causal selection based on counterfactual dependence can be extended to make non-trivial predictions about plural causes consistent with our experimental findings. Third, we show that surprising logical properties of plurals in natural language interpretation might be found in causal reasoning with plural causes.