Explaining Necessary Truths

Knowing the truth is rarely enough---we also ask ``why?'' While much is known about how we explain contingent and probabilistic facts, far less is known about we explain necessary truths, such as those in mathematics, where the outcomes follow by logical deduction rather than causation. We present a framework, inspired by computational complexity, to make sense of how this happens: because cognition is resource-bounded, people prefer explanations that (1) reduce the cost of discovering the conclusion or (2) mark points of cognitive load where contemplation concentrates. In a series of three experiments with n=838 participants using SAT problems, we show how these motives serve as strong and significant predictors of explanatory preferences in both purely abstract logic problems (Study One), and in problems where the underlying logic is presented in the style of classic detective stories (Studies Two and Three, the latter preregistered). When story framings are used, preferences for causal drivers appear, but do not displace the the computational ones. Our results demonstrate that cognitive resource limitations are central to explanation-making; they can both coexist with causal explanation and can yield satisfying reasons-why judgments even in the absence of cause.