The modern scientific study of war relies primarily on correlational studies aimed to uncover the recurring patterns of actions, events, and conditions associated with interstate conflict. The goal is an explanatory theory of war grounded in reproducible evidence, free of the subjective biases that have plagued traditional analysis. But to date the myriad findings within the correlational project remain unintegrated and the promise of a scientific breakthrough unfulfilled. This article suggests that the problem of nonintegration in the study of war is rooted at least partly in an overly restrictive epistemology that cannot ground causal theory. The argument summarizes an understanding of integrative causal reasoning drawn from some of the natural sciences, and suggests that future research into the question of war be aimed to develop such causal theory. Without some such vision of integrative reasoning, this analysis implies, the correlational project is unlikely to transcend its current state of empirical fragmentation.