The variety of topics studied in the field of international relations contrasts sharply with the relatively anemic theoretical framework that IR scholars use to study those topics. This pattern is different from, say, biology or astrophysics, where evolution and gravity provide powerful overarching theoretical guides to inquiry. In International Relations (IR), there is instead a vast proliferation of explanatory variables that “matter” for specific questions, often demonstrated by their statistical significance in relation to a dependent variable. David Lake described this pattern as “midlevel theory,” and celebrated it as an improvement over what had become a tired debate in IR over the “isms,” referring to realism, liberalism, constructivism, and other paradigms.1 One virtue of midlevel theory is that it facilitated a shift from IR’s traditional focus on the causes of war to a broader set of topics, including human rights, energy, migration, norms, and much else.
A significant disadvantage of the field’s theoretical eclecticism, however, is that it leads to an intellectual blizzard: an excess of proposed variables and concepts. Computers and generative AI might thrive on hundreds of variables, but the human brain does not. Humans are what psychologists call “cognitive misers,” looking for efficient ways to understand problems.2 True, institutions matter in IR. Interests matter, too. Ideas matter. Perceptions matter. Nuclear weapons matter. Oil matters. Regime type matters. What is needed is not more variables that “matter” in some statistical sense, but a better way of integrating those variables into a meaningful understanding of world affairs.