Social scientists study many types of justice, which they find to have important empirical effects on human behavior and policy. These scientists, however, rarely offer satisfying theories as to why concern for justice ought to influence behavior in the first place, and their peculiar renditions of justice often muddle rather than clarify our understanding of justice, overall. I propose a single definition of justice, hinging on an emergent, variable desert. Whatever the context, justice is the rewarding of desert. Who deserves what and why, conversely, depends critically on context and will change with people, place, and time. Desert, in short, is a social institution that communities erect around multiple-equilibria problems concerning the distribution of socioeconomic resources and responsibilities. As an institution, statements expressing desert and its reward can be codified in the standard Institutional Grammar, thereby prescribing, demanding, or forbidding certain actions, with built-in incentives to conform. Desert, therefore, provides a theoretical grounding for justice’s influence on human behavior. Additionally, centering justice on desert clarifies the oft-abused language surrounding justice, and bridges otherwise distinct conceptions of justice now prevalent in academic literature.