In virtually every democratic nation in the world, political representation is defined by where citizens live. In the United States, for example, Congressional Districts are drawn every 10 years as lines on a map. Why do democratic governments define political representation this way? Are territorial electoral constituencies commensurate with basic principles of democratic legitimacy? And why might our commitments to these principles lead us to endorse a radical alternative: randomly assigning citizens to permanent, single-member electoral constituencies that each looks like the nation they collectively represent? Using the case of the founding period of the United States as an illustration, and drawing from classic sources in Western political theory, this book describes the conceptual, historical, and normative features of the electoral constituency. As an institution conceptually separate from the casting of votes, the electoral constituency is little studied. Its historical origins are often incorrectly described. And as a normative matter, the constituency is almost completely ignored. Raising these conceptual, historical and normative issues, the argument culminates with a novel thought experiment of imagining how politics might change under randomized, permanent, national electoral constituencies. By focusing on how citizens are formally defined for the purpose of political representation, The Concept of Constituency thus offers a novel approach to the central problems of political representation, democratic legitimacy, and institutional design.