This article examines the central theory in the sociology of aging-disengagement theory. It asks why the last decade of research bearing on it has been so inconclusive. The answer, the author suggests, lies (a) in the overlooked flaws in its underlying logic-the escape clause problem, (b) in the overly inclusive nature of its central variables-the omnibus variable problem and (c) in the level of reality it selects for study-the assumption of meaning problem. These three problems appear not only in disengagement theory but in much of the later research bearing on it. In an attempt to avoid these problems, the author sketches an alternative theory. This theory redefines disengagement and proposes new structural determinants of it.