The article examines the methodological possibilities and the inherent limits of the structural and the institutional approaches to the study of social inequality. The author proceeds from the thesis that the opposition between them, as it is ordinarily presented, is false, since the two approaches answer different questions — the first explains why a given distribution persists and reproduces itself, the second explains why distributions differ across societies and how they change — but that their relation is not one of simple complementarity either, since each possesses a characteristic failure that the other cannot repair from outside. Employing comparative-analytical, systemic and logical-methodological analysis, the study demonstrates that the structural approach tends towards functionalist tautology and underdetermines mechanism, as the classical critique of the functional theory of stratification established, while the institutional approach encounters the problem of endogeneity, since institutions are themselves products of the distribution of power they are invoked to explain. It is shown that the most productive methodological strategy is the mechanism-based one developed by C. Tilly, whose categorical pairs and the four mechanisms associated with them are neither purely structural nor purely institutional but organisational, and that the duality of structure formulated by A. Giddens supplies its ontological warrant. The operational consequences for research design are determined.