Comparing Networks of Regime Transitions in Ancient Greece

This paper introduces a novel network approach to political analysis, focusing on transitions in political systems. Drawing inspiration from graph theory, we conceptualize political regime changes as transition sequences within a finite state space, where each state represents a distinct political regime. This perspective builds on the millennia-old tradition of analyzing cyclical regime changes (anacyclosis) as explored by thinkers like Plato, Aristotle, and Polybius, and enriches it with mathematical rigor and empirical data. Graph theory provides precise definitions of motifs (like paths, walks, circuits, and circles), offering a robust foundation for studying regime trajectories. Despite the prevalence of terms like "conjunctures", "trajectories," and "bifurcations" in political science, these concepts often lack mathematical clarity. Our approach addresses this gap, formalizing political developments as sequences of transitions between regime states—e.g., shifts from monarchy to tyranny or democracy—captured in transition networks. Empirically, we draw on a comprehensive dataset of regime changes in ancient Greek city-states compiled by a Danish research group (Hansen & Nielsen, eds., 2004). Our analysis focuses on 46 city-states with at least two regime transitions, and further examining longer trajectories. Using Aristotle’s typology of regime forms (based on the number and quality of rulers) and formal concepts from graph theory, we integrate spatial dimensions into our analysis. Additionally, we adapt Charles Tilly’s framework of regime trajectories, which maps transitions onto a two-dimensional space defined by government capacity and democratization, while addressing the challenges of imprecise temporal data in the Danish dataset. From a theoretical perspective, we explore a range of approaches to interpret transition networks, including conservative theories of political decay, progressive theories of democratization, and equilibrium-seeking evolutionary processes. We also test the law-like nature of regime transitions, revisiting long-standing debates about cyclical versus linear development in political systems. Our paper represents the first network analysis of this historical dataset and contributes to a broader understanding of political regime changes. Beyond ancient Greek city-states, our findings suggest broader applications for studying democratization and de-democratization processes. Notably, we respond to Philippe Schmitter’s critique of unidirectional theories of democratic development by illustrating how complex, multi-directional pathways can be systematically analyzed as transition graphs. By integrating state space modeling, graph theory, and empirical historical data, this paper bridges conceptual gaps in political science and offers a versatile framework for analyzing political transitions.