Loading field map
Why place two or more historical trajectories side by side? The practice of comparison promises to reveal patterns that a single case cannot show—why some societies industrialised and others did not, why revolutions broke out in some places but not others, how different systems of rule produced different forms of inequality. Yet comparison also carries risks: it can flatten the internal texture of each case, impose categories that belong to one society onto another, and treat nations or civilisations as self-contained units that can be lined up like specimens. The history of comparative history as a subfield is the story of how historians have wrestled with these promises and risks, producing a sequence of frameworks that have narrowed, expanded, and sometimes fundamentally redefined what it means to compare.
The first major framework to place comparison at the centre of historical method emerged not from a single manifesto but from a journal. The Annales School, founded in 1929 by Marc Bloch and Lucien Febvre, broke decisively with the event-focused political history that dominated French universities. Bloch and Febvre argued that historians should study deep structures—demographic regimes, landholding patterns, mentalities—that changed slowly and shaped the possibilities within which individuals acted. This commitment to the longue durée made comparison almost unavoidable: if you wanted to know why serfdom took different forms in different parts of medieval Europe, you had to place those regions side by side.
Bloch's own work provided the model. In Feudal Society (1939–1940) and in his programmatic essay "A Contribution towards a Comparative History of European Societies" (1928), he distinguished between two kinds of comparison. The first, which he called "generalising comparison," looked for regularities across societies that shared similar structures—for example, comparing feudal institutions in France, England, and Germany to identify a common type. The second, "historical comparison," examined neighbouring societies that influenced each other, tracing how institutions and practices crossed borders. Both approaches required the historian to define units of analysis carefully and to resist the temptation to treat each society as a unique, incomparable case.
Annales historians also pioneered serial quantitative history—counting wills, grain prices, or baptism records over long stretches of time—as a way to make comparison systematic. By the 1960s, the school's influence had spread well beyond France, shaping the work of historians in Latin America, Eastern Europe, and the United States. Yet the Annales framework had limits. Its focus on long-term structures made it difficult to explain sudden transformations such as revolutions or state breakdowns. And its preference for regional or civilisational units left the nation-state largely intact as the container of comparison, even as the school's rhetoric challenged political history.
By the 1960s and 1970s, a new generation of scholars—many of them sociologists rather than historians—began to push comparison in a more explicitly explanatory direction. Comparative Historical Sociology (CHS) took the Annales interest in large-scale structures and added a sharpened concern with causal inference. The central question became: how can we use comparisons to test explanations of major historical outcomes?
CHS drew heavily on John Stuart Mill's methods of agreement and difference. In the method of agreement, the researcher identifies two or more cases that share the same outcome and looks for a single factor that all have in common—that factor is the likely cause. In the method of difference, the researcher compares cases that differ in outcome and looks for a factor that is present in one and absent in the other. Barrington Moore Jr.'s Social Origins of Dictatorship and Democracy (1966) used a version of this logic to argue that the balance of power between landlords and peasants determined whether a society became democratic, fascist, or communist. Theda Skocpol's States and Social Revolutions (1979) compared France, Russia, and China to argue that revolutions occurred when international pressures combined with agrarian class conflict and state breakdown.
Skocpol's 1980 article "The Uses of Comparative History in Macrosocial Inquiry" codified the CHS approach. She distinguished three logics of comparison: the "contrast-oriented" logic, which highlights what is distinctive about each case; the "parallel comparative" logic, which tests causal hypotheses across similar cases; and the "macro-causal" logic, which uses Mill's methods to establish necessary and sufficient conditions. Process tracing—the detailed reconstruction of causal chains within each case—was added as a complementary technique to ensure that correlations identified across cases actually reflected real historical mechanisms.
CHS narrowed the comparative enterprise in two ways. First, it privileged cases that could be treated as independent of one another, making it difficult to study interactions, borrowings, or entanglements between societies. Second, it tended to treat nation-states as the natural unit of comparison, even when the processes under study crossed borders. These limitations did not prevent CHS from becoming institutionally dominant: it remains a thriving tradition in sociology and political science, and its methods are taught in graduate programmes across the social sciences. But within history departments, the bounded-unit assumption increasingly came under fire.
The 1990s brought a framework that did not reject comparison outright but fundamentally reframed what was being compared. Connected History, developed by historians such as Sanjay Subrahmanyam and Serge Gruzinski, argued that the units historians had been comparing—civilisations, nations, regions—were not pre-existing containers but products of the very connections that comparison was supposed to study. Instead of placing two societies side by side and asking how they differed, Connected History traced the movements of people, goods, ideas, and institutions across boundaries, showing how those movements constituted the societies themselves.
Subrahmanyam's work on early modern Eurasia, for example, showed that millenarian movements in India, Iran, and Europe were not parallel phenomena that could be compared as separate cases. They were connected through trade routes, diplomatic networks, and the circulation of texts. To compare them as isolated units would miss the very thing that made them historically significant. Connected History thus preserved the comparative impulse—it still placed different trajectories in relation to each other—but it replaced the logic of side-by-side comparison with a logic of entanglement.
This shift had consequences for method. Connected historians often worked at smaller scales than CHS or the Annales School, following a single commodity, a family network, or a religious text across multiple regions. They drew on archival research in several languages and paid close attention to the asymmetries of power that shaped who could move and who could not. The framework did not offer a single causal logic like Mill's methods; instead, it offered a way of seeing that made the bounded-unit assumption of earlier comparison visible as a limitation.
Just as Connected History was gaining traction, a group of French and German historians—Michael Werner and Bénédicte Zimmermann most prominently—proposed an even more radical rethinking of comparison. Histoire Croisée, or "entangled history," shared Connected History's suspicion of bounded units, but it went further by interrogating the categories that historians themselves bring to comparison.
Werner and Zimmermann argued that comparison inevitably imposes categories—"religion," "state," "economy"—that have their own histories and are themselves products of the entanglements being studied. When a historian compares "French" and "German" nationalism, for example, the very concepts of France and Germany as distinct national entities were shaped by centuries of cross-border interaction. To treat them as pre-existing units is to naturalise what should be explained. Histoire Croisée therefore calls for a reflexive approach: the historian must examine how the categories of analysis are constituted through the same historical processes that the comparison aims to illuminate.
This reflexivity extends to the historian's own position. Werner and Zimmermann insisted that the researcher's institutional location, disciplinary training, and national context shape what questions seem worth asking and what comparisons seem natural. A French historian comparing French and German labour movements will see different things than a German historian doing the same comparison, not because one is more objective but because the categories of comparison are embedded in different national historiographies. Histoire Croisée does not abandon comparison; it multiplies the levels at which comparison must operate, adding a critical self-awareness that earlier frameworks lacked.
These four frameworks do not form a simple succession in which each replaces the last. The Annales School declined as an organised movement after the 1980s, but its core commitments—the longue durée, interdisciplinary collaboration, the study of mentalities—have been absorbed into the broader practice of social and cultural history. CHS remains institutionally strong, especially in sociology and political science, where its causal-inference toolkit is taught as a standard method. Connected History and Histoire Croisée have reshaped how many historians think about comparison, but they have not displaced CHS; rather, they have created a division of labour.
Today, the leading frameworks agree on several points. All four reject the old practice of treating nation-states as natural, self-evident units of comparison. All four insist that comparison must be problem-driven rather than case-driven: the research question should determine what is compared and at what scale. And all four recognise that comparison is not a neutral technique but a historically situated practice with political implications.
Where they disagree is on the goal of comparison. CHS aims at causal explanation: it wants to identify the conditions that produce revolutions, wars, or economic development. Connected History and Histoire Croisée are more sceptical of causal claims, arguing that the entanglement of cases makes it difficult to isolate independent variables. They prefer to trace processes, show connections, and expose the constructedness of categories. The Annales School, in its classic form, sits somewhere in between: it sought structural regularities but was less interested in the kind of formal causal testing that CHS demands.
This pluralism is likely to persist. CHS provides tools that are indispensable for answering certain kinds of questions—why some countries industrialised and others did not, why some states collapsed and others survived. Connected History and Histoire Croisée provide tools for questions about circulation, translation, and the formation of categories themselves. The most sophisticated comparative work today draws on multiple frameworks, using CHS-style causal logic for some parts of the analysis and entanglement-sensitive methods for others. The subfield has not settled on a single method, and it probably never will. What holds it together is the conviction that placing histories in relation to each other—whether through structural comparison, causal inference, or the tracing of connections—remains one of the most powerful ways to understand how the world came to be the way it is.