Ancient Tyranny and Modern Dictatorship

Abstract

This article traces the conceptual history of key terms used to describe and criticize bad political regimes, focusing on the displacement of “tyranny” by “dictatorship” and “authoritarianism.” Classical Greek thought understood tyranny primarily in terms of the character of rulers, whereas the modern idea of dictatorship emerged from a Roman conceptual framework that focused on authority and its legitimation. New problems of legitimation in the eighteenth and nineteenth centuries diminished the utility of the character-centric concept of tyranny and increased the fruitfulness of dictatorship for political analysis. The emergence of the modern state in the nineteenth century shaped the conceptual field by increasing the salience of problems concerning the appropriation or usurpation of sovereignty, the distortion of popular legitimation and accountability, and the incentives for submission to illegitimate orders. I conclude that the use of “authoritarianism” is likely to increase in prominence, but that retaining multiple regime concepts enriches analysis.

Like many other European languages, English uses words derived from classical Greco-Roman thought—e.g., “tyranny,” “despotism,” “autocracy,” and “dictatorship”—to describe “bad” political regimes.¹ While these terms are still in use, the central concept for these regimes in classical political thought was tyranny, whereas modern thought favors dictatorship and authoritarianism. Indeed, political science today operates within a conceptual landscape that contrasts “democracy” with “dictatorship” and “authoritarianism” as residual concepts for all forms of nondemocracy, and rarely refers to “tyranny” as a distinctive evil.

Recent examples include Jennifer Gandhi, who uses “dictatorship” preferentially, and explicitly defends the use of the term by looking at its historical usage,² and Steven Levitsky and Lucan A. Way, who prefer “authoritarianism” instead.³ But many use both terms interchangeably, including Barbara Geddes, Joseph Wright, and Erica Frantz,⁴ Milan Svolik,⁵ Steffen Kailitz,⁶ and Mike Alvarez, José Antonio Cheibub, Fernando Limongi, and Adam Przeworski.⁷ Julian G. Waller argues that authoritarianism has become, and should remain, the preferred term for the residual concept of “nondemocracy,” yet sometimes uses “dictatorship” interchangeably with “authoritarianism.”⁸ In the 351 pages of Juan Linz’s classic Totalitarian and Authoritarian Regimes, which discusses a variety of regime concepts, there is only a single incidental mention of “tyranny.”⁹

This seemingly unimportant linguistic shift represents a significant change in the frameworks we use to diagnose and evaluate political forms. Building on work on the history of these concepts by Andrew Arato,¹⁰ Markus J. Prutsch,¹¹ and Melvin Richter,¹² among others, I describe the roots of this change in the “age of revolution” in the eighteenth and nineteenth centuries, when new concerns with “popular sovereignty” led to a process of conceptual displacement that pushed the idea of tyranny away from its previously central place in theoretical reflection on what constituted a bad political regime. I summarize this as a movement from character to legitimation.

As I show, tyranny in classical political thought is always personalized. The individual tyrant is the central focus of theorists, and the badness of tyranny derives ultimately from the tyrant’s character rather than from his possession of unaccountable or illegitimate power. While questions of what we call “legitimacy” were not ignored, they appeared in the context of a concern with “just rule” that did not postulate a distinctive mechanism for the illegitimacy of tyranny beyond injustice (which other regimes could also display), and that therefore did not center illegitimacy as the key problem of political rule.

Because the modern idea of dictatorship emerged from a Roman conceptual framework that focused on authority and its legitimation independently of just rule and the character of rulers, it was better suited to describing and diagnosing the problems posed by political regimes since the nineteenth century than the earlier concept of tyranny. Thus, though “dictatorship” sometimes still shares connotations of personalized power with the classical concept of tyranny, the character of the dictator has receded in importance relative to the forms of his authority and the legitimation of his position, and “dictatorship,” “authoritarianism,” and related terms have come to be understood as deviations from forms of government based on popular legitimation.

Conceptual genealogy cannot tell us how we should use concepts, only how they have been used. It does not contribute to current debates over concept formation in political science, but it can draw attention to the grain of their construction, and thus to the kind of usages for which they are best suited. As I argue, the emergence of the modern state in the nineteenth century, with its rationalized legal order and dependence on certain understandings of “popular” legitimation, shaped the conceptual field of political science by increasing the salience of terms like “dictatorship” and “authoritarianism” that point to the appropriation or usurpation of sovereignty, the distortion of popular legitimation and accountability, and the incentives for submission to illegitimate orders. Hence the importance of forms of legitimation in our current regime concepts.¹³

But the kinds of nondemocratic regimes that political science primarily studies in the twenty-first century—regimes whose institutions increasingly resemble those of democracies—are less well described by the concept of dictatorship, with its connections to the assumption or usurpation of sovereignty in emergency conditions. Instead, I suggest they are best conceptualized in terms of certain violations of authority relationships whose normative standards are implicitly provided by conceptions of democracy, that is, as “authoritarian” forms. I am therefore in accord with Waller about the central place of authoritarianism in the contemporary study of political regimes.¹⁴ The study of these forms is thus likely to produce further differentiation within the concept of authoritarianism even as “dictatorship” remains important as an alternative conceptualization: authoritarianism “with adjectives” just reflects the fact that there are a multiplicity of ways in which political regimes fail to preserve the authority relations typical of democracy.¹⁵

My argument does not entail that the concept of “tyranny” is no longer used or useful. Indeed, tyranny in the classical sense—the personalized power of a ruler, exercised in his private interest—is still with us, and the word is still used frequently to describe multiple phenomena. The personalization of power remains a danger to political life, and the processes by which power becomes personalized have become an object of study for modern political science.¹⁶ But the shift from character to legitimation that I discuss here added new problems to the study of political regimes, and the emergence of new conceptual terminology allows us to better grasp their variety and distinctiveness, from the appropriation of sovereignty characteristic of “dictatorship” to the restricted forms of authority and accountability characteristic of “authoritarianism.”

I first establish that there has been a change in the usage of the political terms connected with defective political regimes. I draw freely on the history of concepts (Koselleck’s Begriffsgeschichte) as a methodological framework to argue that changes in usage represent changes in the salience of particular political problems.¹⁷ I then show that classical Greek political philosophy thought of tyranny primarily in terms of character, and in the third section discuss how first “despotism” and then “dictatorship” displaced it from its central role in discussions of bad political regimes from the eighteenth century onwards as new political problems of legitimation and sovereignty, demanding different conceptual tools, emerged. I trace more briefly the careers of “totalitarianism” and “authoritarianism” and indicate how each relates to the broader shift in ideas of legitimation during the nineteenth and twentieth centuries. I conclude that the multiplication of these concepts has enriched our grasp of political regimes but suggest that “authoritarianism” in particular is likely to increase in prominence given the current forms taken by nondemocratic political regimes.

Concepts as Problems

Tyranny, dictatorship, and similar concepts are political concepts, their meaning developed in political struggles. These are not just essentially contested, but deployed against opponents, deflected with adjectives to neutralize their stigma, and so on.¹⁸ By this I do not just mean that they are difficult to define, but that insofar as they have strong valences, they necessarily become instruments of coalition building, legitimization, and delegitimization, preventing meanings from stabilizing. Nevertheless, political concepts, and in particular regime concepts like those under study here, are also attempts at description.¹⁹ New terminology, or significant changes in meaning to existing terms, emerge when new situations call for new descriptions. Political regime concepts are often diagnostic: they represent attempts to describe what is wrong with a political situation, and the salience of the problem determines the salience of the concept.²⁰

Novel situations call for novel concepts, and novel concepts generate novel terminology.²¹ Accordingly, we should expect major events—the French and American Revolutions and their aftermath, the rise of Fascism and Communism—to be especially fertile for the development of new meanings, the introduction of new terminology, and the recrudescence of disputes over the application of existing terms: revolutions in language are reflections of revolutions in action. At such times, we may see a linguistic ferment where old concepts appear to lose their meaning and new terminology proliferates. Over time, concepts that are too specific to a given situation will tend to fall out of use (though they can also be revived when analogies to a new situation are found), whereas more generally applicable concepts can become more permanent features of the conceptual landscape, even as their valence remains contested.

Terms like “tyranny” may endure because the concentration of political power for private gain is not uncommon, and writers have found the idea of “tyranny” useful to diagnose it, even as they have repurposed the term in a variety of ways. Consider the “tyranny of the majority,” a concept which becomes increasingly important in the nineteenth century after the phrase was coined by John Adams in 1788.²² The idea of the “tyranny of the majority” also illustrates the fact that concepts occur in networks. It summarized a particular diagnostic of potential problems with the new republics of America and France in the late eighteenth and early nineteenth centuries, and thus required a concept of the “majority,” which in turn was part of a network of concepts including “representation” and “democracy.”

“Tyranny” nevertheless became less central within these networks during the nineteenth century. To be sure, “tyrant” and “tyranny” remain widely used terms in both scholarly and popular discourse. Their marginalization is relative rather than absolute. Figure 1 tracks the percentage of texts in the large Hathi Trust corpus of books mentioning a set of regime-specific terms related to tyranny.²³ “Tyrant” and “tyranny” are consistently the most widely used terms, indicating their staying power, but their usage peaked at the time of the French Revolution, and has since declined relative to other terms, while the frequency of “dictatorship” and “dictator” (as well as “authoritarianism” and “authoritarian”) has been consistently increasing since then.

Figure 1. Frequency of regime terms, 1650–2020, in the Hathi Trust digital library corpus (17 million volumes). NB: There are very few digitized volumes before 1800. Shaded areas represent 1787–1810 (the American and French Revolutions) and 1914–1945 (WWI and WWII).

These trends are more clearly visible when we look at word frequencies relative to “tyranny.” From Figure 2, we can roughly identify three periods of conceptual innovation: the American and French Revolutions, when “despotism” became important relative to “tyranny”; the middle of the nineteenth century, when “dictatorship” and “Caesarism” become key concepts to diagnose the pathologies of European politics; and the second half of the twentieth century, when “authoritarianism” and “totalitarianism” have their heyday. Throughout, “dictatorship” maintains its steady trajectory upwards, indicating its increasing centrality in the conceptual networks underpinning the modern understanding of political regimes.

Figure 2. Frequency of regime terms relative to “tyranny,” 1650–2020, in the Hathi Trust digital library corpus. Shaded areas represent 1787–1810 (the American and French Revolutions) and 1914–1945 (WWI and WWII). The red horizontal line represents the relative frequency of “tyranny.”

Further confirmation of the diminishing centrality of “tyranny” can be found in word frequency data from JSTOR journals, books, and chapters from 1900–2023 (in subjects political science, history, philosophy, law, and other social sciences—see Figure 3).²⁴ Here the trends are less marked, though we can see the increase in the usage of “dictatorship” in the interwar era, followed by a flat trend, as well as the larger increases in the usage of “authoritarianism” in the postwar era. Throughout, “tyranny” remains at a relatively steady level (about 3 percent of all works mention the term, comparable with the frequency of “dictatorship”), while other concepts become more central (“authoritarianism” most recently).

Figure 3. Frequency of regime terms 1900–2023, in articles, books, and chapters in JSTOR classified as political science, history, law, social sciences, and philosophy. Shaded areas represent WWI, WWII, and the end of the Cold War.

Although the term “tyranny” has not disappeared from our popular and scholarly vocabulary, it has disappeared from rigorous use in political science. As we saw above, major studies of political regimes today use “dictatorship,” “autocracy,” and “authoritarianism” and rarely mention “tyranny.”²⁵ The explicit rejection of a term like “tyranny” in scholarly work is found as early as the 1950s; in “Notes on the Theory of Dictatorship” Franz Neumann complained that “tyranny and despotism have no precise meaning,” and preferred to use “dictatorship” instead.²⁶ The literature on “totalitarianism” from the same period, while not altogether avoiding “tyranny,” developed new terms for this “novel form of government.”²⁷ Earlier in the twentieth century, Max Weber’s Economy and Society did not use the term except in passing, preferring to discuss regime forms with a variety of alternative concepts; and while the nineteenth century was more likely to use “tyranny,” it also saw the development of conceptual alternatives, including “despotism” and more short-lived terms like “Caesarism,” “Usurpation,” and “Bonapartism.”²⁸

Though the meaning of “tyranny” changed over time, initial meanings gave the concept a “diachronic thrust,” as Koselleck put it, against which conceptual innovators needed to work.²⁹ When attempts to fit the concept to new situations demanded too much adjustment, other concepts came to take its place. In the case of tyranny, its strongly individualized and moralized thrust was increasingly difficult to reconcile with the question of legitimacy and illegitimacy that came to occupy center stage for political thinkers after the early nineteenth century.

Tyranny in Classical Political Thought

As Victor Parker notes, “tyranny” originally appears to have been simply another term for monarchy in Greek cities in the seventh and sixth centuries BC, without a specifically negative connotation.³⁰ He argues that the earliest use of the term in literary texts is found in a fragment from the poet Archilocus (ca. 680–645 BCE),³¹ who stresses the wealth and power of the tyrant as something that stimulates envy and is conventionally desirable and does not clearly distinguish tyrant and monarch. Clearly negative usages are only attested beginning in the mid-sixth century; Solon’s poetry mentions tyranny as something bad that some people would nevertheless envy.³²

By the fifth century, however, the negative usage of tyranny is well established, and there is a robust “anti-tyrannical” ideology in Athens, as Kurt Raaflaub points out.³³ While the classical tragedians occasionally still used the term in a neutral way, Richard Seaford observes that they increasingly associated it with hubris and violence,³⁴ and Athenian democratic culture praised tyrannicide, as Josiah Ober argues.³⁵ Yet Athenians remained fascinated with the idea of the tyrant as someone who had great wealth and power. Tyranny, while collectively to be avoided, appeared as something individually desirable, both for the individual and for the polis as an actor in a world of poleis. ³⁶

At the same time, “tyranny” as a specific theoretical term with a negative connotation emerged from the efforts of Greek thinkers to classify and evaluate political arrangements: Are particular politeiai good or bad for people’s ability to live good lives? Though classifications varied, the general trend from Herodotus to Aristotle was to divide political arrangements in terms of the number of people who exercised power (one, a few, many, sometimes as a proxy for social classes like the poor or the rich), the purposes for which power was used (the common good vs. the private interest of the ruler), and the form in which it was wielded (according to law or arbitrarily). In Plato’s Statesman (301a6–c9), these distinctions led to a sixfold characterization of political arrangements, including three good forms (monarchy, aristocracy, and law-constrained democracy) and their corresponding bad forms (tyranny, oligarchy, and lawless democracy), a classification further systematized by Aristotle in book 3 of the Politics (1279a22–b10).

As I have argued elsewhere, these classifications presented ideal types which were often just a starting point for further analysis and evaluation, including the analysis of “mixed” forms.³⁷ Aristotle’s political sociology countenanced a variety of forms of democracy described in Politics 4, and later writers introduced other mixtures (though the general tendency was towards simplification). But while individual writers differed in their detailed evaluations of the merits or demerits of particular regimes, they all agreed that tyranny, understood as the government of a bad or unjust “monarch” (a single ruler), was the worst form—the rule of one person, for their own private good rather than the public good, and arbitrarily rather than by law. Both theoretical and popular conceptions of tyranny thus converged not only on its badness but also on its association with particular individuals and their desires.

But because tyranny appeared as the most extreme scope given to the desires of a single individual, it also appeared as potentially desirable in popular culture; ambitious individuals could find it very much worth pursuing, as the Republic and the Gorgias most famously show. Many classical writers thus sought to demonstrate not only that tyranny is bad for those who are ruled, but that it is bad for the rulers themselves. In Xenophon’s Hiero,³⁸ the poet Simonides and the tyrant Hiero discourse on whether the tyrant is happier than a private individual. Surprisingly, Hiero argues that the tyrant cannot be happy so long as he enslaves others to his will, while Simonides encourages him to rule more like a king than like a tyrant if he wants to be happy. And Plato’s Republic 9 is famous for its description of the tyrannical individual as someone who is dominated by his nonrational, lawless desires, and hence is unable to be truly happy (571aff.).

The tyrant is bad for those he rules because he treats them unjustly as slaves, subjecting them to his private desire. Hence the common opposition between tyranny and freedom, eleutheria, the status of free men. But here we also see the beginning of the parallel career of that other term for political pathology: the despotēs, the master of slaves, which Aristotle dismisses as a nonpolitical form of rule in Politics 1.1255b17ff., and associates with “Oriental” (non-Greek) rulers.³⁹ While tyranny is an individual deviation from just rule, despotism is a “system” thought to naturally emerge from the condition of natural slavery found among barbarians. And because tyranny was dependent on particular individuals rather than being a systemic result of impersonal causes like national character, the concept could entail a full-blown ideology of tyrannicide and resistance to tyrants. The idea that tyranny can be cured by “tyrannicide” continued to be regularly debated well into the Renaissance.⁴⁰

Portrayals of tyranny, while well aware of the tactics that tyrants used to maintain power (Aristotle, Politics 5.1313a33ff.; Xenophon, Hiero 6 on mercenaries), and of the overriding concern of the tyrant with its preservation, nevertheless tended to focus less on the mechanics of power than its moral effects: the disfigurement of desire, the loss of rational control, and the warping of the soul that made happiness impossible for both ruler and ruled. Moralists and philosophers aimed to depict tyranny as an unattractive prospect for potential tyrants by showing that being placed in a position that seemed to promise the satisfaction of all desire without constraint was ultimately self-defeating. These diagnostics were thus typically paired with a prescription of virtue as the cure for the unhappiness of tyranny: the ruler must rule in ways that are for the benefit of his subjects, not himself, with moderation and self-control, and in doing so he will indirectly benefit himself also (Aristotle, Politics 5.1314a35ff.; Xenophon, Hiero 9–11). The distinction between monarchy and tyranny was thus not institutional or systemic but characterological: the king was the tyrant become virtuous.

Moreover, despite some loose talk of “voluntary obedience” in some passages of Plato (Statesman 291e), the problem of tyranny is never primarily characterized as one of rule over “unwilling subjects” (Statesman 292a) but as one of rule against the common good. Voluntary obedience was pragmatically seen as a desideratum of rule, but not a measure of legitimacy. Similarly, though many writers take for granted that tyrants do not have proper right to their office as they typically achieve it by violence, their right to use power is never a subject of extended discussion. What distinguishes the king from the tyrant is the justice of his rule, not the fact that the king is more likely to be voluntarily accepted by the subjects or that he has a right to rule. Indeed, in the Platonic tradition the only legitimate title to rule is knowledge or virtue, not consent or hereditary right; and while Plato’s position on this point is especially extreme, no classical writer makes much of the distinction between the monarch’s right to rule and the tyrant’s illegitimate use of power. Contra Arato, I do not think Greek reflection on tyranny ever made much of “the problem of legitimacy.”⁴¹

Tyranny was not only associated with monarchy but was also often understood as a specific pathology of democracy. In Plato’s famous description of the cycle of regimes in the Republic (8.564ff.), the tyrannical individual emerges from the corruption of a form of democracy. Since democracy was understood as the direct government of “the many,” their poverty and lack of education were thought to make them especially susceptible to deceptive persuasion to support potential tyrants. The popular, demagogic leader was always a threat to democracy. But although much later writers—Tocqueville being perhaps the most famous—came to speak of the “tyranny of the majority” (the majority’s unconstrained use of its power for unjust purposes), tyranny was not uniquely identified with democracy or its destruction. Moreover, though Greek writers understood the idea of collective tyranny, Ivan Jordović shows that they conceived of it as a specific intensification of the rule of a small group, speaking sometimes of “oligarchy” or “dynasty.”⁴² Collective tyranny was not impersonal.

Contrast this “classical” understanding with Locke’s claim that tyranny “is the exercise of power beyond Right, which no Body can have a Right to. And this is making use of the Power anyone has in his hands; not for the good of those, who are under it, but for his own private separate advantage.”⁴³ While Locke still stresses that the tyrant uses his power for private advantage, the emphasis is now on the right to use power rather than on the specific individuals misusing power. One immediate consequence of this is that tyranny can be understood as a potential property of groups, not just individuals: “where-ever the Power that is put in any hands for the Government of the People, and the Preservation of their Properties, is applied to other ends, and made use of to impoverish, harass, or subdue them to the Arbitrary and Irregular commands of those that have it: there it presently becomes Tyranny, whether those that thus use it are one or many.”⁴⁴ This new emphasis on right and authority, characteristic of writers in the social contract tradition, illuminates a larger shift in political terminology, best seen in the development of the concept of dictatorship, which eventually displaces tyranny.

Dictatorship and Sovereignty

The terminology of “dictatorship” has its roots in Roman legal thought. As Mark Wilson observes, in early Republican times the dictator appears to have been a magistrate appointed to an office with enlarged powers for a specific and limited purpose, requiring the solution of a distinct political, military, or religious emergency.⁴⁵ In the three-hundred-year period between the founding of the Republic (~500 BCE) and the end of the Second Punic War (202 BCE), eighty-five dictators were appointed for a variety of purposes.⁴⁶ Wilson indicates that while the office was not intrinsically time limited (contra later interpretations from the first century BCE), these archaic dictators typically resigned their offices as soon as possible after resolving the crises for which they were appointed.⁴⁷ Dictatorship was thus a form of emergency government, heavily constrained by customary understandings of the role, even if in principle unaccountable to other Roman officeholders.⁴⁸ After the battle of Zama, however, the Romans did not call for a dictator for one hundred and twenty years. Instead, they made use of a variety of other devices to empower other officials to deal with emergencies, including the prorogation of commands (e.g., the creation of proconsuls) and the use of the senatus consultum ultimum.

The civil conflicts of the late Republic led to conceptual innovation, both by political figures intent on reviving the dictatorship and by writers who criticized these usages. First Sulla revived the office in 82 BCE in an attempt to fundamentally reform the constitution in a conservative direction, engineering his appointment as dictator legibus faciendis et reipublicae constituendae causae, dictator for making the laws and (re)constituting the republic. As Wilson notes, Sulla thus extended the “mandate” of the office beyond what had normally been the case in the archaic period, though he still resigned his office after he considered his task completed.⁴⁹ Caesar’s multiple dictatorships in the wake of further civil conflict in the 50s further expanded the temporal scope of the office, eventually culminating in his appointment as dictator perpetuus in 44 BCE. By this point the authority of the office hardly added much to Caesar’s personal power; it was one of many increasingly sycophantic honors bestowed on him by the Senate, and condemned by later writers as signs of tyranny.⁵⁰ Nevertheless, Wilson argues that something of the structure of authority of the ancient dictatorship as emergency government remained even then, adding weight to Caesar’s efforts to reform Roman institutions and stabilize its empire in conditions of permanent crisis.⁵¹ By the 20s the office was irredeemably tarred by Caesar’s “tyrannical” actions as dictator, and Augustus refused it when offered to him in 22 BCE, preferring princeps, a term “encumbered by neither ancient nor recent precedent.”⁵²

The experience of Caesar’s dictatorships did not, however, turn the term into a synonym for “tyranny.” While the dictator could turn into a tyrant, dictatorship remained the name of an office, not a regime type or a pattern of character; legally speaking, a dictator turned tyrant had exceeded his authority. As Andreas Kalyvas notes, in the early Principate Dionysius of Halicarnassus and Appian of Alexandria discussed dictatorship as a “temporary tyranny” and the tyrant as a “permanent dictator,”⁵³ but with the end of the Republic the point of using “dictatorship” to refer to a potentially tyrannical ruler was lost; “tyrant” was a perfectly serviceable alternative applicable to monarchs who misused their power. Though later writers occasionally used “dictatorship” to refer to a temporary magistrate, deriving its authority from another body but not bound by normal legal restraints, and given the task of resolving an emergency situation (see, for example, Rousseau’s discussion of dictatorship in On the Social Contract 4.6), it was rarely part of a live evaluation framework for political pathology. The terminology of dictatorship became what we might call “academic,” appearing primarily in discussions of the history of the Roman Republic, as Ernst Nolte demonstrated in his history of the concept.⁵⁴

Nevertheless, the idea of a constitutional device to resolve emergency situations remained useful in the later age of absolutist state-making in Europe. Locke’s theorization of the “prerogative power” was perhaps the closest to the ancient dictatorship (Second Treatise, chapter 14).⁵⁵ Such prerogative often took the form of royal and imperial “commissars” and declarations of a “state of siege,” catalogued by Schmitt in his 1921 Dictatorship. ⁵⁶ The terminology of “dictatorship,” though not commonly used, retained its importance for Romanophilic theorists as a plausible model for how to integrate such devices into a constitutional structure.⁵⁷ And as Schmitt noted, the idea of dictatorship became more important with the rise of the discourse on sovereignty in the seventeenth century.

Discussions of sovereignty are concerned with the justification of authority: not whether a ruler uses his power for good or ill, but whether a ruler has the authority to use power in any particular way. The absolutist concept of sovereignty represents an extreme on the spectrum of answers to this question: the sovereign has the authority to use power in any way, good or ill. Though not every theorist agreed with this understanding—as Locke argued, there are some uses of power that nobody has a right to—the majority view (exemplified by thinkers as diverse as Bodin, Hobbes, and Filmer) of sovereignty as supreme authority meant that tyranny and monarchy could not be distinguished from one another on that basis. Thus, for Hobbes tyranny is “monarchy misliked”;⁵⁸ both tyrant and monarch are equally sovereign. A similar point was made by Filmer from a very different theory of the basis of sovereignty—a theory which nevertheless accepted the premise that sovereignty was equivalent to an unlimited right to use power.

This is not to say that “tyranny” lost all critical purchase. Defenders and critics of absolutism continued to make distinctions between good and bad uses of power in monarchy,⁵⁹ but as several studies have noted,⁶⁰ while the primary critical term between the sixteenth and seventeenth centuries continued to be “tyranny,” “despotism” gained traction in the eighteenth century, thanks in particular to the influence of Montesquieu, who developed a contrast between the “despot” (ruling by fear, without intermediate bodies, associated with “Oriental” polities) and the “monarch” (ruling through honor, with many intermediate bodies) while simultaneously reducing “tyranny” to a specific pathology of democracy. As Richard Koebner showed, other writers had already drawn on the “Orientalist” connotations of “despotism” to criticize aspects of the French monarchy,⁶¹ but Montesquieu systematized this, focusing attention not just on the person of the king but on the system of monarchy.⁶² The contrast between an aristocratic monarchy with many intermediate bodies and a despotic system without them was particularly congenial to aristocratic members of the French Parlements who resented monarchic interference, but “despotism” came to be used in many different ways as a systemic criticism of monarchy. Conceptual innovation in the eighteenth century thus tended to depersonalize the problem of absolutist monarchy.

Yet both “despotism” and “tyranny” proved insufficient to describe the new politics of nineteenth-century Europe. While “despotism” remained a key critical term during the revolutionary age, it also lost its distinctiveness, as many authors came to use it interchangeably with “tyranny.” The period even saw inversions in its valence, as in Robespierre’s “despotism of liberty,” and in the later nineteenth century some historians used “enlightened despotism” to refer positively to a particular set of eighteenth-century monarchs.⁶³ More importantly, influential authors thought that both concepts were inadequate to diagnose the problem rulers like the Bonapartes or later Bismarck posed. These rulers claimed popular legitimation and governed in a systematic fashion (rather than arbitrarily), but their legitimacy was nevertheless suspect. “Tyranny” with its focus on character was manifestly inadequate to the description of the phenomena; but so was “despotism,” with its accent on fear and lack of connection to popular legitimation.⁶⁴ Hence it is not surprising that we see a steady decline in the usage of “despotism” and “tyranny” from the revolutionary era onwards, as noted above, and the introduction of a variety of alternative terms, from “Bonapartism” to “Caesarism” to “dictatorship.” The overall tendency was toward increased generality; as Prutsch has shown, Bonapartism was popular earlier in the nineteenth century (when the problem was Napoleon), Caesarism increased in popularity in the middle of the nineteenth century (after Louis Napoleon), and dictatorship finally triumphed at the beginning of the twentieth.⁶⁵

One reason for the emergence of “dictatorship” as a key post-Revolutionary term was the new understanding of sovereignty as grounded in “consent.” Though defenders of absolutism did not generally intend to ground sovereignty on consent, Hobbes did; and the eighteenth and nineteenth centuries increasingly follow Hobbes in seeing sovereignty as dependent on “the people’s” implicit authorization of the sovereign body. Indeed, explicit authorization, through ratification of a constitution, election, or some other representative mechanism, takes on an ever greater role in the theory and practice of sovereign authority. But if sovereignty, by construction, comes from the consent of the people, then the people retain an ultimate and absolute authority, which Sieyès called the pouvoir constituant during the early debates of the French Revolution—the power to create a constitution, or a state, which could in principle be exercised at any moment and to set aside preexisting constraints. Thus, while sovereignty is typically exercised in some mediated form (through laws and constituted bodies), it is characteristic of this discourse of sovereignty that the ultimate sovereign (be it the king or the people as the pouvoir constituant) can make new law, set aside law, or even intervene in ad hoc ways to address a crisis that threatens its existence.

From this point of view, the dictator can “represent” the sovereign people, either by explicit delegation, on the classical model, or by appropriating its powers for the sake of resolving an emergency situation, on the “Caesaristic” model. Schmitt argued that dictatorship was understood in terms of explicit or “commissarial” delegation until the late eighteenth and early nineteenth centuries; and these forms of commissarial delegation are still with us, despite the fact that we do not use the terminology of dictatorship.⁶⁶ But from the nineteenth century onwards it became possible to understand dictatorship as a potentially illegitimate, though in principle temporary, appropriation of sovereignty justified by reference to an emergency situation. As Arato argues, dictatorship as such comes to be increasingly understood in “sovereign” terms,⁶⁷ while other legal mechanisms (e.g., legislation under urgency, states of siege or emergency) took the place of the “commissarial” dictatorship.

The important point for present purposes is that dictatorship remained tied to questions of authority, rather than character; and as noted, these questions came to focus increasingly on the “consent of the people” from the seventeenth century onwards. Dictatorship could either be seen as a necessary use of authority to “restore” popular sovereignty in the face of the recalcitrance of partial intermediate bodies or internal enemies, or as a “usurpation” of that same sovereignty.⁶⁸ And to the extent that the dictator is understood to have a right to normatively unlimited power, it is increasingly as a representative of the pouvoir constituant of the people rather than of a monarch, even if the dictator is never explicitly authorized.

The emergence of republicanism in nineteenth-century Latin America also gave the term “dictatorship” a new lease on life and a contested (and often positive) valence. As Moisés Prieto observes, the acute problems faced by the new Latin American republics made the Roman magistracy an obvious reference point for people steeped in the Latin classics to justify the assumption of supreme power and resolve existential crises without abandoning the presumption of popular legitimacy.⁶⁹ The early rulers of many of these republics occasionally took on the title of “dictator” without embarrassment: Simón Bolívar was formally dictator three times during the wars of independence in the territories that eventually became Venezuela, Colombia, Ecuador, Bolivia, and Peru, and the first leader of independent Paraguay, José Gaspar Rodríguez de Francia, declared himself its “perpetual dictator” in 1816, ruling the country until his death in 1840.⁷⁰ Moreover, though the term “dictator” eventually re-acquired a negative connotation, the possibility of “constitutional dictatorship” in order to “save the republic” was not seen as illegitimate. Stephan Ruderer suggests that Latin American political thought even appropriated the term “caudillo” to refer to “good dictators,”⁷¹ and in the early twentieth century the idea that social disorder justified dictatorship became the basis for Venezuelan sociologist Laureano Vallenilla Lanz’s influential theory of the “necessary gendarme.”⁷²

In Europe, the valence of dictatorship was also contested throughout the nineteenth century, especially after the revolutions of 1848, which made clear the continuing instability of the European political order, as Prutsch argues.⁷³ Conservatives like Juan Donoso Cortés invoked dictatorship as a restorative force, hoping for the use of repression to promote a Catholic reaction in Spain.⁷⁴ In his view, dictatorship was extra-ordinary government to “save” society, likened to God’s extraordinary and miraculous interventions, and thus outside the law.⁷⁵ Donoso Cortés identified dictatorship not with particular individuals or constitutional mechanisms, but with the “reserve” powers of sovereignty, as did later thinkers. He thus claimed to find dictatorship not only in the Roman institution of the same name, but also in the Athenian power of ostracism, the French National Assembly, the 1830 Spanish Carta, and the supremacy of the British parliament.⁷⁶ But as Brian Fox shows, Donoso Cortés, for all his influence on later thinkers from Romieu to Schmitt,⁷⁷ remained ambivalent about this power.⁷⁸ Dictatorship was a neutral instrument, capable of coming from below as well as from above, and of fomenting revolution as well as order.⁷⁹ If he had thought religion sufficiently strong, he would have preferred liberty to dictatorship.

Marxists also developed a positive understanding of dictatorship that stressed its transformational agency in the second half of the nineteenth century. For Marx, and later Lenin,⁸⁰ the “dictatorship of the proletariat” was not the problematic rule of a single individual or group, but the sovereign power of the working class, organized to crush its enemies without the constraints of law to accelerate the transformation of the old order, but supposed to vanish with the coming of the new.⁸¹ Dictatorship thus remains tied to democracy as a “transitional” phase towards a deeper, fuller democracy; and pragmatically it is even associated with the main organizational innovation of republican politics, the political party. The Marxist dictatorship is no longer oriented towards the restoration of the past, but towards the construction of a future order; yet it remains in principle the agent of the sovereign, interpreted as the working class. As Nolte notes, the key contrast is not between dictatorship and democracy, but between true democracy (the dictatorship of the proletariat) and bourgeois democracy (the dictatorship of the bourgeoisie).⁸² Yet the word remained a provocation, and its valence could easily be inverted; by 1922 Kautsky was using it in a negative sense to criticize the Leninist regime as a dictatorship of a party and its leader.⁸³

In the early twentieth century Fascists and conservative nationalists also spoke of “dictatorship” in positive terms, drawing on its classical sense as a tool for resolving what they saw as great national crises.⁸⁴ A general sense of the crisis of constitutional democracy during the interwar era justified the establishment of dictatorships in many countries as forms of “emergency government,” even as the protagonists of these new regimes (e.g., Piłsudski, Primo de Rivera, or King Alexander of Yugoslavia) did not always use the term. Yet observers quickly converged on the idea of “dictatorship” to describe them, using it sometimes in relatively positive ways.⁸⁵ Schmitt initially justified Article 48 of the Weimar constitution as enabling a much needed “commissarial” dictatorship, though by 1933 he supported a “sovereign” dictatorship.⁸⁶ And as Arato shows, although Fascists had a much greater estimation of the leader figure than Marxists, they mostly still understood the dictator as essentially representing the popular (or community) “will.”⁸⁷

At the same time, the ambiguous valence of “dictator” (both restorer and usurper in Roman history) still required some justification, and the most notorious Fascist leaders preferred titles that did not mean “dictator” (“Duce” and “Führer”). Thus Mussolini, who while in power had not disclaimed being a dictator, thought it necessary to say in 1945 that “strictly speaking, I was not even a dictator, because my power to command coincided perfectly with the will to obey of the Italian people.”⁸⁸ Hitler claimed in Mein Kampf to be a “Führer” of the people (Volk) rather than a “dictator,” which he associated with violence, Judaism, and Marxism, despite the fact that the idea of a Mandatar des Volkes had been the essence of the national-conservative understanding of dictatorship up to the 1920s.⁸⁹ The ambiguity of “dictator” and “dictatorship” enabled opponents of both Marxists and Fascists to invert its valence. Marxists were uncomfortable with the associations of the term after the war and tended to stress the opposition between the “People’s Democracies” and the Fascist dictatorships and bourgeois democracies.⁹⁰ Moreover, because the key characteristic of the dictator is that he need not consult or require the consent of those he acts upon, and can set aside established rules to resolve an emergency, dictatorship came to signify an impairment of consent and constitutionality, either temporary or permanent, and eventually became a catch-all term for undemocratic patterns of rule, understood as appropriations of popular sovereignty. Arato’s attempt at a modern definition that distinguishes “dictatorship” from premodern tyranny thus stresses the power to set aside a rationalized legal order and to appropriate sovereignty in defiance of constitutional authorization.⁹¹

Totalitarianism and Authoritarianism

While by the early twentieth century “dictatorship” had effectively displaced “tyranny” from its previous centrality in the network of concepts used to discuss bad political regimes, new phenomena still called for conceptual development. Concepts form networks; their generative potential (whether for polemics or explanation) depends on their relationships with other concepts. Consider “totalitarianism.” Coined in 1923 by the antifascist journalist and Communist deputy Giovanni Amendola to criticize the manner in which Fascists attempted to monopolize power, the word “totalitarian” was appropriated by Mussolini in 1925 as a description of his movement, but as Bruno Bongiovanni argues, it retained a critical sting that made the term easier to use by opponents of Fascism, Nazism, and later Soviet Communism.⁹² Many scholars in the postwar era found “totalitarianism” useful to understand the peculiarities of these new regimes, using a term that expressed the “totalizing” nature of their political projects. We need not delve deeply here into the controversies over the use of the term during the Cold War, except to note that the concept involved stretching the (Marxist) idea of sovereign dictatorship as a “transformative” form of government rooted in “mass politics” and legitimated by an ideological project.

While writers in this tradition did not always embrace the terminology of dictatorship—Arendt in particular was skeptical of its utility for understanding totalitarianism, though she still grudgingly used the term⁹³—it was evident to most of them that the ancient category of “tyranny” was inadequate to understand the phenomenon, much less to effectively criticize it.⁹⁴ “Totalitarianism” was often viewed as a radicalization of tendencies present in the discourse of popular sovereignty of the French Revolution,⁹⁵ and writers argued that the totalitarian regimes could only be fully understood and criticized by attending to the forms of their ideological legitimation rather than to the character of the tyrant or the brutality of their rule. The category of “totalitarianism” thus remained firmly in the orbit of “dictatorship”; the term’s enduring appeal and utility lie precisely in its concern with the process of ideological legitimation, and thus with the forms of its authority.⁹⁶

“Totalitarianism” could not, however, replace “dictatorship.” Its polemical usage multiplied its meanings and made its application vague; and most dictatorships could not be called “totalitarian,” as even proponents of the totalitarian paradigm understood. Nevertheless debates over “totalitarianism” during the Cold War produced a multiplicity of alternative conceptualizations of bad political regimes, among which “authoritarianism” has become the only one able to partially displace “dictatorship” from its centrality in such discussions today. The term, first attested in the 1860s,⁹⁷ initially referred to the use of authority without sufficient justification (mostly in nonpolitical contexts) or which violates individual liberty. Left-wing (especially anarchist) writers used it commonly in the 1910s to refer to a variety of institutions, including the church, that inculcated servile dispositions or unjustifiably restricted individual freedom.⁹⁸ Authoritarian rule is thus not simply rule with authority—all regimes require some form of authority, as Przeworski rightly notes in a critique of the concept⁹⁹—but rule that abuses its authority by, for instance, violating certain rights and liberties, or by failing to give an account of itself; the authoritarian only commands, without explaining why something must be done. To be “authoritarian” is thus to claim illegitimate or unaccountable authority.¹⁰⁰

In the 1960s, Juan Linz repurposed the term to distinguish “totalitarian” dictatorships reliant on totalizing ideological forms of legitimation from “authoritarian” regimes whose main characteristic was their “limited pluralism.”¹⁰¹ His redefinition put the focus on the result of violations of individual liberties and public accountability, rather than on the violations themselves. In a regime of “limited pluralism” certain political forces have been prevented from competing for positions of authority, and certain justifications for authority have been systematically censored or marginalized. Under authoritarianism, potentially authoritative people and ideas are prevented from holding authority without good reason, while other people and ideas acquire authority without sufficient justification. Linz also wished to restrict the usage of “dictatorship” to forms of crisis government, leaving only “nondemocracy” or “autocracy” as the broader category for the political regimes he was studying.¹⁰²

While this was not Linz’s intention, his reformulation of authoritarianism enabled “totalitarianism” to be conceived merely as an extreme instance of the former, with “democracy” as an ideal standard where open competition for office diminished the degree to which authority lacks justification: totalitarianism has no ideological pluralism, authoritarianism limited pluralism, and democracy sufficient pluralism. The low number of “true” totalitarian regimes also eroded the utility of a separate “totalitarianism” category. Accordingly, “authoritarianism,” like “dictatorship,” eventually lost its distinctiveness as a specific kind of autocratic regime and became synonymous with any form of nondemocracy.

But “authoritarianism,” in contrast to “dictatorship,” does not have a clear connection to popular sovereignty or emergency rule. Instead, as Marlies Glasius suggests, the term shifts the focus from rulers and the basis of their authority to their practices.¹⁰³ While the concept of an authoritarian regime (a regime that systematically limits or degrades political pluralism, primarily by preventing fair electoral competition) remains cogent, we can also say that even democratically elected leaders or institutions may act in authoritarian ways by abusing their authority, or evading accountability for their actions, without saying that they are dictators (able to use sovereign, unlimited power) or tyrants (using their power for the pursuit of their private desires). And while dictatorship is typically understood to be authoritarian, not all authoritarianism is dictatorial, since authoritarian practices are not all justified by reference to the concrete features of an emergency situation or retain a link to some real or imagined pouvoir constituant.

Conclusion

Some degree of terminological chaos is the norm for political concepts. They are, after all, political, their meanings enlisted in struggles over legitimation. But the development of the meaning of such concepts is not arbitrary and responds in part to their intrinsic potentialities and “fit” with other concepts. “Dictatorship” displaced “tyranny” as the central concept in the study of bad political regimes because it fit better within the emerging understanding of popular sovereignty since the early nineteenth century and responded better to the new problems caused by this understanding in the new republics of Europe and Latin America. The emerging network of concepts surrounding dictatorship—including “totalitarianism” and “authoritarianism”—extended this understanding by focusing attention on specific problems of legitimation, from ideology to abuses of authority to its personalization.

Today it is characteristic of most forms of “dictatorship” (as the term is used by political scientists) to be dressed in democratic garb, and their rulers are not obviously “tyrants.” Such regimes use elections, sometimes relatively competitive ones; their leaders spend much time courting popular support; their public sphere is lively if restricted; their violations of individual rights are limited to particular minorities or individuals; and most of them sport constitutional documents containing a multitude of individual rights and where they claim to be democracies. And while it is possible to find examples of more straightforward tyrannies in recent history, one of the key problems for political scientists today has been to characterize how a wide variety of political arrangements that appear democratic in fact significantly impair popular sovereignty, constrain individual consent and choice, evade genuine accountability, and violate important rights.

The increasing replacement of “dictatorship” by “authoritarianism” in scholarly work, including adjectival forms like “competitive” authoritarianism, probably reflects in part a desire for a characterization of these phenomena that is not tied to dictatorship’s conceptual baggage (such as its connection to emergency rule or its “transformational” dimension), but instead stresses the many possible ways in which access to political authority can be limited and its uses made less accountable. Following Hans Kelsen, some scholars also use “autocracy,” but the term has not been widely adopted, and offers few advantages over dictatorship (indeed, autokrator was sometimes used in Greek to translate the Latin dictator—Polybius, Histories 3.86.7), though without the connection to emergency government.¹⁰⁴ John Keane has proposed reviving “despotism,” drawing on Etienne de La Boétie’s sixteenth-century Discourse on Voluntary Servitude (which does not contain the term in the original French—La Boétie used tyran and tyrannie, not despotisme) and on Montesquieu’s eighteenth-century Spirit of the Laws (which does) to focus on how citizens are reduced to servitude and made dependent on the government through the manipulation of their consent.¹⁰⁵ The key problem here is why people submit to systems of apparently illegitimate domination. The idea of despotism suggests that such submission is in their interests in some sense. The focus is not on the authorization of “unlimited” power (dictatorship), or the character of the tyrant, or on abuses of authority, but on the mechanisms that make it possible for unaccountable authority to operate as such.

Yet there is little need to choose. Dictatorship, authoritarianism, autocracy, tyranny, and despotism all refer in different ways to the problems posed by the impairment of popular consent, choice, and sovereignty in modern societies shaped by the understanding of political legitimacy that emerged in the sixteenth and seventeenth centuries. Dictatorship points to the appropriation of sovereignty; authoritarianism to the abuse of authority without right or accountability to those over whom it is exercised; and despotism to the manipulation of interest and passion to ensure voluntary submission to these forms of power. Each of these concepts adds new layers to our understanding of political regimes. We can condemn the tyrant based on the ancient understanding of tyranny, as someone who subjects the ruled to his desire, but also note how the dictator who claims to represent popular sovereignty in effect usurps it, or how authoritarian practices violate our rights or evade accountability, or how the mechanisms of despotism prevent popular organization. Moreover, these new concepts make it possible to understand how democracies also fail. The imaginary of political pathology is enriched by having a multiplicity of conceptual tools that can make sense of the different ways both injustice and illegitimacy are implicated in particular political regimes.