Worlds of Civil War: The Twentieth Century

“All European wars, said Voltaire, are civil wars. In the twentieth century his formula applies to the whole earth. In our world, which shrinks progressively as communications become swifter, all wars are civil wars: all battles are battles between fellow-citizens, nay more, between brothers.”1The words are those of Jaime Torres Bodet (1902–74), the Mexican scholar, poet, and diplomat who served as the second director general of the United Nations Educational, Scientific, and Cultural Organization after World War II. He was speaking in 1949, soon after the foundation of the UN and its sibling organizations and in the wake of the momentous events of 1947–48, among them, Indian independence and partition, the foundation of the State of Israel and the first Arab-Israeli conflict, and the promulgation of the Universal Declaration of Human Rights: a pivotal moment that one of his contemporaries, the émigré German political scientist Sigmund Neumann (1904–62), called—with the addition of the Chinese civil war and burgeoning anticolonial nationalism in the Middle East and Southeast Asia—both “the age of revolutions” and “the international civil war.”2

The title as well as the message of Torres Bodet’s speech, delivered in Paris at the United Nations Day ceremony on October 24, 1949, was “Why We Fight”: not in the sense of the global military conflict that had ended four years earlier, but in a different struggle—the fight for peace. On this occasion, Torres Bodet’s sentiments about civil war were sounder than his scholarship. Although Voltaire did famously argue that Europe was a “kind of great republic divided into several states,” all with “the same principle of public law and politics, unknown in other parts of the world,” he did not stretch his vision of European community to imagine its wars as civil wars.3 That leap was taken instead by Voltaire’s more courtly predecessor the French Catholic archbishop and political writer François de Salignac de la Mothe Fénelon (1651–1715). In his immensely popular work of advice for a young prince, the Dialogues of the Dead (1712), Fénelon has the character of Socrates offer an eloquent pacifist argument based on the cosmopolitan principle of common humanity:

All Wars are properly Civil Wars [Toutes les guerres sont civiles], ’tis still Mankind shedding each other’s Blood, and tearing their own Entrails out; the farther a War is extended, the more fatal it is; and therefore the Combats of one People against another, are worse than the Combats of private Families against a Republick. We ought therefore never to engage in a War, unless reduced to a last Extremity, and then only to repel our Foes.4

All such wars are not just civil wars; they are, as Lucan might have said, worse than civil wars, precisely because they ensnare ever larger circles of humanity.5 It was one of the many paradoxes of the civil war’s intellectual history that as the world came closer to the cosmopolitan ideal of universal humanity, the more intimate would international and even global wars become. More acute pain, not more assured peace, might be the unintended outcome of the world’s progressive shrinkage as cosmopolitan empathy and interaction grew.6 Enlightened thinkers like Fénelon who believed in Europe’s cultural unity feared that all wars between Europeans would become civil wars, because they were fought within the bounds of a community of fellow citizens who recognized one another as such. Under Immanuel Kant’s later conception of cosmopolitan right in “Toward Perpetual Peace,” the realm of mutual recognition itself became global, because “the (narrower or wider) community of the nations of the earth has now gone so far that a violation of right on one place of the earth is felt in all.7

Over the century of near-constant warfare that was Europe’s “Second Hundred Years’ War” (1688–1815), European civil war had, rather ironically, come to signify, both on the Continent and in its imperial outposts, a degree of cultural unity as well as civilizational difference from the rest of the world. In his own Project for Perpetual Peace (1761), Jean-Jacques Rousseau judged the wars between the powers of Europe to be “much the more deplorable, as their combinations are intimate;…their frequent quarrels have almost the cruelty of civil wars.”8 Four decades later, in 1802, Napoleon reportedly told the British minister Charles James Fox, during the negotiations for the Anglo-French Treaty of Amiens, that “Turkey excepted, Europe is nothing more than a province of the world; when we battle, we engage in nothing more than a civil war.”9 And in 1866, the French historian Henri Martin saw no end to these European civil wars, which for him included the recent Crimean War (1853–56), because he viewed Russia, too, as part of the same civilization.10 The saying that all European wars were civil wars would again become popular in the moment between the twentieth century’s two great wars, when it was usually attributed to Napoleon, perhaps recalling his bon mot of 1802.11

The second half of the twentieth century would indeed witness the globalization of civil war, but not quite in the form Torres Bodet or his predecessors in the Enlightenment might have anticipated. This new world of civil war emerged with three overlapping features. First, civil war (now technocratically rebranded as “international conflict of a non-international character”) gradually came under the jurisdiction of international institutions, especially international humanitarian law, in the wake of World War II, but with subsequent modifications during the age of decolonization and then during the internal conflicts of the 1990s. Second, and closely related to the first, civil wars became seemingly ubiquitous, distributed across most parts of the world (especially Africa and Asia, but in the 1990s in an otherwise pacified Europe as well), and then gradually came to supplant wars between states as the world’s most common and widespread form of large-scale organized violence. And third, the communities within which civil wars were imagined as taking place—the polities, civitates, or spheres of human commonality—became ever wider and more capacious, until the idea of “European civil war” gave way to various conceptions of “global civil war” in the present century.

At the same time, the era’s great transnational conflicts, from World War I to the Cold War and then on to the “Global War on Terror” early in the twenty-first century, were often seen as civil wars in the realms of political and legal argument. But as we shall see, by the 1960s and 1970s social scientists and philosophers had begun to take a more focused interest in civil war as a topic of analysis, speculation, and definition. Traces of all these arguments endure in ideas of civil war today and are likely to linger into the future.


Shortly before Torres Bodet delivered his speech in October 1949, a humanitarian conference aimed precisely at ameliorating the ever-expanding effects of war concluded its deliberations in Geneva in August 1949. The Diplomatic Conference, as it was known, drew representatives from across the world to revise the Fourth Hague Convention of 1907 and the 1929 Geneva Convention with particular regard to the status of civilians in time of war. The most pressing issue on the minds of many of the assembled delegates was how to extend the protections guaranteed to recognized combatants in conventional international warfare to “the victims of conflicts not of an international character.” Not all of them could agree that this was necessary or desirable; some, including the British delegation, thought the incursion of international law into domestic disputes was an infringement of national sovereignty. (Recall that it was, in fact, on just such grounds that the original Geneva Convention of 1864 had not been extended to civil wars.) Others, however, argued successfully that “the rights of the State should not be placed above all humanitarian considerations” because “civil war was more cruel than international war.” The result of these deliberations was Common Article 3 of the Geneva Conventions (1949), now finally applied to what was precisely termed “armed conflict not of an international character” (later compressed to “non-international armed conflict” or, still more succinctly, “NIAC”).12

The 1949 discussions leading to Common Article 3 built on proposals set forth in Stockholm in 1948 by the International Committee of the Red Cross to make application of the existing Geneva Conventions “obligatory on each of the adversaries” in “cases of an armed conflict which are not of an international character, especially cases of civil war, colonial conflicts, or wars of religion.” After much discussion, the revised draft presented in Geneva in 1949 omitted the last qualifying clause and specified only “armed conflict not of an international character.” That became the preferred formulation thereafter among international lawyers and international organizations, despite early objections that it could cover too wide a range of violent acts within the frontiers of a single state: not just “civil” wars, but any enemies of the state, whether legitimate freedom fighters, brigands, or even common criminals—in fact anyone engaged in riots or coups d’état rather than actions recognizable as wars. Did they all deserve the protection of the Geneva Conventions, even if their actions were illegal under domestic law?13 Most civil wars were wars “not of an international character”; however, only some wars “not of an international character” were civil wars. Trying to draw a line between the two overlapping categories would remain a continuing source of controversy and confusion up to the present.14

As finally adopted, Common Article 3 of the Geneva Conventions was minimalist in its ambitions. It provided that civilians and members of armed forces who were no longer combatants (for instance, because they were wounded or ill) were “in all circumstances to be treated humanely”; that “the wounded and sick shall be collected and cared for”; that the Red Cross would be permitted to minister to anyone involved in the conflict; and that the parties to the conflict should endeavor to apply as many remaining provisions of the Geneva Conventions as possible.15The article permitted great latitude of interpretation, not least because no attempt was made to define more exactly what constituted an “armed conflict not of an international character,” thus avoiding “the dangers of under- and over-inclusivity.” What resulted was not so expansive that it might encompass a range of internal police actions (or the threats to national sovereignty provoking them), for example, yet not so restrictive that too many conflicts might be excluded from any regulation or amelioration. On the other hand, it gave states ample discretion to decide whether or not conflicts crossed the threshold from rebellion to civil war and therefore to decide for themselves whether their own actions against rebels would be subject to Common Article 3 and the rest of the Geneva Conventions. That latitude seemed especially precious to states with overseas colonies that might demand self-determination. So it was for Portugal, which in 1949 “reserve[d] the right not to apply the Provisions of Article 3, in so far as they may be contrary to the provisions of Portuguese law, in all territories subject to her sovereignty in any part of the world.”16

Common Article 3 was drafted and approved in 1949 largely in response to concerns about the inadequacy of the existing Geneva Conventions during recent conflicts such as the Spanish Civil War (1936–39). In the decades after World War II, the heightened incidence of “non-international” conflicts demanded greater precision in the application of the conventions. Amid the proxy wars of the Cold War, and the wreckage of disintegrating empires, intervention into internal conflicts became more common, tarnishing the luster of the Long Peace then emerging in Europe. These pressures led to a program of updating and revising the Geneva Conventions between 1974 and 1977. In this context, the Institute of International Law (Institut de Droit International)—the leading professional body of the global community of international lawyers—met in 1975 in the German city of Wiesbaden to draft a document titled “The Principle of Non-intervention in Civil Wars.” This Wiesbaden protocol noted “the gravity of the phenomenon of civil wars and of the suffering they cause” and expressed concern that such conflicts could readily escalate into international conflicts if any side sought foreign involvement, thereby triggering the opposing side to do likewise. External parties were urged not to intervene, except to offer humanitarian, technical, or economic aid “not likely to have any substantial impact on the outcome of the civil war.” In the course of setting conditions for noninterference, the institute sought briefly to define “civil war” as “any armed conflict, not of an international character, which breaks out in the territory of a State” and in which either an insurgency aiming to take over the government, or to secede, opposes the established government; or two or more groups vie to control the state when no government exists. Crucially, the Wiesbaden Protocol also set limits specifying what was not a civil war; “local disorders or riots,” “armed conflicts between political entities separated by an international demarcation line,” and “conflicts arising from decolonization” were all beyond the pale.17

The outcome of these discussions was a set of further protocols, of which the second—Additional Protocol II (1977)—applied to conflicts of a non-international character. The limits set at Wiesbaden continued to apply, because AP II excluded riots and also wars of decolonization, these being covered instead by Additional Protocol I, which first brought international humanitarian law to bear directly on anti-imperial struggles. The second Additional Protocol expanded the protections and prohibitions relevant to civil wars, remaining today the chief component of humanitarian law relevant to them.18 The application of those protections depends on the judgment that a conflict “not of an international character” is in progress. If the conflict is held to be “international”—that is, between two independent sovereign communities—then the full force of the Geneva Conventions applies. If it is “non-international,” then it will be covered by Common Article 3 and Additional Protocol II.19 But if the violence has not been deemed a conflict of either kind—perhaps because it is a riot or an insurgency—it remains within the scope of domestic jurisdiction and hence subject to police action. In these cases, a great deal hangs on the determination of whether a conflict is “not of an international character” or, in plain speech, whether it is a “civil” war.

The legal boundaries of what is or is not a civil war have continued to be flexible and dynamic.20 The institute’s next major resolution concerning non-international conflicts (1999) reflected the experience of the Balkan Wars: “Considering that armed conflicts in which non-State entities are parties have become more and more numerous and increasingly motivated in particular by ethnic, religious or racial causes,” with especially devastating consequences for civilian populations, the institute recommended that international humanitarian law should apply to “internal armed conflicts between a government’s armed forces and those of one or several non-State entities, or between several non-State entities.”21 This shift in turn reflected the jurisprudence of the International Criminal Tribunal for the Former Yugoslavia, or ICTY, which, over the course of the 1990s, had attempted to apply international humanitarian law to internal conflicts.

In 1996, the ICTY ruled that the Bosnian war had mutated from an international war into a civil war at the point in 1992 when the former Federal Republic of Yugoslavia had withdrawn its support from the ethnic Serbs. This shift was of special significance because the defendant before the tribunal, the Bosnian war criminal Duško Tadić, claimed it had no jurisdiction over his actions because the statute creating the tribunal applied only to international armed conflicts. The tribunal’s 1996 ruling was later reversed on appeal, but it revealed how much can hang on the definition of a conflict as civil war—in this case, whether or not Tadić could be held liable for breaches of the Geneva Conventions.22 The ICTY Appeals Chamber made the stakes very clear in its ruling on the Tadić case:

Why protect civilians from belligerent violence, or ban rape, torture or the wanton destruction of hospitals, churches, museums or private property, as well as proscribe weapons causing unnecessary suffering when two sovereign States are engaged in war, and yet refrain from enacting the same bans or providing the same protection when armed violence erupted “only” within the territory of a sovereign State?23

The international institutions created in the last decade or so have tried to answer such questions convincingly, building on earlier efforts, like those of the Red Cross, to apply the constraints increasingly standard in conventional warfare to civil wars.24 The 2004 revision of the British Ministry of Defence’s manual on the laws of war for military personnel, for example, has served to bring civil wars within the theoretical reach of war.25 The U.S. Army’s groundbreaking Counterinsurgency Field Manual of 2007, written with events in Iraq and Afghanistan very much in mind, likewise recalled the need to apply the relevant provisions of the Geneva Conventions and acknowledged that “although insurgencies can occur simultaneously with a legal state of war between two nations, they are classically conflicts internal to a single nation.”26

Yet recent efforts to bring civil war within the pale of civility remain frustratingly and lethally incomplete; as the ICTY put it, there has been no “full and mechanical transplant” of the laws of war to civil warfare, nor could there be until all parties to such conflicts agree to be bound by international humanitarian law. The ICTY nonetheless laid down a vital principle for effecting the desired “transplant”: “What is inhumane, and consequently proscribed, in international wars, cannot but be inhumane and inadmissible in civil strife.”27 As that principle is put into legal operation, the world might come closer to the “civilization” of civil war.

But matters are never quite so simple. Take the case of Syria in 2011–12. Ordinary Syrians understood the contention with Bashar al-Assad’s regime throughout 2011 and the first half of 2012 to be civil war. Outside Syria, however, interested parties considered the matter more debatable. In December 2011, the U.S. State Department’s deputy spokesperson Mark Toner demurred when asked if he agreed with a UN official that Syria was experiencing civil war. “We think violence needs to end in Syria. And that includes among the opposition elements,” he said. “But there’s no way to equate the two, which, in my view, is implied in using the term ‘civil war.’ ”28 As for the Assad regime, it naturally saw only rebellion. The opposition said they were engaged in resistance. Meanwhile, powers like Russia and the United States held the threat of declaring a civil war over each other’s head as they jousted over intervention and nonintervention.29

It took the International Committee of the Red Cross until July 2012—more than a year into the conflict, and after some seventeen thousand deaths—to confirm that what was taking place in Syria was, in fact, an “armed conflict not of an international character.”30 Only with that determination could the parties be covered by the relevant provisions of the Geneva Conventions.31 The reluctance to call the conflict a civil war has become typical of international organizations in the twenty-first century because so much—politically, militarily, legally, and ethically—now hangs on the use or withholding of the term. Thus, a set of legal protocols designed to humanize the conduct of civil war—to bring to bear humanitarian constraints on its practice and to minimize some of the terrible human cost—may have served only to constrain international actors from any such effort. To see how this paradoxical situation arose, we need to review a brief genealogy of the treatment of civil war in the social sciences, beginning in the 1960s.


“When today’s social science has become intellectual history, one question will almost certainly be asked about it: Why did social science, which has produced so many studies of so many subjects, produce so few on violent political disorder—internal war?”32 Here is one of those rare moments when a historian can hear someone from the past talking to them directly from the sources. It is a little unsettling to imagine that the actor here was speaking in 1963 and already waiting for an intellectual historian to map his field, but the question of Princeton professor Harry Eckstein (1924–99) was—and remains—perceptive.

As Eckstein was aware, the academic consensus had long been that civil war was good for absolutely nothing. Accordingly, civil war had been a Cinderella subject, of apparently equal irrelevance across all the academic disciplines. Yet starting in the 1960s, and inspired first by the Cold War and then by the wars of decolonization, American social scientists, often with the backing of the RAND Corporation and similar institutions of the military-academic complex, became increasingly invested in scrutinizing what was called broadly “internal warfare,” a category that encompassed everything from guerrilla warfare and insurgencies to civil wars, coups, and revolutions.33 Eckstein’s call was not heeded as quickly or as wholeheartedly as he might have wished, despite the efforts of a research group on internal war that he ran at Princeton University, which included political scientists, sociologists, and even the odd historian. Progress was slow. “The crucial conceptual issues about internal war are still in the pretheoretical stage,” lamented one of its first systematic analysts in 1970. “Satisfactory theories of internal war have neither been compiled nor evaluated.”34

The continuing confusion about the meaning of civil war could be seen publicly when, in the spring of 1968, the U.S. Senate Foreign Relations Committee held a remarkable series of hearings during the Vietnam War titled The Nature of Revolution. These were chaired by Senator J. William Fulbright and called distinguished academic witnesses, including the eminent Harvard historian of revolution Crane Brinton (1898–1968) and his younger colleague Louis Hartz (1919–86), the political scientist and student of liberalism. On the last day of the hearings, a young Princeton political scientist, John T. McAlister, tried to explain the intractability of the conflict by noting that the United States was not “fighting a civil war that is a purely internal matter” but had instead become embroiled in “a revolutionary war involving all of the Vietnamese people.” Senator Fulbright immediately wanted to know if there was a distinction between a revolutionary war and a civil war. McAlister argued there was: “In civil wars, including our own, there are secessionist goals on the part of those who are fighting. In a revolutionary war, by contrast…the primary goal is unification…[and] there are very distinct political goals about the whole reconstitution of the basis of political order involved.” There then followed a bizarre exchange between the two southerners, Fulbright of Arkansas and McAlister of South Carolina:

THE CHAIRMAN. Well, with that definition is our own War Between the States a civil war or a revolutionary war?

DR. MCALISTER. I would say that was a civil war.

THE CHAIRMAN. They were seeking to secede?

DR. MCALISTER. We were seeking to secede; yes.

THE CHAIRMAN. We were seeking to secede. [Laughter.] But we failed.

DR. MCALISTER. That is right.

THE CHAIRMAN. And since then it has been a revolutionary war as to who controls it?

DR. MCALISTER. That is right.

THE CHAIRMAN. Is that right?

DR. MCALISTER. That is right.

THE CHAIRMAN. All right.35

Such southern humor might have been awkward in the era of the civil rights movement, but it did reveal a persistent confusion, even among political scientists, about the definitions of civil war and revolution.

Even Harvard’s preeminent political philosopher was confused. In the spring of 1969, also during the Vietnam War, John Rawls (1921–2002) gave an undergraduate lecture course titled Moral Problems: Nations and War.36 Two years later, he would publish A Theory of Justice, the work usually credited with reviving Anglo-American political philosophy in the late twentieth century. That work was notoriously reticent on matters of international justice, but in his Harvard lectures Rawls faced squarely the questions swirling around American campuses, including Harvard’s, about the ethics of war, conscription, and civil disobedience. Discussions of just war—both the just causes of war (jus ad bellum) and the justice of actions in war (jus in bello)—bulked large in the lectures. Rawls distinguished between different types of wars in order to define the principles that would best apply in each case. The initial typology found in his lecture notes proposes nine kinds:

1. Wars between existing via states (WW I + II)

2. Civil wars (of social justice) within via states or society (French Rev);

3. Wars of secession of minorities within region: American Civil War.

4. Colonial Wars of secession (from Empire): Algerian War; American Rev War?

5. Wars of intervention (humane intervention)

6. Wars of national unification (War of Roses; Tudors)

7. Wars of conquest, of Empire (Wars of Rome).

8. Wars of Crusade, religious or secular

9. Wars of national liberation (in present sense); guerilla wars37

Rawls’s categories are as revealing as his applications. Civil wars were to be distinguished both from wars between states and from wars of secession, and wars of secession were further divided between intrastate and anti-imperial secessions. By implication, a civil war could only be considered a just war if its aim was what Rawls called “social justice”—that is, comprehensive internal reform directed toward the well-being of all the inhabitants of a viable state or society, such as France after 1789. Wars of secession might be thought just on the grounds that they aimed at the relief of an oppressed population—for example, that of a minority within an established state or of a colonized people within an empire. In common with contemporary international lawyers and political scientists, he separated civil wars from “wars of national liberation” and guerrilla wars.38

Rawls’s distinctions were lucid, but his examples were rather less clear-cut. Initially, he was not sure what kind of war of secession the American Revolutionary War had been. In the body of the lectures, he included both the American Civil War and the American Revolution under “wars of secession of minorities.” This was not evidence that Rawls, who pointedly did not include the Civil War under wars of social justice, equated the two great conflicts but it was perhaps an indication that he did not want to assimilate American Patriots with Algerian colons; in the American Revolution, it was the European settler population that sought to escape from empire, not the indigenous population or the enslaved, for example. And as an example of a civil war of social justice, he cited not the French Revolution but the Spanish Civil War.

Later in his lecture series, as he treated the jus ad bellum in more detail, Rawls briefly examined the question of whether intervention in a civil war might be justifiable, taking John Stuart Mill’s “Few Words on Non-intervention” and the Vietnam War as his points of reference. Rawls was quite dismissive of what he called the “troublesome” faults in Mill’s argument and noted “it would not justify our intervention in Vietnam,” because none of the arguments Mill rehearsed for British intervention in the nineteenth century could be applied to American policy in Vietnam: “We have not intervened neutrally in a protracted civil war; nor have we intervened to help a people throw off a foreign despotism.” Indeed, Rawls concluded, intervention in such a case could only be “under international auspices, where the intervention is impartial…and undertaken for clearreasons of humanity.”39

For Rawls in these lectures, civil war was at least temporarily helpful for clarifying the limits of humanitarian intervention and for elucidating the differences between various instances of national liberation and revolution. For the French philosopher and historian Michel Foucault (1926–84) a few years later, civil war was even more useful in helping to define what he was just beginning to call the “physics” of power. Each year, Foucault had to deliver a course of public lectures on the research he had undertaken while holding a professorship at the prestigious Collège de France in Paris. In 1973, his lecture was titled “The Punitive Society” (“La société punitive”), a topic that would become central to his conception of modern regimes of power. Like many commentators in the 1960s and 1970s on both sides of the Atlantic, Foucault found civil war “philosophically, politically, and historically, a rather poorly developed notion,” not least because most analysts viewed it as what he called “the accident, the abnormality…the theoretical-practical monstrosity.”40 With his characteristic theoretical dexterity and historical daring, Foucault proposed to bring civil war from the margins of analysis to the center by arguing that it was hardly marginal or irrelevant to the understanding of power; civil war, he would argue, was, in fact, the matrix of all power struggles.41

Foucault’s dazzling account of civil war in the 1973 lectures executed three particularly illuminating volte-faces in relation to standard historical accounts. First, he showed that civil war could not be identified with Hobbes’s war of all against all in the Leviathan; indeed, Foucault argued that civil war was the very opposite of the Hobbesian state of nature. Second, he confronted the assumption that civil war was the antithesis of power because it represented its dissolution and breakdown, arguing that civil war was in effect the very apotheosis of power; politics was civil war by other means. And third, he contended that civil war had, in fact, not gradually disappeared in Europe in the transition from the early modern period of wars of religion and monarchical succession to the more stable world of modernity. There had been no progressive movement from an era of civil wars to an age of revolutions; rather, civil war had endured as the fundamental characteristic of what Foucault famously called the “disciplinary society,” in which the structures of power constantly shape human beings.

In accusing Hobbes and his followers of conflating civil war with the war of all against all (bellum omnium contra omnes), Foucault argued, to the contrary, that they could not be further apart, in their character (the one collective, the other individual), their motivation, or, crucially, their relation to sovereignty; the war of all against all was the condition that preceded, and indeed necessitated, the constitution of a sovereign in Hobbes’s political theory, while civil war marked the collapse of sovereignty, the dissolution of the sovereign itself.42 Civil war was directed at seizing or transforming power and therefore “unfolded in the theatre of power”; it haunts power, even to the point that the daily exercise of power can be considered as like a civil war. In this sense, Foucault concluded, with a twist on Clausewitz’s famous dictum that the master of the art of war himself would never have approved, “Politics is the continuation of civil war.”43

While Rawls and Foucault were worrying about the theory of civil war, social scientists, especially in the United States, had already begun a decades-long effort to come up with an operational definition. Its major crucible of innovation was the Correlates of War Project which was then based at the University of Michigan. This was the most systematic attempt by the empirical social sciences to measure the incidence of conflict across the globe by collecting and analyzing data on wars since 1816. Initially, the bulk of the work focused on interstate warfare, as had earlier research programs on conflict, such as A Study of War (1942) by the American political scientist Quincy Wright and Statistics of Deadly Quarrels (1960) by the eccentric British meteorologist Lewis Fry Richardson.44 But bracketing internal war from international war could not continue indefinitely, because, as the leaders of the Correlates of War Project admitted, “civil wars, insurgencies, and foreign interventions have come to dominate the headlines in our generation and now play as important a role in the international community as traditional interstate war.”45

Once the compass of the Correlates of War Project had expanded to include internal wars, the team needed to develop criteria for civil war—as against other forms of conflict—that could be used to sort the masses of data they had collected on conflicts stretching back to the Vienna Settlement in 1815. They sought a quantitative, rather than a qualitative, definition, as they put it, “to minimize subjective bias” and, more pointedly, to “facilitate the construction of a data set,” as a means of getting out of what they saw as the conceptual morass of competing and inconsistent definitions. So the definition of civil war devised for the Correlates of War would have a numerical cutoff point, a set of boundary conditions, some empirical criteria, and a great many problems:

sustained military combat, primarily internal, resulting in at least 1000 battle deaths per year, pitting central government forces against an insurgent force capable of…inflict[ing] upon the government forces at least 5% of the fatalities the insurgents sustain.46

This “deceptively straightforward” definition was designed to allow political scientists and others to create the large sets of data needed to analyze the incidence of civil war across time and around the world.47 It also excluded many conflicts that would have blurred the analysis because they did not fit the procrustean definition.

The core of that definition was empirical, not experiential: combatants and victims might believe they were trapped in a civil war, but until the death toll reached a thousand or antigovernment forces had killed at least fifty people, social scientists could tell them they were wrong, at least for the purposes of comparative analysis. A conflict had to be militarized, to distinguish it from other forms of internal violence like riots and coups d’état; it was only “primarily internal” because it also had to encompass internationalized civil wars, into which outside powers or forces had intervened; a thousand battlefield deaths annually defined it as a “major” civil war; it had to have two sides (but possibly only two sides), one of which was the existing government; and it had to be militarized on both of those sides, to distinguish it from, say, massacres or genocide.

There were many difficulties that could be discerned in this definition.48 The greatest, surely, was the number of conflicts it does not encompass. Consider the condition of being “primarily internal”—that is, internal to a sovereign state, recognized as such by the international community: this was specified as being “internal” to the metropole, in order, quite deliberately, to exclude postcolonial wars of national liberation, just as international legal protocols did at the time. Like those protocols, this stipulation would mean omitting a conflict like the Algerian War or, to go further back in time, the American Revolution as a civil war.49 A second problem was that the emphasis on metropoles also implied the existence not just of states but of nation-states in the “Westphalian” mold of territorially bounded sovereignty; by these terms, there could not really have been proper civil wars before roughly the early nineteenth century—let alone in classical Greece or ancient Rome, for example—because there were few states of the kind International Relations scholars might identify.50 Without unitary sovereignty and its external recognition, it seems, there can be no civitas,and hence no “civil” war.

Finally, the definition would exclude many conflicts thought of, by at least some of their participants and by external observers, as civil wars—for example, the Swiss Sonderbund War of 1847. One of the shortest and least bloody civil wars on record—it lasted only twenty-five days and claimed 93 lives by the best count—it was nonetheless thought to be, at the time, as it is now, a civil war.51 Likewise, it would exclude the Irish Civil War of 1922–23 (in which an estimated 540 pro-treaty troops died, along with perhaps 800 members of the army and an unknown number of Republicans).52 And it would not encompass the Troubles in Northern Ireland, for which the death toll was around 3,500 between 1969 and 2001, with a peak of 479 in 1972; indeed, the requisite total of 1,000 deaths was not reached until April 1974, five years into the conflict.53

The essential contestability of the Correlates of War Project’s definition became starkly clear during the Second Gulf War, when it was used in 2007–8 to prove both that there was a civil war taking place within the boundaries of Iraq—and that there was not. Vehement disagreement arose as to whether the category fit the facts on the ground. Representatives of the Bush administration and others, mostly neoconservative military strategists and political pundits, denied that the turbulence merited the name. Terrorism? Insurgency? Perhaps. But a civil war? Certainly not. Then, in July 2006, the Yale political scientist Nicholas Sambanis had announced in The New York Times that according to the standard social-science criteria Iraq was indeed experiencing civil war.54 By late in that year, there was no doubt among many inside and outside Iraq what was happening. The UN’s secretary-general, Kofi Annan, for one, told the BBC, “When we had strife in Lebanon and other places, we called that a civil war; this is much worse.”55

At the same time, various sectors of the U.S. media, including the NBC television network and newspapers such as The New York Times and the Los Angeles Times, were also calling the conflict in Iraq a civil war.56 Adel Ibrahim, a young Shi’ite sheikh, vehemently told the Times, “You need to let the world know there’s a civil war here in Iraq. It’s a crushing civil war…We don’t know who our enemy is and who is our friend.”57 When Turkey’s then prime minister, Recep Tayyip Erdoğan, was asked if he thought there was a civil war in Iraq, he replied, “Muslims kill each other just because they belong to different sects. This is a civil war, because I cannot make any other definition.”58 By the end of 2006, half of the American population agreed, according to a Pew Research Center poll in December 2006, which found that “more Americans said the current violence in Iraq was mostly a civil war than said it was an insurgency aimed at the United States and its allies.”59 A few months later, an analyst at Chatham House, the British think tank on international affairs, argued that there was not a civil war in Iraq; in fact, there were “several civil wars and insurgencies, between different communities and organizations”: Shi’as versus Sunnis, Sunnis versus the United States, Shi’as versus other Shi’as, Sunnis versus other Sunnis, Kurds versus non-Kurds, and so on.60

In testimony before the U.S. Congress in September 2006, the Stanford political scientist James Fearon had no doubt that “the rate of killing in Iraq—easily more than 30,000 in three years—puts it in the company of many recent conflicts that few hesitate to call ‘civil wars’ (e.g., Sri Lanka, Algeria, Guatemala, Peru, Colombia).” Fearon defined a civil war as “a violent conflict, fought by organized groups that aim to take power at the center or in a region, or to change government policies.” The violence in Iraq fit that mold, he argued. It had also passed a significant threshold of casualties and could be directly compared with other civil wars around the world since 1945—in Lebanon, Turkey, and Bosnia, for example.61 Such comparison could then help one to imagine what the future held for Iraq and the coalition forces stationed there. The results were not encouraging: civil wars last longer than other wars, ten years, on average; they often end only with a decisive military victory by one side or the other; leaving too precipitately after an intervention can only make matters much worse; power-sharing agreements usually devolve into violence when external guarantors to the agreements leave.

Fearon later revised the casualty figures upward, to 60,000, making “Iraq the ninth-deadliest civil war since 1945 in terms of annual casualties,” and noted that everything that could be learned from earlier civil wars suggested that the Bush administration’s agenda for Iraq was deeply misguided and likely to fail disastrously.62 These were grim prognostications; here was another costly effect of applying the category. Quite predictably, there was the equally adamant view that it did fit the facts in Iraq. In December 2006, the Iraqi prime minister, Nouri al-Maliki, briskly rejected characterization of the conflict as a civil war and accused Kofi Annan of “burnishing the image” of Saddam Hussein as a viable belligerent. What Fearon and others could see clearly, commentators on the other side hotly denied.63

Longue-durée history proved useful in rebutting those who wanted to discover civil war in Iraq. In March 2006, the conservative Iranian journalist Amir Taheri published a short but wide-ranging essay brimming with historical counterexamples. He denied that Thucydides had written about civil war and saw Cicero as the first to popularize the term. He boiled the Roman definition down to its chief feature: “that it pitted one group of Roman citizens against another, with no armed intervention by foreign powers.” This fit the wars of Sulla and Marius and Caesar and Pompey but not any of the many revolts in Roman history, such as Spartacus’s. Taheri inferred from the Roman example that civil war “must be fought on political, not religious and/or ethnic grounds,” and that “the conflict must be over the control of the whole disputed state and not aimed at splitting it into smaller units.” These stringent criteria denied the name of civil war to almost every conflict ever called one, including the U.S. Civil War and the Algerian Civil War (1992–2002), except for the Russian, Spanish, and Lebanese. His conclusion? “Iraq is not in a civil war” and would not be until the multiple overlapping conflicts in the country coalesced into a two-sided, secular struggle for control of the Iraqi state with no foreign assistance or intervention.64

Later in the year, the eminent British military historian and journalist Sir John Keegan and the American commentator Bartle Bull offered a similar and only slightly more expansive dismissal of the term “civil war” to describe the violence in Iraq. For any conflict to earn the designation “civil war,” they argued, “the violence must be ‘civil,’ it must be ‘war,’ and its aim must be either the exercise or the acquisition of national power.” That is, it must be fought within a state by organized bodies of combatants drawn from a single national population who use force either to grasp or to retain overall political authority within the territory. Like Taheri, Keegan and Bull discovered from history that instances of civil war were “extremely rare,” and they counted only five: the English Civil War (1642–49), the American (1861–65), the Russian (1918–21), the Spanish (1936–39), and the Lebanese (1975–90). Because the clashing parties in Iraq were fragmented, partly made up of non-Iraqi insurgents, and fighting for ends more contradictory or simply opaque than seizing national authority, Keegan and Bull concluded that Iraq’s troubles did not qualify as the modern world’s sixth civil war. Instead, they proposed that “the disorders in Iraq…are nearer to a politico-military struggle for power.” They also echoed Erdoğan’s analysis of the sectarian division between Sunnis and Shi’ites by noting, “It might be said that Islam is in a permanent state of civil war.”65

The Iraqi government and representatives of the Bush administration officially denied that Iraq was in the grip of civil war. For the Iraqi government, to acknowledge otherwise would have implied it had lost authority. For the U.S. coalition, the characterization would have entailed a bevy of strategic implications. It could mean deciding which side, Sunni or Shi’ite, the coalition should support as it made its bets on an internal struggle for dominance. It could also imply that the invaders had unleashed sectarian enmities that previously had no outlet and that events were spinning out of their control. If such instability continued, then much higher troop levels would be needed to prevent it from spilling over Iraq’s borders. Alternatively, a rapid but undignified withdrawal might be necessary to avoid being drawn ever deeper into indigenous dilemmas, which an alien presence could inflame but not resolve.

David Patten, an American sergeant with the Third U.S. Infantry Division in Baghdad who also held a doctorate in philosophy, warned in the summer of 2007 that “premature withdrawal could lead to a self-fulfilling prophecy, creating the conditions for a civil war that does not exist.” Loose talk about civil war was little more than posturing and inexactitude, he charged, as when the Democratic representative John Murtha announced in January 2006, “We’re fighting a civil war in Iraq,” or Iraq’s former interim prime minister, Ayad Allawi, lamented in March of that year, “If this is not civil war, then God knows what civil war is.” Yet, Patten warned, “debate should not be political. Precision matters.”66

Patten cited the U.S. Army’s 1990 field manual for low-intensity conflict:

civil war: A war between factions of the same country; there are five criteria for international recognition of this status: the contestants must control territory, have a functioning government, enjoy some foreign recognition, have identifiable regular armed forces, and engage in major military operations.67

This definition was designed to distinguish civil war from other kinds of conflicts on the grounds of being both more formally organized and on a larger scale than other forms of irregular warfare. Yet in few recent civil wars have both sides controlled territory and possessed a “functioning government,” let alone one that has been recognized internationally. The classification also conforms with a peculiar and rather rare instance of civil war, one involving armed forces more typical of interstate wars in the industrial era, the major example being the U.S. Civil War, which was hardly a model for most of the civil wars of the twentieth and twenty-first centuries. The utility of this definition is therefore rather limited for observers other than the U.S. Army, and even for them it is not clear of what use it would be under the present conditions of asymmetrical warfare.

Patten judged that only one of this definition’s five criteria—that the contestants possessed control of territory—had been satisfied in Iraq. When he deployed Melvin Small and David Singer’s social-scientific definition of civil war, however, he found that “Iraq has suffered sevenseparate civil wars in the last forty-five years” alone, a total spectacularly at odds with the five civil wars Keegan and Bull discovered in the whole of modern history since the mid-seventeenth century or Taheri’s three since the end of the Roman republic.68

All such attempts at precision are as doomed as they are illusory for the simple reason that civil war is an essentially contested concept. Indeed, even the separate ascriptions of both “civil” and “war” can be contested and, in most social-scientific analyses, change according to specific factors such as location, intensity, and duration. There is also no agreement about just which features of civil war take priority in its various definitions or even as to how they might be applied consistently to particular conflicts. Being precise, in the sense of using clear definitions, turns out to be inescapably political. The elements of those definitions as much as their application are always matters for principled dispute. This seems to be especially true of civil war—an essentially contested concept about the essential elements of contestation.


By the twentieth century, various conceptions of supranational community had generated dark fears and clear-eyed analyses of civil war as taking place on a regional, continental, and ultimately planetary scale. As the imaginative limits of civil war grew, they coincided with the knowledge that civil wars were themselves becoming more transnational in their form and global in their impact. The rueful cosmopolitanism of Fénelon was belatedly echoed by the Italian anti-Fascist writer Gaetano Salvemini, by the German painter Franz Marc, and by the economist John Maynard Keynes at either end of World War I. In September 1914, Salvemini warned his readers that they were now witnessing not a war among nations but a “global civil war” of peoples, classes, and parties in which no one could remain neutral; two months later, in November 1914, Marc called the Great War “a European civil war” for the first time.69 After the war, in 1919, Keynes regretfully recalled the common civilization in which France, Germany, Italy, Austria, Holland, Russia, Romania, and Poland “flourished together,…rocked together in a war, and…may fall together” in the course of “the European Civil War.”70 That term appealed to liberals and Marxists alike across the century as a means of describing the continuities between the two “world” wars, in Europe at least.71

Intimations of enmity on the eve of World War II had raised the fear of an “international civil war” between “reds and blacks” that cut across Europe’s countries.72 After the conflict came, this “gigantic civil war on the international scale” presented an opportunity for national liberation, according to the Indian Marxist M. N. Roy, writing in 1941–42.73 A similar idea occurred later in the century to the right-wing revisionist German historian Ernst Nolte, for whom the entire period from 1917 to 1945 was a “European Civil War” in the sense of a struggle within a single community torn between the opposed forces of Bolshevism and Fascism.74 The characterization of the entire span of the world wars as a single internal conflict would also find purchase in unexpectedly prominent places, such as when the former U.S. secretary of state Dean Acheson wrote of 1914–45 as a “European Civil War”—in effect, a civilizational war—that had intersected with an “Asian Civil War” in East Asia.75

Such an expansion of the idea of civil war was fostered by the Cold War, a conflict that itself would be called “a global civil war [that] has divided and tormented mankind,” as President John F. Kennedy put it in his second State of the Union address in January 1962.76 Two months later, in March 1962, Carl Schmitt spoke in a lecture in Spain about “the global civil war of revolutionary class enmity” unleashed by Leninist socialism.77 In Schmitt’s case, the expansive idea of civil war was not of Cold War vintage; it had been something of a term of art for him and his followers since 1939, as a critique of the pretensions of all revolutionary universalisms, whether applied to the French Revolution, the Revolutions of 1848, or the “present global world civil war” (as he called it in 1950), for example.78 More sympathetic to those legacies were the American Students for a Democratic Society, whose Port Huron Statement in June 1962 predicted that “the war which seems so close will not be fought between the United States and Russia, not externally between two national entities, but as an international civil war throughout the unrespected and unprotected civitas which spans the world.”79 Also sympathetic was Hannah Arendt, who argued in On Revolution the following year that the twentieth century had seen a new phenomenon arising from the interrelatedness of wars and revolutions: “A world war appears like the consequences of revolution, a kind of civil war raging all over the earth, as even the Second World War was considered by a sizeable portion of public opinion and with considerable justification.”80

“Global civil war” has more recently been used to denote the struggle between transnational terrorists like the partisans of al-Qaeda against established state-actors like the United States and Great Britain. In the hands of some of its proponents, this post-9/11 usage refers to the globalization of an internal struggle, especially that within a divided Islam, split between Sunnis and Shi’ites, that has been projected onto a world scale. As a broader metaphor for terrorism, “global civil war” has also been used to imply an unbridled struggle between opposed parties without any of the constraints placed on conventional forms of warfare, a return to a state of nature in which there are no rules for a war of all against all, and a peculiar species of conflict in which the boundaries between “internal” and “external,” intrastate and interstate, conflict are utterly blurred.81In this vein, the critical theorists Michael Hardt and Antonio Negri wrote in 2004 that “our contemporary world is characterized by a generalized, permanent global civil war, by the constant threat of violence that effectively suspends democracy.”82 This was civil war as what Schmitt had called the “state of exception”: the state of emergency determined by an all-powerful sovereign, in which the rule of law can be replaced by discretionary rule or martial law. “Faced with the unstoppable progression of what has been called a ‘global civil war,’ ” Giorgio Agamben observed in 2005, “the state of exception tends increasingly to appear as the dominant paradigm of government in contemporary politics.”83

Such metaphorical expansions of the ambit of civil war carry with them recognizable features from past ideas of civil war—for example, that of a defined community, a struggle for dominance within it, and an aberration from any normal course of politics or “civilization.” The idea of “global” civil war carries with it, additionally, an idea of universal humanity affirmed by discerning conflict within a single capacious community, that world city or cosmopolis peopled by hostile fellow citizens. In these regards, the recent language of global civil war appears as an intensification of long-standing, originally Roman ideas of civil war that were later broadened and intensified by cosmopolitanism’s expansion of empathy and broadening of horizons. Such a “global” civil war may not be susceptible to analytical measurement, in the way that social scientists trust other forms of conflict might be. Nor is it subject to legal regulation or humanitarian amelioration, as international lawyers believe that other wars not of an international character might be. Yet the internal complexities the term encompasses, the ideological freight it carries from earlier in the twentieth century, and the anti-Islamic connotations implied by some of its users, mark it, like “civil war” itself, as an essentially contested concept. In this regard, the recent discussion of “global civil war” can be seen as an intensification or a qualification of the competing conceptions of civil war that gave rise to it.

The idea of global civil war has gained added currency from the rise of transnational terrorism.84 This terrifying phenomenon brings warlike violence into the domestic sphere, most wrenchingly into the streets of the world’s cities—New York in 2001, Madrid in 2004, London in 2005, Mumbai in 2008, Sydney in 2014, Paris and San Bernardino in 2015, and Brussels in 2016, for example. The attackers are often demonized as alien to the societies they assault—even when they are natural-born or naturalized citizens of the country concerned—and hence they are not identified in the same manner as those who classically constituted the contending parties in “civil” wars or wars among fellow citizens. At the same time, the proliferation of various forms of irregular warfare and the more elastic conceptions of war devised to understand and combat them have helped to loosen and stretch the metaphorical reach of civil war. Finally, the long-term decline of wars between states accompanied by the rise of wars within them—at least proportionately to overall levels of organized violence—has encouraged the belief that there might be no wars but civil wars in the future.85 In the twenty-first century, all wars may indeed be civil wars, but for very different reasons and in senses more slippery—and more chilling—than those imagined by Torres Bodet in 1949.

If you find an error please notify us in the comments. Thank you!