More collective violence was visited on the world (in absolute terms, and probably per capita as well) in the twentieth century than in any century of the previous ten thousand years. China’s Warring States period, Sargon of Akkad’s conquests, the Mongol expansion, and Europe’s Thirty Years War were times of terrible destruction. But earlier wars deployed nothing comparable to the death-dealing armaments and state-backed exterminations of civilians characteristic of twentieth-century conflicts. Between 1900 and 1999, the world produced about 250 new wars, international or civil, in which battle deaths averaged at least 1,000 per year. That means two or three big new wars per year. Those wars produced about a million deaths per year. Altogether, then, about 100 million people died in the twentieth century as a direct result of action by organized military units backed by one government or another. A comparable number of civilians likely died of war-induced disease and other indirect effects.1

The century’s two world wars are an important part of this story of mass violence: battle deaths in World War I amounted to about 10 million across all theaters, and the totals for World War II ran close to 15 million. But the 1990s alone brought catastrophic violence to the Caucasus, former Yugoslavia, Liberia, Sierra Leone, Angola, Rwanda, Congo-Kinshasa, Haiti, Colombia, Iraq, Algeria, Lebanon, Palestine, Israel, Yemen, India, Afghanistan, Sri Lanka, the Philippines, Indonesia, Papua New Guinea, and Laos. Rwanda alone produced close to a million deaths through collective violence during the 1990s.

I will not try here to provide an explanation of this history. Instead, I will sketch four points about a sea change in the nature of collective violence during the second half of the twentieth century, with ominous implications for the future course of global politics.

1. Since World War II, civil war has displaced interstate war as the dominant setting for large-scale violent death.

2. The prevalence of violence by such irregular forces as militias, mercenaries, and paramilitaries has greatly increased, even as the shift away from interstate war has reduced the prevalence of top-down violence by regular national armed forces (including police).

3. Irregular forces have often engaged in sudden attacks on civilian targets—which we can reasonably call bottom-up terror—but the same groups that have employed such terror tactics have also commonly used other political tactics.

4. On the whole, participants in collective violence, including terroristic violence, have struggled for power and profit in ways that overlap with the politics of their less violent counterparts.

In short, violence and terror have maintained close affinities with traditional political struggle. Like conventional war, they represent the conduct of politics “by other means.”

Shifting Patterns of Violence

During the first half of the twentieth century, massive interstate wars produced most of the world’s political deaths, although deliberate efforts of state authorities to eliminate, displace, or control subordinate populations also accounted for significant numbers of fatalities. During the century’s second half, civil war, guerrilla and separatist struggles, and conflicts between ethnically or religiously divided populations increasingly dominated the landscape of collective violence. Between 1950 and 2000, civil wars killing half a million people or more occurred in Nigeria, Afghanistan, Sudan, Mozambique, Cambodia, Angola, Indonesia, and Rwanda. Over the century as a whole, the proportion of war deaths suffered by civilians rose startlingly: according to one estimate, they rose from 5 percent in World War I to 50 percent in World War II, all the way to 90 percent in wars of the 1990s.2 While the precise numbers may be disputed, the shift is clear.

These changes in patterns of violence reflect large-scale political changes after World War II. At first, decolonization and the Cold War combined to implicate the major Western powers in new states’ domestic conflicts: in Indochina for the French and the Americans; in Indonesia for the Dutch; in Malaysia for Great Britain. Most former European colonies began their independence as nominal democracies, then rapidly moved either to single-party oligarchies, military rule, or both at once. Military coups multiplied during the 1960s, as segments of national armed forces bid for their shares of state power.

Coups became less common and less effective from the 1970s onward. With backing from great powers, however, post-colonial rulers began to consolidate their hold on the governmental apparatus, to use it for their own benefit, and to exclude their rivals from power. In the process, dissidents (often backed by international rivals of the power that patronized the ruling party) turned increasingly to armed rebellion; they sought either to seize national power or to carve out autonomous territories of their own. Civil war became increasingly prevalent.

Figure 1 represents the number of large civil wars taking place in any given year between 1960 and 1999.3 While the number of independent states rose from about 100 in 1960 to 161 in 1999, the number of civil wars expanded much more rapidly. An early peak arrived in 1975, with substantial civil wars in Angola, Burma, Cambodia, Ethiopia, Indonesia, Iran, Iraq, Lebanon, Morocco, Mozambique, Pakistan, the Philippines, Vietnam, and Zimbabwe. But civil wars continued to multiply until they peaked in 1992, when fully twenty-eight internal military conflicts were raging across the world. The number of civil wars fell off during the late 1990s, but internecine killing continued at much higher levels than those of the 1960s.

The post-war period also saw dramatic increases in the incidence of genocide (state-directed or state-authorized killing of populations identified by race, ethnicity, and/or religion) and politicide (wholesale killing of populations identified by political affiliation). In many cases, such as the Hutu-Tutsi conflict, genocide and politicide overlap. Large post-war waves of genocide and politicide occurred before 1980 in the USSR, China, Indonesia, Pakistan, Uganda, and Cambodia. During the 1980s they continued on substantial scales in Afghanistan, Uganda, El Salvador, Iran, Syria, Sri Lanka, Ethiopia, and probably Iraq.

In 1989 it looked as though the aging century might see some respite from the business of mass killing. Genocide and politicide seemed to be diminishing. From a peak of twenty-seven wars (both civil and interstate) underway in 1987, only fifteen were continuing, and not one new war above the thousand-death threshold began that year. War-sized conflicts in Angola, Colombia, East Timor, El Salvador, Guatemala, Iran, Lebanon, Ethiopia, Mozambique, Peru, the Philippines, Sri Lanka, and Sudan were winding down. Even in Afghanistan, where war had killed a million of the country’s 17 million people since the 1978 military coup, the carnage was declining as Soviet troops withdrew. The revolutions of 1989 in Eastern Europe were momentous but relatively bloodless; only in Romania did the struggle approach civil war.

But the downward “trend” was only a blip. In 1990–1 war erupted in the Balkans and Persian Gulf, Somalia broke into even more intense factional violence, and civil wars began to sunder Georgia and Azerbaijan. New or renewed conflicts in India, Kuwait, Liberia, Somalia, South Africa, and Tibet all exceeded the thousand-death threshold in 1990. In every instance (except perhaps Kuwait), non-military or paramilitary organizations killed large numbers of unarmed civilians on the basis of their group identification.

Since World War II, then, we have witnessed increased deployment of violence not by officially constituted national armed forces but by paramilitary forces, guerrilleros, death squads, secret police, and other irregulars, and increased direction of state-sponsored and state-seeking violence against civilians, especially whole categories of the population stigmatized for their religious, ethnic, and/or political identities. These trends greatly exceed population growth and the multiplication of independent states; they constitute an enormous increase per capita and per state. Paradoxically, a world war characterized by immense armies, elaborate technologies, centralized planning, and weapons of mass destruction generated a shift away from the efficiently segregated military activity that Clausewitz analyzed and advocated as the essence of rational modern warfare. The result is a series of decisive, frightening steps away from painfully-achieved distinctions between armies and civilian populations, war and peace, international and civil war, lethal and non-lethal applications of force.


From the seventeenth century to World War II, violence generally moved in two directions across the world: toward increasing deadliness of international war, but also toward increasing security and peacefulness of domestic life, including declines in both large-scale and small-scale killing. Both trends resulted from states’ increasing monopolization and perfection of coercive means. To be sure, Western powers continued their forceful conquest of non-Western areas through most of the period, and usually put down resistance to their rule ruthlessly. Yes, in times of war the distinction between international and domestic killing is often blurred, and if we include the effects of state actions on famine and disease the reversal will look earlier and less dramatic. Yet even with these qualifications, the period since World War II stands out for the prevalence of civil war, genocide, and politicide. How and why did these dramatic changes occur? Some of the causes are fairly clear:

More targets. With international backing, decolonization and separatist movements roughly doubled the number of formally independent countries, and therefore the number of governments over which dissidents and opportunists could try to seize control.

Weaker states. Absent support from colonial armies, many post-colonial regimes lacked the means of controlling their territories effectively.

External support. Throughout the Cold War, both the Soviet Union and the United States often subsidized domestic opponents of those regimes that aligned against them.

More weapons. Both Western countries and members of the Soviet bloc greatly increased their shipments—legal and illegal—of arms to the rest of the world.

Financial support. The enormous expansion of international trade in cocaine, heroin, sexual services, illegal migrants, dirty money, rubber, oil, diamonds, and other minerals provided sources of support for rebels, intervening forces from adjacent countries, and merchants who profited from weak and corrupt governments; note that markets for the contraband in rich countries, notably the United States, sustained much of this trade.

Emigrant support. In an era of improved communications and relatively inexpensive travel, increasing numbers of emigrants maintained contact with their home countries, and either supported opposition movements, provided outlets for contraband, or both.

In short, a larger number of weak states faced increasingly well-financed and well-armed opponents. The annual Human Rights WatchWorld Report, the Freedom House report Freedom in the World, and the Stockholm International Peace Research Institute SIPRI Yearbookall provide grisly documentation of the shocking results.4

Take the Human Rights Watch reports for Sierra Leone, Peru, Sri Lanka, Kyrgyzstan, and the United States in 1999. In Sierra Leone, the yearbook counts at least two thousand brutal murders by the Revolutionary United Front, plus widespread amputations of hands, arms, lips, legs, and other body parts—inflicted on people who failed to cooperate with the Front. In Peru it describes Shining Path raids on towns, often followed by gun battles and civilian massacres. In Sri Lanka, Human Rights Watch details how detachments of Tamil separatists hacked Sinhalese villagers to death, and how Tamil and Sinhalese armed forces engaged in a cycle of retributive killings. In Kyrgyzstan, reporters concentrated on hostage-taking by Muslim militants who demanded release of their Uzbek counterparts. And for the United States, the report describes the case of West African Amadou Diallo, who was shot at forty-one times and struck by nineteen bullets fired by officers from the New York City Police Department’s Street Crime Unit. The shooting put the unit’s practices, and the NYPD generally, under increased scrutiny.5

Despite its upbeat title, Freedom in the World is filled with bad news. Of Sri Lanka, for example, the report notes that the civil war has so far killed fifty to sixty thousand people, many of them civilians.

The SIPRI Yearbook, for its part, publishes a comprehensive catalog of the world’s larger interstate and civil wars. Among the countries I mentioned earlier, Sierra Leone, Peru, and Sri Lanka made the SIPRI armed conflict list for 1999, while Kyrgyzstan did not. SIPRI clearly sets a high threshold. Yet its experts still found twenty-seven major armed conflicts raging in 1999, including eleven in Africa, nine in Asia, four in the Middle East, two in the Americas (Peru and Colombia), and two in Europe (Kosovo and Chechnya).

Victims of those wars and other large-scale collective violence concentrate disproportionately in countries where most people live miserably in other respects as well. In addition, availability of valuable, portable resources (such as Sierra Leone’s diamonds and Colombia’s cocaine), large emigrant populations that supply aid to rebels and/or provide outlets for contraband, and external governments’ support for dissidents all increase the likelihood of large-scale violent conflicts.

The correlations of misery and conflict do not result from a general propensity of poor people to lash out in violence. They arise from the tyrannies large and small that flourish in undemocratic regimes in which the state has limited capacity to act for the common benefit.

To be sure, tyranny also arises in high-capacity and democratic regimes, including the United States. But collective violence there accompanies much smaller proportions of all public political interactions. High-capacity undemocratic regimes make a three-way split of collective violence. Their violent episodes divide sharply among external wars, citizen-state interactions, and various scattered attacks, including those that authorities call terrorism. Democratic regimes, whether low-capacity or high-capacity, often engage in external wars, but their collective violence generally takes more intermittent and fragmentary forms than the collective violence of undemocratic regimes. We may deplore the violence that occasionally disrupts domestic politics in democracies, but the overall record gives lovers of domestic peace strong reasons for preferring high-capacity democratic regimes.

Three conditions converge to foster violence in the world’s low-capacity undemocratic regimes. First, they have vulnerable governments that nonetheless yield power and profit to the groups that control them. Second, external advantages are available to rulers in the forms of revenue from valuable commodities and/or of international recognition for local dominance by their ethnic or religious faction. Third, specialists in coercion such as militias and mercenaries can operate more or less autonomously within their territories. With high-value targets that are relatively weak and relatively easy to organize against, the predictable result is widespread collective violence and civil war.

Terror as a political strategy

Violent groups often operate through acts of bottom-up terror: sudden attacks on civilian targets. Over recent decades, different variants of the strategy have killed large numbers of people, not only in the U.S., but also in India, Russia, Rwanda, Ireland, El Salvador, Algeria, Colombia, Uganda, the Caucasus, and South Africa. This roll call makes it clear that this kind of terrorism overlaps extensively with other varieties of politics. Terror often describes a single party’s conflict strategy rather than a distinct extra-political goal. Sudden attacks on civilian targets by irregulars occur across a wide range of political settings. In the United States, after all, we have recently seen not only suicide hijacking, but also anthrax letters, abortion clinic bombings, and eco-terrorism.

The U.S. State Department adds another annual report to the grim yearbooks I mentioned earlier: a document called Patterns of Global Terrorism. The State Department defines terrorism as “politically motivated violence perpetrated against noncombatant targets by subnational groups or clandestine agents, usually intended to influence an audience.” The report singles out attacks on noncombatant targets by other than regularly constituted national military forces, especially when someone broadcasts political claims on behalf of the attackers. The State Department document covers only those attacks that department specialists regard as crossing international lines—because the attackers came from outside the country, because they received substantial backing from outside, or because they assaulted foreigners.Figure 2 shows trends in the number of such incidents from 1980 through 2001.6

The frequency of designated international terrorist incidents reached a high point in 1987, and generally declined thereafter. When they voiced demands, attackers most often called for autonomy or independence for some sub-national population or region, replacement of existing governments, or redress of wrongs done to a particular organization. On the whole, international terrorist incidents rose and fell with the activity of independence movements. Whether the secondary rise that has occurred since 1996 will continue into the twenty-first century, and whether it represents a new manifestation of a political campaign remains to be seen.

It is instructive to look again at the data for 1999. For that year, the State Department’s annual count included 392 attacks, with 233 persons killed and 706 wounded, as compared with 741 killed and 5,952 wounded in 1998. The lower casualties of 1999 resulted from a shift away from massacres and toward hostage-taking and attacks on property, especially business property such as oil pipelines. Bombing (including bombing of pipelines) remained by far the most frequent form of attack, with almost half the total recorded attacks.

The report distinguishes carefully between sites of terrorist attacks and bases for attacks on American interests. In the first regard, during recent years, Asia and Africa each usually suffered far more casualties than all the other continents put together. As for bases, the report singles out Afghanistan, Pakistan, Iran, and Syria as the principal shelters and supports of international terrorist groups. Of Afghanistan, for example, it says, “While not directly hostile to the United States, the Taliban, which controls the majority of Afghan territory, continues to harbor Usama Bin Ladin and a host of other terrorists loosely linked to Bin Ladin, who directly threaten the United States and others in the international community.”

In 1999, then, the State Department anticipated a distinction that took on enormous political force after Muslim suicide squads crashed packed passenger jets into the Pentagon and the World Trade Center in September 2001: on one side, terror affecting U.S. interests; on the other side, essentially local and regional conflicts.

During the 1980s and 1990s, the great bulk of the world’s terrorism, as defined by State Department experts, occurred outside the range of U.S. interests. For the countries we looked at earlier, the 1999 report included these “Significant Terrorist Incidents”: in Sierra Leone, extensive hostage taking with demands for ransom; in Kyrgyzstan, more hostage taking; in Sri Lanka, no officially designated “Significant Incidents,” but plenty of smaller scale massacres and assassinations; in Peru, no “Significant Incidents,” but deadly clashes between Shining Path and government armed forces. Concerned with events elsewhere whose interactions crossed international boundaries, the document reports no terrorism within the United States.

For the countries covered, however, it implicitly makes two important observations. First, no one sort of action, person, group, or political cause dominated the episodes State Department specialists called terror. Second, the events in question ranged widely across different sorts of groups, settings, actions, and demands. Perhaps most impressive is the increasing prominence of kidnapping in the repertoire of those defined as terrorists. It suggests that opportunism is on the rise. In today’s world, terror and profit-taking often converge.

Chechen guerrillas, for example, have created a grisly business on the side: hostage-taking for revenge and (especially) profit. They grab journalists when they can, because the journalists usually work for companies that want their employees back and that can afford to make large ransom payments. The journalist Dmitrii Balburov, for instance, came close to being killed by his Chechen captors:

When they were told a Russian journalist was to be brought to them as hostage, they were all bloody-minded because, after the attack on Dagestan, bombing had begun in Chechnya, with casualties. “Now we’ll press him after full program,” they said meaning torture or even murder. “But when they pulled you out of the car and took off the hood and the ropes and we saw your mug, we lost all such thoughts—look he’s not Russian! Are you a Kalmyk?” Yes, I said, I’m a Kalmyk. “But you, Kalmyks, were also deported, why don’t you fight the Russians, why don’t you rise against them?” Later that guard told me that they would have killed me if I were Russian: they had twice decided to cut off my head and lay it out as a threat, as with the Englishmen, when the bombing got worse and when one of them had lost his home and family. My life, he said, hung on a thread, but each time they changed their mind.7

Balburov’s Asian features placed him on the rebels’ side of the them/us boundary Chechen soldier-bandits had adopted for predation and retaliation; he should have been on the side of the oppressed, against the Russians and Westerners who were oppressing non-Christians across a broad front. But his employment by Russian media pushed Balburov back across the boundary, suspending him between categories. In the new era of violent conflict, that is rarely a comfortable position.

During the late 1990s, the practice of hostage-taking accelerated, especially in the North Caucasus. Across the Russian Federation as a whole, about 1,100 abductions occurred in 1997, 1,400 in 1998, and more than 1,500 in 1999.8 As backing for their demands, hostage-takers often mutilated or executed their prisoners, videotaped the violence, and sent the videos to reluctant payers.

The most organized form of the hostage-taking business operated under the sponsorship of guerrilla commanders. In the end, a significant share (but not all) of the proceeds went to support the guerrilla movement, not merely to fatten the purses of individual hostage-takers. When a major military figure got involved in abductions, he rarely had his own forces perform them. Instead, like a Mafia boss, he provided protection, backup, and a reputation for brutality in return for a cut of the take. Away from the most highly coordinated part of the hostage-taking business, however, many independent entrepreneurs took advantage of civil war conditions to seize captives for ransom, rape, or revenge. Chechen hostage-taking involved large elements of opportunism.

At one point, American journalists singled out the practice as a prime example of terrorism. Yury Zakharovich, Moscow correspondent ofTime magazine, put it this way:

They did not search for an excuse for terrorism, but tried to trace how this terrorism had been provoked by the war. Violence breeds new violence. This circle horrified journalists because it became endless. One can trace this dominant motive when journalists covered a famous Basayev action in Buddenosk.9

Although American journalists typically interpreted terroristic attacks as part of a senseless spiral of action and reaction, a closer look suggests that the terrorism in question combines political bargaining and profit-taking. As Valery Tishkov points out, Chechen captors regularly mutilated their hostages, sometimes subjected them to lingering, spectacular deaths, and usually made sure other people learned about their cruelty. They certainly practiced terror. But they used terror to extort revenue and concessions from their enemies. Not all users of terror, by any means, bargain coolly over the lives of their victims. Some do. That is the point: terror is not a creed but a political strategy to extract resources and increase power.

What Next?

Did the crashes of passenger-packed aircraft into the World Trade Center and the Pentagon on September 11 undermine such conclusions? Certainly they proved that U.S. cities were far more vulnerable to large-scale lethal attacks than most Americans had thought. More dramatically than ever before, they breached the boundary between the generally peaceable domestic politics of high-capacity democracies and the incessant violence of many low-capacity undemocratic regimes. At least in the short run, they have rapidly shifted American foreign and military policy toward direct intervention. But, as we have seen, those attacks actually extended a pattern that has been developing since World War II.

The use of terror spreads across a wide variety of groups, ideologies, and targets. In the U.S. alone, anti-abortion activists, defenders of the environment, and a wide variety of anti-government clusters have used assassination, bombing, and wholesale property destruction against their enemies during recent decades. Terror-inducing attacks often take place as one element of a larger conflict; World War II bombings of Dresden, London, Tokyo, Hiroshima, and Nagasaki all had the tinge of terror. As post-war Ireland demonstrates, the same actors sometimes alternate between relatively conventional and terroristic conflict strategies.

We could make the same case for Sierra Leone, Peru, Sri Lanka, Kyrgyzstan, the United States, and the Basque Country. It is a serious but common error to assume that a class of people called terrorists, motivated by ideological extremism, perform most acts of terror. As the saying goes, one person’s terrorist is another person’s freedom fighter.

Nevertheless, the common sense assumption that users of terror tactics constitute a separate class of people has some foundation. Regimes have often authorized violent specialists such as paramilitary forces, secret police, and subsidized thugs to silence their opponents, but over the last few centuries those killers have usually operated in the shadows. Furthermore, when unauthorized groups have employed terror, they have commonly belonged to two categories of political actors: groups actively aligned with international enemies of the regimes they are attacking (the case of most suicide attackers in recent decades) and factions of larger dissident coalitions that have broken away from moderate control (frequently the case of armed activists in Ireland and the Basque Country). But the central point remains: the same sorts of political processes that generate other forms of violent conflict produce the special forms that authorities and horrified observers call terrorism.

Collective violence occupies a perilous but coherent place in contentious politics. It emerges from the ebb and flow of collective grievances and struggles for power. It interweaves incessantly with non-violent politics, varies systematically with political regimes, and changes as a consequence of essentially the same causes that operate in the non-violent zones of collective political life. Understanding those causes will help us minimize the damage human beings inflict on each other. In our own violent time, advocates of non-violent political struggle need all the help they can get.


1Statistics and catalogs synthesized from Jean-Claude Chesnais,Histoire de la violence en Occident de 1800 à nos jours (Paris: Robert Laffont,1981); Martin van Creveld, The Transformation of War (New York: Free Press, 1991); Kalevi J. Holsti, Peace and War: Armed Conflicts and International Order 1648–1989 (Cambridge: Cambridge University Press, 1991); Holsti, The State, War, and the State of War(Cambridge: Cambridge University Press, 1996); Mary Kaldor, New & Old Wars: Organized Violence in a Global Era (Cambridge: Polity, 1999); Evan Luard, War in International Society (New Haven, Conn.: Yale University Press, 1987); R. J. Rummel, Death by Government(New Brunswick, N.J.: Transaction Publishers, 1994); Charles Tilly et al., “State-Incited Violence, 1900–1999,” Political Power and Social Theory 9 (1995): 161–225.

2Simon Chesterman, ed., Civilians in War (Boulder, Colo.: Lynne Rienner, 2001).

3Paul Collier and Anke Hoeffler, “Greed and Grievance in Civil War,” working paper, World Bank, Washington D.C., 2001.

4Human Rights Watch, World Report 2000 (New York: Human Rights Watch, 2000); Adrian Karatnycky, ed., Freedom in the World: The Annual Survey of Political Rights and Civil Liberties (Piscataway, N.J.: Transaction, 2000); SIPRI [Stockholm International Peace Research Institute], SIPRI Yearbook 2001: Armaments, Disarmament and International Security (Oxford: Oxford University Press, 2001).

5Human Rights Watch, World Report 2000, 391.

6Computed from U.S. Department of State, Office of the Coordinator for Counterterrorism, Patterns of Global Terrorism 1999,< index.html; [17 May 2002]; Office of the Coordinator for Counterterrorism, Patterns of Global Terrorism 2000, <> [17 May 2002].

7Valery Tishkov, “The Culture of Hostage Taking in Chechnya,” inCountering Terrorism Through International Cooperation, ed. Alex P. Schmid (Milan: International Scientific and Professional Advisory Council of the United Nations Crime Prevention and Criminal Justice Programme, 2001), 348.

8Tishkov, “Hostage Taking in Chechnya,” 341.

9Ibid., 346.