In the twentieth century, therefore, war's chickens came home to roost. The attempt to treat mass killing as an extension of rational politics began to founder in the killing swamps of Flanders, the burnt-out German cities and the irradiated ruins of Hiroshima, even before the nuclear age opened up. Warfare, on the grandest scale and in the hands of the most advanced, liberal states, repeatedly degenerated into little more than deliberate mass slaughter, first of soldiers and then of civilian populations. Strategy and politics wore very thin in the face of the enormities that their pursuit revealed.

The depth of the problem for war's viability as a social practice lies in the fact that these outcomes were neither accidental, nor deviations from its inner logic. On the contrary, they were direct and (in a broad sense) inevitable results of a classic process of war in modern conditions. Located within a competitive interstate system, fuelled by the technology, socio-economic organization and politics of mass industrial societies, slaughter was a predictable outcome of war. 

Supplied with unprecedented resources, destructive logic overrode all others, producing truly total war. Supplied with an ever wider range of human targets, warfare swallowed up (or threatened to swallow up) whole populations in a more or less indiscriminate fashion. The dialectic between discrimination and indiscrimination in modern war is instructive as to the meaning of degeneration in modern war. It involved simultaneously 

• the extended definition of the enemy as civilian as well as military;
• the deliberate targeting of elements of the civilian population as well as military forces;
• intensified means of destruction which killed more people more speedily and efficiently; 
• but also increasingly indiscriminate slaughter which killed people across broader areas and with less precision as to their membership of any enemy group.

Thus in the name of defeating each other's armies, First World War generals mowed down a generation of young men from across Europe. To destroy Chinese resistance, Japanese troops massacred civilian populations wholesale. To crush Soviet power, Nazi German forces slaughtered civilians and prisoners of war. To attack German industries in the Second World War, British leaders began to destroy whole cities. To crack German civilian morale, they obliterated tens of thousands, refugees and prisoners alongside citizens, soldiers and war workers, in huge firestorms. To defeat Imperial Japan, American leaders blew up two entire urban populations with atomic weapons.

In the all-out use of massive force there was already in the middle of the last century, therefore, an overwhelming tendency for targeted violence to produce indiscriminate mass slaughter on a previously unimagined scale. Politicians and armies have tried since 1945 (and especially since 1989) to reinvent more limited roles, and means, of military force. To a certain extent they have succeeded. But warfare remains fundamentally compromised by this degenerate tendency, which has led war, I shall argue, to produce what has come to be known as genocide. 

Since this argument is controversial, I shall return to it only after a general examination of the three major forms in which large-scale political violence is identified: war, revolution (including counterrevolution) and genocide. I shall first examine the problems involved in delineating these categories, illustrating them by a discussion of revolution, and then deal more fully with the nature of genocide.

Most writing assumes that the categories of war, revolution and genocide can be clearly delineated, even if we need to investigate the relationships between them. I argue, in contrast, that the categoric distinctions are only partial, and have been all too easily blurred in historical practice. In this sense, the genocidal tendency in war is a manifestation of the internal linkages between the principal kinds of political violence.

War, revolution and genocide are not 'phenomena' in the sense that volcanic eruptions are natural events. All are, in contrast, general ways of representing different kinds of intentional actions by individual and collective social actors. All three concepts have the dual meaning of referring both to the courses of action of specific collective actors, and to the clashes of the actions of opposing actors.

The intentions of those who make war, revolution and genocide are generally supposed to differ sharply. In this sense, there are well established distinctions between the three concepts:

• revolutions are social upheavals that overturn the organization of political power; 
• genocide is the destruction of a social group by political power;
• both of these are clearly different from war, the contest of force between two state-like organizations.

However we should not regard these distinctions as rigid. The tendency to regard revolution, genocide and war as radically different types of social action rests on three fundamental errors. 

The first error is to regard the distinction between 'inside' and 'outside' the state as fundamental. From this point of view, states are established centres of power. It is assumed that contests between two or more states (or groups of states) are of a different kind from the contests that occur between states and social groups within states. However, states are only centres of concentrated social power; they rest on complex supporting relations in society. Conversely, social groups may produce state-like organizations, which form the basis of new state centres, for example in revolutions. Hence the differences between the two types of conflict may not be so great in practice.

The corresponding second error is to regard war as paradigmatically a state-state contest, but revolution and genocide as state-society contests. In reality, most wars are at least partially civil wars, and even the most clearly interstate wars always involve tenser relations and violent clashes between state power and society. In total war, as we have seen, war is normally waged on the enemy society as well as enemy state. Likewise revolutions, at least in their most advanced stages, acquire the character of a conflict between two centres of state power. They often occur in contexts of war and have taken increasingly militarized forms. Genocide is often assumed to involve states inflicting absolute violence on helpless populations. In fact, as we shall see in more detail later in this book, it mostly occurs in the context of struggle - revolution, counterrevolution and especially war - and it is mostly ended by war.

The third error is to regard the asymmetrical characters of power struggle in revolution and genocide as a definitive contrast with the symmetry of war. In fact, wars too, between state centres as well as between established states and insurgents, always involve asymmetries of power. The defining feature of war is not that the opposing forces have similar characters or comparable strengths, but that the two sides share the aim of destroying the enemy's power in a given context.

In this light, revolution and genocide not only involve elements of war and have clear affinities with it. Although both involve more than war, they are in the end variants and often phases of war, in which the relations of state and society, and hence the nature and role of mass killing, are posed in unique ways. But both revolution and genocide are contests that share two defining characteristics of war:

• the opposing sides aim to destroy each other's power;
• they are prepared to engage in mass killing of the enemy in order to achieve their goals.

In revolutions, emergent centres of state power and social groups that support them aim to destroy the power of established state institutions. Established regimes and social groups that oppose change equally seek to destroy revolutionary power. But revolution was classically conceived, like war, as a process. The power necessary to overthrow established state institutions had to be mobilized, through social struggles and political organization. Only in this way would the basis for a new form of state emerge.

Insurrectionary war was classically considered the final stage of revolution, which occurred when the emergent revolutionary centre acquired sufficient strength to openly challenge for power. Insurrection, the military phase, was the conclusion to the process of revolution, and civil war was the probable outcome of the reaction to insurrection by the old power. In this sense, Leon Trotsky could claim that the revolutionary process itself was only minimally violent. The extent of violence depended, he argued, principally on the extent of the resistance to change from counter-revolutionary forces. Certainly counterrevolution has often involved extensive slaughter. 

However revolution itself has also inceasingly taken the form of war, in guerrilla warfare (see box 5). Revolutionaries have pursued armed struggle not as a conclusion to political struggle, but as a central means of that struggle from the outset. Likewise established power has used force not merely to defeat open insurrection, but to stamp out revolutionary forces and terrorize its actual or potential social supporters. As revolution became armed struggle, counter-revolution became counter-insurgency. In this sense there has been a radical change in the character of many revolutionary processes. 

By the mid-twentieth century, revolution had become more clearly a form of war. Moreover the mass killing of civilians as well as combatants was often normal on both sides. Like the mainstream military tradition of total interstate war, the struggle of revolution and counter-revolution became increasingly genocidal. Only in the largely peaceful democratic revolutions of the emerging global era - for example, the 'velvet' revolution in Czechoslovakia in 1989 - in which those seeking change have sought to transform rather than overthrow established states, has the idea of revolution begun to be rescued from its total implication in war.

It might be thought that genocide was completely different from either war or revolution. The conventional definition (see box 6) is the deliberate destruction of a people, above all (but not only) by means of killing members of the group. Uniquely, the definition of genocide depends heavily on an international legal document. The genocide convention was agreed in 1948 by the victors of the Second World War. Accordingly, it took the Nazis' extermination of the Jews as its standard. It implicitly drew a particularly sharp line between genocide and other forms of mass killing.

Thus the convention said that genocide was about the destruction of national, racial and religious groups. It excluded the annihilation of groups defined by other characteristics such as class or political affiliation - so that the extermination of the 'rich' peasants (kulaks) or eastern European political elites by Stalin's Russia could not be counted. It limited genocide to the deliberate destruction of national, racial and religious groups as such. It thereby separated genocide from even the most degenerate war, such as the annihilation of civilian populations for strategic reasons. As we have seen, this was what the Allies themselves had practiced, and it was about to become the norm of the new nuclear age.

It is clearly true that extermination of Germans and Japanese as such was not, for the Allied governments, an end in itself. But mass killing of civilians was an intended and desired consequence of the bombing to produce German and Japanese surrenders. This fine line, which separated Allied mass slaughter from Auschwitz, remained sacrosanct in the convention. Subsequent debate has tended to emphasize, moreover, the uniqueness of the Holocaust's horror, in comparison to other mass slaughters, even those committed by the Nazis. This has helped to enshrine a narrow idea of genocide in legal, popular and even academic understanding.

There are grave problems with this model. In the same year that the United Nations adopted the genocide convention, it also adopted a universal declaration of human rights. From a universal human standpoint, it is untenable to lay down that the destruction of some sorts of human groups, such as races, nations or religions, should be regarded as a particularly heinous crime, while that of others, such as classes, professions or political groups should not. 

The restricted international definition gives special status to the former groups. Given that perpetrators of mass killing often target both kinds of group, in some combination, the restriction often makes little sense in practice. Are we to say, for example, that when the Khmer Rouge targeted people because of their Vietnamese origins, they were practicing genocide, but when they killed people because of their education or social class, they were not? Such a legalistic distinction would fragment the significance of the Cambodian 'killing fields' and stop us understanding the process as a whole.

From the standpoint of historical analysis, this definition of genocide is also in danger of taking the standards of its perpetrators, the genocidists, for reality. Certainly, we cannot explain what these people do without grasping the particular mentality that defines certain social groups as enemies, to be defeated by mass killing and other victimization. And yet genocide, like war, is not just a course of action pursued by power, but a clash of social power and experience. We should define genocide, therefore, by the experience of the victims as well as the mentality of the genocidists. 

The common experience of genocide is of cumulative discrimination of targets leading to indiscriminate slaughter. For victims, the threat to their social, personal and physical existence is very similar regardless of the particular reasons for it in the minds of their killers. The reality of killing in practice has a deeply arbitrary aspect, often appearing 'senseless'. If Nazism is our model of genocide, we need to put alongside the clinical images of selection for the gas chambers the roadside massacres and mass burnings of villagers - often Jews and non-Jews alike - by the Einsatzgruppen in Soviet territory. From the standpoint of the victims, genocide is deeply arbitrary. Whether the Nazis were killing you because you were a Jew, a Slav, a Soviet citizen or a Communist, or all or none of these, was not so important as the fact that you were going to be killed.

All genocidists kill for a range of reasons and at the same time, seemingly, for none in particular. As we shall see later in this book, they target young men because they are potential fighters or producers; old people for the opposite reason, their uselessness; women for reasons of the significance of sexual power. But in the end, killing is killing and tends to lead to more killing. Slaughter can appear utterly casual, arbitrary and senseless. There is (il)logic in terror and murder that surpasses the pseudo-rational discrimination dictated by strategy and political ideology, even as it fulfils them.

Our understanding of genocide needs to grasp this indiscriminate character as well as the categorical discrimination of genocidal killing - as of war. It is rarely the case that killers focus only on one kind of social target. The logic of terror tends towards the multiplication of enemies. Political enemies and social classes (commonly, educated elites) are often targets of genocidal regimes at the same time as ethnic or national groups. Jews were not only a racial enemy for the Nazis, but represented social privilege and political opposition. At all stages they were only one among the social enemies of Nazism, and not always the most threatened. Since genocide often takes place in the context of war, social enemies are often identified with international state enemies. Jews were linked to international finance and Soviet Bolshevism.

In the way in which mass killing partially transcends political logic, genocide turns out to be very similar to war. Just as 'strategic' bombing has led to relatively indiscriminate mass slaughter, so has 'racial' targeting. This is hardly a coincidence. The connection between war and genocide is not merely a strong causal correlation, which means that most genocides take place during or around wars. The link is not just external but internal.

The simplest way to express this is to say that genocide is a form of war in which social groups are defined as enemies. 

• In genocide, it is the aim of genocidists to destroy the power of social groups as such, rather than of organized state-like entities or of armed forces (although the two may go together). 
• As the genocide convention recognized, albeit in the imprecise reference to 'in whole or in part', the destruction of a group does not necessarily entail its total elimination. The point of genocide is to destroy the social power of a particular group - usually to eliminate or drastically reduce its presence in a particular territory. 
• Typically this is achieved by some combination of measures - from discrimination, expulsions, and terror to killing. Together these measures produce the destruction of communities on a large scale.
• However as in war generally, the deep destruction of the enemy that constitutes genocide is defined by extensive mass killing among its members. 

Like war in general, and also revolution, genocide needs to be understood as a process. Genocide is prepared in complex ways within society, and many social forms have genocidal potential. However, genocidal policies, unlike war or even revolution, are generally regarded as illegitimate. Therefore its preparation is not openly celebrated, and the link between genocidal potential and the actual preparation of genocide is not always direct or conscious.

Indeed, institutions and ideas that are mobilized by genocidists are widely accepted in pre-genocidal periods. Many political ideologies and laws have genocidal implications, more or less explicit, as we shall see in more detail in Chapter 4. Racism, chauvinism, religious and class hostilities are very widespread, and expressed in state policies. Discriminations against particular groups are almost ubiquitous. Although they all have genocidal potentials, for the most part these potentials remain abstract. Such practices are not necessarily seen as means of genocide.

A major difference between genocide and war is, therefore, the way in which they are prepared in society. Formal institutions and practices of war-preparation are universal, socially recognized and culturally justified. Genocide-preparation, even more than revolution, exists on the margins of social acceptance. The specific machinery of genocide, like the machinery of insurrection in classic revolutions, is usually a novel construction, begun only in the period prior to its use. Indeed, this machinery generally develops in contexts of pervasive violence, linked to war and revolution. Although perpetrating genocide usually requires specific kinds of organization, it is mostly carried out by or through the general state machine. Thus it often utilizes the machinery of war and takes place under its cover.

We can use genocidal as an adjective referring to genocide, just as we use military referring to war. Genocidal practices are those that treat social groups as enemies whose power may have to be destroyed. It is important to identify the conditions under which forms of hostility towards specific social groups (e.g. racism), move from a general, pre-genocidal to an actively genocidal significance. This is, however, a difficult area. The change occurs when states, or state-like movements, begin to organize the conditions for widespread violence against certain groups. In this context, any hostility against enemy groups begins to take on directly genocidal significance. For example, we know with hindsight that the previously abstract genocidal potential of German anti-semitism started to actively threaten genocide, when it was mobilized in the propaganda and street violence of the Nazis.

Moreover we can see now that terroristic episodes involving relatively small numbers of killings, such as Kristallnacht on 11 November 1938 (when Jewish shop windows were smashed and about 90 Jews killed) often threaten, all they do not make certain, much larger-scale slaughter. As a result of such experience we can now read genocidal dangers in other movements, even if full-scale genocide may still be some way off. Genocide may be discerned, therefore, in relatively limited mass killing, short of 'all-out' slaughter. The concept of genocidal massacre has been proposed to cover smaller incidents, which are often a prelude to larger-scale genocide. These are particularly indicative where they are linked to widespread targeting of given social groups. It is particularly important to identify the early stages of genocide, in order to halt the escalation in policies designed to destroy social groups - which may lead to mass killing.

Thus in the broad sense of genocide proposed here, Nazi policy was already becoming genocidal in the 1930s, in relation to Jews - as well as to the homosexuals and mentally defective people who were physically attacked in this period. It was clearly so during the period 1939-41, when Jews were transported across the territory of conquered Poland, and gradually forced into ghettos. Not only did maybe half a million Jews, as well as many others, die of ill treatment, starvation and disease during this period. These were already policies designed to comprehensively destroy the social power, and existing conditions of social life, of the Jews, well before any comprehensive plans for slaughter were evolved. Moreover, it was not only Jews, but Poles in general, who were targets of genocide in this period.

If genocidal policies and limited genocidal massacres do not lead to all-out mass slaughter, this may be because the aims of the perpetrators can be achieved with limited violence. It may also be, however, because genocidal processes are interrupted. This in turn reflects the fact that genocide, however successfully perpetrated in the short term, is never uncontested. However helpless many of the victims may be, much of the time, there are always resisters, always opposition - either from the population that is directly attacked, or from other forces that see their interests implicated in what is going on.

This, in the end, is another reason why we should understand genocide as a form of war. However unequal the initial struggle between the state that wishes to destroy and the groups whose power and lives are attacked, in the end that struggle is likely to become more balanced. Groups targeted for destruction cannot but resist, and often fight. And since genocide usually occurs in the context of wider warfare, victims will look for and often find more powerful allies. Genocide is generally ended in war, by more conventional military forces: as indeed Soviet and Anglo-American armies ended the Holocaust.

I have argued that war, revolution and genocide are closely related modes of organized political violence. In the end, revolution and genocide can be seen as distinct forms of war. The distinctions are often narrow, and they should not be regarded as absolute categorical differences. Nevertheless they have important continuing significance.

Between strategic bombing and the Holocaust, for example, there is not a simple difference of war and genocide. But there is a difference: between degenerate total war (that has become genocidal), and genocide as a distinctive form of war in the context of larger war. The difference boils down to who was the enemy - whose power the perpetrators of killing intended to destroy. For the Allies, the prime enemy was still other states; enemy civilians were derivative enemies and their mass killing was (however terribly) incidental to the major goals of the war. For the Nazis, Jews - and many other social groups - were themselves enemies, whose destruction ranked as an aim alongside that of the states that they fought.

The importance of such a distinction is, first, that it mattered in practice. When the Allies defeated the Nazi German and Imperial Japanese states, they stopped killing Germans and Japanese civilians. The Nazis would probably not have stopped killing Jews until they had eliminated virtually all Jews in Europe. The paradox of course is that the Allies, through their degenerate war, stopped the open genocide of the Nazis.

The other importance of the distinction is that it continues to reverberate. Mainstream war-preparation became hyper-genocidal with the prospect of all-out nuclear exchanges. But this has led to partially successful attempts by the dominant West to renew military power in more limited forms of warfare, with more discriminating technology and strategy. There is a continuing genocidal implication in strategic bombardments, like those of Baghdad and Belgrade in 1991 and 1999 respectively, but it is muted by tactics and technologies that aim for precise targeting. Arguably UN non-military measures that have indirectly targeted civilian populations, such as economic sanctions, have produced more genocidal results than has Western military power as such in recent years.

In contrast, direct forms of genocide have come to dominate the war strategies of many local states and state-like movements in the non-Western world. From Bosnia to Rwanda, genocide has become a large part of the practice of war. The aim of state elites has been to destroy the power of certain groups, plural urban as well as ethnic communities, and with this the lives of large numbers of their members. In Rwanda, this aim was unprecedentedly open, literally broadcast over the airwaves to simultaneously mobilize perpetrators and threaten victims. In other cases it has been only slightly veiled by more familiar practices of denial.

The contemporary argument about war is framed, therefore, by a new version of the old dilemma. In its simplest form, the question appears to be, can war end genocide? Our discussion should lead us to question this categorical opposition:

• We could reframe it as, can one kind of war end another? 
• Can mainstream military forces create forms of war, or of military force short of war, in which its degenerative tendency is controlled? 
• How else can genocide, the war of states against social groups, be halted? 
• Can the degenerative tendency of war be overcome within the military tradition, or only by far-reaching political change that will gradually render war superfluous?

These questions will reverberate throughout this book. In showing how war has degenerated and is linked to genocide, I am posing once again the legitimacy of war in human affairs. Strategic theorists and military buffs may find the durable features of warfare the most important. However war is not just a fact of life. It has become a huge problem. Mass killing as a key means of resolving political issues is increasingly unacceptable. War is not yet over, but realizing that age-old aspiration is a central issue of our times.