The War of the World: History's Age of Hatred (101 page)

The War of the Worlds remains science fiction. The War of the World is, however, historical fact. Perhaps, like Wells’s story, ours will be ended abruptly by the intervention of microscopic organisms like the avian influenza virus, which could yet produce a worse mutation and pandemic than that of 1918. Until that happens, however, we remain our own worst enemies. We shall avoid another century of conflict only if we understand the forces that caused the last one – the dark forces that conjure up ethnic conflict and imperial rivalry out of economic crisis, and in doing so negate our common humanity. They are forces that stir within us still.

In the introduction, I make the claim that ‘The hundred years after 1900 were without question the bloodiest century in history, far more violent in relative as well as absolute terms than any previous era.’ It seems worth substantiating that assertion, which is by no means beyond dispute. To attempt to do so is to enter a realm of great statistical confusion. Estimates for death tolls in twentieth-century conflicts are unreliable enough. Those for earlier wars are worse. Dividing such figures by estimates for population only tends to widen the range of possible error.

There are conceptual as well as empirical problems. The notion of violent death – as opposed to natural death – may seem straightforward to a modern reader. Yet many of the millions of victims of war, genocide and other acts of organized violence were not directly killed by a weapon operated by another human being. They died in famines or epidemics that (it seems probable) would not otherwise have happened had it not been for antecedent acts of ‘direct’ violence. Many of the death tolls calculated by historians are therefore the products of subtraction sums: the population before a war or other violent event minus the population after it, where census figures or credible estimates are available. Clearly, however, figures obtained in this manner are bound to include some deaths by natural causes. Moreover, it is debatable how far even those deaths authentically due to war-induced starvation or disease should be regarded as equivalent to deaths due to weaponry. It is not always possible to say for sure whether or not such indirectly caused mortality was an intended consequence of the original acts of aggression. And what of the unborn? Sometimes historians calculate the net demographic impact of a particular event by estimating a counterfactual population, that is the population as it would have been if there had been no war. Yet here too there is a tendency to inflate the death toll, by counting among the victims
‘people’ who were never in fact born. It is clearly a dubious procedure to juxtapose figures calculated in this way with figures based on, say, the number of soldiers recorded by military authorities as killed in action. Matters are further confused when aggregate casualty figures – including men missing (but not dead), captured or wounded – are confused with figures for battlefield mortality. In some wars, being taken prisoner or wounded amounted to a death sentence; in others it was a reprieve from the much more dangerous business of combat. As medical science has advanced, so soldiers’ chances of surviving battlefield injury have improved. But there has been no such progressive trend in the way prisoners have been treated. Finally, there is the problem of disorganized violence. In times of war and revolution, opportunities are more plentiful for individual acts of murder than in times of peace and political order. Yet this kind of violence is generally treated as a separate phenomenon from organized violence, rather than just another form of ‘deadly quarrel’, in the phrase of L. F. Richardson, perhaps the most methodologically careful statistician of modern violence.

When expressing non-natural mortality in percentage terms in order to allow for variations in population size, the choice of denominator is also problematic. Is it a worthwhile exercise to express the estimated death tolls of eighteenth-and nineteenth-century great-power wars as percentages of estimates for
world
population, when none of these conflicts was strictly speaking a world war (with the possible exception of the Seven Years War)? Would it not be more worthwhile to have country populations as denominators, so that we may compare, say, the proportion of Germans killed by the Thirty Years War with the proportion killed by the Second World War? Here, too, there are difficulties, not the least of which is the changing character of the entity called Germany. The Holy Roman Empire was a very different thing from the Third Reich. A large proportion – perhaps as many as one in thirteen – of the men killed while fighting on the side of the Third Reich were not German citizens, but members of other nationalities that had been recruited or drafted into the Wehrmacht, the SS and other auxiliary formations. Should we therefore narrow the political or geographical unit down still further, and compare mortality rates in regions or cities? Perhaps, but to do so is to risk concluding that, say, the massacre of the population of a village in German South-West Africa during the Herero Uprising was a more violent act than the destruction of Warsaw during the Second World War, since the dead in the former accounted for a larger percentage of the population
than the dead in the latter. Small denominators can produce large percentages, because it is on the whole easier to kill a hundred villagers than to kill a hundred thousand city-dwellers.

This, in turn, raises the question of destructive technology. Should we somehow adjust for the greater ‘bangs per buck’ of twentieth-century weaponry? Does it require a greater quantity of violence (though I remain unsure in which unit violence should be measured) to kill a hundred people with a machete than with a bomb? Finally, does intention matter? Is it worse to kill people out of racial or religious prejudice than to kill them in pursuit of a strategic objective? Should we allow for the fact that in some cases organized violence is asymmetrically perpetrated against defenceless civilians, while in others it is reciprocally inflicted by well-matched armies? To put it differently, is ‘genocide’ merely a term for a civil war in which only one side is armed? None of these questions is easily answered, as
The War of the World
makes clear.

There is no question that the twentieth century witnessed a mind-bogglingly large number of deaths by organized violence. Estimates for the total number, which can rest only on some heroic if not downright reckless assumptions, range from 167 million to 188 million. One survey of the available published death tolls concludes that one in every twenty-two deaths during the century was caused by the action of other human beings. But – as I seek to show in this book – lethal organized violence was highly concentrated in both space and time. Indeed, a distinctive feature of twentieth-century warfare noted in the introduction is precisely that it was much more intense (in terms of battle deaths per nation year) than warfare in previous centuries. So the interesting question is not really, ‘Why was the twentieth century more violent than the eighteenth or the nineteenth?’ but, ‘Why did extreme violence happen in Poland, Serbia and Cambodia more than in England, Ghana and Costa Rica?’; and ‘Why did so much more extreme violence happen between 1936 and 1945 than between 1976 and 1985?’ Altogether, the best available estimates suggest, somewhere in the region of 58 or 59 million people lost their lives as a result of the Second World War. That can be expressed as a percentage of the pre-war world population (2.6 per cent), though it should be borne in mind that many of those people who were living in 1938 died of natural causes by 1945 and some of the babies born after 1938 were killed in the war. Military and civilian death tolls varied widely from country to country in absolute and in relative terms, however. In absolute terms, as is
well known, many more Soviet citizens died violently between 1939 and 1945 than people of other nationalities – perhaps as many as 25 million, if not more. This suggests that more than one in ten Soviet citizens was a victim of the war, though it might be more accurate to say that one in ten was a victim of totalitarianism between 1939 and 1945, given the number of lives lost to Stalin’s domestic policies.
*
In percentage terms Poland was the country hardest hit by the war. The Polish mortality rate (total military and civilian fatalities as a percentage of the pre-war population) amounted to just under 19 per cent, of whom a large proportion were Polish Jews killed in the Holocaust. Among other combatants, only Germany (including Austria) and Yugoslavia suffered mortality rates close to 10 per cent.
†
The next highest rates were for Hungary (8 per cent) and Romania (6 per cent). In no other country for which figures have been published did mortality rise above 3 per cent of the pre-war population, including a number of Central and East European countries, Czechoslovakia (3 per cent), Finland (2 per cent) and Bulgaria (0.3 per cent). For four of the principal combatants, France, Italy, the United Kingdom and the United States, total wartime mortality was less than 1 per cent of the pre-war population.
‡
For the three West European countries, the First World War was, at least by this measure, a more costly conflict. Turkey was of course far worse affected by the First World War (some put the total mortality rate at 15 per cent, including the Armenian genocide), since it remained neutral during the Second. Note, too, that Japan’s mortality rate during the Second World War (2.9 per cent) was significantly lower than Germany’s, as was China’s (at most, 5 per cent). These differentials reflect two important features of the war. War itself was waged at a much
higher human cost in Central and Eastern Europe than anywhere else. The Germans fought to kill. Soviet commanders were also wasteful of the lives of their men. This region also witnessed exceptionally systematic violence against civilians.

The incidence of violent death in Central and Eastern Europe between 1939 and 1945 was high, but other conflicts came close. Between 9 and 10 million men were killed in the First World War, with Serbia and Scotland suffering the highest mortality rates, though the mortality rate was also high in the campaigns between the Entente and the Ottoman Empire, where disease was worse and reserves fewer. Estimates vary widely for the number of deaths in China attributable to Mao’s policies, but they must certainly have run to several tens of millions. The total victims of Stalinism within the Soviet Union may have exceeded 20 million. Mortality rates in excess of 10 per cent have also been estimated for Pol Pot’s reign of terror in Cambodia, as well as for the civil wars in Mexico (1910–20) and Equatorial Guinea (1972–79), and the Afghan War that followed the Soviet invasion of 1979. By one estimate, sixteen twentieth-century conflicts – wars, civil wars, genocides and sundry mass murders – cost more than one million lives each; a further six claimed between half a million and a million victims; and fourteen killed between a quarter and half a million people. In all, according to the Correlates of War Project, there were at least two hundred inter-state or civil wars between 1900 and 1990. Using slightly different criteria, the Stockholm International Peace Research Institute estimated that there were over a hundred armed conflicts in the last decade of the century, of which more than twenty were still in progress in 1999.

It might be argued that there are precedents in human history for such high rates of lethal organized violence. First, it is clear from archaeological and anthropological studies that pre-historic and pre-modern tribal societies were very violent indeed. The percentage of male deaths due to warfare among the Amazonian Jivaro Indians is known to have been as high as 60 per cent within the recent past. Rates in excess of 20 per cent have been recorded for at least five other tribes.

Secondly, there is reason to believe that two or three Asian tyrants perpetrated mass murder on a scale comparable with that inflicted by their twentieth-century counterparts. The exemplary violence meted out by the thirteenth-century Mongol leader Chingis (Genghis) Khan is said to have resulted in a decline in the populations of Central Asia and China of more
than 37 million – a figure which, if correct, is equivalent to nearly 10 per cent of the world’s population at that time.
*
Timur (Tamburlaine)’s late fourteenth-century conquests in Central Asia and Northern India were also notably bloody, with a death toll said to be in excess of 10 million. The Manchu conquest of China in the seventeenth century may have cost the lives of as many as 25 million people. It is important to emphasize, however, that the majority of victims of these conquerors almost certainly died from famines and epidemics arising from their disruptive incursions. The populations of the regions affected lived perilously close to subsistence, so that vandalism of irrigation systems or destruction of harvests could have devastating effects, particularly for urban centres. Nevertheless, these figures help to set the death toll inflicted by the Japanese during their conquest of north-eastern China (which is said to have exceeded 11 million) in some kind of long-term perspective. It seems likely that the hundred years after 1900 were the bloodiest century in European history, in relative as well as absolute terms. It is less certain that the same can be said for Asia, especially if wilfully causing a famine is counted as a form of bloodshed.

Thirdly, moreover, several pre-1900 Chinese rebellions and their suppression caused human suffering on a scale that may have matched or exceeded that inflicted on the people of China by twentieth-century civil wars. The eighth-century An Lushan Revolt is believed to have cost the lives of more than 30 million people. The mid-nineteenth-century Taiping Rebellion (1851–1864) – a peasant revolt led by the self-proclaimed younger brother of Christ against the Qing dynasty, which the rebels accused of capitulation to Western commercial penetration – was estimated by Western contemporaries to have claimed between 20 and 40 million lives. Also devastating to the provinces affected were the roughly contemporaneous Nien and Miao Rebellions and the Muslim rebellions in Yunnan and north-western China. Here, once again, death tolls have to be inferred from provincial and local censuses taken before and after the rebellions. In some cases the declines seem to imply mortality rates ranging from 40 to 90 per cent. At least some part of these declines in population must surely have been due to emigration from, and reduced fertility in, ravaged areas. Still, there clearly was very large-scale
organized violence, not least in the way the rebels were systematically exterminated by Qing commanders. Famine was a direct consequence of the scorched earth policy used against the Taiping rebels’ ‘Celestial Kingdom of Great Peace’ centred in Nanking. One hypothesis in
The War of the World
is that the worst time for an empire – in terms of the loss of human life – is when it begins to decline. This is the period when rebellions are most likely, but also when the authorities are most likely to resort to exemplary brutality. The evidence suggests that this was already painfully obvious in China a century before it became apparent in the rest of the world. Another way of thinking about the twentieth century, then, may be to see it as a Western version of Qing China’s nineteenth-century death throes.