History Statistics

Colonial Atrocity Porn

Thanks to Adam Hochschild, King Leopold of Belgium (1835–1909) is now known as one of history’s greatest mass murderers. At 10 million, he keeps company with Joseph Stalin (perhaps 30 million), Mao (perhaps 60 million) and that little house painter with the silly mustache (usually credited with 6 million).

Carrying out a genocidal plundering of the Congo [Leopold] looted its rubber, brutalized its people, and ultimately slashed its population by ten million.

King Leopold’s Ghost (1998), front flap.

It’s all in King Leopold’s Ghost (1998), so college students are taught and untold millions believe. The book has sold well over 600,000 copies, possibly over a million by now. Only it isn’t really so.


Hochschild was quite candid when describing his motives for writing the book. The civil war in the Congo had torn the country apart in the 1990s after Big Man Mobutu’s deposal. Irredentist Tutsis and other foreign adventurers were ransacking the country. Savage massacres and destruction dominated the news. The Rwandan genocide oozed down to the oceans. So Hochschild decided to put it all “in context”. Sure, things look bad now, in the mid-1990s. The crocodiles look fat, but you should have seen them bulging in King Leopold’s day! Rwanda was child’s play. The colonials murdered millions of Congolese. Actually, ten times that. So Hochschild claimed. It caught on like wildfire.

Before going any further, consider the units involved. Given his purpose, Hochschild is really forced to use units in millions. That is because the Rwandan genocide was initially calibrated at around a million. With that number in circulation, stating that King Leopold had murdered 10,000 people would risk bathos. Nor would 100,000 do, since that is still not even in the same league. Hochschild has to see 1,000,000 and raise it. He multiplies it by ten. Far better to go for broke.

Even back in the early 1900s, when Maxim guns had subdued the plains of Africa, King Leopold could find few to defend his personal colonial project in the Congo, with its forcibly staffed rubber plantations. Henry Morton Stanley, an undeniably great explorer but nothing if not ruthless, had been happy to get it started in the early 1880s. But by the 1900s the Boer War had given imperial projects a bad odor. An international campaign, led by the likes of Mark Twain and Roger Casement, had a field day with reports of gratuitous cruelty on the plantations and economic looting of the country for personal gain. Leopold’s paid defenders added some ineffective counter-blast. Eventually Leopold relented and in 1908 ceded control of the Congo to the Belgian state. A year later he was dead. Commissions of Enquiry were launched and duly reported.

It had all blown over by the end of the First World War. But all that written material was left lying around, waiting to be reused at an opportune moment. Enter Hochschild. Nothing in his book shows any evidence of original archival research, or any command of the subsequent literature, aside from Conrad’s Heart of Darkness. Everything he felt the need for had already been produced before 1920 and was to hand. He selected what suited him.

Which brings up the talk about genocide. This is a curiously slippery subject when one starts delving into the text to find out what, exactly, is being claimed and with what evidence. Genocide is a deliberate attempt to wipe out an entire race. It is from the beginning an absurd idea, since neither Leopold nor the Belgian state had any motive or known desire to do that. Hochschild recognizes this and inserts a concession to cover the objection.

Furthermore, although the killing in the Congo was of genocidal proportions, it was not, strictly speaking, a genocide. The Congo state was not deliberately trying to eliminate one particular ethnic group from the face of the Earth. Instead, like the slave dealers who raided Africa for centuries before them, Leopold’s men were looking for labor.

Except, the charge of genocide and 10 million murders is repeated numerous times in the book anyway, dangled before the jury but quickly withdrawn when objections would be obvious. Recall the claim that Leopold slashed the population by 10 million. One has to plough through nearly to the end of the text to track the supposed basis of that slashing claim.

It isn’t easy to kill 10 million people. The Hutu in Rwanda seem to hold the world record for the largest number of people killed systematically in the shortest space of time, using machetes (they really were slashing) and other hand-weapons, not even bombs. If King Leopold’s agents had killed that many, the logistics need to be explained. Unlike the Hutu, there simply were not that many colonial agents in the country. How did they do it? In fact the question does not need to be answered since, as one discovers later, Hochschild is not even claiming that the King or his men murdered all those people. Though he believes that many people were indeed murdered, and spends many pages rehearsing the details, he is conscious that the numbers involved from those sources are modest even if one concedes every single case he corrals. Recall that he needs units in the millions. The vast bulk died instead, we are told, from disease!

As with the decimation of the American Indians, disease killed many more Congolese than did bullets. Europeans and the Afro-Arab slave-traders brought to the interior of the Congo many diseases previously not known there. The local people had no time to build up immunities—as they largely had to malaria, for instance.

(Technical note: the Congolese did not “build up immunity” to Malaria, they inherit a sickle-cell genetic mutation that protects them at significant functional cost). Yes, disease. Nothing daunted, Hochschild shakes that off and talks everywhere else of murder and genocide anyway, going so far as to repeatedly drag the Gulag and the Nazi concentrations camps into the discussion at the same time. The techniques of yellow journalism.

It gets worse. Hochschild’s argument is really at an extra level of indirection: yes, the incursion of outsiders into the Congo brought the usual infectious diseases like smallpox, sleeping sickness and the flu, but maltreatment by Leopold weakened the locals so much that they were unable to withstand the consequences. And so Congolese deaths were greatly multiplied. Well then, one asks, by how much were these deaths supposedly multiplied? It is curious that the author makes no attempt to estimate this effect, which is really what he is claiming. The only figure bruited about is the total death count, not the part Leopold might be accountable for. Alright then, let’s be generous and overlook this sleight of hand. How many died overall? Where did the 10 million figure come from?

Now things get really strange. The Congo is a very large place, roughly the size of Western Europe. How does one actually count how many people died of disease over a period of say 20 years? The answer lies, Hochschild believes, in the first Congo census. In 1924 the inaugural census produced an estimate of the population at 10 million (we will pass over the vexed question of how accurate that figure is). We are told that a colonial official, Major Charles C. Liebrechts, guesstimated in 1919 that the Congo population might have halved since 1880, a year for which nobody has any solid idea how many people inhabited the country. There’s your figure anyway: 10 million missing. The reader may wish to read that again.

How, you may ask, can we know that the population halved since 1880, a rather precise starting date? Aside from the offhand guesswork of one man, who did not show his work, what evidence do we have to go on? Why had it halved since 1880 and not, say, 1865? Hochschild is aware that this is a potential problem for his argument, so he marshals some support.

The most authoritative judgment today comes from Jan Vansina, professor emeritus of history and anthropology at the University of Wisconsin and perhaps the greatest living ethnographer of Congo basin peoples. He bases his calculations on “innumerable local sources from different areas: priests noticing their flocks were shrinking, oral traditions, genealogies, and much more.” His estimate is the same: between 1880 and 1920, the population of the Congo was cut “by at least a half.”

page 233

Vansina has no systematic analysis either, which is why he resorts to vague hand-waving like “innumerable local sources”, inscrutable “oral traditions”, and imponderable “genealogies”. Innumerable local sources? There is a simple problem with his method, such as it is: ascertainment bias. You need a random sample, or close as you can get it, before you start making these inferences. If you deliberately track down reports of declining populations, you will find … declining populations. Were they declining all over? By how much? What was the population basis from which they declined? How do you tie that basis to 1880? This before even asking: were the flocks of priests declining because the population was declining, or because the power of religion was ebbing and flowing? Were people dead or migrating? The appeal to authority, “greatest living ethnographer”, is not going to paper over basic defects in method.

Mull over the supposed use of genealogies to measure population growth and decline, which is what Vansina claimed to have done, without telling us the crucial details. How would that work? Obviously you would construct family trees by explicitly asking people who their ancestors and relatives were, given that there were no written records, especially no marriage registers. Then you have to estimate missing survivors. A lot of people are involved in this data gathering, enough to account for an area the size of Western Europe. Enough to fairly represent 10 million people. You need to incorporate inbreeding estimates, and the well-known fact that most people ultimately leave no descendants since their family lines die out. Anyone who has constructed real genealogies can tell you that it is a laborious exercise for a single family, let along many thousands. Relatives die from natural causes, and you need some estimate of that attrition rate, etc. etc.

Invoking Dr. Johnson, let us have no more of this. Vansina simply repeated Liebrechts’ guess.

On the question of the spread of diseasesmallpox, sleeping sickness from tsetse fly, and later Spanish Fluat least two sources of infection are known. Arab slave traders like Tibboo Tip from Zanzibar were there by the 1850s and likely much earlier than that. Coastal dwellers and colonists followed rather later. The dynamic was repeated all over the world in much the same way. The destruction was a function of the lack of previous exposure and lack of natural resistance. Thus the island of St Kilda in the Hebrides was almost obliterated by smallpox in the early 1700s when passing ships transmitted it to the islanders. That so many were wiped out by it is a reflection of previous isolation. No special explanations are required to realize large mortalities. And it is certain that no matter how historical events played out, uninfected areas would have been infected over time.

Let us concede for the sake of argument every single one of the depredations gathered by the author, including punitive and by all accounts fairly brutal raids against those who refused to work for minimal reward on the company plantations. Hochschild really has no idea to what extent these events had anything at all to do with the impact of disease at a population-wide scale. Just as he has no real idea how many people at that population-wide scale died from disease in the first place. In any event we can be sure, and again Hochschild slyly concedes deep within the text, that King Leopold and the colonists had no intention of causing the massive deaths alleged, whether by neglect or not, since that would only rob them of greatly-needed labor on their rubber plantations. That labour was not easily replaced in those days, quite unlike the example of the Gulag so gratuitously introduced by the author, where endless supplies could be produced by troika.

Someone who wants to get an accurate estimate actually worries about being accurate, and goes to considerable lengths to test competing estimates. An honest investigator would have to account for reductions in deaths due to suppression by the state of internecine conflicts, endemic to the area before an effective state existed there. Hochschild went through no such exercise, which would be extensive. He had the figure he wanted to hand, so he banked it.

Constructing incendiary charges like genocide in this unscrupulous way, from unknown proportions of unknown quantities, is just a kind of information pollution for profit. But if the grounds for this claim of genocide and figure of 10 million murders are so flimsy, and the methods used so devious, how is it then that it has been repeated so often? In the age of Wikipedia, that is easily answered: because it could be. However it goes further than that. There is widespread demand for this type of material. A film runs through it. Numerous awards for court jesters are on offer, and the author has received them gladly. If Hochschild had not fed the demand, someone else would have been glad to. The best description for the genre is colonial atrocity porn. Intellectual smallpox.

Further Reading

Hochschild is no historian and his book shows absolutely no concern for accuracy. For a more balanced view of Leopold by a real historian, based on research, see Leopold II Of The Belgians: King Of Colonialism (1979) by Barbara Emerson.