Numbers (Unproofed Draft, not for quotation or citation)

Several reviewers have raised questions about the figures I have used throughout The Killing Age, at times part of a wider criticism—allegation really—of exaggeration and hyperbole. This is not surprising. (Generally, I avoid the temptation to respond to reviews.) Some of this is shaped by politics and ideology. Some, mostly on the right of the political spectrum, don’t want to recognize the dark side of capitalist development, or that globalized violence may have unfolded during an era usually associated with human progress, liberalism, and democracy. But as I argued throughout, this is one of the great contradictions of the age. One only needs to look at the rapid increase in the Atlantic Slave Trade from about 1750—much though not all of it powered by the British—to begin getting a sense of the issue. In other words, we have to hold two opposing observations in our heads at the same time.

It is also just easier to think of the period 1914-45 as uniquely destructive of human life, which of course it was. The killing fields of WWI, the bombing of Dresden and the atomic bombing of Japan in WWII, and of course the Holocaust, are all clear reminders of the violence of the “short” twentieth century (1914-45). (I would add Stalin here as well.) And remember there are survivors of that ghastly period.

At the same time, it is simply very difficult for many of us to fathom that so many people, and in so many areas such a large percentage of populations, perished in the period c. 1750-c. 1914. Or that violence was so much a part of the lives of so much of humanity.

At one level, this is a good thing. Most of us do not tolerate violence the same way our ancestors did. The period roughly from 1945-1970s, what some have termed the Great Acceleration, saw real progress, if not everywhere. But across large areas of the world life improved: people lived longer, and in better health, and they lived more secure lives.

So, I thought I would pen a few words here about numbers. I am also going to ramble a bit, so that readers can get a sense of how I went about making the calculations, the method in the madness, what I did and didn’t include, and so on. This is a first draft, so I will need to proof everything later.

At the very highest level of generality, over about the last half century population estimates have tended to trend upwards. This is not always the case, certainly, but generally where one number was in 1975 is larger today. So that’s the first thing to keep in mind. One reason is that what we used to consider was natural (say famine and disease) we now know was often linked to other processes, for example war. Basically, if you displace a population, you render that population more susceptible to disease, malnutrition, even famine. And that means people die unnecessarily. Two recent—and ongoing—examples are Gaza and Sudan. Going back in time, even mortality estimates of one of the best known and most studied modern famines, Ireland in 1845-52, have drifted upwards to somewhere above 1 million people. (NB. I did not include the Irish famine in my calculations when writing The Killing Age, as I was largely interested in areas outside of Europe. In retrospect, I probably should have.)

Second, generally our ability to measure has improved, so estimate for the more recent past are far better than they are for distant ones.

Slavery. Now let’s get more specific and examine on example, the Atlantic Slave Trade (AST). I begin with slavery because it forms a large part of The Killing Age. My argument was simply that the spread of weapons was associated with a spread of slavery and a rise in violence. (Slavery of course is ancient and emerged when one group was able to dominate another, and people were seen as a scarce good. Invariably it is associated with an unequal distribution of the means of destruction.) The gunpowder/slave connection is most clearly demonstrated for Africa and specifically the Atlantic Slave Trade.

Let’s dive in.

In 1969, Philip D. Curtin published his The Atlantic Slave Trade: A Census. (Disclosure. I studied with Curtin. This was not a pleasant experience. Thank the Lord for my advisor!) At the time of its publication the book generated a huge amount of controversy, precisely over the politics of counting. Curtin’s estimate was 9.6 million enslaved people left Africa’s shores for the Americas. He estimated mortality rates on ships at 15-20%. I number of people accused Curtin of deliberately undercounting (remember this was 1969). With dramatically improved research, partly pioneered by David Eltis and others, estimates are now above 12.5 million and, overall, with similar mortality rates. (Disclosure. Eltis and I were colleagues at Emory; I also worked with some of his students, and at least one student was under my direct supervision.) So, this is an increase of about 30% over Curtin’s 1969 estimate. That’s a big increase. I imagine the numbers will continue increasing, but rather incrementally.

Estimates for the Indian Ocean Slave Trade have also generally trended upwards, though it is very important that research here is far less developed than for the Atlantic World and this this vast area poses multiple challenges. Much the same can be said for slave trades in other parts of the world.

One of the conclusions we can draw from this research is that slavery was very common across vast reaches of the planet. And we also know that while there are some instances of people selling themselves into slavery, overwhelmingly enslavement is a violent process. This process was made vastly easier with the spread of guns, so long as your intended victims didn’t have them.

The process is rather simple. Enslavers used guns to intimidate or to shoot people to capture slaves. A contemporary example is Sudan, where there is considerable enslavement perpetrated by those with superior weapons, including drones. Of course, most enslavers didn’t, in effect, say, “well, I killed x people to produce y slaves.” So, some guesswork is inevitable if we want to explore the destruction costs of enslavement. In The Killing Age I offered a wide range.

Famine. Generally, estimates of famine deaths have, like slavery, trended upwards, for a variety of reasons. (See above mention re Ireland.) First, scholars have approached colonial records more critically. These records tended to be excessively narrow in how they reported famine deaths. Second, scholars have used more sophisticated tools to estimate mortality, a kind of “up-streaming” using census materials. Third, and quite importantly, scholars now include famine-related deaths from various diseases such as cholera and malaria. We know that severely malnourished people have weakened immune systems and easily succumb to infections. Crudely, if a person is suffering from severe food shortage in the context of famine, catches cholera and then dies, scholars include this death as a “famine death.” Finally, one must think about timespan, when a particular crisis finally ends. Did the South Asian famines beginning in 1876 end in 1878, or persist for another year or more? How you answer this has an impact on one’s estimates.

In other words, if you go by colonial records, you will get A, but if you take into consideration the issue noted above you get A+; the question is just how much more? And that “just how much more” changes from famine to famine. (One problem in the scholarship is that often it is not clear how scholars are doing the “counting.” This can produce considerable confusion.)

In South Asia, for example, it is just very difficult to get a good handle on eighteenth century famines, but we have a better sense of the famines unfolding in the next two centuries. There were three famines in the eighteenth century, and about 6-8 famines in the 1800s; this increase for the nineteenth century is crucial to my overall argument in The Killing Age. The famines in the eighteenth century may have killed as many as 23 million, though again the level of uncertainty is higher here than for later periods. Mike Davis, in Late Victorian Holocausts, came up with an estimate of two famines between 1876 and 1900 ranging from 12.2-29.3 million. Davis did not include other nineteenth century famines, I suspect because of his overall arguments about ENSO. His estimates for China ranged from 19.5-30 million, again only for the period from 1876, though we know there were other famines going back to c. 1750 (where mortality figures are hotly debated).

If we take just Davis’s numbers, we get a range for Asia of 31.7-59.3 million. (NB. Some have considered his book to be very controversial and have questioned his numbers and some of his arguments. My own criticisms of Late Victorian Holocausts are that Davis romanticizes the pre-colonial state and has a far too narrow accounting of famine due to his linking them to ENSO.)

As I mentioned, Davis did not consider the other famines, nor did he consider events like the Taiping Rebellion, loss of life at the hands of the British East India Company in India, and so on. In India, violence prosecuted by the BEIC likely killed anywhere from 2-5 million; once we consider deaths indirectly related to company violence these numbers certainly increase. And as I mentioned in The Killing Age, there were numerous wars prosecuted by various polities across the subcontinent. As for China, a recent study of the Taiping Rebellion estimates that “twenty million people lost their lives” (Reilly, The Taiping Heavenly Kingdom: Rebellion and the Blasphemy of Empire, introduction). Some estimates, including my own, are higher.

Of course, Asia includes much more than India and China. Some areas never suffered famine, for example Thailand, where wet rice cultivation may have expanded. Other areas, especially Indonesia, were dramatically different; here millions died in the nineteenth century. Still other areas, for example Burma, experienced considerable hardship though the scholarship here is rather uneven.

In other words, just working with these figures, and just China, we have 39.5-50 million dead. (Note, I haven’t even discussed the Boxer Rebellion or the Dungan Revolt.) This represents about 60% to 42% of the total estimates I made (p. 550) for all of Asia. Looked at this way, 66-118 million strikes me as a reasonable range; the numbers may well go higher, and not lower.

I am the first to admit there is lots to argue about here. Some might consider, for example, my categories are too expansive (famine, war, the sequalae from both, and so on). Note, as well, that I don’t discuss the twentieth century (the subject of another book), a period that saw extraordinary famines especially in India, China (esp. under Mao), and elsewhere. To return to my introductory remarks, what about genocide, particularly the Holocaust? Historical demography is contentious work, if only because it raises strong emotional responses. Work tends to elicit responses that one is exaggerating for political purposes, or underestimating for political and other reasons, for example appeal to that first great demographer, Malthus, who saw famine and disease as an inevitable part of human history.

But let me be clear here. What I was attempting to do in The Killing Age was to focus on violence and its direct and indirect role in the development of global capitalism and European dominance. And, crucially, I wanted to bring together human costs as well as the destruction of other large mammals.

So, let’s return to numbers. My total numbers for the period 1750-1900 ranged from 100.6 to 228.15 million, with the suggestion that these are conservative estimates. Davis, working on a much shorter time span of just 24 years instead of my 150 years, had a range of 31.7-61.3 million; these related to famine mortality, only included India and China, and in the Americas just Brazil. (Interestingly, his numbers range by the same percentage as mine do, the higher number twice the smaller one.) As I mentioned, my estimates did not include many areas of the world, such as Oceania, nor many formal wars, revolts, and so on. A more comprehensive census would include, to take just one example, the 200,000 odd dead in the Haitian Revolution of 1791-1804, and the tens of thousands who died outside Europe in the Seven Years’ War. And let’s not forget the Dungan Revolt (1862-77) in Northwest China; here total mortality ranges from 8 to 20 million, or deaths in French Indochina roughly from the 1880s through the 1910s.

In other words, my estimates were far from complete, very far.

Let’s look at Africa. Here some data are pretty darn solid, other data very flimsy. Many areas we just don’t have much of a clue. So, what do we know, or think we know:

Over 7 million enslaved people moved across the Atlantic;

About 2.5-3.5 million for the Indian Ocean slave trade;

About 1-2 million across the Sahara

(Recall that these numbers are tending to trend upwards.) So, just in terms of people lost to slavery, we are talking about 10.5-12.5 million. To which we have to add the number to died before they got to a port or the equivalent, and the mortality associated with the enslavement process. Now we know two things. First, enslavement is invariably a violent process. Second, guns were associated with enslavement during this period. This is where things get debatable. As I mentioned in the book, scholars have estimated that 20-50% died before reaching port. So, at the low end we are talking another 2 million to 2.5 million people. At the high end we are looking at more than 5 million. But let’s be conservative and assume a total of 12.5 million. And we haven’t added mortality costs (how many died to produce a single enslaved person); we will return to this later.

Now let’s discuss ever so briefly the colonial period. Scholars estimate 5-10 million died in just the Congo Free State in the early colonial period. To be conservative, let’s just take a middle point of 7.5 million. Note that we are already at 20 million people. Some other relatively easy ones: about 300-450,000 dead as the direct or indirect result of German colonialism in German East Africa, and anywhere from 60-80,000 dead in Southwest Africa. In the Sudan, we are talking about 1.5-2.5 million. Now we are up to about 22 million people. Add .5-1 million in Algeria, so our running total is 22.5-23 million.

If we bring in the mortality costs from enslavement, we add anywhere from 12-36 million people, bringing our running total to 34-59 million. And we haven’t begun to discuss West, East, Northeast and Southern Africa. According to a recent Stanford Ph.D. thesis, “In 1867, up to eight hundred thousand Algerians starved in a famine….One French witness described the sight: ‘On the roads, in the ditches, we encountered lifeless unfortunates already cadavers. Others breathed still…occasionally allowing a breath to escape, a groan, they shook with convulsive jolts in the cruel trance of affliction’.” (Slobodkin, “Empire of Hunger,” 1) Some areas, for example Oran in the west, lost half their population to the famine. Again, it is important to remember that the famine came on the heels of military conquest.

As to the rest of the French empire in Africa, there was the “Great Famine” of 1913 in the Sahel; anywhere between 400-800,000 people perished. Here is another quote from page 153 of Slobodkin’s study:

 

Censuses taken throughout the Ouaddaï showed that most districts had lost at least 50% of their population, and certain areas, including the city of Abéché, suffered a reduction of as much as 86%. Well over half the villages in certain districts ceased to exist entirely. One example among many was the village of Koutoul, which in 1913 counted 54 men, 96 women, and 111 children, but in May 1914 was reduced to 4 men, 5 women, and 6 children, and “in such a condition that they will probably not make it to the next harvest.”52 The devastation among cattle was enormous, and certain districts lost 95% to 100% of their herds.53 According to an estimate by Hilaire, the population of the Ouaddaï dropped from 700 thousand to 400 thousand. The administration of Chad, while acknowledging the devastation, seemed perfectly pleased with its own response. Largeau wrote, “One asks oneself what would have become of the Wadai if a cataclysm of this magnitude had occurred in the times when the improvident and badly organized Sultans surrounded themselves with thousands of idle people.”

 

So, between Algeria and the Sahel, we are talking about another 1.2-1.6 million deaths from famine. But note that we haven’t even discussed deaths from colonial conquest.  As for the Sahel, another catastrophe came in 1938.

Our running total is now somewhere between 36-60 million people. Again, we have not even begun counting other unnecessary deaths, as I note above.

As for the French empire, there are multiple place we have yet to consider, including areas where there was rubber extraction.

The British were just as bad. In The Killing Age I discussed in some detail areas like Bunyoro, in Uganda, and a few other areas in East Africa. Let me add a few additional words here on the British empire in Africa. Wherever you see white colonial settlement you also see very high levels of violence, if only because settlers were confiscating land from indigenous peoples. In colonies like Nigeria, there were not settlers; in South Africa, Rhodesia and elsewhere there were. The association of colonial settlement and violence has been explored by legions of scholars. At the same time, national and imperial mythmaking has invariably downplayed violence.

I will return to the British in considerable detail in two weeks. In the meantime, a few points seem apposite:

1. My estimates tended towards the conservative and never attempted to be comprehensive.
2. The period c. 1750-c. 1900 was nasty! It was also exceedingly messy.
3. It is difficult to get one’s head around the destruction. (Remember, as well, that this period was all before the wonders of modern medicine.) We want to say “No way, it couldn’t possibly have been that bad.”
4. The long nineteenth century makes the short twentieth century all the more fascinating and perplexing. Generally, non-state violence declined but state violence increased….tremendously. One important story of the twentieth century is the massive increase in the ability of modern states to destroy life. At the same time, there was real human progress—across large areas of the world people lived safer, better lives— punctuated by utter catastrophe (1914-18, 1939-45). For example, despite two world wars life expectancy increased in parts of Europe. Even in Africa, life expectancy slowly crept upwards. Other metrics also indicate slow upwards growth

 

Again, I am trying to be conservative. But I hope one thing is becoming clear. At the same time, we have to remember that the capacity to destroy has increased remarkably; this is the story of the twentieth century, which I have begun to address elsewhere.

 

More later. At this early juncture, readers might be interested in taking a peek at another attempt to catalog (mostly) wars in the nineteenth century: https://necrometrics.com/wars19c.htm. I haven’t had much of a chance to look at this very closely, and the site is at least 14 years old so out of date. The author concludes:

“EXTREMELY PRELIMINARY TOTAL:

• Adding all the events listed here gives a very tentative total of 45 Million unnatural deaths during the 19th Century — and I probably missed a lot.
• You might want to add an additional 45M, depending on whether you feel that the El Niño famines of 1876 and 1896 were man-made or not. That would boost the total to 80M unnatural deaths for the 19th Century.
• Using Table 1.2 of A Concise History of World Population, 2d by Massimo Livi-Bacci, I determined that there were 8661 million deaths between 1750 and 1950. As the 19th Century covers the middle half of that, let’s assume roughly 4330 million deaths during the 1800s.
• Another source would be “How Many People Have Ever Lived on Earth?” by Carl Haub [http://www.prb.org/Content/ContentGroups/PTarticle/0ct-Dec02/How_Many_People_Have_Ever_Lived_on_Earth_.htm]
  1. We can calculate the number of deaths between 1850 and 1900 as [1850_population] + [Births_1850-1900] – [1900_population] = [Deaths_1850-1900] = ca. 2509 million.
  2. We can calculate the number of deaths between 1750 and 1850 as [1750_population] + [Births_1750-1850] – [1850_population] = [Deaths_1750-1850] = ca. 3576 million.
  3. Adding [1] and half of [2] would give us some 4297 million deaths in the 19th Century.
• Forty-five million unnatural deaths would be 1% of 4.3 billion deaths (or 1 out of every 96), considerably less than the percentage for the 20th Century. Counting the famines would bring the percentage to 2% or 1 out of 48.”

 

 

He also did the same for the eighteenth century (https://necrometrics.com/wars18c.htm), concluding:

“EXTREMELY PRELIMINARY TOTAL:

• Adding all the events listed here gives a very, very (very) tentative total of 18 Million unnatural deaths during the 18th Century.
• Using “How Many People Have Ever Lived on Earth?” by Carl Haub [http://www.prb.org/Content/ContentGroups/PTarticle/0ct-Dec02/How_Many_People_Have_Ever_Lived_on_Earth_.htm]
  1. We can calculate the number of deaths between 1750 and 1850 as [1750_population] + [Births_1750-1850] – [1850_population] = [Deaths_1750-1850] = ca. 3576 million.
  2. We can calculate the number of deaths between 1650 and 1750 as [1650_population] + [Births_1650-1750] – [1750_population] = [Deaths_1650-1750] = ca. 2877 million.
  3. Adding half of [1] and half of [2] would give us some 3226 million deaths in the 19th Century.
• Eighteen million unnatural deaths would be 0.6% of 3.2 billion deaths (or 1 out of every 179). This is a lot lower than either the 19th or 20th Centuries death rates, but we shouldn’t take this as proof that the 18th Century was more peaceful. With a childhood mortality rate far higher than later eras, fewer people would live long enough to get caught up in wars and tyranny.”

 

 

Readers might also want to check out: https://correlatesofwar.org/data-sets/cow-war/.

 

See also: https://ourworldindata.org/conflict-data-how-do-researchers-measure-armed-conflicts-and-their-deaths, and https://ourworldindata.org/search?q=deaths+in+war&resultType=all

This data is also highly problematic and should be used with caution, as with all the others. A good deal of the ourworldindata material goes to the research of Peter Brecke at Georgia Tech. See https://brecke.inta.gatech.edu/research/conflict/

You will see in all these cases just how incomplete things remain. We are working very much in a world of guestimates. In the USA especially, research funding has dried up considerably; works in progress have stalled. We have to take these numbers with a grain of salt, including my own.