Frequency of conflict initiation worldwide

That is from the new and interesting Only the Dead: The Persistence of War in the Modern Age, by Bear F. Braumoeller, which is largely a critique of Pinker on trends toward peacefulness (Pinker gives only the more optimistic data on Europe).  And from the text:

…there is variation in the rate of conflict and war initiation over time, and it’s pretty substantial.  Leaving aside the two jumps during the World Wars, the median rate of conflict initiation quadruples in the period between 1815 and the end of the Cold War, after which it abruptly drops by more than half.

The “falling rate of conflict” is thus not entirely reassuring.

How about the deadliness of occurring conflicts?:

Analyzing the two most commonly used measures of the deadliness of war, I find no significant change in war’s lethality.  If anything, the data indicate a very modest increase in lethality, but that increase could very easily be due to chance…Worse still, the data are consistent with a process by which only random chance prevents small wars from escalating into very, very big ones.

Overall, the arguments in this book are strong, and the discussion of data issues is subtle throughout.  You can buy the book here, its arguments seem fundamentally correct to me.


Does evolution have a quota that compels us toward the equilibrium?

'(Pinker gives only the more optimistic data on Europe)'

Of course - most people who aren't European do not understand just how brutally bloody Europe has been historically. Particularly if we now think of the break up of Yugoslavia as history, too.

Oh I think most people have heard of Germany

And Napoleon...

And they've at least heard of the 30 Years War.

You don't have to be a very good historian to have a strong impression that Europe was more peaceful between the Congress of Vienna in 1815 and the outbreak of World War I than it was before (or, for a while, after). It's the Middle Ages and the early modern period that were unbelievably bloody, as was much of the rest of the world during that time.

We do, however, know more about conflict initiation after 1800 so it's understandable the retrospective numbers go up around then.

'that Europe was more peaceful between the Congress of Vienna in 1815 and the outbreak of World War I'

Well, if you think of war as being between major empires.

This list highlights how peaceful Europe was -

The Greek War of Independence was not a trivial war in cultural terms, and what happened in 1848-49 can also be considered war (it certainly is in this region, where the last major battle was fought in Germany, involving thousands of troops). Then there is the entire creation of the modern Italian state - easy to overlook those conflicts, isn't it?

Europe is not peaceful, and never really has been. Of course, one does need to draw a line between war and those events where military force is used to kill people - the Troubles simply did not reach a large number of dead in military incidents, either in individual events or in total, to be considered war.

None of the conflicts you mention were particularly deadly - certainly not compared to e.g. the Taiping Rebellion in China, where your favorite source states that "Estimates of the war dead range from 20–70 million to as high as 100 million"

I'd also venture that no European conflict between 1815 and 1914 were nearly as gruesome as the Mfecane (or Difaqane, in Sesotho) in Southern Africa. It was set in motion by Shaka's conquests, and some of the tribes fleeing from his armies (most notably, Mzilikazi's Matabele) in turn killed everyone they encountered. According to your favorite source, "the death toll has never been satisfactorily determined, but the whole region became nearly depopulated. Normal estimates for the death toll range from 1 million to 2 million". . Eventually, Mzilikazi's reign of terror on the highveld was stopped by those evil white Boere. He and his people then invaded Zimbabwe.

Taiping rebellion was led by a messianic Christian leader. People forget that before Marxism murdered 100 million people in the 20th century, Christianity also slayed 100 million people the century before. In terms of genocidal ideologies, radical Islam of today is small potatoes.

Nope. Even if you include the concurrent revolts of the Panthay, Dungan, and Nian rebellions you still do not break 50 million.

Of those dead, the greatest numbers were from the Qing's habits of scorched earth warfare and river control demise. Or and the other main source of civilian deaths were the Qing reprisal genocide against the Haka. To whit, the vast majority of the dead in the Taiping conflict came from the Qing side. Which should not surprise us, they were the ones who controlled the territory, burned more crops, and engaged in genocidal retaliatory massacres after the military conflict was over.

And of course calling the Taiping "messianic Christians" is historically illiterate. The Taiping were less Christian than the Mormons of this era, what with opting for the whole secret society thing, Jesus's brother, and other fun things.

But this is par for the course. sadly. Why learn history when you can regurgitate talking points?

"Estimates of the war dead range from 20–70 million to as high as 100 million"

Propagandists, anti-Marxists and anti-Christians, tend to pick the high water figures to make their opponents look bad but such is the nature of the game. The 19th century had less than 2 billion people in total across the globe and China's population in particular was less than 400 million. Even taking the lowest figures, that is still devastating and very much a genocide.

The death tolls from the Taiping rebellion like 20th century Maoism mostly came from famines. Mao was actually a proponent of the Taiping rebellion citing it as an early attempt by the people to revolt against the oppression by a feudal state and also because the Taiping Heavenly Kingdom abolished private property, redistributed land, declared equality of the sexes, abolished social class, and prohibited vices like slavery, prostitution, and opium which were all in line with Maoist thought.

The problem with this is that one of the worst famines preceeded the rebellion. Taking the Famine deaths of 1849 into account makes some sense for measuring the devastation of China. It, however, is nonsensical for attributing blame to a rebellion which began in 1850. Likewise, attributing all the famine deaths to Taipings is pretty silly, huge numbers of deaths occurred in provinces that the Taipings never captured.

The biggest problem I have is that the Taiping Rebellion was an active war with active contestation of control. Mao had a unified country with full control. It is roughly equivalent comparing Jeffry Dahlmar to George Custer.

The Taipings represented a calamity in Chinese history. Yet saying that Christianity, a religion where all the missionaries refused to baptize Hong Xiaquan, "slayed [sic] 100 million" is bogus.

Not sure why you object to calling it messianic Christianity when it was called exactly that by historical sources. From University of Cambridge:

"The 'Taiping Heavenly Kingdom' (Taiping Tianguo) (1850-1864) was an armed insurrection based on a messianic cult of Christian inspiration"

Well firstly because the Taiping's objected to it, viewing the Christian churches extant at the time as heretics themselves. Secondly, the Christian Powers (self-identified) at the time opposed the Taipings and backed the Qings. And lastly Taiping doctrine had exceedingly little in common with Christian doctrine; hence why your source quite correctly calls it "of Christin inspiration".

To whit, Taiping religion was inspired by Christianity and was not connected to extant Christian organization, doctrine, or practice. This much how Liberian government is of American Inspiration, but is not actually, you know, American.

Samuel: Any good suggested reading on this?

Autumn in the Heavenly Kingdom

The period between the War of the Spanish Succession and Great Northern War and the French Revolutionary Wars was also an era of limited intensity wars.

It looks like the graph is a raw number of conflicts, not adjusted for population. But the world population has increased at least by a factor of 20 over the span of this graph. Doesn't this matter?

Also, how well can worldwide conflicts be measured from 300+ years ago?

Not very convinced by this piece of evidence.

Another issue (in addition to having information about conflict and adjusting it for population) is how to define conflict initiation. So the graph captures conflict initiation. Does the 100 Years' War count as one conflict or many? Is a conflict a particular battle in a war? If we go back to the Punic Wars, how would we measure those? Do we have good enough info on them to actually distinguish between the First and Second Punic Wars, or should we just consider them the same conflict?

This seems more of an exercise in "how can I define my variables to get the answer I want?" To be fair, I think Pinker is sometimes engaged in the same thing.

Agreed. I found this conflict data from Max Roser with rates of death adjusted for the population much more informative:

Convincing? I'll have to read the book, but the arguments quoted I find not at all convincing.

What's the point of counting something as varied in size as "conflicts"? How is a conflict defined anyway? Does WWII count as one conflict or as many?

What does "lethality" means in the second quote? Number of deaths per conflict, in which case it makes not much sense since it depends on how you count conflicts, or total number of deaths due to war? If it is the second, that seems a better measure, but can we trust the figures we have before 1800 outside of Europe and some regions in Asia? Do we know much about the war inside Africa before 1800? Do we have even a reliable Fermi estimate of the number of deaths in internal war in pre-columbian America? Hey, historians don't even agree on how many people there were in America in 1492.

For me this was the most obvious flaw too, especially given the increase in the size of the average political unit over this time. For example, it is hard to imagine hundreds of seperate conflicts happening at once in North America now, but if there were to be single conflict, it could lead to just as many deaths.

Seconding the argument above. An inter-tribal conflict in Africa is very unlikely to be counted in this data, but a civil or interstate war in Africa today certainly would.

"OurWorldInData" have many plots supposedly showing global declines in violent conflict. Unless there's a clearer argument why that data is wrong and this is right, it's not clear who we should believe.

That was also my first thought!

Interstate wars of conquest in Africa are not hugely common. Has any African country enlarged its borders in the last 50 years? Perhaps Morocco absorbing Spanish Sahara? There have been a few attempts to take some desert away from Ethiopia by Somalia and Eritrea but I don't recall either succeeding. In general, most African states have their hands full exerting control over just the territory left to them by the colonialists and don't have much appetite for conquest of more territory.

"Interstate wars of conquest in [continent X] are not hugely common"

Most wars in every part of the world are civil wars

Yeah, I have my questions about what constitutes as conflict and how it is measured. I look at this graph and 1700-1800 is unusually conflict free. Yet I remember that the Great Northern War, one of the most devastating wars for Eastern and Northern Europe was just at this time. I am no historian, but I remember reading that approximately 20% of Russian Tsardom's population died back then. That is higher percent than in WWII. I would think that losses for other actors in this war were also incredibly high.

Doesn't count. Basically, Americans aren't taught much about Europe in general (though more than anywhere else), and tend not to be aware that each European nation has its own history of violent conflict.

If you look at the wikipedia link above, about the only war Americans would likely be aware of involving Russia would be the Crimean War. The expansion of the Russian Empire in the Caucasus over two generations? The fairly regular conflicts involving Turkey? Nope, those are not part of how peaceful the 19th Century was in Europe.

During the periods 1700-1800, at least the Americans are aware (or should be) of the seven years' war (aka French-Indian war), one of the most important war of modern times in term of geopolitical consequence.
It was quite deadly too, with more than one million deaths.

I hate to say this but most Americans have no idea what that is let alone why it was important.

The French and Indian war is part of basic American history teaching. Every kid learns about it. Of course, people forget lots of basic stuff. However, I would be surprised if "most Americans" weren't even aware of it.

The geopolitical significance and global scale of the war are largely unknown to Americans, I would think. What is taught focuses primarily on the events in North America.

Here's the French and Indian War as it's taught to Americans (or at least what is retained).

There was a war, mostly in the north, by Canada. America wasn't a country yet, so who cares.

The French and Indian War had more than 1,000,000 deaths? Source? The usual number is 1,512 killed in action on the British side, somewhat more on the losing side. Maybe you are thinking of the Indian wars more broadly, over decades and extending into the 19th century.

Wikipedia has it at several hundred thousand, though significantly under a million. Maybe those figures you mention are for North America only? It was a global war that spanned Europe, India, Africa, and South America, as well.

Conflict is the antithesis of cooperation. A rational person would resort to conflict if conflict is the most efficient means to an end, such as the pursuit of wealth. A rational person would resort to cooperation if cooperation is the most efficient means to an end, such as the pursuit of wealth. It's not surprising that conflict spiked during the period when conquest was considered the most efficient means of obtaining natural resources, and that conflict receded when cooperation was considered the most efficient means of obtaining the end product of which natural resources are only a part. One might ask: what's the trend today, conflict or cooperation? Those with a zero sum view might conclude conflict, while those with a non-zero sum view might conclude cooperation. Which is the prevailing view today? I don't know. Trump has expressed a zero sum view: in trade, there are winners (exporters) and losers (importers). But he has also expressed a non-zero sum view: the terms of trade can be improved with his extraordinary negotiating skills. Everybody is a winner! My view: shared prosperity promotes cooperation (the rising tide), while rising inequality promotes conflict (economic and political). Across countries trade has produced shared prosperity, while within countries trade has produced inequality. Will the former increase cooperation (reduce conflict), or will the latter increase conflict (reduce cooperation)? It depends. On what? On who chooses the path forward.

There's a theory out there called offense-defense balance, which posits that changes in the amount of war are due to changes in technology over time. Basically, that at some points in history technology gives the defender an advantage and at some points in history technology gives the attacker an advantage, and there are more wars when the attacker has an advantage.

Can you show me that graph on a per capita basis please?

The world population was 1 billion in 1804, 2 billion in 1927, 3 billion in 1960, etc.

I wonder what a map of violent conflict in the pre-Columbian Americas would look like?

Something smells fishy with that graph.

Counting 'conflicts' seems a bit too squishy. How about homicide deaths per capita? That alone would probably reverse the graph and is more relevant.

There's some other things that smell fishy too. The y axis starts at 20 instead of 0. That's a major red flag. The time frame of the graph also is picked so that WW2 and the cold war are on the right but the left is post-Mongols.

Seriously: and would Timur's campaigns count as a single conflict or as multiple/sequential conflicts? (with or without prior Mongol efforts)

The death toll from Timur's cumulative efforts is said to be quite high by informed historians.

This graph corroborates Oswald Spengler's contention that the pinnacle of Western Culture was marked by the death of JS Bach in 1750.

Armies used to be a lot smaller and the territory affected by a military campaign was also a lot smaller, so in absolute numbers wars were less deadly.

I've looked over the Google preview of the book, and the author answers a lot of these data questions (adjusting for population size, number of countries, etc.) in the middle chapters. This graph is toward the beginning and is just meant as a response to one of Pinker's graphs. Basically, Pinker reprinted the very optimistic Europe graph while ignoring the much more pessimistic world graph. Braumoeller writes, "my own take is that neither of these graphs really tells us much of anything. There’s too much going on—changes in population size, erratic and unpredictable changes in the odds that a violent incident will be recorded for posterity, and a transition from a feudal order prior to the Thirty Years’ War to a system of sovereign states afterward, just to name a few—for us to be able to glean much of anything about trends in conflict from either of these graphs. But if Pinker is right and Figure 2.1 really does contain useful information, Figure 2.2 should have been entered into evidence as well."

+1, helpful comment.

+1, informative

I read this book the day it came out. It is one of the best IR books I've ever read; it's ambition, careful, and I agree with Tyler that it seems fundamentally correct. An interesting part of it was how wars follow a power law distrubition.

Now normalize by world population.

Comments for this post are closed