Rise & Fall of War

There seems to be lots of confusion on the history of war, so let me try to clarify. Most confusion comes from seeking a one-way trend, as in “is there more or less war than in ancient times?” Problem is: overall, warfare increased, then decreased.

Since WWII, at least, we’ve seen a dramatic decrease in casualties from wars between states, from civil wars, and from crime. While these are lowest in richer nations, the strongest correlate seems to be the fraction of men aged 15-30. The fewer young men, the less war/crime. Rich societies today likely have the lowest war/murder rates ever. Rich industry does seem to have greatly discouraged war.

Yes, most of the “tribal” societies that anthropologists study have high rates of war.  But most of these are intermediate forms between very distant ancestors and very modern societies, with many relatively modern features. So high rates of war in such tribes does not imply that our very distant ancestors had such high rates.

The rise in density before, during, and after farming seems to have been associated with a huge increase in war. Long ago, strong social norms limited violence within nomadic forager bands, and the fact that one gender typically moved to neighboring bands to find mates greatly discouraged attacking such bands.  War was hard for foragers, as hostile victims were far away, at unpredictable locations, and with few physical goods worth taking; women taken in war could easily escape.  Trading places, with predictable locations and trade worth taxing, made the first good war targets. Increasing density made targets easier to reach and find, and marriage as property made wars to grab women more tempting.  Herding helped attacking armies to travel further and faster, while farming created more tempting and harder-to-defend targets to attack.

War is hell, not an especially modern hell, but also not an especially ancient hell.  War is most distinctly, a farmer’s hell.

GD Star Rating
Tagged as: , ,
Trackback URL:
  • “Since WWII, at least, we’ve seen a dramatic decrease in casualties from wars between states, from civil wars, and from crime. ”

    For homicide, that secular decline goes back to around 1500 in Western Europe (varies by country; in Italy only back to the 1800s). See Manuel Eisner’s reviews of long-term homicide rate data (the one in the Brit J of Criminol has time series pictures), or their summary in a talk on historical violence by Steven Pinker for the TED talks.

  • lemmy caution

    Chimpanzees have a high level of violence:


    It isn’t clear if ancient humans had high levels of violence as well. Maybe not, for the reasons Hawks explains.

  • I’m worried about a small sample size. How far are we really from disintegration of world trade, or disintegration of a national government? As you pointed out in your talk casualties tend to follow a power-law distribution.

  • Sean

    Victor Davis Hanson actually writes quite well about the changing forms of warfare, though many disagree with his politics.

    The interplay between the industrial nature (getting many poor farmers working together), the knowledge and willingness to commit violence (steppe people with a lifetime of butchering), and the ownership of the fight (Greek farmers fighting as full citizens) all have an interesting interplay.

  • This matches my understanding on the progression. Considering how the most devastating conflict in history happened less than a century ago, it’s awfully premature to proclaim peace and prosperity.

    • Doug S.

      I’m not so sure that World War II was really the “most devastating conflict in history”. Sure, it was the war with the biggest death toll, but much of that is because there were a lot more people alive who could fight and die. Compare it to the An Shi Rebellion

      Total world population in 1940: 2.3 billion
      World War II death toll, upper estimate: 72 million
      Percent of world population that died: 3.1%

      Total world population in 756: 220 million
      An Shi Rebellion death toll, upper estimate: 36 million
      Percent of world population that died: 16%

      So yeah.

      • Alex Flint

        It’s not obvious that percentages matter here more than the raw number of human lives lost.

      • I think “devastating” implies percentages matter more, at least more than the specific alternative of raw number of human lives lost.

      • Great point.

      • From that article:

        “However, these figures should be viewed skeptically as the decline of 36 million was in registered population, due to the breakdown of the census system.”

      • Roko

        Alex: percentages matter, at least from the point of view of an individual. If the number, N, of casualties tends to infinity but the fraction f tends to zero, then eventually the probability of you dying in a war becomes less than the probability of you dying of hiccups.

  • Pingback: Tweets that mention Overcoming Bias : Rise & Fall of War -- Topsy.com()

  • Microbiologist

    > hostile victims were far away, at unpredictable locations

    Sedentism may be much older than agriculture.

    Hostiles may not usually have been very far away, either. Bear in mind that the foragers we have studied are a non-representative sample, consisting mostly of peoples that inhabit lands unsuitable for agriculture/husbandry under pre-modern conditions. Often these are extremely sparse ecologies, arctic or desert.

    > women taken in war could easily escape

    But do they? And do they do so frequently enough to disperse the incentive? Why not be more inductive; your deductiveness on this topic seems puzzling. We know that Yanomamo capture women and that their braves told Chagnon this was the biggest motor of their belligerence.

    > Long ago, strong social norms limited violence within nomadic forager bands

    “Limited” it as compared to what? “[T]he Kung San of the Kalahari are viewed as a very peaceful society; indeed, one popular ethnography on them was titled /The Harmless People/. However, their homicide rate from 1920-55 was 4x that of the US and 20-80x that of major industrial nations during the 50s and 60s. Before local est. of the Bechu-[something]/Botswana police, the Kung also conducted small-scale raids and prolonged feuds between bands and against Tswana herders intruding from the east. The Copper Eskimo, who appear as a peaceful society in the cross-cultural surveys just discussed, also experienced a high level of feuding and homicide before the RCMP police suppressed it.”

    And so on; see here

  • I would draw a distinction between war and battle. The Battle of Gettysburg, say, is unthinkable among hunter-gatherers. They’d run away. They’re not crazy.

    On the other hand, ambushes are perfectly natural for hunter-gatherers. After all, they’re hunters, and an ambush is just hunting people you don’t like. In a Malthusian world, there are lots of reasons for disliking people. So, endless chronic wars of ambushes (kind of like gang wars) are common among hunter-gatherers. The Battle of Thermopylae, however, took farmers or herdsmen.

  • agnostic, yes murder rates have long declined

    lemmy, yes chimps have high murder rates

    kevin and summer, yes the future is hardly guaranteed

    micro, yes stable locations precede farming

    Steve, yes foragers can ambush, if they know where targets will be when, and that isn’t too far away.

  • It is true that primitive agriculturalists are much more warlike than hunter-gatherers. But like the commenters above, all the sources I’ve read report that hunter-gatherers (distinguished from primitive agricultural “tribals”) die from homicide at a much higher rate than we do today.

  • John Thacker

    War is most distinctly, a farmer’s hell.

    War is hell on farming. But isn’t it more common in herding societies (not the same as hunter-gatherers or foragers) than farming societies?

    After all, the problem with making war on farmers is that you still have to work the field, and all that war is likely to destroy crops. Sure, you can come by at harvest time, but it still seems most efficient to leave the farmers alive but establish suzerainty or otherwise get the crops via taxation.

    On the other hand, in a herding society, one’s entire possessions are always at risk to a cattle raid or some such. At literally any time in the year, you can steal away all that someone has, in a way that you can’t seize and take away land.

    That’s why Nisbett and Cohen say they found a link between traditional herding societies and violence.

    Perhaps you weren’t distinguishing farming from herding?

  • Right. Hollywood Western movies frequently illustrated the differences between herdsman (cowboys) and farmers. The cowboys lived in a Hobbesian world, but by the end of the movie, the good sheriff had made the place safe for farmers. He rides off into the Western sunset to find a place still Hobbesian enough to need him.

  • After all, the problem with making war on farmers is that you still have to work the field

    Right again. England had lots of war among the highest ranks through the 1640s, but it had remarkably little effect on the lower ranks, other than whom they paid rents to and whom they looked up to.

  • Drewfus

    The trend toward non-violence is even more remarkable given the enormously increased capacity for desctruction – at least at the level of the nation-state.

  • The most intense war scenario was herders attacking farmers. Herders were hard to attack because they could just retreat with their herds, and could concentrate their forces. Farmers, in contrast, could not move or concentrate and so were easy targets. Farmers formed the largest scale political organizations to defend against herder attacks.

  • “Yes, most of the “tribal” societies that anthropologists study have high rates of war. But most of these are intermediate forms between very distant ancestors and very modern societies, with many relatively modern features. So high rates of war in such tribes does not imply that our very distant ancestors had such high rates.”

    Most seem to think it does. We still have uncontacted tribes out there – and so the level of “modernisation” of tribes goes down pretty low. We can still see their body paint and weaponry from the air.

  • Neither body paint nor bows necessarily imply a high rate of warfare. Bows are not specialized weapons for killing people.

  • Two drunken Indian fishermen washed up on North Sentinel Island a few years ago, which is inhabited by an Andamanese tribe that has resisted all contact. They were immediately killed by the natives.

  • Chimpanzees murder for land

    Between 1998 and 2009, John Mitani witnessed 18 murders firsthand, and found circumstantial evidence for three more. But no police were ever called, for these killers were all chimpanzees, from the Ngogo community in Uganda’s Kibale National Park.

    Groups of chimps, mostly male, will mount lengthy aggressive campaigns against individuals from other groups, attacking them en masse and beating them to death. After the chimps picked off their neighbours, they eventually took over their territory. It seems that chimps kill for land.

    The vast majority of these murders were carried out by groups of Ngogo males on patrol. These patrols are stern, single-file affairs. Males march along the borders of their territories, scanning for other chimps and neither feeding nor socialising. They monitor the northeastern part of their territory with particular fervour and indeed, 13 of their 21 kills took place here.

    Of these victims, 4 were adult males and 9 were youngsters. That may seem like a small number, but for chimps, these are severe losses. They were even higher (by around 5 to 17 times) than the death rates due to violence between groups of human hunter-gatherers.

    And because of their aggressive tactics, they have increased the size of their territory by some 22%, expanding into the northeast area that their neighbours once called home. With murder came new real estate to colonise.

  • Pingback: Overcoming Bias : On Sex & Violence()