Apologies for running a fireside so quickly after the gap week, but with the end of the semester coming as the job market gets busier, I haven’t had time to finish the next post on shield walls just yet....
Apologies for running a fireside so quickly after the gap week, but with the end of the semester coming as the job market gets busier, I haven’t had time to finish the next post on shield walls just yet. That will come out next week for sure though, as it is close to done.
I realize Ollie’s enormous eyes make it seem like he is frightened or concerned here, but he is in fact eyeing a toy and using his scratching post as ambush cover. That feather-toy will never seem him coming.
For today’s musing, I wanted to actually give something of a longer-form answer to a question a student asked me recently. To paraphrase, the former student noted the terrible civilian toll of modern warfare on civilian populations and asked, in essence if it is always this way, why such results are so common in American military history and then does any of that mean it must it always be this way? Is there no other way where civilians do not suffer terribly? And I think that is a grim but necessary question as people are trying to make sense of conflicts occurring not in open fields as the battles of old but in dense urban environments where there are lots of civilians.
And because this is bound to be a heated topic I want to remind y’all in the comments that you will be civil. There are many places to have screaming matches about today’s active conflicts, but what I would like to see instead of a sober discussion of what is and is not possible.
I think we have to begin by specifying that the norm that civilians are not supposed to be the subject of violence in war, that war is a matter between combatants only, is a relatively new one. For most of human history it was broadly assumed that ‘enemy civilians’ were fair targets for mass violence or enslavement. Indeed, for the oldest ‘first system‘ warfare, civilians were often the primary target as the goal was to force enemy groups to migrate away. We see the earliest sort of tentative steps to the idea that some ‘enemy civilians’ ought to be exempt from violence coming out of medieval Christianity and Islam, but these efforts had limited effects. Instead, as we’ve discussed standard military practice, particularly foraging, in this period turned armies into what were effectively roving catastrophes that unavoidably burned and pillaged when they moved through enemy territory.
In the early modern period in Europe, a sort of standard code emerged on how to treat ‘enemy’ urban areas. During a siege, at key points the city would be given a chance to surrender; if they did so, treatment was expected to be lenient and the commander of the besieging army was expected to keep his army in check. If they did not surrender and the besieging army was forced to take the city by storm, however, the norm was pillage, rape, and massacre, with the normal expectation being that the attacking general would only even attempt to get control of his troops after about three days. This ‘code’ obviously falls well short of the modern Laws of Armed Conflict (LOAC) and it was mostly a question of expediency: you wanted a fortified city to surrender quickly rather than force you to launch a bloody assault in which the initial waves of attackers could be sure to take extremely heavy losses (the Dutch term for that first wave was a verloren hoop, ‘lost heap’ (which becomes ‘forlorn hope’ in English) which gives a sense of the losses you are trying to avoid).
We really only start to see stronger norms protecting civilian populations emerge in the 1800s, but that too is a slow process of half-measures. In both World Wars, all sides engaged in the bombing of urban areas from the air. In the First, the Germans used Zeppelins and the Allies were developing bombers for the purpose and using them at war’s end. In the Second World War, German terror-bombing in Poland and Japanese terror-bombing in China came first, but were of course followed by far more extensive allied bombing as the weight of industrial might made itself known. There was some sense that this was wrong, but no one was going to desist over that. Curtis LeMay, the commander of US strategic bombing operations over Japan, remarked that if the United States had lost, he’d have been tried as a war criminal, though it is worth noting that of all of the German and Japanese leaders prosecuted for war crimes, none were charged with bombing civilian areas, despite there being many who could have been.
The modern infrastructure of the Laws of Armed Conflict to protect non-combatants is a product of the horrors of the Second World War, codified in the Fourth Geneva Convention in 1949. And even then, the protections are limited: belligerents are not to either intentionally target civilians nor to attack in ways that are indiscriminate. But civilian casualties are permitted due to ‘military necessity’ with the principle being that the damage done ought to be in proportion to the military objective.1 If that seems mealy-mouthed and half-hearted that’s because it is. But the challenge with laws to limit was is that if you make them too limiting, no one will follow them because it would lead surely to defeat and in that case norms would revert back to the slaughter and brutality of the olden days.
Still, as weak as these norms are, it should mean in theory that at least some armies are striving to limit civilian suffering and death as a result of war as much as reasonably practicable (which is not the same as ‘as much as possible.’)
That in turn collides with the reality of battle in urban spaces, which has changed meaningfully since the 1800s. Before the industrial revolution, only a small part of the population lived in dense cities, typically around 10-20% or so. Today more than half of all people globally live in cities and of course the world population is several times larger. Even as late as 1800, there were no cities on Earth very much larger than a million people; today there are more than eighty cities with five million people in them. That means the locus of population, production and infrastructure is in these cities, making them key military targets, while at the same time they become massive concrete warrens – extremely difficult terrain to have an army in.
Looking at battles in urban spaces over the past century or so – that is, in the age of Industrial Firepower – it seems to me we have to hold two difficult ideas in our heads at once. On the one hand, it seems quite clear that there is no way to do urban operations ‘clean.’ Even very scrupulous efforts in urban spaces to follow the LOAC end up with massive infrastructure damage to cities and significant civilian casualties in our age of industrial firepower. On the other hand there is a world of difference between the armies that try very hard to avoid civilian losses and those which do not, and a terrible spectrum between them. This is a case where the difference between ‘bad,’ ‘worse,’ and ‘worst’ is actually a very big and meaningful difference, even though ‘bad’ is still quite bad.
One way to think about this – simplistic, to be sure, but it will do for now – is a very simple metric I’m borrowing from Wesley Morgan: we can consider a ratio between civilian and military deaths in an urban operation. In taking Mosul (2016-2017), the anti-ISIS coalition killed an estimated 10,000 or so ISIS militants at the cost of around 8,000 civilians (numbers very approximate, of course), a ratio of about 1.35:1. In the Second Battle of Fallujah, the United States (and partners) figures it killed around 2,100 enemy militants and the Red Cross estimated 800 civilian deaths, a ratio of 2.6:1. On the other end, in the Battle of Berlin (1945), the Soviet Red Army killed an estimated 92,000 German soldiers and 125,000 civilians, a ratio of 0.736:1. In Mariupol (2022) the Russian Armed Forces claim to have killed 4,200 Ukrainian soldiers2 and while Russia claims there were only 3,000 civilian casualties, that number fairly clearly a lie as the AP identified more than ten thousand mass graves and estimates the true civilian toll around 25,000, a ratio of 0.168:1. At the absolute rock-bottom of this well of misery, the Imperial Japanese Army took Nanjing (1937), killing around 10,000 Chinese soldiers before butchering more than 200,000 Chinese civilians in a horrific massacre that even today some Japanese politicians like to pretend did not happen, a ratio of 0.05:1. We don’t know the military deaths from the Siege of Aleppo (2012-2016), but given that there were never estimated to be more than around 15,000 fighters in the city (many of whom will have survived) and an estimated 31,000 civilians died, it seems safe to assume the ratio there was also horrible, probably on the scale of the Mariupol ratio above.
The IDF claims that they are killing one Hamas soldier for every two civilians they kill in Gaza, a ratio of 0.5, which their spokespeople have claimed is ‘unprecedented in the modern history of warfare,’ but which looking at the figures above is not actually a particularly scrupulous or discriminating ratio – though of course one may well argue the vast differences in circumstances. It is still both far from the best and far from the worst performance for armies operating in civilian spaces.
At the same time, ‘least bad’ is not ‘good.’ The Battle of Fallujah, which in those examples has the highest rate of target-discrimination still destroyed much of the city, as you can see in images like this one. But most of the people in that city (population c. 250,000) survived and that makes a world of difference compared to some of the abject horror on that list above.
.jpg){kind=link}
Where does the United States fit into all of this? While there is ample ground to criticize the US military’s policies towards civilians since 1900, on the whole the United States tends, compared to other major states involved in the same or similar wars, to be on the low end of causing civilian death and suffering, though with some pretty important ‘black marks’ on that record.3 Americans focus – rightly so – on the morality of strategic bombing in Europe and Japan, but the fact is the vast majority of civilians killed in Europe and Asia in WWII were not killed by bombs from the sky but by soldiers on the ground with guns intentionally murdering civilians. Strategic bombing over Germany killed an estimated 400,000 civilians, a terrible figure – but the German government today estimates that Soviet expulsions alone may have killed as many as two million civilians. Both the Nazis and the Red Army engaged in mass-atrocities on the Eastern Front at staggering scale (an estimated twenty million Soviet civilians were killed during the war). American strategic bombing over Japan killed approximately 400,000 civilians as well, but in contrast to an estimate that Japan caused between six and twenty million civilian deaths due to atrocities in the territory it controlled. I am no fan of strategic bombing, but while it did kill hundreds of thousands of civilians, it was not the primary driver of civilian casualties in WWII; intentional mass-murder on the ground was.
That is not to say American soldiers have never engaged in massacres. They have. But as militaries go, the United States military does not seem to be unusually massacre-prone, indeed if anything it appears to be somewhat unusually massacre-avoidant in the post-1900 period. I’ve written at length elsewhere on this topic, but there are certain institutional-culture patterns which produce what I’ve termed ‘atrocity-prone’ armies and the United States military has generally tried to avoid these patterns, whereas some other militaries – I discussed Russia in the link above – seem to embrace them. The United States has traditionally relied very heavily on firepower and that has driven a lot of the ‘collateral’ civilian casualties in modern American wars: soldiers can (if they choose) discriminate between civilians and combatants, but bombs and missiles cannot. But American soldiers, it seems, generally do try to make that distinction, again, not always and this should not be taken to minimize the horror of the times this was not true, resulting in American armies post-1900 generally coming in on the lower-end of the destructiveness scale – as hard as that can be to believe when looking at photos of flattened cities.
Must it be this way? I think there are three seemingly contradictory but true answers to this. On the one hand, yes – so long as we humans fight wars on this increasingly urbanized planet, we’re going to get military operations in dense urban areas in which terrible, gut-wrenching civilian suffering cannot be fully avoided. There is no way to fight ‘clean’ in an urban area. On the other hand, no – while civilian deaths cannot be entirely avoided, I think it’s clear from the statistics above that they can be reduced, that while one cannot fight clean in cities, it is possible to fight cleaner, to be merely bad instead of terrible. Carefully discriminate targeting, a preference for infantry operations over airpower (though of course, infantry need air and artillery support in modern warfare – this is a spectrum, not a binary), and an insistence both in regulations-as-implemented (read: discipline) and training that civilian casualties are to be avoided and failure to do so results in accountability can all keep the terrible toll lower, if not low.
But there is a deeper and more profound no here: we could choose not to war. As we’ve discussed at several points here, I think it is correct to say that both human society and human biology have been evolved for war. For hundreds of thousands of years, the humans that fought the most and the best thrived and passed those traits on to the next generation. But our genius and skill have now created a world in which peace is so profitable (because of industrial production) and war so destructive (because of industrial firepower) that to war is now a maladaptive trait. The advent of nuclear weapons has only intensified this fact by giving us the power to destroy ourselves; such that, as Bernard Brodie put it, “Thus far the chief purpose of our military establish has been to win wars. From now on its chief purpose must be to avert them. It can have almost no other useful purpose.”4
War may never fully and entirely go away, but if we wish to survive as a species, we must tame our instinct for it. And here is where, when I close out the last lecture of my global military history survey, I point out to my students that, because we live in a democracy, it is we who must take to that task of taming our instinct for war.
No one will do it for us. And if we fail, no one will be left to do it after us.
On to recommendations:
For the Tolkien-minded, a fun treat, demographer Lyman Stone has attempted to work out a reasonable population estimate for Tolkien’s Middle Earth during the Third Age. It’s a fun exercise and the fact that figures broadly work out in terms of correlating army size to settlement size to settlement patterns to population once again brings home how effectively Tolkien seems to have grasped all of this even though I do not think that at any point did Tolkien sit down and ‘math out’ the population of these places. He just knew about how many soldiers a kingdom of about a given size might have in the Middle Ages and about how large its cities might be and so everything tends to fall within a fairly plausible range – a skill that some other creators in the fantasy space clearly lack.
I though this Twitter thread by Wesley Morgan (author of the excellent The Hardest Place) discussing the normal range of civilian casualties in urban combat was sobering but valuable reading. As I noted above, there are two kinds of errors here, the first assumes that nothing can be done to avoid civilian casualties in urban environments and the latter assumes that anything higher than zero indicates the intentional targeting of civilians. Morgan’s example – the Second Battle of Fallujah – does not represent ‘perfection’ because perfection in these kinds of messy operations is impossible, but it probably does represent an unusually careful effort at avoiding civilian losses. And on the one hand, civilian deaths were much lower there than in other urban battles…and on the other hand, depressingly, tragically a lot higher than zero.
I have a chapter in the now out Brill’s Companion to Diet and Logistics in Greek and Roman Warfare, ed. J. Donahue and L. Brice (2023). Alas, the book is not priced for mortals, so it won’t get a blog recommendation (I think it is very good, but I try to stick here to books that normal people can afford!), but, due to a generous funding grant Jeremy Armstrong’s excellent chapter on “Diet and Nutrition in the Roman Republican Army” is free to download at the link above, so check that out if you are interested. I may do a public-facing essay at some point covering some of the same ground as my chapter there, but in the meantime I’d note that a lot of my insight – such as it is – already made it into the series on foraging and logistics.
Finally on to the book recommendation, where I am going to recommend Christopher J. Fuhrmann’s Policing the Roman Empire: Soldiers, Administration and Public Order (2012). Fuhrmann’s book presents a series of illustrative studies detailing how Roman power actually worked on the ground in the provinces, particularly for common folk and how Roman governors and lesser authorities did – and failed to – keep their provinces ‘pacified and quiet.’5 While Fuhrmann does discuss the role of higher officials and indeed the emperors themselves in managing the provinces, his focus – in contrast to much of the historiography in provincial administration, which tends to focus on emperors and the senatorial elite who write our sources – is often on lower-level officials and the points of contact between imperial power and regular people.
The book is organized as a series of case studies rather than a single schematic description and I think it gains from this approach. The fact is, our evidence for Roman ‘order’ on the ground is patchy, thin and difficult. By approaching the question as a series of case studies, Fuhrmann can focus on where he does have evidence, but these studies also knit together really well to provide a good sense of the overall principles of the system and how it functioned. Fuhrmann can thus weave out of these case studies an overall thesis, namely that the Roman state was rather more involved in the maintenence of public order in the provinces than is sometimes assumed by scholars, but at the same time that role was mostly about sustaining and coopting existing systems of power and authority: the Romans kept order for Rome first and for the local elite second and for the common populace at best a long distant third, if at all.
This could easily make for a bewildering, confusing topic with a mess of different local officials, self-help regulations and soldiers, behaving both well and badly, but Fuhrmann is very good at keeping the volume readable and accessible. I think the non-specialist reader may have to run for Wikipedia at a couple of points, but only a couple: for the most part Fuhrmann defines his terms (he always translates his Greek and Latin) and gives enough context to figures readers might not know to allow even the non-specialist to understand the argument being made and the strength of the evidence supporting it.