Michael KreponSaving Cities

Aspiring wonks, here is your end-of-semester exam question: It’s not OK to use cluster munitions in metropolitan areas, but it is OK to use nuclear weapons against targets that fall within or close to them. Yes? No? Under some circumstances? Explain.

States that possess nuclear weapons are reluctant to argue whether and how their use applies to the laws of armed conflict. To do so would risk undermining deterrence by nullifying battlefield applications, except as a last resort and for responses in kind. Even here, I suppose legal scholars, like The Hague Court, would have more than a few words to say.

Cluster bombs are not supposed to be used in built-up areas because they can have indiscriminate and long-lasting effects. Nuclear weapons, on the other hand, are widely presumed to be targeted against command and control, war-supporting industry, and leadership targets in and around cities.

Secretary of Defense Robert McNamara tried (briefly) to place metropolitan areas off-limits to nuclear targeting. He gave a commencement speech at the University of Michigan in June 1962 on this subject. Here’s the key passage:

The US has come to the conclusion that, to the extent feasible, basic military strategy in a possible general war should be approached much the same way that more conventional military operations have been regarded in the past. That is to say, principal military objectives, in the event of a nuclear war stemming from a major attack on the Alliance, should be the destruction of the enemy’s military forces, not of his civilian population… In other words, we are giving a possible opponent the strongest possible incentive to refrain from striking our own cities.

McNamara’s good intentions were defeated by the logic of Cold War nuclear deterrence. The imperative of damage limitation meant that high-value targets in or close to cities could not be stricken from targeting lists – particularly since there could be no assurance of comparable restraint by an adversary once the nuclear threshold was crossed.

Can you imagine hearing an address about nuclear targeting at your college graduation? Thankfully, I’m no longer in the business of reading or grading exams. Still, the question lingers, especially when the answers are unpersuasive.


  1. Gregory Matteson (History)

    The fatal arguments for the use of wide area weapons away from the battle front prevailed leading up to and during WWII. The laws of war, such as they are, reflect the abhorrence with which we view much of that use with in hindsight.

    In my view the potential use of nuclear weapons amplifies that abhorrence. First use of nuclear weapons by any nation will certainly earn that nation great infamy. If first use were against civilians, that would certainly be ethically and morally unacceptable; a great crime.

    The major nuclear powers, specifically the US and Russia, have conventional weapons capable of dealing with any reasonable contingency. (I’m basically talking about ‘bunker busters’ here) It is safe to say that any use of nukes between major powers would be the end of the world as we know it.

    If a minor nuclear power were to attack a city, regardless of the real or imagined ‘legitimate’ military targets in the area, it would almost certainly bring down the wrath of the world, and likely of the major powers.

    This would present yet another moral dilemma: should the offending nation be annihilated, or somehow punished? The UN Charter, and other human rights agreements condemn collective punishment, or the purposeful elimination of a population; yet the hideous logic of capital punishment prevails. By this I mean, if the repetition of a crime is seen as truly intolerable, capital punishment has zero recidivism.

    The actual precident on the punishment of nations reflects very poorly on that ‘alternative’. The collective punishment of the losers in WWI may almost certainly be said to have resulted in WWII. Then, at the end of WWII, even Stalin eventually relinquished the desire to eliminate the German State, and the World in general eschewed the punishment of nations, with apparently very positive results.

  2. John Schilling (History)

    I am not aware of any absolute prohibition against using cluster weapons in cities, save for the Convention on Cluster Munitions which A: says you’re not allowed to even have the things in the first place and B: remains conspicuously unsigned my most of the world’s serious military powers.

    Which leaves us with the general rules of warfare, which have always allowed exceptions for military necessity. If you are justified in fighting a war in the first place, and you can’t win without cluster-bombing a city, it is “OK” to cluster-bomb a city. Or nuke it. That neither of these things has happened since 1945 does not establish a new rule, it just means things haven’t gotten that bad yet.

    So the cluster-bomb / nuke analogy may be illustrative, but the premise that cluster-bombing cities is Absolutely Wrong is not. We need to figure out where that line really is.

    • George William Herbert (History)

      There is the standard conventions “proportionality” standard; using a rifle to kill a sniper in an occupied appartment building, ok. Rocket launcher, still ok. Tank gun… Ok, ish. 2000 lb bomb? Not so ok.

      I think 155 mm artillery and Hellfire Missiles are sort of the borderline for that sort of stuff now, and we’re seeing significant efforts to get PGMs smaller to do less collateral damage.

  3. Cthippo (History)

    The problem with Cluster Bomb Units (CBUs) is one of reliability. If they worked like they were supposed to 99.8% of the time instead of leaving unexploded bomblets lying around which become de facto land mines, then no one would have a problem with them. All bombs have a certain fuzing failure rate, but this rate is so bad with CBUs that they have been banned. No other reason I’m aware of.

    I suspect that the first nation to use nuclear weapons will see a Gulf War 2 style regime change in which a massive international coalition of forces will come together to foricbly remove the government that made the use decision, but the population will be spared to the extent possible. I think in such a case there will be the political will suffer the kinds of casualties it would take to accomplish such a goal.

    World War 2 was fought the way it was not because the goals were that different, but because that was the only way we had to fight the war. We wiped out whole cities because we couldn’t get the targets contained within any other way. The goals of the war were always framed in the context of defeating Hitler / Tojo / Mousollini , not annihilating the German / Japanese / Italian people. This is demonstrated by how allied forces treated the people in occupied countries. Even the Russians, who saw the Great Fatherland Patriotic War as more of a war on the German people than the rest of the allies did, and committed some fairly nasty atrocities, never made a policy of wholesale slaughter, genocide, or displacement of conquered civilian populations.

    Times and technology have changed, but the goals of war between civilized states have not. We made war on the Taliban and government of Saddam Hussein, not the Afghan and Iraqi people. Russia attacked the Georgian military, not the citizenry.

    Use of a nuclear weapon would violate this norm of state vs state, government vs government, instead of population vs population. This norm probably goes back at least as far as the 19th century in the west and underlies nearly all international law on warfare. As such, a major violation of it would, in my opinion, be countered by a massive conventional response that was keeping with the norm of removing offending governments, not populations.

    • FOARP (History)

      “the Russians, who saw the Great Fatherland Patriotic War as more of a war on the German people than the rest of the allies did, and committed some fairly nasty atrocities, never made a policy of wholesale slaughter, genocide, or displacement of conquered civilian populations.”

      By modern standards, the USSR policy of deportation of nationalities involved in collaboration with the Germans (e.g., the Volga Germans, the Chechens) to labour camps would very probably be considered genocide. At the very least it is not clear how the treatment meeted out to these peoples differs significantly to the genocide and ethnic cleansing seen in Bosnia.

  4. j_kies (History)

    Really misses the big point:
    Cluster munitions have percieved utility in maximizing effects per carried mass, not targeting civil populations minimizes collateral damage.

    Nuclear weapons applied against hard targets would almost assure ground bursts. Ground bursts will have drastic fallout impacts especially on civil populations. I would rather sit on the intended target locale – lawn chair and dark glasses please – rather than suffer delayed death due to radiation sickness. Dead is dead, prompt >> slow.

  5. Jonah Speaks (History)

    Other approaches to nuclear weapons and cities have been suggested, though little discussed. An alternative approach is to demand evacuation of cities. Leo Szilard, the physicist who invented the atomic bomb, suggested that in a pinch, the U.S. and Soviet Union could bomb one or two or three cities of equal population. They would first provide adequate warning and allow adequate time (perhaps 2 to 4 weeks) for the cities to be evacuated. After that the cities would be nuclear bombed. Szilard intended this as a rational alternative to unlimited nuclear war.

    Independently, I developed a similar idea. Of course, cities could be bombed after evacuation, but normally a better alternative is to accept the evacuations in lieu of bombing. Evacuation imposes costs on each nation, that motivates both sides to find some agreement to resolve the war or other crisis dividing them. After agreement is reached to resolve the conflict, evacuated populations can return to their cities. Ordering evacuation is a rational alternative to nuclear war and it requires no death, destruction, radiation, environmental damage, or nuclear winter. This alternative approach is more sane, more prudent, and more ethical.

    • John Schilling (History)

      Another alternative would be a formalized “open city” status in which a reputable third party verifies through intrusive on-site inspection that the city is not being used to host or support any military activity, whereupon it would become a war crime to bombard that city.

      Both open cities and evacuations have precedent in WWII, though not as formalized as we might like. And prior to WWI, the issue was further simplified by the fact that a city could only be threatened if an army or a fleet fought its way to within artillery range.

      But the general expectation that cities wouldn’t be destroyed and their populations slaughtered in wartime, was historically coupled to fairly specific rules about cities surrendering – and meaning it – when they could no longer be defended. The cities that didn’t surrender when their defenses were proven inadequate, might as well have been cluster- or atom-bombed.

      And sometimes the cities that did surrender got that treatment as well, but that’s a place where we can realistically draw the war-crime line. Insisting that cities remain untouchable even as they continue to produce insurgencies and even armies, is probably not.

    • Jonah Speaks (History)

      John, Thank you for the information on open cities in World War II.

      Just war principles are a curious admixture of moral principles and practical wartime experience. Even though it violates all the other just war principles, the city siege (starve all the civilians until the military surrenders) has been viewed as a general exception to what would otherwise be viewed as an abominable war crime. In World War II we bombed the cities instead of starving them, so perhaps that is simply the modern extension of ancient principles of just war.

      My primary objective in proposing a system of orders for evacuations is to suggest a rational alternative to the devastation of nuclear war. In this or future centuries, it is not unlikely that another great power war may arise between nations that already have (or know how to build) nuclear weapons. Instead of fighting a nuclear war, it is rational and mutually beneficial to agree to fight a war of evacuations. Even if the nations cannot immediately agree on how to resolve the conflict, they can agree to fight the war by rules that avoid mutual destruction.

  6. George William Herbert (History)

    I think that there’s a tactical (and strategic) mistake being made here.

    The various norms of warfare – “Geneva Conventions”, land mine treaty, etc etc. – are all intended to proscribe or focus certain actions in warfare which is expected to keep happening. I.e., limit but don’t stop the underlying conflicts.

    The norm regarding WMD really ought to be evolving towards “this is not supposed to happen”. It has much of a consensus in that in much of the world, but not among all of the capable players. And not consistently across the chemical and biological weapons fronts.

    The west (and Russia) have developed that across the CB areas. But there are still chemical weapons out there.

  7. Paul Bernstein (History)

    Michael: a few thoughts on your post, regarding the history and the morality question (sorry, nothing on cluster bombs). McNamara will always be remembered as the avatar of assured destruction, but it is important to recall, as you have done, that his strongest instinct was to protect cities to the degree possible. I don’t think he harbored any illusion that cities could be kept completely “off limits” either in principle or practice, but as the last line of the quote you offer states, perhaps there was hope to create incentives to minimize the worst attacks on population centers. His thinking was greatly shaped by the “No Cities” work done at Rand by Andy Marshall and Bill Kaufmann, which itself was a culmination, in the operations research world, of the growing challenge to the largely indiscriminate and spasmodic SAC war plan. And with the advent of reconnaissance satellites that helped pinpoint the locations of Soviet missiles, we could, presumably, move toward a counterforce strategy. It turned out there were a lot fewer of these missiles than we had earlier thought.

    McNamara retreated from No Cities for reasons that are well-understood – the interpretation of it in Moscow as a first strike strategy, negative reaction of the NATO allies and, perhaps most important, the emerging budgetary implications. The Allies saw the threat to Soviet cities as the strongest form of deterrence, so yes, this particular “logic of Cold War deterrence” (there are others) contributed to McNamara’s retreat. The fiscal factor – well it just looked too expensive once the Services prepared their wish lists. I’m not sure the logic of Cold War deterrence had anything to do with it. Seeking a more affordable standard for force development and procurement, McNamara shifted toward Assured Destruction and ultimately established the now famous AD metrics as the sufficiency criteria for US forces. These metrics reflected (understated, more likely) extant US capabilities against the Soviet urban-industrial target structure rather than any understanding of Moscow’s tolerance for damage. But it was believed that anything beyond these levels would provide rapidly diminishing returns. So in this sense it was as much a resource management strategy and a means to resist calls for major new strategic force programs as it was a calculated deterrence strategy. McNamara didn’t know if this was the best way to deter the Soviets, but he did know that the alternative was going to be expensive.

    It’s also been argued that McNamara saw the adoption of AD as a way to warn against the dangers of nuclear war. Freedman makes this argument, suggesting that McNamara was concerned that “through euphemistic language, the circumlocutions endemic to military briefings, and ill-disciplined logic, the Air Force and its Congressional supporters were deluding themselves into thinking there was a tolerable way of fighting a nuclear war.” In this sense, the blunt language of AD can be seen as a kind of antidote to forms of strategic analysis and rhetoric that tended to treat nuclear war as an abstraction. This also suggests that McNamara did not view counterforce and damage limitation as a strategy for waging nuclear war, even though there is an irresistible tendency in parts of our community to equate any approach that diverges from holding cities hostage as synonymous with “nuclear warfighting.” This begins to engage the morality question.

    The arguments against holding cities hostage have always reflected a mix of strategic efficacy (what works best?) and morality (what’s more or less ethical?). The moral critique derived in part from Just War Theory, which proscribes deliberately targeting civilians. But the defense of deterrence mounted in the 1980s by the Bishops and others also drew on moral arguments. There were plenty of thoughtful people who accepted MAD not just as a strategic norm, but as an ethical norm as well. I think the explanation (or at least part of the explanation) is, as I suggested above, that MAD was framed as the opposite not of damage limitation, but of nuclear warfighting. And if your objective is to deter and prevent nuclear war, it’s hard to argue that a warfighting strategy is the more effective policy. If less effective in preventing war, then less moral by definition. In this way the efficacy argument trumps the just war proscription against targeting civilians. It was possible to make this case in the early Reagan years because of all the loose talk about surviving and winning a nuclear war. SDI complicated the picture by resurrecting damage limitation. Had all the loose talk about nuclear war been avoided, perhaps the debate about the morality of deterrence would have been different, framed more usefully by the renewed focus on damage limitation. It’s no surprise that the talk of warfighting and the talk of damage limitation were conflated, but on their own terms they are different things, practically and morally. And I’d argue that any discussion of nuclear morality needs to make distinctions between assured destruction, damage limitation, and nuclear warfighting as approaches to or elements of strategy. That may not make the moral question any less of a tangle, but it can’t hurt.

    McNamara did not embrace AD because it was morally preferable, but because it was more practical. This was in no way a repudiation of his basic view on the need to protect cities and move nuclear targeting and employment planning toward greater counterforce emphasis and greater discrimination, control, and decision making flexibility for the President. If there was a moral imperative, it was this, and this is what he started to do in the SIOP process, largely outside of public view and, admittedly, very incrementally. But he remained committed to the principles and objectives of the No Cities paradigm in the actual practice of nuclear planning, even as he embraced AD as declared policy and the criterion for spending. So despite the public retreat from No Cities, I’m not sure this should be characterized as a defeat for McNamara’s good intentions. And yes, as you note, there was only so much that could be done to limit damage to populations as we started to move toward counterforce targeting, given target locations and weapon system accuracies. But we were able to do more as time passed, so here too I would not necessarily frame this as a defeat of his intentions.

    All this may raise some potentially interesting questions about what states say about their nuclear policies and strategies, and what they actually do as a matter of planning and force development, and what this means for deterrence and risk management. Though looking back, maybe not in the Soviet case; it seems they never subscribed to the concepts of limited nuclear war that increasingly shaped US policy in the 60 and 70s. In the event, maybe they would have embraced restraint; this was certainly our stated hope.

    Sorry to be so long-winded. Hope this helps.

    • krepon (History)

      Much appreciated.

  8. Bill (History)

    To elaborate one of Mr. Bernstein’s points, it’s worth showing how “no cities” thinking influenced the major SIOP attack options that developed under McNamara’s auspices. They were nuclear strikes versus 1) Soviet nuclear forces only except those located near cities, 2) non-nuclear forces, and 3) urban industrial targets and those nuclear targets located near cities. The first category was plainly a stand-alone no-cities counterforce option that could be launched preemptively or on a retaliatory basis. Once you get into categories 2 and 3 you’re moving into all-out nuclear war, where thoughts about sparing cities are out the door. Of course, as Stan Norris and others have noted even no-cities options could cause great harm to urban populations from fallout effects.