Michael KreponLimited Nuclear Options

The first impulse of states that acquire nuclear weapons is “simple” deterrence, and the simplest form of deterrence is massive retaliation to existential threats. (India and Pakistan, for example, now embrace declaratory policies of simple deterrence.) The second impulse, made possible by the growth of nuclear capabilities, is to avoid “all or nothing” decisions that could effectively result in the complete destruction of national territory. Thus was born the notion of limited nuclear options.

Here’s how the Nixon administration thought about the problem, as codified in NSDM 242, issued on January 17, 1974:

The United States will rely primarily on U.S. and allied conventional forces to deter conventional aggression by both nuclear and non-nuclear powers. Nevertheless, this does not preclude U. S. use of nuclear weapons in response to conventional aggression.

Should conflict occur, the most critical employment objective is to seek early war termination, on terms acceptable to the United States and its allies, at the lowest level of conflict feasible. This objective requires planning a wide range of limited nuclear employment options which could be used in conjunction with supporting political and military measures (including conventional forces) to control escalation.

Plans should be developed for limited employment options which enable the United States to conduct selected nuclear operations, in concert with conventional forces, which protect vital U. S. interests and limit enemy capabilities to continue aggression. In addition, these options should enable the United States to communicate to the enemy a determination to resist aggression, coupled with a desire to exercise restraint.

Thus, options should be developed in which the level, scope, and duration of violence is limited in a manner which can be clearly and credibly communicated to the enemy. The options should (a) hold some vital enemy targets hostage to subsequent destruction by survivable nuclear forces, and (b) permit control over the timing and pace of attack execution, in order to provide the enemy opportunities to reconsider his actions.

Nuclear deterrence theory works best at the conceptual level. As Hans Morganthau once wrote, “The very purpose of threat and counter-threat is to prevent the test of actual performance from taking place.” But here’s the rub: Deterrence concepts must be credible, which means that they must be translated into operation plans. And when concepts become plans, they strain credulity.

No operational plans are more “incredible” than those for limited nuclear war. In Robert Osgood’s best known book, Limited War: The Challenge to American Strategy (1957), he argues that credibility “requires that the means of deterrence be proportionate to the objectives at stake.” Osgood’s view was challenged at the time by those who believed that credible nuclear deterrence required more punch than the enemy could take. As General Curtis LeMay testified before the Senate’s Airpower Hearings in 1956,

A deterrent force is one that is large enough and efficient enough that no matter what the enemy force does, either offensively or defensively, he still will receive a quantity of bombs or explosive force that is more than he is willing to accept.

The dilemma of proportionality vs. leverage has bedeviled nuclear deterrence theory ever since. Reverse mirror imaging hasn’t helped matters: What is perceived as a deterrent by Country X looks a lot like a nuclear war-fighting force in Country Y. The scales of this debate in the United States tipped in the 1970s as Soviet nuclear capabilities grew. (Aspiring Wonks who want a sense of the debate over “prompt hard-target kill capabilities” might check out Paul Nitze’s Deterring our Deterrent in the Winter 1976-77 issue of Foreign Policy.)

The best and the brightest nuclear deterrence theorists – Nitze, Herman Khan, Henry Kissinger, and Albert Wohlstetter, to name a few – failed to crack this nut. Their arguments were unpersuasive because the entire concept of limited nuclear options rested on heroic and questionable assumptions. As Bernard Brodie wrote in his classic book, Strategy and the Missile Age (1959), “It takes only one to start a total war, but it takes two to keep a war limited.” If the stakes in a confrontation between nuclear-armed states were so great as to warrant a crossing of the nuclear threshold, how would they not warrant upping the ante? How would leaders evaluate threat and damage assessments in the chaos of a nuclear battlefield? Would command and control over nuclear forces remain intact in these circumstances? Would national leaders in such discord as to find themselves in a nuclear war somehow manage to find the same page to keep it limited? And how does the quest for securing advantage accord with the very notion of a limited nuclear war?

Given the stakes involved in potential nuclear confrontations, it would be irresponsible for national leaders to refuse to authorize planning for limited nuclear options. But because so many core questions about crossing the nuclear threshold have no good answers, it would also be irresponsible for national leaders to authorize the employment of limited nuclear options.


  1. Tom (History)

    I always thought it was interesting that even Clausewitz touched on this subject. In the beginning of On War, he discusses why wars don’t immediately escalate to one gigantic decisive battle (besides the fact, which he recognizes, that for conventional armies this is practically impossible):

    “But the possibility of gaining a later result causes men to take refuge in that expectation owing to the repugnance, in the human mind, to making excessive efforts; and therefore forces are not concentrated and measures are not taken for the first decision with that energy which would otherwise be used. Whatever one belligerent omits from weakness, becomes to the other a real objective ground for limiting his own efforts, and thus again, through this reciprocal action, extreme tendencies are brought down to efforts on a limited scale.”

    In the context of a limited nuclear war, I wonder if that “repugnance” is eliminated by the relative ease of escalation: all one has to do is turn a a few more keys a push a few more buttons.

  2. Chris van Avery (History)

    I wrestled with this nut for a bit and the only realistic option I saw was some sort of collective deterrence strategy.

    Any thoughts?

  3. John Schilling (History)

    Collective deterrence tends to suffer a credibility gap, in that it is almost always in the interests of each deterring partner to “cheat” and let their allies do the actual nuking while they play the diplomatic voice of reason. In economic terms, you’ve essentially created a deterrence cartel. Cartels are a wonderful thing for all their members so long as all the members play by the rules – which never lasts.

    Collective deterrence in the form of one nuclear power placing several non-nuclear powers under its protection, is credible to the extent that the allies are perceived as critically important to the nuclear power – you need the cost of losing an ally and losing credibility, both at the worst possible time, to seem greater than the cost of directly waging nuclear war.

    Hmm, just realized that this is one case where it is beneficial to espouse a doctrine of limited nuclear warfighting, even if one privately thinks it is utter nonsense.

    Combining both forms of collective deterrence, several nuclear powers defending many non-nuclear powers, might work better than either alone, or it might not. The NATO example seems pretty good to me – the British and French add a bit of uncertainty to an attacker’s calculations without undermining the credibility of America’s deterrent. And NATO at least traditionally consisted of nations which were mostly vital to each other’s core interests.

    NATO with several small rather than one big and two small nuclear powers, would I think run into the cartel failure mode.

    Extending NATO to e.g. the Baltic states, seriously weakened the credibility of both its conventional and nuclear deterrence, I think. Anything much broader would be a joke, for any collection of nations or other entities that actually exists in the real world.

  4. MarkoB

    Deterrence is an ideology, not a “theory” and Kahn et al were ideologues not theorists. The US maintained nuclear forces during the cold war in order to compel, not to deter. The vast destructive power of nuclear weapons needed to be justified. Deterrence functioned as that justification, but policy had little to do with deterrence.Overkill went well beyond deterrence, even the 2010 NPR states that deterrence is overdetermined. Furthermore deterrence “theory” was also used to justify military buildups domestically that went well beyond the needs of national security. That’s why deterrence is an ideology. It’s very easy to have theological debates that have little to do with reality. Debating “the requirements of deterrence” is like debating how many angels can fit on the end of a pin.

  5. Vijai Nair (History)

    I note you have invented yet another terminology i.e. ‘simple dererrence’! And whats even more confounding, ascribed it to the the doctrines of both India and Pakistan. Nothing could be further from the actual, as can be affirmed by the declared nuclear doctrines of both. A lot has been written about ‘nuclear deterrence’ but there is no precedence that confirms any of the different avtars of this term – at best they are untested hypotheses emanating from different quarters in an inexplicable psychological game to convince concerned governments that they are secured from a nuclear scourge. Deception – of self and the assumed adversary is an unquantifiable mechanism prone to self deception. … As John Schilling puts it ‘collective deterrence’ suffers from a credibility gap. But so do all the others (in varying degrees). My response to that is treaties, agreements and other arrangements between ‘peace time’ coalition partners are unlikely to pass the test of nuclear conflict. Neatly analysed equations are susceptible to drastic changes under actual conflict conditions. Magoo

  6. anon (History)

    General Lee Butler once told me that he bacame an abolitionist (after being the CincSTRATCOM) because, in every exercise they ever conducted, the conflict quickly escalated from the employment of limited options to full out global nuclear war. They could NEVER keep it limited. Hence, as they said in Wargames, the only way to win was not to play.

  7. MK (History)

    Many thanks for these thoughtful comments.
    Magoo: It’s no fun if we can’t make up new terms.
    Agreed: There’s nothing ‘simple’ about deterrence. And deterrence gets more and more complicated as capabilities grow. My point here is that countries typically start this process by announcing their intention to use whatever they’ve got in a massive way. Indian and Pakistani declaratory policy currently use this phraseology.
    Best wishes,

  8. Scott Monje (History)

    I tend to lean toward John Schilling’s point on “collective deterrence.” People used to debate whether nuclear weapons or bipolarity was responsible for the “long peace” during the Cold War (that is, peace with regard to relations among the major powers, of course). As long as the Cold War continued, it didn’t matter because both conditions prevailed. Bipolarity, however, may have forced the two superpowers to concentrate on each other and counter each other’s moves in ways that enhanced predictability, ways that wouldn’t necessarily be replicated in a multipolar nuclear world. Miscalculations of the sort that prevailed in the pre-nuclear, pre-bipolar world could reemerge. I would suspect that a situation with many nuclear powers protecting many nonnuclear powers would be among the least predictable situations as each protector tried to shuffle the responsibility of actually doing something off onto others.

    On the other hand, I have a certain sympathy toward MarkoB’s attitude toward nuclear theorists, or some of them at least. I get the impression that a lot of the more detailed or precise theories were more the result of strategists’ need to justify their salaries than anything that would ever be applied to the real world. (Overkill, however, resulted from the fear that a first strike would take out most of the arsenal before it could be used.)

    Finally, regarding MK’s gratitude for “thoughtful comments,” I have to say that occasional forays into the comments posted at ForeignPolicy.com or (good God!) Politico.com have given me a profound admiration for the readers of this blog.

  9. archjr (History)

    This is a great post, so much so that it led me back, as usual, to Freedman (http://books.google.com/books?id=ruAxF8SBR3IC&pg=PA203&lpg=PA203&dq=escalation+dominance&source=bl&ots=bkySLkOugM&sig=D5iIIQ0soN5xnC0d2y27mGAtG6o&hl=en&ei=Fog_TMrDCYH48Ab8qP24Cg&sa=X&oi=book_result&ct=result&resnum=7&ved=0CDQQ6AEwBjgK#v=onepage&q=escalation%20dominance&f=false), on escalation dominance as a touchstone in thinking about “limited” nuclear war. I certainly can’t add anything insightful to what Sir Lawrence wrote, but it strikes me that these kinds of considerations are being actively contemplated in many places, from Iran and Israel to Pakistan and India, to name but a few. Too bad we don’t know much about what those countries think in this regard.

  10. John Ainslie (History)

    I was recently reading interviews with former Soviet officials involved in nuclear planning (http://www.gwu.edu/~nsarchiv/nukevault/ebb285/index.htm). There were conflicting statements on whether they planned for Limited Nuclear Options. On the one hand they wanted the US to believe that they would make a large scale response to a limited US nuclear strike, while on the other hand some practitioners prepared Limited Options.

  11. MK (History)

    This comment from John Schilling came in during the transition–

    The Soviet Union was almost certainly preparing limited options – they had the staff resources to prepare every option without even trying hard, and it would have been grossly irresponsible and implausibly stupid not to have at least a few file cabinets full of limited nuclear war plans.

    However, there was no reason for them to ever tell anyone about these plans. For the former USSR, massive retaliation in response to any significant attack on the USSR/WP was credible, so it was in their best interest to leave potential attackers believing that massive retaliation was inevitable, that there were no other plans.

    It was NATO that needed to publicize limited nuclear war plans, on account of massive retaliation by the US/UK/France in response to a Russian attack on e.g. Turkey was not credible. Turkey is important enough to NATO to require nuclear-grade protection, not so important that anyone would believe we’d set the world aflame over it, so we needed a limited warfighting plan on the record. If that made an invasion of Western Europe seem a little bit more thinkable from the Russian point of view, well, that’s the trade.