Good morning – or rather, good day, since while it is still morning here in Washington, it will be mid-afternoon for you in Vienna. Anyway, it is a pleasure to be able to take part in this discussion, and I hope our video connection holds up.
I. Debating Deterrence
There is irony in the fact that it may be the very success of nuclear deterrence over the decades of the Cold War that has brought about a situation in which it is surprisingly common to hear it said that nuclear deterrence is a fantasy, and that nuclear weapons are unnecessary for anyone. To my eye, nuclear weaponry does seem to have helped check the traditional tendency of great power competition to escalate into general war. Precisely because there is no counterexample of a postwar world that did collapse again into conflict, however, it is now sometimes claimed that our nuclear efforts were unnecessary all along – and that nuclear deterrence is a quaint illusion of which we should now rid ourselves.
I disagree. There is no escaping the fact that despite the occasional crisis, there has not been a full-up general war between the great powers in the nuclear age, nor indeed has there been one between any nuclear-armed states, and this represents a remarkable departure from the previous few hundred years. The question, of course, is what to make of it. The usual interpretation is to conclude that nuclear deterrence did contribute to the postwar peace, and that "extended" nuclear deterrence has helped prevent additional countries from acquiring nuclear weapons. Is there, however, a persuasive reason to throw out the idea of nuclear deterrence, the conceptual foundation of the nuclear security with which so many have lived for so long?
I do not believe that there is. We should be very wary of revisionist reexamination of what seems to have worked for so long, especially when the evidence for a contrary view is so scant and tendentious. Nuclear deterrence was never a panacea, no one ever claimed that it would or could perfectly, and we indeed face significant challenges in applying it in today's world. Nevertheless, it would be a great mistake to dismiss its relevance, in toto, without very good cause indeed. And there is no strong case for doing so.
II. Analytical Challenges
Analytically speaking, of course, part of the challenge in assessing these questions is that history provides us with a terrible data set for "scientific" assessment. At best, we only know what actually happened, we can't control for different variables, and we can't run the experiment over again to see how different approaches affect outcomes. All we really have to work with, in effect, is a body of anecdotal evidence, to which we can do little more than good sense in good faith.
Using simple good sense can still be useful, however, since not all lines of argumentation stand up to it very well. I would contend, for instance, that a common-sense approach to analyzing the history of nuclear deterrence should pay little heed to historical cases that do not seem actually to have involved questions of deterrence in the first place. This is why I think – with apologies to Ward, whom I know has – that the much-discussed issue of the atomic bombings of Hiroshima and Nagasaki is no more than an analytical red herring. The use of nuclear weapons in 1945 was about how to coerce one side in a massive and prolonged general war into giving up the fight. It was, in other words, not about deterring war but about war termination, and those are notably different things. The question concerned defeating Japan, and not just keeping it from doing something that it was not yet doing anyway. As a result, it wasn't about "deterrence" at all, from which it follows that whatever role atomic weapons really played in bringing about Japan's surrender, this answer is of little relevance to the study of nuclear deterrence theory today.
That said, of course, it is not entirely clear precisely how relevant our long Cold War experience with actual nuclear deterrence really is today either – and here lies what I think is an important point. I have yet to see any persuasive debunking of the generally-accepted notion that nuclear deterrence played a critical role in the strategic stability of that period. Nevertheless, it does not necessarily follow that the same approach that deterred the USSR then is the right recipe with which to deter others today. It might in fact be, but whether or not this is so is a critical question. And this points us to one of the most vexing challenges of the study of deterrence: there probably is no fixed answer to the question of what deters whom. Deterrence is subjective, in that it exists in the eye of the one who is deterred, but what deters him has no necessary connection to the anticipations of his would-be deterrer.
One obvious implication of this subjectivity is that deterrence strategies should be tailored to their targets as specifically as possible, for what deters one country from doing one thing may be somewhat different from what deters a different one from doing a different thing. (What deters any particular target, of course, may also change over time, so these questions will presumably have to be constantly re-asked even if one has gotten the answer "right.") This is obviously difficult to implement in practice – and it also undercuts our ability to translate lessons from one case study to another.
III. Lessons for Nonproliferation Policy
But there may be some lessons for us in these challenges all the same. Let me suggest two that pertain directly to nonproliferation.
First, I think the subjectivity of deterrence provides a powerful reason to limit the number of "players" in the nuclear game. Cold War experience suggests that it is not impossible to achieve adequate deterrent effects when the only real nuclear relationship of consequence is bipolar. This seems reasonable to me, for focusing one's efforts along this single axis of deterrent interest gives participants some opportunity to study each other and to "learn" reciprocal deterrence over time. It would surely be vastly harder, however, to do this well – or perhaps at all – as the number of players grows.
Game-theoretical literature suggests that tripolar systems are much less stable than bipolar ones, and that instability increases geometrically with each new addition to the game. It stands to reason that this is true in deterrence relationships. I might be able to figure out what deters you, for instance, but this may not be the same thing that deters the next guy. Indeed, what deterrence impels me to seek vis-Ã -vis one player might actually end up looking alarmingly provocative, or dangerously weak, to another. We are obviously talking about very complex analytical problems here, and at some point one has to wonder whether they simply become insoluble.
Since one can only have one force posture at a time, it follows that each player may have to choose between multiple different and somewhat inconsistent deterrent "packages" – such as by picking the one felt best to deter the most worrisome potential adversary and hoping that this doesn't cause too many problems elsewhere, or by choosing one that merely "satisfices" in a kind of lowest-common-denominator way against many targets while offering optimal deterrence against none. Since a deterrence failure could mean the slaughter of huge numbers of people, it isn't very reassuring to have to rely overmuch upon this kind of judgment call. This should provide us powerful reasons, therefore, to work very hard to keep the number of nuclear players from increasing: the world is likely to be much more stable if this number can be kept small. (Most of you listening right now are in Vienna in connection with the Nuclear Nonproliferation Treaty Preparatory Committee meeting, so I implore you to keep this in mind and not get distracted by issues ancillary to the imperative of nonproliferation.)
A second lesson for nonproliferation also flows from the difficulty of guessing precisely what it is that deterrence requires. Ideally, one would identify what will deter the target, and then build a force posture around that requirement. Since there is uncertainty in this, however, it is safer to err on the high side, as a strategic "hedge" against getting it wrong. (After all, while having too many weapons is hardly costless, this is less problematic than having fewer than needed to achieve deterrence. The price in that case can be war.) The more players there are in the game, moreover – and thus the more different deterrent axes one has to try to posture along, and the greater the calculative uncertainty – the greater will be each player's incentive to aim high.
And here we see a linkage between nonproliferation and disarmament different from what one usually hears in NPT circles. I do not mean merely that achieving some future nuclear "zero" will be impossible if the international community cannot demonstrate its ability and willingness to preclude the entry of new players into the nuclear game, though this is obviously indeed the case. I mean also that the difficulty of precise deterrent calculations suggests that the more players there are, the harder it will be even to achieve reductions in weapons-possessor arsenals. This logic is simple: the more participants in the game – or at least the more significant players there are, at any rate – the more pressure there will be for each to hedge against deterrent uncertainty along these proliferating axes by aiming high (i.e., building, or keeping around, larger numbers of weapons). This is another powerful reason for countries to take nonproliferation more seriously than many have hitherto done.
IV. The Deterrence Paradigm
But while I thus think that there are lessons to learn from how challenging nuclear deterrence can be, I still see no adequate alternative. Nor do I think there is good reason to conclude that the concept itself is chimerical.
And I am not alone. All nuclear weapons possessors in the real world – and a good many of their friends and allies – seem to think deterrence works, and they indeed have long acted as if it does. And of course there indeed has been no general war between nuclear weapons possessors, nor even the close allies thereof – which in historical terms is quite remarkable. Furthermore, when confronted by a potential adversary with nuclear weaponry or with conventional superiority, non-possessors with the technical option of doing so have shown a notable tendency to seek nuclear weapons themselves. (They are sometimes talked out of it, especially when they are reassured that their interests will be protected by some relationship with a strong ally that often itself possesses nuclear weapons, but to my eye that tends to prove the point.)
Even if one cannot "prove" that deterrence works in a meaningfully scientific sense, this is all highly suggestive. It may not make the case in an impeccably rigorous and scientifically defensible way, but it is very hard to speak coherently about "proving" any such interpretive proposition – one way or the other – when looking back upon the once-through multivariant puzzle of the historical record. There certainly seems to be a vastly stronger case for nuclear deterrence than for the proposition that the entire concept has always been fantastical.
If anything, I would suggest that the argument needs to be turned around. We are not debating what course of action to take as if writing on a blank sheet of paper. We are debating whether to uproot long-established concepts that form the foundation of many countries' most important security policy choices – a framework for which no coherent alternative has yet been offered, which seems to have helped prevent general war for a long time, and which hasn't in any event been shown to be so broken today as to justify its repudiation. Under the circumstances, the burden of "proof," such as it is, should really be reversed: it should lie with anyone who wants to overturn the world's ongoing reliance upon nuclear deterrence by asserting its irrelevance. So far, I have heard nothing to suggest that this burden of disproof, as it were, is anywhere near being met.
In an uncertain and still dangerous world, strategic planners have an obligation to plan against contingencies, hedging against uncertainty and unpredictability in part by avoiding choices that tie us inescapably to assumptions that might turn out to be dangerously wrong. Even if we believe in nuclear deterrence only in the same sense that Pascal famously suggested one should believe in God – that is, because the cost of doing so in error is lower than the cost of not doing so in error – I'd say we have every reason to keep on believing.