Nuclear deterrence is getting renewed attention just now. On the one hand, the Treaty on the Prohibition of Nuclear Weapons (TPNW) — with 58 states now signed on — has demonstrated that a surprising number of states seem to have doubts about relying on nuclear weapons (and by implication nuclear deterrence) for security. On the other hand, new weapons and increasing arsenals in nuclear-armed states, which some see as the beginning of a dangerous new arms race, seems to be evidence that the nuclear-armed states have no intention of foregoing these weapons (and the deterrence they provide) any time soon. The prospect of nuclear weapons being around for the long run naturally prompts a desire to be sure that the deterrence they provide works reliably. And among NGOs, think tanks, and foundations, there is a move afoot to re-examine the theory of nuclear deterrence, apparently in hopes that such a re-examination will rebuild momentum for disarmament. With all this renewed attention, it makes sense to reconsider what we know about nuclear deterrence, examine what we don’t know, and think about what work still needs to be done.
What we don’t know
You could fill a medium-sized library with the books and articles that have been written on the subject of nuclear deterrence since 1945, but despite this enormous scholarly output, there are still fundamental questions that remain unanswered.
The difficulties arise because there is a peculiar lack of facts connected with nuclear deterrence — an unusual situation for such an important policy topic. This dearth of facts flows from four realities: 1) we have very little actual experience using these weapons in war, 2) testing, in this case, is not a reliable substitute for experience, 3) we have little direct access to the mental processes that result in nuclear deterrence decisions and indirect data has significant pitfalls, and 4) the mental processes involved in these decisions are so obscure we lack even a rudimentary model of how they might work.
First, practical experience. It stands to reason that a weapon that is highly effective in war will be a stronger deterrent. If attacks with such a weapon are devastating and decisive, then threats with that weapon are likely to work, too. There are, however, questions about how effective nuclear weapons are as weapons. They have been used only twice, in a single week, in one war, against one adversary, and on only one type of target, more than seventy-five years ago. And haven’t been used since.
For most types of weapons, there is a large, well-attested set of data about their impact derived from extensive use in battle. Weapons assessments are based on use in a multiplicity of circumstances — against different opponents, on different types of terrain, against different types of targets, and so on. History teaches that experience is critical if you want to have a realistic understanding of the impact of a weapon on war.
Consider, for example, the case of city bombing. In the years between World War I and World War II, because large bombers were just developing, no one had experience with large-scale attacks on cities by bombers. Nevertheless, some experts assured governments that if cities were bombed, it would lead to almost immediate surrender. Giulio Douhet, one of the most influential of these strategists, wrote:
And if on the second day another ten, twenty, or fifty cities were bombed, who could keep all those lost, panic-stricken people from fleeing to the open countryside to escape this terror from the air? A complete breakdown of the social structure cannot but take place in a country subjected to this kind of merciless pounding from the air. The time would soon come when, to put an end to horror and suffering, the people themselves, driven by the instinct of self-preservation, would rise up and demand an end to the war—this before their army and navy had time to mobilise at all!
These experts and strategists were certain: bombing would be decisive. And yet, in the event, it turned out that civilian morale was much harder to break than some strategists expected, the damage was more quickly repaired than anyone had thought possible, and the ability of city bombing to win or even significantly affect the outcome of war had been drastically overstated. The dean of nuclear strategists, Bernard Brodie, summed up the lesson this way: “The Allies learned after the war that the attack on enemy morale had been on the whole a waste of bombs . . .”
The moral to the history of city bombing seems to be that without real experience, radical mistakes are possible. Experience is the key for estimating how effective a weapon will be in war (and therefore also as a deterrent).
This lack of data is often overlooked because of the second problem: many people assume that the extensive testing that the United States has done on nuclear weapons can substitute for actual use in war. But test explosions do not always predict all the impacts of a weapon on the battlefield. The history of military innovation shows that even the most extensive testing sometimes fails to uncover what will actually happen. Lynn Eden’s remarkable work in Whole World on Fire illustrates how difficult it is to extrapolate from tests to use in war. She showed — more than forty years after the majority of testing had been completed — that the amount of fire that would likely be created by nuclear attacks on cities had been seriously underestimated. If testing is a reliable indicator of actual impacts in war, how could such a mistake have been made? Fire is one of the five basic impacts of using a nuclear attack on a city. If it isn’t possible to get the most basic impacts right, how can we be confident that estimates of their impact on war — a complex and multifaceted activity — have been correctly estimated?
The third area where we lack key knowledge is perhaps more important than either of these other two. At a practical level, we have hardly any hard data about nuclear deterrence decisions. Deterrence is a type of threat used to dissuade an adversary from taking certain actions. Threats work or fail depending on something that happens inside the mind of your adversary. The difficulty, of course, is that there isn’t any way to accurately measure inside someone’s head. Objective truth, remember, only exists where something can be measured, tested, and others can reproduce the results.
We know a good deal about what nuclear weapons can do to buildings and cars and human bone and tissue. But this physical knowledge is mostly beside the point, because nuclear deterrence is a psychological phenomenon. What we need to know is what is happening inside a leader’s mind. But there is no way to measure that process, both because our ability to measure phenomena inside the brain is rudimentary at best and also because no leader has ever agreed to have his or her feelings and thoughts measured during a crisis. Which means that in terms of objective data, we know as much about nuclear deterrence as we know about dreams.
Most scholars try to offset this lack of objective information by gathering information indirectly — from the after-the-fact accounts by participants. This attempt to use indirect data is, however, highly problematic. Leaders learn not to “tip their hand” when decisions are in the offing. Guessing a leader’s thought process is a little like watching an expert poker player. You can see this inscrutability in action in the varying interpretations of President Kennedy’s state of mind during the Cuban Missile Crisis. He was a hawk pushing for a sudden strike until almost the last moment, according to some. He was a dove almost from the first, covertly pressing for a blockade, according to others. Sixty years later, there is still no agreement.
The more leaders conceal their thinking, the more interpretation has to be used. But the difficulties are considerable. For example, for more than seventy-five years advocates of nuclear weapons have pointed to the radio broadcast of the Emperor’s Rescript (a form of official pronouncement) on August 15, 1945 as proof that Japan’s leaders were coerced into surrendering by the atomic bombings of Hiroshima and Nagasaki. In that Rescript the Emperor mentioned that a “most cruel” bomb had been used on Japan and didn’t mention any other reason for Japan’s surrender. So it seems obvious, they argue, that the radio broadcast explains his thinking about the surrender.
But it turns out that this conclusion is only clear-cut if you have interpreted the data. There was a second Imperial Rescript two days later on September 17, 1945 that was disseminated and read aloud to all of Japan’s soldiers and sailors. In this second official explanation of the surrender, the Emperor talked about the Soviet Union joining the war and didn’t mention the atomic bombings. The Emperor’s two Rescripts actually give no clear indication why Japan surrendered. But by interpreting the Emperor’s radio broadcast as indicative of the Emperor’s state of mind, and ignoring his second Rescript as irrelevant, some scholars have created certainty where none existed. The problem with interpretation is that people often see what they are predisposed to see. Indirect data is sometimes useful. But the obstacles to building a complete and accurate understanding of events are dauntingly high.
Finally, consider how little can be said about how the process works. What happens inside someone’s head when they receive a threat of nuclear attack? What steps occur? Is the person flooded with anger and a desire to lash out (as President Kennedy and his advisors appear to have been after learning that nuclear missiles had been placed in Cuba?) Is the first step strong emotion? If so, what happens next? Does rationality take over? Does rational thought rule the rest of the process? It is well documented that subconscious factors sometimes shape everyday decisions. If there are subliminal processes at work during a crisis, are primitive instinctual reactions like ‘fight or flight” incorporated into the process? How does that happen? Is the process like a conveyor belt — a sequence of considerations? Or a helter-skelter web of ideas, emotions, and instincts all interacting at once?
The powerful emotions that will inevitably be interwoven in the process are important because they introduce an element of uncertainty that often goes unacknowledged in discussions of deterrence. One of the hallmarks of emotion is that it can vary widely and unpredictably. As the Roman poet Catullus wrote before the birth of Christ, “I hate and I love. Why I do this, perhaps you ask. I know not, but I feel it happening and I am tortured.” Emotions well out of some dark part of us that is only dimly understood. This variability undermines certainty at every turn.
I once sat with a distinguished former government official who had dealt with nuclear weapons policy at the highest level, and during our conversation he said, in the offhand manner one uses for something that is beyond question, “Well, we know that [a particular kind of deterrence threat] doesn’t work.” But it seems to me that that is exactly what we don’t know. We don’t know that any kind of deterrence threat will fail. Or work. Even the most unlikely threat has the potential to unexpectedly succeed, just as the most ironclad threat has the potential to unexpectedly fail. Emotions are highly variable, and this means that nuclear deterrence outcomes are also highly variable. A threat that might work one day, might — under exactly the same circumstances — fail the next.
We understand little about the mental process that goes along with nuclear deterrence decision making and most of the attempts to understand this process seriously under-estimate how uncertain the outcome will be.
What we do know
What we do know about nuclear deterrence — the process that something like 4.3 billion people depend on for their lives — is that it cannot last. It seems reasonable that nuclear deterrence ought to work fairly well. After all, a nuclear attack could inflict frightful devastation and any rational leader who imagines such destruction should (one would expect) act cautiously. It is a pretty good bet that on most days nuclear deterrence will work. And nuclear deterrence seems to have been a restraining influence on war for over seventy years. One can make a case that this poorly understood process is generally effective.
That is the good news. The bad news is that we know with certainty that nuclear deterrence is bound to fail one day.
The certainty of that failure is undeniable because the system has a key component that is flawed. A key component with a history of failure. And that component is . . . us. We are the reason that nuclear deterrence cannot work forever. This conclusion is undeniable based on human nature and logic.
First, human beings are fallible. No one is perfect. Second, nuclear deterrence involves human beings. Nuclear deterrence is not an automated process that runs along unattended in a corner. Human beings are involved at every step. We make the threats and other human beings decide how to respond. So, if human beings are prone to folly (and we are) and if nuclear deterrence involves human beings (and it does), then nuclear deterrence is inherently flawed — it will fail. It’s not a question of if. It’s only a question of when.
The eventual endpoint of continuing to rely on nuclear deterrence is catastrophic war. Deterrence will continue to fail periodically. One day, perhaps next month, perhaps next year, perhaps years from now, a crisis will spiral out of control into a nuclear war so devastating that all previous conflicts will pale by comparison.
What still needs to be done
There are steps to be taken to enhance our knowledge and understanding of nuclear deterrence. The first is to put a greater emphasis on psychology and neuroscience. Deterrence is not a military phenomenon nor a physics equation; it is a psychological process. A number of currently unexamined questions could be usefully explored. Is survival always the highest human value? Do communities ever choose extinction? How much can unnoticed emotion impact thinking? Are there limits to rational decision making? How does intense pressure affect rationality?
The second is to examine past failures a little more closely and objectively. A great deal can be learned from failure (which is one reason why, for example, commercial jet crashes are so carefully investigated). There are at least six (arguable) failures of nuclear deterrence that should be examined more closely: the Berlin crisis of 1948, the entry of the Chinese into the Korean War in 1950, the Cuban Missile Crisis of 1962, the Middle East War of 1973, the Falkland Islands War of 1982, and the Gulf War in 1990. Some work has already been done on this subject, but a great deal more is needed.
Finally, there is a rich store of data on deterrence that is rarely used in thinking about nuclear weapons: deterring crime. Critics often say that criminal deterrence cannot be compared to nuclear deterrence. Nuclear weapons are so awesome, they say, so far removed from any other human technology, that no other activity can be compared to nuclear deterrence. Although it is true that the military situation of war has little resemblance to crime, people are involved in both endeavours. The circumstances may be different, but the central actors are the same. Given how little we know about nuclear deterrence, even analogies that are not perfect should be given attention. More work is needed comparing criminal deterrence and nuclear deterrence.
* * *
The people who originally shaped our understanding of nuclear weapons had to build their ideas from a single use in war, whatever indirect evidence was available, and a collection of intuitions and theories. Their ideas, in other words, were derived for the most part not from experience but from assumptions. And since there have been no actual uses of nuclear weapons in the intervening years to fill in this experience gap, the field is still based to a remarkable degree on what amount to hunches.
Looking back over the work that has been done on nuclear deterrence so far, much of it seems to be more about soothing our fears than honestly confronting the difficult realities. Governments that claim they understand nuclear deterrence well enough to manage it safely are not being fully honest. When genuine peril is present, false reassurances are deeply irresponsible.
If that seems like an unnecessarily harsh judgment, consider that reality is hard and unsparing. It forgives no foolish choices and punishes the unwise without mercy. The only way to responsibly handle tools as dangerous as nuclear weapons is to insist on a severe and exacting commitment to reality. If we are to meet this standard in connection with nuclear deterrence, we have a good deal of work to do.
The opinions articulated above represent the views of the author(s) and do not necessarily reflect the position of the European Leadership Network or any of its members. The ELN’s aim is to encourage debates that will help develop Europe’s capacity to address the pressing foreign, defence, and security policy challenges of our time.
Image: Wikimedia commons, Michael