Sorry for the light blogging of late — as usual, I am writing a lot, just not finishing anything.
I notice this very interesting story by Paul Richter in the Los Angeles Times on deliberations over the Nuclear Posture Review. In particular, Richter summarizes the debate over “declaratory” policy — public statements about the role of US nuclear weapons.
The core debate is between those who want to declare that the “sole purpose” of US nuclear weapons is to deter and, if necessary, respond to nuclear attacks. On the other are those who want to limit the role to “existential” threats — whatever that means.
Richter’s summary of the debate is, more or less, also my understanding.
A core issue under debate, officials said, is whether the United States should shed its long-standing ambiguity about whether it would use nuclear weapons in certain circumstances, in hopes that greater specificity would give foreign governments more confidence to make their own decisions on nuclear arms.
Some in the U.S. argue that the administration should assure foreign governments that it won’t use nuclear weapons in reaction to a biological, chemical or conventional attack, but only in a nuclear exchange. Others argue that the United States should promise that it would never use nuclear weapons first, but only in response to a nuclear attack.
Pentagon officials question the value of such public declarations, contending that foreign governments may not even believe them, said the U.S. officials and others.
During the Cold War, Soviet officials declared that they would use nuclear weapons only in response to a nuclear attack. But when Soviet archives were opened, it became clear that “there were scenarios where they would have contemplated first use,” said Charles Ferguson, a former State Department official who now heads the Federation of American Scientists.
The lingering skepticism that resulted could carry over to similar U.S. declarations, limiting their worth, some officials have argued.
A “no-first-use” policy may represent a bigger step than the Obama administration would be willing to take, private analysts said.
Instead, they think the administration might hedge its policy by saying, for instance, that the United States would use nuclear weapons only in situations that threatened its existence.
My view is basically the same as that expressed by Mort Halperin in Survival — an article that draws on work Mort, Arnold Kanter and I have done together at the New America Foundation.
The United States maintains nuclear weapons to deter and, if necessary, respond to nuclear attacks against ourselves, our forces, or our friends and allies.
This is probably the most consequential decision that will come out of the Nuclear Posture Review. It is very important that the President get it right.
The Problem With No First Use
I am temperamentally inclined toward a “no first use” pledge. (I don’t think it would be a huge gain for the United States, though nor do I think it is a huge danger.) But it does suffer from one very specific problem.
As it happens, I don’t think it would ever be in the interest of the United States would initiate the use of nuclear weapons. The late Michael Quinlan, for instance, once said in a meeting that “We do not foresee first use. We do not expect it. We will do everything in our power by our posture to sustain our expectation. But we cannot guarantee” that a situation will not arise that would force us to consider the first use of nuclear weapons.
Sir Michael’s objection, I thought, was quite sensible. Categorical statements are too simplistic for the real world. As a result, others don’t take such pledges seriously. Reassurance must be credible. I often see, in the Chinese case, this particular drawback of a no-first use pledge. Americans and others don’t take it seriously — although I think we should. As a result, Chinese academics and officials often get trapped in silly “what if” games.
Take the case of Chu Shulong, a Chinese academic who ended up in Chinese Military Power for what seems like a relatively innocuous interview:
The Director of Tsinghua University’s Institute of Stratgeic Studies, in an interview with a reporter from Da Gong Bao expressed, China’s promise not to be the first to use nuclear weapons was extremely clear and firm. As of now, their isn’t the slightest indication that China’s government will let go of this promise. ”(I) have not heard any leader on any occasion state China will change or let go of this position. Never.”
[snip]
At the same time Chu Shulong provided a hypothetical, except in the case of a foreign power launching a full scale war against China, using all of their advanced (precision) weaponry except nuclear weapons, and the Chinese nation were facing the danger of extermination, China may let go of this promise. But he considers the possibility not very great.
As a result, Chu Shulong ended up in Chinese Military Power declaring, “China may renounce [no first use] at a time when the country’s fate hangs in the balance.” A very similar thing happened to Sha Zukang regarding Taiwan.
This is a basic problem when statements are categorical — it is too easy for someone to use a “ticking time bomb” scenario (or Martians using non-nuclear lasers to incinerate elementary schools) that twist the speaker up in knots. The Chinese official or academic defending “no first use” has to either admit that, in a hot-blooded moment, that Chinese leaders might not be especially scrupulous about observing past statements or lamely repeat “China undertakes unconditionally not to use or threaten to use…”
Neither is very appealing.
I’ve had several Chinese participants tell me about a recent Track II meeting in Beijing where they explained China’s categorical no-first use pledge. The American participants, to make the classic point, rather clumsily suggested a hypothetical US conventional attack on China’s nuclear forces.
The Chinese participants freaked. [Perhaps I should say, “were disturbed.”]
The Americans went home satisfied that the Chinese weren’t very serious about no-first use; the Chinese left thinking they had been subjected to a very serious threat of coercion. And perhaps wondering if they should start planning for first-use scenarios. I am repeatedly asked about this interaction and was again during my last trip to Beijing. This particular Track II debacle is going to haunt the US-China nuclear dialogue for years.
I happen to agree with not using nuclear weapons first, but as a declaratory policy it does suffer from the problem that Sir Michael identified.
The Problem with Calculated Ambiguity
On the other hand, it isn’t like I think our current policy of calculated ambiguity is a god’s gift to declaratory policy. For one thing, the US policy isn’t actually an instance of ambiguity — in the sense of statement that could mean different things — but rather the current policy consists of two of logically inconsistent statements.
Our current policy is incoherent, not ambiguous.
As I noted this summer, “calculated ambiguity” often devolves into clumsy brandishing of nuclear weapons which is anything but ambiguous:
Particularly compelling, to me, is [Scott Sagan’s] demolition of “calculated ambiguity” using a case study from the Bush Administration. Whatever the appeal in theory, in real life “calculated ambiguity” degenerates into the clumsy brandishing of nuclear weapons.
I had been mulling a similar case study in a memo using the almost comical efforts to maintain “calculated ambiguity” regarding a possible nuclear strike against the Libyan facility at Tarhuna. The Clinton Administration looked like the Keystone Cops armed with nuclear weapons, which says something when you manage to upstage Muammar al-Gaddafi in a black comedy.
No need to, now. Scott executes a much cleaner demonstration of why “calculated ambiguity” is, to my mind, more trouble than it is worth …
Scott’s case study involved then-President Bush’s April 2006 statement that “All options are on the table” with regard to Iran, which was widely interpreted as a nuclear threat and ended with the UK Foreign Minister Jack Straw calling the idea of a nuclear strike against Iran “completely nuts.”
This is also a declaratory policy fail.
A More Sensible Declaratory Policy
These problems suggest two criteria for a declaratory policy — it must clearly articulate a limited role for nuclear weapons, but those limits should not be categorical. They really reflect a single observation: Declaratory policy must be, in part, designed for ease of execution. If it is too complicated, it may be paraphrased. If it is too simple, it may be shown to be foolish.
One option is to find a felicitous phrase — “last resort” or “existential threats” — that conveys the sense of a commitment, without actually making one. That is the essence of the ambiguity approach. (I happen to think that either formulation would improve on the current approach.)
An alternative — which I prefer — is to talk not about when the United States would use nuclear weapons at all, but rather why we have them.
If you think about it, we make decisions every day about our forces, policies and posture that need to be explained. We haven’t dropped the big one since 1945. So, our declaratory policy ought to defend the actions we do take, not ones we might. As it happens, I believe it is a true statement that the United States maintains nuclear weapons, either largely or exclusively, to deal with nuclear threats.
As for talking about nuclear use scenarios — well, the only way to win is not play! Look, you can always come up with an artificial, hypothetical that would compel the first use of a nuclear weapon against a kindergarten. (The kindergarten is sitting on top of a deeply buried bunker containing the Andromeda strain and there isn’t enough time to evacuate…)
No good can come of speculating on such hypothetical scenarios because the deck is stacked against you. Moreover, there is no reason to play these games, because such unlikely scenarios are irrelevant to our nuclear policy, force structure or posture. The very fact of having nuclear weapons, no matter what the President says, provides the appropriate measure of deterrence against these sort of unlikely speculations.
As a result, I tend to think talking about why we have nuclear weapons is a better approach than trying to find a phrase, such as “existential threats,” that explains when the President might use nuclear weapons. The “existential threats” formulation, in particular, will baffle foreign audiences, who in turn will ask what precisely threatens the existence of the United States and others. This discussion can only go badly. For example, are we saying we would forswear nuclear weapons in the event of a limited nuclear attack that didn’t threaten the existence of the United States? Would a North Korean biological attack threaten the existence of Japan? No good can come of answering such questions, yet declining eliminates much of the advantage in making the commitment in the first place.
It’s much better to state that neither of these cases — nor the truly weird cases like asteroids — have anything to do with why the United the States maintains a nuclear deterrent. United States nuclear policies, forces and posture are not shaped by the need to deter biological weapons or deflect asteroids. That’s the implicit meaning of the President’s statement that the United States seeks the security of a world without nuclear weapons: That all plausible non-nuclear threats can be met with conventional forces.
I happen to think that this is statement of fact about US policy, even if it is not the formal declaratory policy, and that the President should say so:
The United States maintains nuclear weapons to deter and, if necessary, to respond to nuclear attacks against ourselves, our forces, or our friends and allies.
Such a statement, by subtly changing the nature of the debate from use to possession, requires a certain amount of discipline. The President and his advisers would have to refuse to elaborate on its meaning, characterizing the statement as a “statement of fact” about why the United States maintains nuclear weapons. They would need to refuse to engage in “irresponsible speculation about hypothetical scenarios” involving the possible use of nuclear weapons. (You can add, “that have nothing to do with why the United States maintains nuclear weapons, which is…” if you really want to.)
The lone exception to the prohibition of discussing the “use” of nuclear weapons would be the existing negative security assurance — which would be politically calamitous to withdraw. In that case, I would revise to remove the “Warsaw Pact exclusion clause” on the grounds that there is no Warsaw Pact. It might read something like:
The United States will not use nuclear weapons against non-nuclear-weapon state-parties to the Treaty on the Nonproliferation of Nuclear Weapons that are in compliance with their nuclear nonproliferation obligations.
And since there are two nuclear weapons states — Russia and China — that the United States needs to address, I would make one last statement, drawn from the recent CFR Task Force co-chaired by Bill Perry and Brent Scowcroft.
With regard to Russia and China, mutual vulnerability (or deterrence) is not a policy choice, but a fact to be managed with strategic stability.
Taken together, I think these three statements would accurately convey the limited role that nuclear weapons play in US security in a manner that would neither alarm allies or comfort potential adversaries.
Update | 10:04 am January 6: Sorry about the truly staggering number of spelling slip ups. I was in a rush. Should be better now.
Paul Nitze wrote long ago that there are two types of nuclear policy (a) declaratory policy which is for “political and psychological effects” and (b) actual policy which is the actual war plans that will govern conduct in the event of hostility. For Nitze (a) is not real, what is real is (b).
The purpose of good declaratory policy from an arms control perspective is to constrain what the war planners can and cannot do. If it is too ambiguous then it won’t really do that. What really matters is not what external states think so much as what internal actors can do.
I personally believe that the debate on declaratory policy doesn’t matter so long as the disconnect between (a) and (b) obtains. This disconnect is at its most pronounced when the Democrats occupy the White House. The good thing about the hawkish GOP linked nuclear strategists is that their declaratory policy matches their actual policy. We should at least respect them for their intellectual honesty.
The most important part in the LA Times article is the bit where it states that Obama needs something to give him credibility for the 2010 NPT Rev Con. That is he needs a declaratory policy that is tailored to desired “political and psychological effects”. Meanwhile (b) won’t change much.
The US President Obama has decided to have a nuclear free world. That is a policy decision. Will the world be a better world if it is devoid of nuclear weapons? Ofcourse it will be. So…Nobody wants such a wonderful decision to be revoked.
Out of this policy flows the purpose of having nuclear weapons. That should be only to deter and, if necessary, respond to nuclear attacks.
Once the purpose is absolutely clear then it becomes very easy to declare a “No First Use” policy.
Once such a great confidence building “No First Use” policy is enunciated by the US government, the faith and respect for the US will increase and more countries will follow suit.
OTOH, let us try to answer a few Qs.
Will any kind of mere maintaining or posturing of nuclear weapons deter Iran from pursuing its nuclear weapons program?
Will China’s naval ambitions be subdued by having a presence of nuclear tipped submarines around Japan?
Will Al-Qaeda or Taliban whether of Afghanistan or Pakistan be deterred from trying to acquire nuclear weapons from religious zealots like A Q Khan?
Will Pakistan stop proliferating for technological or monitory gains?
If the answers to these Qs are vehemently NO, then its time US declared a clear “No First Use” policy.
Re: the Soviets and first use
Consider for a moment two real examples of Soviet thinking which have come to light since the end of the Cold War. Both concern Europe, where first-use was considered possible under the NATO doctrine of flexible response, to use nuclear weapons to repel an invasion from the larger Warsaw Pact conventional forces.
These two examples are from Soviet documents that were never intended to see the light of day, which makes them more interesting.
1.
In 1977, the Warsaw Pact carried out a massive war game in East Germany, known as “Zapad,” according to Marshal Nikolai Ogarkov’s report on the May 30-June 9 maneuvers. Ogarkov’s report is reprinted in the volume, “A Cardboard Castle: An Inside History of the Warsaw Pact, 1955-1991,” edited by Vojtech Mastny and Malcolm Byrne (National Security Archive/CEU, 2005) pp. 406-412.
To summarize, “Zapad” envisioned a situation in which the NATO conventional forces were reeling, on the verge of defeat, and began to consider a first strike with nuclear weapons.
In the exercise, the Warsaw Pact forces went into overdrive to pre-empt. The plan was to weaken the NATO forces as they were getting ready to launch the first nuclear strike. The Warsaw Pact forces were to attack the NATO nuclear forces with conventional means, to disable and weaken them, as well as go first with nukes, including about 300 weapons on the Second Front.
Not everything went according to plan. Ogarkov lamented that on the Second Front, “Many [NATO] targets to be attacked were mobile, and their positions changed constantly.” The Warsaw Pact attackers stumbled in moving their nuclear weapons to hit the targets. There was a “tremendous amount of work in a short period of time” and the Soviets weren’t up to it, and lacked the computer automation, Ogarkov admitted. They also failed to disable the NATO weapons with conventional arms.
Seeing that NATO was minutes away from a nuclear attack, the Eastern forces launched a first strike with nuclear weapons. The West fired back. Both suffered huge losses in weapons and people, Ogarkov wrote, and “along the front lines, huge areas of contamination, destruction and fires developed.”
In the end, reporting on the exercise, Ogarkov said that Warsaw Pact forces “lost almost twice as many divisions” as NATO. The reason was they had failed during the NATO preparations to disable or weaken the West’s nuclear weapons. So a first strike as pre-emption didn’t gain much in the confusion and pressure of the moment.
2.
Twelve years later, in 1989, Gorbachev was watching a debate in Western Europe about whether to modernize NATO’s nuclear weapons. The United States was pushing allies to modernize the 88 short-range Lance nuclear-tipped missiles in Europe. After implementation of the 1987 INF treaty, these shorter-range missiles would be among the remaining battlefield nuclear weapons available to NATO against a possible Soviet conventional attack. (There were also thousands of other weapons on bombers.) West Germany was balking at modernization, since use of the Lance missiles in war would quite probably be on its soil.
When Secretary of State James A Baker III visited Moscow on May 10, Shevardnadze told him Gorbachev was eager to eliminate the whole class of tactical, or short-range, nuclear weapons in Europe. “Do not dodge” the issue, Shevardnadze warned Baker. I tell this story in my book, The Dead Hand.
A day later, Gorbachev announced he was unilaterally withdrawing 500 tactical nuclear warheads from Eastern Europe, and promised even more if the United States would take similar steps. But Baker brushed off the proposal as a political ploy, thinking Gorbachev was simply trying to split the alliance.
What is interesting is the staff work behind Gorbachev’s proposal. Vitaly Katayev, a Central Committee staffer who was deeply involved in these arms control issues, prepared a memo in this period about tactical nuclear weapons, which I found in his files.
The memo shows that the Soviets were, indeed, trying to manipulate the politics of Lance modernization in Western Europe. But it also shows that Gorbachev’s offer was genuine. The Kremlin arms control specialists, Katayev said, “believe that short-range land-based nuclear weapons are the most … dangerous for all countries in the deterrence arsenal.”
Why? The Warsaw Pact short-range nuclear weapons were less than 100 kilometers of the frontier. NATO’s fastest tanks could easily catch up with and overrun them in 40-50 minutes, he said. “The enemy is able to move close to it quickly or destroy it using high precision means,” he wrote.
Moreover, in making a first-use decision, Katayev speculated, the West’s options weren’t very good: either launch quickly and start a nuclear war on the enemy’s territory, or detonate the nuclear weapons on its own territory, or just surrender the missiles to capture or destruction by the enemy. “Which variant is preferable?” he asked.
And then he added:
“The main issue when debating the use of nuclear weapons is the gain one gets when taking such a catastrophic step. Any nuclear attack is followed by retaliation. Nobody can say to what extent the exchange of nuclear strikes can go on after the first nuclear blow.”
The Soviet Union could respond with strategic missiles aimed at the United States if the U.S. delivered a tactical strike on Soviet troops in Europe, he said.
“Is it a good decision for a state to solve some small tactical problems on a local battlefield and on another continent by suicide, and turning the most developed part of mankind to ashes?”
Ogarkov and Katayev saw the same thing – “first use” on the European battlefield was probably a path to disaster. It was not a sure-fire pre-emption strategy. It posed all kinds of excruciating decisions in the chaos of a real battlefield.
Maybe there is still value in ambiguity about first use – at least in declaratory nuclear policy. But these Soviet documents – and they are only a small fragment of the archives – suggest that some on the other side in the Cold War took a hard look at the reality of nuclear first use, and came to good and sober conclusions about the folly.
Refreshing commentary.
A very interesting and useful discussion.
Is a new definition (“The United States maintains nuclear weapons to deter and, if necessary, to respond to nuclear attacks against ourselves, our forces, or our friends and allies.”), though, just another clever way to articulate ‘calculated ambiguity’? I’m not quite sure I got the overall point here.
Where are discussions at on CBMs with Russia and China, such as de-alerting? This would seem to be a useful step, along with substantial post-START US/Russia warhead reductions, that would add real teeth to the idea that nuclear weapons play a limited role in US security.
Good, thoughtful post.
MarkoB wrote: “The purpose of good declaratory policy from an arms control perspective is to constrain what the war planners can and cannot do. If it is too ambiguous then it won’t really do that. What really matters is not what external states think so much as what internal actors can do.”
I am unaware of any link between U.S. declaratory policy on nuclear use and U.S. operational plans for nuclear use, particularly with respect to NFU. The operational plans determine which weapons are placed on which targets. They launch the plans when the President orders them to do so. The plans, to the best of my recollection, don’t contain a reference to timing, i.e. they don’t say do “a” if we are shooting first and “b” if we are shooting second. They say do “a” to destroy this particular target set. Its up to the President to decide if we shoot first or second.
The internal actors plan to achieve military objectives using a range of weapons; the President decides if and when the military will implement its plans. If anyone is constrained by declaratory policy, it would be the President, not the planners. And, to paraphrase Sir Michael, NFU only constrains the President when we don’t need to shoot first. If we really needed to do it (perhaps “under dire circumstances”), the President would not be constratined by a State Department press release.
David Hoffman’s points are well taken. Only a lazy examination of the first WP archives in the early 1990s could lead to the conclusion that “the Soviets lied about NFU”. Back in 1994, I argued in my Ph.D. dissertation that the NATO strategy ran the risk of working too well: the WP would have been persuaded that NATO would start using nuclear weapons after a few days… and would have massively used nuclear weapons.
As David reports, we now know that they realized this would have been a folly.
On the substance of Jeff’s thoughtful post, I think that he does not spend enough time discussing “what deters” – i.e., “what declaratory policy has the best chance of being useful to our security”. Hypothetical scenarios (“would we respond with NW if…”) miss the point; the focus should be on how to avoid the aggression in the first place.
And, frankly, Scott Sagan’s arguments against calculated ambiguity sound pretty hollow to me.
Now you’ve given away your identity, M. Scud!
I tend to think, as you know, that the existence of the weapons does the heavy lifting of deterrence. As long as one doesn’t explicitly rule out specific scenarios (and probably not even then), you preserve the lion’s share of the deterrent. (Indeed, I suspect threatening nuclear weapons against chemical attacks undermines the credibility of conventional options faster than it adds to credibility of nuclear use. I think we should always assert the existence of conventional options for non-nuclear threats.)
I have no objection to various formulae on nuclear use, whether it is the NATO claim that “the circumstances in which they might have to contemplate any use of nuclear weapons are extremely remote” or the delineation of “extreme circumstances of self-defense” — though the latter comes to close to the ICJ language for my taste. (One option, obviously, is to replace “survival at stake” with some other formulation, but I think that just draws attention to difference.
I don’t think, however, that the extreme circumstances in question should be specified, which raises the question of how to get some reassurance, as well as deterrence, from our declaratory policy. Which is why I favor explaining that the “extreme” or “remote” situation for which we “maintain” nuclear weapons is a nuclear attack.
Jeff, thanks for the useful development of your thinking. A lot of what you say makes an awful lot of sense. The question that keeps nagging me is the following: as time passes and self-restraint on the use of nuclear weapons (quite a healthy thing in itself) becomes increasingly entrenched, further declaratory restrictions may end up seriously eroding the very credibility of the threat of nuclear response in any circumstance, and instillate the idea that “we would never dare”. In sum, I believe that the risk may be exactly the opposite of what Scott and you seem to fear when mentioning the “all options are on the table” episod. But we’ll discuss that some day around a glass of grappa.
The general sentiment is dead-on, but I disagree that “the United States maintains nuclear weapons, either largely or exclusively, to deal with nuclear threats”. The most obvious exception, IMHO, is chemical and biological threats. I believe that when the United States dismantled its CBW arsenal, it was with at least an implicit understanding that large-scale use of such weapons against the United States or her allies might merit a nuclear response.
And I think that’s still the case. Clearly the preferred option in that scenario would be precision conventional attacks on the enemy’s CBW facilities, and a lot of work has gone into making that at least sometimes feasible, but it’s not a complete solution. Particularly when the problem isn’t just responding to actual CBW attacks, but reassuring allies who think they might be on the receiving end of such and are wondering if maybe they should be building their own nukes should the US withdraw that part of the umbrella.
So, the obvious fix is to replace your “respond to nuclear attacks” with “respond to WMD attacks”. Except that in other contexts we’ve reduced “weapons of mass destruction” to incoherent ranting on the same level as “existential threats”, so that doesn’t work.
I’m not sure what does. This is hard.
@John Schilling-
I’m sure that Jeffrey has thought about whether to say “respond to WMD attacks” or “nuclear” only. His careful balance is not to assert that the US would never use nuclear weapons first, and one scenario in which one might imagine a nuclear response is a massive attack by chem, bio, or some other hypothetical WMD, if it were unambiguously attributable to a state actor presenting an appropriate nuclear target set. But asserting that the US retains nuclear weapons for replying to “WMD” attacks would seem to imply a threat that in the event of a chem or bio attack the US would respond with nukes. That would make it easy for an intemperate president to use nukes in a situation where it might be better not to, and it might even create political pressure for her to do so, as well as underminining deterrence should she fail to. Remember that the spectrum of possible attacks with chemical or biological weapons is much broader than the nuclear spectrum. What should the US do in the event, for example, of a terrorist anthrax attack that kills a few hundred or a few thousand Americans and that can be traced to some jerk known to be hanging out in some country under the protection of its local regime? Drop a nuke? And risk opening the Gates of Hell, so to speak? Would you want to create such an expectation, even as a deterrent? Think terrorists are deterrable, after all?
The nuclear free world is an illusion. And no first strike is an invitation for conventional salami tactics.
But what probably should be done is a declaration of a policy of no first use of strategic nuclear weapons.
It seems to me that the problem with NFU – getting tangled up in responding to silly hypotheticals – is simply to refuse to engage with silly hypotheticals.
The correct response to the American suggestion at the Track II meeting would have been to respond to a hypothetical with another hypothetical – could the US ever put enough stock in the Chinese NFU pledge to attempt such a thing?
This illustrates the advantage of the NFU pledge – sure, you might say “ultimately we can’t trust that to hold in a severe crisis” – but the doubt in this case is sown in the mind of the potential aggressor, not the defender. It deters aggressors while not intimidating non-aggressors.
Having thought further about this, I’m not completely convinced that characterizing Jeffrey’s position as shifting from talking about actions to intentions is entirely right. There’s something fishy here though, but I can’t put my finger on it.
I guess the mark of a good idea is that it makes you think.