Sonntag, Juli 24, 2005

Thinking the uncomfortable...

This post got me annoyed, and I'm gonna take time off getting something productive done in order to try and pound some sense into this discussion.

So sit back and get a cuppa.

The writer, Jeffrey Lewis, should know better:

Jeffrey Lewis is a Research Fellow at the Center for International and Security Studies at the University of Maryland School of Public Policy (CISSM). His research focuses on the space policy component of CISSM’s Advanced Methods of Cooperative Security Program, which is generously funded by the John D. and Catherine T. MacArthur Foundation. Dr. Lewis is also member of the Editorial Advisory Board of the Bulletin of the Atomic Scientists.

Dr. Lewis received his PhD in Policy Studies (International Security and Economic Policy) from the University of Maryland and his BA in Philosophy and Political Science from Augustana College in Rock Island, IL

Now, I don't have his academic background - I stopped with a MA in political science, economics and philosophy - but I spent a lot of time on Hermann Kahn and his works.

The above linked article includes this:

Most discussions about target sets leave the impression that the decision to use a nuclear weapon here or there is a deeply rational business, with great care taken not just in the selection of each target, but also to ensure each nuclear weapon is really necessary. After all, if we are going to put a nuclear weapon on a tank factory sitting next to a grade school, you'd think that someone made a careful, anguished decision about the lesser of two evils in a morally ambiguous world.

You might think that, but you'd be wrong.

And that is where I stopped and said: gotta fisk this one.

First, the primary assumption is incorrect. Target sets have nothing to do with deciding to use a nuclear weapon: rather, a decision to use a nuclear weapon is first and fundamentally a political decision that has nothing to do with target sets. Target sets and SIOP are all about the military implementing a political decision, to use nuclear weapons at all. Target sets and SIOP can allow a politician making such a decision to achieve the goals of such a political decision with a minimum of force and destruction, if the decision is to, for instance, destroy part of the military infrastructure of an adversary.

Hence if the US were to "put" a nuclear weapon on a tank factory, that decision is not made after a careful, anguished decision of whether there is a school next door or not: that decision is made after a careful, anguished decision of the necessity for using nuclear weapons at all.

Hence Jeffrey Lewis is wrong: the decision to use nuclear weapons isn't fundamentally a rational decision: it's a political decision. Now the nitty-gritty of choosing target x under criteria y is one of deciding to, for instance, use a low-yield surface burst in order to destroy utterly a large tank-making complex, rather than using a medium-yield air burst that would do the same job but also destroy a significantly larger area. But the decision here is: eliminate target x under a set of operating criteria.

Not: hand-wringing "gotta get the tank factory, but oh, no, there's a school next door. What should we do????"

For example, of the 12,500 targets in the SIOP at that time, one of them was slated to be hit by 69 consecutive nuclear weapons. It seems superfluous to say that this is crazy, but it is important to understand how the planning process could result in such a figure. At the level of a presidential directive, a document of a thousand words or so, you will have the reasonable-sounding requirement--if you're thinking about war-fighting at all--to, say, target the political and military leadership. That guidance goes to the Office of the Secretary of Defense, which in a 15- or 20-page document called a NUWEP (for "nuclear weapons employment policy") adds some detail: for example, what sorts of leadership facilities should be targeted. The NUWEP then goes to the Joint Strategic Target Planning Staff of the Joint Chiefs of Staff, which in hundreds of pages of a document called Annex C to the Joint Strategic Capabilities Plan lists specific facilities to be struck and damage requirements to be met. Annex C then goes to STRATCOM, where the targetting staff figures out which weapons, and how many, to apply to each target to meet the required level of damage.


When I mentioned Butler's 69 weapons to Dr. Bruce Blair, a former Minuteman missileer and acknowledged expert on the operational aspects of nuclear warfighting now at the Brookings Institution, he found in his notes a statement by a high official at SAC in the late 1980s that the highest kill probability for the United States' best weapon against deeply buried, sprawling, hardened command posts was less than 5% (how they calculate this is a whole other matter, but the short answer is, they guess). Blair got out a calculator, assumed a kill probability of 4% for one weapon, and started multiplying. To attain a 50% confidence in destroying the target required 17 weapons. When Blair got up to 69 weapons, the "kill probability" had reached 94%.

Yep. That's why you target like you do. Target 69 weapons if you want to have a 94% kill probability of a hardened target under those conditions. Period.

It's part of the calculus of deterrence. The Soviets were really hot on scientific planning of war-making capabilities, and using a complex system of correlation of forces to determine when their war-making capabilities were such that they could be guaranteed crushing success. It's important to understand that "guaranteed" thing: if the Soviets thought they could obtain success via the battlefield, that gave them a psychological edge in their confrontational politics.

And that's why so much of SIOP and various missile-defense systems and planning is strictly targeted at ensuring that the enemy can't do what they want to do: get inside of their loop and you've won the game. Not the wham-bam-thank-you-ma'am knock-down exchange of nukes, but rather of making sure the Soviets never thought they could obtain success via the battlefield. That way you don't have to fight, since the confrontational politics are never implemented.

But here is where I really got annoyed:

The real issue here is that organizations abstract reality to manage it. That abstraction, James Scott pointed out in his book, Seeing Like A State, can produce disasterous consequences such as Soviet collectivization and the Maoist Great Leap Forward.

Here he is jumping to a conclusion: that abstracting reality in order to manage it is comparable to such abominations of human behavior as Soviet collectivization and the Maoist Great Leap.

But more fundamentally, organizations aren't those that abstract reality in order to manage it: people do this. It's how we all work. Reality, that amorphous cacaphony of perceptions and demands, is managed by each and every one of us when we interface with the world. If you can't abstract the real world, how are you supposed to understand it at all?

Abstract reality to manage it. The mere phrase sets me on edge, because it sounds like this is a bad thing: after all, he goes directly to comparing it with the Soviets and Maoists.

Most of us intuitively understand the inhumanity of bureaucracies - a perhaps necessary evil in the modern world. This understanding is why General Butler's narrative is so compelling -- a human being acheives a vantage point from which to survey the madness of an inhuman organization. It's Kafka and Joseph Heller in equal measures.

Bureaucracies aren't inhuman: they are, after all, human organizations. What gets people riled up about bureacracies is that they, the bureaucracies, tend to control access to things that people want. Hence bureaucracies are the necessary logical result of political decisions, for instance, to give people welfare benefits: the politicians want benefits to be under control.

But it's not a madness of an inhuman organization.

Only an organization would target 69 nuclear weapons on a single facility (later revealed to be the Sofrino missile defense radar) outside of Moscow in a strike designed to minimize "collateral damage". To take another example, STRATCOM calculates only blast damage from nuclear weapons. STRATCOM does not calculate the damage from any fires that would be ignited, even though such fires would be far more damaging than any blast effects. Why? Because fire damage is hard to calculate and, therefore, not real.

Now this is where the fundamental error is: targetting 69 nukes on a missile defense radar does minimize collective damage if the alternative is having to target 300 to achieve the same level of guaranteed destruction in the targets the missile defense radar is tasked with defending.

And STRATCOM doesn't calculate other damages not "because fire damage is hard to calculate and, therefore, not real."

It doesn't because calculating fire damage is fundamentally unknowable, given the fact that it makes a huge difference if the attack is in summer in the middle of a drought or in the middle of winter in a blizzard with 3 meters of snow on the ground after two weeks of sub-zero weather. But that doesn't mean it's not real: it just means that you can't take those factors into account in planning for a strike that might happen at any given time of day.

Further, secondary blast effects are very much real, but not to the target in question: a hardened, deep bunkered target ignores surface fires. Only blast and radiation affect such a target.

Again, secondary effects are ignored because they don't change the cold, cold equations of ensuring that the target is destroyed. The link takes you to an analysis of how fire damage can be calculate, but he's begging the question here: the analysis is for a soft target, not a hard one.

Although the details are classified, the contact website makes clear that the ISPAN doesn't change how STRATCOM does business. ISPAN does not address the fundamental myopia of "kitchen sink" target sets, artificial damage expectencies and rigid delivery schedules that encourage the President to use nuclear weapons before an adversary has time to take protective measures.

That's one reason to be worried about efforts by the OVP to plan to strike Iran -- not because there has been a policy decision to execute the plan (there has not), but because nuclear war planning continues to define the President's options in ways that alienate him from the execution.

Sigh. Where to start?

First: the SIOP offers the President a widely nuanced range in the practice of what Tom Lehrer so beautifully called "escalatio" in his song "Who's Next?" It's not a "kitchen sink" of target sets, but if the President decides in a conflict to destroy, for instance, the ability of an adversary to make rocket fuel, that opportunity needs to be available *now* and not after 6 weeks of planning.

Artificial damage expectancies? What SIOP and military planning does is plan on not having a theoretical, ideal performance of weapons and systems, but rather of accepting reality, that they won't always work. That's why you see two conventional bombs hit a target instead of one when the military really, really, really wants to have the target destroyed. The idea that it's not "real" if damage assessment doesn't take into account secondary effects is specious when the planning calls for ensuring a minimum of assured destruction: it's only real if you can be really, really sure that you can do what you claim.

Rigid delivery schedules? Humans are always in the loop: you can stop at any time. The military isn't full of yahoos who simply obey orders. This isn't the Franco-Prussian War or WW1, where the Kaiser was told he couldn't stop the war any more.

And for the very last: SIOP, if not constantly updated and rethought, does fail if it doesn't offer politicians a nuanced, controlled usage. But that's not got anything to do with alienating any President from deciding to use nuclear weapons.

Again: using nukes is first and most fundamentally a political decision. The use of military force is always an expression of political will. The President shouldn't be involved in the technical details of whether to use 45 or 12 weapons, but rather makes the decision to eliminate the enemy's war-making capabilities.

And if the SIOP fails to give him this option, then it's failed in its mission.

I've gone on here for longer than I planned to. I don't want to rag on the man, but this was sloppy and shallow.

And I'll reiterate: using nuclear weapons is a political decision. SIOP serves the President, not the other way around.

And it is one of the fundamentals of our age: if you don't want nuclear weapons to be used, then make sure that everyone involved understands what their use means. If North Korea threatens to use them, make sure that the North Korean leadership knows that one nuke from them means three hundred on them and the complete dismembering of the entire North Korean military apparatus in a hail of complete and utter devestation.

And this is the problem with countries like Iran developing nuclear weapons: that you introduce non-rationality into the nuclear calculus, that an obsession with the destruction of Israel means that deterrence will no longer work. That's when the world becomes very, very dangerous: if you have people who are deeply, completely convinced that Israel must be eliminated, that Israel and the US are indeed direct manifestations of Satan, and you allow them to obtain nukes, then it's gonna be hard to keep them deterred, since they prove day after day that they love death and its transformation to martyrdom more than the lives of their children.


Jay Denari hat gesagt…

Hi, John,

Target sets have nothing to do with deciding to use a nuclear weapon: rather, a decision to use a nuclear weapon is first and fundamentally a political decision that has nothing to do with target sets.

What Dr. Lewis is pointing out is that such a decision SHOULD include agonizing over the effect on the people (esp. civilians) near the target. Taking into account the effects on people is the definition of a "political" decision, and refusing or being unable to do that may not be "inhuman," but it's certainly inhumane.

These devices are not and cannot be precision weapons whereby we can take out that tank factory with minimal civilian deaths. If a nuke goes off, it's all but guaranteed to vaporize that factory AND the school AND obliterate everything else within several square miles, minimum, with the VAST MAJORITY of the dead being civilians.

When nukes are in play, the expression a minimum of force and destruction is simply a denial of reality. Minimum destruction would be an attack that destroyed the tank factory without touching the school or any nearby homes, something that is impossible with nukes.

Target 69 weapons if you want to have a 94% kill probability of a hardened target under those conditions. Period.

What exactly is the point of such massive overkill? Doing that only serves to turn a wide swath of land into black glass forever and kicks up massive quantities of fallout that poisons millions downwind. That's not "minimal" destruction by any sane definition of the term.

Even if the target is technically not destroyed, it only takes one or two explosions to ensure that nobody will be using the place for the foreseeable future.... b/c anyone who tries will cook themselves getting there (or getting out). Survivors inside will be effectively trapped, probably until they starve or suffocate.

A perfect example is the US's own NORAD HQ -- The Russians or anyone else wouldn't need to pummel it into dust (which might indeed take dozens of warheads); they just have to drop one or two beside the main entrance and nobody inside is getting out. NORAD is, de facto, dead.

Of course, so is most of Colorado Springs, which goes right back to my original point. There is no such thing as a "nuanced" use of nuclear weapons. They are designed to destroy indiscriminately and to inflict as much terror as possible.

You're right when you say that everyone abstracts various aspects of life and decision-making to some degree, and it is almost necessary when comprehending geopolitical issues. But the level of abstraction and demonization necessary to actually use nuclear weapons is pathological given what we already know about the deaths, destruction and long-term suffering "small" nukes caused in Hiroshima and Nagasaki.

**This response cross-posted to my blog. **

John F. Opie hat gesagt…

Hi Jay -

Nice post, thanks. Sorry to be so late responding, was in London for 9 days.

But I must disagree.

What you really are trying to argue, I think, is that nukes are terrible weapons. No argument there: but you are making the common mistake of demonizing them.

I'll repeat: the decision to use nuclear weapons is a political decision. Everything else is derivative.

There is no massive overkill of targeting 69 weapons IF the potential enemy has an ABM system in place that may well knock down the first 20 missiles, or if the target is so hardened that nothing less than a CEP of 0 meters will take it out.

But you're missing another point: you don't target with 69 weapons arriving in some sort of blind automatism, but rather sequentially until your target is finally destroyed: that is the purpose, to a large extent, of having satellites up there to do damage assessment. Nukes are too valuable to simply expend them.

And the point of destroying a technical target is not to ensure that no one will use the place for the foreseeable future, but rather that no one can use it **now**: many targets are time-sensitive. If I am targetting a heavily defended target, then I will attack first the defenses. When the US went into Iraq in GW I, helicopters were used to take out major air defense radars right before the air attacks began. No point in taking them out after the attacks have taken place and they have been used.

And blocking the entrance to NORAD doesn't mean it's dead! Multiple independent and redundant communications systems - just what do *you* think the design of the Internet was for??? - means that even if the folks at NORAD can't get out, they can still run the show.

That means that NORAD **isn't** dead. That is why NORAD is buried the way it is, in the heart of a mountain and designed to survive such a blast, even in front of the doors (which aren't where you think they may be): it would take a direct hit with a massive multi-megaton weapon to directly remove NORAD.

And NORAD doesn't have to survive long: it just needs to survive long enough to get the orders out. Those orders might be simple enough as a 4-digit code to trigger the appropriate response: if a potential attacker *really* wants to cripple any US response to a nuclear attack, they *have* to try and take out NORAD in order to cripple any response.

And I would severely disagree with your stance of there being no nuanced use of nuclear weapons: are you saying that there can only be a nuclear spasm?

And nukes are not designed to destroy indiscriminately and inflict as much terror as possible: to do that, you need to do things like putting cobalt blankets around your warhead to make it as dirty as possible, among other things.

And what you're trying to say, ultimately, is that any decision to use nuclear weapons in pathological. That is too simplistic: the use of any weapon could then be said to be pathological, which leads to extreme pacificism and an abondonment of the idea that anything is worth defending.

Which is nothing less than the path that leads to abject appeasement.

I think you have no recourse except to enter the world of geo-military-political calculus, of laying out the choices and making the decisions based on abstractions.

Jay Denari hat gesagt…

Thanks for your response, John, but we clearly disagree on a lot of things here.

First off, although I certainly prefer peace to war and think the vast majority of problems can be solved by serious diplomacy, I do believe there are certain circumstances where it's necessary to fight.

WW2, for example, was a just war, and I can even accept our use of nukes on Japan, if only on the grounds that we can't change it now and we didn't really know what their full effects would be.

But now that we do, there's NO justification for using such weapons. The costs are far too great, in both the short and long term. That's not demonization or "appeasement," it's concern for the survival of fellow human beings and hope for a long-term future. Your "calculus" that my opposition to nukes is pacifistic and "an abondonment of the idea that anything is worth defending" is completely untrue.

The very existence of nuclear weapons is itself a source of terror for many people, and dropping one -- even without the extra "dirty" fallout-producing material like cobalt -- IS indiscriminate. Sure, planners can choose to use the Bomb vs. a conventional bomb on target X for any number of reasons -- it's "discriminate" in that sense. But the effects are not discriminate at all, and those are what will determine who the survivors are, if any.

Let's assume the targeters are 100% accurate: ground zero is in the dead center of the tank factory. Using a laser-guided conventional missile, a typical tank shell, or similar weapon in such an attack will have a far more limited radius of destruction than will a nuke. Any of the conventional weapons certainly will cause some shrapnel/debris that might injure civilians & damage civilian buildings nearby. But their potential for "collateral damage" (I hate that term, its purpose is to allow military folks to distance themselves from the mass murder of civilians) is substantially less than a nuke's. Even a Hiroshima-sized nuclear warhead (a very small one by modern standards) WILL, without doubt, destroy nearly everything for several square miles, regardless of whether it's the factory, the school, hospitals, everyday businesses, homes, bridges, etc. The neighboring school stands no chance of survival with a nuke, but it does with a conventional bomb. How can that effect be considered "nuanced?"

When nukes go off, their thermal pulse, blast wave, and subsequent radiation including fallout have much broader & longer-lasting effects than does a conventional high explosive. Fallout is especially random & pernicious: it can go anywhere the wind goes, endangering civilians in the enemy nation, our own troops, and even citizens of completely neutral countries for extended periods of time, often without the affected people even knowing a bomb has been dropped. Likewise, radioactive particles last for years and can seep into the food chain and water supply. Conventional weapons only affect those in the immediate area for a short period of time. (Some psychological effects do overlap in individuals, although they're likely to be far more widespread after being hit by nukes than by conventionals simply due to the sheer power of the nukes.)

Much of this is probably stuff you already know, but clearly it needs to be repeated. None of these effects are debated by military planners or scientists. Because these effects are so terrible, they CANNOT be "derivative;" they MUST be central to the decisionmaking process. If they don't play such a role, then they are meaningless... or the decisionmaker is sociopathic and seriously lacking in concern for other human beings. Just as chemical and biological weapons are largely banned, so nukes should be, and for many of the same reasons.

You wrote:
you don't target with 69 weapons arriving in some sort of blind automatism, but rather sequentially until your target is finally destroyed:

Really? Once they're launched, you have ZERO control over them. The first one is successful, or the tenth, and that makes all subsequent hits "overkill" by definition.

I understand your point about NORAD, but it doesn't take into account the fact that in all situations but complete surprise, NORAD's orders have already been transmitted by the time it gets hit. By then, it's too late; in fact, ironically, by nuking NORAD the enemy might be depriving itself of a remote chance that someone there might be able to stop the carnage. (Yes, I know the Net is based on a DOD system.)

You raise the issue of an ABM system. While a legit point in theory, nobody actually HAS a system that's very effective. Same's true of the satellites: that works if the enemy is NOT remotely our equal; but if a war were ever to reach the point of ICBM launches, those satellites would likely be destroyed almost immediately for the very same reason radar stations are.

12:03 AM