68 comments

[ 4.6 ms ] story [ 133 ms ] thread
A good read if you are interested in the topic. I do take exception with the constant tonal use of absolutes which seem to be trying to lead the reader to a false conclusion based on that false fear.

If the president (Biden or Trump) ordered a first strike nuke against Toronto or London there is a 0% chance it would happen. I personally doubt that an order against Moscow or Pyongyang would be followed either, but I wouldn’t argue with someone who wanted to take a position that those first strike orders would be followed.

My point is just that there is obvious a lot more...”human nuance” in the situation that the author is describing.

Yes. This is not an automated process, the president can't just hit a button to launch. While the generals have to obey the rules - they still can just say "no" if there is no logical explanation for the strike. Adding more safeguards could slow down the response in case of an actual attack and could do more harm than good. https://www.bbc.com/news/world-us-canada-42065714
and independent, critical thinking is exactly what the military trains for
I presume that was sarcasm, but it has some truth to it. Yes, the military trains for obedience. It also trains for independent thought. Your comms can be compromised, meaning you could be receiving information and orders from a hostile party. Or they could be completely cut off, and you could be receiving no orders, meaning you have to figure out what to do on your own. Or you could have a window of opportunity that lies outside the scope of your orders, and it could be gone by the time you get permission or approval. Or...

So, yes, the military does train for a degree of independent, critical thinking.

International law is one consideration but under US law, the generals are not allowed to refuse a launch order. Possibly they would anyway, but it's a poorly designed system that in some circumstances relies on people stepping outside the system in order to prevent WWIII.

Retaliation is another matter, but we at least need more safeguards against a president launching first. A good argument can be made that we shouldn't allow first strikes at all.

> International law is one consideration but under US law, the generals are not allowed to refuse a launch order.

That is not my understanding from talking with people in the army. What I understand is that anyone in the armed forced is not only allowed, but required, to refuse a unlawful order given by anyone, even the president. Executing on a unlawful order is a crime and you can end up court marshalled or even executed if the outcome is bad enough. The Nuremberg defense ("I was just following orders") does not fly in the US military.

I wonder how someone who receives an order is expected to determine if it's lawful or not, especially within 60 seconds? Doesn't that require a court of law to decide?
I asked a similar question, my understanding is it's a matter of degrees. If you follow a unlawful order and say file paperwork the wrong way that probably won't result in a court marshal, if you burn down a village in a friendly country you most certainly would.

That said Congress is supposed to declare war so most of the wars we have been fighting for the past two decades are likely unlawful. I doubt you would be successful refusing to execute a drone strike on that basis. However if the president called up drunk one night and ordered your to bomb London, I suspect you'd be safe in that refusal.

They are obligated to refuse illegal order.

For a use of nuclear force to be legal, it must satisfy customary requirements of necessity, distinction, proportionality, and avoidance of unnecessary suffering.

> Adding more safeguards could slow down the response in case of an actual attack and could do more harm than good.

Retaliating doesn't actually do any good. It increases the body count on the other side, but does nothing to help those in the US about to be atomized. It's not like the missiles will collide in mid-air.

What matters is the believed threat of retaliation because that may prevent an attack in the first place. The optimal system in terms of human life and safety is:

1. All foreign governments 100% believe that the US has complete capability to retaliate with devastating force at a moment's notice and will do so with zero hesitation.

2. But the US actually has no such capability so that an accidental first strike on a false alarm is impossible.

The tension between those two is the hard part. :)

I wouldn't count on military commanders employing "human nuance" within 30 seconds of receiving a direct presidential order. On the contrary, it's pretty amazing that with 2, 3, 4 or more nations on a nuclear hair trigger for multiple decades that the systems of deterrence and avoidance have so far been successful.
The military drills into its officers and enlisted personnel the fact that they not only can disobey an unlawful order, it's their duty. A direct presidential order is not inherently lawful, and I'd expect most commanders (especially, in contrast to lower ranking enlisted or officers) to be able to respond appropriately.
Even granting your position, but while the set of “orders to use nuclear weapons that the public would, in retrospect, consider catastrophically undesirable” and that of “orders to use nuclear weapons that would be clearly illegal” do overlap, the former is not at all fully contained in the latter.
Even granting your position for the sake of argument, while the set of “orders to use nuclear weapons that the public would, in retrospect, consider catastrophically undesirable” and that of “orders to use nuclear weapons that would be clearly illegal” might overlap, the former is clearly not fully contained in the latter.

But I’m not sure we can take your position as true, anyway:

First, the historic lack of accountability for illegal Presidential orders or obedience to them both reduces the clarity (through absence of case examples) of the legality of such orders and reduces, due to incentive structures, the probability of any person choosing to be the one that takes a stand on the perceived illegality of such an order even were the legality relatively clear.

Also, given the War Powers Act, its not clear how any order to employ the military by the President (except domestically in a manner violating the Posse Comitatus Act) would be illegal under domestic law initially (which, for a nuclear attack, is all that is going to matter.) It might violate norms of international law like waging a war of aggression, but I think recent history shows that even if maybe you can trust the US military to resist some set of illegal orders, “orders to wage of war of aggression that is, nonetheless, authorized by US domestic law” is pretty clearly not within the scope of “illegal orders” for which that is true.

There's a bit of nuance to this.

An order is presumed to be lawful under military law. The burden of proving it unlawful rests on the person receiving the order. The default is to execute the orders given.

In the absence of some direct reason that it would be unlawful, a military person must carry out the order they are given.

Now of course, as you say, it is the duty of the one ordered to do something that is unlawful to refuse to carry out that order.

The entire system is designed to eliminate human nuance once the order is given. Exactly how successful it is is impossible to say, but it's designed to make carrying out the order both fast and straightforward.

The only tricky bit is if the President wanted to execute a nuclear plan that hadn't been devised yet. But if there's an existing one (attacking NK, Russia, Iran) then ordering a first strike is really straightforward with an absolute minimum of people in a position to refuse.

This article doesn't address important differences in the scenario in which a launch order is given out-of-the-blue, and when a launch order is issued following certain events.
> I do take exception with the constant tonal use of absolutes which seem to be trying to lead the reader to a false conclusion based on that false fear.

I don’t think that’s the purpose. It’s describing features of the protocol, which are the defined rules. Yes, humans deciding not to follow the rules can result in different results, but when we are discussing what the rules for use of nuclear weapons use ought to be, you probably don’t want to avoid considering improvements on the basis that that status quo is acceptable because, in the event that the rules currently in place would result in an outcome viewed as undesirable by certain key actors, there are points in the process where they would be able and likely to execute an unconstitutional coup d’état.

Speaking to people who have worked in Titan II silos (admittedly, long ago in the 70s), the opposite is true. They were conditioned to believe if they received a launch order, the US was already a cinder. The person I spoke to said he would "absolutely" fire if he had received an order.
The President doesn't directly speak to those in the silos, there's people in between.
Obviously - the point being, this command chain is optimized and conditioned for compliance.
Why fire? If America is already gone, why retaliate? Revenge? What is there to gain? Retaliation would just make things that much more difficult for the rest of humanity, assuming the first strike didn't already make the earth uninhabitable...
The purpose of a retaliatory posture is the foundation of MAD. If the posture, instead, is: We'll just rollover and die, then (the theory goes) there's reduced reason for others to avoid a first strike.
Yeah, I understand the reasoning of MAD, but on like a thinking individual level... when you get the order and are under the impression that this is a second strike, why hit the button? MAD clearly didn't work at that point, so why fire? On the off chance that it was a small first strike and this is a limited second strike? I, at least, would rather run the risk of upsetting the doctrine than run the risk of destroying all of humanity.
Perhaps there are 2 kinds of people: those who think the way you're describing, and those who would hit the button to retaliate. I wonder if the people who end up sitting in the missile silo are selected for the job by asking them how they would respond.
They are absolutely selected for that.
Did you by any chance read the _Three Body Problem_ novels? It is partly about this problem, but I won't describe the details because of spoilers. Highly recommended.
If someone in Russia has decided my family can be incinerated, I don't see why the favor can't be returned. I mean, if we had a mindset of "one human family", nuclear weapons wouldn't even exist.
There are American citizens in Russia, you would be choosing to kill Americans for petty reasons.
Revenge is probably reason enough.

But even a full Russian strike probably wouldn't 100% destroy the United States (and a full Chinese strike probably wouldn't even kill half of us unless society totally collapsed afterwards).

Contrary to popular belief, in the event of nuclear war we wouldn't all suddenly be vaporized. Especially today when the number of nuclear weapons is severely limited and there are credible missile defenses, in a first strike only a few major cities and military installations are going to be targeted. It's more important to hit DC 10 times to make sure some of them actually get through than it is to destroy, for example, Albuquerque. While you might seriously be looking at casualty numbers comparable to the whole of WW2 in a matter of hours, the vast majority of both populations would still be alive and relatively unscathed. If the other side still has their military capabilities while we do not, those survivors would be at their mercy. This is to say nothing of allied nations which may not have been targeted in the nuclear exchange but rely on our military for protection. Nuclear exchange is unfortunately merely the first part of WW3.
(comment deleted)
simple explanation; orders must be lawful.
The people receiving the orders are trained to understand that they do not/will not have full situational awareness. The president issues the orders and they carry them out.

The presidential football is basically a menu of strike options that are ready to execute. To launch a strike all the president needs to do is call the number in the book, authenticate himself using the a two letter challenge code, and tell them which strike from the menu to order. Targeting is preprogrammed and designed to eliminate any delay or nuance in the process.

The US has spent billions (trillions?) of dollars on its nuclear strike capabilities and the systems that order its use. To presume that these systems will fail precisely when they are called upon seems short sighted.

Word has it (supposedly sources inside the white house) due to Trump's habit of making impulsive statements / demands that sound like orders, the folks at the Pentagon put in place the expectation that orders went through some white house staff / filters / had to include written orders before they did anything extraordinary.
I will breathe a lot easier a week from now. This system definitely needs to be reformed.

Daniel Ellsberg's recent book The Doomsday Machine supports a lot of this, but adds that historically, theater military commanders have also had independent launch authority, and as far as he can tell, they still do.

(Ellsberg is best known for the Pentagon Papers but his job was actually nuclear strategy. He'd intended to publish documentation on that as well, but thought if he did that first, nobody would even care about the Vietnam stuff. The nuclear papers were lost before he published them.)

Since I've read 'Three Body Problem', I'd be breathing a lot less easier a week from now.
Seems like a rogue presidential order wouldn't get past the "war room at the Pentagon", they need to actually send the message to the silos.

I assume they are fully aware of the threat level and aren't going to level London on a Tuesday afternoon for no reason.

Yeah. You might be court martialed afterwards, but there are plenty of links in the chain, down to the guys in the silos saying “but really?”.

I also find it an odd standard to treat nukes differently than any other military action. You think we don’t have enough traditional options to level London on a Tuesday afternoon? It seems like the President is either the CinC or he isn’t...

> You think we don’t have enough traditional options to level London on a Tuesday afternoon?

I actually don't think there are any that are feasible. Leveling a city the size of London is no small feat, and amassing the troops and weaponry required to carry it out seems like a complicated enough exercise that it would take longer than "Tuesday afternoon".

Not that I don't think it can be done, but just that moving a large traditional military apparatus, or even mustering enough conventional missiles to flatten an area that large seems like it would take far longer than that.

The "shock and awe" campaign waged on Baghdad might be similar in scale within about an order of magnitude, and it certainly took a heck of a lot longer than an afternoon, although the goal was never to level the city in its entirety.

Why flatten? Just a few supersonic booms near ground level would shatter much glass. The facades may be rated for storms and thunder, but the pressure gradients of low flying supersonic jets?
Nukes level London in 5 minutes (and the launch alone initiates MAD from China and Russia before they realise what the target it)

That's not enough time for the VP and cabinet to 25th amendment and cancel the orders.

Is that so? I heard there are cruise missiles able to do the job. No need for ICBM glaring in infrared and the visible spectrum all over the skies.
And that also elevates the risk since countries that might launch nuclear arms at the US understand that needing to second-guess the command to launch our weapons gives them an advantage.
I imagine a second-strike is going to get a lot less hesitation than a first strike without obvious provocation. If nothing else, there's no harm and potentially an immense benefit in delaying a first strike to get confirmation, whereas there is a time crunch if enemy missiles are inbound.
(comment deleted)
The U.S nuclear protocols are secret and very little is known about their current situation. Optimal deterrence position is when enemy is uncertain what the protocol is but the US actually has launch after confirmation. Land launched ICBM's are not absolutely necessary for US deterrence anymore, so launching them before they are destroyed is not necessary.

While the US certainly maintains the technical ability to have launch on warning, it seems that it might not be the policy.

Clinton Issues New Guidelines on U.S. Nuclear Weapons Doctrine https://www.armscontrol.org/act/1997-11/news/clinton-issues-...

> ... while the United States has always had the "technical capability" to implement a policy of launch on warning, it has chosen not to do so. "Our policy is to confirm that we are under nuclear attack with actual detonations before retaliating," he said.

And presumably the choice to fire back before our ICBMs are destroyed by incoming nuclear weapons would take into account the size of the confirmed launches by an enemy.

If you get confirmation that Russia launched 1-10 ICBMs, well maybe you are reading it wrong, and the loss of 100-200 missiles silos if you are wrong won't make much difference.

But if you get confirmation Russia is launching 100 ICBMs, its bombers are in the air, and its subs are rising to launch depth, well lets just hope that's hard to fuck up that confirmation.

The US has enough submarine launched ICBM's to retaliate.
Well, congratulations.

The real winners will be the ones who never saw it coming.

> But if you get confirmation Russia is launching 100 ICBMs, its bombers are in the air, and its subs are rising to launch depth, well lets just hope that's hard to fuck up that confirmation.

Will nuking Russia in that situation prevent the US from being nuked? Will it help any American citizen anywhere in the world?

Yeah this stuff is ridiculous, but I think the idea is not that it would help anyone but that the “policy” would deter the first aggression.
In an all out nuclear exchange with Russia, the US is estimated to lose 15-30 million people. Terrible, but only 5-10% of the population, mostly from a few major population centers. Even if the government and military collapses, the American people will still exist after getting nuked, and presumably they won't want the Russians to have a large, unscathed military.
Unless the Russians were total psychopaths, their first strike would probably be aimed mostly at our nuclear missiles and the military at large (i.e., a counterforce strike), and not on purposefully destroying our population (i.e., a countervalue strike). So our nukes would reduce their ability to do second, third, etc. nuclear strikes. It would also destroy their ability to continue to make war, which would prevent an invasion of Europe or the USA after the radiation levels drop.

If we had a policy to wait a long time, the Russians could strike our missile bases, large military bases, aircraft carriers, naval bases. But unless we acted fast, our military forces would be in shambles even though the eastern seaboard, Chicago, SOCAL, the major texas cities, etc. are all still standing. We'd still have second strike weapons but they are unusable if we still had domestic civilian targets to worry about. No sane President or General would order a coutervalue strike when faced with a counterforce strike from an enemy.

If we let the enemy successfully first strike the United States, Mutual Assured Destruction goes out the window.

TL;DR; the Russians could "win" a nuclear war unless we shoot back soon.

And if the Russians were total psychopaths who target our cities in a first strike? We have to be prepared to respond in kind or else we incentivize a one sided attack.

Well, the threat of nuclear annihilation has helped all the nuclear powers in the world from actually using their weapons, even their low-yield tactical weapons. It might not be logical to bomb them to smithereens after they launched against you but you better hope they assume you will do so in retaliation. As long as everyone thinks you will make good on your promise to launch if attacked, the incentive to attack you is close to zero.

And yes, I know the nuclear powers have fought proxy wars outside of their own territory, That is not what this discussion is about, this is about keeping conflicts from going nuclear. The Soviets first and the Russians after were and are rational countries, unlikely to invite their own destruction by launching against any other the other nuclear powers. The same goes for nearly all other nuclear powers with the possible exception of the North Koreans. The real danger with nuclear weapons is the possibility of them ending up in the hands of irrational actors, pick and choose from religiously motivated terrorist groups to ecofascist organisations and anything in between. It is somewhat surprising that there has not yet been a nuclear bombing by such a group given the supposedly sizeable number of nuclear munitions which have gone missing over time. I can only assume that this means that it is harder than thought to make this actually work.

It would be more than that, a launch without any pre-existing escalation is unlikely, I’m not sure what the US or Russia for that matter would do in case a “rogue” launch.

We know from a few previous close calls that the Russians didn’t launch when they erroneously detected launches.

Quite likely any nuclear engagement would be preceded by an adversary trying to destroy or minimize your second strike capabilities first.

The US keeps its SLBM fleet in a rotation that would always be optimal for a second strike against Russia.

Russia increasing its anti submarine activity or even destroying US subs would likely be seen as a prelude to a first strike.

So Russia spending 3 months pushing US subs into a kill box further and further form optimal launch positions and then the US detects a ground based launch no matter how small would likely invoke a ground based launch from the US before anything actually hits.

What do you mean by launch depth? And how would that detection supposedly work?

I'm asking because AFAIK they don't need that anymore, at least not up to periscope depth. (Unsure here, something like 30 to 20meters in my mind)

edit: By that I mean I'm seriously doubting the capability of anyone to have sufficient coverage by whichever means to detect that in realtime.

ICBMs are rockets... which means the heavier, the more fuel for a given range, even heavier with the extra fuel, etc. (aka the tyranny of the rocket equation).

SLBMs also need to withstand water pressure, increasing with depth. Withstanding pressure requires reinforcement, which means weight, which... see above.

So while there are no "official" numbers as to what constitutes launch depth, physics dictates there is a maximum practical number.

So what to do about it?

Well, nuclear attack submarines generally exist to try and follow nuclear ballistic missile submarines.

Any any given point in time, >0 and <all ballistic missile submarines are being shadowed, likely without knowing it.

Consequently, were a substantial launch order given... a lot of those missile subs would start a very synchronized and quick rise to shallower depths, as well as possibly creating mechanical noise related to launch preparations.

After that, well, that's what preset plans of action in different threat scenarios are for...

More about the lack of or simply bad protocols, but if you want a good read and scare yourself “Command and Control” by Eric Schlosser is highly recommended.
Overall, I enjoyed listening to that as an audiobook. I thought it was a bit odd that the story of the damaged Titan-2 was interleaved with the part about protocols and security, which I considered the more important part of the book. Whenever the story switched back to the Titan-2, I found myself wishing that they would just get it over with so I could get back to the more interesting part. However, the Titan-2 story might make for a good movie by itself.
Grey on white background is the nuclear strike of web design.
Thank goodness for Reader Mode and Outline. I hate such websites with a passion.
Really good article.

I don't know much about nuclear policy myself, but I know this guy @DavidSantoro1 in Oahu who works on that, and recommend following him if you are interested in nuclear issues.

This is especially relevant to people in Hawaii, which is the only state, as far as I know, to have sent all its residents an emergency warning that a missile was about to hit (https://en.wikipedia.org/wiki/2018_Hawaii_false_missile_aler...). Luckily it was a mistake and the only affect was a million funny stories about how everyone reacted (btw: someone should write a book cataloging the best reactions!).

@RadioFreeTom is also a good source for this stuff.
Oof. We made it this far, can we not talk about this until next week?
If you're ever around Tucson AZ, I recommend a visit to the Titan Missile Museum. (No affiliation, just a fan.) The tour includes a simulated launch order, two volunteers are chosen to turn the keys, and the ancient control panel will blink its lights in sequence.

https://titanmissilemuseum.org/