Questions and Answers About No Launch on Warning

Questions and Answers on “RLOAD” and De-alerting

This page is based on comments by Dr. Alan Phillips, then a retired medical radiation specialist* .
See also the RLOAD page here. Direct interview with Alan Phillips here.

1. Would the U.S. be able to convince its military that, due to the occasional unreliability of its technology, they should confirm a nuclear attack by waiting for evidence of a nuclear explosion having occurred?

(a) The military definitely do not want an accidental nuclear war. The advantage of RLOAD (Retaliatory Launch Only After Detonation) to them, as to everybody else, is to reduce that risk which they know about as well as anyone. They have gone to immense trouble over the years to minimize the risk of accidental or unauthorized launch, as well as to make their warning systems reliable. They have not managed to eliminate false warnings, and they know it.

2. Wouldn’t the military object to “sitting back and taking it”, before retaliating?

(b) They are going to have to “take it” anyway. The warning has shown that the attack, if real, is on the way. Retaliation does not stop or mitigate the first salvo. I hope and think it probable that at the top they are professional enough to be willing to wait those few minutes, which they know makes no difference to the force of the first attack. Answer (a) above shows their very strong motivation not to launch on false warning. It has saved us a few times already; notably when a Russian officer refused, on a hunch, to send on up the line a warning that he only suspected was false, and it was. Although he may have saved the world, his career was ruined.

3. Would the public understand the need for the U.S. to take a hit to avoid hitting the Russians unnecessarily, and that after apparently being willing to drop a bomb on the Russians for the last 50 years?

(c) The public is not consulted or told about such matters as LoW, nor details of the planned retaliation. It takes a lot of effort by peace activists to find out any of these things. The fact is, though, that if the attack is real they are going to have been hit anyway, long before the retaliation reaches Russia. Half an hour or more. Retaliation does nothing to stop missiles that are already on the way.

“Deterrence” is threatening retaliation so inevitable that the attack is never started. Once an attack has started, deterrence is history – an endeavour that failed. And if we hit the Russians “unnecessarily” (i.e. because of an error in our warning system, not their fault) we get hit back, fatally. Those who have thought about it know that if we drop a bomb on Russia we suffer a devastating nuclear attack from which we do not recover. They do not want that to start by accident; and those who are socially aware have a healthy scepticism for technology. They can easily understand that warnings are not infallible.

The question to ask when an attack is reported to have started, is: “Is it wise to retaliate at all?” The military may not ask that question, but it is a real one. All that retaliation might do for us is to reduce the number of times “the rubble bounces” – or it might increase that number.

However, if we have a campaign and make a big plea to adopt “RLOAD”, it could easily be misrepresented (intentionally or unintentionally) to appear as giving up an advantage in combat. It isn’t that at all, but it might give up a tiny element of deterrence (see answer (i) below). We have to remember that, on the normal assumption that deterrence is what prevents nuclear war, maintaining deterrence is not only for the benefit of one side, but of both, and of the whole world. If either side is tempted to attack, it is a disaster for the whole world. The same with an accidental war. In my view, the gain here is far greater than the loss, because, as shown in (i) there is practically no loss of deterrence to set against the big gain of eliminating the risk of a purely accidental war. Note that nuclear weapons have resulted in occasional collaboration between adversaries: the “hot line” was a cooperative measure to reduce the risk of unintended war.

4. They have the “hot line” to resolve misunderstandings, so why do we still have to give up LoW?

(d) In the situation of a possible false warning, the “hot line” could probably not be used quickly enough. It does not run from one president’s office to the other, but between the Pentagon and the Kremlin, and messages both ways need translation. Also, both sides know that if the other side were really trying to do a surprise attack they would certainly have rehearsed what to say on the “hot line” to keep their enemy confused.

5. Would the military be more readily convinced that they must have the launch confirmed by 2 or 3 independently operated radars or satellites before the “warning” is considered a reality and a response authorised?

(e) The U.S. Strategic Command say they operate like that already, requiring two independent warnings. My “20 Mishaps…” paper includes cases where that practice (whether policy or not) has saved us. But it has been dicey several times.

As to authorization of response, the U.S. military have always officially planned on following the law, and no launch or use of a nuclear weapon is permitted without the President’s authorization. To some extent we believe authority to use nuclear weapons has been conditionally pre-delegated, so they may not always be bound to wait for his or her authorization.

On the Russian side, command of nuclear weapons is said to be very centralized and strictly controlled. However, they have not in recent times had both the radar and the satellite warning systems available all the time because too few satellites are orbiting and some of the radars built by USSR are now in independent States; so they must be relying on only one system for part of the time. They also have a “dead hand” system codenamed ‘Perimetr’, which comprises non-armed rockets that can be launched automatically and fly over Russia broadcasting launch codes and launch orders to the missile silos. This is meant to be activated automatically (after an enabling action by the high command) if Moscow is destroyed and communication by the high command to the nuclear forces is lost. It has been said that the system could be activated inadvertently at a moment of crisis. [The working of Perimetr is better described in our recent paper “Replace LoW Policy”]

6. “The Nuclear Posture Review claims that US strategic nuclear forces are not on ‘hair-trigger’ alert. How does this square with persisting reports that “some 2,000 strategic nuclear warheads on each side are still held at ‘launch-on-warning’ readiness?”

(f) It squares in this way. The Dictionary of Military Terms found at contains some 100,000 definitions. Neither ‘hair-trigger’ nor ‘launch on warning’ is among them. However, ‘Launch under Attack’ is defined there as: “(DOD) Execution by National Command Authorities of Single Integrated Operational Plan forces subsequent to tactical warning of strategic nuclear attack against the United States and prior to first impact. Also called LUA.” This definition is identical with peace activists’ conception of ‘launch on warning’.

The lack of a definition of ‘launch on warning’ could explain the fact that US military spokespeople have sometimes told us that the strategic forces are not at ‘launch on warning’, but at ‘launch under attack’ (which is defined). However, correspondence with a retired British Naval Commander has indicated that the term “Launch on Warning” has had a meaning in deterrence theory. It seems to have meant, or to have included in its meaning, the launch of a pre-emptive disarming attack on strategic warning of a nuclear attack believed to be imminent, but not yet launched. Strategic warning would include information from interception of Russian military messages, satellite observation of preparations on the ground, information from spies, and the like.

I have no doubt that the drafter of the NPR statement in Q.#6, expected that the sentence “US strategic nuclear forces are not on ‘hair-trigger’ alert” would be taken by most people as a denial of Launch-on-Warning and of LUA. However, that is not what it says. It says they are not at hair-trigger alert, without defining what that means.

‘Hair-trigger’ is a familiar term in connection with shooting for sport. It would indicate a very much quicker response, on sighting a target, than what would happen in the context of Launch under Attack by the strategic nuclear forces, namely: calling a threat conference, consultation with the President to obtain his order to retaliate, and giving the order to launch rockets perhaps 20-30 minutes after the first warning. So the NPR statement is not a definite lie, and it does not deny Launch-under-Attack nor Launch-on-Warning.

7. If the US was convinced to allow time for the first explosion to be reported before responding, would they not be concerned that the first warhead would be so well directed that it would eliminate key elements of the decision-making apparatus, thus causing chaos? A hit on the White House, for example, could cause extraordinary problems…”

(g) That is the whole reason for LoW. However, switching now to RLOAD would not disable the greater part of the mechanism. As soon as the warning was seen to show a serious possibility of being real the president would be warned and the launch crews alerted. At the final conference (while the perceived attack was still apparently on the way), when the president ordered retaliation it would be “subject to the attack being real”. The order to start the launch process would be given immediately by the general in command. The launch crews would stop before ‘turning the key’. When the first nuclear explosion was reported the launch keys would be turned immediately. If a detonation did not come within a certain short time after the predicted time of arrival of incoming missiles, the preparatory steps would be reversed and the launch crews would revert to peace-time readiness.

Nuclear explosions are detected and reported automatically (at the speed of light) to all command centers. Under RLOAD they would also be reported to the launch crews, and the retaliatory launch would be done (I believe) within a few seconds. It seems to me inconceivable that either side could launch such an attack (at 8,000 kilometers range) as to prevent retaliation. It would have to would disable all the land-based missiles within seconds. The attacking missiles start from silos many kilometers apart and the target silos are similarly spread over a wide area, so the times of flight would not all be the same.

The president would be much less stressed under RLOAD than he would be under the existing LoW policy, because he would know the launch would not be done if there was no attack.

8. Would the decision-making not be required to continue over a period of time? Do they launch one missile, or everything at once?

(h) I believe the “Single Integrated Operational Plan” (SIOP) has a number of choices. Most if not all would be for a considerable list of targets at the first salvo, like missile silos, submarine docks, command posts, communication nodes, or cities. The Chief of Strategic Command would advise the President at the final conference (during flight time of the incoming attack) which SIOP option to order, according to his assessment of the situation.

9. What are the odds that all key people could be airborne in Air Force One within 15 minutes?

(i) It’s been done at least once for an actual (false) warning – all except the president. They couldn’t find him in time but the airborne command plane (which is not Air Force One) took off anyway. Ultimate launch authority is already conditionally pre-delegated to someone who was there. In any case, my answer (f) above shows that getting the decision-maker airborne before the bang is not necessary for retaliation, and (c) shows that actual retaliation is not necessary. All that is necessary to prevent intentional nuclear war is for the Russians to be fairly sure that retaliation will happen if they start it; and vice versa.

10. If the first explosion was air burst, would it not result in an electromagnetic pulse (EMP) which would destroy communications and response capability?

(j) That could possibly be done by a very well-planned first strike with several special bursts engineered to maximise EMP. (It is not “air bursts” that do it so much as bursts far above the palpable atmosphere, several hundred kilometers up.) It is the possible damage to satellites that gives me some concern that this may be a weak point in my position. I am pretty sure such a first strike could not stop the retaliatory missiles being launched. I do not think they all rely on satellites for navigation, but I am not 100% sure of that. However, as in (c) above, no-one gains by the retaliation; the only possible disadvantage from adopting RLOAD is the remote possibility that the other side might think they could do a first strike like this, with maximised EMP, and survive. But to try it they would have to be *sure* that the EMP would be completely successful. Half a dozen retaliatory warheads exploding near their targets (one MIRVed missile) would be a terrible disaster for the attacking country; a complete failure to prevent retaliation would be the end of their country. They know this. They also know that submarines and the missiles they carry are unaffected by EMP. (P.S. One hydrogen bomb exploding over a city would be 100 times as bad as the attack on the World Trade Center, and then there is the radiation from fallout.)

11. I know that the US military is apparently fully hardened against EMP, but are the cars they drive, the traffic lights they must go through to get to the base, etc?

(k) Anticipation of failing transportation might bother the military, who may think they will be fighting for a long time. From the “deterrence” point of view, and the citizens’, there are plenty of people at their battle stations all the time, on both sides, for a very effective nuclear holocaust to be made.

12. It seems peculiar to propose a system that is based on mutual distrust. I would prefer a system based on trust which, over time, serves as a confidence-building measure. Thus I’m very anxious to get on with the examination of means of de-alerting.

(l) I agree entirely. I am not proposing RLOAD as an alternative to other more radical measures to prevent nuclear war, which are certainly needed. The big thing about RLOAD is that it can be done at once, as soon as the decision-makers see the point. Trust is slow to develop where it has been absent for decades, which is part of the reason why RLOAD can be put in effect so much sooner than verified symmetrical de-alerting. The latter does require some degree of trust.

RLOAD would be a very valuable step that can be done at once, and I think it is what we should be campaigning for now. I am convinced that real de-alerting (with enforced delay) will take a long time to get through all its hoops, and I am scared that accidental nuclear war will put an end to all our endeavours before the objectives are achieved. I am pretty sure that LoW represents more than 90% of our present risk, and that the risk is real.

13. A dream – for Canada to host an international conference to consider preliminary steps toward a nuclear weapons convention. This would presumably include a discussion of de-alerting methods among technical people from the nuclear weapons states and key non-nuclear states.

(m) Yes, keep working on it. But how about RLOAD to protect you a bit while you work?

*Dr. Alan Phillips graduated with honours in physics at Cambridge University in 1941. He spent the rest of World War II doing radar research for the British Army. After the war he qualified in medicine at Edinburgh University and specialized in the treatment of cancer by radiation. He retired in 1984. His retirement activities included the study of nuclear armaments and the risks of accidental nuclear war. Alan Phillips died in 2008.