Menu Style


Nuclear command and control: from fail-safe to fail-deadly

Public concern focuses mainly on the muscles of the nuclear arsenal, the missiles themselves. A recent book by Paul Bracken of Yale University, The Command and Control of Nuclear Forces, draws our attention to its nervous system, known as Command and Control. Data on the military environment flow continuously into the nervous system through its numerous sensory organs -- satellites, aircraft and submarine probes, communications monitoring systems, radars and other spying devices. The data are fed along hundreds of channels into the system's brain, the "fusion centers" where computers process the flood of data into what passes for intelligence and early-warning information, and on into the command centers where other computers control the responses to this information. The responses range along a continuum of successively higher alert levels, at which nuclear weapons are maneuvered into place and the physical and procedural safety catches on their use are progressively loosened.

Only over the last few years has Command and Control taken mature shape as a fully integrated system. The integration is "vertical," directly linking intelligence to nuclear control, and "horizontal," linking forces in different parts of the world. So when North American Aerospace Defense Command in Colorado received a signal in 1980 (from a faulty microchip) that the United States was under nuclear attack, the airborne command post of the U.S. commander in the Pacific took off from its base in Hawaii.

Levels of nuclear alert

The mature system has yet to be tested in a nuclear crisis. It is so complicated that no one can predict its behavior with any degree of confidence. But there is reason to fear the existence of a threshold level of nuclear alert beyond which the interacting systems of the two sides would drive one another into a frenzy of war preparations, ending up in a holocaust neither side had intended.

The current rise in Soviet alert levels as Pershing-IIs are deployed [in Western Europe] makes a practical test of this theory more likely. Until now the Soviet Union has maintained its forces at much lower levels of alert than the United States. This is why there has never been any intensive interaction between the alerts of the two sides: even during the Cuban missile crisis of 1962 only American forces were put on alert. The shock of a Soviet high alert would trigger hundreds of preprogrammed reactions from an American system that is adapted to a passive Soviet force posture.

Political leaders would attempt to maintain control over an unfolding crisis. But they would probably lack a clear picture of the fast-changing situation, as the system is an inadequate and accident-prone interpreter of the data it collects. (It failed to give the United States advance warning of the Yom Kippur War or of Argentina's invasion of the Falklands.) In any case, leaders would be overwhelmed by the large volume of information reaching them and by the need to make immensely complex decisions very quickly. Control of a nuclear alert would be a job for which leaders are poorly prepared and have no real-life experience.

As political control is lost, even before war actually breaks out, two things would happen to nuclear forces in proportions difficult to foresee. Some forces might "hang," waiting for orders that do not come. Control of other forces would be taken over by automatic responses to signals received and by standard operating procedures that might be irrelevant to the real situation. At some stage, battlefield nuclear weapons in Europe would be dispersed from their storage "igloos" to low-level military units, which would have to assume operational control over the weapons for logistical reasons.

Shattered into "islands"

If things come to war, the first nuclear attack will disrupt communications so badly that surviving nuclear forces will be shattered into disconnected "islands." The commander of each island will have to assess the situation and act independently. There will therefore be no way either to coordinate retaliation or to implement a ceasefire.

The complex official NATO strategies of nuclear war are impracticable. It will be impossible to exert the central control that they require. Among those who recognize this to be the case, Bracken explains, there are three schools of thought concerning what to do about it.

One school seeks ways to make nuclear war more controllable by strengthening Command and Control. This, however, is not a realistic goal.

Another school takes the attitude that there can be no such thing as a good nuclear strategy. Bracken dismisses this view as irresponsible escapism.

Bracken belongs to the third school. Its members believe that "the best hope of escaping annihilation" may be an attack aimed at decapitating the enemy's Command and Control, followed up by a massive pre-emptive counterforce strike. The enemy, it is hoped, will be so paralyzed that his retaliation will be "ragged and ineffective," if not averted altogether.

To those of us reared on the idea that nuclear war is deterred by fear of mutual assured destruction (MAD), this suggestion comes as something of a shock. Is it pure fantasy? Or is it reasonable to conjecture that decapitation has some chance of success?


A decapitation strike would combine about ten huge bursts in the upper atmosphere with attacks on about 100 key points of the enemy's Command and Control (command posts, fusion centers, etc.). 

The purpose of the atmospheric bursts would be to generate the strongest possible electromagnetic pulses (EMP) with a view to knocking out the enemy's communications and electric power facilities. Some communication links are insulated against EMP, but insulation on a really large scale is too expensive to contemplate. The impact of these bursts cannot be estimated reliably, especially as the detailed physics of EMP are not understood.

A decapitation strike would ideally destroy not only primary command centers, but also the secondary and tertiary command centers to which launching authority is pre-delegated in the event of destruction of the primary centers. The results of an attack therefore depend crucially on how much the attacker knows about the structure and locations of the victim's Command and Control. This information, of course, is shrouded in deep secrecy, but it is conceivable that the intelligence services of the two sides have penetrated this secrecy.

From fail-safe to fail-deadly

The paralysis of one's forces can be guarded against, and a decapitation strike by the enemy thereby made futile, by taking measures to ensure that retaliation will be automatic. For example, the blocks that normally prevent low-level officers from firing nuclear weapons without authorization can be removed. Reliance on orders to fire (fail-safe) can be switched to reliance on orders not to fire (fail-deadly). Thus, nuclear-armed submarines might be sent a "no-go" signal every four house. If a signal is not received at the expected time, it fires its missiles on the presumption that headquarters has been destroyed or incapacitated.

These steps, however, would entail a great danger of accidental nuclear war. For that reason, they would be taken only at a late stage in a crisis, under the pressure of fear of imminent enemy attack. So by launching a decapitation strike at a relatively early point in a crisis, an attacker may still hope to paralyze retaliation.

Confusion, indecision, and accidents may contribute to the paralysis of the victim of a decapitation strike. So may an unsuccessful attempt on his part to avert war by unilaterally reducing his level of alert.

In view of all these uncertainties, decapitation may seem to have some chance of minimizing or even preventing retaliation. The situation is destabilized even if this chance seems a small one, because as a crisis escalates the chance of averting war may come to seem smaller still. Each side contemplates whether a decapitation strike may not be the best hope of escaping annihilation, and believes that the other side is contemplating the same thing. At a certain point, the question of whether or not war can be averted is overshadowed by the question of who will strike first.

All these calculations, however, are based on outdated methods of predicting the consequences of nuclear war. Both the speculations of nuclear strategists and the computer programs of Command and Control essentially take into account only the direct and immediate effects of nuclear explosions. But scientists now know that the global environmental consequences even of a one-sided nuclear war -- the prolonged darkness and extreme cold of "nuclear winter," depletion of the ozone layer, etc. -- would be severe enough to rule out the survival even of a "successful" decapitator (see Norman Mayers' article in The Guardian, November 3, 1983). Destruction might not be "mutual" but it would nonethelessl be "assured."

Thus, nuclear war cannot be prevented merely by the vague and generally accepted belief that it would be a terrible catastrophe. The holocaust machine can only be stopped if policy makers on both sides arrive at the full realization that it is capable of one thing and one thing only -- destroying the world. It follows that there is indeed no "good" or even rational nuclear strategy. 

Note. I wrote this article for Sanity, the journal of the (British) Campaign for Nuclear Disarmament, at the beginning of the 1980s, at the height of the Cold War. Although the nuclear standoff between Russia and the West is less tense today, the issue remains pertinent, especially to regional nuclear confrontations -- for example, between India and Pakistan. Paul Bracken has made a more recent work on nuclear command and control available on the internet here. See also this site on nuclear de-alert.

You are here: Home Themes War and disarmament Nuclear weapons Nuclear command and control: from fail-safe to fail-deadly


This email address is being protected from spambots. You need JavaScript enabled to view it.