The unauthorized use of nuclear weapons is not the only way in which a cataclysm might be initiated by accident. Although the United States deployed thousands of nuclear missiles that could be launched almost instantly—in a single salvo—some influential missile experts feared the enemy could destroy our whole force before it got off the ground. Accordingly, they urged that the “retaliatory” salvo ought to be launched upon receipt of warning that an enemy attack had been started—the “warning” in this case being an interpretation of radar screens and perhaps other signals that an attack appeared to be on its way. Many advocates of this perilous “launch-on-warning” policy misrepresented it as arms control. They argued that the United States and Soviet Union would be condemned to an indefinite arms race unless the situation was “stabilized” by threats of mutual assured destruction—which became known as MAD—meaning each side could instantly inflict devastating retaliation and thus always deter an attack.8
The doctrine held that (1) defenses against missiles had to be banned lest they prevent retaliation, and (2) each side had to keep its missiles poised for instantaneous launch lest they might be destroyed on the ground. This double-barreled position—“yes” to launch-on-warning, “no” to missile defense—became the accepted dogma for quite a few liberal politicians.9 I am convinced both ideas are deeply flawed.
At the time these arguments were gaining currency, I was Director of the U.S. Arms Control and Disarmament Agency (1973–1976) and could use my inside-the-government position to stress the danger of launch-on-warning. Shortly before President Nixon nominated me to head the agency, I had published an article on this danger.10 (The article almost cost me Senate confirmation because it had somewhat unflatteringly quoted an endorsement of launch-on-warning by Senator William Fulbright, then chairman of the Foreign Relations Committee.) I had further encounters with this issue when I served President Ronald Reagan as Undersecretary of Defense for Policy from 1981 till 1988. One of my areas of responsibility in support of Defense Secretary Weinberger was nuclear strategy. During those years in the Pentagon, I learned a good deal more about the far-flung U.S. command-and-launch system and its complex interactions with what the Soviets were doing. Bruce Blair has written compelling warnings about these risks of accidental nuclear war.11 I was appalled to learn how close the Carter administration came to relying on launch-on-warning procedures as an acceptable way to make mutual deterrence more “stable.” And I became more convinced than ever that my earlier condemnation of launch-on-warning was prescient.
We must keep in mind that this dangerous Cold War legacy is still with us: highly destructive U.S. and Russian missiles remain capable of being launched within minutes. Over the years, there have been false U.S. radar signals indicating a Soviet nuclear attack, errors in the use of critical computers (illustrated by the mistaken Pentagon warning in 1980 of an incoming Soviet missile attack), and more hidden problems of unauthorized acts. Dangers have surfaced even from routine maintenance procedures (illustrated by the Chernobyl accident). In 1995, a bizarre incident occurred in Moscow when President Boris Yeltsin reacted with public bluster to an innocent Norwegian missile used for weather research.
Lastly, there is a persistent disconnect between, on the one hand, the specialists who know how to design and maintain different nuclear weapon systems, and on the other, the authority to examine the overall system in its entirety and, if needed, to order remedial action. The few most senior officials who would have the authority to look into every facet of this far-flung system have neither the time nor technical understanding to do so. But a review commission of subordinates large enough to include all the required expertise is unlikely to get full access to every piece of this secretive domain. To my knowledge, the various commissions tasked to conduct a “complete” review could never get to the bottom of all serious problems. The members either were not given access to, or were not told about, this or that arcane risk. For example, applying locks on missile launch controls or on the detonation mechanism of weapons is a frequently praised safety measure (which my 1958 RAND study had first recommended). These locks—it is said reassuringly—can only be opened with a numeric code that is available to no one but to the highest national authority. In reality, however, these codes have to be installed and maintained by many technicians, some of whom might inadvertently, or deliberately, leave some weapons or missiles unlocked. So far, luck has been with us.
Lesson Four: What Reagan Taught
Ronald Reagan had a different philosophy about the nuclear peril: he was deeply troubled by the risk of mistakes or accidents that could lead to Armageddon. I was a foreign policy advisor in his election campaigns (from pre-primary to the final campaign), and in July 1979 I participated in a small meeting that my friend Richard V. Allen had arranged with Reagan to review nuclear strategy. As we addressed the vulnerability of our missile forces to Soviet attack, Reagan heard us out patiently and then remarked that an acquaintance of his (whom I knew as a frantic industry executive) had told him there was no vulnerability problem at all—we could simply launch our missiles as soon as we had warning of an attack. Reagan paused briefly to let the point sink in (while I anxiously held my breath) and then added with firm conviction: “But I think, this is the wrong thing to do!” He had made my day.12
Reagan’s judgment on this
