Would nuclear war endanger civilization or even the human species? Does that possibility require us to subordinate all considerations of freedom to survival and to dismiss any possibility of responding justly to a nuclear attack—or at least without committing suicide? The question has a familiar ring to anyone whose memory stretches back as far as 1958 to the Campaign for Nuclear Disarmament, and the famous controversy between the philosophers Bertrand Russell and Sidney Hook over whether it was better to be "Red or Dead."
The question, at any rate, is not dead. It is implicit in nightmare visions like "The Day After" that continue to flood our television and movie screens, and it appears in even more gruesome form in the cold and dark post-nuclear world described by some scientists. The horrors revealed by science can exceed, it seems, those of science fiction: the sense of one meeting, in 1983, was that there may not be any world after nuclear war. As biologist Anne Ehrlich summarized it:
In the northern target regions, it is unlikely that more than a tiny fraction of the original population could survive the first few months after a nuclear war of appreciable scale. . . . Nuclear war would render all but uninhabitable the only known habitable planet in the universe. Nothing of value to anyone alive today is likely to survive such a catastrophe—and least of all, the ideologies that supposedly motivated it. The virtues of freedom—or communism—pale when survival is not an available option and there may be no future generations to whom it can be bequeathed.
This particular meeting was based in good part on a draft study by the scientists Richard Turco, Brian Toon, Thomas Ackerman, James Pollack and Carl Sagan (now generally called "TTAPS," an acronym of the authors’ last names). Though they have greatly overstated the scientific consensus, TTAPS does have prestigious support. It extended the results of an international study in 1982 sponsored by the Royal Swedish Academy of Sciences (published in its magazine, Ambio) which indicated that smoke from fires ignited by a nuclear war might bring about a world of "twilight at noon." It seems to be confirmed by the conclusion of our own National Academy of Sciences at the end of 1984 (NAS 85) that there is a "clear possibility" a major nuclear exchange would obscure the sun for an extended period and bring on freezing temperatures with catastrophic long-term effects in the northern hemisphere. To this formidable international array of scientists Admiral Noel Gayler, former deputy director of the Joint Strategic Targeting and Planning Staff, has added some American military authority. Such massive scenarios, he has said, are "quite reasonable":
Whatever our rhetoric or theirs, in a general nuclear war cities will be struck, and they will burn. Will not at least some of the smaller cities be spared? Not likely.
The idea that any major nuclear conflict would mean the destruction by each side of the other’s cities dominates the European and American political establishment, including the press. And apparently the concern goes beyond that of major nuclear attacks or responses. In fact, TTAPS has estimated that even a "small" nuclear war burning 100 cities might trigger a global climatic disaster. And the editors of The New York Times have taken that to mean that any nuclear conflict on an appreciable scale would leave no survivors. The Times concludes that, so long as there are "those in the United States or the Soviet Union who believe there is any point in ever risking nuclear war, defining degrees of destruction is not an empty exercise."
Like most of our political establishment for the last 20 years, the Times firmly believes that "deterrence works because it is based on horror"; and apparently the less restraint and the more horror the better. Hence the enthusiasm for the latest flurry of calculations and TV spectaculars on the nuclear apocalypse which indicate that the horror would be total; and hence the drift of the Western establishment toward threatening to use, but resolving never actually to use, nuclear weapons, first or second, early or late. "Deterrence Only" means giving up if deterrence fails; it assumes, as some make explicit, that to acquire weapons in order to be able to respond to such attacks discriminately is wrong and hopeless, but that even "under communist rule . . . as long as life exists, there is hope. A nuclear holocaust would wipe out that hope."
It appears then that we may be faced with a choice between darkness at noon in the political sense of Koestler or darkness at noon in the literal sense.
Recent calculations of the possible global effects of a nuclear war are not the first. They are only the latest in a long line, though they may be the grimmest since the early 1940s.
The possibility of a total nuclear cataclysm came up even before the formal start in August 1942 of the Manhattan Project. Two months earlier, Edward Teller had expressed concern that the enormous heat he had calculated would build up in an atomic bomb might be enough to "melt" the Coulomb barrier between light nuclei. Robert Oppenheimer suggested that, if so, a single explosion of an A-bomb might ignite and destroy the atmosphere of the entire planet. A detailed theoretical investigation of the basic theory of thermonuclear reactions disposed of this appalling possibility before Trinity, the first test. It showed that there was virtually no danger that the test would set the atmosphere on fire. Definitive theoretical work on this point must have been quite a relief.
Nonetheless, in the immediate aftermath of Hiroshima, Manhattan Project scientists talked in apocalyptic terms of the alternatives as being "One World or None." They argued typically and without equivocation that "if war breaks out, atomic bombs will be used and they will surely destroy our civilization." It was this catastrophic early prospect of "no world" that lent urgency to the pursuit of "one world." Literally hundreds of other statements by Manhattan Project scientists issued shortly after Hiroshima were very similar. Though few conjured up a disastrous global change in climate, Edward Teller did caution that "it was not even impossible to imagine that the effects of an atomic war fought with greatly perfected weapons and pushed by utmost determination will endanger the survival of man."
A basic assumption of that period was that in an atomic war both the attacker and his victim would aim their atomic bombs primarily or exclusively against the other’s cities. Scientists asserted typically that "atomic bombs are weapons used only against large cities and industrial centers." This was a natural, though mistaken, inference from the circumstances in which the first two A-bombs were used and from the experience of strategic bombing in World War II when thousands of non-nuclear weapons were delivered in single raids, some of which destroyed more civilians than the Hiroshima or Nagasaki bombs.
The alternative of "no world," then, contemplated a war that began with the destruction of the major cities of one side, followed a few hours later by the destruction of the cities on the attacker’s side. This is the way an atomic war appeared in the writings of the eminent scientists associated with the Metallurgical Laboratory (Met Lab) at the University of Chicago. Met Lab scientists, like James Franck and Leo Szilard, recognized that if the attacker expected his own cities to be destroyed only a few hours later than his victim’s, this might act as a deterrent to attacking the victim’s cities. However, they did not give much weight to the stability of such a deterrent. They thought that nuclear war was sure to break out through some irrational or unauthorized act or accident, "sooner or later." This lent great urgency to making drastic changes in the international system.
Events soon indicated that "one world" was not a realistic alternative. The Soviets turned down the Baruch-Acheson-Lilienthal plan for international control of atomic energy at the end of 1946. A series of crises followed over Berlin, Czechoslovakia and other key points, culminating in the first Soviet explosion of a nuclear weapon in August 1949. In the course of these crises, and later in debating the hydrogen bomb, the major Manhattan Project scientists began to reexamine their insistence that unless there was one world there would be none at all. They reviewed the possibilities for limiting the catastrophic effects of a nuclear war and devoted about a decade to the intense pursuit of programs of defense that might preserve Western civilization. These efforts included programs for the large-scale use of small tactical nuclear weapons in Europe rather than large strategic warheads (Project Vista at California Institute of Technology); for the fleet defense of the sea lines of communication (Project Hartwell at Massachusetts Institute of Technology); for extensive shelters and other "passive" civil defense measures in the United States (Project Charles and its sequels, leading to the Gaither report); for the interception of Soviet bombers on their way to target and other measures of "active defense" of the continental United States (The Lincoln Summer Study, Project Lincoln at MIT and others).
A consensus of such studies, embodied in a 1953 report by the Panel of Consultants on Disarmament of the Department of State entitled "Armaments and American Policy," was that our preoccupation with a major offensive capacity against the Soviet Union should be, but was not, matched by any corresponding concern for defending the United States against a Soviet atomic attack. The report concluded that the interception of a high percentage (though not all) of enemy bombers on their way to the United States was feasible and important. Civil defense measures could further contain the damage done by those Soviet bombers that got through the defense. Defenses would improve the position of the United States by reducing the damage the Russians could do on any one given date, and by making it more difficult and expensive for the Soviets to achieve a specified result. Defenses would help develop a healthy (but not yet excessive) sense of the dangers of the atom. Such a continental defense would help the United States in negotiating for the regulation of atomic weapons. It could not be read by the Soviets as an aggressive move. And, while geography made such a defense of Europe less practicable, given the state of the art then current, "if the United States can maintain some immunity to a knockout, the American connection may yet serve to protect the Western Europeans and so to quiet their fears." Moreover, research and development were likely to yield still more impressive gains, and might eventually give increased hope for the air defense of Europe.
In brief then, the consensus in 1953 of the major figures of the Manhattan Project, such as Robert Oppenheimer and Hans Bethe, not only reversed their 1945 views, but contrasts even more with almost every stereotype of the Mutual Assured Destruction (MAD) dogma which has dominated our elites, including our scientists, in the decades after the Cuban missile crisis.
Such views on the vital importance of limiting the catastrophe of a nuclear war strongly influenced high officials in the Truman as well as the Eisenhower Administrations, as National Security Council documents of that time show. Indeed, the belief in the importance of assuring the survival of civil society in a nuclear war was a basic tenet of elite opinion outside the government, as common in The New York Times and the rest of the media as its opposite is today. The present notion that we should avoid any element of defense or restraint in offense and should try to assure mutual destruction is the anomaly—a comparatively recent and perverse innovation. It was not, as is widely assumed, the basis for the second-strike theory of stable deterrence from the beginning.
In the intense emotions generated by the dispute over the H-bomb, both the majority faction of the scientists who opposed it and the important minority who favored the H-bomb tended to caricature each other’s views: one side accused the other of being for offense only, and in turn was said to be for defense only. The Panel of Consultants to the State Department rejected the argument that giving attention to continental defense was dangerous because "such a change would require a lessening of our attention to the development of our strategic air capability"; the panel said this argument was "based on the mistaken notion that we must have one or the other and cannot have both." The offense favored by scientists like Hans Bethe would "fit the weapon to the target and . . . attempt the best accuracy . . . to avoid unnecessary destruction."
It was the Soviet intercontinental ballistic missile (ICBM) tests and the Sputnik in 1957 which brought to an end this decade of effort by the Manhattan Project physicists to limit the catastrophic effects of nuclear war. The views they then adopted, reversing their position 180 degrees, paralleled a reversal in the views of the U.S. Navy and followed the development of similar European views which rationalized the spread of small nuclear forces to be aimed exclusively at population centers. These views helped shape the declaratory policy of Mutual Assured Destruction adopted with substantial incoherence by the U.S. government five or six years later. They have persisted as the prospects for discriminating military from civilian targets changed drastically for the better. The average yield of the nuclear weapons we might "fit to targets" was 15 times larger in 1957 than in 1982, and "the best accuracy" we expected then has seen revolutionary improvement. It is ironic that for about a decade, when discrimination among targets was most difficult, major figures of the Manhattan Project called for it most intensely, and now that it is more possible than ever, many oppose it.
Early in this period of emphasis on defending and sparing civilians rather than killing them, the decision to go ahead with the H-bomb briefly generated a new and rather different discussion of a global catastrophe leaving no survivors. For some, the possibility of an ultimate nuclear catastrophe came up in a more ambiguous and, in fact, more ambivalent form: it appeared that one might actually want to rig the possibility of a total nuclear catastrophe on purpose. Opponents of the H-bomb opposed it because of its enormous expected yields which made it possible for one warhead to destroy an entire large city, and which made it seem useful only for that purpose. Multi-megaton weapons seemed bad enough for the direct harm they might do, but Hans Bethe in 1950 raised the possibility that carbon-14 from exploding H-bombs might poison the air for 5,000 years and so make life impossible.
Bethe’s calculations did not satisfy Leo Szilard because "it would take a very large number of bombs before life would be in danger from ordinary H-bombs." Szilard had something else in mind:
What I had in mind is this: the H-bomb, as it would be made, would not cause greater radioactivity than that which is due to the carbon; but it is very easy to arrange an H-bomb, on purpose [my emphasis], so that it should produce very dangerous radioactivity.
Szilard did not estimate how many megatons of TNT equivalent yield the total number of H-bombs exploded might require in order to produce enough cobalt-60. (My colleague, Gregory Jones, has estimated roughly that it might come to 100,000 megatons—much more than all the megatons in the world’s nuclear arsenals and, therefore, not really very cheap.) I have stressed Szilard’s phrase "on purpose" because his concern was very different from the first scare about the A-bomb burning up the atmosphere, and very different from Bethe’s concern about the H-bombs. Szilard was considering rigging the bomb deliberately so as to make it possible to end life on earth with a moderate number of H-bombs. Szilard said that his Doomsday Machine (to anticipate Herman Kahn’s name for a similar conceptual device) had a practical utility and was likely to be adopted by both the superpowers:
You may ask . . . who would want to kill everybody on earth? Let us suppose that we have a war and let us suppose that we are on the point of winning the war against Russia. . . . The Russians and others can say: ‘You come no farther. You do not invade Europe, and you do not drop ordinary atom bombs on us, or else we will detonate our H-bombs and kill everybody.’
Faced with such a threat, I do not think that we could go forward.
Asked whether any nation would really be willing to kill all people on earth rather than suffer defeat, and specifically whether we would be willing to do it, Szilard answered:
I do not know whether we would be willing to do it, and I do not know whether the Russians would be willing to do it. But I think that we may threaten to do it, and I think that the Russians might threaten to do it. And who will take the risk then not to take that threat seriously?
Always a pioneer, Szilard in this response clearly foreshadowed the argument that was to become common among proponents of Mutual Assured Destruction, that we could threaten mutual destruction and even the end of the world (even though we would be unwilling to execute the threats), and that the threat would be enough for purposes of deterrence. This was the line of argument that eventually culminated in the "Deterrence Only" school of strategic thought.
When Herman Kahn invented a conceptual device he called the Doomsday Machine in 1960, he intended it as a reduction to absurdity of the view—increasingly popular in the academy after Sputnik—that targeting population was the cheapest and best way to deter nuclear attack. However, as was frequently the case with Szilard, it was not easy to tell whether his modest 1950 proposal for destroying everybody inexpensively was not also tongue-in-cheek. Szilard pointed out that his Doomsday Machine would make it cheaper to kill everybody in the world rather than only in enemy territory. The global apocalypse, in short, would come at bargain prices.
Here again, Szilard was ahead of his time. The perverse arguments that began to pop up in academia after 1958 for avoiding military targets and focusing on cities took the implausible line that threats against military targets led to exponential increases in spending on both sides, since military targets could be more cheaply multiplied than cities. Targeting cities, therefore, was the only way to escape an arms race.
Today many theorists of global disaster seem unable to decide whether threatening cities and assuring mutual destruction is an evil or a good thing. Carl Sagan, for example, at one point in his recent essay on the implications for policy of nuclear winter, seems to express my own concerns about apocalyptic threats: "To the extent that these are not credible, they undermine deterrence; to the extent that they are credible, they set in motion events that tend toward apocalyptic conclusions." Sagan, however, rejects weapons like "burrowing low-yield warheads," even though they would be much less likely to cause a climatic disaster, as "provocative"; "they are the perfect post-TTAPS first-strike weapon."
The jargon about "provocation" and "first strike" casually misuses the second-strike theory of deterrence. No single performance characteristic of a weapon, whether precision or warhead yield or speed of delivery, establishes whether a nuclear force has the capability for credibly performing a second strike, or a preclusive first strike, or both, or neither. In this context, such casual labels are merely code words for preferring large-yield warheads that do enormous collateral harm. Both low-yield, deep penetrating warheads and large-yield warheads exploding at the surface of the earth can destroy underground military targets. The difference is that the small penetrating warhead will do much less collateral harm, perhaps two orders of magnitude less.
In the intellectual confusion of contemporary MAD doctrine, it is the reduced risk of collateral damage, local and global, that is regarded as evil. And in spite of a welcome bow in the direction of abandoning apocalyptic threats, the basic thrust of many nuclear winter theorists is exactly the opposite of the earlier rejection of large H-bombs because of the enormous collateral harm they might do. The thrust, as foreshadowed by Szilard, is toward rigging things to make sure the harm would be global. On the other hand, the possibility of a global climatic disaster has recently made even some supporters of MAD threats doubtful about strikes against population centers. Now even they contemplate the need to avoid cities—or at least some cities.
But on the whole, theorists of nuclear winter, like supporters of MAD in general, continue to oppose the development of more discriminately effective offense and defense weapons or any preparations to respond to attack in ways that would confine destruction to military targets. That has influenced their choice of cases to investigate in ways that cloud their conclusions even more than the numerous uncertainties about the density of fuel near specific targets and the like in models of nuclear winter. They have selected cases calculated to create enormous collateral damage both local and global.
The many uncertainties that shroud nuclear winter come in several distinct kinds. The first has to do with whether, when and how an attacker, such as a Soviet planner, might choose to use nuclear weapons, and how many and what types of weapons he would use. He might choose to attack in the summer when concentrations of fuel are dry and most easily ignited and when his own crops are in their growing season and therefore likely to be affected most drastically. He may include in his initial attack many targets like steel mills that have no time urgency since they could not affect the course of a war for months or even a year, and do this even though their rapid initial destruction along with urgent targets would greatly increase the amount of smoke generated in an interval short enough to make nuclear winter effects far more likely.
He might decide to attack cities and other targets with high densities of fuel, such as oil refineries outside cities, and he may explode high-yield weapons at altitudes that would maximize the thermal pulse over combustible areas and so send smoke in huge quantities into the atmosphere. He may blindly barrage mobile or concealed military targets with little military but large environmental effect. And he may use multi-megaton weapons at or near the surface of the earth in ways that would maximize the chance of sending submicron dust in large quantities into the stratosphere.
On the other hand—far more likely—a Soviet planner might use nuclear weapons to accomplish some military purpose in the course of a war, and do it in a way that takes account of the fact that destruction extraneous to that purpose could cause a nuclear winter and make that military purpose idle. He could try to avoid these self-defeating effects. If he is involved, for example, in a conventional war on the critical northern or southeastern flanks of NATO, or in the Persian Gulf or in the center of NATO, and has suffered unexpected reverses, he might use nuclear weapons against selected targets whose destruction or paralysis could turn the tide of battle. He could do this perhaps by destroying or putting out of action, for the duration of the battle or the war, most aircraft and the maintenance facilities on main operating bases, munition stockpiles, defense radar and communications and the like; and he could block reinforcements from inside or outside the theater. Moreover, he could do this in a way that would least interfere with the movements of his own military forces and his other military efforts, and would also confine the generation of smoke or dust to levels well below the diffuse zone of uncertainty for severe global effects. Most precautionary measures taken to prevent harm locally to his military effort would also be useful in staying well below that zone.
A second sort of uncertainty concerns how the victim of an attack, such as NATO, might respond. If a Soviet attacker had used nuclear weapons with effects largely confined to military targets in a local theater of war, and if NATO had prepared no response except for devastating cities and risking a nuclear winter, it might not respond at all. Or, conceivably, it might respond in a way deliberately to assure mutual destruction and, incidentally, the ruin of the hemisphere. Or, far more likely, it would follow the Soviet attacker in restricting itself to measures that stopped key military operations but kept things from getting out of hand and destroying the planet.
There is the other case, in which the Soviets had launched an attack generating enough smoke and dust to have a substantial probability of bringing on severe global effects; NATO might then respond by generating still more smoke and dust and increasing the likelihood of even more severe effects and danger to the species. This would differ from the "tit-for-tat" response in the stereotype of mutual escalation. It would be tat-for-tat, NATO bombing itself in response to Soviet bombing of NATO. Or—again more likely—it might choose a form of response that would serve a military purpose but did not substantially further increase the probability of a ruin that would encompass the West as well as the East. Here too, boomerang effects may influence choice.
A third type of uncertainty has to do not with choice but with matters of fact. Some of these have been deplorably neglected but should yield to further empirical study, such as the density of fuel at various locations; related issues as to how the fuel would burn and generate various kinds of smoke and soot in varied circumstances may be harder to resolve.
All these first three sorts of uncertainty, those that involve the choices of the two sides and those that have to do with the local concentrations of fuel of various sorts, have to do with the amount of smoke and submicron dust which would be generated and lofted into the troposphere and stratosphere during a nuclear conflict.
A fourth sort of uncertainty—one which will be under investigation for many years—is more complex than the third category. It has to do with how the smoke and dust are likely to be transported vertically and horizontally in the troposphere and stratosphere, how clouds of smoke particles might form and be reduced by rain-out, how much solar radiation would reach the earth through the smoke clouds, and how much infrared radiation would escape and the resulting light and heat at the earth’s surface. The first-generation models of the atmosphere after a nuclear war were designed by experts on planetary atmospheres. They were more appropriate for studying nuclear war on desert planets like Mars than on the earth, most of whose surface is ocean. Current models indicate considerably smaller, though still substantial, effects for the huge exchanges.
Finally, there are the biological effects of possible patterns of change in temperature and light at the earth’s surface. Biologists and physicians have been among the most prominent prophets of a global nuclear winter. Their work on global effects, however, has focused only on cases even more extreme than the enormous baseline cases looked at by the Swedish study in Ambio, TTAPS and NAS 85, and so far has been supported by less evidence than the work on climatic changes.
Of these five types of uncertainty, the first and second—those that involve choice—have been least satisfactorily addressed. Yet they are of immense importance and can dominate the rest. The weapons and strategies that adversaries choose make quite a difference. And alternatives that can reduce smoke and dust even more dramatically have not been much explored. These first two sorts of uncertainties differ greatly from the others precisely in that they are a matter of choice. They are choices—partly independent and partly interlocking—made by the antagonists.
Nuclear winter theorists, however, tend to treat these uncertainties as if they were simply matters of chance uninfluenced by choice, like the collision of an asteroid with the earth or the impact of a comet that lofted devastating dust. Nuclear winter theorists treat antagonists as rather like asteroids and comets, or, at least so far as the application of intelligence is concerned, like the dinosaur that may have become extinct as the result of such a collision. They presume explicitly, at any rate, that the antagonists will make their choices of targets, methods of attack and timing without any intelligent consideration as to the likely implications of such choices for their own destruction by a nuclear winter. The NAS 85 study, for example, assumed that neither side would refrain from attacking military or economic targets located in cities in spite of the dangers of igniting their dense fuel. And, in fact, its baseline case involved explosions over 1,000 cities in proportion to their population—attacks in which each side’s explosions are well designed to contribute to its own destruction.
Attacks on population, or attacks which ignore collateral harm to population, of course, have had many advocates in the Western establishment. And even more members of the establishment consider that any use of nuclear weapons will end in devastating cities on both sides even if we try to avoid it. Nuclear winter theorists cite as justification for their assumptions statements not only by some Western strategists but by a good many former high officials—defense secretaries, chairmen of the Joint Chiefs of Staff and deputy directors of the Joint Strategic Targeting and Planning Staff.
What is novel in nuclear winter theory, what makes it capable of exhibiting with a particular clarity the incoherence and implausibility of much establishment doctrine, is that it assumes that each side will use weapons to bring about its own destruction not merely as part of a process of mutual "escalation," but directly with its own weapons. The rebound of one’s own weapons eliminates the middleman in self-deterrence. Even if nuclear winter should ultimately turn out to be a less substantial danger, it will nonetheless have served a useful function. Its proponents carry one step further the assumption widespread in Western elites that in a nuclear conflict neither side would persist in choosing to keep the destruction done by its own weapons within bounds short of self-destruction.
Nuclear winter theorists also make clearer some of the absurdities in the Western view of Soviet behavior. Even apart from nuclear winter, one need not suppose, as some members of our foreign policy establishment assume, that only "gallantry" or some courtly interest in Western welfare would lead the Soviets to place limits on their use of nuclear weapons. The Soviets have always had strong reasons of self-interest not only to be wary about using nuclear weapons at all, but to try, if they should feel that the risks of using them in the course of a war are less than the risks of not using them, to keep the risks from getting completely out of hand. It is absurd to suppose that the Soviets would totally disregard the risk of disaster to themselves. Yet that may be a canonical assumption about Soviet attacks.
Moreover, it is one thing to say that political and military leaders sometimes mindlessly take the most self-destructive course. It is quite another thing to suppose that one’s adversary will always either do nothing or mindlessly attack in a way that will do himself the most harm; and it is still another thing to recommend mindlessly suicidal behavior on our side, and to rig our preparations so that we could not respond against the source of attack without killing ourselves. When strategists rely on Mutual Assured Destruction, they assume intelligence can have essentially no influence at all.
Recent studies of a potential global nuclear disaster seem to generalize about the outcome of nuclear conflict—whether fought in a theater of operation in Europe or Asia; between the superpower homelands; with past, current or future weapons, of small or large yield, in modest or huge numbers; in accordance with present secret war plans or any future plans developed with the knowledge of these new global effects—and independently of whether the antagonists have prepared and try to restrict the damage done to each other or to themselves. In fact, the principal studies starting with a National Academy of Sciences report in 1975 (NAS 75) of the effect of nuclear war on the ozone, to the recent NAS 85 study on nuclear winter, take as their baseline case a conflict in which each adversary directs thousands or tens of thousands of weapons, many of very large yield, at locations and at burst altitudes calculated to do immense direct damage to the civil society of its antagonist; each thus also helps to bring on a global disaster destroying itself. The studies do not consider cases where either antagonist tries to confine damage to military targets or in any way tries to prevent damage to his own political, military or civilian resources.
The TTAPS baseline case assumed 5,000 total megatons, and 10,400 warheads with a yield between 0.1 and 10 mt. It assumed 1,000 mt would explode over urban or industrial areas, generating over 130 million tons of smoke. The remaining 4,000 mt exploding over non-urban land areas was calculated to generate an additional 95 million tons of smoke and to loft 65 million tons of submicron dust into the stratosphere.
NAS 85 assumed 6,500 mt and 25,000 nuclear explosions (the largest number of nuclear explosions in any of the studies); 3,500 of these, with a total yield of 1,500 mt, were assumed to be detonated over the largest 1,000 cities in NATO and the Warsaw Pact, generating 150 million tons of smoke. Explosions outside of cities were calculated to generate 30 million tons of smoke and to loft about 15 million tons of submicron dust into the stratosphere.
Are these "scenarios" of nuclear winter "reasonable" as is claimed? Not really. Neither the baseline cases nor the smaller excursions are. I have never been an enthusiast for the scenarios favored by many strategists and planners. They seldom have much relevance to the uncertain evolution of sequences of events likely in real contingencies. They seem more closely related to the movies or TV dramas that have given them their name. But the cases on which nuclear winter calculations have been based make little consistent sense even as scripts for movies. They hardly describe a sequence of events at all. Rather they list a collection of simultaneous explosions by the two antagonists well-designed to bring on a global disaster. In particular, such sets of explosions seem dubious as aspects of conflicts fought under the shadow of a nuclear winter.
The baseline cases are, first of all, huge. In fact, some may be infeasibly huge even if the contestants used the entire world arsenal and cooperated to detonate as many weapons as possible. The authors of these calculations seem only intermittently aware that in the normal course of a conflict some weapons would be duds or carried on vehicles that had some mechanical or operational failure preventing them from getting off the ground; some weapons would destroy other weapons (for example, weapons in stockpiles or weapons placed in vehicles that were hit on the ground or on their way to target), or both the attacking weapon and the weapons destroyed might explode at sea where they could do damage to naval targets without any substantial global effect on climate, and so on. The chairman of the panel that produced NAS 85 has suggested that, in assuming the explosion of 25,000 nuclear weapons (about half the world’s arsenal), the panel "didn’t want to take an extreme position," that it wanted to leave "plenty of room on either side" of the estimate. But it is more than doubtful that the two antagonists, even with much collaboration, using the entire world’s stockpile, could arrange to explode that many nuclear warheads of the required range and distribution of yields.
Second, the explosions do seem to require a collaboration of the antagonists. Take the countersilo part of the NAS 85 baseline: the United States and the Soviet Union each explode two bursts in the megaton range at the surface of the earth near each silo of the other side in order to assure that the missiles in silos will be destroyed. But with singular lack of success—since each side launches its missiles at the other side. In fact, this example illustrates a characteristic of these odd scenarios: attacks which are nominally directed at military targets seem to do very little direct harm to their military targets, but do manage a large contribution to global disaster.
Third, even if one were to assume that it is reasonable to direct a great many large-yield weapons at highly combustible cities, the choice of targets made by the pair of superpowers acting in unison seems rather bizarre. The Swedish Academy, for example, presumes that 15 nuclear explosions with a total yield of ten megatons will occur over each one of such cities as Manila, Jakarta, Bombay, Calcutta, New Delhi, Madras, Dacca, Sydney and Hong Kong. In a war between the United States and the Soviet Union? The Swedish Academy study in fact presumed that the two superpowers together would explode 173 mt in the southern hemisphere. The World Health Organization (WHO) topped that in its 1982 study on the biological effects of nuclear war: it assumed 1,000 mt would explode in the southern hemisphere.
Fourth, these baseline scenarios involve attacks on population centers on such an enormous scale that the direct destruction done locally seems almost unimaginably large. The Swedish Academy scenario would directly kill or maim over a billion persons; WHO’s scenario, over two billion. Professor Sagan makes much of the fact that these two billion casualties would be just the beginning: the global effects of smoke and soot might take care of the other half of the world’s population.
Moreover, the "small" or "limited" scenarios may be even more absurd than the baseline cases. The TTAPS 100-mt exchange is the one most widely referred to as showing that any nuclear conflict, even a "limited" nuclear war, would trigger a devastating global change in climate. The 100-mt exchange is also the basis for estimates of the "threshold" number of missiles that would trigger a drastic climatic change, and therefore the maximum number of nuclear weapons that could safely be left undestroyed and unreplaced out of the world’s stock of about 50,000 nuclear weapons—if we were to be sure that no combination of powers, no matter how mindless their behavior, could bring about a nuclear winter. But TTAPS’s 100-mt case, which involves one thousand 100-kiloton weapons exploding over the 100 most populated cities in NATO and the Warsaw Pact, is not "small." And it is certainly not "limited" in the relevant sense of illustrating the results of attempts to confine collateral damage to military targets.
Both sides, including whichever side may be thought of as striking first in this odd collaboration, ignore military targets—which makes this "limited" case even more absurd than the massive baseline scenarios. My colleague, Richard Brody, calculates that this 100-mt attack might cause as many as 300 million casualties directly and locally. This does not count the global effects which are worse for this 100-mt countercity attack, as TTAPS itself points out, than those of their 3,000-mt countersilo war. In fact, by changing the value of its fuel parameters, TTAPS derived about the same amount of smoke and soot (130 million tons) from this 100-mt case as from the 1,000 mt exploded over cities in its huge baseline case.
Professor Sagan has indicated the threshold suggested by the TTAPS 100-mt case in several ways: from 500 to 2,000 weapons; or one percent of the world’s nuclear arsenal (about 500 weapons); or one percent of the strategic offense warheads (about 180 weapons). However, the notion of a "threshold" here is quite deceptive—and not merely because there is no fine line nor even a well-defined band of uncertainty about the number of weapons that might trigger a global disaster. More important, the location of the diffused zones separating safety and disaster would depend on whether the weapons exploded were of high yield, detonated over cities or other dense concentrations of fuel, burst at high altitude or on the surface of the earth or well beneath it, and so on. So long as one assumes that nuclear weapons will be used mindlessly or deliberately to maximize the likelihood of disastrous global effects, the number of explosions would have to be extremely small to stay below the zone of uncertainty. The threshold numbers named by Professor Sagan are extremely small. No enforceable arms agreement is ever likely to reduce the world’s arsenal by the required 99 or more percent. (The superpower agreement that Sagan suggests would beat swords into plowshares by using the fissile material in bombs for nuclear electric power would actually spread plutonium quickly useable in bombs. )
Nuclear disarmament will never be 99 and 44/100 percent pure. The most extensive nuclear disarmament feasible will not permit the luxury of the mindless use of those nuclear weapons that may remain hidden. And disarmament far short of 99 percent is a long way off. Nuclear winter theorists propose such Utopian solutions because they reject the idea that an attacker could use nuclear weapons for some military purpose and yet confine the effects substantially to military targets. They do not calculate the outcome for the global environment of such a selective attack.
In fact, the critical unexamined claim made by nuclear winter theorists in drawing implications for policy is that a Soviet or other prospective attacker would deter himself without any need for a suicidal decision to respond on the part of his victim; that the attack itself would cause global harm engulfing the attacker even if the victim did not respond. If the claim were true it would fill a void at the very center of MAD doctrine.
Since the dogma of Mutual Assured Destruction persists in the West, it is worth spelling out clearly what makes it hollow:
(1) If a responsible political leader, contemplating a strike in aggression, expects that he (and his political and social order or his military force or whatever he values most) will be harmed vastly more than if he does not strike, he is unlikely to strike.
(2) That is the basic theorem of the second-strike theory of deterrence, phrased as it should be in terms of comparing the risks of striking and not striking rather than in terms of "acceptable damage." Where the harm he might suffer in striking is at the extreme, we say the aggressor will not attack because he does not want to commit suicide.
(3) But it is equally true and relevant that if a sensible political leader, contemplating a strike in response to aggression, expects that he and his entire political and social order will suffer a vastly greater disaster than if he does not respond, he is unlikely to decide in favor of responding. That follows from the same root assumption as the second-strike theory of deterrence. Where the harm he might suffer in responding is at the extreme, we can say he will not want to commit suicide. At any rate, he will not commit suicide in order to avoid death.
(4) If the aggressor can see that his victim has no response other than suicide, and that the victim is aware of that and says so all the time, then the aggressor may expect no response. If there is no response, he will suffer no harm. He may then find that his best way out of a palpable disaster, say, during a conventional invasion, is to launch a selective nuclear attack.
Some political leaders tend now to concede that a "deterrent" that promises no response is incoherent, but they cling nonetheless to the belief that it works and that it has worked for 40 years. But it has not. In the mid-1960s when Robert McNamara introduced the threat of Mutual Assured Destruction, he also made clear that if deterrence failed, the United States would respond with an attack designed to limit the damage, not to commit suicide. The drift toward not responding and the consequent erosion of the military balance and of deterrence is a quite recent phenomenon.
When the Soviets talk primarily for Western ears, they indicate that if they ever attack they will do it massively and indiscriminately, even if it means the end of the world—either to frighten us into believing that it is futile to prepare to use nuclear weapons even in response to their use of nuclear weapons, or to lull us into believing that they would never use them. Or both. Such Soviet statements are in good part disinformation. They are a main support for the drift toward not responding, toward "Deterrence Only."
When Soviet military planners write primarily to inform each other, they may demonstrate an interest both in using force "massively," that is "decisively," to accomplish a key military purpose, and in using force selectively so as not to defeat that or any wider purpose. ("Size would reflect only the desire to achieve strategic results. [It] can be as few as several score weapons" to deliberately "reduce unwanted and unnecessary damage." ) The development of their military forces confirms this double interest. They are increasingly capable of selecting some targets to be immobilized or destroyed with important military effect while leaving others essentially untouched. They have been moving toward more precise, lower-yield weapons. This ability to reduce unintended harm locally lessens also the chance of bringing on a global disaster. Moreover, while the Soviets show little interest in the purely symbolic, Western-style use of nuclear weapons "to demonstrate resolve" and "to send messages" in a process of "escalation control," their force development shows a strong interest in keeping the battle under their control.
Any Soviet leadership which we could deter by threatening disaster would also be deterred from an unlimited strike by a prospect that the rebound from their own weapons could, by itself, without our intervention, bring disaster to the Politburo, Soviet military forces and Soviet society. Soviet leaders would obviously prefer to extend their control without having to fight at all. If they have to fight, they would prefer to risk less and to win by conventional means. If they do use nuclear weapons, they would prefer to use them in a way that would not destroy their purpose in attacking. But in any case, they want to avoid suicide and the end of life in the northern hemisphere. Members of the Politburo do not believe in the hereafter. The Utopia the true believers have sought—and the privileged rewards of the nomenklatura—have to be realized here on earth. But advocates of MAD, in their extremity, defy common sense. In their zeal to show that the Soviets will always nullify any defense we can construct, some scientists prophesy that if the United States were to attempt any "serious" protection of its cities, a "likely response" by the Soviet Union would be "to target its missiles so as to maximize damage to the U.S. population," even though that would "pose a serious danger of triggering a climatic catastrophe (the nuclear winter phenomenon)."
Are the Soviets really possessed by so single-minded and pure a passion to kill our harmless bystanders that they would be willing to destroy themselves and any future for "communism" or for life itself in the Soviet Union? If so, they have an even better countermeasure to our ballistic missile defense. This one does not require them to penetrate our elaborate defenses at all: they could explode their warheads over their own cities in large enough numbers to bring on the death of the biosphere. Showing how hard it is for our advocates of MAD to beat Soviet efforts to make protection against Soviet attack seem hopeless, Izvestia recently printed a piece by Valentin Falin (former ambassador to West Germany) saying that the Soviets might very well counter our anti-ballistic missile defense in just that way: "No ABM options," Falin wrote ominously on December 14, 1984, "will change the fact that a precisely known quantity of nuclear devices detonated simultaneously on one’s own territory would have irreversible global consequences [emphasis added]." If the Soviets were really that crazy, the prospects for arms control would be even dimmer than the skeptics believe.
Quite apart from the dangers of a nuclear winter, the Soviets have compelling incentives to use nuclear force selectively if at all. They have recognized this interest, as can be abundantly documented both from recently available materials of the Voroshilov General Staff Academy as well as from a fresh evaluation of Soviet military writings during the last 20 years. Moreover, they are able to use nuclear force selectively and keep it under control, and have been greatly increasing this capability. I can only illustrate these points.
The most elementary concern the Soviets show for avoiding the indiscriminate use of nuclear weapons has to do, first, with local effects on their own military forces. They want to avoid "hindering the actions" of their "own troops," or having them "suffer heavy casualties" and "lose their combat capability," or having "their rear support area . . . suffer heavy damages" and their "communications routes . . . destroyed." And in particular they are concerned about "a break in the originally outlined plans for combat actions" and "an interruption in troop control."
Second, the Soviets would prefer a Soviet Western Europe intact rather than in ruins. "The objective is not to turn the large economic and industrial regions into a heap of ruins," but, among other things, "to sharply reduce the enemy capability to conduct strikes." Some "targets and regions should be left intact [to] strengthen the economic potential of our own country." After World War II, the Soviets transferred whole factories from occupied lands and many German scientific and engineering cadres to the Soviet Union.
Third, Soviet military authorities have recognized the stake of both sides in exercising caution. "The accumulation of nuclear missiles," according to a contemporary statement, has "reached such extremes that their massive use could turn into catastrophic consequences for both sides." This is not new; an earlier authority said that "the risk of destruction of one’s own government" is so heavy that the belligerents will use other means to attain their objectives or, if not, will "limit themselves to inflicting some selective nuclear strikes on secondary objectives."
Fourth, "in modern conditions, such [selective] nuclear attacks would primarily be the consequences of the expansion and development of a conventional war. . . ." The conventional balance in the center of Europe is already very favorable to the Soviets. If they met unexpected reverses there, they could make up for them with a rather small number of low-yield nuclear weapons, precisely delivered and with quite confined damage. This sort of militarily decisive contribution at the margin differs greatly from the "demonstrations of resolve" that figure in the Western strategic debate.
Fifth, the conventional war leading to the use of nuclear weapons may take place in a weakly armed but critical flank of NATO, such as southeastern Turkey or its unstable neighbor Iran, and may be confined to that theater of military operations. That is not the canonical NATO scenario. But it is more likely to divide the allies and to give some the incentive and opportunity to opt out. The Soviets would risk less, and the autonomy of Europe would nonetheless be at stake. In World War II, they did not attack all their enemies at once but, successively, a weak Poland under attack also by Hitler, little Finland, the Baltic states, Bulgaria, then a Japan already beaten by the Allies. Hitler, a risk-taker, attacked the Soviets. Their military authorities do not exclude "the limited use of nuclear means in one or several theaters of military operations." Southeastern Turkey and the Persian Gulf are part of the Soviet southern theater.
Sixth, even an intercontinental attack might be directed at disrupting the scheduled deployment from the United States of the military airlift, tactical aircraft and ground forces most urgently needed for reinforcing NATO. The essential facilities involved are few, poorly protected and distant from any large population center; disrupting them would require only a modest number of air-burst weapons of moderate yield. Yet that could alter the "correlation of forces" decisively for a Soviet combined-arms invasion of Western Europe. Such an attack could generate a thousand times less smoke than the baseline nuclear winter scenarios and essentially no stratospheric dust. The Soviets are greatly improving their ability to deliver weapons at long range precisely enough to employ low-yield nuclear weapons for such a purpose. (In fact, Marshal Nikolai Ogarkov has stressed that the ongoing revolution in precision at very extended ranges will permit the use of conventional weapons deep in an adversary’s homeland for some "strategic" purposes. )
The same incentives that impel the Soviets to be able to use nuclear weapons selectively impel them to keep the use of nuclear weapons under continuing control. In fact, the notion that Soviet political leaders would casually let nuclear weapons slip out of their hands is even less plausible than the notion that they might use them indiscriminately. For the Soviets have often been as brutal as they now are in Afghanistan; they might use nuclear weapons without much discrimination if it were a question only of compassion rather than their most elementary self-interest. Letting things get out of their political control, however, control that could decide the life or death of the party and their political order, is quite another matter. It has nothing whatsoever to recommend it in the Bolshevik canon.
The Politburo does not encourage spontaneity in the use of nuclear weapons. Nor is there any evidence that, after a few nuclear weapons were used, the Politburo would allow everyone in physical possession of them to fire at will. The Soviets will, of course, use threats of uncontrollability. We have seen some outstanding examples. But the threats were quickly followed by a demonstration that the Soviet political leaders had no intention of letting things get out of control.
During the Cuban missile crisis, for example, when confronted with the American threat to intercept, board and search a Russian freighter, Nikita Khrushchev asserted that this "would make talk useless" and bring into action the "forces and laws of war"; it would have "irretrievably fatal consequences." However, when it became plain that President Kennedy was nonetheless proceeding to intercept Soviet freighters, Khrushchev changed his tune from ominous notes of uncontrollability to hasty reassurance that there was no occasion for the United States to act desperately. Khrushchev made clear he had no intention of deliberately responding in a massive way, of letting the laws of war take over, or of letting those romantic Cubans precipitate a nuclear conflict. Good, solid, stolid, sensible Russians were guarding the safety catches on the missiles in Cuba:
The means which are located on Cuba now, about which you are talking and which as you say concern you, are in the hands of the Soviet officers. That is why any possibility of accidental usage of those means, which might cause harm to the United States, is excluded.
After the crisis, Khrushchev made a good deal of the fact that he had "shown restraint" to save the world from imminent peril.
The Soviets clearly would want to maintain control. But could they? Arguments to the contrary proceed like arguments about collateral damage by assuming what they are supposed to be proving. In fact, the argument reduces to an unsupported or circular assertion about collateral damage and an equally circular assertion about "pressures for escalation." Collateral damage from an initial attack, it is said, would be essentially the same as from a direct attack on cities, now or in the future, no matter with what weapons or what targets. But evidence cited always seems to be based on calculations which assume the combatants do not try to reduce collateral damage. They simply ignore attacks designed both to achieve a military purpose and to reduce unwanted harm. Much the same can be said of assertions about "escalation." These are usually spiced by statements that those in possession of a weapon would have to "use it or lose it." But using it could be a lot worse than losing it. On the nuclear winter theory, using a missile force indiscriminately might mean losing the hemisphere. All the enormous pressures to keep destruction within less than suicidal bounds are totally neglected.
Everything that we know about how the Soviets store and handle their nuclear weapons suggests that control is both extremely tight and well protected. For good reason. First, even more than the Western powers, the Soviets have to be concerned about the seizure of nuclear weapons by dissidents who might use them to help establish their independence from the Soviet government. The Soviets are in potentially hostile territory, not only in Eastern Europe, but in some parts of the Soviet Union. Second, the leadership is not unaware of the dangers of the "use it or lose it" syndrome. Marshal Ogarkov, to take one example, has rejected policies of "launch on warning," that is, launching one’s own missiles on the basis of electromagnetic signals that enemy missiles are on their way to target. Third, several Soviet military writers have made clear that not only is the decision to use nuclear weapons "the exclusive prerogative of the political leadership [but] it is the political leadership . . . not the military leadership . . . who select the primary targets and the moment of the infliction on these strike targets."