The startling technological advances of the last two decades, notably the marriage of nuclear and thermonuclear weapons with long-range delivery systems, have brought about fundamental changes in the nature of military power. The range and carrying capacity of strategic delivery systems, the speed with which they can be brought to bear, and their awesome potentialities when employed in mass all give to the country possessing them a destructive capability unequalled in history. Moreover, while it is possible to blunt such attacks and to ameliorate their effects, there does not at present seem to be any way of preventing one nuclear power from devastating the homeland of another should it really desire to do so.
The achievement by both the United States and the Soviet Union of such capabilities has initiated a fundamental re-examination of the role and the purpose of force. While this re-thinking is far from complete, it has apparently led to increased inhibitions against the large-scale application of military power and to a prudent desire to achieve national objectives without the use of force—so far as this is practicable. In turn, concern over the possible consequences of a nuclear war has focused attention upon the potential rather than the actual use of force, upon deterrence rather than defense as a means of securing national interests.
In itself the concept of deterrence is not new. Flavius Vegetius said centuries ago, “Let him who desires peace, prepare for war.” What is new is both the tendency to rely very heavily, if not absolutely, upon the effectiveness of the threat to employ deterrent forces, and the attempt to structure these forces so that they can inflict a given—and therefore calculable—level of damage upon another country.
Webster defines deterrence as “that which serves to hinder or to prevent action by fear of consequences, or by difficulty, risk, unpleasantness, etc.” As applied to military policy, this usually means discouraging an enemy from initiating military action (or from taking other steps which may lead to armed conflict) by posing the prospect that the costs or consequences of such actions might outweigh any possible gain. This prospect can be direct and explicit—as witness Secretary of Defense Robert S. McNamara’s statement that the United States could, even after absorbing a Soviet surprise attack, launch a retaliatory strike “fully capable of destroying the Soviet target system . . . ”—or it can be indirect and implicit, as when the Secretary of State announces that the United States will “fight” to maintain its “rights” in Berlin. In either case the policy involves not only a presumed capability to take effective counteraction under certain (more or less) prescribed circumstances but the understanding by the enemy that this is (or might be) so, and his consequent decision that the risks and the possible costs of taking a particular step exceed its potential gains or advantages.
One of the most authoritative statements of U. S. deterrence policy was that made by President J. F. Kennedy in his 1961 budget message:
Our strategic arms and defences must be adequate to deter any deliberate attack on the United States or our allies—by making clear to any potential aggressor that sufficient retaliatory forces will be able to survive a first strike and penetrate his defenses in order to inflict unacceptable losses upon him.
Leaving aside for the moment the admittedly difficult problem of rendering credible our intent to respond to attacks on our allies, when such response involves initiating a thermonuclear war, we should consider the task of designing and developing those forces required to hinder or prevent aggression “by fear of consequences” or, as Mr. Kennedy put it, by “their ability to inflict unacceptable losses. ...”
At first glance, this task would seem simple in concept (if difficult in execution): determine the level of losses which is “unacceptable,” calculate the number of nuclear weapons on target required to inflict such losses, allow for duds and misses, and for delivery vehicles which abort or are destroyed by enemy action, estimate the number of missiles and aircraft which could survive enemy attack under the worst of circumstances, add a safety factor, and presto! As any analyst knows, however, a vast number of variables and uncertainties affect such calculations, beginning with the determination of what constitutes “unacceptable losses.”
This seemingly simple and straightforward determination is in practice complex and subjective, since it depends in large measure upon the importance attached to the accretion of power, to the preservation of human life, and to the maintenance of peace as an end in itself. Different people not only will give differing weights to such aims, but also will tend to disagree over the value-judgments of those whose actions they are seeking to deter. For example, some analysts argue that the Soviet Union, which lost 25 to 30 million people in World War II, would be willing to accept proportionately heavy losses in order to eliminate the United States as a major power; others argue from the same premise that the Soviet Union, having once suffered such losses, would be unwilling to suffer them again, and consequently would be deterred by the prospect of many fewer casualties. Similarly, estimates of the casualties the United States might be willing to suffer in a nuclear exchange have ranged from two million to almost 100 million—although the question is admittedly posed in the context of an extreme Soviet provocation and not in terms of a deliberate U. S. decision to eliminate the Soviet Union as a viable entity.
Even assuming that a level of damage presumably sufficient to deter aggression could be determined precisely, how can we know that our estimate of this level will be the same as that held by a potential enemy, who has values, interests, and attitudes differing from our own? Moreover, how can we be sure that his views tomorrow will be the same as those he holds today? Levels of damage sufficient to deter attack are not immutable, but may change with the perceived threat to the national existence or to the vital interests of one power or the other. Indeed, under some circumstances, even assured knowledge of damage levels may not have sufficient emotional impact to alter decisions for peace or war, particularly since it is difficult to comprehend —and feel—all the implications of a massive nuclear attack.
Moreover, even if a precise figure, or a narrow range of figures, for damage levels could be set, this would not automatically determine the size of strategic strike forces. For one thing, these forces have missions other than retaliation against the enemy, missions ranging from the destruction of specific military targets to the back-up of diplomatic demarches designed to secure U. S. interests in areas such as Berlin. For another, there is no direct and assured relationship between the level of damage and the retaliatory forces required to inflict it. Despite the attractive simplicity of many calculations, there are dozens of variables whose potential impact on force requirements can be assessed only within wide ranges of probability. The selection of those variables which are to be considered, the probabilities assigned to them, and the weight attached to the resulting conclusions depend on human judgments, which, like all such judgments, are subject to errors, imprecisions, and biases. Thus at best, estimates of forces needed to inflict a given degree of damage can only be uncertain and approximate, while the relation of these forces to the level of damage which presumably will deter an enemy attack is even more remote and indeterminable.
The premise that one country can be deterred from following a given course of action by fear of another country’s counteraction has other weaknesses, in that it requires the individuals or the nations involved to base their decisions upon the relative utility of the choices afforded them. The determination of utility is a difficult task, even in a simple two- person game with complete communication and a desire for co-operative action. To suggest that it is possible for a nation to reach accurate and objective assessments of the relative utility of various courses of action is stretching the premise of deterrence almost beyond credibility. As Sidney Verba, author of “Assumptions of Rationality, and Nonrationality in Models of the International System,” points out:
In order for a decision-maker to maximize a particular value or set of values, he must be aware of his own values and be able to order them in terms of their significance to him. Such clear self-awareness is rare. This is true not merely because values conflict—because peace may conflict with prosperity or defense may conflict with deterrence—but because individuals do not have a clear set of value preferences that exist independently of the situation and can be matched against a variety of alternatives to see which gives the best value outcome.
More importantly, effective deterrence presumes that the enemy arrives at the same set of value judgments as you do, i.e., that the utility to him of a given action does not outweigh the risks and potential costs which you have posited in order to deter him from taking that action. Such an assumption tends to impute to another power the same scale and weight of interests which motivate oneself— or at least to imply that one can accurately predict the outcome of that power’s assessment. While predictions of another country’s behavior and responses are possible, in fact essential to the conduct of foreign affairs, these predictions are rarely quantifiable, seldom assured, and frequently wrong. To give only one example, the United States apparently did not believe that the Chinese Communists would intervene in the Korean conflict.
It is, of course, true that the nation seeking to deter another is not concerned with the enemy’s objective evaluations of utility, only with his subjective reactions and final decisions. However, if pushed to its utmost, the concept of deterrence presumes that avoidance of nuclear war is an interest which overrides all others—and this is not necessarily so.
In the first place, both the United States and the Soviet Union have other interests and objectives which may conflict with the desire to avoid a nuclear exchange. For example, the United States apparently regards the preservation of a non-hostile Western Europe as so essential to its national well-being and long-run security that it is prepared to wage nuclear war rather than to permit that area to come under Communist control. Even where interests are not so demonstrably vital, they may be regarded as worth fighting for. Thus economic goods such as Middle East oil, places like Berlin and Quemoy, and even symbols such as national prestige may have an importance in the eyes of one power far beyond their intrinsic value as judged by an impartial observer. In fact, if history teaches us anything it is that the causes of war are frequently disproportionate to the investment of national resources and the losses suffered during the conduct of the war..
In the second place, the question of avoiding nuclear war is seldom posed directly or alone but usually arises in connection with a conflict of objectives or a clash of interests. Under these circumstances the issues at stake may assume a symbolic importance disproportionate to their actual value, and influences such as domestic political pressures, the desire to maintain solidarity with one’s allies, or the felt need for “standing firm” may lead to decisions which risk nuclear war. Although fear of nuclear war may induce prudence, it is by no means certain that prudence will always prevail over these other influences, especially where the issue is not the initiation of nuclear war but the risk of war and where this seems a lesser evil than the sacrifice of national interests.
This is not the only respect in which subjective elements enter into a policy of deterrence. Most theorists and many practitioners argue that deterrence depends upon a “calculus of risks,” in which the aggressor (and deterrer) weigh the expected value of a given action, the costs of possible counteractions, the probability of each counteraction occurring, and the results of these counteractions in terms of further costs, risks, and benefits. Although such an hypothesis provides a useful model for the study of decision-making, it is (as the model-builders themselves acknowledge) both over-simplified and too abstract. How, for example, does one accurately weigh the influence upon Soviet foreign and military policy of continuing struggles for power within the Soviet political elite?
Even if the model did approximate reality, and if meaningful values could be assigned to given actions, there are a number of other matters to be considered before one can argue that there is a determinable rationale and a predictable logic to deterrence. The first of these is that, while the rational process of means-ends calculation can be described, it is doubtful whether it has meaningful application outside purely intellectual areas such as mathematics. In human affairs, limitations of knowledge, personality problems, conflicting values—and even differing interpretations of what is rational—all obstruct the process of rational decision-making and affect the resultant judgments. Moreover, there is no such thing as a wholly “rational” man, who can consciously and objectively evaluate information, weigh risks, calculate possibilities, and assess consequences. For one thing, each of us is in some measure a creature of his environment. As Gabriel Almond puts it:
The political behavior manifested within cultures, subcultures, nations, social classes, and similar groups is not only to be understood as rational effort directed at objective problem-solving, but as affected by these largely unconscious, culture-bound “views of the world,” and the emotional tonous [sic] of the social environment.
Further, both individuals and groups have views of the world which tend to influence their thinking-processes, their perceptions, and their judgments. Nikita Khrushchev’s image of America was such that he apparently saw no difference between Messrs. Nixon and Kennedy, whom he characterized in 1960 as both representatives of big business, and “as alike as a pair of boots.” Finally, since some judgments affect an individual directly and personally, he tends to inject himself and his own interests into these judgments, so that, according to Marbury B. Ogle, he is “unable to arrive objectively at his decisions in these ‘personal cases.’ ”
As every student of international affairs realizes, these factors do have a marked influence upon the behavior of nations, and upon the ability of one country or another to calculate in advance the costs, the risks, and the benefits of a particular course of action. Raymond A. Bauer, one of our most perceptive analysts, has admirably summarized the difficulties of applying to U. S.-Soviet relations the assumptions of rationality which bulk so large in many discussions of deterrence:
Game theory is based upon the assumption that one’s opponent has complete information and has calculated his strategy with complete rationality. These may be optimal assumptions for setting strategy in a competitive situation, but they are scarcely a reasonable description of conditions that exist in the real world. Yet, years of reading American analysis of Soviet behavior and Soviet analysis of American behavior, have convinced me that each party attributes to the other a degree of omniscience and omnipotence that he knows is manifestly impossible in his own situation.
A further consideration in the practice of a deterrent policy is the effect of uncertainty- Even if due allowance is made for subjective and non-rational factors in the “strategic calculus,” a potential aggressor is confronted by a wide range of uncertainties which may affect the outcome of his decision. These arise in part from the sheer inability of human beings to store, analyze, and digest the information required for anything but the simplest choices in decision-making, and in part from the fact that information may be inadequate, variables too numerous to examine or not easily quantifiable, and decisional methods simply not advanced enough to assure reasonable certainty of outcome. They also derive from the wide range of alternatives which should theoretically be considered before taking action. During the first move of a chess game, the number of possible strategies open to Black is 1026, a number so large that if 100 strategies were written on a single page, and 100 pages stacked into a one-inch pile, the stack of paper so filled would be 40,000 light years thick!
In real life, the range of responses open to an opponent may be considerably narrower, but the difficulties of thinking through the possible outcomes of each response are immeasurably greater than in chess. Moreover, some of the significant factors are subject to such great variations as to render calculations suspect. As P. M. S. Blackett points out in “Critique of Some Contemporary Defence Thinking,” estimates of the effectiveness of first-strike forces, and consequently of the levels of damage which might result from retaliatory attack, can vary widely with only minor changes in assumptions—and the probability that any particular assumption will be valid is essentially indeterminate.
Uncertainty also extends into estimates of the consequences of lesser military actions and extreme provocations. Even in those situations where the risk of a nuclear riposte is deemed remote, there is always the possibility that the unpredicted may come to pass. This is particularly true since there is a tendency on the part of both the great powers to regard the boundaries between East and West as sacrosanct, and to extend the protection of the nuclear deterrent to areas which are of marginal utility or are indefensible per se. Such coverage, even when implicit rather than explicit, commits the prestige of the power concerned, engages its good faith, and affects the credibility of other warnings, so that a potential aggressor can never completely rule out the possibility of a seemingly non-rational rejoinder to severe pressures. In short, ambiguity concerning one’s intentions may be almost as effective as an iron-clad guarantee—as well as more prudent, should deterrence fail.
Intentions, however, whether precise or ambiguous, require communication and perception, which raises another problem. The process of communication between two persons is an imperfect one—how much more that between nations? When one country increases its military budget in the interest of “peace,” another may deem this evidence of aggressive intentions. When one country alerts its strike forces to reinforce the deterrent, its opponent may interpret this as a prelude to war—and be tempted or panicked into launching a pre-emptive attack. Further, measures one country regards as stabilizing and reassuring may convey quite an opposite impression. The fleet ballistic missile submarine, which is widely touted in the United States as a non-provocative second strike weapon, is apparently regarded by the Soviet Union as an instrument for surprise attack, since it can proceed stealthily to within range of Soviet targets and launch its Polaris missiles without warning.
As just illustrated, the difficulties of communicating intent are aggravated by differences in outlook between individuals and by correspondingly different patterns of perception and reception of communications. Since these perceptions are further affected by the culture, the world outlook, and the myths of different nations, communications between these nations are subject to additional warping and misunderstanding.
This difficulty is intensified by the fact that, when individuals are under stress, they tend even more to assess information in terms of their preconceptions, to narrow the range of alternatives laid out for examination, and to act on the basis of stereotyped patterns of conduct, which may not be at all applicable to the situation at hand. For instance, one study of behavior in the period July-August 1914 has indicated that in time of crisis national leaders may disregard relative military capabilities, become obsessed with evidences of hostility, and on this basis initiate courses of action which completely fail of their desired intent. In part at least, such behavior reflects a further diminution of the “rational” process resulting from the physical and emotional strains of crisis situations, but it also shows the difficulty of communicating intent, even between allies.
The argument has been set forth above that it is very difficult to determine precisely the kind and level of forces needed to deter even a direct and premeditated nuclear attack, in part because it is impossible to say with any degree of certainty what threatened punishment would suffice for deterrence. There is no guarantee that the values held by a potential enemy, or the choices among alternatives which he might make, will correspond with one’s own. As the phrase “better dead than Red” suggests, there are some values which presumably would lead a country to initiate war even when the consequences might be disastrous. Further, I have maintained that there is no such thing as a wholly “rational” man, since, as decision-makers, every one of us is influenced by his cultural environment, his world outlook, and his ego-involvement, all of which affect the thought process and color the judgments reached.
It has also been noted above that the maintenance of a deterrent posture is both helped and hindered by other limitations on human thought processes. Thus, inability to store and utilize available information, to consider the full range and the complete implications of alternative strategies, and to weigh accurately significant variables whose probability of occurrence is indeterminate, all serve to create uncertainty about the outcome of military adventures and to render reasonably credible even ambiguous threats of reprisal or of counteraction. Conversely, problems of communicating intent necessarily arouse concern over the effectiveness of deterrence, particularly in time of crisis; without either side planning or desiring it, both may find themselves at war. And this prospect is heightened by the fact that strategic strike forces are designed not merely to deter direct attack but to support national interests in a variety of ways, not all of which are compatible with stability of deterrence.
These findings would indicate that the continued effectiveness of a deterrent posture depends at least as much upon an understanding of human nature as upon “objective” assessments of relative military capabilities—which, after all, can never be either precise or certain. Although man may be a rational animal he is also, as Aristotle said, a political animal. Any view which fails to take into account the political—and the animalistic—nature of man will necessarily give a false perspective of deterrence and misleading promises of stability.
This is true even in the determination of force levels and force structures. The uncertainties, ambiguities, and irrationalities intrinsic to the conduct of human affairs cast grave doubts upon the premise that we can define the requirements of deterrence and then construct a military posture which will completely and unequivocally satisfy those requirements. Thus I, for one, would question the theory that the stability of deterrence can be assured by x number of relatively invulnerable strategic delivery vehicles, whether x equals 50, 500, or 5,000. Moreover, I would argue that the ranges of uncertainty in both calculations of military capabilities and estimates of human behavior render suspect any judgment that the stability of deterrence depends on undertaking (or not undertaking) a particular military program, whether it be nuclear testing, construction of the RS-70, or production in quantity of Nike-Zeus.
This uncertainty concerning the impact on deterrence of changes in military postures has obvious implications for efforts to achieve stability of the military environment through the control and reduction of armaments. Since it is impossible to evaluate objectively and specifically the relative utility of particular weapon systems or the effects of changes in force levels, the parties to the negotiations find it difficult to agree on “balanced, phased, and safe-guarded measures ...” for disarmament, much less to relate these measures in such a way that they “will not affect adversely the security of any state, whether or not a party to an international agreement or treaty.”
Thus, even aside from the differences resulting from clashes of interest and diverse military requirements, it is futile to expect that the great powers will readily agree on formulae for reductions in force levels, for the destruction of nuclear stockpiles, or even for the location of inspection points, however “scientifically valid” these formulae may seem to one party or another.
Perhaps the greatest relevance and importance of these findings is with respect to the conduct of international affairs. If, as I have maintained, stability of the military environment is affected as much by political factors as by military ones, then deterrence can be successful only if antagonists having incompatible goals and values agree to “live and let live.” Neither side can hope, under the shield of nuclear deterrence, to eradicate the other by political action or to nibble away at his vital interests by various kinds of military operations, lest these activities generate frustrations, emotions, or political and psychological pressures which could precipitate a disproportionate military reaction, perhaps even a thermonuclear strike. Nor can either side count upon using its military power to preserve the status quo, since changes will take place despite the wishes of the great powers and under circumstances in which the level or balance of strategic nuclear forces may be utterly irrelevant—as is evidenced by the splitting off of Yugoslavia from the Soviet bloc or the coming to power of Castro in Cuba.
These conclusions about the nature and the limits of stability are not intended to invalidate the concept of strategic deterrence. As Albert Wohlstetter points out, deterrence, with all its inadequacies, “is not dispensable.” Such conclusions do, however, have important implications for the conduct of a deterrent policy by the United States and, hopefully, by the Soviet Union as well.
The first of these is that the military basis for a policy of deterrence, as it is now conceived, depends upon the validity of computations which are essentially incalculable. Insufficient recognition has been given to the imprecisions and the uncertainties inherent in determining requirements for deterrent forces. Although no one would deny the utility of strategic models and computerized games as aids to decisionmakers, they cannot and should not be expected to provide all the answers. They not only suffer from inherent limitations in methodology but are largely inapplicable to many weapons systems and types of units, such as amphibious forces.
Secondly, deterrence has been considered too largely in terms of differences in military capabilities, despite the fact that such differences are, so far as strategic deterrence is concerned, largely inconsequential. As Secretary of Defense McNamara has testified, even very elaborate and costly defensive and counterforce programs, on top of the present U. S. capabilities, “could not preclude [U. S.] casualties counted in the tens of millions.” This argues for greater consideration of the political and psychological implications of military measures, as well as a better appreciation of the limited applicability of military estimates and calculations.
Thirdly, too little attention has been paid to the constraints on national behavior imposed by a policy of deterrence. If neither side wishes to fight a thermonuclear war, then both must be chary of undertaking movements, actions, and measures which risk such a war—notable among them, overt military operations in areas important to the other. Although theoretically such operations may be conceivable—indeed, may even be justified by hypotheses concerning the calculus of deterrence—their initiation could generate political and psychological pressures which would invalidate all concepts based upon rational consideration of costs, risks, and benefits.
Finally, there has been a tendency in some quarters to assume that only unyielding obduracy on all points at issue will maintain the creditability of the deterrent. However, as the recurring crises over Berlin illustrate, accommodation of conflicting objectives and adjustment to change are essential to the long- run effectiveness of a deterrent policy. Neither side can afford to stake its all upon the maintenance of the status quo, lest in the process it precipitate the very conflict it is trying to avoid. In the final analysis, the stability of deterrence depends upon walking the thin line between “appeasement” and inflexibility, between the sacrifice of vital interests and the assertion that every interest is vital.
In short, the results of the “strategic calculus” would be more valid if the political and psychological factors which are sometimes slighted were given due weight. This requires that the “rational” decision-maker give prudent consideration to the weaknesses, the foibles, and the illogicalities of an opponent, and not push him too far on the basis of some presumed military advantage. It suggests that the policy and practice of deterrence be based less on essentially irrational “logical” processes and more on what I have, to borrow a phrase, called “the rationality of non-rationality.” If this is done, then perhaps deterrence may have a longer future than any of us might otherwise envisage.