Deterrence Resurrected:

Revisiting Some Fundamentals


From Parameters, Summer 1991, pp. 13-21.

Of recent years several cottage industries of deterrence studies have resumed active life following a hiatus of several decades. Specifically, academics have endeavored to reduce deterrence to a sub-specialty within political science; defense analysts have been seeking quasi-scientific truth about "stable deterrence;" strategic and moral philosophers have been going "back to basics;" and the defense community writ large has been wondering whether its long-standing deterrence creed was not really overly Soviet-looking in orientation and evidential base, and certainly unduly nuclear in inspiration, application, and assumed context.

The Cold War may be over, but the United States still has a strategic relationship with a very heavily nuclear-armed Soviet superpower. The political future of the USSR is unknown, indeed is unknowable, but the historical and cultural material from which that future will be fashioned does not provide much fuel for optimistic speculation. All things are possible, but the apparent reversion of recent months to authoritarian rule and the collapse of anything seriously resembling perestroika probably should not be read as only a temporary setback. The Soviet Union currently coercing the Baltic republics, evading some of the key terms and purposes of the CFE (Conventional Forces in Europe) treaty, and modernizing its strategic nuclear forces comprehensively is indeed the real Soviet Union.

Mikhail Gorbachev, on the evidence to date, is not committed to the political, economic, and social reform of the USSR as an end in and of itself. Instead, he is committed first to the preservation of his personal power, and second to the protection of the core values of the Soviet patrimony which he inherited. Whatever Gorbachev released, or has appeared to release, occurred as a result of the miscalculation that the devolution of power would be slow and only partial (in east-central Europe and the peripheral republics at home). The Soviet Union, either with Gorbachev as the reforming-then-repressive czar or with Gorbachev's replacement, assuredly will continue to pose a superpower level of strategic competition for the United States. Soviet American relations in the future may be less antagonistic than was the case for many of the years from 1946 until Gorbachev's accession to power in 1986; nonetheless, the United States has a prospectively permanent need of policy, strategy, and military posture to help manage the security relationship with a Soviet/Great Russian superstate.

What follows is a series of propositions on deterrence, along with critical commentaries. Their purposes are to help people to think appropriately about deterrence and to distinguish strong from weak arguments.

Propositions on Deterrence

. The requirements of deterrence are indeterminate or, at best, very uncertain.

Judgment is possible and necessary, but not quantified judgment. Not only is it difficult if not impossible to calculate in a quantified way what will or should deter, but even the intended deterrees will not know how much of a politically and technically credible threat it would take to deter them. Deterrence, or perhaps deterrent effectiveness, is not a quality that can be purchased discretely and reliably by incremental amounts. This reality tends to be obscured by the fallacy of making reference to the deterrent, as if a force posture itself deters. The US strategic nuclear triad is the agent of a deterrence policy, but candidate deterrees have to choose to be deterred. In a deterrence relationship, intended deterrees can always decide that they are not deterred, a situation which contrasts sharply with a relationship of physical coercion wherein the enemy's behavior is controlled directly by our forces.

An intended deterrent may not deter if the intended deterree believes we will not fight, or will not fight very hard (regardless of the size of our forces); if he is confident (however foolish such a belief) of securing military victory; if he believes (incorrectly) that war is absolutely inevitable anyway; or if he is totally or substantially indifferent either to threats in general or to the particular threats he receives from us.

Even if our deterrent policy is nominally effective, we may trip accidentally or inadvertently into war, perhaps courtesy of the catalytic action of a third party. Further, individual leaders and leadership groups can be subject to mood swings---optimism/ pessimism--and also are liable to redefine policy objectives (the "stakes") quite abruptly. In short, an adequate deterrent at 0900 hours may not be adequate by 1600 hours.

In sum, an adequate deterrent, let alone an enduring condition of stable deterrence, cannot be calculated with precision. Deterrence is a realm of educated guesswork. Scholars talk of the "calculus of deterrence"[1] and of "risk/benefit analysis," but those are just overblown figures of speech.

. To deter, an ounce of will is worth a pound of muscle.

This proposition conveys an important truth. It does not, however, provide an adequate basis for development of an attractive theory or policy of deterrence. There are at least three outstanding problems.

First, military muscle can be acquired and sustained much more reliably than can political will. Will may indeed be to muscle as 16 is to one, but defense planners cannot know that their political leaders, or their country more generally, truly would generate and demonstrate the necessary will when and for as long as it would be needed. Thus, in the early 1990s the United States can plan and invest in a reasonably well balanced strategic-forces triad for the beginning of the next century--START or no START. But, if one takes the will-to-muscle proposition as a candidate guide to defense planning, what would be its implications? What kind and quantity of strategic force posture generates perceptions abroad of the requisite "will" if the force posture is divorced from serious military designs for campaign plans?

Second, it is not sufficient to have what we think ought to be a deterring level of political will. That will needs to be perceived for what it is by the intended deterrees. The stuff of which wars are made includes a determination to fight which is underappreciated abroad. It would be difficult to improve upon Herman Kahn's observation that "if we want to have our strategic forces contribute to the deterrence of provocation, it must be credible.... Usually the most convincing way to look willing is to be willing."[2]

Third, even a credible determination to fight might avail little if the quantity and quality of combat power threatened falls short of some critical threshold of effect as seen by the intended deterrees. In 1939 the Poles had no credibility problem as to their will to resist Germany. Their problem, rather, was a shortage of modern military muscle. Will may be more important than muscle for deterrence, but it is not as reliable--so, acquire the latter and hope for the former.

. The art of commitment is critically important for deterrence.

Commitment, a close but distinguishable relative of will, may be a policy art; but its utility is jeopardized to the extent that the depth of the commitment exceeds the depth of national interests at stake. The implications of this proposition were explored near exhaustively by Thomas C. Schelling in the mid 1960s.[3] A commonsense validity pervades the idea. Deterrent effect will be secured only if a would-be aggressor believes, or suspects, that (for example) US policy is truly committed to the protection of the regional values at issue.

Some American strategic studies of the late 1950s and early 1960s chose to treat commitment as a flexible quality which lent itself to more or less agile manipulation by suitably schooled, tough-minded policymakers. Without denying the important facts that foreign perceptions of US national commitment will be a variable and that those perceptions can be influenced by US behavior and other signals, important constraints also obtain.

First, the subject of political commitment, and foreign perception of the same, is not a level playing field on which American policymakers can weave whatever game plans they wish and hope to succeed. For an actual case in point, US policy toward Vietnam was motivated partially by a determination to be seen as a reliable protecting power. The problem was that there was an inherent softness at the center of US policy. The political integrity of South Vietnam did not have the intrinsic or exemplary value to the United States which US policy, incredibly, insisted on declaring. In other words, it is one thing to treat commitment as an art in support of desperately vital national interests; it is quite another to treat commitment as a social-scientific tool for use by skillful policymakers. The latter will tend not to work (for long). If the scale of asserted commitment fails to bear some plausible proportionality to the worth of values at stake, intolerable criticism on the home front should be expected.

Second, crisis-time posturing as well as actual conflict entails the clash of opposing wills, of contending commitments. The question is never just, "Are we committed credibly?" In addition one must ask whether, and if so to what degree, the would-be aggressors are committed (and whether they have a feasible exit route).

. A warfighting deterrent is more user-friendly for policymakers than is a punitive deterrent.

All would-be deterrents are warfighting in character--to punish the enemy is a way of fighting a war. If classical-style military victory is unobtainable, then US warfighting strategy actually is punitive in nature. The history of US targeting policy since 1961 attests to policymakers' agreement with this proposition.[4] It is important to recognize, however, that many of the distinctions drawn between deterrence and defense or, rephrased, between a punitive and a warfighting deterrent, lack for real substance.

For several reasons of potential self-deterrence as well as other deterrence, the United States has never settled for a strategic-forces posture which could only punish Soviet (inter alia) urban areas. Nonetheless, the influence of assured-destruction thinking on plans and posture should not be dismissed as entirely trivial, while the salience of a warfighting vision easily can be overstated.

A classical warfighting approach to nuclear strategy would seek to devise a US military force capable of winning a nuclear war.[5] That is not, and long has not been, the character of US nuclear strategy. Indeed, on close inspection it transpires that the alleged warfighting trend in US targeting policy from the early 1970s to the present actually has been a variant of a punitive approach to the subject. With differing emphases, the Nixon, Ford, Carter, Reagan, and now Bush administrations have endorsed nuclear targeting ideas which have as their leitmotiv the punishment of the highest values of the Soviet state--values which happen to include military assets.[6] However, the operational purpose of such a warfighting approach would not be to win a war in a classical sense, but rather to punish the Soviet polity in ways which that polity should find intolerable in prospect--hence a stable deterrence should be the happy result. A true warfighting strategic force posture would be designed to seek to defeat the enemy's strategic forces and limit damage to North America by physically denying him the ability to do our society harm.

Just as the US warfighting approach to nuclear deterrence and actual combat is not really of the character suggested by the "warfighting" label, so its nominal contrary--a punitive deterrent--is much less contrary than its label might suggest. There is no such beast as a non-warfighting nuclear deterrent. All targeting designs express, or imply, a theory of and policy for the actual conduct of war. City bombardment or counter-military strikes are both examples of warfighting.

. A nuclear deterrent is graded abroad.

For policy to be effective as a deterrent, US military power has to speak persuasively to the values of foreigners as well as to the values and judgments of Americans. Nonetheless, all defense policy is made at home, somewhere, hence its first test of adequacy has to be domestic. Over the course of a long peace, however, it has not been unknown for defense planners and their political masters to forget that the needs of "real deterrence" can differ from those of "deterrence on paper" (to adapt Carl von Clausewitz's distinction between "real war" and "war on paper").[7]

In peacetime, foes tend to become abstract Red Teams and frequently are assumed to think and behave in ways which exercise our plans and forces, rather than truly test them. Projection, as psychologists warn us, can be a problem. In the absence of genuine knowledge about, for example, the probable Soviet modus operandi for a central nuclear war, American theorists (and planners) are wont to discern, or invent, a conveniently accommodating enemy.

The error of projecting contemporary Western ideas upon other strategic cultures is not by any means confined to nuclear planners guessing about the Soviet Union. A leading American strategic theorist, Edward N. Luttwak, for example, has quite brilliantly imputed to the Roman Empire what almost certainly is a wholly fictitious grand strategy.[8]

To be conducted successfully, a US national security policy has to play to American strengths and be domestically tolerable. But to be successful, that national security policy also must generate the strategic effectiveness necessary to prevail abroad in the teeth of an opposing national security policy. In the special case of deterrence, when US policymakers are seeking to persuade foreign political leaders not to do something, that persuasion has to work within the mental universe and political-organizational context and routines of those foreign political leaders. Value systems can differ markedly from political culture to culture, and certainly among different historical periods. Many of the people whom US leaders may need to deter (or actually compel to disgorge illicit gains) in the years ahead do not share American, or Judaeo-Christian values. They may be sensitive to loss of face or honor or dignity or possibly financial loss, but they may be less than impressed with the cries of widows and orphans.

. The officially perceived need for deterrent effect is fickle, fluctuating between near zero and an insistence upon heroic performance.

Armed forces need to be sized and shaped for international security storms, not for fair weather or conditions of permanent peace. It is understandable for defense communities not to plan against truly worst cases, for example a Soviet Union so desperate as to be literally beyond deterrence. Nonetheless, it is instructive to recall that close-to-worst cases do occur. For example, British military planners recognized in the early 1930s that their most awful nightmare would be simultaneous hostilities with a great power in north-central Europe, a great power in the Mediterranean, and a great power in the Far East. Germany, Italy, and Japan, respectively, saw to it that the British nightmare came true, albeit alleviated by the eventual facts of Russian and American alliance.

The first proposition on deterrence discussed in this article argued the uncertainty of how much is enough to deter at any particular time. The present proposition warns that a US polity which could not and did not foresee the demise of the Cold War in the late 1980s, and which (generally--as always, there were some honorable exceptions) was less than far-sighted concerning 1990s-era perils in the Persian Gulf, should not trust itself to comforting assumptions about the likely need for nuclear-force deterrent support over the next decade and more.

An individual may look at a bright blue sky, decide strangely that it will never rain again, and proceed to discard all his rainwear. When he is proved wrong he can hasten to a shop and instantly replenish his stock of bad weather equipment. But if the United States should judge the currently sunny climate of superpower relations to be a permanent and irreversible condition, and accordingly decide to discard much of its current and virtually all of its future strategic nuclear forces with or without benefit of START, it may find there is no shop to repair that deficiency in the event of a nasty turn in the weather. Unfortunately, US policy is not the problem, contrary to the views of some arms control ideologists and others who have seen the enemy and decided he is us. Instead, the problem is that US defense planners today have no way of knowing what Soviet policy will be five to ten years hence, nor what burdens, as a consequence, US policymakers then will wish to place upon what remains of US nuclear forces.

Nuclear deterrence is an unprovable phenomenon whose inferable success works, paradoxically, through political channels to undermine the plausibility of the need for its continuance. After all, what can one show to skeptics as the demonstrable security benefits of a policy of deterrence? Events which did not happen? But how can one be sure it was nuclear deterrence that caused those events not to happen?

Philosophical questions such as those above cannot prudently be dismissed as simply academic. For their own safety, many people need to be persuaded that the American policy demand for strategic effectiveness of the long-range nuclear forces could vary during this new decade from near zero to the heroic. Moreover, there is no way of knowing, of really knowing, just how much dissuasive or compellent effect might be asked of those forces. Marshal of the Soviet Union Nicolai Ogarkov has been telling veterans groups in the USSR that "order must be brought to the country."[9] An illiberal restoration of central discipline is the most likely among the possible futures for what, today, remains an exceedingly heavily nuclear-armed USSR.

The requirement for deterrent effect has an unhappy tendency to fluctuate massively between the pole of near irrelevance and the pole of maximum need. As with the case of strategic air and sea lift, the United States finds that it needs either very little or the maximum that the defense community can muster in a hurry. Legislators and "reasonable" defense officials are wont to gravitate in a time of long peace toward the "golden mean" of providing for a modest, compromise level of military capability. Such a level is plainly excessive for the needs of peace, yet would fall embarrassingly if not fatally short of requirements should (or rather when) the trumpet sounds.

. General deterrence and immediate deterrence are not synonymous.[10]

The effect of deterrence is specific to time, place, issue, and adversary. US military power in support of a general framework of order requires the flexibility to be applied so as to achieve specific, immediate, regional effects. The deterrent effect generated by strategic nuclear forces often, if inadvertently, is treated somewhat in the abstract (and in ways analogous to the effects of airpower, seapower, landpower, and spacepower). In practice, however, deterrence is always specific and has reference to concrete stakes. It is probably correct to argue that Soviet-American competition, particularly in the military realm, has a broad mutual deterrent effect. This effect is typically expressed in terms of perceptions of a general "correlation of forces" or "balance of power" helping to shape detailed policy choices. But superpower competition manifests itself in particular ways, in particular places, and over particular issues. The balance of relative advantage in general deterrence is not reliably translatable into immediate deterrence. Somehow, broad structures, frameworks, and concepts have to provide usable coin of the realm in crises that have some local focus (if more general meaning).

For instance, in 1956 it was unarguable that there was only one first-class superpower, the United States. However, that US superpower judged that it lacked the immediate deterrent clout to sustain the free Hungary of Imre Nagy, while it certainly declined to consider nuclear coercion as a policy course to oblige the Soviets to withdraw from Hungary a second time.

For a further example, the US armed forces today are in the very early stages of being reconfigured, or reconsidered, as a post-Cold War mainstay for global and regional order. It is predictable with high confidence that the United States will produce general deterrence in seemingly massive quantities. However, deterrence is never really general; rather, is it always specific. The general deterrence which the United States should provide may or may not function locally at the immediate level. In 1990, Iraq was not deterred from invading Kuwait (for the third time). In the future, can India, or Pakistan, be deterred from military adventure?

Such considerations are not to suggest that military power in both absolute and relative senses is unimportant. But it is to suggest that to be effective when and where it is needed--which frequently will not be predictable long in advance--US military power has to be flexibly focusable for immediate deterrence. Immediate deterrence in the Gulf, or in Kashmir, or in Translyvania, is not an elementary, existential consequence of a general deterrence.

Deterrence for the 1990s

Questions of deterrence theory and nuclear strategy are currently attracting again a degree of attention that many people will find surprising. Four reasons combine to explain this revival.

First, the apparent imminence of a START treaty, and early discussion of a START II follow-on, has stirred the defense community to pose unusually basic questions about the merit in, say, phased strategic-force reductions down to a real level of 3000 weapons. Second, the transformation of Soviet-American relations from a Cold War character to something else as yet unformed is promoting a healthy questioning of the strategic utility of strategic-nuclear capabilities of different kinds. Third, US-Iraqi relations in 1990-91 demonstrated first a massive failure by the United States to recognize when deterrence was required, and second a subsequent but no less massive failure to compel by coercive threat an Iraqi withdrawal from Kuwait. It has begun to dawn on many American officials and theorists that the understanding of deterrence gleaned largely from four decades of experience with the USSR might not apply to regional conflicts without some substantial amendment. Finally, Iraq's demonstration that the missile age truly has arrived has encouraged American policymakers and theorists to revisit their beliefs on the subject of the offense-defense relationship. It looks very much as if President Bush's new world disorder will require some active missile defense.


1. For example. see Morton A. Kaplan, "The Calculus of Nuclear Deterrence," World Politics, 11 (October 1958), 20-43.

2. Herman Kahn, On Thermonuclear War (Princeton, N.J.: Princeton Univ. Press, 1960), p. 287.

3. Thomas C. Schelling, Arms and Influence (New Haven, Conn.: Yale Univ. Press, 1966), chaps. 2-3.

4. For the best brief treatment in the open literature, see Scott D. Sagan, Moving Targets: Nuclear Strategy and National Security (Princeton, N.J.: Princeton Univ. Press, 1989), chap. 1. 5. On the range of choice among nuclear strategies, see Colin S. Gray, Nuclear Strategy and National Style (Lanham, Md.: Hamilton Press, 1986), chap. 9.

6. For a recent strongly worded critique, see Desmond Ball and Robert C. Toth, "Revising the SIOP: Taking War-Fighting to Dangerous Extremes," International Security. 14 (Spring 1990), 65-92.

7. Carl von Clausewitz, On War, trans. Michael Howard and Peter Paret (Princeton, N.J.: Princeton Univ. Press, 1976; first pub. 1832), p. 119.

8. Edward N. Luttwak, The Grand Strategy of the Roman Empire: From the First Century A.D. to the Third (Baltimore: Johns Hopkins Univ. Press, 1976). The most punishing criticism of Luttwak's thesis to date is in Benjamin Isaac, The Limits of Empire: The Roman Army in the East (Oxford. Ecg: Clarendon Press, 1990).

9. "Marshal Ogarkov Speech Defends Veteran's Rights," Moscow Veteran (in Russian), No. 29 (16-22 July 1990), p. 3, trans. in FBIS-SOV-90-140 (20 July 1990), p. 44.

10. This distinction is developed in Patrick M. Morgan, Deterrence: A Conceptual Analysis (Beverly Hills. Calif.: Sage Publications, 1977), chap. 2.

Dr. Colin S. Gray is Chairman of the National Institute for Public Policy in Fairfax, Virginia. Dr. Gray formerly was Director of National Security Studies at the Hudson Institute and, prior to that, an Assistant Director at the International Institute for Strategic Studies. He received his D. Phil. degree in international politics from Oxford University in 1970. Dr. Gray's most recent book is War, Peace, and Victory: Strategy and Statecraft for the Next Century (Simon and Schuster, 1990). Currently he is completing a book on Seapower and Landpower in War: An Exploration in Strategy.

Reviewed 29 August 1997. Please send comments or corrections to