Nuclear Strategy and Small Nuclear Forces: The Conceptual Components

Rajesh Rajagopalan, Research Fellow, IDSA

 

Introduction

The primary purpose of much of the nuclear strategising since 1945 has been the maintenance of nuclear deterrence. But, despite the length of the debate, there has been little agreement on the basic question of how and why nuclear deterrence works.1 All nuclear deterrence strategies have to begin with an answer to the question: "what deters?" Even if this question is not explicitly addressed or answered, this is the unavoidable beginning point of any discussion on nuclear deterrence. Most of the debates about nuclear strategies have, ultimately, been about the different answers to this fundamental question. Addressing the question, thus, allows us to get beyond the interminable debates about specific strategies.

The problem of what deters is particularly important for Small Nuclear Forces (SNFs) because the margin of error for SNFs is much thinner than that for larger nuclear forces.2 When aiming to deter, more may be unnecessary and wasteful, but it does not diminish the capacity to deter, while less might reduce that capacity. But understanding what deters might permit more efficient deterrence, if deterrence can be achieved with smaller numbers, or avoid the dangers of having an ineffective deterrent, if deterrence requires the large arsenals that the two superpowers accumulated during the Cold War.

There are possibly two ways of addressing the question of what deters. One way is to examine how scholars and practitioners have conceptualised the answer to this question. The second is empirical: examining previous cases of attempted deterrence and their successes and failures can point out useful lessons. But empirical inquiries have their limitations because, with the possible exception of the Cuban missile crisis, there have been no well documented cases of confrontation between nuclear powers.3 Another problem is that it is easier to analyse deterrence failures than successes; in fact, it is difficult to even know if deterrence is working in the normal course of events. Because of the weaknesses of the second method, the first is even more important. In this paper, I use only the first method. Subsequent parts of this project will attempt to use the empirical method to give clues to what deters.

This paper is divided into two sections. In the first section, I examine three ways in which the question of what deters has been answered by the theoreticians and practitioners of deterrence. In the second section, I examine the implications of these views for deterrence with SNFs.

Three Views of Nuclear Deterrence

Less than a year after the atomic bombings of Hiroshima and Nagasaki, two books set out the broad outlines of the nuclear debate that continued through much of the Cold War.4 In The Absolute Weapon, Bernard Brodie argued that because the actual use of nuclear weapons could not be harnessed to any meaningful military objective, the relation between weapons and war had been fundamentally altered.5 Because all major urban and industrial centres of any nation could be destroyed in a matter of hours, and because both adversaries might have the capability to conduct such campaigns, the fear of a punishing retaliation would ensure deterrence against aggression.6 On the other hand, Borden's There Will Be No Time argued that while these weapons were revolutionary in their destructive potential and would change the way wars are fought, they were, nevertheless, ultimately weapons of war, and if they differed from other weapons, this was a difference of degrees rather than of kind.

Much later, these two contrasting ideas would be debated as the difference between deterrence by punishment and deterrence by denial.7 Though new wrinkles would be added to these approaches, and even new insights revealed, no radically different conception of nuclear deterrence has emerged since Brodie and Borden produced their seminal works.8 Every nuclear power would follow some version of these two approaches.9 Some nuclear powers, the United States in particular, tended to oscillate between these approaches, the choice of approaches apparently dictated more by the changing fortunes of domestic politics rather than by any strategic imperative. In the case of the other nuclear powers (and arguably, even in the case of the US), scholars would debate inconclusively whether their nuclear strategies fit one or the other of these two approaches. These two approaches have remained the most prominent explanations of deterrence, and ones that most closely match the policy practices of nuclear deterrence.10

A less prominent idea, born out of the experience of some of the participants of the Cuban missile crisis, suggested that it was threat of nuclear war itself, rather than either the fear of retaliation or difficulty of launching a successful nuclear first-strike, that ensured what Thomas Schelling called "the tradition of non-use" of nuclear weapons.11 In this paper, I first examine the logic of these three different views of nuclear deterrence. Subsequently, I look at the implications of these views for SNFs.

Deterrence by Punishment

Briefly, deterrence by punishment seeks to prevent aggression by threatening unacceptable damage in retaliation, by the threat of punishment. Though deterrence is itself not a new concept in the military relations of states, deterrence prior to the arrival of air power and long-range bombers was almost synonymous with denial. The arrival of the strategic bomber during the inter-war years, and the promise of large scale destruction away from the battle-area and the resultant capacity to affect the policy of the adversary, highlighted, for the first time, the possibility of deterrence through the threat of punishment.12 The inter-war years also witnessed the emergence of deterrence with weapons of mass destruction in the form of chemical warfare weapons.13 But airpower was never as devastating as its proponents hoped, either in affecting civilian morale or in effecting industrial destruction, and the presence of large chemical weapons arsenals could only deter the use of other chemical weapons, not war itself.14 It was only the atomic bomb that had the potential to create sufficient destruction as to make their use a serious enough punishment to deter war itself.

Some American nuclear strategies, such as Massive Retaliation and Assured Destruction provide good examples of both the logic of deterrence by punishment and the practical problems and requirements in adopting such a strategy. Massive Retaliation, as the earliest "popular" exposition of American nuclear strategy, was explicit in its threat of punishment as a means of deterring the Soviet Union.15 The strategy was severely criticised, but the criticism was about the expansive reading of the capacity of the nuclear deterrent, rather than of the purpose of the deterrent itself.16 Because the criticism of Massive Retaliation was not so much in what it sought to achieve (deter the Soviet Union) or in the manner it sought to achieve it (by the threat of punishment) but in the methodology of executing the threat of punishment, it is not surprising that the logic behind the American nuclear strategy did not change with the shift to Assured Destruction under McNamara. The primary objective of Assured Destruction, "to deter deliberate nuclear attack upon the United States or its allies," was no different than the objectives of the doctrine of massive retaliation.17

The threat of a punishing retaliation with nuclear weapons might appear to be a straightforward one, but comparing Massive Retaliation with Assured Destruction suggests that clarity of purpose and the correlation of means and ends is important in ensuring the credibility of the threat of punishment, at least in the minds of those holding out the threat. As the Soviet nuclear force began to grow,18 American strategists became increasingly worried about the vulnerability of American missile and bomber bases to a Soviet first-strike, and hence about the credibility of the strategy of Massive Retaliation.19 The threat of punishment would be credible only if the forces that are to do the punishing survived to complete their mission, thus, introducing into the lexicon of nuclear strategy the idea of "second-strike," the capacity of a nuclear force to retaliate after being struck first. This would become a basic concern in the subsequent build-up of the American nuclear force.

Another concern in McNamara's strategy of Assured Destruction was the extent of damage that constituted punishment. Accepting that no a priori judgements can be made about the capacity of any society to tolerate punishment, McNamara, nevertheless, fixed the level of destruction required as being the capacity to kill 20 to 25 per cent of the Soviet population and destroy 50 per cent of Soviet industry, which could be achieved with about 400 one-megaton-equivalent delivered warheads. This figure was settled upon because American analysts estimated that increasing the numbers of warheads delivered beyond this amount brought no appreciable increase in the level of destruction that could be inflicted on the Soviet Union.20 At this level, the United States simply faced the problem of diminishing returns. This is, of course, not the only way to conceive of punishment: the French strategist Pierre Gallois suggested that the amount of punishment can be proportional to the stake being defended. In other words, a less worthy objective can be defended with a smaller nuclear force and a smaller threat of punishment.21

The two fundamental concerns in nuclear deterrence through the threat of punishment can be seen in the criticism of Massive Retaliation and the conception of Assured Destruction strategies: the need to have forces that can survive an enemy attack and a conception of the level of destruction that is required for deterrence to operate. However, neither of these issues is a settled one: the capacity of a force to survive in the unlikely event of an enemy first-strike, and the level of destruction that constitutes punishment, are both open to debate.

Deterrence by Denial

Deterrence by denial is more complex. It is difficult to classify the denialists as deterrence theorists because at a very fundamental level, the denialists exhibit a profound mistrust in the possibility of deterrence. The logic of denial begins at the point when deterrence fails.

The "victory theorists" of the 1970s American nuclear debate provide the best example of deterrence by denial. The victory theorists assumed that deterrence can fail, and they insisted on the need to plan for that eventuality. They argued that deterrence could fail to deter a Soviet Union if it was in the midst of a political crisis,22 because of irrationality and misperception, because the Marxist-Leninst leaders of the Kremlin had a Clauswitzian view of war as a continuation of politics, and because Marxist ideology, which was held to be true even in the nuclear age, predicted victory for socialism.23

Deterrence might also fail simply because of the vulnerability of the nuclear arsenal. This theme, an important one for the victory theorists who argued that the emphasis on strategic stability and arms control had taken the United States from a position of strategic "superiority" to "sub-parity", echoed the worries of Albert Wohlstetter and the Gaither Committee in the 1950s about the possibility that the presumed vulnerability of the American strategic force might itself invite a Soviet first strike by presenting too tempting a target.24 Ultimately, as Colin Gray puts it, " (no one) can guarantee that deterrence will always work."25

Indeed, the true distinction might not be between deterrence by punishment and deterrence by denial but, as Glenn Snyder originally suggested, between deterrence and defence, because defence rather than deterrence is the true objective of the denialists.26 For the denialists, deterrence is a concession to the nuclear age, but a concession that should not be carried too far. The devastation of nuclear war makes deterrence the first choice, but not an alternative to defence. Preparing for defence, even in the context of nuclear war, makes sense to the denialists because good defence both reinforces deterrence and it also accounts for the possibility of the failure of deterrence. Denialists, thus, attempt to prevent aggression by convincing the aggressor through strong defence that such aggression would surely fail and that the aggressor would be defeated if he/she tried it.27 The "victory theorists" of the 1970s, who argued that the United States should plan to fight and prevail in a nuclear war, provide the best example of such thinking during the Cold War nuclear confrontation.28 The difference from punishment should be clear: Assured Destruction suggested that the Soviet Union would be foolish to attack because nobody could possibly win a nuclear war, while victory theorists proposed that the Soviet Union would not attack only if it was convinced that not only would it lose the war, but that the US would win it.

The requirements for a strategy of denial were severe, and encompassed strategic defensive forces, strategic offensive forces, the command and control capacity to maintain central authority through nuclear exchanges and sufficient civil defence measures to ensure societal survival after a nuclear war.29 The elaborateness of these plans was much criticised by other strategists, but a good part of it became official American policy in the 1980s.30

Existential Deterrence

Existential deterrence, a concept suggested by McGeorge Bundy, and based on his experience as a member of the Kennedy Administration during the Cuban missile crisis, argues that nuclear deterrence is the function primarily of the presence of survivable thermonuclear arsenals in the hands of both superpowers. Though he did not call it "existential deterrence", Andre Beaufre had suggested even earlier that "the basis of nuclear deterrence is the certainty of the damage which would result from the use of these weapons...It is this threat of destruction which generates deterrence because of the undoubted level of risk implied."31 Bundy has defined the concept thus: "As long as each side has thermonuclear weapons that could be used against the opponent, even after the strongest possible preemptive attack, existential deterrence is strong and it rests on uncertainty about what could happen."32 Bundy had suggested this idea much earlier, though he did not then characterise it as existential deterrence, arguing in a 1969 essay in the journal Foreign Affairs that in the light of the certain prospect of retaliation, "there has been literally no chance at all that any sane political authority . . . would consciously choose to start a nuclear war."33 Much more recently, Bundy repeated his earlier characterisation of existential deterrence as the result of the "existence of the enormous arsenals of the two superpowers."34 For Bundy, a crucial element in the creation of existential deterrence was, as the above quotation makes clear, the presence of survivable thermonuclear forces.35

But can existential deterrence work only in the presence of arsenals such as the ones that the US and the Soviet Union developed during the Cold War period? Bundy is unhelpful on this point because there is a good amount of confusion in Bundy's thinking about the kind of nuclear forces that are required for achieving existential deterrence. The selections from his writing above would tend to suggest that existential deterrence cannot work in the absence of robust nuclear forces. This is strengthened in his discussion of other, smaller nuclear forces such as those of the French. He is sceptical of the deterrent capabilities of such forces because "the armaments of the middle-level nuclear powers are...vulnerable to an obliterating first-strike, and that situation may not entirely disappear even if they shift to seaborne missiles."36

But, at other times, Bundy appears to suggest that the balance of nuclear forces is irrelevant to the working of nuclear deterrence, which could be taken to mean that the deterrent effect of smaller nuclear forces is no different from that of the Cold War superpower arsenals. In his 1969 essay, he argued that "(I)n the real world of real political leaders --whether here or in the Soviet Union-- a decision that would bring even one hydrogen bomb on one city of one's own country would be recognized in advance as a catastrophic blunder..."37 If, as he suggests, the destruction of just one city would be considered catastrophic, there would be little difference between having two hundred warheads and ten thousand because even with such a lopsided balance of forces, there could be no complete guarantee that the larger force can carry out a successful nuclear first strike, or that the smaller force would not be able to retaliate with at least a few nuclear warheads. Such a reading is supported by the experience of the Cuban missile crisis, when, despite a 17-to-1 advantage in strategic nuclear warheads, the United States gained no measurable advantage in dealing with Khruschev.38 Bundy, and many of his colleagues during the Cuban missile crisis, credit American superiority in local conventional forces rather than superiority in strategic nuclear forces for compelling Khruschev to back down during the Caribbean confrontation.39 Additionally, Bundy has repeatedly asserted that nuclear superiority was and is inconsequential in ensuring nuclear deterrence: "...mutual deterrence," he has argued, "has been steady through more than three decades of dramatic variations in the relative numbers of warheads, in delivery systems, and in the vulnerability of particular systems on each side."40 Moreover, Bundy's appreciation of the deterrent capability of smaller nuclear forces is inconsistent: though he is sceptical of the deterrent capabilities of the British and French nuclear forces, he seems more convinced of the capabilities of the equally small Chinese nuclear force, arguing that "...the Chinese deterrent is untested precisely because it exists."41

Given Bundy's own confusion about the concept, it is not surprising that other scholars have also misinterpreted the concept. Most scholars have described existential deterrence as simply the consequence of the existence of any nuclear force, rather than a survivable one, an important distinction. Lawrence Freedman, for example, characterises existential deterrence as "almost wholly impervious to the location and capabilities of nuclear weapons and the doctrines that would notionally govern their use,"42 a characterisation that ignores completely the importance of the survivability of nuclear forces, something that Bundy has stressed with uncharacteristic consistency. Though second-strike capability might be important, on the other hand, the strength of the second-strike might not be particularly important. Beaufre, for example, who also suggests that SNFs can deter much larger nuclear forces, ascribes such capability to the capacity of even very small forces to hold out the threat of a nuclear "riposte", however weak such riposte might be.43 If it is the threat of nuclear war that deters, then the threat of a second-strike capability is what is important, and the strength of the second-strike might be irrelevant.

Unfortunately, scholars sympathetic to the idea of existential deterrence have been equally prone to such mischaracterisation of the concept, leading to further confusion. Marc Trachtenberg, for example, in an otherwise excellent study of the Cuban missile crisis, defined existential deterrence thus: "The mere existence of nuclear forces means that, whatever we say or do, there is a certain irreducible risk that an armed conflict might escalate into a nuclear war. The fear of escalation is thus factored into political calculations: faced with this risk, states are more cautious and more prudent than they would otherwise be."44 Completely absent from this definition is any reference to a survivable, second-strike force. Others have carried this idea even further, suggesting that existential deterrence works even in the absence of openly-acknowledged nuclear forces, as long as the adversaries believe that the opponent has nuclear forces.45

Trachtenberg's own work tended to support the Bundy thesis that the balance of nuclear forces made little difference to the eventual outcome of the Cuban missile crisis. Nevertheless, his explanation of existential deterrence, which might be called a "pure" view of existential deterrence, left out a crucial component of the concept: the presence of survivable nuclear forces on both sides, in other words, a second-strike capability. Other theorists have argued similarly: Beaufre, argued that it was uncertainty and risk that are inherent in any contemplation of a nuclear first-strike that create deterrence; thus, a small nuclear force, even one without an assured second-strike capability can still deter.46 But for Bundy, a survivable second-strike force was an essential ingredient for creating the risk of nuclear war that induces so much caution among the leaders of nations with nuclear arms.

Implications for SNFs

What are the implications of these three answers to the question "what deters?" for SNFs? The implications can be studied with two further questions. First, what do these answers say about the deterrence capability of SNFs? Second, what do these answers say about an appropriate nuclear strategy for SNFs?

From the perspective of denial strategies, SNFs have no deterrent value. If you cannot deter unless you can fight and prevail in a nuclear war, then, in a world populated by both SNFs and larger nuclear forces, SNFs are by definition incapable of deterrence because they would not be able to fight and prevail against a larger nuclear force. Would the answer be different if the two larger nuclear forces, the American and the Russian, were removed from the equation? Even then, SNFs might not deter because the difference between the nuclear deterrent forces is not great enough to give any one force a sufficiently large advantage. Moreover, in the absence of defensive measures, the absolute level of destruction might be too large to prevent any clear determination of victory. In short, from the perspective of denial strategy, SNFs are incapable of deterrence.

The deterrence capability of SNFs is suspect from the perspective of punishment strategies also. Though this perspective does not deny that SNFs can deter, the emphasis on a second-strike, and more importantly on a credible second-strike, would make it difficult for SNFs to have any capacity to deter larger nuclear forces. Even if punishment was defined as less than destruction (the McNamara definition), the emphasis on a substantial punishment still remains (such as Gallois' notion of "proportional" punishment).

But the burden of punishment, and, hence, of deterrence, might be less of a problem if the United States and Russia were removed from the calculus of deterrence for SNFs. If SNFs only had to deter other SNFs, as the situation is in South Asia, punishment, even substantial punishment, might be less difficult. In essence, then, deterrence through punishment suggests that SNFs might have credibility, as long as they are confined to deterring other SNFs.

Existential deterrence is the most optimistic of the three about the deterrence prospects of SNFs. But here also, Bundy's approach, which comes close to a punishment strategy, would be less hopeful of the deterrence capacity of SNFs. On the other hand, what I have called a "pure" existential deterrence perspective would consider SNFs to have a deterrence capacity strong enough to deter even the American and Russian nuclear arsenals.

And what do these three perspectives say about an appropriate nuclear deterrent strategy for SNFs? At the very outset, it would appear that denial strategies are irrelevant for SNFs because denial strategies require a level of force capability that would simply be absent in an SNF. Denial strategies require, at the minimum, the capacity to absorb an enemy first-strike, and to have enough forces left over not only to retaliate but to defeat the enemy. Alternatively, it requires a first-strike force that would be sufficiently strong to carry out a successful assault on an enemy nuclear force. These are clearly beyond the capabilities of an SNF in a confrontation with a large nuclear force, and it might be beyond the capability of SNFs in confrontations with other SNFs also. The capability requirements of a denial strategy are so great that it is clearly beyond the capability of SNFs.47

Denial strategies might be theoretically possible for an SNF in a nuclear confrontation with another SNF. In a confrontation that is limited to SNFs, the symmetry in capabilities will permit, theoretically, the adoption of denial strategies. For example, in the South Asian context, India might theoretically be able to build a first-strike force against Pakistan, and China might be able to have such a capability against India. But even in an SNF-SNF dyad, there will be significant practical problems for an SNF first-strike force.

The first problem is that any attempt to adopt a denial strategy will immediately ignite a nuclear arms race, which could possibly threaten the strategy. The second problem is that any such SNF will need to be dedicated to only one threat. A Chinese attempt to adopt a denial strategy vis-a-vis India, for example, would require such a great extent of Chinese nuclear focus on India that it would cripple Chinese capabilities in any other direction. The third problem is that such a strategy is likely to increase the problems of instability. The requirements of any such strategy, especially in terms of command, control, communications and intelligence (C3I) facilities, are so complex (even when pitted against another SNF) that they are likely to increase the risks of inadvertent war.

A punishment strategy can be less demanding on resources, and, thus, more appropriate for an SNF, though, as the American build-up of forces during the 1960s showed, even a punishment strategy can be pursued with extravagance. But a punishment strategy still requires quite an extensive nuclear infrastructure, primarily because punishment also requires some forces that could survive a first-strike. Another factor in deciding on a punishment strategy is determining what constitutes punishment: if punishment is assumed to be "destruction" as McNamara did, the force requirements would be substantial. But "destruction" is only one way to conceptualise punishment. Retaliation is another way in which punishment can be conceptualised, and retaliation does not require as extensive a second-strike force.48

The least demanding of all the approaches is existential deterrence. However, existential deterrence itself, as was seen earlier, could mean both a nuclear force with a robust second-strike capability, in Bundy's conception, or the mere existence of a nuclear capability which depends on uncertainty as a factor, as Beaufre conceptualised it. Bundy's "assured retaliation" perspective is still less demanding than a punishment strategy, since it includes no determination of the extent of damage that needs to be inflicted to deter. But it is still more demanding than a "pure" existential deterrence view, because a force based purely on the logic of existential deterrence requires not even the assurance of survival, only uncertainty in the mind of the enemy who contemplates a first-strike.

Conclusion

This discussion suggested three answers to the question "what deters?" It has also suggested the implications of these three answers to both the prospects of having an effective nuclear deterrent with an SNF and the framing of deterrent strategies for SNFs. Nuclear deterrence through denial, and even punishment, is suspicious of the capacity of SNFs to create an effective nuclear deterrent, while the existential deterrence perspective is optimistic about it. As regards an appropriate nuclear strategy for SNFs, the requirements of a denial strategy are too large for such a strategy to be adopted by SNFs. The force requirements of punishment strategies is less, but that of existential deterrence is the least of all, and, therefore, is most affordable for SNFs.

The limitations of the above analysis must also be emphasised: decisions about nuclear strategy have never entirely been determined by the right answer about deterrence. Such questions are also influenced by everything from domestic politics to bureaucratic bargaining and the possibilities of budgetry and technological capacities, issues that are not considered here. But answering the basic question of how to achieve deterrence is an essential starting point.

 

NOTES

1. Or, indeed, even whether it works at all. John Mueller, for one, has argued that nuclear weapons played little part in the prevention of major war after 1945. See John Mueller, "The Essential Irrelevance of Nuclear Weapons: Stability in the Postwar World," International Security, 13:2, Fall 1989, pp. 55-79. I do not address Mueller's argument here because, even assuming his arguments are valid, they addressed only the reason for the absence of system-wide war, and are not applicable to most of the developing world, where there have been frequent wars.

2. For the purposes of this paper, I use SNF to refer to nuclear arsenals of states other than the US and Russia/Soviet Union. For a different definition see, Rodney W. Jones, "Small Nuclear Forces," in Rodney W. Jones, ed., Small Nuclear Forces and US Security Policy: Threats and Potential Conflicts in the Middle East and South Asia (Lexington, Mass.: Lexington Books, 1984), p. 1.

3. Even on the Cuban missile crisis, there are considerable disagreements among scholars and even the participants in the crisis about what happened during the crisis and what it means. See, for example, Raymond L. Garthoff, "The Havana Conference on the Cuban Missile Crisis," The Cold War International History Project (hereafter CWIHP Bulletin) 1, Spring 1992, pp. 2-4; Mark Kramer, "Tactical Nuclear Weapons, Soviet Command Authority, and the Cuban Missile Crisis," CWIHP Bulletin 3, Fall 1993, pp. 40, 42-46; and James G. Blight, Bruce J. Allyn and David A. Welch, "Kramer vs. Kramer: Or How Can You Have Revisionism in the Absence of Orthodoxy," CWIHP Bulletin 3, Fall 1993, pp. 41, 47-50.

4. The two books were Bernard Brodie's, The Absolute Weapon (New York: Harcourt Brace, 1946) and William Borden's, There Will Be No Time: The Revolution in Strategy (New York: Macmillan, 1946). See Gregg Herken, Counsels of War (New York: Allied A. Knopf, 1985), pp. 8-17.

5. The analysis of both Brodie's The Absolute Weapon and Borden's There Will Be No Time is based on secondary material. On the influence and importance of these two seminal works, see, in addition to Herken's Counsels of War, Roman Kolkowicz, "The Rise and Decline of Deterrence Doctrine," Journal of Strategic Studies, 9:4, December 1986, pp. 3-12; Robert Jervis, "Strategic Theory: What's New and What's True," Journal of Strategic Studies 9:4, December 1986, pp. 135-162; Marc Trachtenberg, "Strategic Thought in America, 1952-1966," in his History and Strategy (Princeton, New Jersey: Princeton University Press, 1991); Barry H. Steiner, "Using the Absolute Weapon: Early Ideas of Bernard Brodie on Atomic Strategy," Journal of Strategy Studies, 7:4, December 1984, pp. 365-393; and Barry Steiner, Bernard Brodie and the Foundation of American Nuclear Strategy (Lawrence, Kansas: University Press of Kansas, 1991).

6. Brodie was not the only one to propose the deterrent powers of atomic arms, even as early as 1946. See, Steiner, Ibid., especially pp. 366-367, and fn. 1 and 7.

7. The distinction between deterrence by denial and punishment was first suggested by Glenn Snyder. See Glenn H. Synder, Deterrence by Denial and Punishment (Princeton: Woodrow Wilson School of Public and International Affairs, Research Monograph No. 1, 1959).

8. A useful overview of the evolution of the nuclear debate can be found in Robert Jervis, "Deterrence Theory Revisited," World Politics, 31:2, January 1979, pp. 289-324.

9. It should be noted that a number of scholars, including Barry Buzan, deny that any such distrinction exists, arguing that the real distinction is between deterrence and compellence. See Barry Buzan, An Introduction to Strategic Studies: Military Technology and International Relations (London: Macmillan/International Institute for Strategic Studies, 1987), p. 135.

10. Jervis, n. 5, pp. 135-162.

11. Thomas C. Schelling, The Strategy of Conflict (Cambridge, Mass: Harvard University Press, 1980), p. 260.

12. George H. Quester, Deterrence Before Hiroshima: The Airpower Background of Modern Strategy (New York: John Willey & Sons, Inc., 1966).

13. Richard M. Price, The Chemical Weapons Taboo (Ithaca: Cornell University Press, 1997).

14. Jeffrey W. Legro, Cooperation Under Fire: Anglo-German Restraint During World War II (Ithaca: Cornell University Press, 1995).

15. See, John Foster Dulles, "The Doctrine of Massive Retaliation," in Richard G. Head and Ervin J. Rokke, eds., American Defence Policy, 3rd edition (Baltimore: Johns Hopkins University Press, 1973), pp. 62-64.

16. Massive Retaliation, of course, sought to deter not only Soviet nuclear forces but also Soviet "aggression" in the post-colonial world. Despite such optimistic views of the power of nuclear deterrence, the primary mission of American nuclear forces remained the deterrence of Soviet nuclear forces.

17. Robert S. McNamara, The Essence of Security: Reflections in Office (London: Hodder & Stoughton, 1968), p. 52. See also, Alain C. Enthoyen and K. Wayne Smith, How Much Is Enough? Shaping the Defence Programme, 1961-1969 (New York: Harper and Row, 1971), p. 174; and Jerome H. Kahan, Security in the Nuclear Age: Developing US Strategic Arms Policy (Washington, DC.: The Brookings Institution, 1975), p. 94.

18. American intelligence capability about the Soviet strength during this period was weak, and this lack of information added to the sense of threat. Soviet nuclear forces were, throughout this period, much weaker than American strategists had assumed. See John Prados, The Soviet Estimates: US Intelligence Estimate & Russian Military Strength (New York: The Dial Press, 1982).

19. Strobe Talbott, The Master of the Game: Paul Nitze and the Nuclear Peace (New York: Vintage Books, 1989), pp. 64-74, provides an excellent behind-the-scene view of this "vulnerability" debate. See also, Herken, n. 5, chs. 9-12; and Lawrence Freedman, The Evolution of Nuclear Strategy (London: International Institute for Strategic Studies/Macmillan, 1987), pp. 134-154, 160-165.

20. Enthoven and Smith, n. 17, pp. 175, 207-208.

21. Freedman, n. 19, pp. 314-317.

22. Keith B. Payne, Nuclear Deterrence in US-Soviet Relations (Boulder, Color: Westview Press, 1982), pp. 57-58.

23. Ibid., and Richard Pipes, "Why the Soviet Union Thinks it Can Fight and Win a Nuclear War," Commentary, 64:1, July 1977.

24. See n. 17.

25. Colin S. Gray, "Perspectives on Fighting a Nuclear War," (Correspondence) International Security, 6:1, Summer 1981, pp. 185-186.

26. Glenn Snyder, Deterrence and Defence: Towards A Theory of National Security (Princeton: Princeton University Press, 1961).

27. Colin Gray has made a distinction between a strategy based on denying the Soviet Union victory in a nuclear war and a strategy that deters by planning for an American victory in a nuclear war. See Colin S. Gray, "War-Fighting for Deterrence," The Journal of Strategic Studies, 7:1, March 1984, p. 5. I do not find the distinction particularly useful.

28. Rajesh Rajagopalan, "The Advocates of Armageddon: Nuclear War and the Victory Theorists," Strategic Analysis, 11:8, November 1987, pp. 933-944.

29. Gray, n. 27; Colin S. Gray, "Nuclear Strategy: The Case for a Theory of Victory," International Security, 4:1 Summer 1979, pp. 54-87; Payne, n. 22; and T.K. Jones and W. Scott Thompson, "Central War and Civil Defence," Orbis, 22:3, Fall 1978.

30. Les Aspin, "How to Look at the Soviet-American Balance," Foreign Policy, 22, Spring 1976, pp. 96-106; Richard Burt, "Reassessing the Strategic Balance," International Security, 5:1, Summer 1980; Bernard Brodie, "The Development of Nuclear Strategy," International Security, 2:4, Spring 1978, Michael Howard, "On Fighting A Nuclear War," International Security, 5:4, Spring 1981, pp. 3-17; and Steven Kull, "Nuclear Nonsense," Foreign Policy, 58, Spring 1985.

31. Andre Beaufre, Deterrence and Strategy, trans by R.H. Barry (London: Faber and Faber, 1965), p. 35. Original italics.

32. McGeorge Bundy, "The Bishops and the Bomb," The New York Review of Books, June 16, 1983, as quoted in Lawrence Freedman, "I Exist; Therefore I Deter" (Review Essay) International Security, 13:1, Summer 1988, p. 184. See also, Michael Howard, "Nuclear Danger and Nuclear History," (Review Essay) International Security, 14:1, Summer 1989, pp. 176-83.

33. McGeorge Bundy, "To Cap the Volcano," Foreign Affairs, 48:1, October 1969, p. 9.

34. McGeorge Bundy, "Nuclear Weapons and the Gulf," Foreign Affairs, 70:4, Fall 1991, p. 84. My italics.

35. Curiously, Bundy does not refer to the concept in his history of the nuclear age. See, McGeorge Bundy, Danger and Survival: Choices About the Bomb in the First Fifty Years (New Delhi: Affiliated East-West Press/in cooperation with Random House, New York, 1989).

36. Bundy, n. 33, p. 12. His views on the deterrent capability of Britain and France appear to have remained unchanged. See Bundy, Ibid., pp. 501-503.

37. Bundy, n. 33, p. 10.

38. Bundy, n. 35, p. 453; and Marc Trachtenberg, "The Influence of Nuclear Weapons in the Cuban Missile Crisis," in his History and Strategy, n. 5. This article was originally published as an easy in International Security in summer 1985. For an opposing perspective on the impact of the nuclear balance during the crisis, see Freedman, n. 32, pp. 182-83.

39. Twenty years after the event, a group of Kennedy advisors stated bluntly: "American nuclear superiority was not, in our view, a critical factor (in the resolution of the crisis in American favour)...No one of us ever revived the nuclear balance for comfort in those hard weeks." Dean Rusk, Robert McNamara, George W. Ball, Roswell L. Gilpatrick, Theodore Sorensen, and McGeorge Bundy, "The Lessons of the Cuban Missile Crisis," Time, September 27, 1982, pp. 85-86, as quoted in Bundy, n. 35, p. 447.

40. Bundy, n. 35, p. 592. He had expressed similar views in his earlier essays also. See Bundy, n. 22, pp. 9-12.

41. Bundy, n. 35, p. 535.

42. Freedman, n. 32, p. 184.

43. Beaufre, n. 31, pp. 37-39.

44. Trachtenberg, n. 38, p. 237.

45. Devin T. Hagerty, "The Power of Suggestion: Opaque Proliferation, Existential Deterrence, and the South Asian Nuclear Arms Competition," Security Studies, 2:3/4, Spring-Summer 1993, pp. 256-83.

46. Beaufre, n. 31, pp. 37-39.

47. This situation could, of course, change if the US and Russia were to conduct further reductions in their arsenals to bring them close to or even below 1,000 strategic nuclear warheads. In such case, China, for example, could build up to a theoretical denial capability.

48. Indeed, McNamara had originally considered a strategy of assured retaliation rather than assured destruction, though this did not mean any smaller build-up. See, Freedman, n. 19, p. 246.