All Categories

Kenneth Waltz, “The Spread of Nuclear Weapons: More MayBetter,” Adelphi Papers, Number 171 (London: InternationalInstitute for Strategic Studies, 1981)INTRODUCTIONWhat will the spread of nuclear weapons do to the world? I say ‘spread rather than proliferation’because so far nuclear weapons have proliferated only vertically as the major nuclear powers haveadded to their arsenals. Horizontally, they have spread slowly across countries, and the pace is notlikely to change much. Short-term candidates for the nuclear club are not very numerous. and theyare not likely to rush into the nuclear military business. Nuclear weapons will nevertheless spread,with a new member occasionally joining the club. Counting India and Israel, membership grew toseven in the first 35 years of the nuclear age. A doubling of membership in this decade would besurprising. Since rapid changes in international conditions can be unsettling, the slowness of thespread of nuclear weapons is fortunate.Someday the world will be populated by ten or twelve or eighteen nuclear-weapon states (hereafterreferred to as nuclear states). What the further spread of nuclear weapons will do to the world istherefore a compelling question.Most people believe that the world will become a more dangerous one as nuclear weapons spread.The chances that nuclear weapons will be fired in anger or accidentally exploded in a way thatprompts a nuclear exchange are finite, though unknown. Those chances increase as the number ofnuclear states increase. More is therefore worse. Most people also believe that the chances thatnuclear weapons will be used vary with the character of the new nuclear states—their sense ofresponsibility, inclination toward devotion to the status quo, political and administrativecompetence. If the supply of states of good character is limited as is widely thought, then the largerthe number of nuclear states, the greater the chances of nuclear war become. If nuclear weapons areacquired by countries whose governments totter and frequently fall, should we not worry moreabout the world’s destruction then we do now? And if nuclear weapons are acquired by two statesthat are traditional and bitter rivals, should that not also foster our concern?Predictions on grounds such as the above point less to likelihoods and more to dangers that we canall imagine. They identify some possibilities among many, and identifying more of the possibilitieswould not enable one to say how they are likely to unfold in a world made different by the slowspread of nuclear weapons. We want to know both the likelihood that new dangers will manifestthemselves and what the possibilities of their mitigation may be. We want to be able to see thefuture world, so to speak, rather than merely imagining ways in which it may be a better or a worseone. How can we predict more surely? In two ways: by deducing expectations from the structureof the international political system and by inferring expectations from past events and patterns.With those two tasks accomplished in the first part of this paper, I shall ask in the second partwhether increases in the number of nuclear states will introduce differences that are dangerous anddestabilizing.

I. DETERRENCE IN A BIPOLAR WORLDThe world has enjoyed more years of peace since 1945 than had been known in this century—ifpeace is defined as the absence of general war among the major states of the world. The SecondWorld War followed the first one within twenty-one years. As of 1980 35 years had elapsed sincethe Allies’ victory over the Axis powers. Conflict marks all human affairs. In the past third of acentury, conflict has generated hostility among states and has at times issued in violence among theweaker and smaller ones. Even though the more powerful states of the world have occasionallybeen direct participants, war has been confined geographically and limited militarily. Remarkably,general war has been avoided in a period of rapid and far-reaching changes—decolonization; therapid economic growth of some states; the formation. tightening, and eventual loosening of blocs;the development of new technologies; and the emergence of new strategies for fighting guerrillawars and deterring nuclear ones. The prevalence of peace, together with the fighting ofcircumscribed wars, indicates a high ability of the post-war international system to absorb changesand to contain conflicts and hostility.Presumably features found in the post-war system that were not present earlier account for theworld's recent good fortune. The biggest changes in the post-war world are the shift frommultipolarity to bipolarity and the introduction of nuclear weapons.The Effects of BipolarityBipolarity has produced two outstandingly good effects. They are seen by contrasting multipolarand bipolar worlds. First, in a multipolar world there are too many powers to permit any of them todraw clear and fixed lines between allies and adversaries and too few to keep the effects ofdefection low. With three or more powers, flexibility of alliances keeps relations of friendship andenmity fluid and makes everyone's estimate of the present and future relation of forces uncertain. S6long as the system is one of fairly small numbers, the actions of any of them may threaten the secur -ity of others. There are too many to enable anyone to see for sure what is happening. and too few tomake what is happening a matter of indifference.In a bipolar world, the two great powers depend militarily mainly on themselves. This is almostentirely true at the strategic nuclear level, largely true at the tactical nuclear level, and partly true atthe conventional level. In 1978, for example, the Soviet Union's military expenditures were over90% of the total for the Warsaw Treaty Organization, and those of the United States were about60% of the total for NATO. With a GNP 30% as large as ours, West Germany's expenditures were11.5% of the NATO total, and that is the second largest national contribution. Not only do we carrythe main military burden within the alliance because of our disproportionate resources but alsobecause we contribute disproportionately from those resources. In fact if not in form, NATOconsists of guarantees given by the United States to her European allies and to Canada. The UnitedStates, with a preponderence of nuclear weapons and as many men in uniform as the West Europeanstates combined, may be able to protect them; they cannot protect her.Because of the vast differences in the capabilities of member states, the roughly equal sharing ofburdens found in earlier alliance systems is no longer possible. The United States and the SovietUnion balance each other by ‘internal’ instead of ‘external’ means, relying on their own capabilitiesmore than on the capabilities of allies. Internal balancing is more reliable and precise than external

balancing. States are less likely to misjudge their relative strengths than they are to misjudge thestrength and reliability of opposing coalitions. Rather than making states properly cautious and for-warding the chances of peace, uncertainty and miscalculation cause wars. In a bipolar world,uncertainty lessens and calculations are easier to make. The military might of both great powersmakes quick and easy conquest impossible for either, and this is clearly seen. To respond rapidly tofine changes in the military balance is at once less important and more easily done.Second, in the great-power politics of a multipolar world, who is a danger to whom. and who can beexpected to deal with threats and problems, are matters of uncertainty. Dangers are diffused,responsibilities blurred, and definitions of vital interest easily obscured. Because who is a danger towhom is often unclear, the incentive to regard all disequilibrating changes with concern and respondto them with whatever effort may be required is weakened. To respond rapidly to fine changes is atonce more difficult, because of blurred responsibilities, and more important, because states live onnarrow margins. Interdependence of parties, diffusion of dangers, confusion of responses: These arethe characteristics of great-power politics in a multi polar world.In the great-power politics of a bipolar world, who is a danger to whom is never in doubt.Moreover, with only two powers capable of acting on a world scale, anything that happensanywhere is potentially of concern to both of them. Changes may affect each of the two powersdifferently, and this means all the more that few changes in the world at large or within each other'snational realm are likely to be thought irrelevant. Self-dependence of parties, clarity of dangers,certainty about who has to face them: These are characteristics of great-power politics in a bipolarworld. Because responsibility is clearly fixed, and because relative power is easier to estimate. abipolar world tends to be more peaceful than a multipolar world.Will the spread of nuclear weapons complicate international life by turning the bipolar world into amultipolar one? The bipolar system has lasted more than three decades because no third state hasdeveloped capabilities comparable to those of the United States and the Soviet Union. The UnitedStates produces about a quarter of the world's goods, and the Soviet Union about half as much.Unless Europe unites, the United States will remain economically well ahead of other states. Andalthough Japan's GNP is fast approaching the Soviet Union's, Japan is not able to compete militarilywith the super-powers. A state becomes a great power not by military or economic capability alonebut by combining political, social, economic, military, and geographic assets in more effective waysthan other states can.In the old days weaker powers could improve their positions through alliance by adding the strengthof foreign armies to their own. Cannot some of the middle states do together what they are unable todo alone? For two decisive reasons, the answer is ‘no’. First, nuclear forces do not add up. Thetechnology of warheads, of delivery vehicles, of detection and surveillance devices, of commandand control systems, count more than the size of forces. Combining separate national forces is notmuch help. Second, to reach top technological levels would require lull collaboration by, say,several European states. To achieve this has proved politically impossible. As de Gaulle often said,nuclear weapons make alliances obsolete. At the strategic level he was right.States fear dividing their strategic labours fully—from research and development throughproduction, planning, and deployment. This is less because one of them might in the future be atwar with another, and more because anyone's decision to use the weapons against third partiesmight be fatal to all of them. Decisions to use nuclear weapons may be decisions to commit suicide.

Only a national authority can be entrusted with the decision, again as de Gaulle always claimed.Only by merging and losing their political identities can middle states become great powers. Thenon-additivity of nuclear forces means that in our bipolar world efforts of lesser states cannot tilt thestrategic balance.Great powers are strong not simply because they have nuclear weapons but also because theirimmense resources enable them to generate and maintain power of all types. military and other, atstrategic and tactical levels. Entering the great-power club was easier when great powers werelarger in number and smaller in size. With fewer and bigger ones, barriers to entry have risen. Theclub will long remain the world's most exclusive one. We need not fear that the spread of nuclearweapons will turn the world into a multipolar one.The Effects of Nuclear WeaponsNuclear weapons have been the second force working for peace in the post-war world. They makethe cost of war seem frighteningly high and thus discourage states from starting any wars that mightlead to the use of such weapons. Nuclear weapons have helped maintain peace between the greatpowers and have not led their few other possessors into military adventures.5 Their further spread,however, causes widespread fear. Much of the writing about the spread of nuclear weapons has thisunusual trait: It tells us that what did no, happen in the past is likely to happen in the future, thattomorrow's nuclear states are likely to do to one another what today's nuclear states have not done.A happy nuclear past leads many to expect an unhappy nuclear future. This is odd, and the oddityleads me to believe that we should reconsider how weapons affect the situation of their possessors.The Military Logic of Self-Help SystemsStates coexist in a condition of anarchy. Self-help is the principle of action in an anarchic order, andthe most important way in which states must help themselves is by providing for their own security.Therefore, in weighing the chances for peace, the first questions to ask are questions about the endsfor which states use force and about the strategies and weapons they employ. The chances of peacerise if states can achieve their most important ends without actively using force. War becomes lesslikely as the costs of war rise in relation to possible gains. Strategies bring ends and means together.How nuclear weapons affect the chances for peace is seen by considering the possible strategies ofstates.Force may be used for offence, for defence, for deterrence, and for coercion. Consider offence first.Germany and France before World War 1 provide a classic case of two adversaries each neglectingits defence and both planning to launch major attacks at the outset of war. France favoured offenceover defence, because only by fighting an offensive war could Alsace-Lorraine be reclaimed. Thisillustrates one purpose of the offence: namely, conquest. Germany favoured offence over defence.believing offence to be the best defence, or even the only defence possible. Hemmed in by twoadversaries. she could avoid fighting a two-front war only by concentrating her forces in the Westand defeating France before Russia could mobilize and move effectively into battle. This is what theSchlieffen plan called for. The Plan illustrates another purpose of the offence: namely, security.Even if security had been Germany's only goal, an offensive strategy seemed to be the way to obtainit.

The offence may have either or both of two aims: conquest and security. An offence may beconducted in either or in some combination of two ways: preventively or pre-emptively. If twocountries are unequal in strength and the weaker is gaining, the stronger may be tempted to strikebefore its advantage is lost. Following this logic, a country with nuclear weapons may be tempted todestroy the nascent force of a hostile country. This would be preventive war, a war launched againsta weak country before it can become disturbingly strong. The logic of pre-emption is different.Leaving aside the balance of forces, one country may strike another country's offensive forces toblunt an attack that it presumes is about to be made. If each of two countries can eliminate or dras-tically reduce the other's offensive forces in one surprise blow, tlien both of them are encouraged tomount sudden attacks, if only for fear that if one does not, the other will. Mutual vulnerability offorces leads to mutual fear of surprise attack by giving each power a strong incentive to strike first.French and German plans for war against each other emphasized prevention over preemption - tostrike before enemies can become fully ready to fight, but not to strike at their forces in order todestroy them before they can be used to strike back. Whether pre-emptive or preventive, anoffensive first strike is a hard one. as military logic suggests and history confirms Whoever strikesfirst does so to gain a decisive advantage. A pre-emptive strike is designed to eliminate or decisivelyreduce the opponent's ability to retaliate. A preventive strike is designed to defeat an adversarybefore he can develop and deploy his full potential might. Attacks. I should add, are not plannedaccording to military logic alone. Political logic may lead a country another country to attack evenin the absence of an expectation of military victory, as Egypt did in October of 1973.How can one state dissuade another state from attacking? In either or in some combination of twoways. One way to counter an intended attack is to build fortifications and to muster forces that lookforbiddingly strong. To build defences so patently strong that no one will try to destroy or overcomethem would make international life perfectly tranquil. I call this the defensive ideal. The other wayto inhibit a country's intended aggressive moves is to scare that country out of making them bythreatening to visit unacceptable punishment upon it. 'To deter' literally means to stop someonefrom doing something by frightening him. In contrast to dissuasion by defence, dissuasion bydeterrence operates by frightening a state out of attacking, not because of the difficulty of launchingan attack and carrying it home, but because the expected reaction of the attacked will result in one'sown severe punishment. Defence and deterrence are often confused. One frequently hearsstatements like this: 'A strong defence in Europe will deter a Russian attack'. What is meant is that astrong defence will dissuade Russia from attacking. Deterrence is achieved not through the abilityto defend but through the ability to punish. Purely deterrent forces provide no defence. The messageof a deterrent strategy is this: 'Although we are defenceless, if you attack we will punish you to anextent that more than cancels your gains'. Second-strike nuclear forces serve that kind of strategy.Purely defensive forces provide no deterrence. They offer no means of punishment. The message ofa defensive strategy is this: 'Although we cannot strike back, you will find our defences so difficultto overcome that you will dash yourself to pieces against them'. The Maginot Line was to serve thatkind of strategy.States may also use force for coercion. One state may threaten to harm another state not to deter itfrom taking a certain action but to compel one. Napoleon III threatened to bombard Tripoli if theTurks did not comply with his demands for Roman Catholic control of the Palestinian Holy Places.This is blackmail, which can now be backed by conventional and by nuclear threats.

Do nuclear weapons increase or decrease the chances of war? The answer depends on whethernuclear weapons permit and encourage states to deploy forces in ways that make the active use offorce more or less likely and in ways that promise to be more or less destructive. If nuclear weaponsmake the offence more effective and the blackmailer's threat more compelling, then nuclearweapons increase the chances of war—the more so the more widely they spread. Lf defence anddeterrence are made easier and more reliable by the spread of nuclear weapons, we may expect theopposite result. To maintain their security, states must rely on the means they can generate and thearrangements they can make for themselves. The quality of international life therefore varies withthe ease or the difficulty states experience in making themselves secure.Weapons and strategies change the situation of states in ways that make them more or less secure,as Robert Jervis has brilliantly shown. If weapons are not well suited for conquest, neighbours havemore peace of mind. According to the defensive-deterrent ideal, we should expect war to becomeless likely when weaponry is such as to make conquest more difficult, to discourage pre-emptiveand preventive war, and to make coercive threats less credible. Do nuclear weapons have thoseeffects? Some answers can be found by considering how nuclear deterrence and how nucleardefence may improve the prospects for peace.First, wars can be fought in the face of deterrent threats, but the higher the stakes and the closer acountry moves toward winning them, the more surely that country invites retaliation and risks itsown destruction. States are not likely to run major risks for minor gains. Wars between nuclearstates may escalate as the loser uses larger and larger warheads. Fearing that.states will want todraw back. Not escalation but de-escalation becomes likely. War remains possible. but victory inwar is too dangerous to fight for. If states can score only small gains because large ones riskretaliation, they have little incentive to fight.Second, states act with less care if the expected costs of war are low and with more care if they arehigh. In 1853 and 1854, Britain and France expected to win an easy victory if they went to waragainst Russia. Prestige abroad and political popularity at home would be gained. if not much else.The vagueness of their plans was matched by the carelessness of their acts. In blundering into theCrimean War they acted hastily on scant information, pandered to their people's frenzy for war,showed more concern for an ally's whim than for the adversary's situation, failed to specify thechanges in behaviour that threats were supposed to bring. and inclined towards testing strength firstand bargaining second. In sharp contrast, the presence of nuclear weapons makes Statesexceedingly cautious. Think of Kennedy and Khruschev in the Cuban missile crisis. Why fight ifyou can't win much and might lose everything?Third, the question demands a negative answer all the more insistently when the deter rentdeployment of nuclear weapons contributes more to a country's security than does conquest ofterritory. A country with a deter-rent strategy does not need the extent of territory required by acountry relying on a conventional defence in depth. A deterrent strategy makes it unnecessary for acountry to fight for the sake of increasing its security, and this removes a major cause of war.Fourth, deterrent effect depends both on one's capabilities and on the will one has to use them. Thewill of the attacked, striving to preserve its own territory, can ordinarily be presumed stronger thanthe will of the attacker striving to annex someone else's territory. Knowing this, the would-beattacker is further inhibited.

Certainty about the relative strength of adversaries also improves the prospects for peace. From thelate nineteenth century onwards the speed of technological innovation increased the difficulty ofestimating relative strengths and predicting the course of campaigns. Since World War II,technology has advanced even faster, but short of an antiballistic missile (ABM) breakthrough, thisdoes not matter very much. It does not disturb the American-Russian equilibrium because one side'smissiles are not made obsolete by improvements in the other side's missiles. In 1906 the BritishDreadnought, with the greater range and fire power of its guns, made older battleships obsolete.This does not happen to missiles. As Bernard Brodie put it: 'Weapons that do not have to fight theirlike do not become useless because of the advent of newer and superior types”. They do have tosurvive their like, but that is a much simpler problem to solve (see discussion below).Many wars might have been avoided had their outcomes been foreseen. 'To be sure,' Georg Simmelonce said, ‘the most effective presupposition for preventing struggle, the exact knowledge of thecomparative strength of the two parties, is very often only to be obtained by the actual fighting outof the conflict'. Miscalculation causes wars. One side expects victory at an affordable price, whilethe other side hopes to avoid defeat. Here the differences between conventional-multipolar andnuclear-bipolar worlds are fundamental. In the former, states are too often tempted to act onadvantages that are wishfully discerned and narrowly calculated. In 1914, neither Germany norFrance tried very hard to avoid a general war. Both hoped for victory even though they believedtheir forces to be quite evenly matched. In 1941, Japan, in attacking the United States, could hopefor victory only if a series of events that were possible but not highly probable took place. Japanwould grab resources sufficient for continuing the conquest of China and then dig in to defend alimited perimeter. Meanwhile, the United States and Britain would have to deal with Germany,which, having defeated the Soviet Union, would be supreme in Europe. Japan could then hope tofight a defensive war for a year or two until America, her purpose weakened, became willing tomake a compromise peace in Asia.Countries more readily run the risks of war when defeat, if it comes, is distant and is expected tobring only limited damage. Given such expectations, leaders do not have to be insane to sound thetrumpet and urge their people to be bold and courageous in the pursuit of victory. The outcome ofbattles and the course of campaigns are hard to foresee because so many things affect them,including the shifting allegiance and determination of alliance members. Predicting the result ofconventional wars has proved difficult.Uncertainty about outcomes does not work decisively against the fighting of wars in conventionalworlds. Countries armed with conventional weapons go to war knowing that even in defeat theirsuffering will be limited. Calculations about nuclear war are differently made. Nuclear worlds callfor and encourage a different kind of reasoning. If countries armed with nuclear weapons go to war,they do so knowing that their suffering may be unlimited. Of course, it also may not be. But that isnot the kind of uncertainty that encourages anyone to use force. In a conventional world, one isuncertain about winning or losing. In a nuclear world, one is uncertain about surviving or beingannihilated. If force is used and not kept within limits, catastrophe will result. That prediction iseasy to make because it does not require close estimates of opposing forces. The number of one'scities that can be severely damaged is at least equal to the number of strategic warheads anadversary can deliver. Variations of number mean little within wide ranges. The expected effect ofthe deterrent achieves an easy clarity because wide margins of error in estimates of probabledamage do not matter. Do we expect to lose one city or two, two cities or ten? When these are thepertinent questions, we stop thinking about running risks and start worrying about how to avoid

them. In a conventional world, deterrent threats are ineffective because the damage threatened isdistant, limited, and problematic. Nuclear weapons make military miscalculations difficult andpolitically pertinent prediction easy.Dissuading a would-be attacker by throwing up a good-looking defence may be as effective asdissuading him through deterrence. Beginning with President Kennedy and Secretary of DefenseMcNamara in the early 1960s, we have asked how we can avoid. or at least postpone, using nuclearweapons rather than how we c:an mount the most effective defence. NATO's attempt to keep adefensive war conventional in its initial stage may guarantee that nuclear weapons, if used, will beused in a losing cause and in ways that multiply destruction without promising victory. Early use ofvery small warheads may stop escalation. Defensive deployment, if it should fail to dissuade, wouldbring small nuclear weapons into use before the physical, political and psychological environmenthad deteriorated. The chances of de-escalation are high if the use of nuclear weapons is carefullyplanned and their use is limited to the battlefield. We have rightly put strong emphasis on strategicdeterrence, which makes large wars less likely, and wrongly slighted the question of whethernuclear weapons of low yield can effectively be used for defence, which would make any war at allless likely still.Lesser nuclear states, with choices tightly constrained by scarcity of resources, may be forced tomake choices that NATO has avoided, to choose nuclear defence or nuclear deterrence rather thanplanning to fight a conventional war on a large scale and to use nuclear weapons only whenconventional defences are breaking. Increased reliance on nuclear defence would decrease thecredibility of nuclear deterrence. That would be acceptable if a nuclear defence were seen to beunassailable. An unassailable defence is fully dissuasive. Dissuasion is what is wanted whether bydefence or by deterrence.The likelihood of war decreases as deterrent and defensive capabilities increase. Whatever thenumber of nuclear states, a nuclear world is tolerable if those states are able to send convincingdeterrent messages: It is useless to attempt to conquer because you will be severely punished. Anuclear world becomes even more tolerable if states are able to send convincing defensivemessages: It is useless to attempt to conquer because you cannot. Nuclear weapons and anappropriate doctrine for their use may make it possible to approach the defensive-deterrent ideal, acondition that would cause the chances of war to dwindle. Concentrating attention on thedestructive power of nuclear weapons has obscured the important benefits they promise to statestrying to coexist in a self-help world.Why Nations Want Nuclear WeaponsNations want nuclear weapons for one or more of seven reasons. First, great powers always counterthe weapons of other great powers, usually by imitating those who have introduced new weapons. Itwas not surprising that the Soviet Union developed atomic and hydrogen bombs, but rather that wethought the Baruch-Lilienthal plan might persuade her not to.Second, a state may want nuclear weapons for fear that its great-power ally will not retaliate if theother great power attacks. Although Britain when she became a nuclear power thought of herself asbeing a great one, her reasons for deciding later to maintain a nuclear force arose from doubts thatthe United States could be counted on to retaliate in response to an attack by the Soviet Union on

Europe and from Britain's consequent desire to place a finger on our nuclear trigger. As soon as theSoviet Union was capable of making nuclear strikes at American cities, West Europeans began toworry that America's nuclear umbrella no longer ensured that her allies would stay dry if it rained.Hugh GaitskeIl, as Leader of the Opposition, could say what Harold Macmillan, as Prime Minister,dared not: 'I do not believe that when we speak of our having to have nuclear weapons of our own itis because we must make a contribution to the deterrent of the West'. As he indicated, no con-tribution of consequence was made. Instead, he remarked, the desire for a nuclear force derives inlarge part 'from doubts about the readiness of the United States Government and the Americancitizens to risk the destruction of their cities on behalf of Europe'. Similar doubts provided thestrongest stimulus for France to become a nuclear power.Third, a country without nuclear allies will want nuclear weapons all the more if some of itsadversaries have them. So China and then India became nuclear powers, and Pakistan will probablyfollow.Fourth, a country may want nuclear weapons because it lives in fear of its adversaries' present orfuture conventional strength. This is reason enough for Israel's nuclear weapons, which mostauthorities assume she either has at hand or can quickly assemble.Fifth, some countries may find nuclear weapons a cheaper and safer alternative to runningeconomically ruinous and militarily dangerous conventional arms races. Nuclear weapons maypromise increased security and independence at an affordable price.Sixth, countries may want nuclear weapons for offensive purposes. This, however, is an unlikelymotivation for reasons given below.Finally, by building nuclear weapons a country may hope to enhance its international standing. Thisis thought to be both a reason for and a consequence of developing nuclear weapons. One mayenjoy the prestige that comes with nuclear weapons, and indeed a yearning for glory was not absentfrom de Gaulle's soul. But the nuclear military business is a serious one, and we may expect thatdeeper motives than desire for prestige lie behind the decision to enter it.Mainly for reasons two through five, new members will occasionally enter the nuclear club. Nuclearweapons will spread from one country to another in the future for the same reasons they havespread in the past. What effects may we expect?Relations among Nuclear NationsIn one important way nuclear weapons do change the relations of nations. Adversary states thatacquire them are thereby made more cautious in their dealings with each other. For the most part,however, the relations of nations display continuity through their transition from non-nuclear tonuclear status.Relations between the United States and the new nuclear states were much the same before and

after they exploded atomic devices, as Michael Nacht points out. Because America's relations withother nations are based on complex historical, economic, political, and military considerations, theyare not likely to change much when lesser parties decide to build nuclear forces. This continuity ofrelations suggests a certain ambivalence. The spread of nuclear weapons, though dreaded, promptsonly mild reactions when it happens. Our 'special relationship' with Britain led us to help heracquire and maintain nuclear forces. The distance tinged with distrust that marks our relations withFrance led us to oppose France's similar endeavours. China's nuclear forces neither preventedAmerican-Chinese rapprochement earlier nor prompted it later. American-Indian relations worsenedwhen America 'tilted' toward Pakistan during the India-Pakistan War of 1971. India's nuclearexplosion in 1974 neither improved nor worsened relations with the United States in the long term.Unlike Canada, we did not deny India access to our nuclear supplies. Again in 1980, PresidentCarter approved shipment of nuclear fuel to India despite her refusal to accept safeguards on all ofher nuclear facilities, as required by the Nuclear Non-Proliferation Act of 1978, a provision that thePresident can waive under certain circumstances. In asking Congress not to oppose his waiving therequirement, the President said this.' 'We must do all we reasonably can to promote stability in thearea and to bolster our relations with States there, particularly those that can play a role in checkingSoviet expansionism'. Nor did Pakistan's refusal to promise not to conduct nuclear tests prevent theUnited States from proposing to provide military aid after the Soviet Union's invasion ofAfghanistan in December of 1979.Stopping the spread of nuclear weapons has had a high priority for American governments, butclearly not the highest. In practice, other interests have proved to be more pressing. This is evidentin our relations with every country that has developed nuclear weapons, or appeared to be on theverge of doing so, from Britain onwards. One may expect that relations of friendship and enmity,that inclinations to help and to hinder, will carry over from the pre- to the post-nuclear relations ofnations.What holds for the United States almost surely holds for the Soviet Union. The Soviet Union hasstrongly supported efforts to stop the spread of nuclear weapons. She has good reasons to do so.Many potential nuclear states are both nearby and hostile from West Germany through Pakistan toSouth Korea. Others, like Iraq and India, are nearby and friendly. In international politics, however,friendliness and hostility are transient qualities. No doubt the Soviet Union would preferconventional to nuclear neighbours whatever their present leanings may be. But also, after thediscredit earned in occupying Afghanistan, the Soviet Union would like to repair relations withthird-world countries. If we had refused to supply nuclear fuel to India, would the Soviet Unionhave done so? Secretary of State Edmund Muskie and others thought so. For the Soviet Union, asfor the United States, other interests may weigh more heavily than her interest in halting the spreadof nuclear weapons.One may wonder, however, whether the quality of relations changes within alliances as some oftheir members become nuclear powers. Alliances relate nations to one another in specific and welldefined ways. By acquiring nuclear weapons a country is said to erode, and perhaps to wreck, thealliance to which it belongs. In part this statement mistakes effects for causes. Alliances areweakened by the doubts of some countries that another country will risk committing nationalsuicide through retaliation against a nuclear power that attacks an ally. Such doubts caused Britainto remain a nuclear power and France to become one, but it did not destroy NATO. The Allianceholds together because even its nuclear members continue to depend on the United States. Theygain strength from their nuclear weapons but remain weak in conventional arms and continue to bevulnerable economically. In an unbalanced world, when the weak feel threatened, they seek aid andprotection from the strong. The nuclear forces of Britain and France have their effects on the

Alliance without ending dependence on the United States.Nuclear weapons were maintained by Britain and acquired by France at least in part as triggers forAmerica's strategic deterrent. Given a sense of uncertainty combined with dependence, Europeansunderstandably strive to fashion their forces so as to ensure our commitment. They also wish todetermine the form the commitment takes and the manner of its execution. After all, an Americanchoice about how to respond to threats in Europe is a choice that affects the lives of Europeans andmay bring their deaths. Europeans want a large voice in American policies that may determine theirdestiny. By mounting nuclear weapons. Britain and France hope to decide when we will retaliateagainst the Soviet Union for acts committed in Europe. Since retaliation risks our destruction, weresist surrendering the decision.Alliances gain strength through a division of military labour. Within NATO, however, British andFrench duplication of American strategic nuclear weaponry on a minor scale adds little to thestrength of NATO. The most striking division of labour is seen in the different ways Europeancountries seek to influence American policy. Whether or not they are nuclear, lesser powersfeeling.threatened will turn to, or remain associated with, one or another of the great powers. Solong as West European countries fail to increase and concert their efforts, they remain weak and feelthreatened. Countries that are weak and threatened will continue to rely on the support of morepowerful ones and to hope that the latter will bear a disproportionate share of the burden. WestEuropean states have become accustomed to depending on the United States. Relations ofdependency are hardest to break where dependent states cannot shift from reliance on one greatpower to reliance on another. Under those circumstances, alliances endure even as nuclear weaponsspread among their members.From NATO'S experience we may conclude that alliances are not wrecked by the spread of nuclearweapons among their members. NATO accommodates both nuclear and conventional states in waysthat continue to evolve. Past evidence does not support the fear that alliances, which havecontributed an element of order to an anarchic world, are threatened by the spread of nuclearweapons. The Soviet Union won't permit the East European countries to become nuclear powersand the United States has accommodated two of her allies doing so, though uneasily in the case ofFrance. The spread of nuclear weapons among members of an alliance changes relations amongthem without breaking alliances apart.II. THE FURTHER SPREAD OF NUCLEAR WEAPONSContemplating the nuclear past gives grounds for hoping that the world will survive if furthernuclear powers join today's six or seven. This tentative conclusion is called into question by thewidespread belief that the infirmities of some nuclear states and the delicacy of their nuclear forceswill work against the preservation of peace and for the fighting of nuclear wars. The likelihood ofavoiding destruction as more states become members of the nuclear club is often coupled with thequestion who those states will be. What are the likely differences in situation and behaviour of newas compared to old nuclear powers?Nuclear Weapons and Domestic Stability

What are the principal worries? Because of the importance of controlling nuclear weapons—ofkeeping them firmly in the hands of reliable officials—rulers of nuclear states may become moreauthoritarian and ever more given to secrecy. Moreover, some potential nuclear states are notpolitically strong and stable enough to ensure control of the weapons and of the decision to usethem. If neighhouring, hostile, unstable states are armed with nuclear weapons, each will fear attackby the other. Feelings of insecurity may lead to arms races that subordinate civil needs to militarynecessities. Fears are compounded by the danger of internal coups in which the control of nuclearweapons may he the main object of the struggle and the key to political power. Under these fearfulcircumstances to maintain governmental authority and civil order may be impossible. Thelegitimacy of the state and the loyalty of its citizenry may dissolve because the state is no longerthought to be capable of maintaining external security and internal order. The first fear is that statesbecome tyrannical; the second, that they lose control. Both these fears may be realized, either indifferent states or, indeed, in the same state at different times.What can one say? Four things primarily. First, Possession of nuclear weapons may slow arms racesdown, rather than speed them up, a possibility considered later. Second, for less developed countriesto build nuclear arsenals requires a long lead time. Nuclear power and nuclear weaponsprogrammes, like population policies, require administrative and technical teams able to formulateand sustain programmes of considerable cost that pay off only in the long run. The more unstable agovernment, the shorter becomes the attention span of its leaders. They have to deal with today'sproblems and hope for the best tomorrow. In countries where political control is most difficult tomaintain, governments are least likely to initiate nuclear-weapons programmes. In such states,soldiers help to maintain leaders in power or try to overthrow them. For those purposes nuclearweapons are not useful. Soldiers who have political clout, or want it, are less interested in nuclearweapons than they are in more immediately useful instruments of political control. They are notscientists and technicians. They like to command troops and squadrons. Their vested interests are inthe military's traditional trappings.Third, although highly unstable states are unlikely to initiate nuclear projects, such projects, begunin stable times, may continue through periods of political turmoil and succeed in producing nuclearweapons. A nuclear state may be unstable or may become so. But what is hard to comprehend iswhy, in an internal struggle for power, any of the contenders should start using nuclear weapons.Who would they aim at? How would they use them as instruments for maintaining or gainingcontrol? I see little more reason to fear that one faction or another in some less developed countrywill fire atomic weapons in a struggle for political power than that they will be used in a crisis ofsuccession in the Soviet Union or China. One or another nuclear state will experience uncertainty ofsuccession, fierce struggles for power, and instability of regime. Those who fear the worst have notshown with any plausibility how those expected events may lead to the use of nuclear weapons.Fourth, the possibility of one side in a civil war firing a nuclear warhead at its opponent'sstronghold nevertheless remains. Such an act would produce a national tragedy. not an internationalone. This question then arises: Once the weapon is fired, what happens next? The domestic use ofnuclear weapons is, of all the uses imaginable, least likely to lead to escalation and to threaten thestability of the central balance. The United States and the Soviet Union, and other countries as well,would have the strongest reasons to issue warnings and to assert control.Nuclear weapons and regional stability

Nuclear weapons are not likely to be used at home. Are they likely to be used abroad? As nuclearweapons spread, what new causes may bring effects different from and worse than those knownearlier in the nuclear age? This section considers five ways in which the new world is expected todiffer from the old and then examines the prospects for, and the consequences of, new nuclear statesusing their weapons for blackmail or for fighting an offensive war.In what ways may the actions and interactions of new nuclear states differ from those of old nuclearpowers? First. new nuclear states may come in hostile pairs and share a common border. WhereStates are bitter enemies one may fear that they will be unable to resist using their nuclear weaponsagainst each other. This is a worry about the future that the past does not disclose. The Soviet Unionand the United States, and the Soviet Union and China, are hostile enough; and the latter pair sharea long border. Nuclear weapons have caused China and the Soviet Union to deal cautiously witheach other. But bitterness among some potential nuclear states, so it is said, exceeds thatexperienced by the old ones. Playing down the bitterness sometimes felt by the United States, theSoviet Union, and China requires a creative reading of history. Moreover, those who believe thatbitterness causes wars assume a close association that is seldom found between bitterness amongnations and their willingness to run high risks.Second, some new nuclear states may have governments and societies that are not well rooted. If acountry is a loose collection of hostile tribes, if its leaders form a thin veneer atop a people partlynomadic and with an authoritarian history, its rulers may be freer of constraints than, and havedifferent values from, those who rule older and more fully developed polities. Idi Amin andMuammar el-Qaddafi fit into these categories, and they are favourite examples of the kinds of rulerswho supposedly cannot be trusted to manage nuclear weapons responsibly. Despite wild rhetoric;aimed at foreigners, however, both of these 'irrational' rulers became cautious and modest whenpunitive actions against them might have threatened their ability to rule. Even though Amin lustilyslaughtered members of tribes he disliked, he quickly stopped goading Britain once the sending ofher troops appeared to be a possibility. Qaddafi has shown similar restraint. He and Anwar Sadathave been openly hostile since 1973. In July of 1977 both sides launched commando attacks and airraids, including two large air strikes by Egypt on Libya's el Adem airbase. Neither side let theattacks get out of hand. Qaddafi showed himself to he forbearing and amenable to mediation byother Arab leaders. Shai Feldman uses these and other examples to argue that Arab leaders aredeterred from taking inordinate risks not because they engage in intricate rational calculations butsimply because they, like other rulers, are 'sensitive to costs'.Many Westerners who write fearfully about a future in which third-world countries have nuclearweapons seem to view their people in the once familiar imperial manner as 'lesser breeds withoutthe law'. As is usual with ethnocentric views, speculation takes the place of evidence. How do weknow, someone has asked, that a nuclear-armed and newly hostile Egypt or a nuclear-armed andstill hostile Syria would not strike to destroy Israel at the risk of Israeli bombs falling on some oftheir cities? More than a quarter of Egypt's people live in four cities: Cairo, Alexandria, Giza, andAswan. More than a quarter of Syria's live in three: Damascus. Aleppo, and Homs. Whatgovernment would risk sudden losses of such proportion or indeed of much lesser proportion?Rulers want to have a country that they can continue to rule. Some Arab country might wish thatsome other Arab country would risk its own destruction for the sake of destroying Israel, but thereis no reason to think that any Arab country would do so. One may be impressed that, despite amplebitterness, Israelis and Arabs have limited their wars and accepted constraints placed on them byothers. Arabs did not marshal their resources and make an all-out effort to destroy Israel in the yearsbefore Israel could strike back with nuclear warheads. We cannot expect countries to risk more inthe presence of nuclear weapons than they have in their absence.

Third. many fear that states that are radical at home will recklessly use their nuclear weapons inpursuit of revolutionary ends abroad. States that are radical at home. however, may not be radicalabroad. Few states have been radical in the conduct of their foreign policy, and fewer have remainedso for long. Think of the Soviet Union and the People's Republic of China. States coexist in acompetitive arena. The pressures of competition cause them to behave in ways that make the threatsthey face manageable, in ways that enable them to get along. States can remain radical in foreignpolicy only if they are overwhelmingly strong—as none of the new nuclear states will be—or iftheir radical acts fall short of damaging vital interests of nuclear powers. States that acquire nuclearweapons will not be regarded with indifference. States that want to be freewheelers have to stay outof the nuclear business. A nuclear Libya, for example, would have to show caution, even in rhetoric,lest she suffer retaliation in response to someone else's anonymous attack on a third state. That state,ignorant of who attacked, might claim that its intelligence agents had identified Libya as the culpritand take the opportunity to silence her by striking a conventional or nuclear blow. Nuclear weaponsinduce caution, especially in weak states.Fourth, while some worry about nuclear states coming in hostile pairs, others worry that the bipolarpattern will not be reproduced regionally in a world populated by larger numbers of nuclear states.The simplicity of relations that obtains when one party has to concentrate its worry on only oneother, and the ease of calculating forces and estimating the dangers they pose, may be lost. Thestructure of international politics, however, will remain bipolar so long as no third state is able tocompete militarily with the great powers. Whatever the structure, the relations of states run invarious directions. This applied to relations of deterrence as soon as Britain gained nuclearcapabilities. It has not weakened deterrence at the centre and need not do so regionally. The SovietUnion now has to worry lest a move made in Europe cause France and Britain to retaliate, thuspossibly setting off American forces. She also has to worry about China's forces. Such worries atonce complicate calculations and strengthen deterrence.Fifth, in some of the new nuclear states, civil control of the military maybe shaky. Nuclear weaponsmay fall into the hands of military officers more inclined than civilians to put them to offensive use.This again is an old worry. I can see no reason to think that civil control of the military is secure inthe Soviet Union given the occasional presence of serving officers in the Politburo and some knownand some surmised instances of military intervention in civil affairs at critical times. And in thePeople's Republic of China military and civil branches of government have been not separated butfused. Although one may prefer civil control, preventing a highly destructive war does not requireit. What is required is that decisions be made that keep destruction within bounds, whetherdecisions are made by civilians or soldiers. Soldiers may he more cautious than civilians. Generalsand admirals do not like uncertainty, and they do not lack patriotism. They do not like to fightconventional wars under unfamiliar conditions. The offensive use of nuclear weapons multipliesuncertainties. Nobody knows what a nuclear battlefield would look like, and nobody knows whathappens after the first city is hit. Uncertainiy about the course that a nuclear war might follow,along with the certainty that destruction can he immense, strongly inhibits the first use of nuclearweapons.Examining the supposedly unfortunate characteristics of new nuclear states removes some of one’sworries. One wonders why their civil and military leaders should be less interested in avoiding self-destruction than leaders of other states have been. Nuclear weapons have never been used in aworld in which two or more states possessed them. Still, one’s feeling that something awful willhappen as new nuclear powers are added to the present group is not easily quieted. The fear remains

that one state or another will fire its weapons in a coolly calculated pre-emptive strike, or fire themin a moment of panic, or use them to launch a preventive war. These possibilities are examined inthe next section. Nuclear weapons may also back a policy of blackmail, or be set off anonymously,or be used in a combined conventional-nuclear attack.Consider blackmail first. Two conditions make for the success of nuclear blackmail. First, whenonly one country had nuclear weapons, threats to use them had more effect. Thus, PresidentTruman’s nuclear threats may have levered the Soviet Union’s troops out of Azerbaijan in 1946.Second, if a country has invested troops and suffered losses in a conventional war, its nuclearblackmail may work. In 1953, Eisenhower and Dulles may have convinced Russia and China thatthey would widen the Korean War and intensify it by using nuclear weapons if a settlement werenot reached. In Korea, we had gone so far that the threat to go further was plausible. The black-mailer’s nuclear threat is not a cheap way of working one’s will. The threat is simply incredibleunless a considerable investment has already been made. Dulles’s speech of 12 January 1954seemed to threaten massive retaliation in response to mildly bothersome actions by others. Thesuccessful seige of Dien Bien Pbu in the spring of that year showed the limitations of such threats.Capabilities foster policies that employ them. But monstrous capabilities foster monstrous policies,which when contemplated are seen to be too horrible to carry through. Imagine an Arab state threat-ening to strike Tel Aviv if the West Bank is not evacuated by Israelis. No state can make the threatwith credibility because no state can expect to execute the threat without danger to themselves.Some have feared that nuclear weapons may be fired anonymously—by radical Arab states, forexample, to attack an Israeli city so as block a peace settlement. But the state exploding the warheadcould not be sure of remaining unidentified. Even if a country’s leaders persuade themselves thatchances of retaliation are low, who would run the risk? Once two or more countries have nuclearweapons, the response to nuclear threats, even against non-nuclear states, becomes unpredictable.Although nuclear weapons are poor instruments for blackmail, would they not provide a cheap anddecisive offensive force against a conventionally armed enemy? Some people think that SouthKorea wants, and that earlier the Shah’s Iran had wanted, nuclear weapons for offensive use, Yetone cannot say why South Korea would use nuclear weapons against fellow Koreans while trying toreunite them nor how she could use nuclear weapons against the North, knowing that China andRussia might retaliate. And what goals could a conventionally strong Iran have entertained thatwould have tempted her to risk using nuclear weapons? A country that takes the nuclear offensivehas to fear an appropriately punishing strike by someone. Far from lowering the expected cost ofaggression, a nuclear offence even against a non-nuclear state raises the possible costs of aggressionto incalculable heights because the aggressor cannot be sure of the reaction of other nuclear powers.Nuclear weapons do not make nuclear war a likely prospect, as history has so far shown. The pointmade when discussing the domestic use of nuclear weapons, however, bears repeating. No one cansay that nuclear weapons will never be used. Their use, although unlikely, is always possible. Inasking what the spread of nuclear weapons will do to the world, we are asking about the effects tobe expected as a larger number of relatively weak states get nuclear weapons. If such states usenuclear weapons, the world will not end. And the use of nuclear weapons by lesser powers wouldhardly trigger them elsewhere, with the US and the USSR becoming involved in ways that mightshake the central balance.

Deterrence with Small Nuclear ForcesA number of problems arc thought to attend the efforts of minor powers to use nuclear weapons fordeterrence. In this section, I ask how hard these problems are for new nuclear states to solve.The Forces Required for DeterrenceIn considering the physical requirements of deterrent forces, we should recall the differencebetween prevention and pre-emption. A preventive war is launched by a stronger state against aweaker one that is thought to be gaining strength. A pre-emptive strike is launched by one state toblunt an attack that another state is presumably preparing to launch.The first danger posed by the spread of nuclear weapons would seem to be that each new nuclearstate may tempt an old one to strike preventively in order to destroy an embryonic nuclearcapability before it can become militarily effective. Because of America’s nuclear arsenal, theSoviet Union could hardly have destroyed the budding forces of Britain and France; but the UnitedStates could have struck the Soviet Union’s early nuclear facilities, and the United States and theSoviet Union could have struck China’s. Such preventive strikes have been treated as more thanabstract possibilities. When Francis P. Matthews was President Truman’s Secretary of the Navy, hemade a speech that seemed to favour our waging a preventive war. The United States, he urged,should be willing to pay ‘even the price of instituting a war to compel cooperation for peace’.The United States and the Soviet Union considered making preventive strikes against China early inher nuclear career. Preventive strikes against nuclear installations can also be made by non-nuclearstates and have sometimes been threatened. Thus President Nasser warned Israel in 1960 that Egyptwould attack if she were sure that Israel was building a bomb. ‘It is inevitable’, he said, ‘that weshould attack the base of aggression, even if we have to mobilize four million to destroy it’.The uneven development of the forces of potential and of new nuclear states creates occasions thatseem to permit preventive strikes and may seem to invite them. Two stages of nuclear developmentshould be distinguished. First, a country may be in an early stage of nuclear development and beobviously unable to make nuclear weapons. Second, a country may be in an advanced stage ofnuclear development, and whether or not it has some nuclear weapons may not be surely known.All of the present nuclear countries went through both stages, yet until Israel struck Iraq’s nuclearfacility in June of 1981 no one had launched a preventive strike. A number of reasons combinedmay account for the reluctance of States to strike in order to prevent adversaries from developingnuclear forces. A preventive strike would seem to be most promising during the first stage ofnuclear development. A state could strike without fearing that the country it attacked would return anuclear blow. But would one strike so hard as to destroy the very potential for future nucleardevelopment? If not, the country struck could simply resume its nuclear career. If the blow struck isless than devastating, one must be prepared to repeat it or to occupy and control the country. To doeither would be difficult and costly.In striking Iraq, Israel showed that a preventive strike can be made, something that was not indoubt. Israel’s act and its consequences however, make clear that the likelihood of usefulaccomplishment is low. Israel’s strike increased the determination of Arabs to produce nuclear

weapons. Arab states that may attempt to do so will now be all the more secretive and circumspect.Israel’s strike, far from foreclosing Iraq’s nuclear future, gained her the support of some other Arabstates in pursuing it. And despite Prime Minister Begin’s vow to strike as often as need be, the risksin doing so would rise with each occasion.A preventive strike during the second stage of nuclear development is even less promising than apreventive strike during the first stage. As more countries acquire nuclear weapons, and as morecountries gain nuclear competence through power projects, the difficulties and dangers of makingpreventive strikes increase. To know for sure that the country attacked has not already produced orotherwise acquired some deliverable warheads becomes increasingly difficult. If the countryattacked has even a rudimentary nuclear capability, one’s own severe punishment becomes possible.Fission bombs may work even though they have not been tested, as was the case with the bombdropped on Hiroshima. Israel has apparently not tested weapons, yet Egypt cannot know whetherIsrael has zero, ten, or twenty warheads. And if the number is zero and Egypt can be sure of that,she would still not know how many days are required for assembling components that may be onhand.Preventive strikes against states that have, or may have, nuclear weapons are hard to imagine, butwhat about pre-emptive ones? The new worry in a world in which nuclear weapons have spread isthat states of limited and roughly similar capabilities will use them against one another. They do notwant to risk nuclear devastation anymore than we do. Preemptive strikes nevertheless seem likelybecause we assume that their forces will be ‘delicate’. With delicate forces, states are tempted tolaunch disarming strikes before their own forces can be struck and destroyed.To be effective a deterrent force must meet three requirements. First, a part of the force must appearto be able to survive an attack and launch one of its own. Second, survival of the force must notrequire early firing in response to what may be false alarms. Third, weapons must not be susceptibleto accidental and unauthorized use. Nobody wants vulnerable. hair-trigger, accident-prone forces.Will new nuclear states find ways to hide their weapons, to deliver them, and to control them? Willthey be able to deploy and manage nuclear weapons in ways that meet the physical requirements ofdeterrent forces?The United States even today worries about the vulnerability of its vast and varied arsenal. Will notnew nuclear states, slightly and crudely armed, be all the more worried about the survival of theirforces? In recent years, we have exaggerated the difficulty of deterrence by shifting attention fromsituations to weaponry and from weapons systems to their components. Some Americans areconcerned about the vulnerability of our strategic system because its land-based component can bestruck and perhaps largely destroyed by the Soviet Union in the middle 1980s. If the Soviet Uniontried that, we would still have thousands of warheads at sea and thousands of bombs in the air. TheSoviet Union could not be sure that we would fail to launch on warning or fail to retaliate later.Uncertainty deters, arid there would be plenty of uncertainty about our response in the minds of theSoviet Union’s leaders.In McNamara’s day and earlier the term ‘counterforce’ had a clear and precise meaning. Country Awas said to have a counterforce capability if by striking first it could reduce country B’s missilesand bombers to such small numbers that country A would be reluctantly willing to accept the fullforce of B’s retaliation. In this respect, as in others, strategic discourse now lacks the clarity andprecision it once had. Whether in a conventional or a nuclear world, one cannot usefully compare

some components of a nation’s military forces without taking account of what other componentscan do. Both the United States and the Soviet Union have strategic nuclear weapons that candestroy some of the other sides strategic nuclear weapons. Neither the United States nor the SovietUnion can reduce the other side’s strategic forces to the point where it no longer retains an immensecapability for striking at cities and a considerable capability for striking at military targets as well.That we have ten thousand warheads to the Soviet Union’s six thousand makes us no worse and nobetter off than we were when the ratio was even more favourable. That the throw-weight of theSoviet Union’s missiles exceeds ours by several times makes us no better and no worse off than itwould be were the ratio to be reversed.Deterrent forces are seldom delicate because no state wants delicate forces and nuclear forces caneasily be made sturdy. Nuclear weapons are fairly small and light. They are easy to hide and tomove. Early in the nuclear age, people worried about atomic bombs being concealed in packingboxes and placed in holds of ships to be exploded when a signal was given. Now more than everpeople worry about terrorists stealing nuclear warheads because various states have so many ofthem. Everybody seems to believe that terrorists are capable of hiding bombs. Why should states beunable to do what terrorist gangs are though to be capable of?It is sometimes claimed that the few bombs of a new nuclear state create a greater danger of nuclearwar than additional thousands for the United States and the Soviet Union. Such statements assumethat pre-emption of a small force is easy. It is so only if the would-be attacker knows that theintended victim’s warheads are few in number, knows their exact number and locations, and knowsthat they will not be moved or fired before they are struck. To know all of these things, and to knowthat you know them for sure, is exceedingly difficult. How can military advisers promise the fullsuccess of a disarming first strike when the penalty for slight error may be so heavy? In 1962,Tactical Air Command promised that an American strike against Soviet missiles in Cuba wouldcertainly destroy 90% of them but would not guarantee 100%. In the best case a first strike destroysall of a country’s deliverable weapons. In the worst case, some survive and can still be delivered.If the survival of nuclear weapons requires their dispersal and concealment, do not problems ofcommand and control become harder to solve? Americans think so because we think in terms oflarge nuclear arsenals. Small nuclear powers will neither have them nor need them. Lesser nuclearstates might deploy, say, ten real weapons and ten dummies, while permitting other countries toinfer that the numbers are larger. The adversary need only believe that some warheads may survivehis attack and be visited on him. That belief should not be hard to create without making commandand control unreliable. All nuclear countries must live through a time when their forces are crudelydesigned. All countries have so far been able to control them. Relations between the United Statesand the Soviet Union. and later among the United States, the Soviet Union, and China, were at theirbitterest just when their nuclear forces were in early stages of development, were unbalanced, werecrude and presumably hard to control. Why should we expect new nuclear states to experiencegreater difficulties than the old ones were able to cope with? Moreover, although some of the newnuclear states may be economically and technically backward, they will either have an expert andhighly trained group of scientists and engineers or they will not produce nuclear weapons. Even ifthey buy the weapons, they will have to hire technicians to maintain and control them. We do nothave to wonder whether they will take good care of their weapons. They have every incentive to doso. They will not want to risk retaliation because one or more of their warheads accidentally strikesanother country.Hiding nuclear weapons and keeping them under control are tasks for which the ingenuity of

numerous states is adequate. Nor are means of delivery difficult to devise or procure. Bombs can bedriven in by trucks from neighbouring countries. Ports can be torpedoed by small boats lying offshore. Moreover, a thriving arms trade in ever more sophisticated military equipment providesready access to what may be wanted, including planes and missiles suited nuclear warhead deliveryLesser nuclear states can pursue deterrent strategies effectively. Deterrence requires the ability toinflict unacceptable damage on another country. ‘Unacceptable damage’ to the Soviet Union wasvariously defined by Robert McNamara as requiring the ability to destroy a fifth to a fourth of’ herpopulation and a half to two-thirds of her industrial capacity. American estimates of what is requiredfor deterrence have been absurdly high. To deter, a country need not appear to be able to destroy afourth to a half of another country, although in some cases that might be easily done. Would Libyatry to destroy Israel’s nuclear weapons at the risk of two bombs surviving to fall on Tripoli andBengazi? And what would be left of Israel if Tel Aviv and Haifa were destroyed?The weak can deter one another. But can the weak deter the strong? Raising the question of China’sability to deter the Soviet Union highlights the issue. The population and industry of most Statesconcentrate in a relatively small number of centres. This is true of the Soviet Union. A major attackon the top ten cities of the Soviet Union would get 25% of its industrial capacity and 25% of itsurban population. Geoffrey Kemp in 1974 concluded that China would probably be able to strike onthat scale. And, I emphasize again, China need only appear to be able to do it. A low probability ofcarrying a highly destructive attack home is sufficient for deterrence. A force of an impreciselyspecifiable minimum capability is nevertheless needed.In a 1979 study, Justin Galen (pseud.) wonders whether the Chinese have a force physically capableof deterring the Soviet Union. He estimates that China has 60 to 80 medium-range and 60 to 80intermediate-range missiles of doubtful reliability and accuracy and 80 obsolete bombers. Herightly points out that the missiles may miss their targets even if tired at cities and that the bombersmay not get through the Soviet Union’s defences. Moreover, the Russians may be able to pre-empt,having almost certainly ‘located virtually every Chinese missile, aircraft, weapons storage area andproduction facility’. But surely Russian leaders reason the other way around. To locate virtually allmissiles and aircraft is not good enough. Despite inaccuracies, a few Chinese missiles may hitRussian cities, and some bombers may get through. Not much is required to deter. What political-military objective is worth risking Vladivostock, Novosibirsk. and Tomsk, with no way of beingsure that Moscow will not go as well?Prevention and pre-emption are difficult games because the costs are so high if the games are notperfectly played. Inhibitions against using nuclear forces for such attacks are strong, although onecannot say they are absolute. Some of the inhibitions are simply human. Can country A findjustification for a preventive or pre-emptive strike against B if B, in acquiring nuclear weapons, isimitating A? The leader of a country that launches a preventive or preemptive strike courtscondemnation by his own people, by the world’s people, and by history. Awesome acts are hard toperform. Some of the inhibitions are political. As Bernard Brodie tirelessly and wisely said, war hasto find a political objective that is commensurate with its cost. Clausewitz’s central tenet remainsvalid in the nuclear age. Ultimately, the inhibitions lie in the impossibility of know ing for sure that adisarming strike will totally destroy an opposing force and in the immense destruction even a fewwarheads can wreak.The Credibility of Small Deterrent Forces

The credibility of weaker countries’ deterrent threats has two faces. The first is physical. Will suchcountries be able to construct and protect a deliverable force? We have found that they can readilydo so. The second is psychological. Will an adversary believe that retaliation threatened will becarried out?Deterrent threats backed by second-strike nuclear forces raise the expected costs of war to suchheights that war becomes unlikely. But deterrent threats may not be credible. In a world where twoor more countries can make them, the prospect of mutual devastation makes it difficult, orirrational, to execute threats should the occasion for doing so arise. Would it not be senseless to risksuffering further destruction once a deterrent force had failed to deter? Believing that it would be,an adversary may attack counting on the attacked country’s unwillingness to risk initiating adevastating exchange by its own retaliation. Why retaliate once a threat to do so has failed? If one’spolicy is to rely on forces designed to deter, then an attack that is nevertheless made shows thatone’s reliance was misplaced. The course of wisdom may be to pose a new question: What is thebest policy once deterrence has failed? One gains nothing by destroying an enemy’s cities. Instead,in retaliating, one may prompt the enemy to unleash more warheads. A ruthless aggressor maystrike believing that the leaders of the attacked country are capable of following such a ‘rational’line of thought. To carry out the threat that was ‘rationally’ made may be ‘irrational’. This old worryachieved new prominence as the strategic capabilities of the Soviet Union approached those of’ theUnited States in the middle 1970s. The Soviet Union, some feared, might believe that the UnitedStates would be self-deterred.Much of the literature on deterrence emphasizes the problem of achieving the credibility on whichdeterrence depends and the danger of relying on a deterrent of uncertain credibility. One earliersolution to the problem was found in Thomas Sche!ling’s notion of ‘the threat that leaves somethingto chance’. No state can know for sure that another state will refrain from retaliating even whenretaliation would be irrational. No state can bet heavily on another state’s rationality. BernardBrodie put the thought more directly, while avoiding the slippery notion of rationality. Rather thanask what it may be rational or irrational for governments to do, the question he asked, and repeatedin various ways over the years, was this: How do governments behave in the presence of awesomedangers? His answer was ‘very carefully’.To ask why a country should carry out its deterrent threat once deterrence has failed is to ask thewrong question. The question suggests that an aggressor may attack believing that the attackedcountry may not retaliate. This invokes the conventional logic that analysts find so hard to forsake.In a conventional world, a country can sensibly attack if it believes that success is probable. In anuclear world, a country cannot sensibly attack unless it believes that success is assured. Anattacker is deterred even if he believes only that the attacked may retaliate. Uncertainty of response,not certainty, is required for deterrence because, if retaliation occurs, one risks losing all. In anuclear world, we should look less at the retaliators conceivable inhibitions and more at thechallenger’s obvious risks.One may nevertheless wonder, as Americans recently have, whether retaliatory threats remaincredible if the strategic forces of the attacker are superior to those of the attacked. Will anunsuccessful defender in a conventional war nave the courage to unleash its deterrent force, usingnuclear weapons first against a country having superior strategic forces? Once more this asks thewrong question. The previous paragraph urged the importance of shifting attention from thedefender’s possible inhibitions to the aggressor’s unwillingness to run extreme risks. This paragraphurges the importance of shifting attention from the defender’s courage to the different valuations

that defenders and attackers place on the stakes. An attacked country will ordinarily value keepingits own territory more highly than an attacker will value gaining some portion of’ it. Given second-strike capabilities, it is not the balance of forces but the courage to use them that counts. Thebalance or imbalance of strategic forces affects neither the calculation of danger nor the question ofwhose will is the stronger. Second-strike forces have to be seen in absolute terms. The question ofwhose interests are paramount will then determine whose will is perceived as being the stronger.Emphasizing the importance of the ‘balance of resolve’, to use Glenn Snyder’s apt phrase, raisesquestions about what a deterrent force covers and what it does not. In answering these questions, wecan learn something from the experience of the last three decades. The United States and the SovietUnion limited and modulated their provocative acts, the more carefully so when major values forone side or the other were at issue. This can be seen both in what they have and in what they havenot done. Whatever support the Soviet Union gave to North Korea’s initial attack on the South wasgiven after Secretary of State Acheson, the Joint Chiefs of Staff, General MacArthur, and theChairman of the Senate Foreign Relations Committee all explicitly excluded both South Korea andTaiwan from America’s defence perimeter. The United States, to take another example, could fightfor years on a large scale in South-East ASia because neither success nor failure mattered muchinternationally. Victory would not have made the world one of American hegemony. Defeat wouldnot have made the world one of Russian hegemony. No vital interest of either great power was atstake, as both Kissinger and Brezhnev made clear at the time. One can fight without fearingescalation only where little is at stake. And that is where the deterrent does not deter.Actions at the periphery can safely be bolder than actions at the centre. In contrast, where much is atstake for one side, the other side moves with care. Trying to win where winning would bring thecentral balance into question threatens escalation and becomes too risky to contemplate. The UnitedStates is circumspect when East European crises impend. Thus Secretary of State Dulles assured theSoviet Union when Hungarians rebelled in October of 1956 that we would not interfere with effortsto suppress them. And the Soviet Union’s moves in the centre of Europe are carefully controlled.Thus her probes in Berlin have been tentative, reversible, and ineffective. Strikingly, the long borderbetween East and West Europe—drawn where borders earlier proved unstable—has been free evenof skirmishes in all of the years since the Second World War.Both of the nuclear great powers become watchful and wary when events occur that may get out ofcontrol. The strikes by Polish workmen that began in August of 1980 provide the most recentillustration of this. The Soviet Union, her diplomats privately said, was ‘determined to find apeaceful solution’. And a senior Carter Administration specialist on the Soviet Union was quoted asfollows: ‘it is a very explosive situation. Everyone is aware of it, and they are all reluctant to strikea match’. Even though many steps would intervene between workers’ strikes and the beginning ofany fighting at all in the Centre of Europe, both the Soviet Union and the United States showedgreat caution from the outset. By political and military logic, we can understand why nuclearweapons induce great caution, and we can confirm that they do by observing the differences ofbehaviour between great powers in nuclear and great powers in conventional worlds.Contemplating American and Russian postwar behaviour, and interpreting it in terms of nuclearlogic, suggests that deterrence extends to vital interests beyond the homeland more easily thanmany have thought. The United States cares more about Western Europe than the Soviet Uniondoes. The Soviet Union cares more about Eastern Europe than the United States does.Communicating the weight of one side’s concern as compared to the other side’s has been easilyenough done when the matters at hand affect the United States and the Soviet Union directly. For

this reason, Western Europe’s anxiety over the coverage it gets from American strategic forces,while understandable, is exaggerated. The United States might well retaliate should the SovietUnion make a major military move against a NATO country, and that is enough to deter.The Problem of Extended DeterrenceHow far from the homeland does deterrence extend? One answers that question by defining theconditions that must obtain if deterrent threats are to be credited. First, the would-be attacker mustbe made to see that the deterrer considers the interests at stake to be vital ones. One cannot assumethat countries will instantly agree on the question of whose interests are vital. Nuclear weapons,however, strongly incline them to grope for de facto agreement on the answer rather than to fightover it.Second, political stability must prevail in the area that the deterrent is intended to cover. It thethreat to a regime is in good part from internal factions, then an outside power may risk supportingg one of them even in the face of deterrent threats. The credibility of a deterrent force requires boththat interests be seen to be vital and that it is the attack from outside that threatens them. Giventhese conditions, the would-be attacker provides both the reason to retaliate and the target forretaliation. Deterrence gains in credibility the more highly valued the interests covered seem to be.The problem of stretching a deterrent, which has so agitated the western alliance, is not a problemfor lesser nuclear states. Their problem is to protect not others but themselves. Many have fearedthat lesser nuclear states would be the first to break the nuclear taboo and that they would use theirnuclear weapons irresponsibly. I expect just the opposite. Weak states find it easier than strongstates to establish their credibility. Not only will they not be trying to stretch their deterrent forcesto cover others, but also their vulnerability to conventional attacks lends credence to their nuclearthreats. Because in a conventional war they can lose so much so fast, it is easy to believe that theywill unleash a deterrent force even at the risk of receiving a nuclear blow in return. With deterrentforces, the party that is absolutely threatened prevails. Use of nuclear weapons by lesser states willcome only if survival is at stake. And this should be called not irresponsible but responsible use.An opponent who attacks what is unambiguously mine risks suffering great distress if they havesecond-strike forces. This statement has important implications for both the deterrer and thedeterred. Where territorial claims are shadowy and disputed, deterrent writs do not run. As StevenJ. Rosen has said: ‘It is difficult to imagine Israel committing national suicide to hold on to AbuRudeis or Hebron or Mount Hermon. Attacks on Israel’s occupied lands would be imaginable evenif she admitted having nuclear weapons. Establishing the credibility of a deterrent force requiresmoderation of territorial claims on the part of the would-be deterrer. For modest states, weaponswhose very existence works strongly against their use are just what is wanted.In a nuclear world, conservative would-be attackers will be prudent, but will all would-be attackersbe conservative? A new Hitler is not unimaginable. Would the presence of nuclear weapons havemoderated Hitler’s behaviour? Hitler did not start World War II in order to destroy the ThirdReich. Indeed, he was surprised and dismayed by the British and French declaration of war onPoland’s behalf. After all, the western democracies had not come to the aid of a geographicallydefensible and militarily strong Czechoslovakia. Why then should they have declared war on behalfof a less defensible Poland and against a Germany made stronger by the incorporation of

Czechoslovakia’s armour? From the occupation of the Rhineland in 1936 to the invasion of Polandin 1939, Hitler’s calculations were realistically made. In those years, Hitler would almost surelyhave been deterred from acting in ways the immediately threatened massive death and widespreaddestruction in Germany. And, if Hitler had not been deterred, would his generals have obeyed hiscommands? In a nuclear world, to act in blatantly offensive ways is madness. Under thecircumstances, how many generals would obey the commands of a madman? One man alone doesnot make war.To believe that nuclear deterrence would have worked against Germany in 1939 is easy. It is alsoeasy to believe that in 1945, given the ability to do so, Hitler and some few around him would havefired nuclear warheads at the United States, Great Britain, and the Soviet Union as their armiesadvanced, whatever the consequences for Germany. Two considerations, however, work against thispossibility. When defeat is seen to be inevitable, a ruler’s authority may vanish. Early in 1945Hitler apparently ordered the initiation of gas warfare, but no one responded. The first considerationapplies in a conventional world; the second in a nuclear world. In the latter, no country will pressanother to the point of decisive defeat In the desperation of defeat desperate measures may be taken,but the last thing anyone wants to do is to make a nuclear nation feel desperate. The unconditionalsurrender of a nuclear nation cannot be demanded.Dreaming up situations in which someone may have ‘good reason’ to strike first has plaguedstrategic thought ever since Herman Kahn began writing scenarios. Considering one such scenariois worthwhile because it has achieved some popularity among those who believe that deterrence isdifficult. Albert Wohlstetter imagines a situation in which the Soviet Union might strike first. Herleaders might decide to do so in a desperate effort to save a sinking regime. The desperation couldbe produced, Wohlstetter thinks, by ‘disastrous defeat in peripheral war’, by ‘loss of key satellites’,by the ‘danger of revolt spreading—possibly to Russia itself’, or by ‘fear of an attack by ourselves’.Under such circumstances, the risk of not striking might appear very great to the Soviets’.Imagination places the Soviet Union in a situation where striking first is bad, but presumably notstriking first is even worse.One common characteristic of scenarios is that they are compounded of odd elements. How can theSoviet Union suffer disastrous defeat in a peripheral war? If the war is peripheral, defeat may beembarrassing, but hardly disastrous. Another common characteristic of scenarios is the failure to sayhow the imagined act will accomplish the end in view. Some rulers will do anything to savethemselves and their regimes. That is the assumption. But how a regime can hope to save itself bymaking a nuclear strike at a superior adversary, or at any adversary having a second-strike force, isnot explained. Why is not striking first even worse than doing so, and in what way does it entail asmaller risk? We are not told. The most important common characteristic of scenarios, and oftentheir fatal flaw, is also present in this one. The scenarist imagines a state in the midst of a terriblecrisis in which the alternatives are so bad that launching a first strike supposedly makes some sense,but he does not say how this situation might come about. How could the Soviet Union get into sucha mess, and what would other states be doing in the meantime? Scenarios often feature just oneplayer, keeping others in the background even though two or more states are necessarily involved inmelting and in preventing wars. To think that the Soviet Union would strike the United Statesbecause of incipient revolt within her borders is silly. To think that the Soviet Union would strikefirst believing that we were about to do so is not. One must then ask how the US would behave ifthe USSR were seen to be in a perilous condition. It is sometimes surprisingly difficult forstrategists to think of the actions and interactions of two or more states at the same time. No countrywill goad a nuclear adversary that finds itself in sad straits.

When vital interests are at stake, all of the parties involved are strongly constrained to be moderatebecause one’s immoderate behaviour makes the nuclear threats of others credible. No one wouldwant to provoke an already desperate country it that country had strategic nuclear weapons. Equally,a regime in crisis would desperately want to avoid calling nuclear warheads down upon itself. Whatscenansts imagine seldom has much to do with how governments behave. The bizarre qualities ofvarious scenarios that depict a failure of deterrence strengthens one’s confidence in it.Three confusions mark many discussions of deterrence. First, that nuclear weapons affect thedeterrer as well as the deterred is often overlooked. The many who fear that a country will foolishlylaunch missiles in a moment of panic overlook the care other countries will take in order not tomake a nuclear country excessively nervous. Second, those who are sceptical of deterrence easilyslip back from nuclear logic, by which slight risk of great damage deters, to conventional logic, bywhich states may somewhat sensibly risk war on narrowly calculated advantages. Thus someAmencans fear that the Soviet Union will strike first—destroying most of our land-based warheads,planes on the ground, submarines in port, and much else besides. The strike would be made on thechance that we would not strike back with some of our thousands of remaining warheads. But statesdo not risk immense losses unless the odds on succeeding are overwhelmingly high. No one can saywhat the odds might be. Third, the quality of states’ external behaviour is commonly inferred fromtheir internal characteristics. Thus many emphasize the importance of who the new nuclear stateswill be and dwell on the question of whether their rulers will be ‘rational’. They have failed tonotice that radical states usually show caution in their foreign policies and to notice that nuclearweapons further moderate the behaviour of such states when vital interests are at issue. Nuclearpeace depends not on rulers and those around them being rational but on their aversion to runningcatastrophic risks.Arms Races among New Nuclear StatesOne may easily believe that American and Russian military doctrines have set the pattern that newnuclear states will follow. One may then also believe that they will suffer the fate of the UnitedStates and the Soviet Union, that they will compete in building larger and larger nuclear arsenalswhile continuing to accumulate conventional weapons. These are doubtful beliefs. One can inferthe future from the past only insofar as future situations may be like present ones for the actorsinvolved. For four main reasons, new nuclear states are likely to decrease rather than increase theirmilitary spending.First, nuclear weapons alter the dynamics of arms races. In a competition of two or more parties, itmay be hard to say who is pushing and who is being pushed, who is leading and who is following.If one party seeks to increase its capabilities, it may seem that the other(s) must too. The dynamicmay be built into the competition and may unfold despite a mutual wish to resist it. But need thisbe the case in a strategic competition between nuclear countries? It need not be if the conditions ofcompetition make deterrent logic dominant. Deterrent logic dominates if the conditions ofcompetition make it nearly impossible for any of the competing parties to achieve a first-strikecapability. Early in the nuclear age, the implications of deterrent strategy were clearly seen. ‘Whendealing with the absolute weapon’, as William T. R. Fox put it, ‘arguments based on relativeadvantage lose their point’. The United States has sometimes designed her forces according to thatlogic. Donald A. Quarles argued when he was Eisenhower’s Secretary of the Air Force that‘sufficiency of air power’ is determined by ‘the force required to accomplish the mission assigned’.

Avoidance of total war then does not depend on the ‘relative strength of the two opposed forces’.Instead, it depends on the ‘absolute power in the hands of each, and in the substantialinvulnerability of this power to interdiction’. To repeat: If no state can launch a disarming attackwith high confidence, force comparisons are irrelevant. Strategic arms races are then pointless.Deterrent strategies offer this great advantage: Within wide ranges neither side need respond toincreases in the other side’s military capabilities.Those who foresee nuclear arms racing among new nuclear states fail to make the distinctionbetween war-fighting and war-deterring capabilities. War-fighting forces, because they threaten theforces of others, have to be compared. Superior forces may bring victory to one country; inferiorforces may bring defeat to another. Force requirements vary with strategies and not just with thecharacteristics of weapons. With war-fighting strategies. arms races become difficult, if notimpossible, to avoid. Forces designed for deterring war need not be compared. As Harold Brownsaid when he was Secretary of Defense, purely deterrent forces ‘can be relatively modest, and theirsize can perhaps be made substantially, though not completely, insensitive to changes in the postureof an opponent’. With deterrent strategies, arms races make sense only if a first-strike capability iswithin reach. Because thwarting a first strike is easy, deterrent forces are quite cheap to build andmaintain. With deterrent forces, the question is not whether one country has more than another butwhether it has the capability of inflicting ‘unacceptable damage’ on another, with unacceptabledamage sensibly defined. Once that capability is assured, additional strategic weapons are useless.More is not better if less is enough.Deterrent balances are also inherently stable. if one can say how much is enough, then within widelimits a country can be insensitive to changes in its adversaries’ forces. This is the way Frenchleaders have thought. France, as former President Giscard d’Estaing said, ‘fixes its security at thelevel required to maintain, regardless of the way the strategic situation develops in the world, thecredibility—in other words, the effectiveness—of its deterrent force’. With deterrent forces securelyestablished, no military need presses one side to try to surpass the other. Human error and folly maylead some parties involved in deterrent balances to spend more on armaments than is needed, butother parties need not increase their armaments in response, because such excess spending does notthreaten them. The logic of deterrence eliminates incentives for strategic arms racing. This shouldbe easier for lesser nuclear states to understand than it has been for the US and the USSR. Becausemost of them are economically hard pressed, they will not want to have more than enough.Allowing for their particular circumstances, lesser nuclear states confirm these statements in theirpolicies. Britain and France are relatively rich countries, and they tend to overspend. Their strategicforces are nevertheless modest enough when one considers that their purpose is to deter the SovietUnion rather than states with capabilities comparable to their own. China of course faces the sametask. These three countries show no inclination to engage in nuclear arms races with anyone. Indiaappears content to have a nuclear military capability that may or may not have produced deliverablewarheads, and Israel maintains her ambiguous status. New nuclear states are likely to conform tothese patterns and aim for a modest sufficiency rather than vie with each for a meaninglesssuperiority.Second, because strategic nuclear arms races among lesser powers are unlikely, the interestingquestion is not whether they will be run but whether countries having strategic nuclear weapons canavoid running conventional races. No more than the United States and the Soviet Union will lessernuclear states want to rely on the deterrent threat that risks all. And will not their vulnerability toconventional attack induce them continue their conventional efforts?

American policy as it has developed since the early 1960s again teaches lessons that mislead. Foralmost two decades, we have emphasized the importance of having a continuum of forces thatwould enable the United States and her allies to fight at any level from irregular to strategic nuclearwarfare. A policy that decreases reliance on deterrence increases the chances that wars will befought. This was well appreciated in Europe when we began to place less emphasis on deterrenceand more on defence. The worries of many Europeans were well expressed by a senior Britishgeneral in the following words: ‘McNamara is practically telling the Soviets that the worst theyneed to expect from an attack on West Germany is a conventional counterattack’. Why risk one’sown destruction if one is able to fight on the ground and forgo the use of strategic weapons?The policy of flexible response lessened reliance on strategic deterrence and increased the chancesof fighting a war. New nuclear states are not likely to experience this problem. The expense ofmounting conventional defences, and the difficulties and dangers of fighting conventional wars, willkeep most new nuclear states from trying to combine large war-fighting forces with deterrent forces.Disjunction within their forces will enhance the value of deterrence.Israeli policy seems to contradict these propositions. From 1971 through 1978, both Israel andEgypt spent from 20% to 40% of their GNPs on arms. Israel’s spending on conventional armsremains high, although it has decreased since 1978. The decrease followed from the making ofpeace with Egypt and not from increased reliance on nuclear weapons. The seeming contradictionin fact bears out deterrent logic. So long as Israel holds the West Bank and the Gaza Strip she has tobe prepared to fight for them. Since they are by no means unambiguously hers, deterrent threats,whether implicit or explicit, will not cover them. Moreover, while America’s large subsidiescontinue, economic constraints will not drive Israel to the territorial settlement that would shrinkher borders sufficiently to make a deterrent policy credible.From previous points it follows that nuclear weapons are likely to decrease arms racing and reducemilitary costs for lesser nuclear states in two ways. Conventional arms races will wither if countriesshift emphasis from conventional defence to nuclear deterrence. For Pakistan. for example,acquiring nuclear weapons is an alternative to running a ruinous conventional race with India. Anddeterrent strategies make nuclear arms races pointless.Finally, arms races in their ultimate form—the fighting of offensive wars designed to increasenational security—also become pointless. The success of a deterrent strategy does not depend onthe extent of territory a state holds, a point made earlier. It merits repeating because of its unusualimportance for states whose geographic limits lead them to obsessive concern for their security in aworld of ever more destructive conventional weapons.The Frequency and Intensity of WarThe presence of nuclear weapons makes wars less likely. One may nevertheless oppose the spreadof nuclear weapons on the ground that they would make war, however unlikely, unbearably intenseshould it occur. Nuclear weapons have not been fired in anger in a world in which more than onecountry has them. We have enjoyed three decades of nuclear peace and may enjoy many more. Butwe can never have a guarantee. We may be grateful for decades of nuclear peace and for thediscouragement of conventional war among those who have nuclear weapons. Yet the fear is

widespread, and naturally so, that if they ever go off, we may all die. People as varied as the scholarRichard Smoke, the arms controller Paul Warnke, and former Defense Secretary Harold Brown allbelieve that if any nuclear weapons go off, many will. Although this seems the least likely of all theunlikely possibilities, it is not impossible. What makes it so unlikely is that, even if deterrenceshould fail, the prospects for rapid de-escalation are good.McNamara asked himself what fractions of the Soviet Union’s population and industry the UnitedStates should be able to destroy in order to deter her. For military, although not for budgetary,strategy this was the wrong question. States are not deterred because they expect to suffer a certainamount of damage but because they cannot know how much damage they will suffer. Near thedawn of the nuclear age Bernard Brodie put the matter simply: ‘The prediction is more importantthan the fact’. The prediction, that is, that attacking the vital interests of a country having nuclearweapons may bring the attacker untold losses, As Patrick Morgan more recently put it: ‘Toattempt to “compute” the cost of a nuclear is to miss the point’.States are deterred by the prospect of suffering severe damage and by their physical inability to domuch to limit it. Debate over the Soviet Union’s civil defence efforts calls attention to this inability.Defensive measures can reduce casualties, but they would still be immense were either of the greatpowers launch a determined attack. Moreover, civil defence cannot save the Soviet Union’s heavilyconcentrated industries. Warheads numbered in the hundreds can destroy the United and the SovietUnion as viable societies no matter what defensive measures they take. Deterrence works becausenuclear weapons enable one state to punish another state severely without first defeating it.‘Victory’, in Thomas Schellings words, ‘is no longer a prerequisite for hurting the enemy’. Coun-tries armed only with conventional weapons can hope that their military forces will be able to limitthe damage an attacker can do. Among countries armed with strategic nuclear forces, the hope ofavoiding heavy damage depends mainly on the attacker’s restraint and little on one’s own efforts.Those who compare expected deaths through strategic exchanges of nuclear warheads withcasualties suffered by the Soviet Union in World War II overlook this fundamental differencebetween conventional and nuclear worlds.Deterrence rests on what countries can do to each other with strategic nuclear weapons. From thisstatement, one easily leaps to the wrong conclusion: that deterrent strategies, if they have to becarried through, will produce a catastrophe. That countries are able to annihilate each other meansneither that deterrence depends on their threatening to do so nor that they will do so if deterrencefails. Because countries heavily armed with strategic nuclear weapons can carry war to its ultimateintensity, the control of force, in wartime as in peacetime, becomes the primary objective. If deter-rence fails, leaders will have the strongest incentives to keep force under control and limit damagerather than launching genocidal attacks. If the Soviet Union should attack Western Europe, NATO’Sobjectives would be to halt the attack and end the war. The United States has long had the ability toplace hundreds of warheads precisely on targets in the Soviet Union. Surely we would strike mili-tary targets before striking industrial targets and industrial targets before striking cities. The intentto do so is sometimes confused with a war-fighting strategy, which it is not. It would notsignificantly reduce the Soviet Union’s ability to hurt us. It is a deterrent strategy, resting initially onthe threat to punish. The threat, if it fails to deter, is appropriately followed not by spasms ofviolence but by punishment administered in ways that convey threats to make the punishment moresevere.For several reasons, then, deterrent strategies promise less damage than war-fighting strategies.First, deterrent strategies induce caution all around and thus reduce the incidence of war. Second,

wars fought in the face of strategic nuclear weapons must be carefully limited because a countryhaving them may retaliate if its vital interests are threatened. Third, prospective punishment needonly be proportionate to an adversary’s expected gains in war after those gains are discounted forthe many uncertainties of war. Fourth, should deterrence fail, a few judiciously delivered warheadsare likely to produce sobriety in the leaders of all of the countries involved and thus bring rapid de-escalation.A deterrent strategy promises less damage, should deterrence fail, than does the Schlesinger-Brown‘countervailing’ strategy, a strategy which contemplates fighting a limited, strategic nuclear war.War-fighting strategies offer no clear place to stop short of victory for some and defeat for others.Deterrent strategies do, and that place is where one country threatens another’s vital interests.Deterrent strategies lower the probability that wars will be fought. If wars are nevertheless fought,deterrent strategies lower the probability that they will become wars of high intensity.A war between the United States and the Soviet Union that did get out of control would becatastrophic. If they set out to destroy each other, they would greatly reduce the world’s store ofdeveloped resources while killing millions outside of their own borders through fallout. Even whiledestroying themselves, states with few weapons would do less damage to others. As ever, thebiggest international dangers come from the strongest states. Fearing the world’s destruction, onemay prefer a world of conventional great powers having a higher probability of fighting lessdestructive wars to a world of nuclear great powers having a lower probability of fighting moredestructive wars. But that choice effectively disappeared with the production of atomic bombs bythe United States during World War II. Since the great powers are unlikely to be drawn into thenuclear wars of others, the added global dangers posed by the spread of nuclear weapons are small.The spread of nuclear weapons threatens to make wars more intense at the local and not at theglobal level, where wars of the highest intensity have been possible for a number of years. If theirnational existence should be threatened, weaker countries, unable to defend at lesser levels ofviolence, may destroy themselves through resorting to nuclear weapons. Lesser nuclear states willlive in fear of this possibility. But this is not different from the fear under which the United Statesand the Soviet Union have lived for years. Small nuclear states may experience a keener sense ofdesperation because of extreme vulnerability to conventional as well as to nuclear attack, but, again,in desperate situations what all parties become most desperate to avoid is the use of strategicnuclear weapons. Still, however improbable the event, lesser states may one day fire some of theirweapons. Are minor nuclear states more or less likely to do so than major ones? The answer to thisquestion is vitally important because the existence of some States would be at stake even if thedamage done were regionally confined.Looking at the situation of weaker nuclear states and at the statements of stronger nuclear states,one suspects that weak states are less likely to use nuclear weapons first than are strong ones. Manyhave worried about conventional wars between minor nuclear states becoming nuclear wars as oneside loses. It is NATO, however, that plans to use nuclear weapons in battle if conventional troopscannot hold. Moreover, after the Soviet Union invaded Afghanistan in December of 1979, Americanofficials considered using nuclear weapons in the Middle East if need be. At various times, someAmericans have thought of reasons for making limited counterforce strikes—firing a few missiles atthe Soviet Union to show our determination—an idea revived by James R. Schlesinger when he wasSecretary of Defense. Among others, Generals Earle G. Wheeler and George Brown, formerchairmen of the Joint Chiefs of Staff, have talked of our emerging from a nuclear war with a‘relative advantage’ over the Soviet Union by targeting their ‘war recovery capabilities’.

Presidential Directive 59, signed by President Carter in July of 1980, contemplates fighting alimited nuclear war, perhaps a prolonged one, if deterrence should fail. And some of the SovietUnion’s military leaders have publicly discussed using nuclear weapons to win wars.The United States and the Soviet Union have more readily contemplated the use of nuclear weaponsthan lesser nuclear states have done or are likely to do. But planning is distinct from deciding to act.Planners think they should offer Presidents a range of choices and a variety of nuclear weapons tocarry them through. In the event, Presidents, like Party Chairmen, will shy away from using nuclearweapons and will act with extreme care in dealing with situations that might get out of control, asthey have done in the past. New nuclear states are likely to be even more mindful of dangers andmore concerned for their safety than than some of the old ones have been. Ordinarily, weak statescalculate more fearfully and move more cautiously than strong ones. The thought that fear andcaution may lead insecure countries to launch pre-emptive strikes has amplified anxieties about theinstability of regions populated by lesser nuclear powers and about the extent of destruction theirweapons may bring. Such worries rest on inferences drawn froom the behaviour of conventionalstates and do not apply to nuclear ones, for reasons already discussed.Nuclear weapons lessen the intensity as well as the frequency of war among their possessors. Forfear of escalation, nuclear states do not want to fight long or hard over important interests—indeed,they do not want to fight at all. Minor nuclear states have even better reasons than major ones toaccommodate one another peacefully and to avoid any fighting. Worries about the intensity of waramong nuclear states have to be viewed in this context and against a world in which conventionalweapons become ever costlier and more destructive.The Roles and Reactions of the Great PowersShould a great power help a lesser one improve on its force once it has shown the will and theability to build one? Will great powers be drawn into the nuclear confrontations of lesser ones, orwill they draw away from them to avoid involvement? Will small nuclear powers cut themselvesadrift from the great powers and follow independent policies? Will small countries’ nuclear forcestrigger an arms race between the great powers? These questions suggest four ways in which big andsmall nuclear powers may interact.Small and crude forces tempt pre-emption, so it is thought, and maybe used in reckless andunintended ways because of inadequate command and control. These dangers can be removed bygreat powers assisting lesser ones in building and managing their forces. Nevertheless, neither theUnited States nor the Soviet Union will want to help much, lest countries come to believe that theycan build insufficient and unreliable forces and rely on one of the great powers to turn them intosomething substantial. Such hindrance is unfortunate, if improving others’ forces serves widerinterests. Is help required, not just for the sake of the recipient, but also to avoid nuclear imbalancesbetween states that might prompt wars and to reduce the chances of accidents that might set themoff? We saw earlier that these are minor worries. Because they are minor, the United States and theSoviet Union are not likely to be tempted to give technical help to countries entering the nuclearmilitary business.Nuclear weapons in the hands of six or seven states have lessened wars and limited conflicts. The

further spread of nuclear weapons can be expected to widen those effects. Should the United Statesthen promote the spread of nuclear weapons for the sake of peace, even though we need not for thesake of stability? To do so would replace one extreme policy with another. Present policy workshard to prevent additional states from acquiring nuclear weapons. My examination of the effects ofnuclear weapons leads to the conclusion that our policy is wrong without supporting the propositionthat true proliferation—the rapid spread of nuclear weaponry—is desirable. Rapid change may bedestabilizing. The slow spread of nuclear weapons gives states time to learn to live with them, toappreciate their virtues, and to understand the limits they place on behaviour.Will the United States and the Soviet Union be drawn into the struggles of lesser nuclear states?This question loses much of its urgency given the aversion of states to crises that raise the spectre ofnuclear war and the care they take in crises that do so. Will they then draw away from other states’crises rather than being drawn in? The United States and the Soviet Union will continue to haveinterests in various parts of the world for all of the old political. economic, and military reasons. Ina region where nuclear powers are locked in dispute, the great powers will move cautiously inattempting to tend to their separate and to their common interests. We can hardly expect the UnitedStates or the Soviet Union to risk more in other people’s crises than they have risked in their own.Neither the United States nor the Soviet Union want a regional nuclear confrontation to become aglobal one. If that risk hangs over them, their strong mutual interest is to withdraw.Will lesser nuclear powers want to edge away from their great-power patrons in order to be able tochoose their policies more freely? To do so would be risky. A nuclear Israel, for example, maythreaten to fire missiles at her attackers’ cities if ever their victory in war seems likely. In the faceof’ possible Russian opposition, signs of American acquiescence in Israel’s policy would help tomake the prospect of retaliation credible. Any lesser power contemplating the use of nuclearweapons even for deterrent or defensive purposes will expect opposition from at least one of thegreat powers. An alliance or some other kind of connection with one of them may stay the hand ofthe other. This is another way of saying that even with nuclear weapons weaker states continue todepend on stronger states in various ways.By acquiring nuclear weapons a state changes one variable in a complex equation of forces. Thatvariable is the most important one. Nuclear weapons increase the ability of states to fend forthemselves when the integrity of their legitimate boundaries is at stake. Thus an Israeli deterrentforce would enable Israel to maintain her legitimate boundaries while reducing her extremedependence on the United States. In recent years our aid has amounted to a seventh or an eighth ofIsrael’s GNP yearly. Such dependence will substantially lessen only it military security becomesless of a concern or can be more cheaply provided. Nuclear weapons and strategies, however, do notcover all of the military problems of new nuclear states nor are military problems the whole of theirconcerns. Israeli dependence on the United States will not disappear so long as she remains a smallcountry in a hostile world. Similarly, the deterrent effect of China’s nuclear weapons makes her lessdependent on others militarily, without much reducing her need for economic and technicalassistance. Nuclear weapons are useful against threats to a state’s territorial integrity, but most ofthe doings of states fall far short of this extreme. Independent nuclear forces reduce dependency bylesser powers on others without eliminating it.Will the nuclear arms of lesser powers stimulate the great powers to further exertion? And will armscontrol and disarmament agreements be harder to reach? Consider arms racing first. A faster racebetween the great powers may come about in the following way, and to some extent already has.The United States or the Soviet Union builds more missiles and more defences against missiles as

she perceives a growing threat from China. The increased effort of one of the great powers promptsthe other to try harder, and the effects become reciprocating causes. Action, reaction, and over-reaction by the United States and the Soviet Union have formed a pattern too familiar to disregard.The pattern is likely to repeat itself, but it need not.Consider a historical case. In 1967, McNamara half-heartedly proposed deploying a cheap ($5-billion) ABM system designed to handle an attack by China even though, as he said, we had ‘thepower not only to destroy completely her entire nuclear offensive forces, but to devastate hersociety as well’. Whatever his political and bureaucratic reasons, be publicly argued that a lightABM system offered four advantages:1) China might miscalculate.2) America would be showing Asian states that she would not let China blackmail them and wouldthus dampen their desires to have their own nuclear weapons.3) America would gain marginal protection for Minuteman sites against an attack by the SovietUnion.4) America would be safe against accidental launchings.Had the United States persisted in building a ‘Chinese’ system, this might have prompted furtherefforts by the Soviet Union. The United States and the Soviet Union can react to third countries’nuclear forces in ways that stimulate their own competition in arms, but they need not do so, as isshown by examining McNamara’s four reasons. His fourth reason applies to any and all nuclearcountries. It raises the question of the value of taking out an ABM insurance policy againstaccidental firings whether by third countries, by the Soviet Union, or by the United States. His thirdreason applies explicitly to the Soviet Union and not to third nuclear countries. His second reasonrests on a false belief about the feasibility of nuclear blackmail. Only the first of McNamara’sreasons applies specifically to the forces of China or of any lesser nuclear country. It raises thisquestion: Under any imaginable circumstances, might a lesser nuclear country’s miscalculation leadit to launch an attack on the United States? The miscalculation would have to be monumental.Building missile defences against China would imply that great powers can deter each other butcannot deter minor ones. The weakness of the proposition is apparent.Nor need more missiles be added to either great power’s arsenal in order to deter lesser nuclearpowers, even should the great powers fail to deter each other. In 1978, the United States had about2,150 warheads on land-based launchers, about 5,120 on sea-based launchers, and about 2,580 inbombers. These numbers had changed little by 1980. One study estimates that the Soviet Union’sbest attack, launched in the mid-1980s and coming at the worst time for us with our forces only onnormal day-to-day alert, would leave us with about 6,400 warheads and about 1,800 equivalentmegatons. After such an attack, the Soviet Union would have about 6,000 warheads left and 6.000equivalent megatons. We would still have more than we need since 1.000 Poseidon warheads (theforce loading of some six submarines) ‘can destroy about 75 percent of the Soviet industrialtargets’. With our present force we can absorb a first strike and still destroy the Soviet Union as ‘amodern industrial society’, and do so with missiles to spare for counterforce attacks. And we andthey would have more than enough left over to deter third countries. This plenitude of deliverablewarheads is sometimes referred to as ‘sufficiency’. The great powers scarcely need get into an armsrace because of what lesser powers do.

Still, the United States and the Soviet Union do race from time to time, and the racing has beenfuelled in part by what third countries have done. The Soviet Union, for example, argues that manyof her intermediate and medium-range ballistic missiles are needed because of the threat posed byChina. Some of the NATO countries then conclude that because these missiles threaten WesternEurope, cruise missiles and Pershing IIs must be deployed there. This then further worries theSoviet Union. No one doubts these effects.Strategic arms races between the United States and the Soviet Union, however, are produced mainlyby the strategies they follow and by the kinds of forces they build. In their strategies, dissuasion bydeterrence has always been alloyed with defensive and war-fighting policies and capabilities. Thenumber of Russian cities worth striking is finite and indeed quite low. We have long had moresurvivable warheads than Russian cities to strike. If only cities were aimed at, many warheadswould lack targets. The quantity of warheads on hand and their increased accuracy constitutearguments for a counterforce strategy. In the last two decades, the balance between deterrentstrategy and war-fighting strategy has tilted towards the latter. Available weapons affect the strategya country adopts, and the strategy that is fashioned in turn calls for the further development ofweapons. If each side views the other’s strategic forces as designed for fighting wars as much as, ormore than, for deterring them, then arms races become very difficult to avoid. Such perceptionsvary with changes in the strategies and forces of the great powers, not of the lesser ones. Greatpowers engage in arms races mainly because of what other great powers do. That was true in amultipolar, conventional world; it remains true in a bipolar, nuclear world.CONCLUSIONThe conclusion is in two parts. After saying what follows for American policy from my analysis, Ibriefly state the main reasons for believing that the slow spread of nuclear weapons will promotepeace and reinforce international stability.Implications for American PolicyI have argued that the gradual spread of nuclear weapons is better than no spread and better thanrapid spread. We do not face a set of happy choices. We may prefer that countries have conventionalweapons only, do not run arms races, and do not fight. Yet the alternative to nuclear weapons forsome countries may be ruinous arms races with high risk of their becoming engaged in debilitatingconventional wars.Countries have to care for their security with or without the help of others. If a country feels highlyinsecure and believes that nuclear weapons will make it more secure, America’s policy of opposingthe spread of nuclear weapons will not easily determine theirs. Any slight chance of bringing thespread of nuclear weapons to a full stop exists only if the United States and the Soviet Unionconstantly and strenuously try to achieve that end. To do so carries costs measured in terms of theirother interests. The strongest means by which the United States can persuade a country to forgonuclear weapons is a guarantee of its security, especially if the guarantee is made credible by thepresence of American troops. But how many commitments do we want to make and how manycountries do we want to garrison? We are wisely reluctant to give guarantees, but we then should

not expect to decide how other countries are to provide for their security. As a neighbour of China,India no doubt feels more secure, and can behave more reasonably, with a nuclear weaponscapability than without it. The thought applies as well to Pakistan as India’s neighbour. We damageour relations with such countries by badgering them about nuclear weapons while being unwillingto guarantee their security. Under such circumstances they, not we, should decide what theirnational interests require.If the United States and the Soviet Union lessen their opposition to the spread of nuclear weapons,will not many states jump on the nuclear bandwagon? Some have feared that weakening oppositionto the spread of nuclear weapons will lead numerous states to make them because it may seem that‘everyone is doing it’.Why should we think that if the United States relaxes, numerous states will begin to make nuclearweapons? Both the United States and the Soviet Union were more relaxed in the past, and theseeffects did not follow. The Soviet Union initially furthered China’s nuclear development. TheUnited States continues to help Britain maintain her deterrent forces. By 1968 the CIA hadinformed President Johnson of the existence of Israeli nuclear weapons, and in July of 1970 RichardHelms, Director of the CIA, gave this information to the Senate Foreign Relations Committee.These and later disclosures were not followed by censure of Israel or by reductions of assistance toher. And in September of 1980 the Executive Branch, against the will of the House ofRepresentatives but with the approval of the Senate, continued to do nuclear business with Indiadespite her explosion of a nuclear device and despite her unwillingness to sign the Nuclear Non-Proliferation Treaty.Assisting some countries in the development of nuclear weapons and failing to oppose others hasnot caused a nuclear stampede. Is the more recent leniency towards India likely to? One reason tothink so is that more countries now have the ability to make their own nuclear weapons, more thanforty of them according to Joseph Nye.Many more countries can than do. One can believe that American opposition to nuclear armingstays the deluge only by overlooking the complications of international life. Any state has toexamine many conditions before deciding whether or not to develop nuclear weapons. Ouropposition is only one factor and is not likely to be the decisive one. Many countries feel fairlysecure living with their neighbours. Why should they want nuclear weapons? Some countriesfeeling threatened, have found security through their own strenuous efforts and througharrangements made with others. South Korea is an outstanding example. Many South Koreanofficials believe that South Korea would lose more in terms of American support if she acquirednuclear weapons than she would gain by having them. Further, on occasion we might slow thespread of nuclear weapons by not opposing the nuclear-weapons programmes of some countries.When we oppose Pakistan’s nuclear programme, we are saying that we disapprove of countriesdeveloping nuclear weapons no matter what their neighbours do. Failing to oppose Pakistan’sefforts also sends a signal to potential nuclear states, suggesting that if a country develops nuclearweapons, a regional rival may do so as well and may do so without opposition from us. Thismessage may give pause to some of the countries that are tempted to acquire nuclear weapons. Afterall, Argentina is to Brazil as Pakistan is to India.Neither the gradual spread of nuclear weapons nor American and Russian acquiescence in this hasopened the nuclear floodgates. Nations attend to their security in ways they think best. The fact that