Designing Effective Social Control for Nuclear Safeguards:Matching Strategies to Context or "One Size Does Not Fit All"

Ronald B. MitchellDepartment of Political Science

University of OregonEugene OR 97403-1284

INTRODUCTION

Almost every scholarly book or journal articleon nonproliferation written since 1989 begins bydeclaring that the end of the Cold War requiresthat we adopt new strategies for managing nuclearproliferation in the decades ahead /19,31/. As theUS SecretaI}' of Defense recently noted, "thesimple threat of retaliation that worked during theCold War may not be enough to deter terrorists oraggressive regimes. . .. This new danger requiressome new thinking and new leadership on how toprevent, deter and, if necessary, respond to thisthreat" /46/. Essentially, the argument rests on theassumption that new proliferation threats requirenew nonproliferation strategies. And, indeed, theend of the Cold War has seen deployment of newnonproliferation tools, most notably in the use ofpositive inducements for the Ukraine, Belarus, andKazakhstan to return Soviet nuclear weapons toRussia and for North Korea to forgo developmentof an indigenous nuclear weapons capability.

Thus, recent theory and practice both suggestthat effective international proliferation policyrequires a matching of specific strategies to specificthreats. As Richard Betts has noted, "no solutionsat all . . . are applicable across the board" /9/.Reviewing the scholarly literature onnonproliferation policy reveals, however, a lack ofa dominant theoretical modelor even a field ofcontending theoretical models for matchingstrategies to threats. The categorizations of boththreats and strategies appear ad hoc, theoreticallyunnuanced, and fundamentally unsatisfying. Forexample, one recent book concludes by dividingnonproliferation strategies into incentives anddisincentives even while recognizing that thissimple dichotomy fails to capture adequately thevariance in available policy instruments /50/, whileanother divides nonproliferation strategies along

the rather unsophisticated lines of diplomatic andmilitary instruments Ill/.

In this paper, I take Betts' declarationseriously: one size of policy cannot fit all threats.Yet, I seek to go beyond such a generic claim todevelop a theoretically informed model formatching nonproliferation strategies toproliferation threats. Of course, policy makersappear to have done quite well without suchtheoretical guidance at devising and deployingdifferentiated policies, either implicitly orexplicitly, shaped to best respond to the characterof the specific threat. However, I believe that animproved modelof the threat-strategy nexus wouldnonetheless improve our understanding of why agiven strategy sometimes succeed and sometimesfails and thereby improve our ability to devise andapply strategies that fail less often. I develop themodel in three steps: developing a taxonomy ofproliferation threats, a taxonomy ofnonproliferation strategies, and a process for usingthe traits of the former to choose a strategy orstrategies from the "menu" created by the latter.

A TAXONOMY OF THREATS

In addressing any given proliferation threat, apolicy maker will want to know two things:whether to respond to the threat and how torespond to it. Providing that information requirescategorizing threats according to two salientfeatures: their urgency and their character. Ashort-term, static perspective leads us todichotomize states into "threats" and "non-threats"based on their current intention and capability toacquire nuclear weapons /461. A longer-term,dynamic perspective leads us to recognize that astate's nuclear intentions and capabilities positionit along a threat spectrum between urgent currentthreats and unlikely future threats, with that threatposition changing as intentions and capabilities

57

change. Especially if we seek to devise a proactivenonproliferation regime for the next severaldecades, we must establish a comprehensive mix ofstrategies that addresses immediate "extant" threatsas well as the more medium and long term "latent"threats posed by those who may develop thecapacity and/onhe intention to "go nuclear" in thefuture.

Most analysts identifY the crucial determinantsof a state's decision to go nuclear as depending onits motivations and capabilities. The former havebeen referred to as the political determinants ofnuclear intentions, ambitions, inclinations,preferences, or incentives, and the latter have beenreferred to as the technical determinants of nuclearcapability, capacity, and potential 14, 9, 201.

MotivationsStates appear to seek nuclear weapons for two

primary reasons: "defensive security, and assertiveprestige and power" 19, 38, 52/. That is, states gonuclear for either security or status reasons. Israel,Pakistan,. and North Korea appear to be "puresecurity-seekers" who face regional securityenvironments involving immediate and severesecurity threats. Brazil and Argentina appear tomore closely approximate "status-seekers" whofaced less credible immediate threats of attack fromneighbors and whose motivations were "stronglydominated by an ambition for status and regionalpower but without any genuine aggressive orexpansionist designs" 143/. India and South Africa(at least through the 1980s) would appear to sitsomewhere in between, having clear --althoughless pressing --security concerns but clearly alsodesiring nuclear weapons "as instruments of powerand influence in world politics" that wouldimprove their prestige and standing in theinternational community of states /14, 27, 43/.Other states such as Australia, Canada. and NewZealand seem to have clear anti-proliferationmotivations arising from domestic political factorsand chosen roles in the international system.

Because national security is a core goal ofstates, security-seekers tend to have both strongerand less unseatable nuclear ambitions than status-seekers. Israel, Pakistan, and North Korea haveunderstandable 'reasons for believing that attacksby regional adversaries are both likely andpotentially devastating 125/.

A state's nuclear motivations may be lessenedby the alternatives the state has available forresponding to the security environment they face.As the above dichotomy between security-seekersand status-seekers implies, nuclear weapons are notnecessarily an end in themselves. If we accept arational actor modelof state behavior, we shouldexpect that states will strive to acquire nuclearweaponry only if doing so proves to be the bestavailable alternative for achieving security or statusgoals. In a given security environment, the choiceto pursue nuclear weaponry will depend upon both

the availability and the relative attractiveness ofsuch options as conventional weaponry andexternal balancing through alliance formation.

In assessing just how threatening a particularstate's proliferation threat is, the intention toacquire nuclear weapons has a logical primacy overthe capability to acquire nuclear weapons. Nuclearintentions inherently entrain efforts to developnuclear capabilities. In contrast, the capability toproduce or procure nuclear weapons does notappear to entrain efforts te, develop nuclearintentions. As Betts notes, "there is still no knowncase of a country that stumbled into interest ingetting nuclear weapons because technologyacquired to produce energy for civilianconsumption made the weapons option more easilyavailable" 191. The "technological determinist"argument "that countries will acquire nuclearweapons if they are capable of doing so suffersfrom obvious empirical limitations" in the face ofover 40 countries that have been unambiguouslycapable of developing nuclear weapons for yearswho have equally unambiguously refrained fromdoing so 123, 38/.

Cal)abiliticsLike intentions, we can identify two major

factors that determine the capability to acquirenuclear weapons: technological resources andfinancial resources. This categorization simplycaptures the fact that states can either buy nuclearweapons from foreign sources or develop themusing indigenous technical expertise. Besidesbeing security-seekers or status- seekers, states alsocan be "buyers" or "makers." States like Germanyand Japan that are technologically sophisticated arelikely to be makers, whereas states that areresource-rich but technologically less sophisticated(e.g., OPEC members) could be expected to bebuyers. Of course, these are not completelysubstitutable factors: even a country that bought acompletely assembled nuclear weapon (a morecredible concern since the demise of the SovietUnion) would need considerable technical expertiseto make use of that procurement, and, similarly, acountry must spend considerable financialresources to develop nuclear weapons even if it hasindigenous technical expertise.

These different determinants of a state'scapability to acquire nuclear weapons haveimportant implications for designing appropriatenonproliferation policies because they vary in boththeir implications for the speed of nuclearacquisition and the degree to which they aremanipulable by other states. The ability of a stateto transform technological expertise and resourcesinto nuclear weapons is less susceptible to theinfluence offoreign governments than the ability ofa state to transform financial resources into nuclearweapons.

58

Thrcat Ur2;cncv and Threat Chal"acter

Taken together, intentions and capabilitydetermine a threat's urgency or magnitude, and thefactors that drive intentions and capabilitydetermine a threat's character or type.Conceptually, we can identify the nuclearintentions and capacity of a given state and placethem on the illustrative "map" of threat urgencydepicted in Figure 1. States in the northeast cornerof this map have strong nuclear ambitions andstrong indigenous technological capabilities andpose the most urgent threats States in thesouthwest corner of the map have neither theintention nor the capabilities to acquire nuclearweapons and pose the least urgent threat. States inthe northwest corner have the capability but lackincentives to acquire nuclear weapons and, for thereasons just noted, do not pose a significant currentthreat. States in the southeast corner have strongnuclear ambitions but lack the capability to acquirethem and pose a near to medium term concern, ifnot threat.

Empirically, concern about proliferationthreats appears to be driven primarily by intentionand only secondarily by capacity. That is, threatsseem to be ranked as follows: intention andcapacity, intention but no capacity, capacity but nointention, neither capacity nor intention. To givebut one example, nuclear-capable states that lacknuclear weapons ambitions are not mentioned inthe most recent Defense Department assessment ofproliferation 1.t6/. "Threat lists" generally consistof the subset of "nuclear-ambitious" states that are"nuclear-capable" rather than the subset of"nuclear-capable" states that are "nuclear-ambitious." Thus, a line attempting to divide"threats" from "non-threats" would run morevertically than diagonally through thisconceptualmap.

.

This logical map of threat urgency may provemost useful in conceptualizing dynamic changes inthreats. Consider the reduction in the proliferationthreat posed by Brazil and Argentina during thelate 1980s. Both states technological capacity todevelop nuclear weapons remained unchanged assuccessful efforts at rapprochement reduced theirmotivation to use that capacity. In contrast, thegreatest proliferation concerns involve movementof states from the southeast corner to the northeastcorner or the problem of "opaque proliferation"where a state is actually closer to the northeastcorner than commonly known /28/. As alreadynoted, the strong motivation of states in thesoutheast corner lead us to expect them to bestriving to move towards the northeast, consistentlyseeking to improve their technological capabilitiesfor developing nuclear weapons indigenously orseeking ways to transform financial resources intopurchased nuclear components. In contrast, thereis usually less concern that states in the northwestcorner will mO\'e towards the northeast. However,

the map itself highlights the potential future threatposed by states such as Germany, Japan, Sweden,Taiwan, or the tens of other technologically andfinancially nuclear- capable states in the world thatmay develop nuclear intentions in the future eitherbecause of changes in their security environment orin their domestic political environment 153/. "Alladvanced industrial countries with sophisticatednuclear power industries might be considered'latent' nuclear weapons states: anyone of thesecountries could, in theory, acquire a nuclearweapon within six to twelve months" 1221.

Although threat severity is represented interms of the proximity of a state's position to thenortheast comer of the map, threat character orthreat type is captured by considering which of thetwo scales for each axis places the state in thatposition. In particular, the relalive dominance ofsecurity vs. status as the basis of a state's nuclearmotivations and of technological vs. financialresources as the basis of a state's nuclearcapabilities influence the appropriateness andeffectiveness of any given nonproliferation policy.States already sufficiently close to the northeastcomer of this conceptual map or who seem likelyto move in that direction pose four different typesof threat: security-seeking makers, security-seekingbuyers, status-seeking makers, or status-seekingbuyers.

Predictions and expectations about the factorsthat lead, or will lead, a state to become aproliferation threat, Le., that place it in or movingtowards the northeast comer, as well asassessments about how close the state already is tothe northeast corner provide the criteria necessaryfor identifying the most appropriate strategies toeither retard or reverse proliferation efforts.Identifying the most appropriate and effectivestrategies depends on identifying both the threaturgency and threat character.

A TAXONOMY OF STRATEGIES

The nonproliferation policy literature presentsa plethora of policy options for forestallingproliferation. Recent work ranges from remarkablylong menus of carefully differentiated policies tolimited lists of four or five broad policy optionswith little theoretical bite to broad andoversimplified dichotomizations of policies asdiplomatic or military, or incentives anddisincentives /9, 13, 201. This section develops ataxonomy of nonproliferation strategies whichseeks to classify the policy options that have beenused or are available into a theoretical structurethat provides a foundation for identifying whether,when, and how a given policy will succeed atreducing the proliferation threat posed by a givenstate.

In the most general terms, the nonproliferationregime seeks to either maintain or shift the

59

behavior of potential proliferators along a spectrumtoward "desirable" behaviors in which proliferationprograms and aspirations are explicitly,permanently, and sincerely renounced from"undesirable" behaviors in which nuclear weaponsdevelopment and deployment are actively andaggressively pursued. I use the terms "desirable"and "undesirable" behavior change la caplure lhenotion that the NPT and the associatednonproliferation regime have norms of behavior,however vague, that regime members, even ifreluctantly, acknowledge as operative. I use"effective" to refer to the ability of a strategy toinduce behavior which moves toward, even iffalling short of, regime goals. I distinguish regime"members," Le., the governments as well asnongovernmental actors involved in establishingand implementing a regime's slrategy, from regime"targets," Le., those actors whose behavior theregime seeks to aller, including membergovernments, nonmember governments,corporations, nongovernmental actors, and privateindividuals.

The first pair of strategies, punitive andremunerative strategies -- sticks and carrots --manipulate the consequences a target faces to makedesirable behavior more attractive or undesirablebehavior less attractive. The second pair ofpreclusive and generative strategies alter a target'sopportunities, reducing opportunities forundesirable behavior or creating opportunities fordesirable behavior. The final pair of cognitive andnormative strategies alter lhe target's perception ofa given reality, either by altering the informationtargets have or the value that targels attach lacertain behaviors and consequences. Thista;xonomy seeks to differentiate policies based onthe mechanisms by which they inOuence theproliferation decision and behavior. I recognizethat, "at any given' time, those seeking to forestallproliferation not only will, but should, promulgatesets of policies that combine features of thedifferent ideal-types delineated here. Thetaxonomy is intended not to serve to classifycomplete policies or policy sets as to clarify thedifferent mechanisms and pathways by which theyinfluence proliferation behavior so that we candevelop a better understanding of what policieswork to slow proliferation and why they work.

PunitivePunitive strategies, a particularly common and

intuitive model for altering behavior, rely onsanctions, threats, coercion, or other efforts todiscourage undesirable behavior by increasing ilScosts or reducing its benefits /24, 34. 491. Calls for"treaties with teeth" and for beller monitoring,verification, and enforcement highlight the appealof deterrent strategies and the common view thatmost regimes would be more effective if only statesproperly implemented them" Downs, Rocke, andBarsoom /21/ recently have argued that deterrent

strategies including credible commitments tosanction noncompliance are the only strategies thatcan ensure high levels of compliance. The successof this punitive or deterrent strategy depends onconvincing the potential proliferator that any effortto develop or procure nuclear weapons will a) behighly likely to be detected, b) be highly likely tobe punished, and c) the punishment either willdestroy the nuclear weapons acquisition efforts orwill impose sufficiently costly other sanctions as tooutweigh any benefits the proliferator ex-pected toachieve by acquiring nuclear weapons. In standarddeterrence language, detection and sanctioningmust both be "credible," and the sanctions imposedmust be "potent."

The multilateral safeguards system under theInternational Atomic Energy Agency (IAEA) andthe unilateral proliferation detection efforts ofmany states constitute the first element of such astrategy. Recognizing that most states \\ill resistthe intrusion of any type of foreign surveillance,

"

these policies provide access to peaceful nucleartechnologies in exchange for rights to inspectnuclear facilities to ensure that nuclear equipmentand material are not being diverted to militaryprograms. If we distinguish policies of nuclearexport control from policies of nuclear exportdenial, the former involve punitive/deterrentstrategies that rely on receiving assurances that theimporter will not use the technology transferred formilitary purposes, transferring the technology,using national systems of accounting andinspections to detect any diversions for militaI}'purposes, and threatening some form ofpunishment if such diversions are detected /~~/.

Nonproliferation regime members haveexpended considerable effort to develop asafeguards system capable of detecting diversion ofany significant quantil)' of nuclear material in atimely fashion /441. To the extent this threat ofdetection is sufficiently credible, it is oftenimplicitly assumed that this will alert other statesin time for them to respond in ways that removeany potential benefits of diversion /44/. Mostnotably, such responses would involve military"programs, alliances, and securil)' guaranteesintended to offset the hoped-for military advantagethat motivated the nuclear weapons program /9,45/. Essentially, this part of such strategiesinvolves reducing the benefits the potentialproliferant can ex-pect to "derive from goingnuclear. Significant component of USnonproliferation strategy involves attempting toalter the calculus of potential proliferators toconvince them that acquiring nuclear weapons willnot achieve their objective of threatening the US."One of the core objeclives in [American]proliferation protection policy is to convincepotential and actual proliferants that.

". [the US

will] deny or limit the polilical and militaI}" utilityof [nuclear] weapons" /461.

60

As many analysts have noted, safeguards andinspections alone "may not suffice to deterviolations. NPT parties' must be aware thatdetected violations will be met with resoluteresponses. Verification would lose its raison d"treif violators could get away with impunity" /32/. Aresponse must followafter safeguard activitiesidentify proliferation attempts and the responsemust "outweigh any potential benefits" 1l8, 46, 47/.Potential punitive responses cover a wide range. Atthe milder end of the spectrum, IAEA,EURATOM, and the NPT can suspend current andplanned technology transfers, recall prior transfers,suspend other membership privileges /32, 44/.Because such threats appear insufficiently potent asa deterrent to a state that has decided to acquirenuclear weapons, many analysts haverecommended far harsher responses. "Offensivemilitary options," "forcible disarmament," or"imposed arms control" involve explicit threats touse military means to destroy a state's nuclearcapabilities before they can be used /20, 55/. Suchthreats have been urged and contemplated on bothmultilateral and unilateral bases /13. 37. SOI.

When carried out, as in Israel's 1981 bombingof Iraq's Osirak reactor, Iraq's 1980s bombings ofIran's Bushehr reactor, and the American bombingof Iraq's AI Tuwaitha facility during the PersianGulf War, they have "demonstrated thatproliferation by a 'crazy state' bore serious risk andwould not be tolerated" /43/. Such imposition ofsanctions serves to incrcase the crcdibility that apotential proliferator accords a dcclared policy ofsuch sanctions and demonstrate the potcncy ofthose sanctions. These examples "strengthendeterrence of other would-be nuclear proliferators,and there is tentative evidence that the Gulf Warmay have already had this cffect" 1551.

Punitive strategies face problcms in tcnns ofdetection credibility, sanction credibility, andsanction potency. Two factors make detectiondifficult and limit the credibility of such a strategy.States have strong incentives (and capabilities) tokeep militaI)' programs, especially nuclear ones,secret since doing so enhances national security byincreasing the adversary's uncertainty and hencecaution. The nature of a punitive strategy alsoinherently exacerbates these incentivcs bccausesuccess at avoiding detection promises success atavoiding being sanctioned. Punitive strategiesencourage governments and corporations "to keepsecret the veI)' types of information that can bcsignificant" to preventing proliferation / lOI. Thcsedynamics create requircments for extcnsive andexpensive monitoring programs that lcad somc towonder whether "the intelligence community cankeep pace with the threat?" IS Il.

o The threat that sanctions will bc imposcd mayalso lack credibility. Thc more drastic thethreatened sanction, the grcater thc required"exertion of political and military power, "andstates may simply lack thc political will to impose

the threatened sanctions 113, 55/. The costs ofimposing economie or military sanctions aredifficult enough to overcome unilaterally, and arefurther exacerbated by collective action problemswhen multilateral sanctions are attempted /2, 33/.Experience with India, Israel, Pakistan, and evenIraq demonstrates that states committed topreventing proliferation find it politically difficultto respond to detected proliferation with sanctionsadequate to alter a target state's behavior.

Sanctions may fail to appear potent. Harshthreats of attack may fail to deter a state fromembarking on a nuclear weapons program, and,indeed, may reinforce the state's commitment tosuch a program by confirming that they face ahostile security environment 145, SOI. Milderthreats, such as those the UN or IAEA mightimpose, may simply have little impact on thingsthe target state values. The sanction of domestic orinternational opprobrium will likely proveinsufficient to maintain nuclear abstinence ifsecurity challenges are significant 133/. Especiallywhen faced with a very threatening securityenvironment, the incentives t9 proliferate willusually dwarf any countervailing sanctions otherstates can be expected to impose.

RemunerativeRemunerative strategies reduce the costs or

increase the benefits of desirable behavior to makeit more attractive. Rewards are assumed toinfluence targets who lack any commitment toregime norms, have the capacity to fulfill thosenorms, but would otherwise not do so. Rewardseffect behavioral changes either when targets viewcompliance as desirable but costly, or when they donot value compliance but do value the reward.Two different remunerative strategies adopted toalter the consequences of abjuring nuclear weaponsinvolve either direct financial transfers, as evidentin financial transfers for North Korean andUkrainian nuclear restraint, or contingentlyenhancing a state's strategic environment,providing security guarantees or reducing securitythreats, in exchange for proliferation restraint.

Remunerative strategies address thefundamental point that, "to demand that agovernment forgo nuclear weapons is to demandthat it compromise its own sovereignty. To expectit to do so without.any compensating quid pro quowould not only be condescending but naive" 191.The offers of both direct rewards in the form ofnew nuclear power facilities and improvements tothe security environment in the form of removal oftactical nuclear weapons from the region providedthe foundation for the current agreement by NorthKorea to forego its nûclear weapons program /43/.The Ukraine similarly received direct financial"compensation for the fissile material contained inthe warheads and satisfactory assurances ofsecurity" in exchange for acceding to the NFT andreturning Soviet nuclear weapOns left on Ukrainian

61

soil to Russia /43/. In many cases during the ColdWar, "an alliance with one of the nuclearsuperpowers meant that that superpower waswilling to use its nuclear forces on behalf of theminor coalition partner. This willingness on thepart of a superpower obviated the minor partner'sneed to develop independent nuclear forces" /29/.

Just as veiled and ambiguous threats maynonetheless deter, so vague and ambiguous rewards

'.may lead certain states to forego their nuclear goalsifdoing so seems to promise sufficient rewards.Many states "explicitly or implicitly acceptnonproliferation as a condition of joining theWestern community of states" and enjoying theeconomic and political benefits of integrating into"the Western core community" /13/. Most recently,China appears to have signed the NPT at least inpart "because it desires the benefits of trade andeconomic assistance from the West" /13/.

Remunerative strategies also faceimplementation difficulties. Just as the costs ofpunitive strategies make states reluctant to imposesanctions and thereby undercut their credibility, thecosts of remunerative strategies make statesreluctant to offer rewards to potential proliferators.When multilateral rewards are involved, collectiveaction problems, not surprisingly, appear. Forexample, the US has funded both the Ukraine andNorth Korean programs while oth&rnonproliferation supporters have proved more thanhappy not to contribute. Regime supporters alsoresist rewarding states that break the internationalnann against proliferation both because it seemsdistasteful and politically unpopular to rewardundesirable behavior and because it createsincentives for extortion and blackmail attempts.These factors make states reluctant to provide largescale rewards to proliferators, even if doing sowould effectively alter their behavior.

Lastly, the fact that a target accepts eitherfinancial transfers or security guarantees does notpreclude them from continuing to pursueproliferation. As the extensive monitoringcomponents of both the Ukrainian and NorthKorean deals make clear, neither financial .ormilitary assistance nor security guarantees "crasethe attractiveness of nuclear weapons as anautonomous deterrent" /9/. The acceptance of anoffered reward need not imply that the recipienthas renounced nuclear weapons. Whether withrespect to financial transfers or security guarantees,the target state may well accept the rewards whilecontinuing in a more clandestine way, their nuclearprogram. "Structural constraints, especially thecondition of anarchy, encourage self-regardingbehavior that undermines confidence ininternational guarantees. even ostensibly firmmilitary alliances, as a means for ensuring security.For practical strategic and normative politicalreasons, states prefer autonomy to dependence and,therefore, searchfor ways to provide [or their ownsecurity" /33/.

Preclusive

If punitive and remunerative strategiesincrease the costs of going nuclear or increase thebenefits of staying nonnuclear, preclusive andgenerative strategies decrease the opportunities forgoing nuclear and increase the opportunities forstaying nonnuclear. Strategic trade embargoes andtechnology denial attempt to prevent (rather thandeter) "have-nots" from acquiring certaintechnologies. Such strategies rely on a process ofprohibiting precursor behaviors that regimemembers can more readily control and that, ifprevented, also prevent the ultimately importantundesirable behavior, namely nonproliferationitself /49/. Preclusive strategies tend to clearlyproscribe precursor acts that themselves are notdirectly undesirable and use "premonitorysurveillance" to detect acts before, rather thanafter, they occur /46, 49/. Preclusive strategies canreduce monitoring costs relative to a punitive ordeterrent strategy by proscribing those behaviorsmost inherently .transparent and most costly toconceal. Thus, efforts at technology denial aremotivated in part by the difficulties of detectingnuclear weapons development and deployment.

P.reclusive strategies have been a majorelement in the nonproliferation regime's effort toforestall nonproliferation by nuclear aspirants sinceits inception /53/. Unlike the punitive strategyunderlying efforts to regulate and safeguardexports of nuclear technology under the NPT andthe Nuclear Exporters Committee or "ZanggerCommittee," the Committee for Multilateral ExportControls (CoCom), the London Suppliers Group,and individual national regulations hav!; sought therather different goal of establishing "an embargo asconcerns the most sensitive parts of the fuel cycle"/44/. The EURATOM agreement requires priornotice of expected shipments of nuclear material asa means to prevent, rather than punish diversion/44/. Such technology denialseeks to control"militarily critical technology, such as uraniumenrichment, as a means to control the spread ofnuclear weapons" /53/. Such denial consists ofeither complete bans on e:\'Ports of certain nucleartechnologies, or bans on e:\'Ports to any country thathas not accepted full-scope lAEA inspections /8,40/. Preclusive strategies were urged as early asthe mid-1970s in cases in which punitive strategiesim'olving "end-use deterrents" seem likely to fail"to prevent the diversion of strategic technologies"/40/. For example, in response to the Sovietinvasion of Afghanistan in 1979, the US forbidexports to the Soviet Union of any items on theCoCom list because of "the emerging consensusamong U.S. officials that Soviet end-use assurancescould not be trusted to protect against the diversionof U.S. exports to militaI}' use" /40/.

A particular advantage of preclusive strategiesis that they do not "require understanding themotivations of the states that seek nuclear

62

weapons" /38/. Indeed, technology denialstrategies are not susceptible to the risk that exportcontrol strategies face in which the intentions ofthe recipient state either change or aremisinterpreted. The success of preclusivestrategies, not surprisingly, depcnds on theindigenous capabilities of the target state and thestrength of that state's motivation. Export controlsappear to have significantly delayed, if notcompletely halted, "the diffusion of military ordual-use technologies" /17/. CoCom's strategicembargo of the Soviet Union allowcd the West tomaintain or widen military technology gaps,especially "in areas subject to tighter controls" anddespite the embargo being far from complete /40/.Although much of nuclear weapons technology hasbecome commonly available, critical parts remaindifficult if not impossible for most states to produceindigenously /32/. Thus, even states with strongnuclear ambitions will be "significantly dependenton technological infusions from abroad" 1l2, 43/.The London Suppliers Group made it"extraordinarily difficult for Pakistan and Iraq toget their weapons programs in full swing, andcertainly delayed them. That said, such restraintsdid not prevent Pakistan from eventually achievingthe objective, and would not have prevented Iraq ifthe 1991 war had not intervened" /9/. "Technologydenial has not reversed the nuclear weaponsprograms in states that strongly perceive thcirsecurity as requiring nuclear weapons. although ithas slowed progress in programs forced to rcly onindigenous capabilities and covert activities" /53/.The weapons programs of states \lith strongnuclear ambitions driven by threatening securityenvironments are unlikely to be slopped by atechnology denial strategy /29, 53/.

Implementation difficulties of preclusivestrategies revolve around the need for the regime toestablish a set of "metasanctions" to inducegovernments to enforce export bans against theirown corporate actors Il/. Indeed. much concernover sanctions in nuclear nonproliferation has beenat this metalevel, i.e., seeking to induce countrieslike China to adopt and implement export contra!and related nonproliferation policies /36/.Preclusive strategies also founder when regimesupporters lack the capacity or commitment toidentify and prevent targets from engaging in theprecursor activities or the undesirable behavioritself. Preclusive strategies are impotent againsttargets that have autonomous controIover thetargeted behavior and its precursor activities, aparticular problem in dealing with states that haveor can develop indigenous capabilities to designand build nuclear weapons. Thus. wealthier andmore powerful states will be less susceptible tosuch strategies lhan developing Slates. Finally,preclusive strategies must rely on reinforcingstrategies to address cases in \\'hich preventionfails. As the Toshiba sale of submarine propellermilling machines in violalion of CoCom rules

clarified, any preclusive effort to prohibit certainexports must be "back-stopped" by a punitivepolicy that seeks to ensure that those technologytransfers which are not prevented areat leastidentified after the fact /10/.

GenerativeGenerative strategies complement preclusive

ones, creating rather than removing opportunitiesfrom targets' choice sets. Such strategies addressundesirable behaviors that result from targetincapacity; cases in which threats, rewards, orprevention will be unlikely to alter behavior.Generative strategies attempt to create new, moreattractive opportunities which the target willchoose on their own over the undesirable behavior.The American Cooperative Threat Reduction(CTR) program has sought to provide formerSoviet states with otherwise-unavailable services,tools, and technology to facilitate the eliminationof nuclear weapons and installation of safeguardsagainst proliferation /46/. Without such resources,these states would be unlikely to be able to removeand dismantle these nuclear weapons as safely /46/.In an even more novel generative strategy, the UShas funded "defense conversion industrialpartnerships" and "international science andtechnology centers" in Moscow and Kiev toprovide over 5,000 "job alternatives for weaponsscientists who might otherwise be tempted to selltheir nuclear expertise abroad" /42, 46, 56/. Ofcourse, at the margin, generative strategies mergeinto remunerative strategies, but the former alterbehavior by creating newalternatives rather thanchanging the consequences of existing alternatives.Generative strategies differ from remunerative onesin targeting capacity deficits, focusing on targetswho can't, rather than won't, alter their behavior.Generative strategies assume targets want to butcannot fulfill regime commitments. The newalternatives tend to be made in a noncontingentfashion, with the assumption being that the targetswill choose the new alternatively voluntarily.

In nonproliferation, during the 1950s, theUnited States successfully undermined themomentum of the Swedish nuclear weaponsprogram by making weapons-incompatible civilianlight water reactor technology and fuel available tothe Swedes, thereby undercutting its weapons-compatible civilian heavy water reactor program.By providing Sweden with the new, and otherwiseunavailable, alternative of light water reactortechnology, the US policy delinked the civilian andmilitary nuclear programs, thereby increasing thecosts and undercutting the relative attractiveness ofthe latter program which the Swedes S00l1abandoned /39/. Sales of high technologyconventional weaponry that would not otherwise beavaiJable to a state can help potential proliferantsachieve their security goals without resort tonuclear weapons development. As with forwarddeployments of forces made on a noncontingent

63

basis (as opposed to the contingent securityguarantees discussed as remunerative strategies),such strategies provide otherwise-unavailableimprovements to a country's defense capabilities,creating strong incentives to refrain fromdeveloping nuclear weapons /461. One analyst hasurged that, India be oITered "permanentmembership in the UN Security Council in returnfor signing the NPT" to provide India with analternative to nuclear weapons development as ameans of achieving desired international prestige/14/.

Generative strategies face similarimplementation problems to remunerativestrategies. Because they invoh'e the provision ofresources to make the new opportunities available,wealthier regime members will simultaneously berequired and reluctant to fund these opportunities.As evident in the programs to provide new jobopportunities for ex-Soviet nuclear personnel, mostprograms have not been funded throughmultilateral collaborative efforts but rather byunilateral efforts. As ,yjth remunerative strategies,creating new opportunities for socially desirablebehavior maynot reduce socially undesirablebehavior if the two are not mutually exclusive. Forexample, generative programs "intended to make[nuclear power plants] slightly safer for a shortperiod of operation until Eastern [European] statescan implement plans for shutdown has theunintended effect of prolonging the operations ofthe more dangerous" nuclear power plants /16/.Equally important, the success of generativestrategies depends crucially on accuratelypredicting what choices the targeted actors willmake iffree to choose bctween existing options andthe new options made a\'ailable

COl!.nitivcRegimes incorporating cognitive strategies

provide targets with new, more complete, and moreaccurate information "in order to facilitateintelligent choices" that favor socially desirablebehaviors /2/. The information can relate to thealternatives avaiIajJle.. the causal relationshipbetween behaviorS' and consequences, the costs andbenefits of diITerent behaviors. the current state orlikelihood of various important decisionparameters in the world. or the likely behavior ofother actors. Diplomats and gO\'ernment officialsmay, to a limited extent. be able to provide ncwinformation as well as manipulate "means-ends"belief systems. "These strategies can rcly on regimemembers to generate and disseminate theinformation themselves. to encourage others togenerate and disseminate the information, or tomandate that others provide information in thecourse of private transactions. ln thenonproliferation arena. at leas! pan of thediplomacy has involved the perhaps heroic andoften implicit erfon to provide nonnuclear states\vith informatioI) to convince them that adding

nuclear weapons to their arsenal will not actuallyincrease their security, and that they can increasetheir security more through other means /52/.

Cognitive strategies assume targets supportregime norms and can fulfill regime requirements.The model assumes that targets engage in sociallyundesirable behavior because they mistakenlybelieve it to be privately desirable or beneficial, andthat new information will induce targets torenounce undesirable behavior. Monitoring targetbehavior becomes unnecessary since actors serve as"their own ubiquitous inspectors, tailor their 0\\11standards to particular risks, and invoke their 0\\11sanctions" /3/. Implementation costs decline sincealtering information about consequences oropportunities usually costs far less than alteringthose consequences themselves.

Rather than attempting to manipulate costsand benefits through punitive or remunerativestrategies,. .diplomatic efforts often attempt toreassure states that their security needs can be metthrough means other than nuclear proliferation andstress the political, economic, and military risks ofnuclear weapons development, production anddeployment /38, 46/. No-first-use pledges, forexample, increase the "awareness of the limits ofnuclear weaponry" /20/. Cognitive strategies alsoinvolve provision of specific, accurate intelligenceregarding the status of regional proliferation sothat states do not develop nuclear weapons becauseof unwarranted fears that their neighbors areproliferating /46/. Indeed, such strategies reducethe chances for "missile gap" type arms racesdriven by inaccurate information. Such strategiesconvince states not to undertake nucleardevelopment programs that would otherwiseappear desirable. Such cognitive strategies mayalso' be useful in the post- proliferation stage aswell to ensure that countries have the necessaryknowledge to make intelligent choices regardingnuclear weapons control and safety /41, 45/.

Whether states perceive acquiring nuclearweapons facilitating their goals depends, of course,on actions as well as words. However, diplomatshave some leeway in influencing how events areinterpreted. Both the failure of India and Pakistanto gain in.tcmational stature with their nuclearprogr<],lTlsand Germany and Japan's ability to gaininternational without such programs (indeed,perhaps because they have not expended resourceson such programs) can be highlighted to increasethe perception that nuclear weapons areunnecessary and may be counterproductive toachieving international status. Such perceptionsmay be reinforced by the fact that developingnuclear weapons no longer carries the scientificand technical prestige it once did /50/. The recentcompletion of a comprehensive nuclear test bantreaty and corresponding rhetoric help strengthenthe argument that nuclear weapons are lessnecessary to national security than previouslycontended. On the other hand, the effort and

64

rhetoric expended by nuclear weapons states insupport of maintaining and relying on iargenuclear arsenals tends to undercut efforts toconvince nuclear-ambitious states that pursuingnuclear weapons is against their interests.

Cognitive strategies tend to fail when regimemembers institute them as cheap and nonintrusiveways to "do something," rather than becauseinadequate information is truly the source ofundesirable behavior. Information-based strategiescan induce a wide range and more dynamic set ofpositive behaviors if governmental, corporate, andprivate actors receive better information that allowsthem to more accurately assess the extent to whichtheir individual interests coincide with regimenorms. However, improving the accuracy of theinformation a state has about the nuclearaspirations and capabilities of its neighbors, aboutthe costs and risks of nuclear weaponsdevelopment, and about nuclear strategies andexperiences of the nuclear weapons states will notprove adequate to alter their decision to "gonuclear." In many cases, states will rationallyassess that their nuclear restraint may be in theinterests of other states but not in their ownnational interest. Cognitive strategies will beimpotent to alter behavioral choices unless thosechoices have been driven b~' inadequate orinaccurate information.

Normative

Normative strategies change behavior by'altering targets' deep-seated values rather than theinstrumental incentives that more proximatelydetermine their decisions and actions. Regimescan establish strategies that induce targets to"change their practices because they have come tounderstand the world in a way that promotescertain actions over others" /54/. Normativestrategies involve either collective or hierarchicalefforts at consciousness-raising. During regimenegotiations and recurring meetings, "leader"states may tl}' to com'ince "laggards" to accepttheir norms of behavior, or regime members maywork together to focus attention on a problem,create new collective norms, and increase membercommitment to existing norms. Essentially,normative strategies involve rhetorical attempts topersuade targets not merely to adopt differentmeans to their pre-existing goals, as in a cognitivestrategy, but to adopt new goals. If the precedingfive strategies are built on a "logic ofconsequences," a normative strategy is buill on a"logic of appropriateness" /26/. By encouraging a"transfornlation of interests" so that states "rankadherence to global normsagainst possessingnuclear weapons above preserving nationalautonomy," such a strategy could "decisivelyweaken the incentives to deploy nuclear \\'eapons"/331.

Normative strategies do not alter theopportunities or consequences that targets face, butinstead establish ongoing dialogues among regimemembers and between regime members and targetsto promote regime norms. Normative strategiesassume that targets' values are inconsistent withregime norms but are susceptible to policymanipulation, that targets have the capacity toadopt desirable behaviors, and that they will do soonce their lack of an exogenous commitment toregime norms can be remedied through normativedialogue and education. Normative strategies, ifsuccessful, have the advantage that they "can beginto influence an actor as soon as an act iscontemplated and before it is committed. whereassocial disapproval and forma! punishment can onlybe mobilized after the event and only incircumstances where others acquire evidence ofwho committed the act" /34/.

The debates in the NPT review conferenceshave sought, at least in part, to create a "nucleartaboo" that would convince nuclear "have-nots"that nuclear weapons development would bemorally wrong /45, 48/. "International norms andstandards make an important contribution toproliferation prevention [by] creating anatmosphere of restraint" /46/. Simply inçreasingthe diplomatic and Pul>.lic"relations priorit). givento nonproliferation issues "may dampen nuclearambitions" i50/. Many analysts have argued thatthe negotiation of a comprehensive test ban shouldstrengthen the norm against nuclear weapons bydemonstrating both closer superpowerconformance with the anti-nuclear weapons normand final compliance with Article 6 of the NPT/37, 45, 46/. Of course, the unwillingness ofnuclear "haves" to dismantle their nuclear arsenalstends to undercut such arguments 114/. Thenuclear free zone treaties, building on the lack ofstrong immediate incentives to develop nuclearweapons in certain regions, appear to have beensomewhat more successful in convincing states toforego nuclear development. These normativestrategies can occur at lower, interpersonal levelsas well. "Military-to~military cooperation andcontacts . . . reinforce basic tenets such as civiliancontrol of the military and the honoring ofinternational norms of behavior" /36, 46/. At amore subtle, and more conjectural, level, efforts toincrease the priority given to economic andenvironmental issues while reducing that giventosecurity issues (e.g., in the European Union)appear to involve normative strategies that targetthe relative priority given to different goals thattarget states pursue.

Normative strategies are, by their nature,blunt, soft, and long-term instruments of policy.Unlike the other strategies described above, nosingle state can establish new norms for theinternational community at large. More than withother strategies, the success of a normative strategydepends on a majority of states working

65

consistently and cooperatively through both wordand deed to promote and strengthen such norms.Norms, until they are well- established, are likelyto be vulnerable to even few and small instances ofnonconformant behaviors. Normative strategiesface inherent difficulties of attempting la alterdeeply held beliefs, resistance to "imperialist"efforts at normative education, and the time neededto induce normative change and any correspondingbehavioral change. Since states will not alter theircommitment to the norm of self-defense andsurvival, it requires developing and strengthening anorm about what constitutes "appropriate" meansof achieving those goals in the internationalcommunity of states. As with most norms, somestates would be likely to viol~te even the mostrobust and long-standing anti-nuclear weaponsnorm, were it established: This docs not imply,however, that a strong norm \\'ill not lead certainstates that would otherwise choose to proliferate torefrain from doing so.

MATCHING STRA. TEGIES TOTHREA TS

How can we make use of the precedingtaxonomies of threats and strategies to develop anappropriate set of effective proliferalion responses?This section makes an initial :lllempl to use theinsights from these taxonomies la identifystrategies that will reduce the likelihood thatspecific potential proliferants will "go nuclear." AsRichard Betts has noted, no strategy "is universallyapplicable -. each is undermined by costs andqualifications when applied to eoncrele cases. Arealistic polic)' will have to combine ad hocvariations of several of the options in regard todifferent candidates" 191. The following discussionseeks to go beyond stich a general statement tosuggest that policy responses may be ad hoc in thesense of being unique to the country at hand, butneed not be ad hoc in the sense of theoreticallyuninformed. I begin by examining the mostpressing and urgent threats but then look at how anintegrated set of strategies can help keep long-termlatent threats from becoming more urgent.

States that have both strong motivations andcapabilities to acquire nuclear weapons obviouslypose the most urgent threats. Having said that,however, not all these threats are alike. They differin character based on whether their motivationderives from security or stalus concerns andwhether their capabilily rests upon lïnaneial ortechnical resources. The effectivcness of a givenstrategy will depend on lhe strategy's ability tosatisfy the concerns that motivate the state and onto take advantageof the states capability deficits, ifany.

Security-seekers pose the greatest proliferationchallenge. Such states are "very difficult to dealwith, because their basic belief is that withoutnuclear weapons or at least a nuclear weapon

option national survival may be put into question.For this reason, leadership by example through aprocess of disarmament. . . has little effect as theydo not directly impact upon regional securityconstellations" /43/. To think about how to addressthese states concerns depends on furtherdistinguishing between defensive (or status-quo)security-seekers and offensive (or aggressive)security-seekers. In both cases, security-seekers"perceive their security situation to be so dire" thata remunerative strategy of contingent securityguarantees will not persuade them to foregonuclear development or relinquish existing nucleararsenals /291. Indeed, for offensive but technically-unsophisticated security-seekers, such as Iraq, armssales might well enhance rather than impede thecountry's technical progress.

Defensive security-seekers would seem mostlikely to respond to a combination of strategies thatsimultaneously reduce the perceived security threatand strengthen the normative pressure againstresponding to that threat by acquiring nuclearweapons. Reducing the perceived threat could beaccomplished by combining cognitive \\lthgenerative strategies. A cogI}itive strateg)' ofproviding better information regarding the actuqlthreat the country faces would counter tendencies .

to plan policyan worst-case scenarios, therebyreducing the perceived threat. The US undertooksomething along these lines during the Gulf Warby providing Israel with timely and detailedinformation on Iraqi militaI)' positions,capabilities, and activities to encourage it not torespond to Iraqi attacks. A generative strateg)' ofactually reducing the threat environment couldcreate the conditions that would further reduce themotivations for the state to pursue the nuclearoption. The power of reductions in a state's threatenvironment to reduce nuclear ambitions is evidentmost clearly in the reduction in the US arsenal inthe wake of the demise of the Soviet Union. Thesestrategies to reduce the motivations to go nuclearcould be reinforced by more proactive efforts tostrengthen normative pressures on the state toprovide for security through nonnuclear means.Although such pressures are unlikely to workquickly, a strengthened norm against nuclearownership would establish a "logic ofappropriateness" that would over the long term,lead fewer states to pursue and retain nuclearweapons.

Cognitive and generative strategies aimed atreducing the perceived threat environment \\111 notreduce the nuclear ambitions of an offensivesecurity seeker, however, such as Iraq. With suchstates, punitive strategies of "forcible disarmament"like that implemented during the Gulf War mayactually prove effective at reducing overalllikelihoods of nuclear proliferators by both settingback the nuclear ambitions of the targeted state andincreasing the credibility of the punitive deterrentto other states.

66

Status-seekers face less severe securityenvironments and a looser linkage exists betweennuclear-capability and their goal of internationalstatus. Punitive strategies involving "forcibledisarmament" would be particularlycounterproductive with these states because itwould increase the perceived threat environment,thereby transforming the state from a status-seekerinto security-seeker. Cognitive and normativestrategies may prove use[ul in establishing thelonger-term conditions to persuade such states thatnuclear weapons either will not enhance theirstatus as they desire or are not an appropriatemeans for doing so. Remunerative strategies thatoffer financial rewards for foregoing a nuclearprogram will be likely to succeed if the status goalsare not particularly strong, whereas generativestrategies that provide the state with alternativemeans of increasing their international status mayprove more effective. Generative strategies wouldseem likely to prove particularly elTective withstatus-seekers because, unlike security- seekers,such states have strong disincentives to pursuenuclear weapons once they have achieved greaterrecognition by other means.

Obviously, countries will not fit neatly andcompletely into these security-seeker or status-seeker categories. As Chubin notes, "the mostdifficult proliferator is Jran, in part because itsincentives to acquire nuclear weapons are sodiffuse, being animated by global ambitions asmuch as by regional threats. They arc sought asmuch for an expression of the Revolution's vitalityand for a defiant assel1ion of equality as for anyspecific purpose" 115/. Such cases require effortsto evaluate how these the security threats suchstates face can, if possible, be reduced whilefacilitating their achievement o[ non-securityobjectives through means other than the acquisitionof nuclear weapons.

The character of the potential proliferant'scapabilities provide further guidance as toappropriate policies. "Makers," i.e.. states such asIsrael and Jndia with indigenous nucleartechnological capabilities present those seeking toprevent proliferation with [ewer points forintervention and behavioral control. Preclusivestrategies, such as technology denial, will provelargely irrelevant in such cases because indigenoustechnological capabilities remove the need for suchstates' programs la buy nuclear technology.Punitive slrategies, even including attempts al"forcible disarmament," are also likely to proveeither ineffective or counterproductive because they,\'ill induce a reactive resistance that only increasesthe sense of threat and enhances the motivation toacquire nuclear weaponry while failing to damagethe basic source of their nuclear weapons program,namely, their technological sophistication.

By contrast, for "buyers." i.e.. states such asIraq or Pakistan that depend in the short lamedium term on procuring rather than developing

crucial nuclear components, the initial line ofdefense must be a preclusive strategy of technologydenial. For these states, a slrategy can hope toslow nuclear progress even if the state's strongmotivations preclude the strategy from halting italtogether. For these states, "barring transfer ofenrichment or reprocessing plants would be themost decisive way to hold back proliferation in theshort run" /9/. "Argentina and Brazil are obviouscases in which technology denial, may have slowedtechnical advancement until domestic politicalchanges resulted in a reversal of their nuclearprograms" /53/. If preclusive strategies fail, whichthey eventually will, punitive strategies that attacknuclear development facilities and programs mayhave the double virtues of directly slowing nuclearprogress while also enhancing the credibility ofthreats to punish such progress and reinforcingnorms against such progress. These benefits mustbe weighed against the impetus they give to thestate's nuclear ambitions by increasing theperceived threat environment.

A longer-term vision of nonproliferationmanagement requires that we look beyond thiscategorization of urgent threats to more remotethreats. It requires establishing strategies thataddress currently remote threats before theybecome urgent threats. Such threats include thoseposed by states such as Libya and Syria that arenuclear-ambitious but technologically relativelyunsophisticated, states like Germany and Japanthat are technologically- capable but currentlyunmotivated to acquire nuclear weapons, and themyriad states who currently lack both themotivation and capability to acquire nuclearweapons.

All three types of remote threats seem to bereduced by far-sighted efforts to strengthen thenorm against nuclear weapons acquisition.Negotiation of the CTBT, e:\1ension of the NPT,ongoing reductions in superpower nuclear arsenals,and continued conformance with the norm againstnuclear use can all help legitimize and reinforcethis norm. Likewise, generative strategies aimed atimproving both the global and specific regionalsecurity environments by finding nonmilitarysolutions to conflicts and by providing securityguarantees designed to reassure rather thanantagonize adversaries will help reduce the factorsthat transform states into nuclear-ambitioussec~rity- seekers. Cognitive strategies thatimprove communications between adversaries alsohelp reduce the misperceptions and worst-casemilitary planning that can trigger arms spiralswhich increase militarization but fail to increaseeither side's security.

The technical threats posed by states alreadyfacing severe security environments but currentlyfar from being nuclear-capable can best be met bycontinued efforts to preclude them purchasing anytechnologies or personnel that would further theirnuclear capabililies. Preclusive policies of

67

technology denial should be maintained so long asthese states continue to show nuclear ambitions.Generative, cognitive, and normative attemptsshould be made to provide alternative means ofachieving existing goals, and to engage these statesin diplomatic dialogue to reexamine whetheracquiring nuclear weapons provides the cognitivelymost effective and normatively most appropriatemeans of doing so.

The political threats posed by states that arenuclear-capable but not currently nuclear-ambitious are frequently ignored. However, thenuclear abstinence of states like Germany andJapan "will enduré only until serious threats tovital interests are perceived" /33/. That is, unlesspolicies are put in place now to discourage theacquisition of nuclear weapons as the means torespond to such threats. Evidence from Japan,South Korea, and Taiwan suggests that specificallyidentifying ways to enhance the securityenvironment faced by such stales, includingthrough superpower security guarantees, can helpstop nuclear programs before they get wellunderway /29/. These proliferation threats can alsobe kept latent by facilitating as much directinvolvement of these states as possible incontributing to the formation of strong normsagainst nuclear proliferation. Keeping these states(including both their gO\'ernmental representativesand their publics) engaged in the process of normformation will increase their commitment to thosenorms even as pressures la \'io!ale those normsdevelop.

The most remote threats posed by statescurrently lacking in both nuclear ambition andnuclear capabilities, the dif1ïcuh choice is facedbetween technology engagement and denial. Apolicy of technology engagement based on theexisting safeguards regime that encourages nucleartrade while seeking to prevent diversion formilitary purposes seems a prudent strategy but onethat has both benefits and risks. Safeguard policieswill never be perfect and so run some risks ofproviding a false sense of security that lhe transferof militarily-valuable technology is beingprecluded. On the other hand, "nuclear exports,even to countrie~ that had not signed the NPT,could in the long run advance nonproliferationgoals more than shortsighted technology denialsbecause export contracts and nuclear cooperationbetween supplier and customer countries helped tointegrate the recipients of nuclear technology intothe existing international safeguards ,system tocontrol the spread of nuclear \\eapons" /351. Theappropriateness of a policy of tcchnology denial ortechnology engagement requires an ongoingattempt to assess the strength of the recipientsmotivations for proliferation: as security or statusmotivations flow, policy should shift towardstechnology denial and as these mo!i\'ations ebb,policy should return towards a greater degree ofsafeguarded technology engagement.

CONCLUSION

This paper has identified the types ofproliferation threats we will face in the decadesahead, identified the available range of strategiesby which we can respond to those threats, andmade a very preliminary attempt to combine thesetwo taxonomies to identify appropriate policies forparticular threats. Throughout I have taken theperspective of a policy maker concerned foremostwith halting proliferation. This has meantignoring the important real-world tradeoffsbetween this security objective and the economicobjective of fostering the development of theinternational trade in peaceful nuclear technologywhich sits at the core of the international nucleartechnology trade.

Efforts to prevent proliferation proYide awealth of experience for evaluating the conditionsunder which particular policies facilitated orfrustrated nonproliferation goals. Indeed, thenonproliferation regime has faced a variety ofthreats and has brought to bear a variety ofstrategies to address them. To date, thenonproliferation regime has done remarkably wellat containing proliferation, with the actual numberof nuclear weapons states being far below thenumbers predicted two or three decades ago. Ashas been frequently noted, the end of the Cold Warprovides a historical moment for identifying thenature of the threats the nonproliferation regime islike!y to face in the decades ahead and forattempting to assess how best to deal with them.Without a theoretically-informed means ofevaluating why certain nonproliferation strategiessucceed at preventing proliferation in one case andfail in another, we are unlikely to achieve theimportant dual goals of discouraging nonnuclearweapon states from acquiring nuclear weapons andencouraging existing nuclear weapon states todismantle theirs.

BIBLIOGRAP HY

11/ Axelrod, Robert. 1984. The :Evollltion ofCooperation. New York: Basic Books.

12/ Axelrod, Robert, and Robert O. Keohane.1986. "Achieving Cooperation Under Anarchy:Strategies and Institutions." ln Cooperation UnderAnarchv, ed. Kenneth Oye. Princeton, NI:Princeton University Press.

13/ Bardach, Eugene and Raben Kagan. 1982.Going Bv the Book. Philadelphia: TempleUniversity Press.

/4/ Bernauer, Thomas, and Dieter Rulofr. 1997."Analytical Framework." ln Soliciting Cooperationfrom Critical States: The Politics of PositiveIncentives in Arms Control. UnpublishedManuscript, Version No.7., ed. Thomas Bernauer,Jozef Goldblat, and Dicter Rulofr. 5-44. Zürich:Center for International Studies. ETH-Zentrum.

68

151 Bernauer. Thomas. Jo/.ef Golclblat. andDieter Ruloff. cds. 19<)7. Soliciling Cooperationfrom Critical States The Politics of PositiveIncentives in Arms Control. UnpublishedManuscript, Version NO.7. ZÜrich: Center forInternational Studies, ETH-Zentrull1.

161 Bernauer, Thomas, Jozef Goldblat, andDieter Ruloff. 1997. "Introduction." In SolicitingCooperation from Critical Stales: The Politics ofPositive Incentives in Arms Control. UnpublishedManuscript, Version No.7.. cd. Thomas Bernauer.Jozef Goldblat, and DiCIer Rulofr. 1-4. Zürich:Center for International Studies. ETH-Zentrum.

/71 Bertsch, Gary K.. Richard T. Cupit!. andSteven Elliott-Gower. 1994. InternationalCooperation on Nonproliferation EXPo!1 Controls:Prospects for the I990s and Bevonel. Ann Arbor:The University of Michigan Press.

181 Bertsch, Gary K.. Richard T. Cupitt, andSteven Elliott-Gower. 1994. "Multilateral ExportControl Organizations." In InternationalCooperation on Nonproliferation Export Controls:Prospects for the I990s and Bevonel. cel. Gary K.Bertsch, Richard T. Cupitt. anelSteven Elliott-Gower. 33-59. Ann Arbor' The University ofMichigan Press.

191 Betts, Richard K. 1993 "Paranoiels.Pygmies, Pariahs and Nonproliferation Revisited."In The Proliferation Puzzle: Why NuclearWeapons Spread and What Results. eel. Zachary S.Davis and Benjamin Frankel. 100-24. London:Frank Cass and Co.

110/ Bialas, Jeffrey P 1989. "The DetectionandDisclosure of Violations of U.S Export Con!rols:Management Choices and Their National SecurityImplications." f./rginia Journal 0/ InternationalLall' 29 (56-620)

111/ Black\\"Îll. Robert D.. anel AlbertCarnesale. 1993 "Conclusions andRecommendations" ln Ne". Nuclear Nations:.Consequences for US. Polie\'. cel. Robert D.

Blackwill and Albert Carnesale, 20-50. New York:Council on Foreign Relations.

1121 Bueno de Mesquita, Bruce, James D.Morrow, and Samuel S. G. Wu. 1993. "Forecastingthe Risks of Nuclear Proliferation: Taiwan as anIllustration of the Method." In The ProliferationPuzzle: Why Nuclear Weapons Spread and WhatResults, cd. Zachary S. Davis and BenjaminFrankel, 311- 31. London: Frank Cass and Co.

1131 Chafetz, Glenn. 1993. "The End of theCold War and the Future of Nuclear Proliferation:An Alternative to the Neorealist Perspective." InThe Proliferation Puzzle: Whv Nuclear WeaponsSpread and What Results, cd. Zachary S. Davisand Benjamin Frankel, 127-58. London: FrankCass and Co.

1141 Chellaney, Brahma. 1994. "India." InNuclear Proliferation After the Cold War, ed.Mitchell Reiss and Robert S. Litwak, 165-90.Washington, D.C.: The Woodrow Wilson CenterPress.

115/ Chubin, Shahram. 1994. "The MiddleEast." ln Nuclear Proliferation After the Cold War,cd. Mitchell Reiss and Robert S. Litwak, 33-66.Washington, D.C.: The Woodrow Wilson CenterPress.

1161 Connolly, Barbara, and Martin List. 1996."Nuclear Safety in Eastern Europe and the FormerSoviet Union." In Institutions for EnvironmentalAid: Pitfalls and Promise, ed. Robert O. Keohaneand Marc A. Levy. Cambridge, MA: MIT Press.

1l71 Cupitt, Richard T., Gary K. Bertsch, andSteven Elliott-Gower. 1994. "Introduction." InInternational Cooperation on NonproliferationExport Controls: Prospects for the 1990s andBeyond, ed. Gary K. Bertsch, Richard T. Cupitt,and Steven Elliott-Gower, 3-32. Ann Arbor: TheUniversity of Michigan Press.

118/ Davis, Zachary S. 1993. "The RealistNuclear Regime." In The Proliferation Puzzle:Why Nuclear Weapons Spread and What Results,cd. Zachary S. Davis and Benjamin Frankel, 79-99.London: Frank Cass and Co.

1191 Davis, Zachary S., and Benjamin Frankel,cds. 1993. The Proliferation Puzzle: Whv NuclearWeapons Spread and What Results. London: FrankCass and Co.

120/ Daly, Paul, and Steven Flank. 1993."Arms Control for New Nuclear Nations." In NewNuclear Nations: Consequences for U.S. PoliC'!" ed.Robert D. Blackwill and Albert Carnesale, 53-76.New York: Council on Foreign Relations.

/21! Downs, George, David Rocke, and PeterBarsoom. 1996. "Is the Good News AboutCompliance Good News About Cooperation?"In/emotional Organization 50:3 (Summer): 379-406.

/22/ Dunn, Lewis A. 1993. "New NuclearThreats to U.S. Security." In New Nuclear Nations:Consequences for U.S. Policy, ed. Robert D.Blackwill and Albert Carnesale, 20-50. New York:Council on Foreign Rclations.

/231 Elbaradei, Mohamed. 1996. "OnCompliance with Nuclear Non-proliferationObligations. " Security Dialoglle 27( I): 17-26.

/241 Etzioni; Amitai. 1961. A ComparativeAnalvsis of Complex Organizations: On Power,Involvement, and Their Correlates. Ncw York: TheFree Press of Glencoe, Inc.

/251 Feldman, Shai. 1994. "Israel." In NuclearProliferation After the Cold War, ed. MitchellReiss and Robert S. Litwak, 67-88. Washington,D.C.: The Woodrow Wilson Center Press.

/261 Finnemore, r...1artha. 1996. NationalInterests in International Society. Ithaca: CornellUniversity Press.

/271 Fischer, David. 1994. "South Africa." InNuclear Proliferation After the Cold War, cd.Mitchell Reiss and Robert S. Litwak, 207-30.Washington, D.C.: The Woodrow Wilson CenterPress.

/281 Frankel, Benjamin, cd. 1990. Journal of

69

Strategic Studies 13 (September) Opaque NuclearProliferation: Methodological and PolievImplications. Special Issue of.

/29/ Franke!' Benjamin. 1993. "The BroodingShadow: Systemic Incentives and NuclearWeapons Proliferation." ln The ProliferationPuzzle: Why Nuclear Weapons SlJread and WhatResults, cd. Zachary S Davis and BenjaminFrankel, 37-78. London: Frank Cass and Co.

/30/ Fujita, Edmundo. 1989. The Prevention ofGeogralJhieal Proliferation of Nuclear WealJons:Nuclear-weapon-free Zones and Zones of Peace inthe Southern Hemisphere. UNIDIR Research PaperNO.4. New York: United Nations.

/3 1/ Gill, Bates, Kensuke Enbata. and MatthewStephenson. 1996. "Japan's Expon ControlInitiatives: Meeting New Proliferation Challenges."The Nonproliferation Rel'Îell' 4 (Fall):30-42.

/32/ Goldblat, Jozef. 1992. The Non-proliferation Treatv: Ho,,' to Relllove the ResidualThreats. UNIDIR Research Paper No. 13. NcwYork: United Nations.

/33/ Goldstein. Avery. 1993. "UnderstandingNuclear Proliferation: Theoretical Explanation andChina's National Experience." In The ProliferationPuzzle: Whv Nuclear Weapons Spread and WhatResults, ed. Zachary S Davis and BenjaminFrankel, 213-55. London: Frank Cass and Co.

/34/ Hawkins, Keith. 1984. Environment andEnforcement: Regulation and the Social Definitionof Pollution. Oxford socia-legal studies. Oxford.England: Clarendon Press.

/35/ Hofbanse!, Clat!s. 1994. "GermanPerspectives on Expon Control Policy." InInternational Cooperation on NonlJroliferationExport Controls:' Prospects for the 1990s andBevond, ed. Gary K. Bensch. Richard T. Cupitt.and Steven Elliott-Gower. 179-200. Ann Arbor:The University of Michigan Press.

/36/ Kan. Shirley. and Zachary Davis. 1994."China." In Nuclear Proliferation After the ColdWar, cd. Mitchell Reiss and Raben S. Litwak. 129-64. Washington, D.C.: The Woodrow WilsonCenter Press.

/37/ Kay, David. 1994. "The IAEA: How CanIt Be Strengthened?" In Nuclear Proliferation Afterthe Cold War, ed. Mitchell Reiss and .Robert S.Litwak, 309-33. Washington, D.C.: The WoodrowWilson Center Press.

/38/ Lavoy, Peter R 1993. "Nuclear Myths andthe Causes of Nuclear Proliferation." In TheProliferation Puzzle: Why Nuclear WeaponsSpread and What Results, ed. Zachary S. Davisand Benjamin Frankel, 192-212. Port]and, OR:Frank Casso

/39/ Long, William 1. 1996. "Trade andTechnology Incentives and Bilateral Cooperation."International Studies Quarterly 40 (March):77-106.

/40/ Mastanduno, Michael. 1992. EconomicContainment: CoCom and the Po]itics of East-WestTrade. Ithaca: Cornell University Press.

/41/ Mearsheimer, John. 1990. "Back to theFuture: Instability in Europe After the Cold War"International Security 15: 1 (Summer): 1-40.

/42/ Miller, Steven E. 1994. "The FormerSoviet Union." In Nuclear Proliferation After theCold War, ed. Mitchell Reiss and Robert S.Litwak, 89-128. Washington, D.C.: The WoodrowWilson Center Press.

/43/ Müller, Harald, David Fischer, andWolfgang Kötter. 1994. Nuclear Non-proliferationand GlobalOrder. Oxford: Oxford UniversityPress.

/44/ Nuclear Energy Agency. 1988. TheRegulation of Nuclear Trade: Non-proliferation.Supply. Safetv. Volume 1 international aspects.Paris: OECD.

/45/ Nye, Joseph S., Jr. 1993. "DiplomatieMeasures." In Ncw Nuclear Nations: Consequencesfor U.S. Policy, ed. Robert D. Blaekwill and AlbertCarnesale, 78-96. New York: Council on ForeignRelations.

/46/ Office of the Secretary of Defense. 1996.Proliferation: Threat and Response. Washington,D.C.: US GPO.

/47/ Pilat, Joseph F. 1994. "Responding toProliferation: A Role for Nonlethal Defense." InNuclear Proliferation After the Cold War, ed.Mitchell Reiss and Robert S. Litwak, 275-89.Washington, D.C.: The Woodrow Wilson CenterPress.

/48/ Price, Richard. 1995. "The Genealogy ofthe Chemical Weapons Taboo." InternationalOrganization 49 (Winter).

/49/ Reiss, Albert 1., Jr. 1984. "SelectingStrategies of Social ControlOver OrganizationalLife." In Enforcing Regulation, ed. Keith Hawkinsand John M. Thomas, 23-35. Boston: Kluwer-Nijhoff.

/50/ Reiss, Mitchell. 1994. "Conclusion:Nuclear Proliferation After the Cold War." InNuclear Proliferation After the Cold War, ed.Mitchell Reiss and Robert S. Litwak, 335-50.Washington, D.C.: The Woodrow Wilson CenterPress.

/51/ Riehelson, Jeffrey T. 1994. "Can theIntelligence Community Keep Pace with theThreat?" In Nuclear Proliferation After the ColdWar, ed. Mitchell Reiss and Robert S. Litwak, 291-08. Washington, D.C.: The Woodrow WilsonCenter Press.

/52/ Samare, Gary. 1994. "Iraq." In NuclearProliferation After the Cold War, cd. MitchellReiss and Robert S. Litwak, 15-31. Washington,D.C.: The Woodrow WiJson Center Press.

/53/ Sands, Amy. 1994. "The Impact of NewTechnologies on Nuclear Weapons Proliferation."In Nuclear Proliferation After the Cold War, ed.Mitchell Reiss and Raben S. Litwak, 259-74.Washington, D.C.: The Woodrow Wilson CenterPress.

/54/ Wapner, Paul. 1995. "EnvironmentalActivism and World Civic Politics." lVorld Politics

70

47 (April):311-40.1551 Zelikow, Philip. 1993. "Offensive Military

Options." In New Nuclcar Nations: Conscquencesfor U.S. Policv, ed. Robcrt D. Blackwill and AlbertCarnesa1e, 162-95. Ncw York: Council on ForcignRelations.

1561 Zinbcrg, Dorothy S. 1995. Thc MissingLink? Nuclcar Prolifcration and thc IntcrnationalMobility of Russian Nuclcar Expcrts UNIDIRResearchPaperNo. 35. NcwYork: United Nations.

A Map of Threat Urgency

LeastUrgent

Nuclear-am

·Security Seeking·Status Seeking ".IJ

,r1

71