December 2011 Strategic Forum National Defense University Vincent Manzo is a Research Analyst in the Center for Strategic Research Institute for National Strategic Studies at the National Defense University Key Points Many weapons systems and most military operations require access to multiple domains These linkages create vulnerabilities that actors can exploit by launching cross-domain attacks the United States may seek to deter such attacks by threatening cross-domain responses However both the U S Government and potential adversaries lack a shared framework for analyzing how counterspace and cyber attacks fit into an accepted escalation ladder The real-world effects of attacks that strike targets in space and cyberspace and affect capabilities and events in other domains should be the basis for assessing their implications and determining whether responses in different domains are proportionate or escalatory Development of a shared framework that integrates actions in the emerging strategic domains of space and cyberspace with actions in traditional domains would give decisionmakers a better sense of which actions and responses are expected and accepted in real-world scenarios and which responses would be escalatory This would support more coherent cross-domain contingency planning within the U S Government and deterrence threats that potential adversaries perceive as clearer and more credible www ndu edu inss Deterrence and Escalation in Cross-domain Operations Where Do Space and Cyberspace Fit by Vincent Manzo In most real conflicts the potential escalation sequence is more like a ladder that has been bent and twisted out of shape with all sorts of extra and odd protuberances added on which vitally affect how the conflict does or does not climb it Controlling escalation will depend crucially on identifying the particular twists and protuberances of that conflict's misshapen ladder 1 Warfare has become even more complicated since Richard Smoke wrote this description of escalation in 1977 The National Security Space Strategy de- scribes space as congested contested and competitive yet satellites underpin U S military and economic power Activity in cyberspace has permeated every facet of human activity including U S military operations yet the prospects for effective cyber defenses are bleak Many other actors depend on continued access to these domains but not nearly as much as the United States For this reason some analysts argue that China's opening salvo in a con- flict with the United States would unfold in space and cyberspace Worst-case scenario assessments conclude that such an attack might render the United States blind deaf and dumb almost exclusively through nonkinetic means although it is unclear how effective attacks in the space and cyber domains would be in an actual military conflict How do concepts such as escalation deterrence and proportionality apply in such a context What odd protuberances would counterspace and cyber attacks create in an escalation ladder What are the salient thresholds for cross-domain attacks And what exactly does cross-domain mean This paper explores these questions using the illustrative example of a hypothetical U S -China conflict because both countries SF No 272 1 Center for Strategic ReSEARCH About the Authors possess diverse strategic capabilities that span air land sea space and cyberspace Defining Cross-domain Platforms or Effects Cross-domain is an ambiguous term U S doctrine identifies land air and sea as domains Recent U S national security policy and strategy documents recognize space and cyberspace as distinct domains as well As2 suming that all five are strategic domains there are at least two different ways an action could cross domains cross-domain attacks enable an actor to utilize its strengths and exploit an adversary's vulnerabilities Cross-domain could be defined according to the platform from which an actor launches an attack and contributions from multiple domains One could even argue that a precision conventional strike is a cross-domain attack regardless of whether the attacking platform and target are in the same domain if it utilizes satellites and computer networks By the same reasoning characterizing a cyber attack--as opposed to cyber exploitation--against U S military computer networks as single-domain is misleading If successful such an attack would have important cross-domain effects it would undermine the air ground or naval forces that depend on the degraded computer networks These indirect effects in other domains are often the primary purpose of cyber attacks 3 The same logic applies to attacks with co-orbital ASATs even if the platforms are in the same domain the effects are cross-domain Thus cross-domain can also be defined according to the effects of an operation Under this approach an at- tack is cross-domain if its intended consequences unfold in a different domain than its target This definition illuminates that inter-domain relationships our own and our adversary's create strategic vulnerabilities 4 For example the platform on which the target resides Destroying a U S precision conventional strike operations depend on missile is a cross-domain attack whereas destroying be incapable of destroying U S aircraft or nuclear-pow- satellite with a ground-launched antisatellite ASAT one with a co-orbital ASAT for example a maneu- verable satellite is not Striking a surface ship with a conventional air-launched cruise missile is a cross-domain attack whereas an attack on the same target with a sea-launched cruise missile SLCM is not Defin- ing cross-domain by platforms demonstrates that cross- domain operations are not new Air attacks on naval forces naval attacks on air forces and attacks from both domains on ground forces are common in modern warfare Indeed in many instances a cross-domain operation might simply be the most expedient option As an example a nation under attack by SLCMs might for a variety of reasons be able to attack the adversary's naval assets more quickly with aircraft than with submarines and surface ships This definition might be too simplistic Most U S military forces on land in the air and at sea make use of cyber and space assets and most complex missions integrate 2 SF No 272 access to multiple domains A potential adversary might ered cruise missile submarines but it might be able to at- tack the space and cyber assets that enable these platforms to destroy targets This appears to be the logic underlying China's interest in counterspace and cyber attacks such at- tacks shift the conflict to domains where China's offensive forces have an advantage over U S defenses thereby alter- ing U S capabilities in domains air and sea for example where China would otherwise be at a disadvantage 5 This cross-domain approach would be ineffective if U S air sea and ground forces did not depend heavily upon space and cyber assets Without this link China would be unable to translate U S vulnerability in space and cyberspace into an operational impact in other domains Cross-domain at- tacks thus enable an actor to best utilize its strengths and exploit an adversary's vulnerabilities in some instances Reports that the United States considered launching a cyber attack at the start of North Atlantic Treaty Orga- nization operations in Libya suggest that the U S military www ndu edu inss also perceives cross-domain attacks as useful for exploiting adversary vulnerabilities 6 Cross-domain Operations and Deterrence These definitions highlight the fact that military ac- tors frequently cross domains Indeed U S military posture is inherently cross-domain U S offensive and defensive this does not explain why threats to respond to counterspace and cyber attacks in other domains are considered less credible than cross-domain responses to air land or sea attacks Shared Framework for Assessing Proportionality and Escalation in Space and Cyberspace A concept Thomas Schelling explored in Arms and weapons are distributed across air- sea- and ground- Influence is a useful starting point for answering these U S military operations and engender advantages in other more comprehensible to potential adversaries and thus based platforms space and cyber assets are ubiquitous in domains and it is highly unlikely that future U S conflicts will unfold exclusively within one domain From this per- spective U S deterrence is inherently cross-domain too when the United States threatens to respond to actions that endanger U S and allied interests it threatens albeit implicitly in most cases cross-domain responses The plat- forms the United States employs the targets it attacks and the effect of the attack might be in different domains questions Schelling argued that deterrence threats are more credible if they are proportionate with and connected to the actions they are intended to deter countries lack a shared framework for interpreting how counterspace and cyber attacks fit into an escalation ladder and might differ from the domains utilized in and affected by the adversary's initial attack By the same logic the United States traditionally deters attacks in general without distinguishing between attacks that cross domains and those that do not Naval attacks on naval forces are not inherently more or less dangerous than air attacks on naval forces The United States attempts to deter both and the means target and scale of the U S response to either would depend on the effects of the attack and U S objectives rather than the domains involved Thus the United States deters attacks regardless of whether the attacks cross domains by threatening responses that will likely cross domains and differ from the initial at- There is an idiom in this interaction a tendency to keeps things in the same currency to respond in the same language to make the punishment fit the character of the crime It helps an opponent in understanding one's motive and provides him a basis for judging what to expect as the consequences of his own actions the direct connection between action and response helps to eliminate the possibility of sheer coincidence and makes one appear the consequence of the other 7 Of course such communication requires that coun- tack Given that cross-domain deterrence is neither new tries interpret military actions and reprisals similarly--in topic is How can the United States mitigate vulnerabilities idiom of action nor rare the real question underlying recent interest in the that stem from its dependence on space and cyberspace other words that they communicate through a shared Schelling also acknowledged that breaking a pat- Both are offense-dominant domains where U S defenses tern of behavior that is escalation might be necessary to credibly threaten to impose costs on aggressors and deny to display unreliability and dare the adversary to respond are inadequate and policymakers are uncertain about how benefits of attacks Although potential adversaries depend on space and cyberspace less than the United States does www ndu edu inss in some circumstances to catch an adversary off balance in kind Even then however a shared understanding of limits norms and expected responses creates a necessary SF No 272 3 frame of reference by which actors distinguish between attacks and reprisals that cross from traditional strategic rules is more dramatic and communicates more about The absence of a shared framework within the U S proportionate and escalatory behavior Breaking the domains into these newer ones and vice versa one's intent precisely because it can be seen as a refusal strategic community complicates effective cross-domain The idiom of military action was never as coherent and usable options for responding to attacks in space and to abide by rules 8 communicable and universally recognized in reality as it is in Schelling's prose Nevertheless during the Cold War there was a generally accepted escalation ladder from conventional to chemical and biological to nuclear weapons Within a conventional conflict there has been an understanding that escalation can occur by broaden- ing the geographical area of fighting expanding the targets attacked for example shifting from narrow military contingency planning Developing coherent effective cyberspace requires that military planners in the different Services and combatant commands possess similar assumptions about cross-domain proportionality and escalation For example Principal Deputy Under Secre- tary of Defense for Policy James Miller testified that U S responses to counterspace attacks could include neces- sary and proportional responses outside of the space domain 10 Yet there are a variety of types of counterspace to broader societal targets and increasing the intensity attacks and even more potential nonspace targets for U S shifting to more destructive conventional weapons The determine which nonspace responses best correspond of violence for example using more bombs per sortie or salient thresholds differ in every conventional conflict Unfortunately countries lack a shared framework reprisals A common framework would help planners with counterspace attacks of varying scope and severity The absence of a shared framework between the for interpreting how counterspace and cyber attacks fit United States allies and potential adversaries un- ity in space and cyberspace are new relative to land air miscalculation Effective deterrence requires that U S into an escalation ladder Competition and vulnerabil- and sea Countries have less experience fighting wars in which space and cyberspace are part of the battlefield Unlike conventional and nuclear weapons experts are less certain about the precise effects of attacks in these domains For these reasons a widely shared framework 9 the absence of a shared framework undermines deterrence and increases the potential for miscalculation for judging how counterspace and cyber attacks cor- respond with interactions in other domains and more broadly with political relations between potential adversaries during peacetime in crises and in wars does not yet exist Without one decisionmakers will have diffi- culty distinguishing between proportional and escalatory 4 SF No 272 dermines deterrence and increases the potential for officials influence potential adversaries' perceptions of the likely consequences of the actions the Unit- ed States wishes to deter The United States might threaten to respond to a particular type of attack in space or cyberspace by employing different capabilities against different targets in other domains Such threats however are less likely to resonate as credible with potential adversaries if they do not understand U S assumptions about how domains are linked and why a particular response is a logical and proportional reaction to the initial attack As an example imagine the United States threat- ened to respond to ASAT attacks on U S intelligence surveillance and reconnaissance ISR satellites with attacks against the adversary's air defense network The logic underlying this policy is that the United States might employ ISR aircraft over the adversary's territory to compensate for the lost satellites Attacks on the air defense network would be necessary to enwww ndu edu inss sure that the aircraft could effectively penetrate the legal experts might be a fruitful approach to starting the United States is restoring its lost ISR capabil- What would be the basis for assessing counter- country's airspace This policy is proportional because this conversation ity thereby denying the benefits of the ASAT attack space and cyber attacks in a shared framework Must the adversary's attack Instead of responding in space portional Is a kinetic response to a nonkinetic attack However the U S response would be different from the United States would attack targets on or around the adversary's homeland To further complicate the situation the United States might use conventional weapons to destroy the air defense network even if the initial ASAT attack was nonkinetic Without a shared framework potential adversaries might consider this deterrence threat illogical and therefore not credible If deterrence failed they might perceive such a U S response as arbitrary and escalatory Even with a shared framework they may still consider this response as escalatory but they would also understand it to be a likely consequence of employing ASATs against the United States before authorizing an attack To be clear a shared framework would not and could not prescribe set actions for every imaginable scenario Rather it would define a generic escalation responses to kinetic attacks also be kinetic to be pro- always escalatory Can a cyber attack be proportional to a cruise missile strike How do officials compare attacks that strike targets in some domains and affect capabilities and events in other domains Counter- space and cyber attacks can vary widely in intensity from the equivalent of a tap on the shoulder to a fist in the face Clearly the mere act of extending the con- flict into these domains is an insufficient metric for evaluating attacks and calibrating responses Rather the real-world effects of such attacks both within the domain of the attack and in other domains should determine whether they are escalatory and which responses would be appropriate Variables in a Shared Framework Cultivating a shared framework between potential ladder a tacit or loosely defined code of conduct that adversaries for assessing effects and formulating ap- tions and responses are expected and accepted in real- domains are involved U S and foreign officials interpret would give decisionmakers a better sense of which acworld scenarios and which would cross thresholds that escalate the situation This would pave the way for more coherent cross-domain contingency planning within the U S Government and U S deterrence threats that potential adversaries perceive as clearer and more com- prehensible and credible The United States would also propriate responses is difficult regardless of how many events through different prisms Cultural differences contrasting strategic objectives differences in force structure and doctrine and differing strengths and vulnerabilities can cause decisionmakers in the United States and other countries to reach different conclusions about proportionality and escalation 11 This challenge is not have a better understanding of the calculus of potential new but the uncertainties in emerging strategic domains vating such a shared framework is a constructive goal Imagine that China interferes with U S satellites adversaries in their efforts to deter U S actions Culti- discussed in the previous paragraphs might exacerbate it for the future because deterrence crisis management via nonkinetic means laser-dazzling or jamming dur- tries interpreted proportionality connectedness and conflict The United States might attempt to undermine and escalation control would be easier if different counescalation similarly Engaging the U S strategic com- munity in a thorough dialogue on these issues is the first step toward achieving this goal Forming a deterrence working group of regionalists functionalists and www ndu edu inss ing a military crisis that has yet to escalate into an armed China's ability to attack U S satellites perhaps by scram- bling its space-tracking data through a cyber attack One could argue that this response is proportional because it is limited to systems that China is already employing SF No 272 5 against the United States and does not cross the kinetic in any conflict would grow exponentially after such at- tacking in a new domain is escalatory opening the door be profound As a result U S officials might feel pres- threshold On the other hand one could argue that atto reprisals and counterreprisals in cyberspace and other domains Would Chinese officials distinguish between attacks on military computer networks and computer networks that support the regime's domestic security op- erations If not they might interpret this proportional response as an existential assault especially if they believe that U S cyber attacks will cause collateral damage to computer networks other than the one targeted actions that are escalatory during crises might be proportionate in limited wars and underwhelming responses as the scope and intensity of a conflict increase What if the initial Chinese ASAT attack is kinetic Would U S allied and Chinese officials perceive a nonkinetic response against China's space tracking capability as weak even if it succeeded in protecting U S satellites On the other hand would kinetic attacks on the ASAT weapons China is employing be proportional Or would crossing the geographic threshold assuming the targets tacks occur because the effects in other domains would sure to take preemptive action prior to such an attack or they might take risks to quickly terminate a conflict and punish the adversary in its aftermath The linkage between vulnerabilities in space and cyberspace and the effectiveness of U S capabilities in other domains that makes U S satellites and computer networks high-value targets also makes the threat of a strong reprisal more credible it would be proportionate to the effects of the attack Conveying this to potential adversaries would be a central component of a deterrence strategy Emphasizing this link might even enhance the credibility of the U S commitment to retaliate Alternatively the United States might become ca- pable of denying adversaries the benefits of attacks in these domains through cyber defenses and substituting terrestrial assets for satellites In this case U S deterrence strategy would strive to convince potential adversaries that they cannot affect U S ground air naval and nucle- ar forces by attacking satellites and computer networks Such a message might make U S threats to respond of- fensively appear disproportionate and less credible but this would be a worthwhile tradeoff if the United States developed a defensive advantage in space and cyberspace Decisionmakers will also perceive attacks in space are on mainland China make this response escalatory and cyberspace differently depending on the context At- attack on a Chinese satellite--is proportional However be expected and accepted once a conventional war has One could argue that a symmetrical response--a kinetic if satellites play a smaller role in Chinese military operations one could also argue that such a response is less than proportionate because it does not impose comparable operational costs on China 12 The balance between offense and defense in these domains will also influence perceptions of effects escalation and proportionality and optimal deterrence strategies For example if offense continues to dominate in space and cyberspace and potential adversaries want to attack U S assets in these domains precisely because they are the U S military's soft underbelly U S stakes 6 SF No 272 tacks on military satellites and computer networks might started But similar attacks might trigger a conventional conflict if they occur prior to hostilities when both countries want to prevent a crisis from escalating into a war but are concerned about being left blind deaf and dumb by a first strike in space and cyberspace Propor- tionality and escalation are relative concepts actions that are escalatory during crises might be proportionate in limited wars and underwhelming responses as the scope and intensity of a conflict increase A related issue is whether U S reactions to cyber exploitation during peacetime would affect deterrence in www ndu edu inss crises Though the technology and operations of cyber ex- cyber attacks carry an intolerable risk of misperception fects are different exploitation extracts information from threats that leave something to chance U S officials ploitation and cyber attacks are similar the goals and ef- computers and networks without authorization attacks destroy degrade or alter them to achieve effects in other domains But news outlets frequently describe incidents 13 of cyber exploitation against the U S Government as cyber attacks and evidence of an ongoing war in cyberspace 14 miscalculation and unintended escalation Evoking could credibly argue that they are uncertain about what they would do because such attacks would involve a process that is not entirely foreseen reactions that are not fully predictable decisions that are not wholly deliberate events that are not fully under control 16 Conflating these operations contributes to the impression Of course expressing trepidation about unintended es- ies might conclude that U S threats to respond to cyber threatening such attacks would yield U S concessions that U S deterrence has already failed Potential adversarattacks in other domains lack credibility based on how the United States reacted to previous exploitation operations This perception might affect how they calculate risks and calation could backfire Adversaries may conclude that Conclusion Many weapons systems and most military operations benefits of cyber attacks in crises How can U S officials require access to multiple domains land air sea space different threats and require different responses especially actors can exploit by launching cross-domain attacks the publicly convey that cyber exploitation and attacks pose given the overlap between the two Emphasizing that the real-world effects of attacks and exploitation differ might be a first step toward establishing a threshold between the two This message would reinforce that deterrence has not failed because the effects of exploitation in cyberspace have not yet warranted U S military responses in other domains It clarifies the types of actions that the United States is attempting to deter Some strategists may conclude that proportionate counterspace and cyber responses are impossible because escalation control in these domains is too difficult There are an infinite number of scenarios that are neither in- dicative of a minor harassing incident of jamming nor and cyberspace These linkages create vulnerabilities that United States may seek to deter such attacks by threatening cross-domain responses Yet both the U S Govern- ment and potential adversaries lack a shared framework for analyzing how concepts such as proportionality escala- tion credibility and deterrence apply when capabilities in space and cyberspace not only enable operations in other domains but also are part of the battlefield The real-world effects of attacks that strike targets in space and cyberspace and affect capabilities and events in other domains should be the basis for assessing their implications and determining whether responses in different domains are proportionate or escalatory Integrating actions in the emerging strategic do- Assessing mains of space and cyberspace with actions in traditional sponses amid the stress and confusion of a military crisis toward more coherent cross-domain contingency plan- strategic attack in space and cyberspace 15 the effects of such attacks and choosing appropriate re- might be difficult U S and foreign officials likely will have differing views about the severity of nonkinetic dis- ruptions that defy easy categorization and the obstacles to developing a common framework might be too formi- dable Furthermore the effects of sophisticated attacks on satellites and computer networks might be indiscriminate and too difficult to predict In this case a deterrence strategy could emphasize that limited counterspace and www ndu edu inss domains in a clear escalation ladder would be a first step ning within the U S Government Communicating this framework to potential adversaries would contribute to more effective deterrence and crisis management Notes 1 Richard Smoke War Controlling Escalation Cambridge Harvard University Press 1977 252 2 See Department of Defense DOD Quadrennial Defense Review Report Washington DC DOD February 2010 33-34 37-39 SF No 272 7 The White House National Security Strategy Washington DC The White House May 2010 22 DOD National Security Space Strategy Washington DC DOD January 2011 The White House International Strategy for Cyberspace Prosperity Security and Openness in a Networked World Washington DC The White House May 2011 DOD Department of Defense Strategy for Operating in Cyberspace Washington DC DOD July 2011 3 A 2009 National Research Council report defined cyber attacks as deliberate actions that alter disrupt degrade or destroy adversary computer systems or networks or the information and or programs resident in or transiting these systems or networks whereas cyber exploitation is extracting information from computer systems or networks without authorization The report demonstrates that the intended effects of cyber attacks occur in other domains Direct or immediate effects are effects on the computer system or network attacked Indirect or follow-on effects are effects on the systems and or devices that the attacked computer system or network controls or interacts with or on the people that use or rely on the attacked computer system or network the indirect effect is often the primary purpose of the attack National Research Council Technology Policy Law and Ethics Regarding U S Acquisition and Use of Cyberattack Capabilities Washington DC National Academies Press 2009 80 4 For more on inter-domain relationships and vulnerability see Mark E Redden and Michael P Hughes Global Commons and Domain Interrelationships Time for a New Conceptual Framework INSS Strategic Forum 259 Washington DC National Defense University Press October 2010 5 For a discussion of the role of space counterspace and cyber capabilities in China's military strategy see David C Gompert and Phillip C Saunders The Paradox of Power Sino-American Strategic Restraint in an Age of Vulnerability Washington DC National Defense University Press 2011 chapter 3 James Dobbins David C Gompert David A Shlapak and Andrew Scobell Conflict with China Prospects Consequences and Strategies for Deterrence Santa Monica CA RAND 2011 5-7 Office of the Secretary of Defense Annual Report to Congress Military and Security Developments Involving the People's Republic of China 2010 Washington DC DOD August 2010 22-37 Jan Van Tol with Mark Gunzinger Andrew Krepinevich and Jim Thomas AIRSEA Battle A Point-of-Departure Operational Concept Washington DC Center for Strategic and Budgetary Assessments 2010 17-47 Roger Cliff et al Entering the Dragon's Lair Chinese Anti-Access Strategies and their Implications for the United States Santa Monica CA RAND 2007 51-60 6 For reports of U S debates about launching a cyber attack against Libya see Eric Schmitt and Thomas Shanker U S Debated Cyberwarfare in Attack Plan on Libya The New York Times October 18 2011 and Ellen Nakashima Pentagon Officials Had Weighed Cyberattack on Gaddafi's Air Defenses The Washington Post October 18 2011 7 Thomas C Schelling Arms and Influence New Haven Yale University Press 1966 146-149 8 9 Ibid 150-151 For a discussion of expert uncertainty surrounding intended and cascading effects of cyber attacks see National Research Council 121-128 10 James N Miller testimony for the House Armed Services Committee Subcommittee on Strategic Forces March 2 2011 11 For a theoretical historical study of the causes and implications of doctrinal differences in U S -China relations see Christopher P Twomey The Military Lens Doctrinal Differences and Deterrence Failure in Sino-American Relations Ithaca Cornell University Press 2010 12 This example demonstrates that symmetrical and asymmetrical responses to attacks in space and cyberspace are not synonymous with proportionate and escalatory responses however since symmetrical responses attack the same type of target with the same type of weapon in the same domain as the initial attack it likely would be easier for different countries to reach similar assessments about whether such responses are proportionate or escalatory Assessing asymmetric responses against targets in the same domain as the initial attack--which might attack the same type of target with different types of weapons different types of targets with the same type of weapon or different types of targets with different types of weapons--might be more difficult Assessing asymmetric responses against targets in different domains than the attack would likely be the most difficult because it requires a shared standard for determining equivalencies across domains 13 14 National Research Council 149-152 For example Michael Riley and Ashlee Vance Cyber Weapons The New Arms Race Bloomberg Businessweek July 20 2011 15 Susan J Helms Schriever Wargame 2010 Thoughts on Deterrence in the Non-Kinetic Domain Air Force Space Command High Frontier 7 no 1 November 2010 14 16 Schelling 95 Acknowledgments This Strategic Forum benefited greatly from discussions with and comments by Elaine Bunn and Phillip Saunders from comments by peer reviewers CAPT Mark Redden and Col Michael Hughes and feedback from a presentation for the Office of the Secretary of Defense for Policy Institute for National Strategic Studies The Center for Strategic Research within the Institute for National Strategic Studies provides advice to the Secretary of Defense Chairman of the Joint Chiefs of Staff and unified combatant commands through studies reports briefings and memoranda The center conducts directed research and analysis in the areas of strategic and regional studies and engages in independent and leading-edge research and analysis in related areas Nicholas Rostow Director Center for Strategic Research 8 SF No 272 The Strategic Forum series presents original research by members of NDU as well as other scholars and specialists in national security affairs from the United States and abroad The opinions conclusions and recommendations expressed or implied within are those of the contributors and do not necessarily reflect the views of the Defense Department or any other agency of the Federal Government Visit NDU Press online at www ndu edu press Hans Binnendijk Director INSS Phillip C Saunders Director of Studies Center for Strategic Research www ndu edu inss This document is from the holdings of The National Security Archive Suite 701 Gelman Library The George Washington University 2130 H Street NW Washington D C 20037 Phone 202 994-7000 Fax 202 994-7005 nsarchiv@gwu edu