Sarah B. Snyder Sarah B

CREATING REQUIREMENTS: EMERGING MILITARY CAPABILITIES, CIVILIAN

PREFERENCES, AND CIVIL-MILITARY RELATIONS

Alice Hunt Friend

Submitted to the

Faculty of the School of International Service

of American University

in Partial Fulfillment of

the Requirements for the Degree of

Doctor of Philosophy

International Relations

Chair:

Sharon K. Weiner, Ph.D. Sarah B. Snyder Sarah B. Snyder, Ph.D.

Kathleen H. Hicks, Ph.D.

Dean of the School of International Service 3/20/2020 Date

2020

American University

Washington, D.C. 20016

Alice Hunt Friend

2020

To my compatriots, civilian and military.

CREATING REQUIREMENTS: EMERGING MILITARY CAPABILITIES, CIVILIAN

PREFERENCES, AND CIVIL-MILITARY RELATIONS

Alice Hunt Friend

ABSTRACT

This dissertation explores the relationship between civilian and military preferences in the United States. A standard measure of the health of the civil-military relationship is whether civilian preferences prevail over military preferences in times of disagreement. Generally, the civil-military relations literature focuses on civilian efforts to impose their preferences on the military. But is it possible that the military is able to impose its preferences on civilians as well?

This study asks and answers the questions: Does the military shape civilian preferences, and to what extent? If the military does shape civilian preferences, under what conditions does it do so?

I contend that both purposeful actions by the military and factors natural to the civil-military relationship, each centered on the distribution of information resources, shape civilian preferences. I hypothesize that the less information civilians possess relative to the military, the more civilian preferences are based on military preferences. In three cases of emerging military capabilities, I find support for this hypothesis. Using comparative historical methods and process tracing, I examine the congruence of civilian and military preferences across time and find that military actors frequently framed and constrained civilian thinking about, and in some contexts dictated the purposes of, special operations forces, unmanned aerial vehicles, and cyber capabilities.

ACKNOWLEDGMENTS

Completing a PhD program is a group project. I am indebted to my peerless committee, who supported, cajoled, and improved me as much as they improved the dissertation. Dr. Sarah

Snyder was a steady source of support and motivation. She’s both a model academic and the reason I finally got cracking, and I am forever grateful. Dr. Kathleen Hicks is a mentor, role model, and friend. She has been endlessly giving of her time and intellect and enormously supportive of my career. I owe her so much more than I could ever repay. And finally, my chair,

Dr. Sharon Weiner, defines mentorship. She believed in me even when I didn’t. Mixing ferocity and excellence with compassion and humor, she refused to let me fall short of what I could achieve. She was simply the very best dissertation chair I could have ever asked for and a dear friend.

So many others lent intellectual and moral support along the way I fear I will leave someone out. But here goes: To Meredith (Cheese) Killough, Dr. Mara Karlin, Loren Schulman,

Melissa Dalton, Dr. Erin Simpson, Laura Meissner, Rhiannon Gulick, Abe Denmark, and Matt

Barkan. You are my pit crew. Thank you for the laughs, the help, and the love. Dr. Boaz Atzili, the Director of the SIS PhD program, has got to be the kindest academic on the planet. Dr. Zia

Mian told me I could do it. Dr. Risa Brooks and Dr. Jim Golby are the very best colleagues and friends in civ-mil academia. Shannon Culbertson gave me a passport into the special operations community and gamely read drafts of the SOF case study. Dr. Thomas Ehrhard blazed the trail on UAV research and was generous with his time. Dr. Michael Horowitz and Dr. Paul Scharre both offered support and advice about technology and national security, and robots! Dr. Josh

Rovner and Dr. Jacquelyn Schneider gave me the decoder ring for cyber issues, as did soon-to-be

Dr. Peter Roady. Dr. Kori Schake sent perfectly timed words of encouragement. Dr. Stephen

Tankel and Dr. Carolyn Gallagher were profoundly kind to me about the whole PhD process, and

Dr. Joseph Young gave helpful feedback on developing the research. I am grateful to Dr. Eugene

Gholz for supporting this project and acting as an outside reviewer. I am also appreciative of Dr.

Michael Desch’s thoughtful review of an early draft of the SOF chapter and for hosting me at the

Notre Dame Emerging Scholars in Strategic Studies conference.

My wonderful colleagues at the Center for Strategic and International Studies deserve special mention for being a source of insights and energy. My cohort at the School of

International Service was a great crew for a long journey, especially Brandon Sims and Cherie

Saulter. I may never have done this at all if I hadn’t seen the example of scholarly gumption set by Dr. Cyanne Loyle. The people I interviewed were tremendously generous with their time and memories.

Finally, I could not have done a thing without the love and support of my long-suffering family. My husband, Dr. (the other kind) Kaleb Friend who didn’t blink when I told him I was going to quit my perfectly good job and try academia, who read drafts and took the kids out and reminded me to eat my vegetables, is my hero. And to my beautiful boys, Judah and Ari, I love you so much. Thank you for inspiring me to make myself better. This is the ‘book’ Mommy has been working on.

TABLE OF CONTENTS

ABSTRACT ...... iv

ACKNOWLEDGMENTS ...... v

LIST OF ABBREVIATIONS ...... viii

CHAPTER 1 INTRODUCTION, THEORY, METHODS ...... 1

CHAPTER 2 SPECIAL OPERATIONS FORCES: 1977-2011 ...... 31

CHAPTER 3 UNMANNED AERIAL VEHICLES: 1952-2016 ...... 121

CHAPTER 4 CYBER CAPABILITIES: 1984-2019 ...... 212

CHAPTER 5 CONCLUSION ...... 257

REFERENCES ...... 291

vii

LIST OF ABBREVIATIONS

ACTD Advanced Concept Technology Demonstrator AFO Advance Force Operation ASD Assistant Secretary of Defense CIA Central Intelligence Agency CJCS Chairman of the Joint Chiefs of Staff CND/E Computer Network Defense/Exploitation CT Counterterrorism CYBERCOM Cyber Command DARO Defense Airborne Reconnaissance Office (D)ARPA (Defense) Advanced Research Projects Administration DASD Deputy Assistant Secretary of Defense DoD Department of Defense GPF General-Purpose Force HASC House Armed Services Committee HVT High-Value Target ISR Intelligence, Surveillance, and Reconnaissance JCS Joint Chiefs of Staff JPO Joint Program Office JSOC Joint Special Operations Command LIC Low-Intensity Conflict MFP-11 Major Force Program-11 NRO National Reconnaissance Office NSA National Security Agency NSC National Security Council OSD Office of the Secretary of Defense QDR Quadrennial Defense Review RSTA Reconnaissance, Surveillance, and Target Acquisition SASC Senate Armed Service Committee SEAD Suppression of Enemy Air Defense SEAL Sea, Air, Land Force SOCOM Special Operations Command SOF Special Operations Force UAS/V Unmanned Aerial System/Vehicle UCAS/V Unmanned Combat Aerial System/Vehicle UN United Nations VCJCS Vice Chairman of the Joint Chiefs of Staff

viii

CHAPTER 1

INTRODUCTION, THEORY, METHODS

Civilian control of the military is at the heart of civil-military relations. For those who focus on this relationship in mature democracies, the concern is seldom about the risk of military coups, but of excessive military influence on foreign policy.1 To identify the existence of undue military influence, scholars have established a standard measure of civilian control: if civilian preferences prevail over military preferences most of the time, civilians have adequate control over military institutions and their activities.2 This standard is especially important and useful in times of civil-military disagreement. Contests between civilians and the military must, on average, result in civilians’ favor to assure researchers that the civil-military relationship will not undermine democratic regimes.

The trouble with judging civilian control by how often civilian preferences prevail is that it assumes civilian preferences were not, themselves, the products of earlier cooptation.

According to the scholar Paul Pierson, preferences often reflect the particularly “ideational elements” of power. “Powerful actors,” Pierson argues, “can gain advantage by inculcating views in others.”3 Although the civil-military relations literature tends to believe preferences are

1 Russel F. Weigley, “The American Military and the Principles of Civilian Control from McClellan to Powell,” Journal of Military History 57, no. 5 (1993): 27-58, doi:10.2307/2951800; and Richard Kohn, “Out of Control: The Crisis in Civil-Military Relations,” National Interest 35 (Spring 1994), https://nationalinterest.org/print/article/out- of-control-the-crisis-in-civil-military-relations-343. 2 Michael C. Desch, Civilian Control of the Military: The Changing Security Environment (Baltimore, MD: Johns Hopkins University Press, 1999); Peter D. Feaver, Armed Servants: Agency, Oversight, and Civil-Military Relations (Cambridge, MA: Harvard University Press, 2003); and Mackubin Thomas Owens, “What Military Officers Need to Know About Civil-Military Relations,” Naval War College Review 65, no. 2 (Spring, 2010): 67-87, https://digital- commons.usnwc.edu/cgi/viewcontent.cgi?article=1430&context=nwc-review. 3 Paul Pierson, “Power and Path Dependence,” in Advances in Comparative Historical Analysis, ed. James Mahoney and Kathleen Thelen (London: Cambridge University Press, 2015), 137-139. 1

given by nature or organizational interests, preferences in fact can be constructed through political, bureaucratic, and procedural interaction.

Generally, civil-military relations literature focuses on civilian efforts to impose their preferences on the military.4 But is it possible that the military is able to impose its preferences on civilians as well? Military institutions and actors might also be able to generate convergence between their preferences and civilian preferences. Or, the processes of American government might result in convergence in ways civilians do not intend—and possibly do not even perceive.

This study asks and answers the questions: Does the military shape civilian preferences, and to what extent? If the military does shape civilian preferences, under what conditions does it do so?

The answers to these questions are important because if civilians are adopting military preferences frequently or basing their preferences largely on standards set by military actors, civilians are not controlling the military so much as endorsing it. In this world, civilians may have a veto on military action, but they exercise it much less often than they might if their values, knowledge, and perceptions were not so heavily influenced by military institutions.

I contend that both purposeful actions by the military and factors natural to the civil- military relationship, each centered on the distribution of information resources, shape civilian preferences. I hypothesize that the less information civilians possess about a capability relative to the military, the more civilian preferences are based on military preferences. In three cases of emerging military capabilities, I find support for this hypothesis. Using comparative historical methods and process tracing, I examine the congruence of civilian and military preferences across time and find that military actors frequently framed and constrained civilian thinking

4 Morris Janowitz, The Professional Soldier: A Social and Political Portrait (New York, NY: Free Press, 1960); Arnold Kanter, Defense Politics: A Budgetary Perspective (Chicago, IL: University of Chicago Press, 1979); and Feaver, Armed Servants. 2

about, and in some contexts dictated the purposes of, special operations forces, unmanned aerial vehicles, and cyber capabilities. Sub-groups of the U.S. military were able to shape civilian preferences in large part because of the informational asymmetries between them and civilian elites. Civilians paid far less attention to emerging capabilities than military actors, often had less expertise, and suffered from lower access to information, including access to sources of information that were alternatives to the military.

This study is presented in five chapters. The present chapter introduces the research question and then turns to the literatures on civil-military relations and military innovation to frame the discussion and aid in hypothesis generation. It begins by presenting the definition of

“preferences” that I use here and explores how studies of civil-military relations and military innovation understand the sources of military and civilian preferences. It then explores theories of preference change, presenting the gaps in scholarship and proposing ways that social and bureaucratic processes can construct preferences in the civil-military relations context. It closes with a discussion of the methods used to test the hypothesis and a preview of the empirical analysis that follows.

Literature Review

This study is about whether the military is a substantial source of civilian preferences.

Because the theoretical literature on civil-military relations, particularly those studies focusing on civilian control of the military in the United States, motivate my research question I situate the study in that literature. However, the civil-military relations literature is limited by its assumptions of civil-military preference sources, divergence, and inflexibility. To explore notions of preferences as socially pliable, I turn to other social science literatures, including

administrative science and policy analysis. Then I examine the military innovation literature, which has more explicit models of how actors can shape each other’s preferences within the civil-military context, and in particular for emerging military capabilities.5 Civil-military relations scholars themselves often touch on other social scientific theories, including organizational theory and bureaucratic politics, institutionalism, and behavioral economics.

These literatures, which already engage with each other extensively, form the foundation upon which I build my theoretical approach. Although the below discussion draws mainly on the civil- military relations and military innovation debates, I will note when I recruit literature from further afield to inform my deductions and intuition.

On Preferences and Preference Divergence

Following much of the social science literature, I define a preference as an option that an actor would choose (or privilege) over other available alternatives (e.g., McCarty and Meirowitz

2007). A preference is not, itself, a choice; it precedes and informs choices—a function of preferences that, on top of the literature’s use of them as a measurement device, justifies a deeper understanding of where preferences come from and whether government processes or actors’ choices can alter them.

Both the civil-military relations literature and the literature on military innovation share the above general definition of preferences. But they layer onto this definition the expectation that civilian preferences are generally different and independent from military preferences. The

5 The term “innovation” is not conclusively defined in the literature that relies on it, but it generally refers to major changes in doctrine, operational concepts, and military capabilities and weapons. In his review of the military innovation literature, Adam Grissom divides military innovation studies into four schools of thought. I concentrate much of my analysis here on what Grissom calls the civil-military model of military innovation, although I refer to work Grissom categorizes as the intraservice and cultural models as well. Adam Grissom, “The Future of Military Innovation Studies,” Journal of Strategic Studies 29, no. 5 (2006): 905-934, doi:10.1080/01402390600901067. 4

literature provides two justifications for this expectation. First, scholars argue (some simply assume) civilian and military preferences tend to come from different sources. To civil-military relations scholars, civilians and the military come from different organizational (Builder 1989) and social cultures, which instill in them different values—what Huntington (1957) referred to as conservative (military) and liberal (civilian) mind-sets. Others place more emphasis on the functional differences between the two broad camps, pointing out that the military is, of course, responsible for the operational conduct of war and that civilians concern themselves with broader political and strategic matters.6 The second assumption about preferences is that they are fixed, in part because the cultural and functional sources of preferences are themselves generally unchanging. Functional differences are the result not only of different functions but also the inherent hierarchy between civilians and the military, which also tends to generate differences in perspective.7 Peter Feaver, drawing on principal-agent studies, prominently highlights the difference in both the station and function of the military in his agency theory of civil-military relations, noting that “there is a de minimis difference in perspective that attends agency.”8

Because the military is subordinate to civilians, the two groups will always have a measurable difference in outlooks.

Military innovation scholars also emphasize the unique sources of civilian and military preferences. Studies tend to narrow to two sources in particular: 1) the international threat environment or structure of the international system, and 2) organizational culture. In testing each of these two drivers as explanations of military doctrine, Barry Posen (1984) asserted that

6 Feaver, Armed Servants; Richard Kohn, “Coming Soon: A Crisis in Civil-Military Relations,” World Affairs 170, no. 3 (Winter 2008): 69-80, https://www.jstor.org/stable/20672810; and Janine Davidson, “Civil-Military Friction and Presidential Decision-Making: Explaining the Broken Dialogue,” Presidential Studies Quarterly 43, no. 1 (March 2013): 129-145, https://www.cfr.org/content/member/PSQ2013_Broken_Dialogue_Davidson.pdf. 7 Eliot A. Cohen, Supreme Command: Soldiers, Statesmen, and Leadership in Wartime (New York, NY: Simon and Schuster, 2002). 8 Feaver, Armed Servants, 60. 5

civilian preferences are sensitive to changes in the international environment, whereas military preferences are indexed to their organizational interests. Jack Snyder (1984) comes to similar conclusions about civilian and military preferences, whereas Elizabeth Kier (1997) argues that civilians, too, are motivated by their own culturally defined values and assumptions, although she allows that civilian values are still different from military organizational values. Her study of the sources of doctrine calls into question the assumed difference in motivations for preferences but also the logic of preference formation. Kier’s contention is that there is no reason to expect civilians are more rational or less parochial than the military.

Many scholars observe that civilians at the elite level fundamentally want to maintain political power.9 Their preferences will therefore naturally align with whatever ensures continued access to that power, whether via reelection (for elected officials) or popularity with other civilian gatekeepers (for those who are appointed or hired permanently into their positions). Part of what defines access to power is conditioned by which government institution they are in— particularly the Congress, the White House, or in an executive agency such as the Department of

Defense. Civilians in each of these organizations will have an interest in both their personal political power and their organization’s political power.

But civilians’ interest in maintaining political power does not necessarily dictate particular defense policy preferences. Even for Posen, civilians may recognize that current military doctrine will not protect the country from defeat in war (which is against civilian political interests), but that does not then specify what will cause victory. Civilians may want their militaries to innovate, but their interests will not always constrain them to a particular

9 Elizabeth Kier, Imagining War: French and British Military Doctrine Between the Wars (Princeton, NJ: Princeton University Press, 1997); Amy Zegart, Flawed by Design: The Evolution of the CIA, JCS, and NSC (Stanford, CA: Stanford University Press, 1999); Feaver, Armed Servants; and Risa Brooks, Shaping Strategy: The Civil-Military Politics of Strategic Assessment (Princeton, NJ: Princeton University Press, 2008). 6

innovation pathway—although, depending on the specificity of constituents’ demands, they certainly could.

The military, however, does develop preferences for particular defense policy matters because its survival depends on specific policy choices—namely, those conferring resources and autonomy on the military and on sub-groups within the military.10 These interests can have a strong path dependence, because military sub-institutional structures and the major weapons platforms they often revolve around are dependent on continued resources and autonomy.

Organizational interests in stable budgets and organizational structures thus tend to imbue military preferences with both a specificity and an “inertia.”11 This inertia can cause military institutions to maintain a focus on certain weapons or capabilities long after their initial justifications expire. Military institutions and actors’ preferred policies thus may be at marked variance from what the international environment logically compels or what civilian political interests demand.12 Military preferences may also not be at variance with civilian preferences so much as focused on different things or at a more granular level.

This “irreducible difference” in perspective may mean civilian and military preferences diverge, or it may be simply that attention to military matters will be unequal—an idea explored further below.13 Distinctions between civilian and military preferences that appear more or less

10 Richard K. Betts, Soldiers, Statesmen, and Cold War Crises (New York, NY: Columbia University Press, 1991 [1977]); Barry R. Posen, The Sources of Military Doctrine: France, Britain, and Germany Between the World Wars (Ithaca, NY: Cornell University Press, 1984); and Feaver, Armed Servants. 11 Sharon K. Weiner, “Organizational Interests versus Battlefield Needs: The U.S. Military and Mine-Resistant Ambush Protected Vehicles in Iraq,” Polity 42, no. 4 (October 2010): 461-482, https://www.jstor.org/stable/40865427. 12 Edward L. Katzenbach Jr., “The Horse Cavalry in the Twentieth Century,” in The Use of Force: International Politics and Foreign Policy, ed. Robert J. Art and Kenneth N. Waltz (Boston, MA: Little, Brown and Company, 1971). 13 Feaver, Armed Servants, 59. 7

fixed to the different sources of preferences for each group may not reflect disagreement so much as different emphasis.

Preference Convergence and Change

Not only might civil-military preference divergence be overdetermined, scholars of both civil-military relations and military innovation point out that preferences can and do converge.

Scholars claim convergence happens for political, organizational, and procedural reasons, either because civilians manage to change military preferences or the military changes itself.14 But these studies have perpetuated an assumption of a unidirectional or organizationally contained process where military preferences are the ones doing the changing either because civilians engineer the change or military organizations change themselves. These studies neglect the possibility, introduced by the logic of their own explanations, that civilian preferences can also change by the same social mechanisms.

Civil-military relations scholars treat preference overlap as a desirable condition because it suggests civilians can exercise more control over their subordinates.15 If civilians and the military share preferences, the military is much more likely to act as directed by their civilian managers. Conversely, the greater the gap between civilian and military preferences—the more strongly they disagree—the more civilians will need intrusive monitoring mechanisms, backed up by the threat of punishment, to engineer military compliance with their preferences.16 Because

14 Theo Farrell and Terry Terriff, eds., The Sources of Military Change: Culture, Politics, Technology (Boulder, CO: Lynne Rienner 2002); and Stephen Peter Rosen, Winning the Next War: Innovation and the Modern Military (Ithaca, NY: Cornell University Press, 1991). 15 Desch, Civilian Control. 16 Feaver, Armed Servants; Brooks, Shaping Strategy; Jeff Donnithorne, “Tinted Blue: Air Force Culture and American Civil-Military Relations,” Strategic Studies Quarterly (Winter, 2010), https://www.airuniversity.af.edu/Portals/10/SSQ/documents/Volume-04_Issue-4/Donnithorne.pdf; and Jessica Deighan Blankshain, “Essays on Interservice Rivalry and American Civil-Military Relations,” (PhD diss., Harvard University, 2014), https://dash.harvard.edu/handle/1/12274304. 8

such compliance mechanisms are costly, civilians are generally incentivized to manufacture preference convergence* by promoting military officers who share their preferences already.17

Part of the literature on military innovation explores this notion of civilians seeking preference convergence with the military via alliances or coalitions. This is one of the elements of the debate between Posen (1984) and Stephen Peter Rosen (1991) in particular. The former argues that civilians, whose intervention in military institutions is necessary for military innovation, often must rely on “military mavericks” who share their preferences to champion those preferences within military institutions. These military mavericks buck their own organizations for the national good. But Rosen rejects both that civilians are the key change agents and the idea that the military’s black sheep are at all capable of forcing anything on their services.18 He argues that powerful individuals inside the military are the locus of service change. Although he concedes that civilians may be able to accelerate change by lending officers their support, he insists militaries change only when one military faction wins an internal

“ideological struggle.” Studies in U.S. Navy innovation processes have bolstered Rosen’s claim by emphasizing the importance of intra-service power brokers.19

The debate between Posen and Rosen presents a great deal of empirical evidence that civilians must always work through the military to effect military innovation. Whether civilians must find military “mavericks” who agree with them already or whether civilians simply act as

* Manufacturing convergence between civilian and military views in ways that favor civilian values is anathema to Huntington’s (1957) classic model of objective control of the military. In many ways, this study contributes to the voluminous civil-military relations literature that illustrates the impracticalities of Huntington’s ideal approach. 17 Rosen, Winning the Next War; and Feaver, Armed Servants. 18 Deborah Avant also rejects the notion that civilian intervention is necessary for military change. Deborah Avant, Political Institutions and Military Change: Lessons from Peripheral Wars. Ithaca, NY: Cornell University Press, 1994. 19 Vincent Davis, The Politics of Innovation: Patterns in Navy Cases (Denver, CO: University of Denver Press, 1967); and Bradd C. Hayes and Douglas C. Smith, eds., The Politics of Naval Innovation (Newport, RI: U.S. Naval War College, 1994). 9

external validators for what is fundamentally a military-driven process, they and their preferences must, at best, be endorsed by the military to play a role in military innovation. This suggests that civilian preferences are irrelevant to the military in cases of major military change.

Is that possible? That Rosen refutes Posen’s argument with the notion that the military is a closed ideological system does not square with the American military’s situation in government and society. It is not just civilians who look for allies. For budgetary reasons alone, the U.S. armed services are required on a daily basis to engage in debates about their activities with outsiders. As Allan Millet, Williamson Murray, and Kenneth Watman have shown, a military “must consistently secure the resources required to maintain, expand, and reconstitute itself. Almost always, this necessity requires the military to obtain the cooperation of the national political elite.”20

Empirical work has shown a bi-directional pattern of influence and collaboration in the civil-military relationship. Civilian and military actors seek each other out in the hope of finding like-minded change agents. Civilian and military actors often form coalitions around shared policy preferences*—sometimes called issue networks.21 Such coalitions are possible because there is a great deal of internal fragmentation within the civilian and military camps.22 That civilians and the military often rearrange themselves into mixed coalitions rather than remaining

20 Allan R. Millet, Williamson Murray, and Kenneth H. Watman, “The Effectiveness of Military Organizations,” International Security 11, no. 1 (Summer 1986): 37-17, doi: 10.2307/2538875. * I am grateful to Jim Golby for calling my attention to the prevalence of such civil-military coalitions. 21 Betts, Soldiers, Statesmen, and Cold War Crises; Kimberly M. Zisk, Engaging the Enemy: Organization Theory and Soviet Military Innovation, 1955-1991 (Princeton, NJ: Princeton University Press, 1993); Gregory A. Engel, “Cruise Missiles and the Tomahawk,” in The Politics of Naval Innovation, ed. Bradd C. Hayes and Douglas C. Smith (Newport, RI: U.S. Naval War College, 1994); Desch, Civilian Control; and Christopher P. Gibson and Don M. Snider, “Civil-Military Relations and the Potential to Influence: A Look at the National Security Decision- Making Process,” Armed Forces and Society 25, no. 2 (Winter 1999), doi:10.1177/0095327X9902500202. 22 Samuel P. Huntington, The Soldier and the State (Cambridge, MA: Belknap Press, 1957); and Samuel P. Huntington, “Interservice Competition and the Political Roles of the Armed Services,” American Political Science Review 55, no. 1 (March 1961): 40-52, doi:10.1017/S0003055400124165.

firmly within their institutions suggests that preferences are more diverse and less anchored to

“civilian” and “military” identities and organizational cultures than some assume.

Both logic and the empirical record suggest that convergence works both ways: it is also the case that when civilians agree with the military’s preferences, the military can do what it wants. Despite the hierarchical structure of the relationship, the logic of beneficial preference convergence is not unidirectional. The military agent has just as much incentive to shape the preferences of its civilian principals as vice versa. In a system where civilians control the purse strings, deployment orders, and major strategy formulation processes, the military has every incentive to sell civilians on their interpretations of events. In fact, the formal hierarchy inherent to the civil-military relationship gives the military even more incentive than civilians to generate as many stakeholders in their preferences as possible. It is far better to get orders you agree with than orders you find objectionable or, worse, directly threatening to your existence. Why go through the pain and punishment of openly defying political direction when you can reverse- engineer it? And the more civilians agree with the military perspective, the greater the security of military preferences.

So long as civilians came to their preferences independently from the military, preference convergence is indeed a happy coincidence rather than a problem. But if preference convergence is a product of military efforts or the structure of the situation which presses civilians to adopt military preferences more often than not, such convergence does not favor civilian preferences.

In other words, preference convergence may not be an unadulterated sign of civilian control after all.

Persuasion and Information

The question is, can the military manufacture preference convergence in its favor? They certainly cannot do so in the ways available to civilians—they cannot actually select their civilian controllers like civilians can select their military officers. Instead, the military must rely on other mechanisms to shape the preferences of the civilians they end up with.

The most obvious mechanism is persuasion. Despite many scholars’ commitment to the irreducible differences between civilian and military preferences, the civil-military relations literature also has long explored the capacity of the military to persuade civilians to share military views of policy. Such studies have generated evidence that suggests military actors persuade civilians to share military preferences in at least two ways: First, the military can make conscious efforts to persuade civilians to share their preferences. Scholars argue that the military can constrain civilians’ choices by controlling the “premises” of decisions or the “standards of argumentation.”23 This kind of leverage involves judgments about what is a good or bad decision, changing preferences by controlling the criteria upon which policy decisions are based.

Civilians may end up settling for alternatives that the military has defined as possible. Jeff Legro calls this an “option funnel,” a process whereby “[t]he weapons, plans, and skills developed according to [military] cultural penchant often decided what national leaders would do even before they considered a situation.”24 Social scientists have long dubbed such efforts “frames.”

Frames provide interpretations of the world that work on both individual perceptions and on

23 Betts, Soldiers, Statesmen, and Cold War Crises; Richard Kohn, “How Democracies Control the Military,” Journal of Democracy 8, no. 4 (1997): 140-153, doi:10.1353/jod.1997.0060; and Cori Dauber, “The Practice of Argument: Reading the Conditions of Civil-Military Relations,” Armed Forces and Society 25, no. 3 (1998): 435- 446, doi:10.1177/0095327X9802400306. 24 Jeff Legro, “Culture and Preferences in the International Cooperation Two-Step,” American Political Science Review 90, no. 1 (1996): 118-137, doi:10.2307/2082802. 12

perceptions at group levels.25 Feaver (2003) points out that the military may “shade estimates” of the costs of military operations—a way to frame policy options as less desirable, pushing them down the ranks of preferences.

Framing involves selective presentation of information, which leads me to the second way that the military can shape civilian preferences: by controlling civilians’ access to information. The most obvious way that defense information can be hidden is via formal secrecy and classification. As Feaver writes, “the information classification system vastly eases the task of an agent who wishes to keep inconvenient information from being disseminated.”26 But the military could conceivably limit civilians’ access to information in other ways, including by peppering presentations with jargon or simply withholding knowledge. The point is that when the military has control over the release of information to civilians, it has the ability to withhold or share information in ways that can affect civilians’ views. Manipulating and managing the flow of information are two ways the military can influence how civilians think about policy options.27

Yet there is a broader conceptualization of access to information. Access may also mean that civilians have more than one source of information. Congressional oversight of other agencies, for example, benefits from a range of sources of information and expertise in academia, interest groups, and the media. Although there are fewer such competing sources of information and insight when it comes to national security bureaucracies,28 there are outside

25 Aaron Tversky and Daniel Kahneman, “Prospect Theory: An Analysis of Decision Under Risk,” Econometrica 47, no. 2 (March 1979): 263-292, doi:10.2307/1914185; Dennis Chong and James N. Druckman, “Framing Theory,” Annual Review of Political Science 10, no. 1 (2007): 103-126, doi:10.1146/annurev.polisci.10.072805.103054; and Michael J. Butler, Selling a ‘Just’ War: Framing, Legitimacy, and U.S. Military Intervention (London: Palgrave Macmillan, 2012). 26 Feaver, Armed Servants, 71. 27 Brooks, Shaping Strategy. 28 Zegart, Flawed by Design. 13

sources of information, including the media and think tank sectors.29 For civilians, access to information is not just about having the right clearances but also about the existence of a network of alternative, non-military sources. However, there may be great variance in the range of such sources of information depending on the issue at hand. Additionally, military fragmentation means that sub-groups of the military, especially those dissenting from their own institutions’ dominant preferences, can also serve as alternative sources of information. There may be informational utility for civilians in forming coalitions with sub-groups of the military.

Such reliance on military institutions for information raises the possibility that civilians may adopt military preferences not only because the military intends for them to do so but also because of the natural “information asymmetries” that are outgrowths of the hierarchical relationship between civilian managers and military experts.30 Expertise itself is an information asymmetry built into the design of bureaucracy.31 Civilians generate a military in the first place to ensure that a competent cadre of experts in the use of violence can defend the country from attack while civilian leaders attend to other matters of governance. The military is more expert in warfare than civilians because of this hierarchical division of labor.

What is expertise? Scholars generally refer to expertise as a combination of technical knowledge, skills, and experience. Knowledge and skills are the products of extensive and ongoing training, exercising, and education. The training and education the military provides itself is not generally available to those outside military institutions, but it is far easier to learn about and even witness first-hand than it is for civilians to get direct operational military experience. As Feaver points out, combat experience is a unique expertise that only combat

29 Feaver, Armed Servants. 30 Ibid. 31 Max Weber, Economy and Society (Berkeley, CA: University of California Press, 1922). 14

veterans can acquire.32 Not even the entire military, in fact, will have such expertise, but civilians almost certainly do not—unless they themselves are retired combat veterans. Even then, current operations give information not immediately available to civilians to those deployed, suggesting that civilians are even more disadvantaged in wartime. The hierarchical nature of expertise in the civil-military relationship thus confers on the military a latent power to influence what civilians know and confers on civilians a dependency on the military for privileged information.

But the military can also harness the power of its expertise. Much of the military’s persuasive power comes from its domestic legitimacy with civilian elites, which only bolsters their already high degree of respect for military expertise.33 This combination of legitimacy and expertise gives the military the power to persuade civilians of military views.34 Moreover, the military’s explicit role in strategy formation in the U.S. is to convert its expertise into advice, and this advisory role can also result in the kind of option funnels that Legro (1996) describes and

Golby and Karlin (2018) confirm. The military may thus convert expertise into proactive influence—in ways that may even approach the definition of lobbying.35

The problems of expertise are inherent parts of bureaucracy that are not unique to the civil-military relationship but that have special consequences in the context of civil-military

32 Feaver, Armed Servants, 70. 33 Millet, Murray, and Watman, “The Effectiveness of Military Organizations,” 38-39. 34 Betts, Soldier, Statesmen, and Cold War Crises; Stephen J. Cimbala, “The Role of Military Advice: Civil-Military Relations and Bush Security Strategy,” in U.S. Civil-Military Relations: In Crisis or Transition?, ed. Don M. Snider and Miranda A. Carlton-Carew (Washington, DC: Center for Strategic and International Studies, 1995); Jeh Charles Johnson, “National Security Law, Lawyers and Lawyering in the Obama Administration,” (lecture, Yale Law School, New Haven, CT, February 22, 2012). 35 By lobbying I mean attempts to influence governmental decisions, policies, or laws or the officials making those decisions, policies, and laws. This refers to lobbying in the sense political scientists attribute to interest group advocacy. See Jeffrey M. Berry and Clyde Wilcox, The Interest Group Society (New York, NY: Routledge, 2018); There is also an extensive literature on the relationship between interest group advocacy and the provision of information to decision-makers. See for example, Jan Potters and Frans Van Winden, “Lobbying and Asymmetric Information,” Public Choice 74 (1992): 269-292, https://www.jstor.org/stable/30025608; Matthias Dahm and Nicolas Porteiro, “Informational Lobbying Under the Shadow of Political Pressure,” Social Choice and Welfare 30, no. 4 (May 2008): 531-559, doi.org/10.1007/s00355-007-0264-x; and John M. de Figueiredo, “Lobbying and Information in Politics” Business and Politics 4, no. 2 (2002): 125-29 doi:10.2202/1469-3569.1033. 15

relations due to the special constraints on the possession of information and the importance of civilian control of military agents. There is even some evidence that trends in civilian and military information asymmetries may increasingly favor the military. Drawing on data from the

U.S. between the 1960s and mid-1990s, one study charts a decline of civilian expertise in military matters and experience in government generally, asymmetric expertise the authors find allows the military to make convincing arguments to civilian policymakers.36

However, investigations of the military’s purposeful efforts at persuading civilians to share their preferences and of how natural conditions in the civil-military relationship can make civilians vulnerable to military persuasion in the first place downplay civilians’ role in these interactions. How civilians actually behave with respect to their own preferences and how they respond to military efforts at persuasion suggest civilians themselves may be participating in their own deference to, and adoption of, military preferences.

As discussed above, civilians have their own institutional interests, the fulfillment of which depends on politics and continued access to power. But civilians’ interests may or may not generate in them specific preferences about military policies. Civilians often prefer to keep their options open or are disinterested in or simply unable to track every issue. Public policy theorists note this tendency for preferences to be discovered in the course of government activity rather than forming the basis for that activity.37 James Q. Wilson argues that executive preferences often are “too vague or ambiguous” to be turned into tasks, a condition that government “operators” respond to by acting on the basis of their own values.38 Civilians (the

36 Gibson and Snider, “Civil-Military Relations and the Potential to Influence.” 37 Michael Cohen, James March, and John Olsen, “A Garbage Can Model of Organizational Choice,” Administrative Science Quarterly 17 (March 1972): 1-25, doi:10.2307/2392088. 38 James Q. Wilson, Bureaucracy: What Government Agencies Do and Why They Do It (New York, NY: Basic Books, 1989). 16

executives in this case) may not develop preferences over defense issues at the same rate or with the same intensity as the military (the operators). Now add to the equation the information asymmetries often found between civilians and the military, as well as the military’s own efforts to persuade civilians to adopt their preferences using framing, and a civilian tendency to defer to the military becomes highly probable.

Because the asymmetries of information and expertise enable the military to persuade civilians to adopt their preferences, the clear way for civilians to form their preferences in a process independent from the military would be to acquire alternative information and expertise.

The civil-military relations literature refers to such information-gathering as monitoring or oversight,39 but for my purposes it is more usefully conceived of as attention. Whereas monitoring for compliance suggests a pre-existing preference, attention relaxes this requirement and instead focuses on general awareness. It is somewhat analogous to McCubbins and

Schwartz’s (1984) concept of the “police patrols,” but it hinges on that notion’s surveillance techniques without assuming specific, enforceable preferences that are not necessarily there.

Actors who pay attention know the key players and arguments, who has what interests and agendas, and where power is concentrated. Actors paying sustained attention know how these factors have trended over time. They know the history of the matter. Sustained attention allows the accumulation of expertise and understanding of what information exists and who has it. It also allows actors to perceive efforts to frame information.

Together, these three elements of information—attention, expertise, and access—supply civilians with the substantive building blocks of preferences. The more attention they have paid, the more expertise they have built, and the more sources of information they have as alternatives

39 Avant, Political Institutions and Military Change; Cohen, Supreme Command; Feaver, Armed Servants; and Brooks, Shaping Strategy. 17

to the military view, the more they should be able to give their own meaning to events and situations, including whether they affirmatively agree with military judgments or uncritically accept them. Alternative information resources should be especially useful when a civilian has reason to believe one’s military agent is giving meaning to situations colored by personal interests and biases.

However, as the civil-military relations literature and underlying administrative science point out, information-gathering is costly.40 The whole point of hiring experts is to save the time one would have to spend oneself performing that function. Moreover, intrusive monitoring drives up tension in the principal-agent relationship. What discussion there is about civilians monitoring the military is thus generally about how they can economize on their efforts and minimize their intrusiveness—hence the conclusion that it is easier to simply select officers that share civilian preferences than it is to pursue various more time-consuming and intrusive methods.41 By selecting more minimal monitoring methods, however, civilians may also be exacerbating the problems of information asymmetry and thereby adding to the conditions that cause them to adopt military preferences.

Time is also a factor. Informational asymmetries, including the allocation of attention, also give the military opportunities to form preferences over discrete issues at earlier points than civilians. By the time civilians turn to issues, military actors often have a durable preference of their own and a practiced set of frames and evaluations they can supply to less-expert civilian decision-makers with limited access to information. Together, these dynamics create a path dependence for military preferences that is difficult to re-route. For their part, civilians may have

40 Feaver, Armed Servants; and Colton C. Campbell and David Auerswald, eds., Congress and Civil-Military Relations (Washington, DC: Georgetown University Press, 2015). 41 Feaver, Armed Servants. 18

veto power over military influence, but they must choose to exercise it, and adopting ready-made preferences advanced by the military may often be the most efficient way to understand and evaluate military matters and, eventually, choices.

Theoretical Assumptions

To test for military influence on civilian preferences, I draw a series of six assumptions and one testable hypothesis from the literatures explored above. They are as follows:

1. The international environment and domestic politics will influence both civilian and

military preferences.

2. However, cultures, ideas, and institutional and political constraints germane to military

and civilian communities will cause them to interpret these external factors in different

ways and attach different meanings and degrees of importance to them.

3. Civilians have incentives to engineer convergence between their preferences and the

military’s preferences; the military also has incentives to engineer convergence between

their preferences and civilian preferences.

4. Civilians and the military are fragmented. Both meaning making and preferences will

vary not only between civilians and the military but also between civilians and between

military groups.

5. The military often has more information on military matters than civilians, whether due

to civilian attention, expertise, or access problems.

6. Military expertise, especially that derived from combat experience, has legitimacy that is

difficult for civilians to contest.

Hypothesis

H: Under conditions of low civilian information, civilians tend to adopt military preferences.

Based on theoretical and empirical work in civil-military relations and military innovation scholarship, it is reasonable to expect that military preferences often form the basis of civilian preferences, but only under certain conditions. A key such condition is when civilians have less information than military actors. “Low information” thus is defined relatively, as it connotes asymmetry between civilians and the military, due to at least one of the following elements: expertise, access to information, and attention. Expertise connotes both education and experience, whereas access to information connotes a real-time ability to acquire information that is privately held by the military. Access to information also refers to whether civilians can gain information from alternative non-military sources. Attention is something that civilians choose to allocate and allows them to develop expertise and cultivate information. It includes but is not limited to the concept of monitoring.

Under these conditions, civilians can adopt military preferences by default, by choice, or as a result of military persuasion. Note that military actors do not threaten stonewalling or logrolling. Instead, this hypothesis describes the ambient conditions of relative civilian and military power and expertise and civilian susceptibility to absorbing and adopting military values and choice criteria. Civilians do not automatically adopt military preferences but are highly likely to do so given informational asymmetries and persuasion by military groups.

Methods: Case Selection, Key Concepts, Measurement

To prove that the military has an independent effect on what civilians want, I must also rule out the possibility that another, confounding variable systematically shapes both civilian and military preferences. Yet identifying discrete civilian and military preferences will be difficult given the dynamic nature of preference and policy formulation. This makes the issue of time order vital: I must be able to show that the military began developing a preference prior to civilians either adopting it or beginning to develop a similar preference according to the same rationale. And because this inquiry involves the development and widespread adoption of ideas, the timelines involved extend over years. I must also show that alternative preferences were available to civilians and that they adopted military preferences, nonetheless.

I therefore conduct congruence testing followed by detailed process tracing, which together reveal the presence of the hypothesized relationships between military actions and civilian preferences, the conditions under which the military is able to shape civilian preferences, and the mechanisms of military-driven change to civilian preferences. Within each case I then apply counterfactual reasoning to test the validity of my findings, removing factors and variables that may have had confounding effects on the relationship between civil-military information asymmetries and civilian preferences. To probe the role of the military in civilian preference development and change, and to establish generalizable findings, I complement my in-case analysis with cross-case comparison and comparative historical analysis across three cases: special operations forces, unmanned aerial vehicles, and cyber capabilities.42

42 On cross-case comparison, see: Alexander L. George and Andrew Bennett, Case Studies and Theory Development in the Social Sciences (Cambridge: The MIT Press, 2005); and on comparative historical analysis, see: Derek Beach and Rasmus Brun Pedersen, Process-Tracing Methods: Foundations and Guidelines (Ann Arbor: The University of Michigan Press, 2013); and Pierson, “Power and Path Dependence.” 21

I draw these three cases from the universe of what I call emerging military capabilities.

Emerging military capabilities are suites of tools and techniques in the process of being adopted or re-adopted by one or more military services or have clear adoption potential being actively debated within defense communities. Emerging capabilities transition from niche to widespread use, meaning military services and operational commands are still in the process of developing doctrine, training, and concepts of operation for their use. Critically, the elite defense community is also still debating the strategic utility of such capabilities. It is this latter element of emerging military capabilities that provides opportunities to track the civil-military dynamics of preference influence and change.

Using cases of emerging military capabilities provides several forms of leverage over my research question and hypothesis. First, military capabilities represent the nexus of military and civilian jurisdictions and hence a natural place to find military and civilian preferences intersecting.43 Second, focusing on capabilities that are emerging allows me to trace the development of civilian and military preferences from their origins, giving me greater confidence that I can evaluate whether or not civilian preferences are independent from the military. Third, it is consistent with and contributes to much of the military innovation literature, which tends to examine military organizational concerns such as doctrine and weapons development. Debates over the uses of emerging capabilities are thus good laboratories to test the foundations of civilian preferences and their relationship to the military. As civilians consider how to use new

43 The development of military capabilities is in a real sense part of the traditional civil-military literature’s definition of a military responsibility. The Huntington model dictates that civilians establish strategic ends and the military provides the operational means. Huntington, The Soldier and the State. Nevertheless, in the modern system civilian strategic guidance expressly endorses the mix of military capabilities to be resourced and links those capabilities to objectives civilian leaders want the military to be able to achieve. Capabilities development is thus the area where civilian and military roles are the most ill-defined and where their preferences most often clash. 22

capabilities, the line-up between military preferences at the start of such processes and eventual civilian preferences at the end will signal congruence in civilian and military preferences.

The cases of special operations forces (SOF); unmanned aerial vehicles (UAVs); and cyber capabilities involve all four U.S. military services and a variety of other military organizations, including operational (combatant) commands. The cases also span several events that act as within-case natural experimental treatments on civilian and military preference formation processes: the Cold War and its end; the passage of the Goldwater-Nichols

Department of Defense Reorganization Act of 1986; the first Gulf War; the 9/11 terrorist attacks; and the wars in Afghanistan and Iraq. This span of time also allows me to contribute new data to the existing, Cold War-centric record on civil-military relations scholarship by providing new empirical evidence from the post-9/11 period.

The key variation between the cases is in the amount of information available to civilians about the capability itself: how much attention civilians applied to the emerging capabilities, their comparative expertise, and alternatives to the military for information sources. The latter dimension of information particularly distinguishes the cases. In the case of SOF, civilians had no outside, non-military sources of information. In the case of UAVs, civilians could turn to developers in the private sector and operators in the intelligence community, although both of these groups were highly integrated with military officers and veterans. In the cyber case, civilians had the greatest source of alternative information and ideas because of the booming civilian technology sector—although secrecy in the cyber case was an important intervening variable between non-military sources of information and civilian preferences. Given the third hypothesis, I expect that civilian preferences for SOF will be most directly informed by those provided to them by military actors; civilian preferences for UAVs will be based on military

preferences but may incorporate other views; and civilian preferences for cyber should be the most independent of military preferences.

Concepts, Key Variables, and Operationalization

Following Gerring (1999), my aim is to make the concepts underpinning this research parsimonious, coherent, theoretically useful, and well-differentiated from related ideas. In order for this project to be legible, both the researcher and reader need a clear sense of what is meant by “preference,” how preferences will be identified as “civilian” and “military,” and how I intend to measure civilian and military preferences as discrete variables.

Preference: A preference is an option that an actor (in this case, institutional-level actors) would choose over other available alternatives. It is not the same as an actual choice, which implies an action taken to the exclusion of other options, although preferences are “drawn on when people make decisions.”44

My dependent variable is civilian preferences. My hypothesized factors that condition civilian preferences are collected under the term “information,” which comprises attention, expertise, and access to information, as described above. Military preferences are those preferences held by military institutions and actors, and civilian preferences are those preferences held by civilian institutions and actors. Operationalizing preferences is a matter of finding formal expressions of preference in documents, speeches, or on-the-record comments from senior leaders representing their institutions in their official capacities. An important distinction is that, although I may analyze individual actors’ preferences, my concern is with

44 James N. Druckman and Arthur Lupia, “Preference Formation,” Annual Review of Political Science 3, no. 1 (2000): 1-24, doi:10.1146/annurev.polisci.3.1.1.

preferences that represent institutions and groups (e.g., the Navy, the White House, the Office of the Secretary of Defense). Personal preferences, although they may not be sharply different from an individual’s institutionally based preferences, are not the object of analysis here.

Civilian: Both the concepts of civilian and military should bear a strong family resemblance to the way they are used in the civil-military relations literature if my aspiration is to contribute to that literature. Civilians are the group of elite-level national security policymakers in the executive and legislative branches who are not uniformed members of the military and who have responsibility for civilian oversight of the military.45 In particular, I will focus on the Secretary, Deputy Secretary, Under Secretaries, and Assistant Secretaries of

Defense as well as members of Congress and officials from other agencies who sit as members of policy decision bodies at the National Security Council (NSC).46 Because civilians often fracture along institutional lines, I will be specific about which civilian organization is represented in preferences.

Military: The military is the cluster of uniformed services, combatant and other operational commands, the Joint Staff, and the Chairman of the Joint Chiefs of Staff. For the purposes of this research, the military actors whose preferences are of interest represent military institutions at high levels of command—the Services, COCOMs, and Joint Staff—but also officers from sub-groups of the military that have coalesced around a particular preference. For purposes of operationalization, officers at the heads of military organizations will be treated as speaking for the organization as a whole in terms of documents they approve or comments they make. This includes four-star generals or admirals who are chiefs of their services and junior officers participating in lower-level advocacy for a preference held by a military group.

45 Cohen, Supreme Command. 46 Commonly referred to as the Deputies Committee and the Principals Committee. 25

Consistent with observations in the scholarly literature of intra-military fragmentation, I do not require—and do not expect—total cohesion at the service or Joint Staff level. Rather, I am examining whether preferences expressed by authoritative elements of military institutional groups meaningfully and substantially informed how civilians later constructed their own preferences. For example, if the Army preferred counterinsurgency capabilities in the face of Air

Force skepticism, I am interested in whether the resulting conflict caused civilians to adopt one or the other military groups’ counterinsurgency-related preferences.

Measurement and Data Collection

This project measures preferences in terms of language used and ideas expressed and looks for parallels or departures between military institutional expressions and those used by civilians. An expression of preference is coded as institutional and assigned to a military or civilian institution by virtue of the context in which it is expressed and the role being played by the actor expressing it. For example, statements issued or uttered by the Secretary of Defense reflect civilian, and specifically OSD institutional, preferences. As described above, measurement is combined with process tracing to identify which institution and institutional actors truly “own” a preference. A preference is coded as dominant for a group or sub-group of military or civilian actors if it persists over time and is expressed frequently and at high levels of rank and/or official documents.

There may be significant intra-military and intra-civilian disagreement during the process of preference formation. What concerns me is that the military develops its preferences independently and thus is able to have an independent effect on civilian preferences. Indeed, monolithic civilian and military blocs would be inconsistent with my expectations about a

system that allows for persuasion and change. What is important is that each institutional group coalesces around a dominant preference that can be attributed to that group.

The three dimensions of information—attention, expertise, and access—are also measured qualitatively as well as relatively. What matters is not absolute measures of information but the degree of asymmetry of information possessed by or accessible to civilians and military actors. Attention is measured in the aggregate frequency of comments, hearings held, studies and documents produced, and appearances in strategic documents but also in the quality of such statements and whether parties persist in their interest in the capability. Expertise comprises formal educational background as well as experience, information I derive from biographies of individuals, their own and others’ statements about their expertise. Access is also measured by surveying available alternative sources of information and via comments made in documentary evidence and by interview subjects regarding the effect of secrecy on information flow to military and civilian actors.

This project relies on qualitative methodological triangulation and thus will require collecting different kinds of data. My main source of evidence of military and civilian preferences over the use of military capabilities are primary source government documents authored by the Department of Defense (DoD), military departments, the State Department, and the White House. To capture civilian preferences, I use strategy documents signed by the

President, the Secretary of Defense, or one of the Secretary’s senior deputies, as well as statements issued by senior civilians, including members of Congress, either formally or via news media. To collect military preferences, I rely on a range of government, interview, and secondary sources. Primary sources included publications produced by the service war colleges, articles in service-oriented journals such as Parameters, posture statements, and official

organizational histories (for example, the history of Special Operations Command authored by the command historian).

I also surveyed congressional records pertaining to the cases, primarily the question-and- answer sessions of relevant committee hearings. Because witnesses do not always know the questions in advance, the answers they supply tend to be more candid than their written testimony. I also examined media interviews and questions-and-answer periods after non- congressional speeches for the same reason.

To find both preferences and construct the larger narrative context of their development, I reviewed contemporary media sources, conducting comprehensive searches of the New York

Times and the Washington Post for the years spanning each case study using keywords that referred to the capabilities that made up my cases. I also consulted the memoirs of key government officials. To round out the documentary evidence, I reviewed secondary sources, including histories and books authored by journalists and academics.

Finally, as a means of fact-checking and corroboration (rather than as primary sources for data—see Berry 2002), I conducted sixteen semi-structured elite interviews with former senior officials who served at DoD, the White House, and in the Congress. Both the individuals selected for interviews as well as the interview designs were based on the results of the document analysis described above and followed Aberbach and Rockman’s (2002) recommendations to use open- ended questions. Following King and Horrocks (2010), I selected interviewees with a snowball model, beginning by identifying individuals in the documentary record and then relying on them to suggest additional interviewees.

Together, these sources of data provide me measurements of civilian and military preferences across time, as well as rich descriptions of the processes and mechanisms that shaped and changed civilian preferences regarding the uses of emerging military capabilities.

Remaining Plan of Study

The remaining chapters of this study consist of the case findings and summary conclusions. Chapter two presents the special operations forces case, revealing how military actors first struggled within their own institutions over the justifications for resourcing SOF capabilities, then leveraged their unique expertise and civilians’ limited access to information to persuade civilians in the Pentagon, the White House, and Congress to adopt one or the other of the two main military preferences regarding the purposes and value of SOF. The case also explores the degree to which longtime civilian inattention followed by the exigencies of wartime allowed SOF advocates to persuade civilians to make SOF the dominant counterterrorism tool for the U.S. government.

Chapter three presents the unmanned aerial vehicles case, showing how the military services’ ideational (and generally operational) limitations of UAVs to reconnaissance roles also limited civilian preferences to the same for over sixty years. The UAV case highlights the differential expertise and informational focus of civilians in different institutions within the

Pentagon and Congress and shows how such focuses drive the application of attention in ways that privilege military preferences. The chapter then turns to the surprising liberation of civilian preferences from military influence in the last few years of the case period and explores some of the reasons for civilians’ inability to advance their independent preferences with the military

services, pointing to the path dependence of military preferences once they have informed actual policy and practice.

Chapter four presents cyber capabilities as a shadow case. Although the development of both military and civilian preferences for the military uses of cyber tools is still unfolding, there are patterns in the relationship between military and civilian preferences that are also present in the SOF and UAV cases, suggesting my findings may generalize beyond the detailed cases presented here.

I conclude the study in chapter five with observations of the patterns across the cases— particularly the role played by coalitions of civilian and military actors in forming and changing civilian preferences—and a discussion of the study’s limitations. I also discuss the study’s contributions to the scholarly literature on civil-military relations and propose a research agenda to probe the findings presented here for generalizability and broader implications.

CHAPTER 2

SPECIAL OPERATIONS FORCES: 1977-2011

This case study of Special Operations Forces (SOF) capabilities from 1977 to 2011 shows that low levels of attention and a lack of access to alternative sources of information led civilians to adopt military preferences at multiple points in the re-emergence of special operations capabilities. Even where international environmental and domestic political factors shaped civilian preferences about SOF, military actors frequently mediated the effects of those factors by exploiting civilian inattention and using military expertise and private information to interpret the strategic meaning of important events for civilian policymakers. Civilians almost never developed preferences about SOF without military sources of information shaping them to an important extent.

There were two major, interrelated reasons for this outcome. First, on average civilians were inattentive to SOF. For much of the case, SOF and the kinds of missions they traditionally performed were minor considerations in the contexts of defense strategy and budgeting.

Civilians, especially at the elite levels, simply did not develop much expertise on special operations. For a period of more than twenty years, SOF were the only institutional actors applying sustained attention to SOF capabilities, making strategic and organizational adjustments to SOF in light of the international environment and military operations. Second, civilian inattention gave the military autonomy over SOF capabilities development. This meant that military institutions had the time and space to define SOF’s strategic purpose—to develop interpretations of the strategic environment—that allowed them to supply those answers to under-informed civilians when those civilians were ready to form preferences. This did not mean that such answers were the same across the military, but it did mean that by the time civilians

turned their collective gaze to SOF and its purposes at critical moments, military institutional actors—most prominently the services, the broad SOF community, and, after 9/11, the Joint

Special Operations Command (JSOC)—had spent years shaping strategic assumptions on terms they themselves had defined.

These two conditions meant that military institutional processes drove debate about

SOF’s purposes and provided most of the information available in the defense market of ideas.

Much of the clash of preferences the civil-military relations literature often expects to see between the military and civilians in fact, as Rosen (1991) suggests, happened between military institutions, in this case between special operators and the conventional force. Alternative sources of information thus tended to be supplied by sub-groups in the military taking sides in the debate. Long-inattentive civilians then chose among the alternative positions according to terms defined by military institutions. In a sense, the military conducted the debate over SOF’s strategic purpose for civilians, leaving policymakers to adjudicate among military preferences with military-supplied information and frameworks rather than truly form their own.

This case study is presented in five parts. First, I define “SOF” and explore why they provide an excellent case of an emerging capability to examine the relationship between military and civilian preferences. Next, I outline the overall history of the case, the key institutional actors, and the congruence of military and civilian preferences for the use of SOF over time.

Third, I dive deeper into the history, tracing the processes by which military and civilian preferences interact within three distinct and important case periods, bounded by key inflection points. Fourth, I explore counterfactuals to test the strength of the association between military and civilian preferences. Finally, I re-examine the case as a whole and provide conclusions about

the relationship and the mechanisms that generated the preferences for SOF obtained by the end of the case.

Why SOF? Definitions and Case Justification

What are special operations forces, and what makes them a relevant and useful case study to examine the relationship between military and civilian preferences? “SOF” is a comprehensive reference to every military service’s forces in both the active and reserve components that are organized, trained, and equipped to conduct or support special operations. Each service designates its own special operations capable units. SOF comprise the Army’s Rangers, Special

Forces, Civil Affairs, Psychological Operations, and the Special Operations Aviation Regiment

(160th SOAR). In the Navy, the Sea, Air, Land (SEAL) units and Special Warfare Combatant- craft Crewmen (SWCCs) are designated SOF. The Air Force has special operations wings and groups, while the Marines have the Raiders, organized into three battalions. Special Mission

Units organized under JSOC have personnel from all four services.

SOF have always, in an idealized, doctrinal sense, offered a range of capabilities to strategic leaders. The common elements of SOF capabilities are: limited and precise use of lethal force in extremely hostile environments that pose a high risk of casualties; secrecy and deniability; and performing typically civilian institutional functions in dangerous conflict or post-conflict environments. These capabilities span what defense practitioners call the phases of conflict from peace to war to post-conflict reconstruction. Because of their broad applications,

SOF’s expertise sometimes approaches that of the conventional forces—to such an extent that defining SOF in purely technical terms is difficult, and cultural and other contextual differences are often invoked to underscore the distinction. SOF are more independent, often termed as more

flexible and risk-tolerant than the general-purpose force (GPF) and therefore more useful in ambiguous or politically sensitive situations. SOF often do not perform tasks all that distinct from conventional forces’ skills—or, indeed, intelligence operatives’—but they perform them in distinct combinations and particular legal and political circumstances.

SOF’s fundamentally ambiguous description means that their purpose has been subject to interpretation and debate over time as external events and internal politics (both partisan and bureaucratic) changed. As the historian Mark Moyar has observed, “special operations lack a realm upon which they can lay indisputable claim. The roles and missions of special operations forces have changed frequently, based on the perceptions of political and military leaders of the tactical and strategic environments.”47

This variation in roles over time makes the SOF case study particularly fertile for the research here, as do other factors. Although U.S. SOF were first established during World War

Two and enjoyed a renaissance during the Vietnam War, by the mid-1970s their size and relevance to national defense strategy were at bare minimums. Neglect made SOF a capability that could be reinvented by military institutions and a fresh consideration for many civilian defense practitioners, if not actually new. SOF involve all four military services, the Combatant

Commands, and the Joint Staff, and the time period I have chosen spans changes in the external security environment as well and the internal domestic political environment (including

Democratic and Republican presidential administrations and control of Congress). Finally, there was measurable debate and changes in preferences as to SOF’s purpose over time, supplying the variance in the variables I need to observe relationships. Of note, civilian and military preferences about the purpose of SOF capabilities are what are being examined, but I will also

47 Mark Moyar, Oppose Any Foe: The Rise of America’s Special Operations Forces (New York, NY: Basic Books, 2017), xvii. 34

refer to SOF as a military institutional actor that itself has preferences and is involved in the preference formation processes under study.

SOF 1977-2010: Key Events, Actors, Preferences

Case Organization and Key Events

This case study spans three strategically distinct periods: the last decade of the Cold War, the post-Cold War period of the 1990s, and the U.S. global counterterrorism campaign in the

2000s. It begins with SOF’s organizational irrelevance and ends with its strategic predominance.

SOF’s missions over the entire period always included a mix of warfighting, direct action, counterinsurgency, psychological operations, and civil affairs. But its strategic emphasis changed from a kind of Swiss Army knife for discrete and non-kinetic tasks to direct action counterterrorism (and intelligence-gathering to support that primary mission).

The case does not begin with the founding of American special operations forces during

WWII, but some three decades later in the late 1970s. This choice may seem to violate the terms of case study selection—after all, SOF was not a brand-new concept in the late 70s. However, as this case will show, ideas for the strategic purpose of SOF were reinvented in the 1970s and SOF became a re-emerging capability for the U.S. military. Although SOF was active in WWII and again in the Vietnam War, by the late 1970s the entire American national security community shared an aversion to anything related to Vietnam. In this context both the military and civilians were outright rejecting the approach to U.S. military interventions used in Vietnam— counterinsurgency—that SOF represented. Thus, the beginning of a consistent through-line for ideas about how to use SOF that are still at work today really began in the late 1970s, when SOF was reborn.

The contrast between SOF’s small force structure and low operational tempo in 1977 and its size and high operational tempo in 2011 is stark. In terms of its organizational heft, SOF experienced steady growth throughout the 1980s and 1990s, and then a rapid expansion after

September 11, 2001. In 1979, the Army, whose special operations force dwarfed those of the other services, had only 3,600 personnel “assigned to active duty Special Forces units.”48 By

1996, total numbers of Army SOF reached “approximately 10,000.”49 By 2012, the U.S. Army

Special Operations Command had 28,500 personnel.50 In 1992, the USSOCOM posture statement for Congress claimed that 46,000 active and reserve personnel were special operators, and that some 2,000 of them were deployed to 20 countries on any given day. The statement requested approximately $3 billion for SOF in fiscal year 1993.51 The number of forces then dipped at the end of the 1990s, but the deployment rates increased, and the budget level held.

Then came 9/11. The Government Accountability Office found that in 2001, there were 42,800

“authorized military positions” for SOF with an annual budget of $3.1 billion and an average of

2,900 personnel deployed per week. By 2014, there were 62,800 authorized military positions (a

47% increase), a $9.8 billion budget (a 213% increase), and 7,200 personnel deployed per week

(a 148% increase).52 Given these personnel, deployment, and budgetary trend lines, it is clear

48 Susan Marquis, Unconventional Warfare: Rebuilding U.S. Special Operations (Washington, DC: Brookings Institution Press, 1997), 4. 49 Colonel Rod Lenahan, Crippled Eagle: A Historical Perspective of U.S. Special Operations, 1976-1996 (Charleston, SC: Narwhal Press, 1998), 217. 50 U.S. Special Operations Command, Factbook 2012 (MacDill Air Force Base, FL: 2012). Note the Army’s Special Operations Command is different from the U.S. Special Operations Command. 51 U.S. Special Operations Forces, 1992 Posture Statement (Washington, DC: June 1992), 5. 52 U.S. Government Accountability Office, Special Operations Forces: Opportunities Exist to Improve Transparency of Funding and Assess Potential to Lessen Some Deployments, GAO-15-571 (Washington, DC: GAO, July 2015), 14, https://www.gao.gov/assets/680/671462.pdf. There are also narratives that the chaos was somewhat intentional, given that the services prioritized playing a role in the mission over the operational coherence of the force packages. 36

that over time the national command authority found SOF more and more useful in more and more places.

To track the changing relationship between civilian and military preferences for SOF, I frame the three periods of the case around SOF’s institutional status within the military. I call these periods “survival,” “market share,” and “monopoly.” During its survival period, SOF was threatened with being eliminated as a separate institutional entity at worst, and with irrelevance to national defense strategies at best. But thanks in part to a series of advocacy activities, SOF was able to attain a share of the military market, carving out a set of missions during the second period in which it had an established role, but for which it still competed with conventional forces. Finally, SOF attained a monopoly over counterterrorism missions. The following summary of key events during the case period is divided into those three periods, as is the detailed process tracing that follows.

Survival: At the beginning of 1977, SOF was at an organizational and budgetary nadir.

Each service SOF element struggled to justify its existence in its own parochial context.53 The

Marines had eliminated their Raider SOF units. The Air Force special operators, organized around platforms that were being decommissioned or repurposed, were near extinction themselves. The Navy’s SEAL teams survived by attaching themselves to conventional fleet operations and plans.

The Army took a different approach. It also shrank its Special Forces as Vietnam wound down, but then in the mid-1970s resurrected the 1st and 2nd Ranger battalions to ensure capabilities for hostage rescue, raids, and the ability to secure American embassies.54 Then, in

1977, the Army established Special Forces Operational Detachment-Delta modeled off of the

53 Marquis, Unconventional Warfare. 54 Moyar, Oppose Any Foe, 158. 37

British Special Air Service.55 After two years of training, Delta got their first real-world mission: an attempt to rescue the hostages in the U.S. embassy in Tehran, Iran.

On November 4, 1979, Iranian students overran the embassy and took most of the

Americans—including the Marine security guards—hostage. Persistent diplomatic failures convinced the Jimmy Carter administration to review options for a military operation to rescue the hostages. What was called Operation Eagle Claw unfolded into a now-infamous disaster.56

The operation involved Delta Force, Rangers, Air Force long-range transport and gunship aircraft, and Navy helicopters piloted by Marines because Naval aviators were not trained to fly long distances overland.57 After two of the eight mission helicopters malfunctioned en-route to the first rally point inside Iran and then a third on the ground experienced more mechanical difficulties, the ground commander recommended—and the White House agreed—that the mission be called off. But during the attempt to withdraw the five helicopters and various personnel from the overcrowded airfield, one of the helicopter pilots became disoriented and crashed into one of the fuel-filled C-130 transport aircraft. The two exploded, killing 8 personnel and severely burning others. The task force abandoned the remaining helicopters and evacuated from Iran.58 The subsequent global news coverage was a major embarrassment for SOF, the

Pentagon, and the Carter administration.

A review of the failures that led to the disaster eventually resulted in the creation of a new SOF organization gathering up units from the Army, Navy, and Air Force: The Joint Special

55 Charlie A. Beckwith and Donald Knox, Delta Force (New York, NY: Harcourt Brace Jovanovich Publishers, 1983); see also: Moyar, Oppose Any Foe, 159. 56 For an excellent tactical breakdown of the plan and the numerous equipment failures that occurred during the mission’s curtailed execution, see J. Paul de B. Taillon, The Evolution of Special Forces in Counter-Terrorism: The British and American Experiences (Westport, CT: Praeger, 2001). 57 Moyar, Oppose Any Foe, 167; and Sean Naylor, Relentless Strike: The Secret History of Joint Special Operations Command (New York, NY: St. Martin’s Press, 2015), 4. 58 Moyar, Oppose Any Foe, 167. 38

Operations Command (JSOC). Also partly in response to the disaster in Iran, in 1982 the Army once again expanded its SOF organizational capacity by standing up its own Special Operations

Command.59

In the fall of 1983, these new SOF organizational units were tested in the invasion of

Grenada. A coup on the tiny island nation had launched a Marxist regime, and the Ronald

Reagan administration wanted it out. Operation Urgent Fury was a mix of conventional and special missions, conducted with poor intelligence, little reconnaissance, and poor joint communications. Despite JSOC’s achievements in SOF joint operations, the special operators were not integrated with the conventional force. JSOC was given the lead for the operation and then, toward the very end of planning, forced to cede it to the Marine Corps. Although JSOC’s approach involved launching operations at night, the Marines were accustomed to daylight operations, and so shifted the start time to dawn.60 Things deteriorated from there. Operation

Urgent Fury was a mess of poorly coordinated conventional and special operations forces.

Nineteen Americans died, thirteen of them SOF.61

Concurrent with the invasion of Grenada, spectacular incidents of terrorism and hostage- taking proliferated. The first such event was the April 1983 suicide bombing of the U.S.

Embassy in Beirut, which killed seventeen Americans, and sixty-three people in total. Some six months later, a more famous attack on the Marine barracks in Beirut killed 242 Marines and

Naval personnel.62 In June of 1985, Lebanese terrorists hijacked a Trans World Airlines plane.

The crisis lasted sixteen days, and although JSOC deployed a task force to a nearby staging area,

59 U.S. Special Operations Command (SOCOM), United States Special Operations Command History (MacDill Air Force Base, FL: 2007), 5. 60 Naylor, Relentless Strike, 24. 61 Marquis, Unconventional Warfare, 105. 62 David C. Wills, The First War on Terrorism: Counter-Terrorism Policy During the Reagan Administration (New York, NY: Rowman and Littlefield, 2003), 55, 62. 39

decision-makers never authorized it to launch a rescue operation.63 Four months later, terrorists from the Palestinian Liberation Organization hijacked a cruise ship, the Achille Lauro. Yet again,

JSOC was deployed to a staging area to conduct a hostage rescue operation; yet again, the mission was never launched and the crisis came to an end without the use of military force.64

Meanwhile, Congress had begun working on what would eventually become the

Goldwater-Nichols Department of Defense Reorganization Act of 1986. Three years prior to the passage of that law the Senate Armed Services Committee launched a comprehensive study of

DoD’s organization. Members and staffers on the House side of the Hill also turned their attention to DoD reforms, including for SOF. DoD resisted calls for change but made some attempts to acknowledge Congressional frustrations that SOF was underfunded and poorly coordinated, most notably by establishing a Joint Special Operations Agency (JSOA) in 1984, housed within the Joint Staff—an organization stillborn by a lack of operational and budgetary authority.65

Although SOF reforms were not included in the final Goldwater-Nichols legislation, an amendment to the 1987 National Defense Authorization Act a year later established Special

Operations Command (SOCOM), a new unified combatant command for SOF; a new Assistant

Secretariat for Special Operations and Low-Intensity Conflict in the Office of the Secretary of

Defense (ASD SO/LIC); and a separate budget program for SOF run by SOCOM called Major

Force Program-11 (MFP-11). JSOC remained intact but was subordinate to SOCOM.

Market Share: One of SOF’s first opportunities to demonstrate whether the cumulative reforms of the 1980s would translate into operational performance was 1989’s Operation Just

63 Naylor, Relentless Strike, 31-33. 64 Ibid., 37-38. 65 SOCOM, United States Special Operations Command History, 5. 40

Cause, the invasion of Panama. Although not without its own missteps, the operation was generally considered a success, and SOF was much better integrated with the conventional forces than had been the case in Grenada just six years prior.66 Little more than a year later, however,

SOF struggled to be included in the Gulf War. Only late in the planning stages was it assigned the “scud-hunting” task along with some reconnaissance missions.67 Psychological Operations forces also conducted campaigns to “loosen the Iraqi resolve to fight.”68

By this time, the Cold War had come to an end. Because the conventional forces’ greatest purpose—war with the Soviet Union—was now off the table, and the international security environment was characterized by a proliferation of what had been called “low intensity conflicts,” the entire military became implicated in the kinds of missions SOF had fought to specialize in. Moreover, the uncertainty of the post-Cold War world, mirrored by uncertainty in

U.S. foreign policy, made civilian officials hesitant to take the kinds of risks with military personnel that SOF often embodied.

Such risks became immediately clear to the Clinton administration during the Black

Hawk Down incident of 1993. One of the George H. W. Bush administration’s last major defense policy choices was to send a mix of special and conventional forces to support the

United Nation’s (UN) efforts in the collapsed state of Somalia. The Clinton administration then inherited Operation Restore Hope, which, by the fall of 1993, had been largely successful at responding to the famine in Somalia. But after a Pakistani peacekeeping unit suffered 24 casualties at the hands of warlord Mohamad Farah Aideed’s militants, Admiral Jonathan

Howe—occupying a strange position as the UN Secretary General’s representative but also the

66 Ibid., 44; Tom Clancy with Carl Steiner and Tony Koltz, Shadow Warriors: Inside the Special Forces (New York, NY: Berkley Books, 2002), 391-393. 67 Moyar, Oppose Any Foe, 187. 68 Clancy, Shadow Warriors, 425. 41

nominal head of the U.S. contingent—requested a Delta Force detachment in order to capture or kill Aideed.69 In October, Task Force Ranger, a mix of Delta and Rangers, launched an assault on the warlord and some of his forces. The operation ended in calamity when Somalis shot an

American Black Hawk helicopter out of the sky, precipitating a two-day urban battle that left 18

Rangers and perhaps over 1,000 Somalis dead.70 Internationally broadcast footage of the carnage included images of an American soldiers’ body being dragged through the streets of Mogadishu.

Members of Congress called for investigations into the incident, and Secretary of Defense Les

Aspin ultimately resigned due to the controversy.71

Almost simultaneously, the Clinton administration was also responding to an unfolding crisis on the island nation of Haiti. A military coup had overthrown the democratically elected government, and Clinton was threatening invasion if the junta did not return power to the ousted leader. A plan involving heavy SOF contributions “similar to what SOF had done for the invasion of Panama in 1989” was set to launch in September of 1994 when the Haitian military relented.72 The invasion was converted to a stabilization and peace enforcement mission,

Operation Uphold Democracy, and SOF’s contributions went from a planned assault force to civil affairs and psychological operations. SOF remained in Haiti until March of 1995.

The breakup of the former Yugoslavia also occasioned the use of SOF. The bulk of SOF missions involved combat search and rescue during NATO air operations, tactical air mobility, and psychological operations and civil affairs, both in Bosnia and later in Kosovo.73 JSOC’s most notable mission during the second term of the Clinton administration was a series of

69 Richard Sale, Clinton’s Secret Wars: The Evolution of a Commander in Chief (New York, NY: St. Martin’s Press, 2009), 82. 70 SOCOM, United States Special Operations Command History, 56-59. 71 David Halberstam, War in a Time of Peace: Bush, Clinton, and the Generals (New York, NY: Simon and Schuster, 2001). 72 SOCOM, United States Special Operations Command History, 61. 73 Ibid., 64-69. 42

“snatch-and-grab” operations to apprehend suspected war criminals after the wars in the Balkan states.74 Meanwhile, global terrorism reemerged as a feature of the international security environment thanks to a budding organization called al Qaeda. In 1993, a truck bomb in the

World Trade Center in New York killed six people. In 1996, nineteen Americans died in a bombing at the Khobar Towers in Saudi Arabia. In 1998, simultaneous bombings at the

American embassies in Nairobi, Kenya and Dar es Salaam, Tanzania killed 213 people and injured more than 4,000.75 Despite debating SOF-led counterterrorism operations against senior al Qaeda leaders, including Osama bin Laden himself, SOF was never deployed on such a mission, and instead the Clinton administration conducted air strikes on suspected al Qaeda locations in Sudan and Afghanistan.

Monopoly: SOF was thus still focused on apprehending war criminals in the Balkans when the attacks of 9/11 occurred. On September 11, 2001, nineteen hijackers flew commercial airliners into both World Trade Center towers in New York City, the Pentagon in Washington,

DC, and failed to hit a third target when the passengers on a fourth plane overtook the hijackers and crashed it into a field in Pennsylvania. Tracing the attack to al Qaeda, and knowing that

Afghanistan harbored bin Laden, the Bush administration demanded that the ruling Taliban regime turn bin Laden over to the United States. The Taliban refused, and within a month, the

U.S. invaded Afghanistan with a vanguard force of SOF (particularly JSOC elements) and CIA paramilitary forces operating alongside Afghan insurgents. Conventional forces followed to prevent the Taliban from recapturing the government and to provide security for the population

74 Naylor, Relentless Strike. 75 “Timeline: Al Qaeda’s Global Context,” PBS Frontline, October 3, 2002, https://www.pbs.org/wgbh/frontline/article/timeline-al-qaedas-global-context/. 43

and political processes. SOF remained in Afghanistan to hunt Bin Laden and other members of al

Qaeda, and to conduct strikes against the Taliban.

Just eighteen months after the Afghanistan invasion, claiming that Saddam Hussein had links to al Qaeda and was hiding weapons of mass destruction, the Bush administration invaded

Iraq, with SOF heavily involved in the invasion and efforts to defeat the subsequent insurgency.

SOF had long conducted small-unit deployments across the world to conduct everything from training to intelligence-gathering missions called “advanced force operations.” Now, they focused these dispersed deployments on counterterrorism. In 2004, Secretary of Defense

Rumsfeld designated SOCOM as the lead DoD organization (or “global synchronizer”) for the

Global War on Terror.76 SOF operations in Iraq also persisted for the entire duration of the U.S. involvement in the war—until the end of 2011.

Six months before the U.S. withdrawal from Iraq, SOF’s—and really, JSOC’s—most spectacular publicly-acknowledged operation was the killing of Osama bin Laden in Abbottabad,

Pakistan. The operation featured many of the hallmarks of the post-9/11 SOF capability: it was an ultra-precise manhunt, clandestine, diplomatically sensitive, and high risk in both political and operational terms. It also presented a glimpse at the dramatic transformation in civilians’ views of SOF’s purpose and role in national strategy since the late 1970s. SOF had gone from an after- thought to the premier strategic asset for the nation’s number one national security priority: defeating terrorism.

76 Douglas C. Lovelace, ed. Terrorism: Documents of International and Local Control (Oxford, UK: Oxford University Press, April 2008), vol. 86, no. 176; and U.S. Special Operations Command, 2009 Factbook (MacDill Air Force Base, FL: 2009), 6. 44

Actors

The most prominent institutional actors over the case period were not “the military” and

“civilians,” but instead sub-organizations of those categories. On the military side, the two major players were the conventional forces—represented most frequently in the civil-military dynamic by the Chairman of the Joint Chiefs of Staff, but also by the Service Chiefs—and the special operations community, its own locus of institutional sentiment moving over time between the

Army’s Special Forces, JSOC, and SOCOM. At times, this community also included what other analysts would code as civilians: retired former special operators now in civilian positions. These civilians, largely in the Office of the Secretary of Defense (OSD), a civilian organization within the DoD, but also some on Capitol Hill, continued to share the institutional culture and interests of SOF. In general, institutional preferences were complicated by the tendency of individuals from one sub-group to share preferences with members of other sub-groups—for example, as will be demonstrated below, conventional Army leaders provided much of the early support for developing SOF CT capabilities. Such conditions also obtained when an individual spanned institutional sub-groups, such as being in the Army as a Special Forces officer.77

On the civilian side, three distinct sub-groups stand out as salient actors. First are the civilians in the Office of the Secretary of Defense. This group comprises both political appointees and civil servants (although the former appear more prominently in the history presented below) who not only exercise an element of civilian control over the military, but also represent DoD in the interagency, on the Hill, and to the public. Second are civilians in the White

House, most importantly the President and his National Security Advisor and other staff at the

National Security Council. Third are civilians in the Congress, both the members and their staff.

77 The important thing to attach to key actors is a core preference and to identify its source as a military institution or not. 45

As on the military side, sub-groups of civilians often collaborated with other sub-groups in addition to competing with them.

Preferences

At the beginning of the case study, there were two measurable preferences about what the country should do with SOF capabilities, and they were both held by military institutional actors.

The first preference was for SOF to support the conventional military in its efforts to counter the

Soviet Union on the battlefield. This preference advocated integrating special operators into conventional operational plans. It necessarily excluded ideas of SOF-led operations or modes of warfare that focused on non-state (or irregular state-sponsored) enemy combatants such as counterinsurgency or counterterrorism.

The second preference was for SOF to emphasize missions that occurred outside the context of major war and against non-state or irregular state-sponsored (including Soviet- sponsored) enemies. Operations against Islamist terrorist groups fell under this preference. The former preference emphasized SOF as an adjunct to the conventional force, whereas the latter saw SOF as an element of the military that should operate independently from the larger force outside the context of regular warfare. For brevity’s sake, I will refer to the first preference as

“conventional” uses of SOF and the second as unconventional or irregular uses of SOF.

Conventional uses of SOF were, unsurprisingly, preferred by the conventional military.

But in 1977 the Navy’s SEAL forces and the Air Force’s special operators, elements of SOF, went along with this preference. The SEALs focused on providing reconnaissance support to amphibious operations and working themselves into the Navy’s major operational plans.78 Only

78 Marquis, Unconventional Warfare, 65. The Marine Corps had eliminated its special operator designation, having done away with their Raider units after WWII. Moyar, Oppose Any Foe, xvi, 56, 62. 46

the Army truly held a preference for non-conventional uses of SOF. As will be discussed in more detail below, the Army began to build units focused on counterterrorism in the mid- to late-

1970s. The SEALs, however, also husbanded their counterterrorism capabilities in at least one of their teams. For the Army, interest in SOF CT capacity started at the top, with successive Chiefs of Staff, while the SEALs were circumspect about it as the Navy leadership focused elsewhere.

Thus, the period began not so much with a split between civilians and the military on the purpose of SOF, but a split between elements of the military itself, including between parts of SOF.

At the beginning of the case period civilians showed so little interest in SOF in an institutional sense that a reliable preference is nearly undetectable. The absence of detectable civilian preferences in the late 1970s suggests one of two possibilities: 1) Civilians had not formed a strong or consolidated preference for the uses of SOF or 2) Civilian preferences were reflected in the status quo at the beginning of the case period. Although President Jimmy Carter did express interest in hostage rescue capabilities at one point, and his Secretary of Defense,

Harold Brown, appeared supportive of some SOF institutional reform, the administration and the

Congress showed little consistent interest in re-purposing or even revisiting SOF capabilities in the late 1970s.79 Moreover, the Carter administration’s defense policy focused on reducing tensions with the Soviets and the Chinese, developing sophisticated defense technologies, and limiting the use of military force—overarching preferences that may have precluded entertaining a capability for using discrete force in contexts outside major war.80 Brown himself has written that his experience as Secretary of the Air Force toward the end of the Vietnam War led him to a

79 David C. Martin and John Walcott, Best Laid Plans: The Inside Story of America’s War Against Terrorism (New York, NY: Harper and Row, 1988), 39; Richard Marcinko, Rogue Warrior (New York, NY: Pocket Books, 1992), 236; and Naylor, Relentless Strike, 5-7. 80 Harold Brown with Joyce Winslow, Star Spangled Security: Applying Lessons Learned Over Six Decades Safeguarding America (Washington, DC: Brookings Institution Press, 2012). 47

skepticism of military engagements “beyond our vital security interests” that he brought into the position of Secretary of Defense in 1977.81 In that, he sounds very much like his successor,

Caspar Weinberger, whose well-known “doctrine” of a high threshold for use of military force was eventually adopted by Chairman of the Joint Chiefs of Staff General Colin Powell and influences thinking among military officers and the permanent DoD bureaucracy to this day.82

The connection between civilian leaders’ overall thoughts about low-level warfare and SOF is an inference, however, and the paucity of civilian pronouncements about SOF at the beginning of the case period may reveal that civilians did not think of SOF as a strategic capability, nor as worthy of their time and attention as, for example, ballistic missiles or stealth technologies.

Although a strong civilian preference for the main uses of SOF was not evident in 1977, a preference had clearly emerged by 2010: civilians saw SOF’s main role as countering terrorism and conducting irregular warfare, primarily by eliminating so-called high value targets (HVTs) and training foreign counterterrorism forces. This preference featured many of the elements of the preference that the Army and Navy SOF first formed in the mid-1970s: It focused SOF on non-state (or irregular state-sponsored) enemy combatants; SOF was focused on both lethal and more civilian-oriented activities outside the bounds of major conventional war; and SOF had the leading role in the mission, often operating autonomously from the conventional force. Not only did civilians see SOF this way, but SOF had long held the same preference, expressing the primacy of counterterrorism missions in the official Special Operations Command (SOCOM) mission statements and focusing the full range of subordinate special operations elements on

81 Ibid., 99. 82 Jeffrey Record, “Back to the Weinberger-Powell Doctrine?,” Strategic Studies Quarterly 1, no. 1 (Fall 2007): 79- 95, www.jstor.org/stable/26268385. 48

dealing with terrorists, from special mission (i.e., direct action) units to civil affairs and psychological operations teams.

Civilians’ adoption of the preferences to use SOF as a counterterrorism force was not a linear process, however. Within each period in the case, especially the second period, civilian preferences wavered between conventional and unconventional uses of SOF. Many presidents entered office thinking of SOF as a kind of supporting force to conventional operations but by the ends of their terms in office had adopted the preference of using SOF for counterterrorism and/or unconventional warfare. A president’s successor might then repeat the journey. As the evidence presented below will show, it was the SOF community’s persistent advocacy for the use of SOF for counterterrorism and low-intensity conflict that caused civilian leaders to adopt the same preference for the use of SOF repeatedly.

SOF 1977-2010: Tracing Processes of Preference Formation and Change

Period I: SOF’s Survival (1977-1987)

By 1977, most of the institutional military were aligned in the preference that SOF was a niche capability, useful for special missions in a major combat framework. Neither civilian opposition to nor support for this formulation was evident. The Army, however, had a broader set of preferences for the purposes of their own special operators, supplying them for major operational plans while also husbanding SOF elements for unconventional and irregular warfare and, notably, counterterrorism.

By 1987, the end of the period, an important fraction of civilians had shifted their expressed preference to align more closely with the Army’s original position: SOF was still designed to support conventional operations, but many civilians now also recognized that SOF

was the nation’s specialists in the low-intensity conflict and crisis-related, direct action counterterrorism.83

The shift in civilian preferences was a result of persistent lobbying by the special operations constituency in and outside of government. In the Pentagon, on the Hill, and even in a presidential campaign, current and former SOF persuaded key civilian policymakers, politicians, and lawmakers that SOF was strategically important and institutionally neglected. Over the course of ten years, conceptions of SOF and its purposes, born primarily in the Army Special

Forces community, spread into OSD and the Congress.

SOF leveraged a variety of conditions to make their case to civilians. External events and

U.S. operational performance provided focal points to demonstrate lags in U.S. capabilities, while general ignorance of special operations ceded definitional space to uniformed experts.

Civilians tended toward undefined preferences about SOF and its uses, making them open to the most persuasive argumentation and incentives. The result was that different military communities appealed to different civilians, with those in OSD aligning with the conventional force and those in Congress adopting SOF views.

Details

The 1977 civilian preference for SOF was so weak that it is difficult to find evidence of what it was. SOF received very few mentions in official strategic and budgetary documents, public statements, and even memoirs covering the late 1970s. Secretary of Defense Harold

83 “Direct Action” is defined as “short-duration strikes and other small-scale offensive actions conducted with specialized military capabilities to seize, destroy, capture, exploit, recover, or damage designated targets in hostile, denied, or diplomatically and/or political sensitive environments.” Joint Chiefs of Staff, Joint Publication 3-05: Special Operations (Washington, DC: Department of Defense, July 2014), https://www.jcs.mil/Portals/36/Documents/Doctrine/pubs/jp3_05.pdf. 50

Brown’s first Annual Report to Congress, submitted in February of 1978, did not address or even use the words “special operations” or “special operations forces.”84 What civilian preferences can be detected were derived from the majority preference among the military services and the

Joint Staff. And among military groups, SOF was largely viewed as accessories to the warfighter. In the abstract, its services could be used when general purpose forces (GPF) could not, but such scenarios were underspecified. The Joint Chiefs of Staff’s 1979 definition of special operations, for example, noted that “they may support conventional operations, or they may be prosecuted independently when the use of conventional forces is either inappropriate or infeasible.”85 In any case, SOF had no purpose absent a conventional warfighting context. At a

1983 conference on the future of special operations among defense practitioners and academics,

Secretary of the Army John O. Marsh, Jr. (who became the acting Assistant Secretary of Defense for Special Operations and Low Intensity Conflict in 1987) argued that “a special operations capability is an essential adjunct to an effective conventional force structure.”86

This view even extended to parts of the SOF community. One former SEAL explained the teams’ “job in life was to justify our existence” to the conventional fleet with support to anti-

Soviet operations.87 The Air Force, meanwhile, neglected its special operations aircraft nearly to the point of abandonment, focusing almost exclusively on conventional missions.88

The uniformed Army, however, saw more for SOF than conventional missions or vaguely-defined “things the GPF couldn’t do.”89 The Army believed SOF could also provide

84 Department of Defense, Annual Report to the Congress, FY1978 (Washington, DC: January 1977). 85 Joint Chiefs of Staff, Joint Publication 1, Doctrine for the Armed Forces of the United States (Washington, DC, Department of Defense, 1979), https://www.jcs.mil/Portals/36/Documents/Doctrine/pubs/jp1_ch1.pdf. 86 Frank R. Barnett et al., eds., Special Operations in U.S. Strategy (Washington, DC: National Defense University Press, 1984), 18. 87 Marquis, Unconventional Warfare, 66. 88 Ibid., 76-77. 89 Steven Emerson, Secret Warriors: Inside the Covert Military Operations of the Reagan Era (New York, NY: G.P. Putnam’s Sons, 1988), 38-39. 51

capabilities below the threshold of major war, not only in traditional training and advising to partner military forces, but also to counter state-based and non-state actor terrorism. This preference was organic to the Army itself and arose in at least three different places during the

1970s. In the early part of the decade, Chief of Staff General Creighton Abrams observed the rise in Islamist extremist attacks as a future defining security threat for the United States. He determined that the Army would therefore need a counterterrorism capability: hostage rescue, raids, and the ability to secure American embassies. To develop these capabilities, he resurrected the 1st and 2nd Ranger battalions, a separate entity from the Army’s Special Forces.90

Meanwhile, General Edward Charles “Shy” Meyer, “had long believed that geopolitical pressures from the Muslim movements in the Middle East were a far more likely concern than events in Central Europe.”91 But he also knew the insight was not obvious to others. “It was necessary for us to try to create a requirement,” Meyer told an interviewer in 1988. “It became clear that there was a void that needed to be filled on the low end of the spectrum of warfare.”92

Meyer’s identification of political Islam as a source of conflict that would involve the United

States was novel at the time, and his comment that the Army had to “create a requirement” referred to a campaign of persuasion the Army would have to undertake internally and externally to generate a critical mass of powerful adherents to Meyer’s ideas.

General William DePuy, the Commander of the Army’s Training and Doctrine

Command, already shared Meyer’s assessment as he reimagined the Army’s purpose in the post-

Vietnam era. DePuy and Meyer worked with Special Forces Colonel Charlie Beckwith to develop a counterterrorism-focused element, one with the additional goal of husbanding a near-

90 Moyar, Oppose Any Foe, 158. 91 Marquis, Unconventional Warfare, 62. 92 Quoted in ibid., 63. 52

monopoly on clandestine operations. Beckwith himself had been advocating for such a unit within the Army since the 1960s—and even went so far as to mail a copy of his proposal to his senator, who placed phone calls to the Army to inquire about “the void” Beckwith had “pointed out in Special Forces.”93 It was not until 1975 that another Army officer, General Bob Kingston, asked Beckwith for a version of his earlier paper to shop around the Pentagon, in particular with

General Meyer. By 1976, Beckwith was at a conference with General DuPuy discussing the merits of his proposal, and by 1977 DuPuy took it to Meyer.94 As a result, the Army established

Special Forces Operational Detachment-Delta in November of 1977.95

SOF’s purpose(s) were thus evolving when the Iran hostage crisis unfolded in 1979.

Anticipating a request from the White House for hostage rescue options, the Joint Staff directed

Army Major General James Vaught and Beckwith to develop a plan. Vaught and Beckwith were then summoned to the Situation Room to brief the President and his national security team on the

SOF community’s proposal, which the President approved.96

Delta being one of the only places in the military that trained extensively for hostage rescue, they were an obvious choice for the mission. It was a case where the need for a capability was demanded by actual events, rather than proactive U.S. strategy. Under such conditions, as we shall see again, civilians reach for whatever available tool seems most applicable. This dynamic creates an incentive for military institutions to demonstrate relevance in future such contingencies, and thus to try to anticipate likely national security events and civilian demand signals.

93 Beckwith and Knox, Delta Force, 44. 94 Ibid., 92-95. John Shalikashvili, President Clinton’s future Chairman of the Joint Chiefs of Staff, was also in attendance at the conference, the topic of which was the future of the light infantry. 95 Moyar, Oppose Any Foe, 159. 96 Ibid., 163-165. 53

After Operation Eagle Claw unraveled, the Joint Staff considered the meaning of the failure. The first and best-known review of the operation was the Special Operations Review

Group Rescue Mission Report—better known as the Holloway Report after the Admiral who chaired it. Populated entirely by active duty and retired general and flag officers, the commissioners noted issues with command and control, equipment, and intelligence support.97

The various service SOF personnel were too unfamiliar with each other and their assets to plan effectively or manage the aircraft malfunctions. The Holloway Report recommended a higher- level joint command element be formed to gather up the disparate SOF units—JSOC. JSOC’s main purpose was “to plan and conduct military operations that would counter terrorist acts,” and would gather the SOF of each service under a unified command.98

That JSOC was a response to the failures of Eagle Claw and that its primary mission was counterterrorism may seem incongruous to the modern reader. At that juncture however, terrorism was considered a tool that state competitors used to menace American citizens abroad in order to coerce political concessions from the U.S. government. Taking Americans hostage via proxies was a common terrorist tactic. The military’s role in counterterrorism was narrowly focused on this type of immediate crisis-mitigation scenario.

As JSOC got organized, the Army continued to play a major role in counterterrorism. Not only was Delta Force assigned to JSOC, but the task force’s first commander, Brigadier General

Dick Scholtes, was an Army officer, and the command was headquartered at Fort Bragg, an

Army installation. At the same time, General Meyer had become the Army Chief of Staff and was presiding over a major restructuring of Army SOF, modernizing it and orienting it toward

97 Special Operations Review Group, Rescue Mission Report (Washington, DC: Joint Chiefs of Staff, 1980), https://www.history.navy.mil/research/library/online-reading-room/title-list-alphabetically/i/iran-hostage-rescue- mission-report.html#recom. 98 Marcinko, Rogue Warrior, 236. 54

the unconventional and irregular warfare. Writing in the Army Green Book in 1980, Meyer asserted “low-risk, high-leverage ventures, such as activities on the lower end of the spectrum, are the most likely military challenges to occur” and the Army therefore needed “forces equally comfortable with all the lesser shades of conflict.”99

It was also at this point that the Navy’s contribution to joint SOF missions began to expand, albeit idiosyncratically. The Navy as an institution did not take any particular initiative toward JSOC, but Commander Richard Marcinko, working on the Navy staff at the Pentagon when JSOC was being designed, saw the opportunity to create a brand-new SEAL command

“specifically trained to fight terrorists in a maritime environment.”100 Up to this point, the Navy had maintained counterterrorism training curriculum for a few platoons in each of its SEAL teams. In particular, Team Two, on the East Coast, trained with the counterterrorism forces of a variety of European allies, including Britain’s renowned Special Boat Service.101 In creating the

Navy Special Warfare Development Group, or SEAL Team Six, Marcinko substantially increased the Navy’s role in direct action SOF missions. And he was its first commanding officer.

Thus, by 1981, SOF was gathering momentum toward redefining its purpose. With JSOC and the Army’s SOF restructuring, more of the conventional force saw SOF as a hostage rescue, security force assistance, counterinsurgency, and counterterrorism force along with its utility during major warfighting contingencies.

Ongoing support for these institutional changes, however, needed to rest on a broader foundation. To persuade others that such capabilities were necessary, Army leaders started

99 Quoted in Clancy, Shadow Warriors, 216. 100 Marcinko, Rogue Warrior, 238. 101 Ibid., 237-238. 55

talking more frequently about low intensity conflict (LIC). LIC comprised both state-based and non-state-actor competition and security threats, including many of Special Forces traditional proxy training missions along with the renewed counterterrorism capabilities in Delta Force. It was a way to specify when, exactly, and under what conditions, “the use of conventional forces is either inappropriate or infeasible.” In 1985 the Army began examining the challenges of low- intensity conflict. The Low-Intensity Conflict Project, a civil-military group chaired by general and flag officers, released a two-volume report in August of 1986 that declared that the U.S. was fundamentally unprepared for irregular and unconventional forms of security competition. It also observed the focus on terrorism took away from efforts to develop counterinsurgency and stabilization capabilities. Critically, the project judged that preventive action was called for and that SOF should play such a shaping role. About SOF, Volume I of the report stated: “Special operations forces must develop appropriate strategies and acquire resources to assist or intervene prior to or during the early stages of a conflict.”102

Civilian leaders picked up on the marriage of SOF and LIC, even as they hung onto the earlier preference for SOF to support conventional war. Combining both the JCS’ 1979 language and Meyer’s arguments in the Army Green Book, the 1984 DoD Annual Report to Congress stated that SOF’s purpose was “to meet threats at the lower end of the conflict spectrum—where the use of conventional forces may be premature, inappropriate, or politically infeasible.”103

Within the military, however, arguing for the importance of LIC and capabilities to conduct it was an uphill battle. Despite the recognition that the tactical failures during Eagle

102 U.S. Army Training and Doctrine Command, Joint Low-Intensity Conflict Project Final Report, Volume I: Analytical Review of Low-Intensity Conflict (Fort Monroe, VA, 1986), 8-5, http://www.dtic.mil/dtic/tr/fulltext/u2/a185971.pdf. 103 U.S. Department of Defense, Annual Report to the Congress, FY1984 (Washington, DC: February 1983), https://history.defense.gov/Portals/70/Documents/annual_reports/1984_DoD_AR.pdf?ver=2014-06-24-151113-310. 56

Claw suggested some improvements to interoperability, the prevalent sentiment among the conventional force was schadenfreude. To them, Eagle Claw had demonstrated the unfavorable risk-reward ratio of special operations.104 And although the Iran Hostage crisis raised the profile of terrorism generally, there was still no broad-based military nor civilian appetite for military solutions to what was seen as a criminal and political problem, or an aspect of state-based competition best met with conventional force.105

President Ronald Reagan thus inherited a Pentagon still divided, with a preponderance of opponents to any greater investment in irregular capabilities but a growing cadre of proponents for the same along with a nascent concept marrying SOF and LIC. The fresh civilian leadership became the targets of such argumentation.

As had been generally true of civilian political leaders in the Carter administration,

Reagan and his defense advisors betrayed few strong preferences about SOF or the types of missions it could fulfill. Although Reagan made two major national security speeches during the general campaign, the themes were broad: strength, weakness, peace, aggression. Within that broad rubric were windows of interest in the types of irregular conflict on which the bulk of the

U.S. force did not focus. In a speech before the 1980 election, he outlined a nine-point plan to strengthen the U.S. and her allies abroad which included “a plan to assist African and Third

World development” and “combatting international terrorism.”106 He presented terrorism as a state-centric phenomenon, singling out Soviet and Iranian sponsorship of terrorists, and highlighting the victimization of Israel at the hands of terrorist groups. Reagan’s position on the

104 Moyar, Oppose Any Foe, 171. 105 Richard H. Shultz, Jr., “Showstoppers: Nine Reasons Why We Never Sent Our Special Operations Forces after al Qaeda before 9/11,” Weekly Standard, January 26, 2004, https://www.washingtonexaminer.com/weekly- standard/showstoppers. 106 Ronald W. Reagan, “A Strategy for Peace in the 80s,” (Televised Address, Washington, DC, October 19, 1980), https://www.reaganlibrary.gov/10-19-80. 57

ongoing Iran hostage crisis was that the U.S. should not negotiate with those sponsoring terrorism nor reward terrorism in any way.107

Overall, Reagan believed that compounding the appearance of weakness was actual military vulnerability: that the armed forces had been gutted by neglect. His nine-point plan included, “a determined effort to strengthen the quality of our armed services.” While never taking direct responsibility for Reagan’s campaign-era perspective on defense issues, Caspar

Weinberger, Reagan’s Secretary of Defense from 1981 to 1987, certainly shared his views. In his memoirs, Weinberger stated that he and Reagan had discussed perceptions of American weakness and the need for reinvestment in military strength, as well as his immediate efforts to provide “substantial increases” for conventional and nuclear forces.108 Neither man left meaningful evidence, however, that they reflected substantially on either special operations or irregular warfare.

Yet SOF partisans were present in the Reagan campaign and early administration policy, and there is some evidence that their influence had an impact on the administration’s attitudes toward SOF capabilities. After the former Director of Central Intelligence George Bush signed on as Reagan’s running mate, a collection of retired military officers and Central Intelligence

Agency (CIA) officials joined the campaign, a development that turned out to be a backdoor for

SOF advocates into the Reagan camp.109 Among them was retired General Richard Stilwell.

Stillwell had been a career Army officers, much of his time in Special Forces, and also had a

107 Jonathan Moore, ed., The Campaign for President: 1980 in Retrospect (Cambridge, MA: Ballinger Publishing Co., 1981), 34. 108 Caspar Weinberger, Fighting for Peace: Seven Critical Years in the Pentagon (New York, NY: Warner Books, 1990), 15, 57. 109 Bill Peterson, “Coming in from the Cold, Going out to the Bush Campaign,” Washington Post, March 1, 1980, A2, https://www.washingtonpost.com/archive/politics/1980/03/01/coming-in-from-the-cold-going-out-to-the-bush- campaign/3758ff60-0d13-43a6-9a6f-c692e20d5378/. 58

background with the CIA’s covert operations.110 He helped encourage Reagan to campaign on

“revitalization” of SOF as part of the overall reinvestment in the military that Reagan had made a campaign plank.

Once the Reagan administration took office, Stillwell became the Deputy Under

Secretary of Defense for Policy at the Pentagon, where he continued to press for an emphasis on

SOF. Reagan and Weinberger’s first articulation of defense strategy and policy, the 1981

Defense Guidance (the full text of which has never been released), declared, “We must revitalize and enhance special operations forces to project United States power where the use of conventional forces would be premature, inappropriate, or infeasible.”111 According to Reagan’s

Secretary of the Army, Stillwell had personally seen to it that SOF revitalization was included in the document.112 Thus, Reagan and Weinberger’s first major expression of preference about SOF was one Stillwell, a retired Army Special Forces officer, persuaded them to adopt.

The strength of the civilian preference to “revitalize” SOF was not robust, however, as would be demonstrated in the relative modesty of SOF’s subsequent budget increases despite overall leaps in broader defense spending. Nor did it actually lay out a purpose for SOF beyond what the Joint Staff had already articulated — Meyers was still working to sell the relationship between SOF and LIC. One later observer noted that, “While special operations forces have clearly received increased priority during the Reagan Administration, with a budget increase

110 Reflecting his importance during the campaign, Stillwell later became DUSD for Policy (today the PDUSDP position) at the Pentagon. “General Richard G. Stillwell Dies,” Washington Post, December 26, 1991, https://www.washingtonpost.com/archive/local/1991/12/26/gen-richard-g-stilwell-dies/9187423d-d6d8-417f-aaa7- c6da20b8630e/?noredirect=on&utm_term=.8d570dfcef0d. 111 Richard Halloran, “Military is Quietly Rebuilding its Special Operations Forces,” New York Times, July 19, 1982, A1, https://www.nytimes.com/1982/07/19/us/military-is-quietly-rebuilding-its-special-operations-forces.html. 112 John O. Marsh, “Keynote Address,” in Special Operations in U.S. Strategy, ed. Frank R. Barnett et al. (Washington, DC: National Defense University Press, 1984); and Caryle Murphy and Charles R. Babcock, “Army’s Covert Role Scrutinized,” Washington Post, Nov 29, 1985, A1, https://www.washingtonpost.com/archive/politics/1985/11/29/armys-covert-role-scrutinized/5d3ba408-3469-4742- 9f79-515ad97bd26e/. 59

from $441 million in 1981 to $1.6 billion in 1987, much of this increased spending has gone toward deep-penetration airlift, which has a primary focus of supporting SOF in a conventional war.”113 Nevertheless, the Defense Guidance was an important incremental step toward an important role for SOF in national strategy. For SOF to have a value unique enough to deserve strategic and budgetary emphasis was a new thing for OSD to claim, and a step toward civilians thinking more deeply about SOF as a capability in and of itself.

Nevertheless, the language in the guidance reflected not just SOF preferences to survive but also the alternative preference held jointly by the military services and most civilian defense managers for SOF to be an “adjunct” to the GPF. His 1981 guidance document notwithstanding,

Secretary Weinberger combined conservative views on the use of force generally (and for CT purposes specifically) with a pattern of deference to the military services and the Chairman of the Joint Chiefs, with whom he met almost every day.114 Weinberger wrote in 1985 that “the world consists of an endless succession of hot spots . . . [But] the belief that the mere presence of

U.S. troops . . . could be useful in some way is not sufficient for our government to ask our troops to risk their lives.”115 Apparently, “when Weinberger had to settle conflict between political appointees from the Office of the Secretary of Defense and the JCS, the joint chiefs almost always won.”116 Weinberger himself explained, “my purpose was to ensure centralized control of policy formation direction, but to move toward a more decentralized execution of the policies.”117 For the Secretary to be skeptical of low thresholds for personnel deployment and simultaneously deferential to the military chiefs was a double-whammy for the minority voices

113 Henry L.T. Koren, Jr., “Congress Wades into Special Operations,” Parameters 122 (December 1988), https://apps.dtic.mil/dtic/tr/fulltext/u2/a528146.pdf. 114 Wills, The First War on Terrorism, 31. 115 Caspar Weinberger, “U.S. Defense Strategy” in The Reagan Foreign Policy, ed. William G. Hyland (New York: New American Library, 1987), 194. 116 Wills, The First War on Terrorism, 31. 117 Weinberger, “U.S. Defense Strategy,” 43, emphasis original. 60

advocating SOF capabilities to serve LIC mission needs. Secretary Weinberger’s implacable stance made the Reagan SOF revitalization less revolutionary and more evolutionary. In fact, the opposition between the Defense Secretary and the Chiefs on one hand and SOF and its civilian allies on the other was a pattern during this and the later periods of this case.118

Where were those civilian allies for SOF? No one personified the SOF-oriented civilian cadre quite so well as Noel Koch, the Reagan administration’s Principal Deputy Assistant

Secretary of Defense (PDASD) for International Security Affairs (ISA) from 1981 to 1986. Koch arrived in OSD having served as an enlisted soldier in Vietnam, but without any particular SOF- related agenda. Yet his Military Assistant, Colonel George McGovern, did. McGovern was the former commander of the Army’s 5th Special Forces Group, and according to a historian who interviewed Koch, “McGovern’s mission became convincing Koch that rebuilding the nation’s special operations capability was a fight worth fighting.”119 McGovern joined forces with a civilian in ISA, Lynn Rylander, and together they prevailed upon Koch to lead OSD’s efforts at

SOF revitalization.

In histories and contemporary news pieces, Koch comes across as a zealous convert to the SOF religion, and a clever bureaucratic tactician.120 Realizing his civilian status could put him at a disadvantage when making arguments about needed military capabilities, especially against conventional officers predisposed to be skeptical of SOF, one of the first things he did was assemble a “Special Operations Policy Advisory Group”, or SOPAG, made up of retired

118 Wills, The First War on Terrorism, 32-40. One civilian interested in more robust and possibly even military responses to terrorism was Secretary of State George Shultz. See his “New Realities and New Ways of Thinking” article in Hyland, 111. Shultz wrote about terrorism: “even as the world becomes more secure from the danger of major war, paradoxically the democratic world now faces an increasing threat from this new form of warfare . . . much of it supported or encouraged by a handful of ruthless governments.” 119 Marquis, Unconventional Warfare, 80. 120 By my own standards, counting Koch as a civilian is fraught, since he was in SOF/intel during Vietnam. He was, however, part of the civilian presidential administration and occupying a civilian appointment in OSD, placing him squarely in a civilian institution, something he acknowledged in creating the SOPAG. 61

general and flag officers.121 Indeed, “by early 1982 Koch was sponsoring the first briefings to

Secretary Weinberger and the Defense Resources Board on special operations.”122 He also ensured his chain of command included as few people as possible, orchestrating one circumnavigation of the bureaucracy after another. In addition to the SOPAG, Koch also established a Special Planning Directorate that reported directly to the Under Secretary for

Policy and the Deputy Secretary, allowing him to skip the coordination process among DoD organizations normally required before senior OSD civilian review policies for approval.123 Such

“short-circuits” of the Pentagon allowed Koch to do things such as take a draft directive to the military services to expand their special operations forces and submit master plans for doing so straight to Deputy Secretary of Defense Thayer—who signed it.124

Koch’s guerilla bureaucratic warfare might have come to nothing, however, if it weren’t for the invasion of Grenada in 1983. It was a moment of reckoning for SOF’s strategic purpose.

The invasion itself was entirely consistent with the view that SOF existed to complement conventional operations. Although the invasion succeeded in its operational objectives, and

Secretary Weinberger later mounted a vigorous defense of the necessarily condensed planning process, to special operators it was another study in how not to conduct an operation and to a wider audience became further evidence of the need for broad military organizational reforms.125

Such a review was taking place on Capitol Hill. Although Eagle Claw had gotten many lawmakers’ momentary attention, the needless casualties in Grenada prompted the Congress to take action on SOF reforms. But even though events provided civilian legislators the motivation

121 Ibid., 82. 122 Ibid., 81. 123 Linda Robinson, et al., Improving the Understanding of Special Operations (Santa Monica, California: RAND Corporation, 2018), https://www.rand.org/pubs/research_reports/RR2026.html. 124 Marquis, Unconventional Warfare, 85. 125 Weinberger, Fighting for Peace, 125. 62

for reform, it was active and retired members of the so-called “SOF Liberation Front” who shaped the legislative process and the legislation itself.126 In his own history of the passage of the

Nunn-Cohen amendment, Jim Locher, a leading SASC staff architect of the law—and a later

Assistant Secretary of Defense for Special Operations and Low Intensity Conflict or

ASD(SOLIC)—stated plainly that “current and former special operators and their civilian supporters” were pivotal to negotiations on the Hill.127 Representative Dan Daniel (D-VA),

Senator William Cohen (R-ME), and Senator Sam Nunn (D-GA) were each approached by SOF advocates. Daniel was the chair of the readiness subcommittee of the House Armed Services

Committee. In his history of the legislation that eventually forced Special Operations Command on DoD, then-Colonel William Boykin wrote, “Senator Cohen was approached by a number of credible former special operations people with requests for his assistance in helping to rebuild

SOF.”128 As for Congressman Daniel, Boykin claims he “knew little of special operations, but he was influenced by people who did.”129 The list of influencers included Brigadier General Richard

Scholtes, who had been infuriated by his experience commanding SOF during Operation Urgent

Fury in Grenada, and who had been advocating for a unified combatant command for SOF.130

Nunn and Cohen were members of the Senate Armed Services Committee and already at work on what would later become the Goldwater-Nichols Defense Reorganization Act. Although the thrust of that legislation was not oriented toward SOF issues for political reasons (senators fretted that adding one more thing the DoD hated to the legislation would degrade the law’s chances of faithful implementation), the Senate staff research underwriting the Act surfaced

126 Moyar, Oppose Any Foe, 180. 127 James R. Locher, III, “Congress to the Rescue: Statutory Creation of USSOCOM,” Air Commando Journal 1, no. (Spring 2012). 128 William G. Boykin, Special Operations and Low-Intensity Conflict Legislation: Why It Was Passed and Have the Voids Been Filled (Carlisle, PA: U.S. Army War College, 1991). 129 Ibid. 130 Moyar, Oppose Any Foe, 180. 63

DoD’s shortcomings in LIC—and also the “predominance of Service perspectives in DoD” because of ineffective OSD organization and the Joint Staff’s domination by the services.131

Meanwhile, Cohen and Nunn conducted a series of hearings on gaps in SOF training, resourcing, equipping, and operational performance and became increasingly convinced that DoD’s institutional resistance to special operations and special operators was too great for fundamental change to emerge internally.

Although much of the momentum toward legislation was built during the armed services committee process, a great deal of both the intellectual effort and the eventual compromise that led to the creation of SOCOM and SOLIC was generated by Congressman Daniel. Daniel helped get the House to form a Special Operations Panel to review implementation of the administration’s SOF revitalization goals in 1984, submitted bills calling for SOF reform, and participated in the conference negotiations between the House and the Senate over SOF legislation.

The list of Daniel’s influencers included LTG (ret.) Samuel Wilson, a renowned SOF veteran. Wilson had served on the Holloway Board after Eagle Claw, was part of DoD’s Special

Operations Advisory Panel, had retired from the Army after running the Defense Intelligence

Agency, and acted as an advisor to George H.W. Bush’s 1980 presidential campaign.132 He evaluated the Delta Force in 1979 and recommended a joint task force structure to provide command and control, a recommendation he apparently repeated during the Holloway review.

Perhaps most importantly, he had given Congressman Daniel a battlefield tour in Vietnam and

Daniel had respected Wilson ever since.133 According to researchers at the RAND Corporation:

131 United States Senate, Defense Organization: The Need for Change (Washington, DC, October 1985), 99-86, quote on p. 4; Locher, “Congress to the Rescue,” 2012. 132 Peterson, “Coming in from the Cold.” 133 Boykin, Special Operations and Low-Intensity Conflict Legislation. 64

Daniel asked Wilson to participate in the [SASC and HASC conference] negotiations, which was a critical decision: Wilson, having the trust and confidence of both sides, mediated the disagreements between both teams. He convinced HASC of the merits of a military-led command integrated into the National Command Authority’s policymaking structure, and he negotiated with the SASC to include Daniel’s provision to establish the creation of an SOF- specific major force program (MFP) budget line, now called MFP-11, by the Senate.134

Other key personnel included a member of the HASC staff and later Daniel’s personal staff, Ted

Lunger, who was also a former special operator in Vietnam and worked closely with Daniel and with Koch on the SOF advocacy project, and Ken Johnson, a former Special Forces officer who helped Locher write the defense reform report.135 Efforts on the Hill were also spurred and then cheered on by Noel Koch and his allies in the SOF institutional community. Writing in the pages of Armed Forces International Journal, Cohen, Nunn, Daniel, Koch, and others, spent several months making the case for SOF.136

Many of the articles were part of an artificial debate written by Daniel’s staffer Ted

Lunger and OSD staffer Lynn Rylander.137 The first article was signed by Rep. Daniel and argued for separating SOF from the services entirely and developing a “sixth service”

(presumably counting the Marines and Coast Guard). Another article penned by a Deputy

Assistant Secretary (DAS) in the Air Force (and former special forces group member) retread the argument that what SOF do isn’t sufficiently different from other military missions to justify their separation.138 Senator Cohen also authored an article calling for reform—significant for both the eventual legislation and because he was later to be Secretary of Defense under President

Clinton.

134 Robinson et al., Improving the Understanding of Special Operations, 75. 135 Ibid., 77; and Boykin, Special Operations and Low-Intensity Conflict Legislation. 136 Marquis, Unconventional Warfare, 116-127. 137 Boykin, Special Operations and Low-Intensity Conflict Legislation. 138 Marquis, Unconventional Warfare, 121-122. 65

Although civilians on the Hill were persuaded that the military was neglecting low intensity conflict and, as a direct result, SOF, civilians in the Reagan administration remained wedded to the idea that conventional forces could handle most security challenges and SOF reform and expansion was not only unnecessary but possibly dangerous. One chronicler of the process wrote, “Admiral Poindexter, then the National Security Advisor, wrote a number of members to ‘express the President’s concern’ over the pending SOF legislation.”139 Often,

Reagan officials characterized efforts to reform SOF as working against the ideal of jointness. In

1986, when OSD and the Services were fighting against legislated change, the assistant secretary of defense for international security affairs or ASD(ISA) Richard Armitage told a New York

Times reporter that a new SOLIC/SOCOM-like structure would exacerbate “the subliminal wall between special operations forces and the rest of the military that we have been laboring over the last five years to dismantle.”140

Eventually, Cohen and Nunn became convinced that institutional resistance to SOF was too great for the Pentagon to allow innovation internally. The alarming testimony of just-retired

Brigadier General Scholtes, the most recent commander of JSOC and the man in charge of SOF during the Grenada operation, was apparently pivotal in this regard.141 Not only that, but other uniformed SOF advocates had also testified before the SASC in support of mandated SOF reforms, including General Meyer and Lieutenant General Wilson.142

Scholtes’ closed session testimony was perhaps all the more dramatic because it was preceded immediately by a disastrous hearing for the ASD(ISA) Richard Armitage and the

139 Koren, “Congress Wades into Special Operations.” 140 John H. Cushman, Jr., “Pentagon Plans Unit to Build Up Special Forces,” New York Times, July 19, 1986, A8, https://www.nytimes.com/1986/07/19/us/pentagon-plans-unit-to-build-up-special-forces.html. 141 Marquis, Unconventional Warfare, 144; and Robinson et al., Improving the Understanding of Special Operations, 75. 142 Robinson et al., Improving the Understanding of Special Operations, 74. 66

Chairman of the Joint Chiefs of Staff Admiral William J. Crowe. Crowe and Armitage proceeded to demonstrate the alignment between OSD and the JS on the matter of SOF reform—and their divergence from Cohen, Nunn, and the SOF Liberation Front. First, Crowe repeated the JS talking point on the uses of SOF: “They can support conventional operations in either limited or general hostilities or be employed independently when conventional force is either inappropriate or infeasible.” He then went on to defend SOF remaining the responsibility of their respective military departments, but acknowledged that their resourcing, training, and interoperability with each other and the conventional force could be improved and asserted that DoD was making bureaucratic changes to ensure such resourcing, training, and interoperability. Armitage focused his remarks on the evolving understanding of low intensity conflict and the need for a national strategy to wage it. He argued that LIC was not “solely the province of Special Operations

Forces.”143

Such comments were simply at odds with Congressional advocates for SOF, who by now saw a direct link between SOF and irregular, low-level warfare. Senator Cohen’s opening statement had referred to the changing nature of warfare, guerilla insurgencies and terrorism. “I believe that if the United States is to be spared the bitter frustration we have experienced in earlier efforts to conduct counterinsurgency and counterterrorist operations, we must revamp the organization on which our special forces are built.”144 Prior to the Armitage-Crowe hearing,

Senator Nunn had told a reporter, “In my view, the most likely use of force by the United States in the foreseeable future is by our special operations forces. The threat that we face from terrorism and from other forms of low-intensity conflict mean that we must be prepared to deter

143 U.S. Congress, Senate, To Combat Terrorism and other Forms of Unconventional Warfare, Hearing before the Subcommittee on Sea Power and Force Projection of the Comm. on Armed Services, 99th Cong., 2nd sess., August 5, 1986. 144 Ibid. 67

and respond if necessary with special operations forces.”145 In the same article, an unnamed general officer working in the Pentagon on SOF issues said, “special operations, regardless of what they seek to accomplish, will always be part of a larger effort which requires conventional forces and conventional support.”146 Meanwhile, Congressman Daniel remarked in an interview

“the problem is, they’re trying to conduct unconventional warfare with conventional plans and commanders.”147 Between the contentious hearing with Armitage and Crowe and Scholtes’ and other military officers’ testimony, Cohen and Nunn concluded that legislation was the only thing that would inspire the Pentagon to invest in SOF and, by doing so, also prepare for low-intensity conflict.148

The result was the Nunn-Cohen Amendment to the National Defense Authorization Act of 1987, which established a separate combatant command for SOF, and in so doing gave the

SOF community a permanent and much more powerful position over its own fate and over the uses of SOF. Although the law is usually discussed in the context of defense reorganization, the primary reason for and effect of SOF’s institutional growth was to orient and prioritize SOF capabilities. Codified in Title 10 of the U.S. Code, the legislation stipulated the core missions— or purposes—for special operations forces: direct action, strategic reconnaissance, unconventional warfare, foreign internal defense, civil affairs, military information support operations, counterterrorism, humanitarian assistance, theatre search and rescue, and other activities specified by the President or SecDef.149 To ensure the Department of Defense would

145 Bernard E. Trainor, “Special Military Forces: Congress Sees Room for Improvement,” New York Times, September 6, 1986, A9, https://www.nytimes.com/1986/09/06/us/special-military-forces-congress-sees-room-for- improvement.html. 146 Ibid. 147 Bill Keller, “Conflict in Pentagon is Seen Hurting Elite Units’ Buildup,” New York Times, January 6, 1986, A1, https://www.nytimes.com/1986/01/06/us/conflict-in-pentagon-is-seen-hurting-elite-units-buildup.html. 148 Marquis, Unconventional Warfare, 144. 149 Unified Combatant Command for Special Operations Forces, U.S. Code 10 (2011), § 167. 68

not neglect these missions, Nunn-Cohen created the Assistant Secretary of Defense (ASD) for

Special Operations and Low-Intensity Conflict (SOLIC) and the U.S. Special Operations

Command (SOCOM), and mandated a special budgetary vehicle for SOF, Major Force Program-

11 (MFP-11), to be developed and overseen by SOCOM in coordination with the ASD for

SOLIC.

Taken together, the Nunn-Cohen reforms gave the SOF community more bureaucratic and budgetary power relative to the services and the operational force in the combatant commands by combining the functions of both into one military organization. The law did so without also establishing a civilian service secretariat. Instead, a civilian at the more junior assistant secretary level, in a totally separate organization, was expected to oversee development of special operations forces and policy for the threats that might prompt their use.

Implementation of the law and the working relationship between SOCOM and SOLIC would be critical to SOF’s strategic purpose (and civilian oversight of that purpose) going forward.

The Pentagon’s resistance to a more strategic role for SOF notwithstanding, the parallel tracks for terrorism policy and the purpose of SOF were also beginning to intersect, although just barely. The spectacular terrorist attacks in Lebanon and on commercial air and sea carriers began to demonstrate to U.S. policymakers, particularly Secretary of State George Shultz that “terror was emerging as a new form of warfare for which the United States was poorly prepared,” sentiments that echoed the conclusions Army leaders had made in the late 1970s prior to reviving the Rangers and establishing Delta Force.150 Shultz gathered a group of senior officials to discuss

150 John Arquilla, The Reagan Imprint: Ideas in American Foreign Policy from the Collapse of Communism to the War on Terror (Chicago: Ivan R. Dee, 2006), 179. 69

terrorism on March 24, 1984, including Weinberger, the FBI Director, Koch, and officials from

State as well as the CIA.151

The result was National Security Decision Directive (NSDD) 138, “Combatting

Terrorism”, the administration’s first major policy on the topic.152 The document provides a kind of window into the state of the debate not only about the policy to “prevent, counter, and combat terrorism” itself, but the relevance of terrorism to the Cold War and the tension between conventional military tools and irregular warfare, including paramilitary, capabilities. State sponsors of terrorism were still the prevalent concern, and in the first paragraph, the authors noted “the possibility” that the Soviet Union was sponsoring some terrorist groups. In that vein, and yet also providing a counterterrorism purpose broader than the Cold War rivalry, the document stated that the U.S. should “ameliorate the subversive effect of terrorism on foreign democratic institutions and pro-Western governments” and consider “the practice of terrorism by any person or group in any cause a threat to our national security.” To implement this policy, the document directed improved intelligence collection on terrorist groups; that the Secretary of

Defense continue to improve military counterterrorism capabilities; and that the CIA establish

“capabilities for pre-emptive neutralization of anti-American terrorist groups” along with a

“clandestine service” to respond to terrorism abroad.

151 Wills, The First War on Terrorism, 83. Around the same time, a top CIA covert operative named Bill Buckley was kidnapped by Hezbollah-affiliated terrorists and tortured for 19 months. Director of the CIA William Casey was single-minded about recovering Buckley and loosening constraints on the CIA’s paramilitary counterterrorism program. He pressed for, and got, authority from the President (NSDD-138, see below) for the CIA to grow “capabilities for the pre-emptive neutralization of anti-American terrorist groups.” Casey then created a Counterterrorist Center (CTC) which combined intelligence collection, analysis, and paramilitary capabilities. In many ways, the new CIA organization echoed Delta and JSOC, and in fact collaborated with them. Annie Jacobsen, Surprise, Kill, Vanish: The Secret History of CIA Paramilitary Armies, Operators, and Assassins (New York, NY: Little, Brown, and Company, 2019), 270-271. 152 White House, “National Security Decision Directive 138: Combatting Terrorism,” April 3, 1984, https://fas.org/irp/offdocs/nsdd/nsdd-138.pdf. 70

Despite the document being presidential-level guidance, subsequent events demonstrated an ongoing split among civilians on the matter of counterterrorism policy. Weinberger was unmoved, and Vice President Bush was similarly skeptical that the U.S. should conduct a CT campaign. The difficulty lay, in part, in the fact that civilians—and much of the conventional military—were still thinking of both proactive counterterrorism and retaliatory measures in conventional military terms. A debate over the appropriate response to yet another terrorist attack on a U.S. diplomatic post in Beirut surfaced arguments over targeting. The Pentagon dismissed a special operation as too risky, leading to an examination of airstrikes which were in turn ruled out as too damaging to civilians and U.S. diplomatic interests alike.153 Bush, meanwhile, pushed back on a series of public speeches Shultz had been giving advocating for a more proactive military counterterrorism campaign.154 “We are not going to go out and bomb innocent civilians .

. . I don’t think we ever get to the point where you kill 100 innocent women and children just to kill one terrorist.”155

Additional terrorism incidents in 1985 finally prompted the administration to convene a vice presidential-level task force chaired by the same retired Admiral James Holloway who led the review of Operation Eagle Claw. The review resulted in NSDD 207, which directed that an interagency group craft a policy on the appropriate uses of force in response to terrorism.156 The

Secretary of Defense, specifically, was directed to:

153 Wills, The First War on Terrorism, 85. 154 In a later interview for the Miller Center, Shultz explained, “I thought terrorism was a form of warfare, different from the kind that was always planned on. You had to respond and you had to do more than respond, more than defend yourself. You had to be able to take preemptive action and use your military forces doing it.” George P. Schultz, interview by Stephen Knott, Marc Selverstone, and James Sterling Young, “George P. Schultz Oral History,” Miller Center of Public Affairs, University of Virginia, 2005. https://millercenter.org/the- presidency/presidential-oral-histories/george-p-shultz-oral-history-secretary-state. 155 Bernard Gwertzman, “Bush Challenges Shultz’s Position on Terror Policy,” New York Times, October 27, 1984, A1. 156 Wills, The First War on Terrorism, 135-136; and the White House, “NSDD 207: The National Program for Combatting Terrorism,” January 20, 1986, https://www.cia.gov/library/readingroom/docs/CIA- 71

Review . . . current criteria and procedures for deploying and employing U.S. military CT forces. Consideration should be given to political and legal questions involved and to forward deployment or prepositioning of CT elements in or near areas of most likely employment.157

Directing the Secretary of Defense to consider deploying counterterrorism forces “forward,” or outside the United States, began a change in how DoD had managed terrorist threats up to this point. To have military forces designed to conduct counterterrorism operations stationed closer to places where terrorist attacks occurred made it more rational, from a timing perspective at least, to use military force—and to use those SOF that were forward-deployed—to respond to such attacks.

The Vice President’s Task Force on Combatting Terrorism issued its public report in

February of 1986. The text made it clear that the administration thought about terrorism as individual incidents that required retaliation and that the best way to eliminate adversaries’ reliance on terrorism tactics was to U.S. policy to render those tactics ineffective. This conceptual framing of terrorism made the military’s role narrow. The report referred to military force as a precision instrument: “A successful deterrent strategy may require judicious employment of military force to resolve an incident.”158

Without using the terms “special operations” or “conventional forces,” the task force’s report revealed the strategic tensions between SOF-style use of force and the larger footprints of conventional forces. Training, intelligence, and timeliness were highlighted as essential,

RDP88G01117R000200310001-4.pdf. NSDD 207 listed State, Justice, and the FAA as lead agencies for CT depending on the location and type of crisis (international, national, aviation). The use of force framework was to be developed by one of the interagency task forces and routed through the secretary of state. However, NSDD 207 also lists military, paramilitary, and covert action tools as essential to CT. And it directs the secretary of defense to “develop plans for integrating psychological operations more closely into the overall CT effort” and “review requirements and develop alternatives for more effective intelligence and operational support to CT forces/operations.” 157 The White House, “NSDD 207.” 158 George H.W. Bush, Public Report of the Vice President’s Task Force on Combatting Terrorism (Washington, DC: White House, February 1986), https://www.ncjrs.gov/pdffiles1/Digitization/138789NCJRS.pdf. 72

attributes associated with JSOC’s direct action capabilities. However, a “show of force” could have utility as well, typically a phrase that implies overt, large-footprint conventional operations.

Use of our well-trained and capable military forces offers an excellent chance of success if a military option can be implemented. Such use also demonstrates U.S. resolve to support stated national policies. Military actions may serve to deter future terrorist acts and could also encourage other countries to take a harder line. Successful employment, however, depends on timely and refined intelligence and prompt positioning of forces. Counterterrorism missions are high-risk/high-gain operations which can have a severe negative impact on U.S. prestige if they fail.

A U.S. military show of force may intimidate the terrorists and their sponsors. It would not immediately risk more U.S. lives or prestige and could be more effective if utilized in concert with diplomatic, political or economic sanctions. There are, however, some distinct disadvantages: a show of force could be considered gunboat diplomacy, which might be perceived as a challenge rather than a credible threat; it may require a sizable deployment of support activities; it may provide our enemies with a subject for anti-American propaganda campaigns worldwide; and most important, an active military response may prove necessary to resolve the situation if a show of force fails.159

By the end of the period, in fact, civilian leaders had settled on a shared approach to countering terrorism that used SOF for immediate crisis response and GPF for punishing and deterring governments from sponsoring terrorist groups. This was exemplified by the U.S. response to a Libyan-sponsored terrorist attack that killed two U.S. servicemembers in West

Berlin in 1986. The administration retaliated with air strikes, and explained its response in the

FY1988 DoD Annual Report to Congress (written in 1987):

The Libyan action was not carried out by the kind of special operations forces that are involved in combatting specific terrorist acts while they are in progress and, in a sense, this is even a greater tribute to our conventional forces . . . The objective of the Libyan operation was both to strike at terrorist support bases, and to teach the state of Libya that providing terrorist groups with the support necessary to conduct their international campaign of aggression against the United States carries with it a terrible cost. Thus, our strategy for precluding and combatting terrorist acts involves a range of general purpose forces as well as special operations forces.160

159 Ibid. 160 U.S. Department of Defense (DoD), Annual Report to the Congress, FY1988 (Washington, DC, January 1987), 61, https://history.defense.gov/Portals/70/Documents/annual_reports/1988_DoD_AR.pdf?ver=2014-06-24-151204- 840. A proposal had also come from the CIA to assassinate Libya’s leader, Muammar Qaddafi, but was rejected by 73

SOF’s self-generated orientation toward conducting direct actions against terrorists had made its way into civilian preferences. What SOF did not do yet was address terrorism comprehensively. It was still characterized as an “in emergency, break glass” tool, rather than the element of the force that prevented, responded to, and punished all terrorist violence.

The upshot of the 1977-1987 period was that civilian preferences about SOF came to reflect the uniformed Army’s marriage of LIC and CT, but they were still anchored in the conventional military’s preference to emphasize SOF as supplementing conventional warfighters. Still, SOF had achieved not just bureaucratic distinction from the services in the form of SOCOM but were moving into a strategic distinction as well. The fiscal year 1986 DoD annual report to Congress (submitted in 1985) was the first to have a separate section on SOF.

By the 1987 annual report, the Secretary of Defense was telling the Congress that SOF,

. . . provide the United States a highly flexible, specialized capability to pursue national objectives during peace or war, either independently or in conjunction with conventional forces. In peacetime, SOF, in conjunction with other military forces and federal agencies, participate in security assistance, civic action, and humanitarian assistance operations. They also contribute to combatting terrorism. SOF can play a key role in crises through the use of forces, psychological operations, and the employment of civil affairs units. The importance of these capabilities was clearly demonstrated in 1983 in Grenada. SOF’s unique skills as trainers, derived from cultural orientation and language training, also make them an essential element in counterinsurgency operations. If called upon, these same skills can be employed in support of guerrilla warfare. At higher levels of conflict, the SOF can delay, divert, and disrupt enemy operations, thereby gaining a critical edge for conventional defenses.161

That same year, the White House incorporated the idea of a balance between

“specialized” and “general purpose forces” into the country’s first-ever National Security

Strategy Report (NSSR):

unspecified actors at the White House, possibly because military action was covered by the right to self-defense under the UN Charter. Jacobsen, Surprise, Kill, Vanish, 273. 161 DoD, Annual Report to the Congress, FY1988, 293. 74

The United States must have specialized forces—ranging from those required for nuclear deterrence to forces configured to deal with terrorism; and must also have general purpose forces capable of sustaining high intensity conflict, while maintaining an effective capability for lesser contingencies and special operations.162

Like the annual reports to Congress submitted by the Secretary of Defense, the NSSR formally expressed civilians’ association between SOF and counterterrorism missions as well as between

SOF and “lesser contingencies,” another way to express the military’s “lesser-included case” jargon that effectively meant the missions for which the conventional force was not designed. It was another step toward civilians preferring to use SOF for specific types of missions—for counterterrorism and LIC—rather than as the part of the general purpose force more capable of doing a wide range of difficult but unspecified things.

How influential had military preferences been in shaping civilian preferences on SOF? In

December of 1987, the Reagan administration submitted a statutory report to the Congress on the

U.S. strategy for low-intensity conflict. Annex A of the report addressed the role that DoD played in the strategy and highlighted the use of special operations forces. The discussion of SOF began with the Joint Chiefs of Staff definition of special operations. It goes on to list the critical tasks for LIC, all of which were SOF missions: foreign internal defense (FID), peacekeeping, combatting terrorism, and “peacetime contingency operations.”163 That the Joint Staff’s language framed the ways SOF fit into low-intensity conflict and that SOF itself had informed the statutory missions of its own force suggested military preferences shaped civilian preferences a great deal.

162 U.S. National Security Council, National Security Strategy Report (Washington, DC: The White House, 1987), 19, http://nssarchive.us/NSSR/1987.pdf. 163 FID connotes U.S. training of indigenous security forces. U.S. National Security Council, Report to the Congress U.S. Capabilities to Engage in Low Intensity Conflict and Conduct Special Operations (Washington, DC: The White House, 1987). 75

Perversely, as LIC became a more meaningful U.S. military mission, more of the military wanted a piece of it. The Reagan administration reported to Congress in 1987 that SOF fulfilled

LIC missions, but so did “combat maneuver forces.”164 SOF were still not specialists in any mission area in an exclusive sense as far as civilians were concerned. But they had maneuvered themselves into major strategic documents, adhering special operations to an emerging strategic concept. They had survived.

Period II: Market Share (1988-2000)

In 1988, civilians preferred SOF to fulfill a range of missions spanning the spectrum of conflict. They had essentially split the baby by agreeing with the conventional force’s preference for SOF to be their “adjunct” while also increasing emphasis on SOF’s capabilities in unconventional warfare and counterterrorism. This meant civilians were comfortable with SOF providing FID and conducting low intensity conflict nearly exclusively, while counterterrorism and support to conventional operations were duties they shared with the general purpose forces, in part because SOF and the GPF had not agreed on a division of labor in those two areas and civilians saw no immediate need to adjudicate one.

But unlike the first period, there was no marked change in civilian preferences between the beginning and the end of the second period. Despite the dramatic shifts in international and domestic politics, and a significant post-Cold War downsizing of the military, by 2000 civilian views of SOF’s purposes were generally the same as they had been in 1988. As the 1997

Quadrennial Defense Review stated, SOF “provide a range of unique capabilities that have

164 Ibid. 76

important applications across the full spectrum of conflict.”165 When it came to civilian preferences for their strategic purpose, SOF appeared to be holding their ground.

But this appearance of near stasis belies the pronounced variance within the period in how civilians conceived of SOF and its purposes. In some sense, civilians went on a circular journey of discovery and arrived, by the year 2000, back at the same basic place they had been in

1987. This was partly a consequence of the upheaval in defense strategy and foreign policy caused by the end of the Cold War. The loss of the main U.S. state adversary coupled with the prevalence of regional and intra-state conflicts raised questions about whether the entire military should focus on military operations other than war, in turn raising questions about whether more conventional means, especially air power, could be used to meet unconventional threats, an idea that itself threatened to crowd SOF out of the low-intensity conflict space it had carved for itself in the 1980s.

The variance in civilian preferences also followed the partisan identities of the three presidential administrations that spanned 1988 to 2000. 1988 marked the last year of the

Republican Reagan administration, and civilian views of SOF had already been through seven years (plus the presidential campaign) of military influence. The H.W. Bush administration, also

Republican, made no great adjustments in its overall views of SOF but did give SOF more strategic prominence and in 1991 articulated a detailed vision of SOF’s relevance to “limited engagements in regional contingencies” while still referring to SOF as “crucial adjuncts to larger conventional actions.”166 Civilians in the H.W. Bush administration clearly had absorbed

165 Department of Defense, Report of the Quadrennial Defense Review (Washington, DC: May 1997), https://www.files.ethz.ch/isn/32542/qdr97.pdf. 166 Department of Defense, Annual Report to the President and the Congress, FY1992 (Washington, DC: February 1992), https://history.defense.gov/Portals/70/Documents/annual_reports/1992_DoD_AR.pdf?ver=2014-06-24- 152154-673. 77

arguments that irregular and unconventional warfare would become more prevalent, making SOF broadly useful. “SOF have a role to play in each element of the [defense] strategy – particularly in forward presence and crisis response.”167

The Democratic Clinton team came into office in 1993 with very little knowledge of special operations and few strong preferences about SOF. Moreover, the White House began by delegating management of defense policy to the Pentagon. At DoD itself, there were certainly civilians with sufficient expertise to weave SOF into the first articulation of defense strategy and resources, the Bottom-Up Review, but, perhaps because Jim Locher was still the ASD for

SOLIC, they essentially repeated the H.W. Bush administration’s application of SOF across the spectrum of conflict, to include counterterrorism and countering weapons of mass destruction.168

The 1993 Black Hawk Down incident, however, soured many civilians to SOF direct action, even as stabilization operations in Haiti and later Bosnia sold them on civil affairs and psychological operations. Thus, the Clinton civilian team spent several years emphasizing SOF’s utility as the nation’s post-intervention stabilization force. Yet by the end of the Clinton years, civilian preferences for SOF had rebounded somewhat to incorporate, again, direct action and counterterrorism, while the conventional forces as represented by the Chairman of the Joint

Chiefs of Staff worked to persuade senior policymakers to veto SOF direct action missions. In sum, period two of this case study was another cycle where civilian preferences bounced between those supplied by military institutions.

167 Department of Defense, Annual Report to the President and the Congress, FY1993 (Washington, DC: January 1993), https://history.defense.gov/Portals/70/Documents/annual_reports/1993_DoD_AR.pdf?ver=2014-06-24- 152422-603. 168 I am grateful to Kathleen Hicks for observing the connection between Locher and the Clinton administration’s continuation of the Bush administration’s approach to SOF uses. 78

What buffeted civilian preferences in this way? Certainly, the changes in administration introduced variation in assumptions, ideas, and level of knowledge—as soon as one team’s preferences stabilized, another team came in, sometimes with different starting points in their expertise or preferences themselves. That each team eventually came to the same general preference is telling. How did this result obtain?

The conventional military continued to dominate civilians’ preference formation, but the codification of SOF missions in Title 10 along with the existence of SOCOM and SOLIC and almost a decade of work among defense communities on the growing importance of low- intensity conflict meant that tolerance of SOF’s strategic “market share” had gained widespread acceptance among defense practitioners, whether in uniforms or suits. However, as the Cold War ended and the framing threat from the now-defunct Soviet Union faded, ideas about what generated LIC and how the U.S. should handle it transformed. Insecurity around the world manifested as civil war and Western countries increasingly sought to defend principles and regulate international and sub-national politics in a variety of regions and nation-states rather than constrain the power of one major adversary. In the mid-1990s, concerns about transnational terrorism faded into the greater outcry over genocides, civil wars, and humanitarian crises.169

This made Special Forces A-teams, psychological operations and civil affairs especially relevant to U.S. foreign policy and the more kinetic, confrontational side of SOF (JSOC primarily) less salient.

SOF’s purchase on strategic missions was thus shifting under its feet, and it would need to prove its ongoing relevance. Institutionally, SOF was of two minds. SOCOM was content with

169 A clunkier term than LIC often used by military actors was “Military Operations Other Than War,” or MOOTW, a framing that dovetailed with another bit of military jargon, the “lesser-included case.” Defense analysts also used the terms “post-conflict stabilization” and “complex contingency operations” during this period. 79

SOF’s association with LIC and stability operations, mainly fighting to preserve prominence in these areas in the minds of civilians. JSOC strained against the conventional military and civilians’ shared risk aversion and distraction from the threat of terrorism. Thus, as in the 1980s, there was still a divide within the military about what SOF’s purpose should be. Only now, a substantial portion of the SOF community and the conventional force were aligned, with the conventional forces having moved in the direction of acknowledging that SOF’s domain was unconventional and irregular warfare. JSOC, as ever, thought direct action, especially where terrorist groups were concerned, should be a larger part of SOF’s portfolio of activities—and therefore, what civilians preferred to do with their special operators.

SOF spent the years between 1987 and 2000 developing their market share, alternately convincing policymakers to use them and simply fulfilling the missions the defense community now agreed were theirs. JSOC struggled to keep itself relevant and employed, while SOCOM and the Services negotiated over the appropriate levels of investment in SOF capabilities.

Details

Preferences about SOF may have remained unchanged in the aggregate between 1987 and 2000, in part, because the SOF community spent at least part of that time consolidating its organizational gains. Implementation of the Nunn-Cohen legislation needed to be worked out between OSD, the Services, and SOCOM. Which forces belonged under SOCOM? Would

SOCOM’s new budget authority (the MFP-11) be run as a unique process or would it participate in the Planning, Programming, and Budget System like other COCOMS?170 Would it have the same opportunity to provide the Congress with “unfunded requirements” lists as end-runs to

170 The PPBS is DoD’s process for allocating budget share among its various military and civilian organizations. 80

DoD budgeting? What was the relationship between the Secretary, the Under Secretary for

Policy, ASD(SOLIC), and SOCOM? It took until 1989 for most of these issues to be resolved.171

This was in part because Army General James Lindsay was confirmed as commander of

SOCOM in 1987 but Ambassador Charles Whitehouse was not confirmed ASD(SOLIC) until

1988.

The year gap in military and civilian leadership over special operations policy mattered, not only because it gave General Lindsay a head start substantively and bureaucratically, but also because it meant Whitehouse was only in place for the last year of the waning Reagan administration, a span of time that was not long enough to introduce and implement many new ideas. During that year, Secretary of the Army John O. Marsh was given the acting Assistant

Secretary for SOLIC job—unusually enough by the Congress, which was fed up with what it saw as the Reagan administration’s foot-dragging over filling the role.172 Marsh has been credited with making arrangements for organizational matters such as billets and budgeting, but also had, unsurprisingly, carryover notions of SOF from the Army. It was not, therefore, a period in which the civilians charged with SOF thought leadership were likely to develop preferences independent from the SOF community. Meanwhile the wrangling over bureaucracy meant that much of the SOF community’s energies at the higher levels were consumed by organizational turf battles rather than the substantive strategic purpose of the forces.

Just getting its own house in order actually meant that SOF lost some ground with respect to convincing the services to continue their role in developing SOF. An article in the New York

Times Sunday magazine described a force squandered, still repressed by conventional military

171 Christopher Paul, Isaac R. Porche, and Eliot Axelband, The Other Quiet Professionals: Lessons for Future Cyber Forces from the Evolution of Special Forces (Washington, DC: RAND Corporation, 2014), 15-17. 172 Marquis, Unconventional Warfare, 179. 81

leaders, hindered by intelligence-gathering turf wars with the CIA, and suffering even less mobility support from the Air Force than it had at its disposal in 1980. “Top brass view low-cost alternatives as a threat to big-ticket items, contending that massive firepower is the answer to unconventional warfare.”173

Nevertheless, the mere existence of SOCOM changed the dynamics of influence between military and civilian preferences because it gave SOF much more bureaucratic power. Notably, this had been legislators’ intention. Senator William Cohen told the same New York Times reporter that the purpose of creating SOCOM was to “insure that counterterrorism and the special operations forces had an institutional voice in the Pentagon because it didn’t have its own commercial lobby.”174 With its senior level leadership and new budgetary authority, SOCOM positioned the SOF community to be far more persuasive within DoD and gave it the independence to begin to define its roles and missions for itself—including through operational planning and doctrine. One of SOF’s chief chroniclers, John Collins, wrote that by 1989,

“CINCSOC [Commander SOCOM] and his staff could develop doctrine and tactics in the absence of sound policy guidance from the National Security Council and ASD SO/LIC.”175 SOF was now positioned to develop preferences for its own use under relaxed civilian (and other military) constraints.

Meanwhile, low-intensity conflict as a defense fashion hit its apogee in the late 1980s.

Already in 1986, LIC and SOF were conjoined to the extent that Congress determined the ASD for Special Operations would also be the ASD for LIC. This did not mean that controversy over the ideas underlying LIC had been resolved. Much of the LIC debate was about the threshold for

173 Steven Emerson, “Stymied Warriors,” New York Times Magazine, November 13, 1988, 68-112. 174 Ibid. 175 John Collins, Special Operations Forces: An Assessment (Washington, DC: National Defense University Press, 1994), 13. 82

using force, a tripwire whose location was unclear, or at least highly dependent on context.176

But most civilians now accepted that LIC was a feature of the international security environment and that it presented a demand signal for a different type of force than what DoD was typically prepared to provide. SOF was the only element of the force that wanted such a mission, so it had the opportunity to shape itself for it.

At the beginning of the period, senior civilians used the term LIC frequently, and associated it with SOF often—if not by name, then at least by their primary mission set, as demonstrated by the marriage between SOF and LIC in the White House’s report to Congress on

LIC from 1987. The 1988 national security strategy, for example, stated that “the most appropriate application of U.S. military power [in LIC contexts] is usually indirect through security assistance—training, advisory help, logistics support, and the supply of essential military equipment.”177 H.W. Bush’s Secretary of Defense Dick Cheney noted in remarks before a Congressional committee, “to help deter low-intensity conflicts and promote stability in the

Third World, we must have innovative strategies . . . Our approach for doing this is ‘peacetime engagement’—a coordinated combination of political, economic, and military actions, aimed primarily at counteracting local violence and promoting nation-building.”178 In a separate

176 See Admiral William J. Crowe, “Implications of Low-Intensity Conflict for U.S. Policy and Strategy,” in Low- Intensity Conflict: Old Threats in a New World, ed. Edwin G. Corr and Stephen Sloan (Boulder, CO: Westview Press, 1992). Crowe of course was the chairman of the Joint Chiefs of Staff who opposed the establishment of SOCOM. Amazingly, he goes on to advocate for LIC to be a SOF responsibility. Although he thinks the entire force should develop LIC capabilities, he also writes, “The worst mistake the services could commit, however, is to dedicate the Special Operations Forces solely to direct action missions and declare that the SOFs are the new LIC forces. . . . Deep reconnaissance experience does not help a special forces adviser teach foreign soldiers how to defeat guerrillas. An increasing emphasis on direct action would also put too much emphasis on the military components of LIC, when in fact low-intensity wars involve fare more than simple military considerations.” 297- 298. 177 U.S. National Security Council, National Security Strategy of the United States (Washington, DC: The White House, 1988), 35, https://history.defense.gov/Portals/70/Documents/nss/nss1988.pdf?ver=2014-06-25-121129-393. 178 Quoted in Sam C. Sarkesian, “Special Operations, Low Intensity Conflict (Unconventional Conflicts), and the Clinton Defense Strategy,” in Clinton and Post-Cold War Defense, ed. Stephen J. Cimbala (Westport, CT: Praeger, 1996), 104. 83

appearance on the Hill, he warned of “low-intensity conflicts, including insurgencies, terrorism, and drug trafficking.”179

The problem for SOF was now that it had helped to persuade civilian policymakers that

LIC, terrorism included, was an object of real strategic importance, it seemed like more of the force than just SOF should be able operate in low-intensity environments. Civilians thus began to express a preference for more of the military to operate like SOF, if not entirely adopt its irregular warfare missions. Although SOF were still seen as uniquely capable in LIC contexts, light, agile, and speedy forces became desirable more generally. The 1990 National Security

Strategy, President George H.W. Bush’s first, is especially revealing in this regard. It included a paragraph of discussion that again centered on SOF-like capabilities with speed, smallness, austerity, and sensitivity to political dynamics.180 Thus, although SOF were now clearly associated with LIC, civilian-penned strategic documents had begun to suggest that the rest of the force should also acquire the types of skills typically associated with special operators.

“American forces therefore must be capable of dealing effectively with the full range of threats, including insurgency and terrorism. Special Operations Forces have particular utility in this environment, but we will also pursue new and imaginative ways to apply flexible general purpose forces to these problems,” Bush’s national security strategy read.181

SOF had to continue proving their unique contributions, not only to the strategic challenges of LIC, but also to broader national objectives. Also writing in 1990, the Commander of SOCOM, General Carl Stiner, presented SOF’s ongoing relevance to and mastery of LIC

179 Richard Cheney, Conflicting Trends and Long Term Defense Needs, statement before the Senate Armed Services Committee, February 17, 1991. 180 U.S. National Security Council, National Security Strategy (Washington, DC: The White House, 1990), 27, http://nssarchive.us/national-security-strategy-1990/. 181 Ibid., 28. 84

while also suggesting SOF were a kind of strategic anchor to hedge against future needs: “The strategic capabilities of our nation’s SOF across the operational continuum have been demonstrated daily during the past year . . . deployments for training provided a significant and needed presence in areas where no permanent US military forces are stationed.”182 Subsequent commanders of SOCOM made similar efforts to demonstrate the unique relevance of SOF. In

1996, toward the end of his tour as Commander, General Wayne Downing penned a chapter on

“new national security challenges” in a book about contemporary warfare. In it, he highlighted global forces of social and political disintegration, including “Islamic fascism,” and argued that the U.S. needed military forces who understood the local political implications of their actions.

He also insisted that good intelligence in such contexts required “having human eyes and ears on target,” a subtle endorsement of SOF’s role in intelligence-gathering.183

The conventional force was not enthusiastic about a shift toward LIC. This was especially true during General Colin Powell’s approximately five years in some of the senior-most national security positions in the U.S. government while he was still an active duty Army officer. A career military officer with deployments to Vietnam and a lot of experience in Washington,

Powell was National Security Advisor during the last two years of the Reagan administration and then Chairman of the Joint Chiefs of Staff during most of the H.W. Bush administration and the first nine months of the Clinton administration, until his retirement from the Army in September of 1993.184 Powell was opposed to small scale uses of force under ambiguous circumstances— the hallmarks of both SOF and low-intensity conflict. He saw SOF as more of a force multiplier

182 Carl W. Stiner, “The Strategic Employment of Special Operations Forces.” Military Review 71, no. 6 (June 1991), 3. 183 Wayne A. Downing, “New National Security Challenges,” in Managing Contemporary Conflict: Pillars of Success, ed. Max G. Manwaring and Wm. J. Olson (Boulder, CO: Westview Press, 1996), 92-99. 184 Colin Powell with Joseph E. Perisco, My American Journey (New York, NY: Random House, 2010). 85

for large force packages rather than a set of detachments that could act independently where conventional forces could not. Powell argued for clear exit strategies, public support, and sufficient forces to win overwhelmingly.185 He was generally opposed to military interventions for vague humanitarian reasons or to simply halt a civil war.186 In fact, he can be at least partly credited with reinvigorating the idea that SOF were adjuncts to the conventional force. As

Chairman, Powell devised a four-part U.S. force structure he called the Base Force, for which

SOF were used to reinforce forward-deployed forces in a crisis.187 Powell’s thinking was reflected in the 1991 NSS, which called for “a smaller and restructured force” with SOF a part of the “contingency forces.”188 Notably, and unlike the 1990 NSSR, the 1991 NSSR had no section on LIC.

Nevertheless, the investments in and reorganization of SOF over the previous decade resulted in an operational proof of concept for its use in small-scale contingencies. The

December 1989 invasion of Panama (Operation Just Cause) was celebrated universally as a vindication of the lessons learned since the Grenada debacle in 1983.189 SOF’s much-improved integration with the GPF and the operation’s rapid success had something for everyone.

Proponents of both GPF solutions to political challenges and SOF’s utility during major operations could point to Just Cause.

Both military and civilian experiences during Just Cause also helped make the case for

SOF’s participation in major operations as well. In 1991 the conventional campaign to push Iraq

185 Halberstam, War in a Time of Peace, 2001. 186 Jon Western, “Sources of Humanitarian Intervention: Beliefs, Information, and Advocacy in the U.S. Decisions on Somalia and Bosnia,” International Security 26, no. 4 (Spring 2002), https://www.jstor.org/stable/3092104. 187 John R. Galvin, “Conflict in the Post-Cold War Era,” in Low-Intensity Conflict: Old Threats in a New World, ed. Edwin G. Corr and Stephen Sloan Boulder (CO: Westview Press, 1992), 66. Galvin was SACEUR and commanded Just Cause. 188 National Security Council, National Security Strategy Report (Washington, DC: The White House, 1991), 31, http://nssarchive.us/national-security-strategy-1991/. 189 SOCOM, United States Special Operations Command; and Clancy, Shadow Warriors, 2002. 86

out of Kuwait provided that test. SOCOM and JSOC both wanted to be involved but were limited by the commanding general in charge of the war, Norman Schwarzkopf, who thought special operations were too difficult to control and might trigger premature escalation.190 SOF lobbying went all the way to the Secretary of Defense. Eventually, Secretary Cheney had to order Powell to intervene with Schwarzkopf on SOCOM’s behalf. Cheney later explained to an interviewer that his personal experience with SOF during Just Cause was dispositive: “Wayne

Downing was the guy [Commander of JSOC], I’d dealt with him directly on a number of occasions. We’d done the planning for Panama, for example, and I’d watched him in operation there. I had a lot of confidence in him and I thought they [SOF] were appropriate for us to use.”191 As described above, SOF went on to perform numerous missions during Operation

Desert Storm and subsequent missions supporting Iraqi Kurds.

These were major operations, however, and Powell’s and the conventional force’s suspicion of involving the military in LIC—regardless of whether such missions were undertaken by SOF or GPF—persisted, setting the stage for the events that would set JSOC back in its efforts to woo the Clinton administration into approving direct action missions.

The Clinton team was poised to adopt the same preferences for SOF that the Bush administration had expressed, and the language in Clinton’s first major articulation of defense

190 Moyar, Oppose Any Foe, 187. 191 Richard Cheney, interview by Philip Zellikow, “Richard B. Cheney Oral History,” Miller Center, University of Virginia, March 16-17, 2000, https://millercenter.org/the-presidency/presidential-oral-histories/richard-b-cheney- oral-history-secretary-defense. Cheney also said: “One of the real problems you have with any new Administration, even one as experienced as ours was—I mean, we weren’t a bunch of amateurs, we’d been around there before—it’s hard. You know, there is no training ground for senior civilian political leaders in an Administration.” Also: “The general arrangement I had with General Powell was that in terms of giving directives—the operational use of the force, dealing with the CINC, for example—I would always go through him. Didn’t have to, that was our option. Under Goldwater-Nichols I could go direct to the CINC if I wanted, but I would refrain from reaching around him to go to the CINC when I was passing something down. I wanted it all to go down through him. I wanted him to have some pretty tight control of that. [But] I did not want a single channel coming up. I wanted multiple sources of information coming up. I did not want him to be the screen for information coming to the Secretary from the military. . . . I used my military assistants an awful lot as a source of information or to go pulse the system and generate stuff I needed.” 87

policy, the Bottom-Up Review (BUR), echoed the preceding Bush administration’s own language on the uses of SOF. The 1993 document assigned a variety of missions to SOF, including involvement in the combat phases of conflict, a countering weapons of mass destruction role, interventions and peacekeeping operations, and plugging holes in conventional deterrence including mobility and presence operations. This was, in significant part, an artifact of civilian defense priorities—SOF being toward the bottom of the list rather than the top, and thus not a place for new thinking. Contemporary analysts claimed the BUR did little to re-examine

LIC in light of the changes to the international environment, arguing it “appeared to be in large part a Department of Defense exercise to protect its budget and institutions rather than an effort to discard entrenched historical perspectives and take a broader view of international threats to

U.S. security.”192

Whether that characterization was fair or not, according to the Deputy Secretary of

Defense during the writing of the BUR (and later Secretary of Defense) William Perry, SOF and its missions were subsumed under the so-called lesser-included case category:

We considered then, first of all, a major de-emphasis on nuclear forces, and second, that it provided the forces for the two major regional conflicts, and that the other things—the insurgency operations, the peacekeeping operations—were lesser included cases. That is, we dismissed them in a sense by saying, Yes, they may be important, but the forces we have should cover them. That was, I think you can appreciate, a bit of intellectual laziness. That is, it’s not that simple, even as it wasn’t that simple in the Cold War to talk about the lesser included cases. But that’s the way we thought about it and treated it, not because we were fully satisfied with that answer, but because we had to get on to doing other things. We had to get that Bottom-Up Review done. That answered the mail. When people had questions about lesser included cases, that was the answer to the question.193

192 Edwin G. Corr and David C. Miller, “Organizing for Operations Other Than War in the Post-Cold War Era,” in Managing Contemporary Conflict: Pillars of Success, ed. Max G. Manwaring and Wm. J. Olson (Boulder, CO: Westview Press, 1996).

193 William Perry, interview by Russell L. Riley, Darby Morrisoe, and Andrew Ross, “William Perry Oral History,” Miller Center, University of Virginia, February 21, 2006, https://millercenter.org/the-presidency/presidential-oral- histories/william-perry-oral-history-deputy-secretary-defense. He went on: “So there was a lot of thinking and a lot of talk and some writing in the Defense Department after the Bottom-Up Review about how you really should be 88

In 1994 the Clinton Administration’s first NSS, “A National Security Strategy of Engagement and Enlargement,” reflected the BUR’s various uses of SOF, including mixing them between unique and conventional operations. There were no references to SOF but to “specialized units:”

A number of other tasks remain that U.S. forces have typically carried out with both general purpose and specialized units. These missions include: counterterrorism and punitive attacks, noncombatant evacuation, counter- narcotics operations, nation assistance, and humanitarian and disaster relief operations.194

Thus, while the early Clinton-era civilians’ preference for the use of SOF were substantively similar to those of the H.W. Bush administration, the strength of that preference was weak. SOF were sprinkled into strategy as an after-thought.

There is evidence of civilian preference weakness in the first year of the administration during the progression of the deployment to Somalia. Having withstood public pressure during the waning days of the H.W. Bush administration to get the military involved in either the deteriorating conditions in the former Yugoslavia or Somalia, General Powell had eventually supported a mission in Somalia because he thought it was less likely to mire the U.S. in a protracted deployment.195 Clinton then inherited the deployment, which at the time was humanitarian in scope and in support of UN operations. SOF were involved mostly in transportation security, other force protection, and civil affairs.196

As Clinton’s first year wore on, however, the dynamics on the ground between international forces and the Somalis shifted. Forces loyal to Mohamed Farah Aideed began

thinking about these other cases and how they should influence the way you put your budget together. But that thinking is not articulated in the Bottom-Up Review. So in my judgment, I’ve always regarded it as an incomplete document.” 194 U.S. National Security Council, A National Security Strategy of Engagement and Enlargement (Washington, DC: The White House, July 1994), http://nssarchive.us/NSSR/1994.pdf. 195 Western, “Sources of Humanitarian Intervention.” 196 SOCOM, United States Special Operations Command History, 55-56. 89

attacking UN personnel, making UN and U.S. force protection a greater concern and drawing international personnel into the conflict. American officials on the ground increasingly requested permission to conduct offensive operations and the forces to conduct them. The Clinton White

House, meanwhile, largely neglected the operational aspects of the situation in Somalia, focusing instead on processes at the UN in the hopes of negotiating some form of settlement.197 There is little evidence that civilian managers at the White House understood the various options DoD had for engaging in Somalia, and Secretary Aspin was either unable or unwilling to draw their attention to it.198 The Battle for Mogadishu therefore came as a shock to an inattentive White

House, acting as an educational device about a number of foreign policy and defense issues to include special operations forces. Having not understood the operational risks being run on the ground or the political risk of military casualties outside major combat operations, the Clinton

White House soured against direct action.199

At the same time, however, major stabilization operations were launching in Haiti.

Special Forces teams that specialized in FID and civil affairs fanned out across the rural areas, in many cases supplying basic services to the Haitian population.200 James Dobbins, then the U.S.

Special Envoy for Haiti, recalled that Special Forces “did a good job” establishing law and order

“throughout the countryside.”201 Special Forces also performed well in Bosnia and Kosovo, conducting liaison missions with multinational forces and supplying intelligence on local conditions—and, by helping stabilize post-conflict societies, proving that intervention was a

197 Anonymous former senior OSD official, interview with the author, December 5, 2018. 198 Derek Chollet and James Goldgeier, America Between the Wars: From 11/9 to 9/11 (New York, NY: Public Affairs, 2009). 199 Anonymous former senior OSD official, interview with the author, December 5, 2018. 200 Philippe R. Girard, “Peacekeeping, Politics, and the 1994 U.S. Intervention in Haiti,” Journal of Conflict Studies 24, no. 1 (2004): 20-41, https://journals.lib.unb.ca/index.php/jcs/article/view/290/461. 201 “The World Was Tired of Haiti: The 1994 U.S. Intervention,” Association for Diplomatic Studies and Training, July 11, 2016, https://adst.org/2016/07/world-tired-haiti-1994-u-s-intervention/. 90

workable method for conducting U.S. foreign policy.202 President Clinton’s admiration for them even prompted a reference in his second Democratic presidential nomination acceptance speech in 1996: “The Special Forces are just what the name says; they are special forces. If I walk off this stage tonight and call them on the telephone and tell them to go halfway around the world and risk their lives for you and be there by tomorrow at noon, they will do it.”203

The result was that SOCOM was extremely busy deploying SOF on the full range of their

Title 10 missions—except for direct action and counterterrorism. Overall SOF deployments surged during the 1990s, and SOCOM’s promotion of that fact makes it difficult to disentangle civilian demand from military persuasion.204 But interviewees claimed that the different political experiences Clinton administration civilians had of Somalia and Haiti informed their beliefs.205

JSOC and its direct action orientation took a back seat to civil affairs in civilians’ preferences.

Meanwhile, intra-military dynamics continued to mold concepts of SOF’s purpose in almost infinitely flexible ways. Operationally, conventional military leaders often continued to resist broad applications of special operators. In Bosnia in particular, conventional commanders limited Army Special Forces elements to a liaison function during the first year of operations.206

However, Combatant Commanders (then called Commanders-in-Chief, or CINCS) generated a high demand for SOF in a variety of ways. For example, successive SOUTHCOM Commanders

202 Armando J. Ramirez, “From Bosnia to Baghdad: The Evolution of U.S. Army Special Forces 1994-2004,” Master’s Thesis, U.S. Naval Post Graduate School, 2004, https://fas.org/man/eprint/ramirez.pdf. 203 “Clinton’s Speech Accepting the Democratic Nomination for President,” New York Times, August 30, 1996, https://www.nytimes.com/1996/08/30/us/clinton-s-speech-accepting-the-democratic-nomination-for-president.html. 204 SOCOM’s official history notes that the “number of personnel deployed away from home station per week” doubled between 1993 and 1996 alone (25). 205 Interviews with Collins, Wechsler, Anonymous former senior OSD official. 206 Ramirez, “From Bosnia to Baghdad,” 23-25. 91

(including General Joulwan and General Charles Wilhelm) continually used SOF because the command was focused on counter-drug and FID missions.207

In 1997, the GAO conducted a report, the objective of which was “to determine whether

SOF are being used in a manner that best supports national security objectives.”208 The report observed that SOF were being used at a rate that threatened readiness, in part because SOF were used for “missions that could be handled by conventional forces.” They found that SOF were often “the force of choice” for regional CINCs in day-to-day operations.209

In that event, much as during the Reagan years, terrorism and SOF’s role in countering it became a growing part of civilians’ agenda due to a number of spectacular attacks on American targets. In 1993, a truck bombing in one of the World Trade Center parking garages hinted at the coming pattern of attacks. In 1996, an attack on Saudi

Arabia’s Khobar Towers drew yet more policymaker attention. Then in 1998, near- simultaneous bombings of the American embassies in Kenya and Tanzania propelled high-level policy discussions at the White House.210

Civilian and military preferences about SOF’s role in responding to the threat of terrorism during this period also reveal the extent to which civilian preferences came full circle.

In the early part of the period, as a result of the march of terrorist attacks in the 1980s and multiple high-level strategic reviews, civilians in the H.W. Bush administration had accepted terrorism as a growing threat to U.S. interests. For example, Secretary of Defense Cheney had argued in DoD’s FY1989 annual report to Congress,

207 Dana Priest, The Mission: Waging War and Keeping Peace with America’s Military (New York, NY: W.W. Norton and Company, 2003), 200-205. 208 U.S. General Accounting Office, Special Operations Forces: Opportunities to Preclude Overuse and Misuse, NSIAD-97-85 (Washington, DC: GAO, May 1997), https://www.gao.gov/products/GAO/NSIAD-97-85. 209 Ibid. 210 Shultz, Jr., “Showstoppers.” 92

One type of low-intensity conflict—terrorism—has taken on a new character. What once was largely the activity of small, frustrated extremist groups within countries has become a virtual multinational enterprise, and state-sponsored terrorism has emerged as a new weapon in the arsenal of ambiguous aggression.211

Nevertheless, the Clinton team backed off characterizations of terrorism as warfare. Again, the experience in Mogadishu suppressed proposals for direct action, in turn making the use of special operators for CT unappealing. In a study conducted shortly after 9/11, historian Richard Shultz noted that officials in the Clinton administration linked the catastrophe in Mogadishu to Desert One—the “last time the Democrats had held the White House.” He noted that the Joint Staff was pleased the White House was so gun-shy about SOF after Somalia. He quotes a “Pentagon officer” explaining the Joint

Staff “didn’t want to put special ops troops on the ground. They hadn’t wanted to go into

Somalia to begin with. The Joint Staff was the biggest foot-dragger on all of this counterterrorism business." Shultz also chronicled the hesitancy of civilians in the White

House and officers on the Joint Staff for DoD to focus on CT at all. Michael Sheehan,

Clinton’s Coordinator for Counterterrorism at the State Department, told Shultz that civilians and the Joint Staff alike minimized DoD’s role in CT: “Terrorism was seen as a distraction that was the CIA’s job, even though DOD personnel were being hit by terrorists. The Pentagon way to treat terrorism against Pentagon assets abroad was to cast it as a force protection issue.”212

Nevertheless, as intelligence reports about al Qaeda flowed in, President Clinton himself increasingly asked about the use of SOF against transnational terrorist

211 U.S. Department of Defense, Annual Report to the Congress, FY1989 (Washington, DC: February 1988), https://history.defense.gov/Portals/70/Documents/annual_reports/1989_DoD_AR.pdf?ver=2014-06-24-151548-313. 212 Shultz, Jr., “Showstoppers.” 93

organizations, including al Qaeda.213 A Counterterrorism and Security Group (CSG) was established at the NSC, and in a series of Presidential Decision Directives the administration pressed for more proactive counterterrorism measures.214 Secretary Perry, meanwhile, had asked former SOCOM commander Wayne Downing to chair a commission study on terrorism in the wake of the Khobar Towers attack. Downing characterized the attacks on U.S. targets as part of a campaign of asymmetric warfare and urged DoD to prepare kinetic options to respond.215

Given how hard civilians were leaning into direct action against al Qaeda in formal policy statements, it is striking that the only military action taken against the group were air strikes in Sudan and Afghanistan.216 According to Clinton’s Senior

Director for Counterterrorism at the NSC, Richard Clarke, the Joint Staff representatives at White House meetings “had an answer that they used whenever asked to do something that they did not want to do” that included warnings of the size of the operation, the risks, and their “professional military opinion” that whatever it was—in this case, a SEAL snatch-and-grab operation to capture bin Laden—was a bad idea.217 The Pentagon-based military represented by the Joint Staff clearly still feared that small ground deployments,

213 Ibid. 214 The White House, Presidential Decision Directive 39: U.S. Policy on Counterterrorism (Washington, DC: The White House, 1995), https://www.hsdl.org/?abstract&did=462942; and The White House, Presidential Decision Directive/NSC-62: Protection Against Unconventional Threats to the Homeland and Americans Overseas (Washington, DC: 1998), https://www.hsdl.org/?abstract&did=758094. 215 Shultz, Jr., “Showstoppers.” 216 Raphael F. Perl, Terrorism: U.S. Response to Bombings in Kenya and Tanzania: A New Policy Direction?, CRS- 98-733 F (Washington, DC: Congressional Research Service, 1998), https://digital.library.unt.edu/ark:/67531/metacrs603/. 217 Richard A. Clarke, Against All Enemies: Inside America’s War on Terror (New York, NY: Free Press, 2004), 143. 94

even those designed to be clandestine and exfiltrated immediately, risked greater military involvement in ambiguous environments.218

Regardless, it was during this period of low demand for direct action but high COCOM demand for SOF presence in their operational regions when SOF became a repository of region- specific expertise and strategic narratives. JSOC began to conduct “area familiarization” trips to countries where SOF deployments seemed likely. Also known as “advance force operations” or

AFOs, these missions would scout embassies and airports and other transportation hubs. They gathered massive quantities of data about innumerable places and potential targets, relying also on info from the NSA, CIA, and State Department.219 “One-third of the command’s [JSOC’s]

172 staffers work on intelligence analysis” the New York Times reported.220

These activities were echoed in strategic-level documents. The aforementioned NSDD

207’s direction to the Secretary of Defense to consider prepositioning CT forces and the FY1988

DoD annual report’s discussion of LIC-type operations both provided hints of the opening SOF would later exploit, and had perhaps pried open themselves:

Our specific role is to work with the other appropriate U.S. government agencies and host country organizations, as necessary, to integrate our effort into a comprehensive strategy to combat the insurgency when that is indicated, and, where possible, identify at an early stage those conditions that foster insurgency. [Emphasis added].221

This bias toward steady-state forward-stationing became a way of life for SOF over the course of the 1990s. Working with the regional combatant commands, SOF were the vanguard

218 William Wechsler, interview with the author, October 2, 2018. Wechsler was a speechwriter for General John Shalikashvili when he was chairman of the Joint Chiefs of Staff and a director for counterterrorism at the National Security Council during the Clinton administration. He attributed the cited assertion to broader military sentiment, not Shalikashvili himself. 219 Naylor, Relentless Strike, xii; and Moyar, Oppose Any Foe, 241. 220 Emerson, Secret Warriors. 221 DoD, Annual Report to the Congress, FY1988. 95

element to shape post-conflict, post-Cold War, and post-crisis U.S. relationships around the world. As the journalist Dana Priest explains in her book, The Mission,

When the Soviet Union collapsed, small special forces teams quickly set up in Russia, Poland, and Hungary to develop new military relationships there. A dozen years later, members of an A-team trained to teach de-mining became the first American troops in nearly sixty years to resume military relations with the Vietnamese army. In Africa alone, the 3rd Special Forces Group has taught light- infantry and other military tactics to troops in [22 countries].222

SOCOM, with its institutional permanence and budgetary power, had become a repository of information and policy options. Already a popular force provider among policymakers and

COCOMS alike, their high deployment tempo under “peacetime” conditions gave them global reach and insight that drove the perpetual motion of the SOF enterprise.

JSOC, still a secretive direct action-oriented organization with a far narrower and less-in- demand skill set, had to advertise their services to get work. They demonstrated their capabilities in impressive site visits and role-playing scenarios and “adopted” interested civilians, educating them about SOF—not unlike the way SOF advocates had sought out Noel Koch and members of

Congress in the early 1980s. In interviews, civilian officials reported their experiences on visits to Fort Bragg and the JSOC compound, witnessing high altitude jumps, marksmanship, and hostage raids designed to impress policymakers.223

The lobbying worked—to a point. Although publicly available documents, including memoirs and reportage, state that JSOC was never sent to target terrorists during the Clinton administration, they did get a little exercise. As the wars in Bosnia and later Kosovo concluded,

JSOC teams were deployed to use their unique man-hunting skill set to capture internationally

222 Priest, The Mission, 187. 223 Joseph Collins, interview with the author, September 12, 2018; Interview with anonymous former senior OSD official. 96

wanted war criminals and other counter narcotics-related man-hunting missions have been referred to in press accounts.224

On the eve of 2001, the Clinton admin released their last NSSR, which continued to list

SOF among the range of military capabilities: “Our ability to respond to the full spectrum of threats requires that we have the best- trained, best-equipped, most effective armed forces in the world. Our strategy requires that we have highly capable ground, air, naval, special operations, and space forces.”225 Rather than LIC, the strategy speaks of “smaller scale contingencies” which again mix conventional and unconventional missions:

. . . military operations short of major theater warfare, including peacekeeping operations, enforcing embargoes and no-fly zones, evacuating U.S. citizens, reinforcing key allies, neutralizing NBC weapons facilities, supporting counterdrug operations, protecting freedom of navigation in international waters, providing disaster relief and humanitarian assistance, coping with mass migration, and engaging in information operations.226

By the end of the Clinton Administration the main focus of military policy was on stabilizing countries and regions, not on countering terrorism. SOF were the managers of the lesser- included-case, while the GPF were in charge of the wars national security planners anticipated as likely and used to scale the entire force itself.227 But civilians in the executive branch had rebounded from the post-Mogadishu rejection of SOF’s direct action capabilities, even if the use of SOF’s skills in capturing high value targets remained confined to non-counterterrorism contexts. SOF had achieved a lasting market share in terms of its size and mission scope, and

224 Naylor, Relentless Strike, 63-71; and Priest, The Mission, 214. 225 U.S. National Security Council, A National Security Strategy for a Global Age (Washington, DC: The White House, 2000), https://www.hsdl.org/?abstract&did=460717. 226 The 1997 QDR also used the phrase “smaller scale contingencies,” apparently as a replacement for the concept of “lesser included cases.” Otherwise, that strategy expressed continuity in both SOF missions and force structure, save for some reductions in reserve units. DoD, Report of the Quadrennial Defense Review, 1997. 227 Interview with anonymous former senior OSD official. 97

that share had translated into a stable civilian preference to reach for SOF in a variety of situations.

Period III: Counterterrorism Monopoly (2001-2010)

In 2001, civilians in the George W. Bush administration came into office with an interest in making the general-purpose force more like SOF, and to use SOF to support conventional operations. But the whole military, civilian administration officials believed, should be smaller, more agile, and lethal. The administration’s 2001 QDR, its first major statement on defense policy which was largely completed before the terrorist attacks of 9/11 but published shortly thereafter, articulated its preference for the use of SOF as an “immediately employable supplement to forward forces” that would “achieve a deterrent effect in peacetime,” and be a key element of “swiftly defeating an adversary’s military” in wartime.228 Once again, civilians were seeing SOF as an adjunct to the conventional force; only this time, they wanted the conventional force to look more like SOF.

By 2012, civilians preferred to use SOF as the lead in the nation’s counterterrorism campaign. This preference evolved during the Bush administration and held into the first term of

Barack Obama’s presidency. In its 2006 QDR, the George W. Bush administration announced that the GPF would take on some of SOF’s tasks to free it up for “long-duration, indirect, and clandestine operations.” The 2006 QDR also associated “direct action” with SOF for the first time, specifying that such strikes were for the purposes of counterterrorism and against so-called

High Value Targets (HVTs) or individuals with a strategic role in the terrorist threat to the

United States. The 2010 QDR emphasized SOF’s role in counterinsurgency, stability operations,

228 U.S. Department of Defense, Quadrennial Defense Review Report (Washington, DC: September 30, 2001), 26, https://archive.defense.gov/pubs/qdr2001.pdf. 98

and counterterrorism—although it included the GPF in such activities as well. Both the A-Team side of SOF and the direct-action elements of the force were arrayed against terrorist targets in strategy and in practice, focusing SOCOM and JSOC alike for the same purpose.

The obvious reason for the shift in civilian preferences was the terrorist attacks on 9/11.

And indeed, they had the largest effect. But for SOF to become the lead element in the counterterrorism fight was not inevitable and was not even what the W. Bush administration initially preferred. Indeed, the administration’s pre-9/11 preference as expressed in its first QDR lagged until well into 2002. It was SOF’s—particularly JSOC’s—volunteerism and the conventional force’s hesitancy that persuaded civilian leaders to think of SOF as the counterterrorism force. This advocacy, especially by JSOC, built on years of groundwork laid by the SOF community in educating civilians about special operations and developing intelligence capacity and other direct action-enabling skills and knowledge. And once SOCOM and JSOC were assigned as leads in the War on Terror—and, moreover, began to show tactical and operational successes in Iraq—the bureaucratic, budgetary, and operational momentum carried into the Obama administration. A military preference to focus SOF on counterterrorism capabilities that was born in a small corner of the Army in 1977 and grew after the Desert One imbroglio had finally matured into SOF’s monopoly over CT.

Details

President George W. Bush entered office with a vision for the U.S. armed forces that challenged both SOF and the GPF to be more like each other. His first Secretary of

Defense Donald Rumsfeld stated in his memoir, Known and Unknown that “the President had given [him] explicit guidance to make the Defense Department ‘lethal, light and

mobile.’”229 In a campaign speech at the Citadel, a major military college in South

Carolina, candidate Bush had stated his belief that,

Our forces in the next century must be agile, lethal, readily deployable and require a minimum of logistical support. We must be able to project our power over long distances . . . Our heavy forces must be lighter. Our light forces must be more lethal. All must be easier to deploy. And these forces must be organized in smaller, more agile formations, rather than cumbersome divisions.230

It was a summary of the ideas in a new paradigm for DoD called “transformation.” Rumsfeld was one of transformation’s apostles.231 He re-entered the building (it was his second tour as

Secretary of Defense) with both Bush’s marching orders and his own notions that things such as

“space and high-tech intelligence were the centerpieces . . . for transforming the U.S. military.”232 Thus, while SOF were not irrelevant to the new civilian leadership’s strategic vision, they were on the periphery. And the fact that Rumsfeld essentially wanted the entire force to be more SOF-like threatened SOF’s uniqueness. It was even rumored that Rumsfeld was considering demoting the rank of the ASD for SOLIC to reflect his disinterest in both LIC and special operations.233 Meanwhile, he rode roughshod over the Joint Staff and the services, with a famously brusque and dismissive style conveying an initial unwillingness to bend to their preferences on much of anything.234

Learning of the new boss’s skepticism, JSOC swung into gear. Journalist Sean Naylor chronicled how Rumsfeld “fell for” JSOC, in a series of engagements with its personnel that

229 Donald H. Rumsfeld, Known and Unknown: A Memoir (New York, NY: Penguin Group, 2011), 294. 230 Ibid; and quote from George W. Bush, “A Period of Consequences,” (speech, The Citadel, SC, September 23, 1999). 231 In an essay for Foreign Affairs, Rumsfeld himself chronicled his advocacy of transformation, explaining it as “new ways of thinking and new ways of fighting.” Donald H. Rumsfeld, “Transforming the Military,” Foreign Affairs, May/June 2002, https://www.foreignaffairs.com/articles/2002-05-01/transforming-military. 232 Priest, The Mission, 27. 233 Joseph Collins, interview with the author, September 12, 2018. 234 Priest, The Mission, 23-24. 100

book-ended 9/11. Before the attacks, in March of 2001, Rumsfeld and Wolfowitz took an

“introductory briefing” on the command from its commander, General Dell Dailey, and a few others. Scheduled to last under a half hour, it stretched out to 120 minutes.235 When the 9/11 attacks occurred, the effectiveness of JSOC’s lobbying was evident in the Pentagon’s initial memo to the President about the military response to the attacks. In that document, “Special

Operations Forces” were cited as the key capability for achieving the “strategic theme” of

“aiding local peoples to rid themselves of terrorists.”236 Then, in November of 2001, Rumsfeld visited the JSOC compound at Fort Bragg. Delta and SEAL Team 6 showed off their ambush, hostage rescue, and high-altitude parachuting skills.237

But Naylor also reveals that JSOC’s success meant that Rumsfeld emphasized direct action over—or even at the expense of—SOF’s other capabilities. “Despite—of perhaps because of—his repeated exposure to briefings on the high-end counterterrorism that was JSOC’s forte,”

Naylor writes, “Rumsfeld’s understanding of special operations remained superficial and unbalanced. He did not recognize the value of unconventional warfare and foreign internal defense . . . To Rumsfeld, the value of special operations lay only in the spooky and lethal activities JSOC exemplified.”238

Naylor partially substantiates this claim by citing Rumsfeld’s own memoir, in which the former Secretary of Defense wrote, “Since 2001, I had made a priority of increasing the size, capabilities, equipment, and authorities of the special operations forces . . . We shifted some of the tasks that Special Forces had historically been responsible for, such as training foreign

235 Naylor, Relentless Strike, 160. 236 Donald Rumsfeld, “Memorandum for the President: Strategic Thoughts.” Office of the Secretary of Defense: September 30, 2001, http://library.rumsfeld.com/doclib/sp/272/2001-09- 30%20to%20President%20Bush%20re%20Strategic%20Thoughts.pdf. 237 Naylor, Relentless Strike, 161. 238 Ibid., 162. 101

militaries, to allow regular forces to do them as well. This freed up special operators for more upper-tier tasks—reconnaissance and direct-action missions.”239

Naylor’s own narrative suggests that Rumsfeld’s love of JSOC was partly a meeting of the minds, and partly an abdication on the part of General Charles Holland, then-Commander of

SOCOM. Rumsfeld originally approached Holland to lead the global war that he was envisioning, and Holland balked.240 JSOC was only too happy to step into the void. Thus, SOF to

Rumsfeld became everything JSOC did and nothing it did not. Journalist Thomas Ricks, in his blistering book on the subsequent war in Iraq, quoted multiple mid-grade officers and enlisted SF soldiers complaining that SF were being mis-used in Iraq exactly along those Rumsfeldian lines.

One such officer “criticized the emphasis on raids and other direct action missions, which he felt came at the expense of the training mission . . . ‘We have become locked on kill or capture as a mission statement.’”241 Given how FID-focused much of that first “Strategic Thoughts” memo to

President Bush was, this is a remarkable case of how intra-military bureaucratic politics can affect the preferences of civilian leaders.

Rumsfeld was not the only civilian in the Pentagon to prefer a SOF-heavy approach to warfare more generally. Rumsfeld states that his deputy at DoD, Paul Wolfowitz, was the one who advocated the use of SOF in Afghanistan: “Wolfowitz also suggested that wherever we struck first, our special forces should be a part of the military strategy. He had been impressed by the use of special forces to locate and destroy Iraqi Scud missiles during the 1991 Gulf War.”242

As Naylor notes, Wolfowitz had also been in the JSOC briefing the previous March. Joe Collins,

239 Rumsfeld, Known and Unknown, 654. 240 Naylor, Relentless Strike, 158-159. 241 Thomas E. Ricks, Fiasco: The American Military Adventure in Iraq (New York, NY: The Penguin Press, 2006), 368. 242 Rumsfeld, Known and Unknown, 359. 102

the DASD for Stability Operations from 2001 to 2003, recalled a 1991 trip that then-

Undersecretary for Policy Wolfowitz took to JSOC. During the trip, which began with a jet flight from the DC area to Fort Bragg, JSOC personnel demonstrated their high-altitude parachuting capabilities by actually opening the back hatch of the aircraft and jumping out. The trip sounds remarkably similar to the demonstrations described by Sean Naylor that Rumsfeld himself witnessed some ten years later.243 Rumsfeld also cites a memo that Wolfowitz wrote to him shortly after the 9/11 attacks: “We should consider using those [Special Forces] as a kind of armed liaison with anti-Al-Qaida [sic] or anti-Taliban elements in Afghanistan.” Rumsfeld continues, “We believed our special operations forces could establish links with potential allies in Afghanistan, providing us with better intelligence and demonstrating that we were willing to help those who helped us.”244

Rumsfeld also likely turned to the JSOC arm of SOF for reasons of bureaucratic competition. In the hours after the 9/11 attacks, the Counterterrorist Center at the CIA put together a plan which its head, Cofer Black, presented to President Bush on September 14. The plan called for a CIA paramilitary team-led assault on al Qaeda and the Taliban in Afghanistan, with special operators joining but decidedly not leading the mission.245 A key feature of the plan was its speed, in contrast to DoD’s standard multi-month build-up of forces and logistics. Bush approved the plan. The decision would result in greater collaboration between the CIA and SOF in the years that followed but would also highlight the differences between SOF and the CIA in terms of precision, scale, how the use of force is authorized, and the nature of counterterrorism

243 Joseph Collins, interview with the author, September 12, 2018. 244 Rumsfeld, Known and Unknown, 360. 245 Jacobsen, Surprise Kill Vanish, 343-344. 103

itself.246 For immediate purposes after 9/11, the CIA became SOF’s competition to lead the War on Terrorism.

But SOF had been laying the groundwork for such a competition for years. Germane to competition with the CIA, another feature of SOF that appealed to Rumsfeld, was SOF’s capacity for intelligence-gathering. It was a capability, as explored above, that SOCOM had been growing for over a decade. According to Douglas J. Feith, Rumsfeld’s first Under Secretary for

Policy,

When Rumsfeld complained about risk aversion, he was referring (among other things) to the lack of proposals to use ground forces. The intelligence needed to fight terrorists could not be obtained solely, or even mainly, by satellite. We needed human intelligence . . . [VCJCS] Pace and I used the CAPCOM—the Campaign Planning Committee we cochaired—to connect with SOCOM officers on how to deploy Special Operations Forces, not just for “direct action” against known terrorist targets but also to obtain intelligence.247

The Bush administration would demonstrate its preference to use SOF as an intelligence and targeting device again in Iraq—remarkably, a preference that at least some conventional commanders had come to share. In his memoir, President Bush recalls that the Commanding

General for the 2003 Iraq invasion, Tommy Franks, recommended “multiple, highly skilled

Special Operations Forces identifying targets for precision-guided munitions [so that] we will need fewer conventional ground forces.”248

246 The overlap between CIA paramilitary and DoD special operations missions and capabilities would become a defining feature of the post-9/11 global counterterrorism campaign. Much of the man-hunting approach the CIA would adopt was based on previously developed SOF doctrine, and the two organizations would work together extensively while also competing for missions and authorities. See ibid.; Moyar, Oppose Any Foe; and Christopher J. Fuller, See It/Shoot It: The Secret History of the CIA’s Lethal Drone Program (New Haven, CT: Yale University Press, 2017). 247 Douglas J. Feith, War and Decision: Inside the Pentagon at the Dawn of the War on Terrorism (New York, NY: HarperCollins, 2008), 114. 248 George W. Bush, Decision Points (New York, NY: Random House, 2010), 233-234. 104

Franks appears to have understood that Rumsfeld, under the direction of the president, wanted to see force used in rapid, light, precise packages. Although his initial preference was for the entire military to transform in that direction, Rumsfeld ended by relying on SOF to be the type of force he wanted doing the major military missions at the time. Rumsfeld’s preference for

SOF obtained because SOF offered to fulfill the Secretary’s and the President’s wishes.

So, civilians’ initial preferences, coupled with 9/11 and JSOC’s lobbying produced a civilian preference to use SOF as the lead kinetic element in major operations and counterterrorism. Moreover, SOF’s decade of AFOs and circulating in combatant command areas of operation gave them expertise and credibility about those regions with civilian officials.

Going into unusual detail for a civilian memoir, Rumsfeld laid out the reasoning behind his preference for relying so heavily on SOF for the invasion of Afghanistan:

The Army’s Special Forces, the Navy’s SEALs, and the Air Force’s combat controllers had not been previously entrusted with the lead in such a major mission. The few hundred men who were ready to risk their lives in the service of their country . . . were among the most highly trained, best equipped, and most experienced soldiers on the face of the earth. Some were fluent in the local languages and versed in the cultures they would be encountering. They had trained foreign militaries and understood how to get along with those who thought and fought differently. They were experts in the irregular guerrilla warfare that would be critical to success. They were trained in demolition, hand-to-hand combat, and mountain and desert warfare. American special operators would be the sharp tip of the spear in the first war of the twenty-first century.249

The waging of the wars in Afghanistan and Iraq reinforced civilian preferences about how to use SOF, partly in path-dependent ways, and partly because SOF kinetic actions delivered the rare moments that felt like victory. In his memoir, President Bush refers to JSOC killing Abu Muqab al Zarqawi, the leader of al-Qaeda in Iraq, as “a bright spot” during the

249 Rumsfeld, Known and Unknown, 377-378. 105

otherwise dire news of 2006, “a dramatic sign of progress.”250 Robert Gates, Bush’s second

Secretary of Defense, also noted JSOC’s “remarkably successful” operations that also “played a major role in the success of the surge in Iraq and the counterinsurgency in Afghanistan.”251

As for the path-dependency, that occurred as a consequence of SOF defining the operational terms of the War on Terrorism and then driving activity across the foreign policy enterprise. They did so by framing the periodic policy debates about the best approaches in both wars: SOF-executed direct action counterterrorism missions were seen as a necessary complement to counterinsurgency, but it was also seen as a more viable alternative to counterinsurgency, especially in Afghanistan as the war dragged on late into the Obama administration.252 In turn, SOF advocates framed debates by being present in every major decision and alternative-proposal node. During the Iraq war, SOF led the military colonization of the interagency, putting liaison officers in every intelligence element they could find and, in turn, borrowing analysts from civilian agencies for service at SOCOM.253 During the early years of the Iraq war, SOF’s “outreach” was largely on the initiative of Lieutenant General Stanley

McChrystal, whose theory of “fighting a network with a network” operationalized SOF-led interagency coordination on the ground in both Iraq, when McChrystal was JSOC commander, and in Afghanistan, where McChrystal became Commander of the overall war effort.254 Part of

McChrystal’s theory was that the U.S. needed to act faster than the enemy could process information and launch the next attack. This demanded not only exquisite intelligence but rapid

250 Bush, Decision Points, 365. 251 Robert M. Gates, Duty: Memoirs of a Secretary at War (New York, NY: Random House, 2014), 254. 252 Bob Woodward, Obama’s Wars (New York, NY: Simon and Schuster, 2010), 236. 253 Christopher J. Lamb and Evan Munsing, Secret Weapon: High-Value Target Teams as Organizational Innovation (Washington, DC: Institute for National Strategic Studies, National Defense University Press, 2011), https://inss.ndu.edu/Portals/68/Documents/stratperspective/inss/Strategic-Perspectives-4.pdf; Stanley McChrystal, My Share of the Task: A Memoir (New York, NY: Random House, 2013). 254 McChrystal, My Share of the Task. 106

communication between those gathering information and those leading missions. This meant

JSOC’s target for interagency cooptation was the intelligence community, the long-term SOF partner-and-rival.255 It also meant that operations moved so swiftly that civilian leaders on the ground and back in Washington were in the position of consumer not only of intelligence but of the course of the war itself. One “senior civilian official” told journalist Bob Woodward, “Every single night [in Afghanistan] they [JSOC] are banging on these guys with a pace and fury that is pretty impressive.”256

Echoing the pattern of the first period in this study, the SOF community also counted on retired officers to persuade civilian institutions of their value. No figure better embodied SOF’s institutional influence on civilian preference formation than Michael Vickers. The ASD for

SOLIC from 2007 to 2011, spanning the Bush and Obama administrations, Vickers had been a ten-year Special Forces officer turned CIA analyst. He fought in Grenada—although, since he transitioned into the Agency in 1983, it is unclear whether with SF or the CIA.257 He was best known as the Agency’s lead for the program that armed the mujahedeen in Afghanistan to fight the Soviet occupation.258

Now, as ASD(SOLIC), Vickers was a firm believer in SOF’s potential to be the lead element in the post-9/11 wars. Before being nominated by the W. Bush administration to fill the

ASD role, he wrote a 2004 op-ed arguing that the U.S. force presence in Iraq should be smaller rather than larger, and should focus on training the Iraqis and supporting special operations.259

255 Lamb and Munsing, Secret Weapon; and Robinson et al., Improving the Understanding of Special Operations, 103. 256 Woodward, 2002, 316. 257 Ann Scott Tyson, “Sorry Charlie. This is Michael Vickers’s War.” Washington Post, December 28, 2007, https://www.washingtonpost.com/wp-dyn/content/article/2007/12/27/AR2007122702116.html. 258 George Crile, Charlie Wilson’s War: The Extraordinary Story of how the Wildest Man in Congress and a Rogue CIA Agent Changed the History of Our Times (New York, NY: Grove Press, 2003). 259 Michael Vickers, “For Guidance on Iraq, Look to Afghanistan: Use Fewer U.S. Troops, Not More,” USA Today, June 27, 2004. 107

Once he took the helm at SOLIC, Vickers fought to expand the size of SOF dramatically, and built “a global counterterrorist network” of over 40 countries with U.S. SOF training partner forces in SOF-like capabilities.260 President Bush writes that during a 2006 Camp David retreat to consider outside opinions on the conduct of the war in Iraq, Mike Vickers recommended a larger role for SOF.261 Sounding very much like a 1980s LIC enthusiast, he once told the

Washington Post, “The war on terror is fundamentally an indirect war . . . It’s a war of partners . . . but it also is a bit of the war in the shadows, either because of political sensitivity or the problem of finding terrorists.”262

Vickers’ influence was evident in the QDR and the NSS in 2006, both major reflections of the changes to civilian preferences about the role of SOF:

As general purpose joint ground forces take on tasks that Special Operations Forces (SOF) currently perform, SOF will increase their capacity to perform more demanding and specialized tasks, especially long-duration, indirect and clandestine operations in politically sensitive environments and denied areas. For direct action, they will possess an expanded organic ability to locate, tag and track dangerous individuals and other high-value targets globally.263

The Department of Defense also is expanding Special Operations Forces and investing in advanced conventional capabilities to help win the long war against terrorist extremists and to help dissuade any hostile military competitor from challenging the United States, its allies, and partners.264

Moreover, SOF-related missions came to dominate the “war of partners” Vickers had outlined, a major element in Gates’ 2008 National Defense Strategy:

260 Thom Shanker, “A Secret Warrior Leaves the Pentagon as Quietly as He Entered,” New York Times, May 1, 2015, https://www.nytimes.com/2015/05/02/us/a-secret-warrior-leaves-the-pentagon-as-quietly-as-he-entered.html; and Tyson, “Sorry Charlie.” 261 Bush, Decision Points, 363-364. 262 Tyson, “Sorry Charlie.” 263 U.S. Department of Defense, Quadrennial Defense Review Report (Washington, DC: February 2006), 44, https://archive.defense.gov/pubs/pdfs/QDR20060203.pdf. 264 U.S. National Security Council, The National Security Strategy of the United States of America (Washington, DC: The White House, 2006), https://www.comw.org/qdr/fulltext/nss2006.pdf. 108

Insurgent groups and other non-state actors frequently exploit local geographical, political, or social conditions to establish safe havens from which they can operate with impunity. Ungoverned, under-governed, misgoverned, and contested areas offer fertile ground for such groups to exploit the gaps in governance capacity of local regimes to undermine local stability and regional security. Addressing this problem will require local partnerships and creative approaches to deny extremists the opportunity to gain footholds.265

In 2009, when the incoming Obama administration asked both Michael Vickers and

Secretary of Defense Bob Gates to stay on at the Pentagon, the two men formed part of the path dependency in SOF’s predominance over civilian preference formation for its own use. They were the continuity in civilian expertise on and management of SOF, not to mention the two ongoing wars in which SOF and JSOC in particular played a leading role.

Preferences among the Obama team at the White House, however, did not start—and did not all end—with SOF in the lead. Gates himself notes in his memoirs that he, Under Secretary of Defense Michèle Flournoy, and Vickers “often had to push back hard to keep the White

House and State Department from getting too far into our military knickers” overall.266 But pushback from the White House was more in the category of oversight, of restraining the scale of

SOF activities rather than their scope or purpose.267

In fact, a significant faction within the White House not only embraced SOF’s direct action CT focus, they championed it. During the course of debating the strategy in Afghanistan in 2010, Vice President Joe Biden and his national security team reportedly promoted an approach they dubbed “Counterterrorism Plus,” whereby the U.S. dramatically reduced its

265 Office of the Secretary of Defense, 2008 National Defense Strategy (Washington, DC: Department of Defense, 2008), 2-3, https://archive.defense.gov/pubs/2008NationalDefenseStrategy.pdf. 266 Gates, Duty, 452. 267 The target selection and approval process for CT strikes outside areas of major hostilities is an example of the administration’s continued use of SOF as the main military element of the CT strategy with additional controls on target approval. See: https://obamawhitehouse.archives.gov/the-press-office/2013/05/23/fact-sheet-us-policy- standards-and-procedures-use-force-counterterrorism. 109

footprint to mainly SOF elements that would conduct raids to maintain pressure on al Qaeda.268

Although Obama ultimately rejected Biden’s proposal for Afghanistan, SOF’s identity as a counterterrorism force was clearly cemented in civilians’ minds.

Viewing SOF as a potential main element in what candidate Obama had called “a war that we have to win,” was a remarkable turnaround in civilian preferences for SOF overall.269 By the time the 2010 NSSR was released, the predominance of SOF missions was firmly established. Whereas in previous years, CT and stability operations were “lesser-included cases,” now conventional military operations had become the secondary concern. “We will continue to rebalance our military capabilities to excel at counterterrorism, counterinsurgency, stability operations, and meeting increasingly sophisticated security threats, while ensuring our force is ready to address the full range of military operations.”270 The 2010 QDR also championed another SOF expansion and emphasized special operators’ expertise in CT, counterinsurgency, and stability operations.

Findings and Analysis

SOF capabilities are a classic case of a solution in search of a problem.271 Long before the attacks on 9/11, SOF and its advocates provided a ready frame for civilians’ eventual preference that SOF take over the Global War on Terrorism. Once it had done so and had provided civilian managers with a series of successes, SOF and counterterrorism became synonymous.

268 Woodward, Obama’s Wars, 159-160. 269 Mark Landler, “The Afghan War and the Evolution of Obama,” New York Times, January 1, 2017, https://www.nytimes.com/2017/01/01/world/asia/obama-afghanistan-war.html. 270 U.S. National Security Council, The National Security Strategy of the United States of America (Washington, DC: The White House, 2010), http://nssarchive.us/national-security-strategy-2010/. 271 Cohen, March, and Olsen, “A Garbage Can Model of Organizational Choice.” 110

This outcome was not inevitable. For one thing, civilians long sided with the GPF in thinking of SOF as an “adjunct” to conventional operations. For another thing, when it came to counterterrorism capabilities, civilians had at least two alternatives to SOF. One was the intelligence community, in particular the Central Intelligence Agency’s paramilitary capabilities.

The CIA obviously provided the ability to gather intelligence vital to hunting terrorists and offered the secrecy and some of the lethal precision that SOF could. The other alternative to SOF was the conventional military, particularly the Marine Corps. To the extent that terrorists were affiliated with or supported by states, or behaved like regular paramilitary forces, the conventional military offered deterrence power and offensive combat capabilities; to the extent that terrorists could be wiped off the map with airstrikes, that was another attractive strategic solution. Civilians entertained and experimented with both of these alternatives and even gave parts of the counterterrorism campaign to them. But in the end, SOF became preeminent. Why?

This case study shows an organized sub-group of the military succeeding in persuading civilians to adopt its preferences in part because those civilians were ambiguous as to their own preferences. The SOF advocacy community successfully persuaded civilians to adopt three complementary preferences for the uses of SOF over time. First, a core constituency of SOF always wanted to emphasize the counterterrorism mission as SOF’s key purpose. This constituency first appeared in the Army and shortly thereafter was housed in the Joint Special

Operations Command (JSOC). In the context of the late 1970s and early 1980s, however, SOF as a whole had a second, vital preference: institutional survival. Neglected to the point of near- dissolution, SOF determined that it needed autonomy from the military services.

Yet mere autonomy for SOF’s own sake was not reason enough to attract outside or even inside sponsorship. So, in the expression of their second preference, SOF made the case for its

111

survival, over and over, by pointing to the array of things they could do that no other part of the military could—by arguing that they were a solution to a variety of problems. Some problems civilians had perceived, but most of them they had not. During the 1980s, the most convenient hook for justifying the need for SOF’s unique capabilities was the debate about the future prevalence of low-intensity conflict.

Focusing on LIC was SOF’s third preference. Coupling the assertion of LIC’s future relevance with the abundant evidence that the military services—and DoD writ large—neglected

SOF allowed them also to protect their direct action and counterterrorism capabilities. While basic survival prompted SOF to attempt to anticipate civilian preferences, the proactive persuasion of civilians that LIC would become a key feature of the near-term international environment was a case of military persuasion of civilians with ambiguous preferences. In such a case, because civilians did not have a consolidated preference about LIC, they were open to the institution that most wanted the job.272

SOF leveraged new interest in LIC and the narrative of SOF neglect to persuade civilian stakeholders in the Pentagon to push for implementation of the SOF revitalization agenda and those on the Hill to legislate a permanent and powerful bureaucratic position for SOF within

DoD by creating a new combatant command for special operations. The 1990s were spent institutionalizing SOCOM, overcoming the Clinton administration’s Mogadishu-borne aversion

272 For example, the Marine Corps, with its expeditionary capabilities and experience deploying to “third world” contingencies was a plausible alternative to SOF. What is important in this case is that SOF put early emphasis on their LIC capabilities and the Marines increased their emphasis on it over time, perhaps losing out on being selected by civilians because they did not sell themselves soon or hard enough. As Kathleen Hicks pointed out to me, the “soon enough” part of the previous sentence may have especially mattered, as the Marines became overwhelmed by the demands of being a ground force supplement for Afghanistan and Iraq. See Wayne A. Clemmer for an initial articulation and then Marine Commandant General Krulak’s later argument about the “three block war.” Wayne A. Clemmer, Low-Intensity Conflicts and the United States Marine Corps (Carlisle, PA: U.S. Army War College, 1991); and Charles A. Krulak, “The Strategic Corporal: Leadership in the Three Block War,” Marines Magazine, January 1999, https://apps.dtic.mil/dtic/tr/fulltext/u2/a399413.pdf. 112

to ground force direct action, a new focus on Osama bin Laden’s terrorist network, and chasing war criminals in the Balkans (which turned out to be pretty good training for chasing terrorists elsewhere). All this time JSOC was honing its manhunting skills and selling them to civilian defense leaders every chance it got, designing a kind of marketing strategy it could later use to win competitions for mission assignments.273

SOF’s campaign to persuade civilians to adopt their preferences for the uses of SOF bore fruit after 9/11. When the al Qaeda attacks happened, civilians in DoD and the White House searched for tools that could satisfy their desire for a quick and lethal response. Then, for the invasion of Iraq in 2003 and the subsequent fight against insurgent groups, civilian leaders again needed unconventional military capabilities. In both cases, JSOC stepped in, pulling the rest of the SOF enterprise with it and over time making SOF synonymous with direct action and counterterrorism. As JSOC’s unofficial historian, the journalist Sean Naylor, put it, “JSOC had firmly established itself at the top of not only the U.S. military food chain, but also, arguably, the interagency hierarchy.”274

Importantly and counterintuitively, disagreement within the military actually contributed to the pressures that compelled civilians to adopt military preferences. The internal military debate about the purpose of SOF drove military partisans to seek out civilian allies to prevail in the dispute. Civilians then became trapped inside the strategic frameworks articulated by the military. To some degree it was the reverse of Barry Posen’s claim that civilians recruit “military mavericks” to help them impose civilian preferences on the military—instead, special operators

273 This finding is consistent with some other empirical work in the field. Eitan Shamir and Eyal Ben-Ari, in a survey of special operations forces in democratic states, found that “internal military entrepreneurs . . . promote and market the unique advantages of [special operations] forces to key military and civilian decision-makers.” Eitan Shamir and Eyal Ben-Ari, “The Rise of Special Operations Forces: Generalized Specialization, Boundary Spanning and Military Autonomy,” Journal of Strategic Studies 41, no. 3 (2018): 335-371, doi:10.1080/01402390.2016.1209656. 274 Naylor, Relentless Strike, 427. 113

leveraged the power of civilian mavericks to advance their preferred uses of SOF with legislators and policymakers.275 They then faced-off against another civil-military coalition among the services, OSD, and the White House. The pattern repeated once SOCOM settled into LIC and stability operations and the rest of the national security bureaucracies adopted this preference, and JSOC began its own liberation movement. Importantly, this meant that the alternatives to

SOF preferences were supplied by other military actors, leading civilians to adopt military preferences in any case.

This type of face-off between civil-military coalitions makes the categories of “military” and “civilian” less analytically useful at all levels of government. By the time an idea, which had become an alternative about the strategic purpose of SOF, reached the Secretary of Defense, the

White House, or the Congress, preferences had already cross-pollinated among military and civilian institutional actors for years. At the interagency (White House-State Department-DoD) and inter-branch (Executive-Legislative) levels, there were rarely cases of civilians or military preferences prevailing over each other. Instead, an array of civil-military coalitions faced off based on the ways substantive policy positions reflected joint institutional interests. Even when a controversy featured the Department of Defense and the White House on opposite sides, neither side could truly be called “military” or even purely “civilian.”

Thus, this case study challenges the notion that there is, at the levels of strategy and policy, a wholly civilian preference that prevails over a wholly military preference. Instead, consistent with Gibson and Snider’s argument, there are coalitions of individuals and organizations within DoD, across the interagency, and between the Executive branch and the

275 Posen, Sources of Military Doctrine. 114

Congress that promote shared preferences. One issue network’s preferences prevail, not a majority opinion held by civilians or “the” military.

And yet there were still identifiable phases when military preferences shaped civilian preferences instead of civilians forming preferences based on factors independent of the military—or, for that matter, civilians persuading the military to adopt their preferences. Rather, the overall pattern is of the military establishing the foundational preferences—the links between

SOF and both notional and real-world strategic challenges and events—that eventually become civilian preferences. The military, in this case the SOF community, acted on the basis of its own preferences to persuade civilians of the CT-SOF link over time. Civilians’ adoption of a military preference was important because it played a major role in creating JSOC, writing and passing the legislation that established SOCOM and SOLIC, and shaping the counterterrorism policies of the George W. Bush and Obama administrations. It took civil-military coalitions to codify the preference into policy, but the preference originated with the military. And preferences shape policy.

Testing the Findings: Counterfactuals

It may be that external threats or internal political pressures were what really affected how civilians formed—or adopted—preferences over the purposes of special operations forces.

We can test the effects of military preferences on civilian preferences by considering counterfactuals that remove or change these external and internal pressures.

First, suppose that there had been low external pressure. This would mean there was no

Soviet threat and no Cold War, and no state-sponsored or non-state terrorism. Where would civilians derive their preferences in that case, and would they come from military actors? Two

115

possibilities seem likely: First, as was the case prior to Operations Eagle Claw and Urgent Fury, civilian attention to special operations would have remained so minimal that they did not develop strong preferences. Second, they might have considered SOF at some point and concluded they were not needed. In this case, intra-military controversy about SOF might have existed, but civilians would not have felt moved either to develop a preference or to defer to military debates about the wisdom of different preferences. Regardless, there is little evidence that civilians use moderate pressure such as that of LIC to spend time considering alternative purposes for capabilities. Rather, crisis moments inspired civilian attention. But the urgency of crisis demands quick choices, and therefore the tendency to rely on pre-developed preferences as a heuristic.

Without external pressure that generates a need for immediate action, civilians can afford to have no or ambiguous preferences.

Now reinsert the external pressures of the Cold War and international terrorism but suppose that neither SOF nor the GPF attempted to persuade civilian policymakers of their views or reached out to form coalitions with civilian officials. It is hard to imagine that civilians would have developed SOCOM without the SOF Liberation Front’s activism. There would have been no advocacy within OSD or on the Hill, nor would there have been sources of information and persuasion for lawmakers and staff. Without JSOC selling itself and its direct-action capabilities,

Rumsfeld might have continued to rely on—and to attempt to reshape—the GPF for counterterrorism and interventions.

9/11 cast an especially long shadow. If the attacks on the World Trade Center and the

Pentagon had not occurred, would not JSOC’s desire to conduct more direct-action counterterrorism missions have remained a suppressed one, and the trends of the 1990s simply continued? Here the change in presidential administrations was also clearly salient. W. Bush’s

116

desire to make the GPF itself more like SOF would likely have continued as the dominant debate among military and civilian institutional actors. JSOC still would have conducted its charm offensive, possibly convincing Rumsfeld to maintain or even grow investments in SOF’s direct action capabilities. The 2001 QDR’s expressed preference that one of SOF’s main purposes was to conduct “covert deep insertion over great distances” shows civilian appreciation for SOF’s capacity for accepting risk and undertaking delicate operations. But then, again, this would seem to be a preference shaped by military actors.

Without the war in Iraq, McChrystal’s JSOC would have had no opportunity to expand

SOF’s interagency influence and intelligence activities. Indeed, Iraq, with its high tempo of direct-action counterterrorism missions that seemed for a time to deliver strategic success against the insurgency—suggesting a capacity for success against similar organizations throughout the world—may have been the most meaningful determinant of civilian preferences for the use of

SOF.

The problem exists when, in the course of defining its own purpose or the purpose of a capability, a military community plants valuing the mission in civilians’ minds, or forecloses options that civilians might have been interested in. If there were no 9/11, modern SOF would not be so direct action oriented today. But then, if there had been no JSOC, it is likely that modern counterterrorism would not have been as focused on direct action.

Conclusion

Military organizations do not just want to prevail over civilian decision-makers; they want to persuade them. The military is incentivized to indoctrinate civilian allies so they can be

117

relied upon over the long term to make choices consistent with military preferences. What better way to control policy than to shape civilians’ preferences directly?

This idealistic motivation swiftly encounters practical constraints. The path to persuasion is not free of obstacles, the first of which in the case of SOF was other parts of the military.

Consistent with the theoretical expectations of organizational culture, military factions had divergent preferences over the strategic purposes of SOF while civilians during the first two periods of the case found little political reason to form specific preferences about SOF. Military factions we able to recruit civilian allies, not only to persuade these civilians to rule in their favor, but also to help them survive and prevail in their intramural competitions. Civil-military coalitions thus formed as military actors co-opted civilians into parochial disagreements within the military.

This co-optation process was different for different civilian allies. The SOF case shows how new military capabilities are first debated and tested at lower levels of DoD’s bureaucracy where civilians are acculturated to share military preferences for institutional reasons. The first layer of civil-military coalition building happened almost invisibly within DoD and privileged military preferences simply because DoD is a military institution, especially at the working level.

As the debate over SOF capabilities gained salience, it made its way up the chain of seniority in

OSD, with civilians transmitting their preferences vertically to more-senior civilian policy makers. When resistance was encountered, it spurred the coalition to seek new allies outside of

DoD, across the interagency. Coalitions may eventually span bureaucracies and even branches of the government, as was the case for the SOF Liberation Front. That coalition was built around the preferences of military institutions but leveraged the structural power of civilians. Thus, in shaping congressional perspectives on LIC and the utility of SOF for addressing it, the SOF issue

118

network managed to inculcate powerful civilians with their preferences and see those preferences adopted as policy.

Throughout the case, purposeful SOF persuasive efforts were aided by information asymmetries that resulted from natural functional differentiation and civilian choices. Civilian inattention to SOF capabilities suppressed civilian expertise development, which was followed by civilian reliance on military sources of interpretation of the strategic demand for those capabilities. That military factions supplied ready-made preferences and justifications for them complemented the fact of low political pressure on civilians to develop their own specific preferences about SOF capabilities. Without either the independent information nor the motivation to cultivate it, civilians adopted military preferences. Civilians subsequently became proxies for military factions in debates over the uses of SOF.

What of the arguments that it is really the external environment that determines civilian preferences? Despite Posen’s assertion that external threats affect civilians and the military differently, the SOF case shows how military actors interpreted the international threat environment for civilians, especially in cases lacking immediate urgency from a civilian perspective. Military actors incorporated those interpretations into their justifications for (and against) special operations capabilities. In other words, most civilians viewed the threat environment through SOF’s or the conventional military’s lens. Military preferences mediated the international environment for key civilians. This was especially true during the debates about

LIC in the 1980s and the emerging approaches to counterterrorism in the years after 9/11.

How important was organizational culture? The SOF case underscores the degree to which military institutional cultures shapes information. The Nunn-Cohen Act, by bureaucratically elevating SOF and its civilian allies inside DoD to higher levels of decision-

119

making and a larger and more centralized budget share, helped shift SOF away from its reliance on pure persuasion and allowed it to use bargaining tools as well. Perversely, this may have actually contributed to the inertia of the 1990s, where neither military institutions nor civilians much changed their views of the strategic uses of SOF beyond the sobering Black Hawk Down incident. SOF was no longer at risk of institutional destruction, but their focus on bureaucratic tools and processes sapped the energy the issue network used toward persuasion in the late 1970s and first half of the 1980s.

Disunity within the military helped motivate competitive persuasion of civilian allies.

Civil-military coalitions then competed on the basis of preferences formed by rival military institutions. SOF was desperate to survive, and under its own terms, despite intra-military opposition to their preferences. Later on, JSOC was determined to increase the rate of direct action and its role in implementing it, despite lethargy on the part of SOCOM more broadly.

These disputes drove them to transfer their values, assumptions, and perceptions to powerful civilians in order to affect defense policy in their favor.

The SOF case shows the importance of information—and control over it—in shaping civilian preferences. Much of the civil-military relations literature focuses on situations where civilian and military preferences diverge. I have found situations where civilian preferences were suggestible, whether for reasons of information asymmetry, organizational culture, or political incentives, and military organizations were motivated to persuade civilians to adopt their preferences. It appears that under such conditions, civilians struggle to develop or sustain preferences independent from the military.

120

CHAPTER 3

UNMANNED AERIAL VEHICLES: 1952-2016

In this case study of the emergence of unmanned aerial vehicle (UAV) capabilities, I find that civilian information relative to the military explained civilian preferences over time. From

1952 to 2016 low levels of civilian attention, expertise, and access to information led most civilian elites to adopt military preferences for the uses of UAVs. Moreover, civilian deference to military preferences regarding UAV uses (and therefore development and acquisition goals) increased in wartime. However, civilians also eventually developed a preference independent from military framing and expertise when they sustained attention to UAVs and relied more on their own expertise as the political pressures of wartime faded. The final period of the case clearly demonstrates how civilian preferences can be influenced by the structure of the international environment, but also that they require informational assets free of military framing to do so.

From 1952 to 2000, civilians adopted military preferences for the uses of UAVs, but also held preferences about resourcing UAV capabilities based on their own interests in maintaining political power. Then, for the next decade, military preferences for the uses of UAVs dictated most civilians’ preferences, including by overcoming previous civilian preferences to suppress spending on the systems and enforce joint development programs. Military preferences thus played a major role in the formation of civilian preferences for sixty years. However, this same period, and especially the last few years of the case study, also show emerging civilian preferences for the uses of UAVs that were independent from the military’s.

For the first several decades of the case, military institutions held two principal, contradictory preferences: to use UAVs for reconnaissance purposes or not to use or develop

121

them at all. At the same time, civilian attention to UAVs was sporadic and largely taken up by their roles in espionage. Otherwise, civilian preferences regarding UAVs focused on the domestic political implications of their development rather than military uses. The few civilians who did develop preferences regarding the military uses of UAVs tended to adopt one or the other of the two military preferences. Although civilians also held a range of preferences about the ways UAVs should meet domestic political goals, those preferences always conformed to or complemented one or the other of the alternative preferences for UAV uses supplied by military actors.

Once a strong military advocacy community in the form of deployed forces coalesced around a preference to use UAVs for surveillance and strike missions in irregular warfare contexts, civilians adopted those warfighter preferences for the uses of UAVs, keeping latent an independent preference that had begun to form just prior to 9/11 to consider UAVs for strike missions in conventional warfare contexts. Only in the last few years of the case study period were civilians able to form and strongly hold a preference for UAV uses that was independent of the military’s: to use them for offensive combat purposes in the context of major war, rather than just passive reconnaissance missions against unconventional adversaries.

There were two main reasons for the influential role military preferences played in civilian preference formation. First was the low attention, on average, senior civilians paid to

UAVs during the first several decades of UAV development. Attention allocation is a choice, itself shaped by incentives. Costs, technological performance, and international crises could depress or inspire civilian attention. But as a routine matter, for much of the case period whether or not the U.S. needed UAVs, and what they needed them for, was left to the military to debate internally. By the time civilians at high levels applied their time and attention to UAV

122

development, there were long-held military ideas about what such capabilities should be used for, providing civilians ready-made rationales for UAV development and deployment. The final years of the case show how sustained, high-level civilian attention made civilian preferences for

UAVs more independent of military preferences.

Second whether the U.S. was at peace or at war as well as whether UAVs were in development or deployed to operating forces dictated the structural dynamics of UAV-related civil-military relations. During peacetime UAV development, the services and the Joint Staff wielded the most important military preferences and were motivated by their interests in preserving their shares of the Defense budget, successfully fielding weapons already in development, and protecting service culture. At the same time, the civilian institutions with the most salient preferences regarding capabilities were those involved in budgeting and acquisition

(the Deputy Secretary of Defense, the Office of the Under Secretary of Defense for Acquisition, and congressional budget committees). These civilians tended to be mindful of technological innovations and the costs, timelines, and adherence to the law during the weapons development process; they were less concerned by UAV’s strategic rationale.

Once UAVs were deployed to the operating force, both in peacetime and during wartime, the institutions whose preferences mattered for decisions shifted, and the services became relatively less important than the combatant commands, whose overriding incentive was to prepare for and prevail on the battlefield regardless of service interests or domestic politics.

Civilians in the White House and the policy side of the Pentagon (the Secretary of Defense and the Office of the Under Secretary of Defense for Policy) took over, while Congress assumed its operational and war oversight responsibilities.

123

Moving along the continuum from peace to operations to war made civilians progressively more deferential to military preferences. In peacetime, urgency regarding UAVs was minimal. In wartime, urgency was high, and civilians had political incentives to support (or appear to support) those fighting the war, lest they be blamed for any failures. Yet, even in the absence of war, the deployment of UAVs to operating forces in the Combatant Commands

(COCOMs) began to give military advice a more direct effect on civilian preferences because those forces were preparing for war should it occur. Shifts in the strategic environment that pushed UAVs into actual wartime use elevated the importance of warfighter preferences politically and bureaucratically, limiting civilians’ search for alternative information sources.

Once wartime conditions began to relax, civilians chose, once again, to entertain other factors in their preference formation.

This chapter proceeds as follows: Below I establish definitions and justify the case selection. I then provide an overview of the case history, including the pivotal institutional actors and the preferences they held over time. I then provide the detailed case study, which is divided into four phases: development, accelerant, deployment, and offset. I draw comparisons and contrasts among these processes, uncovering the different arrays of civil-military relations and the conditions that shaped the interactions of military and civilian preferences. I then offer the findings and analysis of the case. To test the findings, I provide a brief counterfactual exploration to reveal how different conditions may have affected the evolution of civilian preferences. I conclude with what the UAV case adds to the larger project.

124

Why UAVs? Definitions and Case Justification

UAVs are a clear case of an emerging capability, one in development for decades that only became persistently deployed beginning in the 1990s. UAVs have been developed by every military service and operated by all of the regional combatant commands, allowing examination of variation among the military components’ preferences, whether those preferences aggregated or not, and how civilian preferences interacted with them. The case begins in 1952 with one of the earliest post-WWII studies that examined the military uses of UAVs and ends in 2016, as the

U.S. was transitioning away from the war on terrorism. The case also allows examination of the relationship between civilian and military preferences for UAVs under a variety of domestic and international conditions, as it spans multiple environments and events, including peacetime and wartime, the Cold War, the post-Cold War period of irregular wars and humanitarian interventions, and the post-9/11 war on terrorism, as well as several re-organizations of DoD, including the changes associated with the Goldwater-Nichols Defense Reform Act of 1987.

The definition of UAVs is simultaneously straightforward and elusive, in part because what the Department of Defense (DoD) thinks of as a UAV has changed over the years, as have the most common names for them.276 Although once a concept so expansive that it covered almost any projectile with at least semi-controlled flight that did not have a pilot on board—a definition that included missiles—the UAV capabilities studied here are distinguished by their reusability, which is to say they are not intended to collide with a target but to land and be used

276 Readers of this case study will not encounter the word “drone” as a general reference to UAVs on my part because it conjures images of the counterterrorism drone program rather than suggesting the full range of uninhabited aerial systems employed by the U.S. military. Nevertheless, “drone” is a popular term for UAVs, including among serious analysts of unmanned aerial systems. For example, see this excellent paper on UAVs: Kelly Sayler, A World of Proliferated Drones: A Technology Primer (Washington, DC: Center for a New American Security, 2015), https://www.jstor.org/stable/resrep06394. 125

again.277 UAVs in this study are also not just aircraft but full capabilities provided by unmanned aerial systems (UAS), which comprise sensors, communications, information-gathering, and weapons. I will often refer to UAV systems or UAV capabilities. Further, UAVs can be flown remotely by a human pilot or autonomously, either by pre-programmed flight patterns or more sophisticated computer decision-making. Either way, a UAV is a capability that enables military personnel to avoid the physical risk associated with military missions and enables technology to perform tasks that human physiology cannot survive.

I therefore use a definition of UAV capabilities here that captures the concepts of aircraft reusability, remote piloting and autonomy, multi-role potential, and reduction of risk to military personnel: A UAV is a capability to send a reusable, remotely or autonomously piloted uninhabited aerial vehicle to conduct combat or combat support missions without exposing military personnel to physical risk.

UAVs 1952-2016: Key Events, Actors, Preferences

Case Organization and Key Events

I present the case in two major and two minor stages.278 The first major stage was development, which I trace from 1952 to 1992. The development stage was marked by a long

277 One UAV historian observes, “UAVs are equipped and employed for recovery at flight’s end and cruise missiles are not.” Laurence R. “Nuke” Newcome, Unmanned Aviation: A Brief History of Unmanned Aerial Vehicles (Barnsley, UK: Pen and Sword Books, Ltd., 2004), 4. 278 The following terms for the stages of the UAV case I borrow somewhat from the formal DoD acquisition process. “Development” here roughly corresponds to the material solution analysis, technology maturation and risk reduction, and engineering and manufacturing development phases. “Deployment” corresponds roughly to the production and deployment phase. See U.S. Department of Defense, The Defense Acquisition System, DoD Directive 5000.01 (Washington., DC: Department of Defense, May 2003), https://www.esd.whs.mil/Portals/54/Documents/DD/issuances/dodd/500001p.pdf; and U.S. Department of Defense, Operation of the Adaptive Acquisition Framework, DoD Directive 5000.02 (Washington, DC: Department of Defense, January 2020), https://www.esd.whs.mil/Portals/54/Documents/DD/issuances/dodi/500002p.pdf?ver=2020-01-23-144114-093.

126

period of service experimentation with a wide variety of UAV systems. Military preferences were not well consolidated, and neither were civilian preferences. Those interested in developing

UAVs at all tended to be a minority among the services. Moreover, a variety of perspectives on what types of systems and platforms would best meet service needs under the umbrella of reconnaissance stoked conflict and competition among the services that delayed technological advancement. The services were clearly the salient military institutions during this phase. On the civilian side, acquisition officials, civilian leaders in the services themselves, and Congressional budgeteers wielded the meaningful preferences.

The next, minor, stage I bound between 1993 and 2000 and call “accelerant,” when technology, leadership, defense trends, and operations all combined to speed the adoption of

UAV capabilities and begin the shift from development to deployment. The accelerant stage still featured civilian acquisition executives and service leaders, but also began to bring the COCOMs into the debate.

Deployment is the second major stage, which I mark between 2001 and 2011, a decade featuring the unprecedented and widespread use of UAVs in the wars in Afghanistan and Iraq and in the counterterrorism campaigns in Pakistan, Yemen, and Somalia. During deployment, deployed forces or “warfighters” became the salient military voice in the civil-military dialogue.

Key civilians were in OSD-Policy, the Secretary of Defense, the White House and on the Hill (in their war oversight capacity). Finally, I briefly cover a fourth, minor phase I call “offset.” During the offset stage, I show how ideas about the use of UAVs that formed part of the Third Offset paradigm correlate with a new independence in civilian preferences that took hold beginning in

2012 and continued through 2016.

127

The broader context for the case study phases and budgetary trends in UAVs occurred in three categories: technology, international crises and military operations, and trends in defense strategy and budgeting. Technologically, UAVs matured quite slowly through the development phase until the 1990s, when dramatic improvements in both navigational control systems and data processing both came on-line and enabled UAVs to be reliable in operational settings for the first time.279 In 1952, UAVs were barely more than target drones and weather balloons, although

Army investment in the 50s put cameras on aircraft that were launched by rocket, controlled via radio and tracked over radar, and landed using parachutes. Each of these elements could be finicky, however, and range and endurance were very limited. High failure and crash rates for

UAVs in the 1950s drove up costs. This persisted in the 1960s, even as the aircraft acquired new capacities such as electronic and communication intelligence collection, electronic countermeasures and suppression of enemy air defense (SEAD).

Meanwhile, the Air Force (with the help of DoD’s Advanced Research Projects

Administration or ARPA and funding from a variety of intelligence agencies) was experimenting with long-endurance UAVs to replace the manned U-2 high-altitude (above 60,000 feet) spy plane. Such experiments persisted through the 1980s, when a few small companies developed aircraft that could fly at medium altitudes (about that of a commercial passenger plane) for more than twenty-four hours and transmit live (or “real-time” in UAV jargon) video of what it was sent to survey. Endurance, the ability to stay aloft for far longer than a manned aircraft, gave

UAVs the ability to loiter over a target for a long period, meaning that if a target moved from its initial position, the UAV could observe where it went. This meant that UAVs could now do

279 This summary of the technological evolution of UAVs is taken from Lawrence R. Newcome, Unmanned Aviation; and Thomas P. Ehrhard, “Unmanned Aerial Vehicles in the United States Armed Services: A Comparative Study of Weapon System Innovation,” (PhD diss., Washington, DC, Johns Hopkins School of Advanced International Studies, 2000), https://catalyst.library.jhu.edu/catalog/bib_2239049. 128

something that neither manned aircraft nor satellites could do: persistently watch a mobile target.

As the 1990s dawned, the invention of the Global Positioning System (GPS) and leaps in computing power helped UAVs process and transmit the data they collected, and also improved the recovery rate of the aircraft. Computing power only improved in the first decades of the 21st century, and airframe innovations, including miniaturization, and developments in artificial intelligence provided UAVs novel capabilities including operating indoors and coordinating with large numbers of other units to “swarm” a target without humans piloting them.280

Military technology often develops concurrent with major military operations and crises, and UAVs are no exception. Signal events in the UAV case were the onset of the Cold War and the U-2 shoot-down incident in 1960, the Cuban Missile Crisis of 1962, the Vietnam War, the

Bekaa Valley raid in 1983, Operation Desert Storm in 1990/91, the interventions in Bosnia and

Kosovo in 1995 and 1999 respectively, the wars in Afghanistan and Iraq, and the counterterrorism campaigns in Pakistan, Yemen, and Somalia.

In the 1950s, the Soviet Union’s hostility and imperviousness to traditional intelligence- gathering methods alarmed Dwight D. Eisenhower’s administration, and drove development of technical intelligence collection methods, including the U-2 spy plane. In 1960, the Soviet Union shot down a U-2 as it conducted a reconnaissance mission over Russian territory. In the weeks of diplomatic drama than ensued, the high stakes and vulnerability of a manned strategic reconnaissance capability became apparent to public and government audiences alike.281 Two years later, the Cubans shot down another U-2 overflying their island, killing the pilot and prompting discussion back in Washington of the appropriate responses and possible escalation

280 Sayler, World of Proliferated Drones; and Newcome, Unmanned Aviation. 281 Associated Press, “Premier is Bitter,” New York Times, May 12, 1960, 1. 129

measures.282 In Vietnam just a few years later, military commanders turned to UAVs because of the “hostile air environment over North Vietnam and [being] anxious about the high losses of pilots shot down by missiles and other means.”283 UAVs were used extensively in that conflict, but quickly abandoned afterward.284

After a decade of neglect, tactical and operational UAVs got a new breath of life thanks to the Israelis and the American experience policing the war in Lebanon. In support of U.S. operations in Lebanon, which had already sacrificed two hundred and forty-one Marines in a bombing in Beirut just two months prior, the December, 1983 raid over the Bekaa Valley resulted in three U.S. aircraft lost, one pilot killed, and one bombardier captured.285 The raid itself was conducted to retaliate against the Syrians for trying to shoot down manned aircraft on surveillance missions. The Israelis then showed Navy and Marine Corps leaders how they used

UAV technology to draw Syrian anti-aircraft missile fire in order to reveal their launch locations so the Israelis could subsequently destroy them. This was the spur for America’s purchase of the

Israeli Mastiff UAV system, re-branded the Pioneer, which remained in the U.S. inventory until

2007.286 In fact, the Pioneer appeared in the first Gulf War, was used by three military services

282 “Notes taken from Transcripts of Meetings of the Joint Chiefs of Staff, October 27, 1962,” Foreign Relations of the United States (hereafter FRUS), 1961-1963, vols. X/XI/XII, American Republics; Cuba 1961-1962; Cuban Missile Crisis and Aftermath, Document 428, https://static.history.state.gov/frus/frus1961-63v10- 12mSupp/pdf/d428.pdf. Such debates over the escalatory dynamics surrounding attacks on U-2 flights persisted beyond the missile crisis. See: “Memorandum from the Joint Chiefs of Staff to McNamara, April 11, 1963,” FRUS, 1961-1963, vols. X/XI/XII, American Republics; Cuba 1961-1962; Cuban Missile Crisis and Aftermath; Document 650, https://static.history.state.gov/frus/frus1961-63v10-12mSupp/pdf/d650.pdf. 283 Richard A. Best, Jr., Intelligence Technology in the Post-Cold War Era: The Role of Unmanned Aerial Vehicles (UAVs), CRS Report No. 93-686 F (Washington, DC: Congressional Research Service, 1993), 8, https://fas.org/sgp/crs/intel/93-686.pdf; and William Wagner, Lightning Bugs and Other Reconnaissance Drones (Fallbrook, CA: Armed Forces Journal International, 1982). 284 Bill Yenne, Birds of Prey: Predators, Reapers and America’s Newest UAVs in Combat (North Branch, MN: Specialty Press, 2010); and Newcome, Unmanned Aviation, 83. 285 Ehrhard, “Unmanned Aerial Vehicles,” 346; and John F. Lehman, Jr., Command of the Seas (Annapolis: Naval Institute Press, 1988), 324-326. 286 The purchase of the Mastiff is a moment of activism on Lehman and Kelley’s part, as it is unusual for the U.S. to purchase entire weapons platforms abroad, even from allied and partner governments. 130

and, according to the official Pentagon analysis, provided critical reconnaissance and combat support functions on land and sea.287 The Gulf War itself “demonstrated that improved intelligence, combined with precision-guided munitions, [made] it possible to conduct air campaigns . . . with far greater accuracy than [had] been possible in the past.”288

But it was the war in the Balkans that propelled UAVs’ adoption by the U.S. military and intelligence community. That war was part of a general shift in the ways the U.S. used the military, itself a product of the international security environment. As the threat posed by the former Warsaw Pact seemingly evaporated, civil wars and humanitarian crises proliferated in

Europe, the Caribbean, and Africa. The fact that these events generated humanitarian crises in the first place drove a desire to minimize any further civilian casualties as a result of U.S. actions. There was also a sense that such operations did not involve vital American national interests, and therefore should pose minimal risk to American personnel as well.289

Accomplishing these two goals required information, precision, and to shield U.S. servicemembers from harm.

During the humanitarian military interventions in Bosnia and Kosovo, UAVs addressed these challenges. Despite a high crash rate (mostly due to operator error) and its inability to fly during bad weather, the intelligence supplied by the UAVs in Bosnia was so useful in discriminating between the perpetrators and the victims of genocide without risk to American pilots that they were deployed again to Kosovo four years later, this time adding a target designation mission in support of the air campaign—improving precision and reducing civilian

287 U.S. Department of Defense, Conduct of the Persian Gulf War: Final Report to Congress (Washington, DC: Department of Defense, April 1992), https://apps.dtic.mil/dtic/tr/fulltext/u2/a249270.pdf; see also: Avery Plaw, Matthew S. Fricker, and Carlos R. Colon, The Drone Debate: A Primer on the U.S. Use of Unmanned Aircraft Outside Conventional Battlefields (New York, NY: Rowman and Littlefield, 2016). 288 Best, Jr., “Intelligence Technology in the Post-Cold War Era,” 12-13. 289 Alice Hunt, “Civil-Military Relations and Humanitarian Intervention,” (unpublished master’s thesis, American University, 2006). 131

casualties.290 As the U.S. began to hunt Osama bin Laden in the wake of the 1998 embassy bombings in Africa, the CIA and the Air Force teamed up to use the Predator to find him in

Afghanistan. Just days after the attacks on 9/11, an armed version of the UAV was flying over the country again, in search of al Qaeda and Taliban leaders. Demand for UAVs then skyrocketed in Iraq, driving a new era of development and procurement of dozens of systems used by tactical and theater-level commanders alike. The total number of hours UAVs flew leapt seven-fold between 2003 and 2006 to over 300,000 hours per year.291

Encompassing both technology and operations were trends in defense strategies and budgeting. UAVs’ relationship to defense budgets was a complicated one, because they followed a combination of normal patterns in weapons investment and their own unusual fate. As with many military capabilities, the more generous the budgeting environment, the less scrutiny services and civilians alike applied to UAV costs and test performance. However, where UAVs threatened to direct resources away from other, higher priority programs, scrutiny and skepticism rose regardless of wider fiscal environments. Moreover, because UAVs were generally believed to be less expensive than manned platforms, they were often of more interest toward the beginning of periods of fiscal constraint—until they proved not to be as cheap as advertised.

The start-and-stop nature of UAV development until its acceleration in the 1990s and lift- off in the post-9/11 period is evident when the investments are displayed in quantitative terms.

Although the Army had as many as 1,400 reconnaissance UAVs in the late 1950s and early

1960s, by the late 1970s in the post-Vietnam era, the services had no operational UAVs and the

290 Richard Whittle, Predator: The Secret Origins of the Drone Revolution (New York, NY: Henry Holt and Company, 2014), 99-118; and Yenne, Birds of Prey, 39. 291 Associated Press, “Military Relying More on Drones, Mostly in Iraq,” NBC News, last updated January 1, 2008, http://www.nbcnews.com/id/22463596/ns/world_news-mideast_n_africa/t/military-relying-more-drones-mostly- iraq/#.XTjTdC2ZOCQ. 132

Air Force only had experimental high-altitude reconnaissance craft. In the 1980s, during the

Reagan defense boom, the Army poured $1 billion into a tactical UAV program, and the Navy bought its first UAV since its ill-fated drone helicopter program of the 1950s and 1960s. The

Pioneer was the only UAV in service, but several were in development. By fiscal year 2000, a decade of reduced Defense spending tempered growing interest in UAVs, and DoD spent just

$284 million on one hundred and sixty-seven UAV systems. In fiscal year 2010, in the wake of

9/11, the war in Afghanistan, and the invasion of Iraq, DoD was spending $3.3 billion on 7,500

UAVs.292 The very next year, the Army alone operated more than 5,800 UAVs.293 As of fiscal year 2017, spending on new drones was at $4.3 billion.294 A series of peaks and valleys in both

DoD and UAV spending had, after 9/11, become an exponential climb for UAVs, a climb that levelled off somewhat after the withdrawal from Iraq and the drawdown in Afghanistan, but certainly did not return to anywhere near its pre-9/11 levels.

Strategically, interest in UAVs also waxed and waned, correlating loosely with the relative emphasis on conventional and unconventional warfare. Throughout the latter half of the twentieth century and into the twenty-first, UAVs evolved in the context of DoD’s and presidential administrations’ approaches to deterring and defeating rivals during the Cold War, which itself spanned strategies and doctrinal trends that swung between modes of warfare; dealing with the uncertainty and low intensity conflicts of the post-Cold War world; responding to the threats posed by non-state actors, especially al-Qaeda-related terrorist groups after the

292 Jeremiah Gertler, U.S. Unmanned Aerial Systems, CRS Report No. 7-5700 (Washington, DC: Congressional Research Service, January 2012), https://fas.org/sgp/crs/natsec/R42136.pdf. 293 Under Secretary of Defense for Acquisition, Technology, and Logistics, Department of Defense Report to Congress on Future Unmanned Aircraft Systems Training, Operations, and Sustainability (Washington, DC: Department of Defense, April 2012), 2, https://fas.org/irp/program/collect/uas-future.pdf. 294 Dan Gettinger, Drone Spending in the Fiscal Year 2017 Defense Budget (New York, NY: Bard College, February 2016), https://dronecenter.bard.edu/files/2016/03/DroneSpendingFy17_CSD_3-1.pdf. 133

attacks of 9/11; and the Congress’s own responses to both strategic conditions and the proposals coming from DoD and the White House.

The case begins in the immediate aftermath of the Korean War, when the considerations of the nuclear battlefield as well as improved conventional methods prompted sweeping reviews of both the quantitative and qualitative aspects of American military capabilities. Eisenhower’s

New Look strategy emphasized economical military balancing against the USSR, while increasing investment in technical (as opposed to human) means of intelligence gathering. This embrace of technology to minimize risk was exemplified by the development of the U-2.295 The subsequent John F. Kennedy administration re-opened the door to conventional competition with the Soviet Union but also saw unconventional warfare as a major means of countering the spread of communism. By the mid-1960s Kennedy’s theories were being tested in the war in Vietnam, where the U.S. military used a range of unconventional and new military capabilities.296 UAVs were a mostly tactical part of this approach, although the period coincided with increased long- range intelligence gathering with unmanned alternatives to the U-2.

The ultimate failure to prevail in Vietnam and the high casualty rate associated with the war led to a decade of relative modesty in both strategy and defense spending. However, the tail end of the Carter administration introduced a little-remembered but highly influential series of concepts collectively called the “offset strategy.” Offset “led to major improvements in stealth, precision strike, battlefield information and communication systems, intelligence systems, positioning and navigation capabilities and training.”297 The idea was to “offset” the USSR’s

295 Gregory W. Pedlow and Donald E. Welzenbach, The Central Intelligence Agency and Overhead Reconnaissance: The U-2 and OXCART Programs, 1954-1974 (New York, NY: Skyhorse Publishing, 2016). 296 U.S. Department of Defense, “The Kennedy Commitments and programs, 1961,” in Report of the Office of the Secretary of Defense Vietnam Task Force (The Pentagon Papers). Washington, DC; and Elizabeth N. Saunders, Leaders at War: How Presidents Shape Military Interventions (Ithaca, NY: Cornell University Press, 2011). 297 Robert Tomes, “The Cold War Offset Strategy: Origins and Relevance,” War on the Rocks, November 6, 2014, https://warontherocks.com/2014/11/the-cold-war-offset-strategy-origins-and-relevance/. 134

conventional superiority and back away from the reliance on nuclear weapons. Although UAVs were not part of offset, the idea set the stage for a new emphasis on technology and an openness to new defense programs. In the 1980s, the Reagan administration pursued just that, putting minimal constraints on defense spending that relaxed concerns about competition among programs and gave the services space to experiment with UAVs—largely in the classified or

“black” budget context.298

The collapse of the Soviet Union between 1989 and 1991 coupled with the seeming technological triumph demonstrated by the first Gulf War, prompted a new conversation about technology as the backbone of U.S. defense. This “Revolution in Military Affairs” or RMA was an approach that built on offset’s technological investments in stealth, precision strike, and rapid information sharing, with the thought that a smaller, cheaper force could nevertheless maintain its technical edge over potential adversaries.299 The RMA relied heavily on the gathering and real-time sharing of battlefield information, the kind of thing that generated interest in reconnaissance capabilities. Meanwhile, the actual use of the military during the 1990s focused on so-called military operations other than war or MOOTW, a mode of warfare “more sensitive to political considerations” including the legitimacy of operations with “more restrictive rules of engagement” and employing restraint along lines more similar to policing than traditional combat.300 MOOTW doctrine codified an aversion to civilian and military casualties alike in the conduct of contemporary conflict.

298 Ehrhard, “Unmanned Aerial Vehicles,” 107. 299 Jeffrey Collins and Andrew Futter, eds., Reassessing the Revolution in Military Affairs: Transformation, Evolution and Lessons Learnt (London: Palgrave Macmillan, 2015). 300 U.S. Joint Chiefs of Staff, Joint Publication 3-07: Joint Doctrine for Military Operations Other Than War (Washington, DC: Department of Defense, June 1995), https://www.jcs.mil/Portals/36/Documents/Doctrine/pubs/jp3_07.pdf. 135

By the end of the William J. Clinton administration, the RMA was giving way to a new name for a similar set of concepts. First proposed by the 1997 National Defense Panel and later adopted by Republican-leaning defense analysts, the new-old idea was called transformation.

Transformation advocates embraced the idea of “technology as a force multiplier,” that quality of forces should be emphasized over quantity, and government should be reformed to make wartime decision-making responsive and agile.301

Although President Bush’s first Secretary of Defense Donald Rumsfeld continued to push a transformation agenda after the 9/11 attacks, the course of the Afghanistan and Iraq wars privileged the ideas of counterterrorism and counterinsurgency. This mix persisted until the final years of the Obama administration, when drawdowns in troop levels across both theaters of war concurrent with successive federal budget crises provided openings for the notion of a “third offset” to seize Pentagon leadership. The Third Offset, much like the preceding defense strategies of the 70s, 80s, and 90s, put its faith in technology to help overcome quantitative and so-called anti-access/area denial (A2/AD) disadvantages.302 Now aimed principally at China rather than Russia (although claiming it still acknowledged the threat of terrorists and other non- state adversaries), it was touted as a deterrence strategy, designed to make the U.S. so capable of winning future wars that no adversary would choose to start one. Although the Third Offset was proposed in the context of threats to reduce the defense budget as a result of national budget compromises, thinking among defense intellectuals was that the international strategic

301 Ricardo A. Marquez, “Transformation Achieved? Revisiting the 1997 National Defense Panel,” Joint Forces Quarterly 61 (2nd quarter, 2011), 72-80. 302 The broader defense community in Washington, D.C. recognized A2/AD challenges for at least a decade before they were included in the Third Offset innovation strategy. See for example Andrew Krepinevich, Barry Watts, and Robert Work, Meeting the Anti-Access and Area-Denial Challenge (Washington, DC: Center for Strategic and Budgetary Analysis, May 2003), https://csbaonline.org/research/publications/a2ad-anti-access-area-denial. 136

environment was changing contrary to U.S. interests and new approaches and investments would be needed to meet the future. Some of these new approaches involved UAVs, used in new ways.

Actors

As noted in the introduction, the important civilian and military institutions and actors shifted during the course of this case study as UAVs moved from development into deployment.

During development, key civilian institutions were OSD-Acquisition, the Defense Advanced

Research Projects Administration (DARPA), and budget authorizers and appropriators in

Congress; key military institutions and actors were the armed services and to a lesser extent the

Joint Staff.303 During deployment, those in the Office of the Under Secretary of Defense for

Policy (OSD-P), the White House, and those with operational oversight responsibilities on the

Hill became the bearers of civilian preferences. The Joint Staff remained relevant, but it was the

Combatant Commands and those fighting under them—what I call “warfighters”—who wielded the salient military preferences.

There is some ambiguity among critical actors that is important to acknowledge. An important civilian institution in the story of UAV preferences is the intelligence community (IC), particularly the Central Intelligence Agency (CIA). The CIA sustained Air Force investment in

UAV capabilities with their (often classified) budgetary support, and later played a major role in the use of reconnaissance-strike UAVs in the war on terror. But the IC is not in a policy-making role as are the civilian institutions of interest to this study, and although it executes certain foreign policy-related intelligence operations it is also not a military institution and thus not a

303 The Acquisition component of OSD was variously titled over the case period, including and importantly the under secretary for Research and Engineering and Acquisition, Technology, and Logistics, or AT&L. Although its current title is under secretary for Acquisition and Sustainment, I use “OSD-Acquisition” here for clarity and brevity. 137

component of the civil-military dynamic. This ambiguity is not fully resolved here; the roles the

IC and CIA played in the civil-military preference dynamic are described but the IC is not treated as an agent in the relationship.

Some individuals in the study also demonstrate the degree to which “civilian” and

“military” institutional identities can blur. Secretaries of Services are civilian political appointees but often act in the interests of the military service they head or, if they served, on the basis of their own military experience. Indeed, civilians with a career background as military officers blur the assignment of preferences in important ways. Moreover, the case is sprinkled with military officers assigned to work with or support civilian organizations who work to spread their institutional or even sub-institutional preferences among those civilians. These actors are important to the story because so often they bring military preferences with them from their past professions and use their power as civilians to advance those preferences. The carryover in institutional interests from, especially, military careers to civilian roles and organizations is a pattern in the evidence rather than a rule, and I work to make clear below when such preference transmission is happening and when it is not.

Purposes and Preferences

During sixty-four years of UAV emergence in the United States, UAV capabilities were dedicated to two general purposes: reconnaissance and strike. The former, which dominated preferences for UAVs throughout the case study, refers to a range of intelligence- and other information-gathering, from collecting photo images and “signals” (e.g., electronic and radio emissions) to capturing and live-streaming video imagery. Reconnaissance may be “strategic” which is to say it is conducted at long ranges in areas deep inside enemy territory to gain broad

138

insights about enemy capabilities and intentions; it can be tactical, referring to short-range information-gathering that can be used to prevail in a battle; or it may be operational, meaning medium-range flights over a target of information that has implications for multiple battlefield engagements. For the bulk of the case, the military reconnaissance role for UAVs was referred to as RSTA missions: reconnaissance, surveillance, and target acquisition. In recent decades, the favored acronym has become ISR, for intelligence, surveillance, and reconnaissance.

“Strike” purposes refer to destroying targets. Such targets may be related to regular military forces and bases, but more often during the case UAVs were used to strike individuals and groups designated by the U.S. as terrorists or “enemy combatants.” In general, the U.S. has only used UAVs for offensive strike missions in this context of irregular warfare—although elements of the air campaign in Vietnam approached conventional warfare conditions. The suppression of enemy air defense mission (SEAD), often performed by UAVs during the case period, straddles the line between offensive strike and combat support because it involved destroying enemy weapons. However, SEAD has typically been considered a combat support mission rather than combat. Nevertheless, it is intriguing here because it demonstrated UAVs’ potential for engaging more directly in combat long before a plurality of defense officials was willing to consider the idea seriously.

Along these lines, and equally important to the case, are the purposes to which UAVs have not been routinely committed and for which development has never advanced to full procurement: as conventional combat weapons platforms. Recently, defense experts in and out of government have imagined UAVs being used for air-to-air combat (colloquially known as dogfighting), swarming tactics during air and naval campaigns, submarine warfare capabilities,

139

and many other offensive uses.304 But throughout their long history, UAVs have been confined to combat-support reconnaissance roles and hunting non-state adversaries with no ability to shoot back at anything targeting them. Only in the last, very short, stage do civilian preferences stray into offensive conventional combat uses. As I will illustrate below, the limitations on civilian preferences for UAVs are the result of military preferences.

At the beginning of the case period, the military services (most importantly the Army and the Air Force) held two competing preferences regarding the purpose of UAVs. First, that UAVs should be developed and used for reconnaissance, surveillance, and target acquisition purposes, referred to by the acronym RSTA. Although experiments in other uses, including combat strike, happened from time to time, the dominant preference and development goal was to use UAVs for RSTA missions. The alternative military preference was simply not to develop UAVs, and instead to continue developing and relying on manned aviation capabilities or satellites. Thus, the services began the case study period holding two contrasting institutional preferences.

Civilians, meanwhile, rarely professed preferences for the military uses of UAVs prior to the wars in the Balkans in the 1990s. Although civilians helped advance the use of UAVs for strategic intelligence gathering, in general they hardly registered a preference for wartime uses of

UAV systems before the mid-1990s. Even when civilians on the Hill took greater notice of UAV development among the services in the late 1970s and 1980s, they focused on costs, casualties, and parochial political interests rather than operational applications.

Beginning in the 1980s, civilians in OSD began to adopt the service-based (but still minority) preference to use UAVs for reconnaissance. Civilians on the Hill accepted reconnaissance as the main purpose of UAVs, but also maintained their cost-and-budget-related

304 See, for example, Paul Scharre, Army of None: Autonomous Weapons and the Future of War (New York, NY: W.W. Norton and Company, 2018). 140

preferences. Civilians in the acquisition realm of OSD bridged the military and congressional preferences, accepting both in order to develop unmanned aerial technologies rapidly.

This complementarity of preferences quickly moved toward a closer overlap in preferences in the early 1990s as civilians on the Under Secretary for Policy’s staff and in the

White House began to learn about UAVs, often for the first time, from deployed forces. This trend continued through the Global War on Terrorism, and by 2012 civilians shared warfighter preferences and had worked to impose these preferences on the services, particularly the Air

Force. Now the generalized preference was for UAVs to perform ISR missions (the modern term for RSTA) and, to a lesser extent, combination reconnaissance-strike missions (or “hunter-killer” operations) in the counterterrorism context. Civilian and military preferences for the major uses of UAVs at this point were indistinguishable.

By 2016, civilian preferences in OSD had changed to embrace the use of UAVs for conventional combat and deterrence missions as well as irregular warfare. The Third Offset strategy did not so much change civilian preferences for UAVs as expand them into general categories beyond ISR, introducing the notions of “manned-unmanned teaming” and “swarm technology”—ideas that reflected new uses of UAVs as sidekick droids and for quantitative advantage. Civilians wanted UAVs to be used for everything from urban warfare to dog-fighting to amphibious assault.305 The service and warfighter preference for UAVs remained committed to UAVs as ISR and hunter-killer platforms in irregular warfare against adversaries without

305 For example, DARPA’s OFFensive Swarm-Enabled Tactics (OFFSET) program boasts: “These swarm tactics for large teams of unmanned assets would help improve force protection, firepower, precision effects, and intelligence, surveillance, and reconnaissance (ISR) capabilities.” Timothy Chung, “OFFensive and Swarm-Enabled Tactics (OFFSET),” Defense Advanced Research Projects Agency, n.d., https://www.darpa.mil/program/offensive-swarm- enabled-tactics. 141

sophisticated air defenses; they preferred not to incorporate UAVs into conventional concepts of operation.

In the following section, I trace the development, accelerant, deployment, and offset processes, showing how civilian preferences for the uses of UAVs became based on military preferences even as civilians maintained preferences for how UAVs should be developed; how military and civilian preferences came to overlap almost entirely; and then how civilians developed preferences that were separate from military preferences.

Tracing the Preference Formation Process for UAVs 1952-2016

Phase 1: Development (1952-1992)

In 1952, the military services—most especially the Air Force and the Army—had two preferences regarding UAVs. The first, held by a minority interested in UAVs’ potential applications, was to explore an intelligence and reconnaissance role for the capability. This general preference to use UAVs to gather information applied in a range of contexts, from the tactical to the strategic and from battle to espionage. This variety of contexts resulted in opportunities for disagreement between and among the services about what UAVs should be able to do technically.

The second preference was negative: that UAVs should not be developed or used at all.

This was the majority opinion in the services. This preference was based generally on the conviction that alternative capabilities, particularly manned ones, were superior to UAVs in terms of flexibility, cost, and culture.

In the beginning of the case, civilians generally did not hold strong preferences for UAVs except as alternatives to manned espionage aircraft, usually in explicitly non-military contexts.

142

At the beginning of the period, policymakers intervened little in military development efforts, and when they did they tended to side with the majority in the services who rejected UAVs.

By 1992, majorities in the services still eschewed UAVs, while a minority advocated their development for reconnaissance purposes. Civilians, meanwhile, had continued to mirror this distribution of preferences, with some willing to support UAV development efforts to improve reconnaissance capabilities, and most skeptical of UAVs’ costs and performance.

Details

The U.S. military had experimented with unmanned flying aircraft as far back as World

War I.306 But it took the Korean War to trigger the modern era of UAV development. As operations in Korea drew to a close and the Cold War persisted, the Air Force commissioned a study of the future of intelligence and reconnaissance by the Massachusetts Institute of

Technology.307 The study, code-named Project Beacon Hill, was the vanguard in a flurry of analyses conducted across the services to examine the Cold War operational environment.308

Despite the civilian—and even non-governmental—character of the report authors, they set an explicit precedent in relying heavily on military expertise, noting that they worked extensively with the Air Force to come to their conclusions. In fact, the Air Force officer assigned as the service’s liaison to the study group was a long-time proponent of strategic reconnaissance using high-altitude aircraft and ran the Reconnaissance Systems Branch of the Wright Air

306 Gertler, U.S. Unmanned Aerial Systems. 307 Incidentally, one of the very best places for a person to earn a PhD, along with Georgetown and American University. 308 The other military department-sponsored projects included Project VISTA, Project CHARLES, and Project LACROSSE; Massachusetts Institute of Technology (MIT), BEACON HILL Report: Problems of Air Force Intelligence and Reconnaissance, Project Lincoln, Massachusetts Institute of Technology (MIT), 1952 (Cambridge, MA: Massachusetts Institute of Technology, June 1952), https://www.governmentattic.org/12docs/USAF- BeaconHillReport_1952.pdf. 143

Development Command. Lieutenant Colonel Richard Leghorn had been involved in internecine

Air Force debates over whether reconnaissance airplanes should be tailored for tactical wartime or covert peacetime uses.309 Now, as the Beacon Hill team’s connection to the Air Force,

Leghorn had the opportunity to help these civilian scientists make the case for strategic aircraft.

In the end, the collaboration established the major capabilities available for Cold War-era reconnaissance: satellites, manned high-altitude aircraft, and UAVs.310 The study thus cemented the Air Force’s association of UAVs with strategic reconnaissance missions.

Meanwhile, the Army tied its interests in UAVs to battlefield purposes. The Army’s

Signal Corps, responsible for communications and intelligence collection, followed the Air

Force’s report with its own analysis in 1953. That project, “The Eyes of the Army” or Project

TEOTA, examined demands for battlefield RSTA. Intended to draw from a broad array of civilian and military perspectives—approximately 100 people from industry, academia, and government, including the Navy and the Air Force, contributed to the report—the authors of

TEOTA, like the Beacon Hill team, took care to explain that the final text was dominated by military perspectives: “Most all of the material we have obtained has come from the military departments or from their contractors or consultants” and “[m]any of our conclusions and recommendations will have a familiar sound to members of the military departments” they wrote.311

One of the key recommendations of Project TEOTA was for the Army to develop UAVs to conduct RSTA. There were two reasons for this. First, and specifically, the report highlighted

309 Pedlow and Welzenback, The Central Intelligence Agency, 5-22. 310 MIT, BEACON HILL Report, 169. 311 S.E. Clements, S. Seely, and F.F. Urhane, eds., Problems of Battlefield Surveillance (Washington, DC: United States Army, Project TEOTA, Office of the Chief Signal Officer, 1953), 3, https://apps.dtic.mil/dtic/tr/fulltext/u2/032532.pdf. 144

humans’ inability to operate in post-nuclear attack environments. Second, and more generally,

UAVs minimized risk to military personnel. “Because of the hazards of aircraft operation in these areas [behind enemy lines], the requirement that the aircraft be unmanned and recoverable, i.e., drone aircraft, is apparent.”312

The course of UAV thinking and development over the next forty years can be traced back to the foundational ideas—and military organizational interests—captured in the Beacon

Hill and TEOTA efforts.313 UAVs were associated, foremost, with intelligence and reconnaissance missions. They were a way to minimize risk while maximizing information. And they were a capability that each service tailored for its own needs. From then on, the Air Force and the Army took the conceptual lead on UAV technology and operational justification, with occasional if important contributions from the Marine Corps and the Navy. By 1957 the Army’s

Signal Corps had five UAV programs dedicated to surveillance in various stages of development and had set up a Combat Surveillance Agency to manage them.314 In 1952 the Air Force established its 645th Aeronautical Systems Group or “Big Safari” program, an office to develop and field “special purpose aircraft” quickly.315 Reconnaissance UAV technology was a major work stream for Big Safari. Meanwhile, the Marine Corps was also linking UAVs to reconnaissance, an idea championed by then-Major (later the Commandant) P.X. Kelley working in the Marine Corps Landing Force Development Center.316 Across these three services,

312 Ibid., 5-7. 313 Ibid., 5-31, 9-1-9-3. 314 These UAVs were: The SD-1 Observer, the SD-2 Overseer, the SD-3 Sky Spy, the SD-4 Swallow, and the SD-5 Osprey. Newcome, Unmanned Aviation, 72. 315 Whittle, Predator, 119-121. 316 Thomas P. Ehrhard, “Seeds of a Revolution: Maritime UAVs in the 1960s,” in Providing the Means of War: Perspectives on Defense Acquisition, 1945-2000, ed. Shannon A. Brown (Washington, DC: United States Army Center of Military History, 2015), 125. The UAV in question was the Bikini, which got its nickname from the commandant at the time, General Wallace M. Greene, Jr., who explained that the UAV was “a small item that covers large areas of interest.” 145

reconnaissance and surveillance needs were the main motivating purpose for developing UAV capabilities.

But the enthusiasm for UAVs did not seize everyone within the services uniformly. By the mid-1960s, the Army’s UAV advocates were encountering a challenge that all the other

UAV advocates in the services would face for the next 35 years: Most officers in positions to make decisions about force structure and budgets believed manned aircraft were simply better able to perform the RSTA mission because they were more reliable. Despite the optimism of the

TEOTA report, and the rise in the defense budget related to the war in Vietnam, in 1964 Army

Chief of Staff General Harold K. Johnson halted Army UAV development because the systems could not match the performance and flexibility of manned reconnaissance airplanes. Simply to bring UAVs up to par with the reconnaissance capability the Army already had—and, he argued, could rely on the Air Force to provide in any case—would be throwing good money after bad.317

The Marine Corps encountered a similar problem. It could not get the technology to work to its satisfaction—at least, not within the narrow limits they imposed on it for weight, costs, and manning. The Corps cancelled its UAV program after seven years of effort and a change in

Commandant, not because the technology would never work, but it would never be made to work when it was a low priority.

The Navy, a persistent outlier in the case, was the first to experiment with a non- reconnaissance role for UAVs, but also divested itself of UAVs for almost two decades thereafter. Chief of Naval Operations Admiral Arleigh Burke was searching for a range of anti-

317 J. H. Brown. R. G. Ollila, and R. D. Minckler, A Survey and Technical Systems Assessment of Drone Aircraft for Tactical Reconnaissance and Surveillance, Vol. 1 (Columbus, OH: Advanced Research Projects Agency, February 1972), C-1, https://apps.dtic.mil/dtic/tr/fulltext/u2/595802.pdf; Ehrhard, “Unmanned Aerial Vehicles,” 236; and War is Boring, “The U.S. Army Got its First Drones 55 Years Ago,” Medium, July 27, 2014, https://medium.com/war-is- boring/the-u-s-army-got-its-first-drones-55-years-ago-fc92ca1c3b3d. 146

submarine warfare capabilities, and the torpedo-firing Drone Anti-Submarine Helicopter or

DASH was one result of his efforts.318 But operational difficulties and Naval aviator and surface warfare officer opposition generated a shift in DASH’s mission from ship-based attack to littoral surveillance.319 DASH was eventually cancelled in 1966 and the anti-submarine warfare missions were thereafter carried out by manned aircraft.320 It was almost another 20 years before the Navy would attempt to develop a UAV again.

In fact, divestment in UAVs despite their strong performance in the war in Vietnam underscored the strength of service preferences to stick with manned capabilities. Vietnam offered UAVs something of a coming out party, as three different types flew during the war itself: two Air Force’s systems and the Navy’s DASH (as a surveillance craft). The Air Force flew more than 3,400 UAV sorties, most with a plane called the Lightning Bug and most for

RSTA purposes and other tactical uses, including as anti-aircraft decoys.321 But UAVs would suffer the same rejection by the American defense establishment as many other things about the war in Vietnam. Aviation historians trace the so-called “white scarf” or pilot opposition to UAVs back to Vietnam, something that helped frame unmanned aircraft as replacements for (rather than complements to or wholly different capabilities from) manned platforms.322

As the war wound down in the mid-1970s, tactical reconnaissance UAV development slowed to a crawl. By 1972, an ARPA study on tactical reconnaissance and surveillance lamented the underwhelming state of American UAV technology for military operations,

318 Ehrhard, “Unmanned Aerial Vehicles,” 309. 319 Rebecca Maksel, “D.A.S.H. Goes to War,” Air and Space Magazine, March 2012, https://www.airspacemag.com/military-aviation/dash-goes-to-war-23369442/?page=1. 320 Ehrhard, “Unmanned Aerial Vehicles,” 326. 321 Newcome, Unmanned Aviation, 83; and Yenne, Birds of Prey. 322 Mark Thompson. “A Killer Drone,” Time Magazine, November 18, 2001, http://content.time.com/time/magazine/article/0,9171,184971,00.html. 147

pointing out that no UAVs geared toward military operations were in use or development.323 The

Air Force’s Mission Analysis Group for Remotely Piloted Vehicles conducted a study in 1973 that found eleven possible uses for UAVs—including for conventional combat. Yet the GAO could not find evidence that the Air Force had pursued any of the investments recommended in the report.324 It was simply the case that the Air Force was no longer concerned with tactical

RSTA. It was focused instead on long-range strategic reconnaissance.

The Air Force put itself in the UAV business, but it stayed and focused its UAV capabilities on strategic reconnaissance in large part because of its relationship with the intelligence community. Around the same time that the Air Force established Big Safari, the CIA was contracting with Lockheed Martin to build the U-2 high altitude spy plane.325 The U-2 was an enormously useful intelligence-gathering aircraft that could fly long distances and capture images of enemy weapons, installations, and other items of national security interests to the

United States. But the plane was also running enormous political risks, flying deep into adversary territory under international legal conditions (espionage) that did not guarantee pilot safety if he was captured. That risk was borne out in May of 1960 when the Russians shot down a U-2 on a mission over Soviet territory. The Soviets took the pilot, Francis Gary Powers, prisoner and put him on public trial for espionage.326 Beyond the peril Powers was in, the incident also revealed the diplomatic risks behind manned reconnaissance flights by derailing a planned meeting between President Eisenhower and Soviet Premier Kruschev as well as

323 Brown and Minckler, Survey and Technical Systems Assessment, 14. 324 Comptroller General, DOD’s Use of Remotely Piloted Vehicle Technology Offers Opportunities for Saving Lives and Dollars, MASAD-81-20 (Washington, DC: General Accounting Office, April 1981), 10, https://books.google.com/books?id=x0ssAAAAIAAJ&printsec=frontcover&source=gbs_ViewAPI#v=onepage&q& f=false. 325 Ben R. Rich and Leo Janos, Skunk Works: A Personal Memoir of My Years at Lockheed (New York, NY: Little, Brown, and Company, 1994), 120. 326 Although Powers has been an Air Force officer for six years, he was a civilian by the time he joined the CIA as a pilot. 148

Eisenhower’s subsequent visit to Moscow.327 When another U-2 was shot down during the

Cuban Missile Crisis two years later, the incident once again illustrated the enormous risks being run with manned reconnaissance over Soviet targets.328 In a background memo prepared a year after the Missile Crisis, the CIA captured the “serious reservations” officials inside the executive branch maintained about U-2 reconnaissance flights.329

For civilian officials, these events supported arguments in favor of alternatives to manned aircraft. In response to the Powers incident, the U.S. government stopped conducting U-2 flights over the USSR for a time and instead performed its intelligence-collecting missions with satellites.330 In 1966 Deputy Secretary of Defense Cyrus Vance prevailed on the reluctant

Lockheed Corporation to continue building an unmanned version of the U-2, declaring, “We need this project to work because our government will never again allow a Francis Gary Powers situation to develop. All our overflights over denied territory will either be with satellites or drones.”331 Director of Defense Research and Engineering (and later President Carter’s Defense

Secretary) Harold Brown also had a change of heart. Although Brown had cancelled an unmanned version of the U-2 in favor of a manned alternative, by 1965 he enthusiastically approved of the same UAV.332

327 Associated Press, “Premier is Bitter.” 328 Wagner, Lightning Bugs, 42; Ehrhard, “Unmanned Aerial Vehicles,” 112-116; and Newcome, Unmanned Aviation, 84. 329 “U-2 Overflights of Cuba, 29 August through 14 October 1962,” FRUS, 1961-1963, vols. X/XI/XII, American Republics; Cuba 1961-1962; Cuban Missile Crisis and Aftermath, Volumes X/XI/XII; Document 623, https://static.history.state.gov/frus/frus1961-63v10-12mSupp/pdf/d623.pdf. 330 Vladimir I. Toumanoff, interview by William D. Morgan, “Vadimir I. Toumanoff,” Association for Diplomatic Studies and Training Foreign Affairs Oral History Project, June 18, 1999, https://www.adst.org/OH%20TOCs/Toumanoff,%20Vladimir%20I.toc.pdf. 331 Rich and Janos, Skunk Works, 267. 332 That alternative was codenamed OXCART and is known publicly today as the high-speed SR-71. Wagner Lightning Bugs, 17; Ehrhard, “Unmanned Aerial Vehicles,” 111; and ibid., 264-265. 149

Meanwhile, the Powers and Cuban incidents also drove intelligence community and Air

Force interest in finding ways to conduct strategic reconnaissance missions without putting a pilot in danger. A year and a half after Powers’ U-2 was shot down, the classified National

Reconnaissance Office (NRO) was established as an integrator of CIA, Air Force, and Navy reconnaissance assets jointly managed by the CIA and the Under Secretary of the Air Force— giving it a technically civilian organizational identity even though most of the staff were uniformed officers.333 The NRO was doubly important not just for the role it gave the Air Force in UAV development, but in the free money it offered to the service: the NRO funded development of UAV projects out of its own (classified) budget, requiring the Air Force to pay only for its own personnel costs.334

Around the same time, both civilians and uniformed officers in the Air Force began to encourage the Ryan Aeronautical Company’s efforts at building a UAV for strategic reconnaissance. As early as 1959, an Air Force Lieutenant Colonel by the name of Lloyd Ryan, who worked in the Air Force’s Reconnaissance Division of the Directorate of Operations at the

Pentagon, was fretting about the ramifications if a U-2 were to be shot down.335 Lt Col Ryan became the “recon drone missionary in the Air Force” and the Ryan Company would persist in developing a number of reconnaissance UAV systems over the following years.336 In 1962, the company got approval from Under Secretary of the Air Force and Chief of the NRO Joseph

Charyk to begin a UAV project under the auspices of Big Safari.337

333 Bruce Berkowitz with Michael Suk, The National Reconnaissance Office at 50 Years: A Brief History (Chantilly, VA: Center for the Study of National Reconnaissance, July 2018), 11-12, https://www.nro.gov/Portals/65/documents/about/50thanniv/The%20NRO%20at%2050%20Years%20- %20A%20Brief%20History%20-%20Second%20Edition.pdf. 334 Ehrhard, “Unmanned Aerial Vehicles,” 108. 335 Wagner, Lightning Bugs, 6; Incredibly, Lt Col Ryan was, in fact, related to the Ryan who founded the Ryan Company, although he was not aware of it until years later. 336 Ibid., 15. 337 Ibid., 23. 150

The result was the aforementioned Lightning Bug, which rolled off the production line in

1964. The Air Force flew “the Bug” extensively in Vietnam, but the service’s original and enduring interest in it was as a strategic reconnaissance aircraft. The NRO flew it over China as a test of Beijing’s air defenses. Although the Chinese shot one of the Bugs down, the lack of diplomatic or domestic political fallout, quite unlike the Powers incident, validated the idea that

UAVs were a lower-risk alternative to manned reconnaissance flights.338 Crisis events and Cold

War pressures, therefore, helped convince civilians the Air Force should indeed, per the service’s own preference, keep using UAVs for strategic reconnaissance. As a mark of the Air Force’s monopoly on UAVs used for strategic reconnaissance, the NRO transferred its entire UAV inventory to the Air Force in 1974 in order to focus exclusively on satellites.339

Meanwhile, interest in UAVs outside the Air Force and for operational and tactical purposes was ebbing. By the mid-1970s, the services had divested from UAV capabilities for military operations despite (or perhaps because of) the uses for UAVs that had been demonstrated in Vietnam. Despite the Carter administration’s emphasis on high technology in defense acquisition, UAV capabilities were not a focus of civilian or military attention. By 1978, even parts of the Hill remarked on the slow-down in military UAV development.340

But at the same time, the Soviet Union began to put nuclear weapons—and the air defenses to keep them from being attacked—on vehicles. Both satellites and vulnerable manned flights were a good deal less useful against mobile targets than they had been against static ones.

The satellites could miss critical weapons movements, and the manned capabilities were

338 Steven Zaloga, Unmanned Aerial Vehicles: Robotic Air Warfare 1917-2007 (New York, NY: Osprey Press, 2008), 4-6. 339 Harrison G. Wolf, Drones: Safety Risk Management for the Next Evolution of Flight (New York, NY: Routledge, 2017). 340 HASC Report, quoted in Comptroller General, “DoD’s Use of Remotely Piloted Vehicle Technology Offers Opportunities for Saving Lives and Dollars,” 4. 151

vulnerable to air defenses. This development moved the Reagan administration to re-invigorate military investment in UAVs for strategic reconnaissance—although not for other purposes. The intelligence transition team argued for increased funding to go to the NRO for a high altitude, long-endurance (HALE) UAV program.341 Two years later, an interagency group recommended the use of UAVs for both tactical and strategic intelligence.342 But by 1983, only the Army was developing a UAV for battlefield purposes and the endurance capabilities were still experimental

DARPA programs. The services simply were not pushing for UAVs, and most civilian officials had other priorities as well.

Once again, outside stimuli piqued civilian and military interest in unmanned aerial vehicles. In this case, the coincidence of a battlefield failure with a proof-of-concept for UAVs helped inspire a few key DoD personnel. First came the failure. By 1983 the United States had become militarily involved in the war in Lebanon as part of a multinational force. One such mission was air patrols, carried out by Navy aircraft launched from ships in the Mediterranean.

In December, the Syrians shot down three aircraft, killing two pilots and taking one as a prisoner of war.343 The incident occurred just two months after 241 Marines were killed in a truck bombing in Beirut, and appalled both the Secretary of the Navy, John Lehman, and the

Commandant of the Marine Corps P.X. Kelley—the same officer who had worked on the Marine

Corps’ drone program some 20 years prior.

During Kelley’s secret visit to Lebanon after the Marine Barracks bombing in October, the Israelis used a so-called mini-UAV system to video the general at the memorial ceremony

341 Ehrhard, “Unmanned Aerial Vehicles,” 138-139. 342 Jack Anderson, “Remote Control Superspy,” Washington Post, September 18, 1983, C7. 343 Lehman, Jr., Command of the Seas. 152

and later, to his astonishment, showed him the video.344 At the same time, a team of Navy officers also had the opportunity to witness Israeli UAV operations which exposed Syrian anti- aircraft launch locations.345 It was a version of the kind of SEAD missions that the U.S. had flown using UAVs during Vietnam. After the losses of the pilots in December, both Kelley and the Naval officers recommended to Secretary Lehman that the Navy acquire the Israeli UAV system, and Lehman agreed.346

In May 1984, Secretary of Defense Weinberger confirmed to the New York Times the

Navy’s purchase of “small, remote-controlled pilotless aircraft from Israel” was prompted by the casualties taken in the raids on Syrian missile sites in Lebanon.347 The system was re-branded as the Pioneer and operated by the Marine Corps. But Kelley and his Navy colleagues were not the only officials at DoD inspired to re-examine the utility of UAVs due to events in Lebanon that demonstrated both the demand for and advances in UAV technology. Civilian and military organizations alike became interested in the capability. Also in 1984, the Defense Resources

Board, “tasked the Joint Chiefs of Staff (JCS) to review requirements for UAVs by the individual services.”348

The events in the Middle East also spurred thinking about UAVs among the broader civilian and military intellectual community. And they started to think about UAVs beyond reconnaissance uses. A September 1985 article in the Washington Post quoted the new Pentagon

344 Ehrhard, “Unmanned Aerial Vehicles,” 348; and Fred Hiatt, “Weinberger Says Navy Has Bought Israeli Drones,” Washington Post, May 24, 1984, A29. 345 John David Blom, Unmanned Aerial Systems: A Historical Perspective, Occasional Paper 37 (Fort Leavenworth, KS: Combat Studies Institute Press, 2010), 72. 346 In most histories of the Pioneer, Lehman gets the credit for championing the Navy’s purchase of the UAVs. P.X. Kelley’s role, though equally if not perhaps more important, is highlighted much less often. See, for example, Norman Polmar, “The Pioneering Pioneer,” Naval History 27, no. 5 (October 2013): 14-15, https://www.usni.org/magazines/naval-history-magazine/2013/september/historic-aircraft-pioneering-pioneer. 347 Richard Halloran, “U.S. Buys Israeli Pilotless Planes,” New York Times, May 24, 1984, A8, https://www.nytimes.com/1984/05/24/world/us-buys-israeli-pilotless-planes.html. 348 Best, Intelligence Technology, 10. 153

RDT&E chief Donald Hicks, former Secretary of Defense Harold Brown, future Secretary of

Defense William Perry, and retired USAF Lt. Gen. Kelly H. Burke (who had been the USAF

R&D chief from 1979-1982) as all in favor of more smart, standoff systems, including unmanned ones. Hicks, citing the Lebanon raid pilot losses, told the Post that “the United States can no longer afford to risk so many lives and airplanes against increasingly lethal Soviet defenses and must put more emphasis on unmanned weapons.” Meanwhile, Burke said, “There may be such a thing as a cheap airplane, but there’s no such thing as a cheap American pilot.” The Post noted that Burke also argued, “it is time to start mass producing unmanned systems.”349

But, despite these ideas and Israel’s use of the Pioneer for SEAD missions, the U.S. military kept talking about UAVs for more narrow reconnaissance purposes. A year after the incidents in Lebanon, military officials were referring to the experience as a reason for Army investments in surveillance drones.350 By 1986 the JCS report on UAVs had provided a series of recommendations to re-invest in unmanned capabilities to conduct reconnaissance missions, and the administration’s fiscal year 1987 DoD budget submission “included renewed UAV procurement” for that purpose.351

Congress was comfortable with linking UAVs and reconnaissance, but not so comfortable with the price tags. And the sore thumb sticking out of the FY1987 budget was the

Army’s Aquila system, a highly capable UAV on paper, but to date a failure in tests with a near-

$1 billion debt to its name.352 Already in 1984, the House Armed Services Committee expressed

349 George C. Wilson, “Unmanned Weapons Gain Backing,” Washington Post, September 5, 1985, A1. 350 George C. Wilson, “U.S. Drone Crashes in El Salvador,” Washington Post, August 24, 1985, A12. 351 Best, Intelligence Technology, 10. 352 Blom, Unmanned Aerial Systems, 70; See also: General Accounting Office, The Army’s Remotely Piloted Vehicle Shows Good Potential but Faces a Lengthy Development Program, C-MASAD-82-8 (Washington, DC: February 1982), https://www.gao.gov/products/C-MASAD-82-8; and General Accounting Office, Aquila Remotely Piloted Vehicle: Recent Developments and Alternatives, NSIAD-86-41BR (Washington, DC: Government Accountability Office, January 1986), https://www.gao.gov/products/nsiad-86-41br. 154

its displeasure with the Army’s handling of the program by recommending the Congress significantly underfund the service’s budget request.353 In October of 1987, the GAO released a report that concluded the Aquila was “not ready for production.” 354 The FY1988 defense appropriations act signaled that Congress had had enough. The law consolidated all service UAV programs under a single office managed by OSD, the Unmanned Air Vehicle Joint Program

Office (JPO, pronounced “Jay-Poe”). Congress gave OSD budget authority but ordered that no money flow at all until DoD provided the congress with a UAV “master plan.”355

Importantly, the Hill was not signaling disapproval of UAVs overall, nor did it challenge the Pentagon’s vision for their reconnaissance uses. The 1984 HASC report even added $3 million to the Air Force budget “to investigate the feasibility of” something called the “High

Spot Surveillance Drone,” a “lighter than air vehicle that can be used as a host platform for surveillance and communication systems.”356 In fact, as Congressional involvement in UAV programs progressed, it was increasingly clear that the Hill preferred not to interrupt UAV development or broadly challenge how the services thought about the strategic justifications for

UAVs; Congress wanted to enable UAV development by forcing DoD to be more efficient.357

Nevertheless, for UAV advocates inside the Pentagon, the Congressional move came at exactly the wrong time. Enthusiasm for UAVs, though far more widespread than it had been prior to events in Lebanon, was still nowhere near the majority sentiment. Even the Navy, which

353 U.S. Congress, House, Committee on Armed Services, Department of Defense Authorization Act 1985 Report, 98th Cong., 2nd sess., April 19, 1984, H. Rep. 98-691, 62. 354 General Accounting Office, Aquila Remotely Piloted Vehicle, Its Potential Battlefield Contribution Still in Doubt, GAO/NSIAD-88-19 (Washington, DC: Government Accountability Office, October 1987), https://www.gao.gov/assets/210/209657.pdf. 355 U.S. Congress, Senate, Committee on Appropriations, Department of Defense Appropriations Act 1988, 100th Cong., 2nd sess., 1988, S. Rep. 100-235, p. 250. 356 Ibid, 156. 357 U.S. Department of Defense, DoD Joint UAV Program Master Plan (Washington, DC: Department of Defense, July 1988), 2, https://apps.dtic.mil/dtic/tr/fulltext/u2/a197751.pdf. 155

had taken the lead in re-introducing a UAV to military operations, was of two minds about the move. The Chief of Naval Operations wanted to stop buying the Israeli system almost the moment Secretary Lehman retired in 1987. The Deputy CNO for Surface Warfare, Vice Admiral

Joseph Metcalf III, stepped in to argue for the benefits of the system, and it survived the CNO’s wrath.358 Metcalf represented those who saw future uses for the platform—a vision that proved true in the Gulf War just a few years later. The CNO represented those who found the systems too costly and delicate, an unpopular Secretary of the Navy’s pet rock, unusable and redundant to existing capabilities besides.

The JPO did little to resolve these tensions. In fact, and despite Congress’ game funding of all but one of JPO’s annual budget requests, “in the years that JPO oversaw DoD development of UAVs, not a single system moved from development to full production.”359 Expressing the general frustration with JPO on the Hill, the House Appropriations Committee wrote in 1991 that

“the JPO has yet to produce a single UAV system . . . the Committee sees a UAV requirements process that is clearly broken.”360

This strange wheel-spinning on fielding UAV systems happened despite the useful performance of the Pioneer during the Gulf War. When, in 1990, Iraq invaded Kuwait and the

H.W. Bush administration decided to come to the country’s defense, the Army, Navy, and

Marine Corps all deployed Pioneer units to the Middle East. And all three of them had to lobby at various levels to be allowed to fly. For one, the Commanding General of the war, Norman

Schwartzkopf, initially banned the use of UAVs but, upon witnessing a Marine Corps flight

358 Ehrhard, “Unmanned Aerial Vehicles,” 360. 359 Blom, Unmanned Aerial Systems, 88. 360 U.S. Congress, House, Committee on Appropriations, Department of Defense Appropriations Bill, 1992, 102nd Cong., 2nd sess., 1991, H. Rep. 102-95, 334. 156

demonstration, allowed them to operate within a constricted zone.361 The Army Intelligence school meanwhile “literally shopped the Pioneer platoon around in the war zone, and after being rejected by the 18th Airborne Corps, found [their] way to General Fred Franks’ VII Corps near the Iraqi border.”362 Meanwhile, the Navy discovered that the Pioneer was an excellent reconnaissance tool “if it loitered over a target”, a new concept of operations for the U.S. military’s use of the UAV.363 In the end, “six operational units from three services flew over 300 combat missions” with the Pioneer system.364 “The Navy used Pioneers launched from battleships to support shore bombardment operations with fire direction and spotting; the Army and Marines used the UAVs for target designation, damage assessment, and reconnaissance and warning.”365 And the operators liked what they saw. As the commander of the Second Marine

Division Lt. Gen. William M. Keys put it, “the RPV [remotely piloted vehicle] is going to be our best tactical intelligence-gathering vehicle in the future, and we need to develop that program.”366

In a final, marked episode of civilian inattention and service desire to ignore their own

UAV capabilities, not many stakeholders seemed to notice the performance of UAVs in Iraq.

The Department of Defense’s interim report to the Congress on the conduct of the war crowed,

“the war marked the dawn of a new technological era,” spending multiple pages on precision weaponry, but struck a more cautious and succinct note on UAVs in just a few sentences, including: “Pioneer proved to be valuable and appears to have validated the operational

361 Ehrhard, “Unmanned Aerial Vehicles,” 373. 362 Ibid., 292. 363 Ibid., 369. 364 Steven E. Reid, “Operational Use of the Pioneer Unmanned Aerial Vehicle System,” Society for Photo-Optical and Instrumentation Engineers, SPIE Proceedings 2829 (November 1989), https://www.spiedigitallibrary.org/conference-proceedings-of-spie/2829/1/Operational-use-of-the-Pioneer- unmanned-aerial-vehicle-UAV-system/10.1117/12.259731.short. 365 Best, Jr., Intelligence Technology, 11. 366 Quoted in ibid., 11. 157

employment of UAVs in combat based on preliminary data.”367 In 1993 the RAND corporation published a report commissioned by the Air Force on the future of airpower that relied heavily on the service’s data from the Gulf War. Perhaps because the Air Force did not operate the

Pioneer, the report did not consider the implications of the UAV’s performance for the future of

American air power.368 Regardless, even amidst a celebration of technology on the battlefield, military institutions did not dwell on the performance of the Pioneer and civilians therefore did not notice it.

From Korea to Vietnam to the Persian Gulf, the military services had not substantially moved away from their dual preferences to either use UAVs for information-gathering or not to invest in them at all. Although experiments with long-range strategic reconnaissance craft continued at DARPA, mainstream service programming for UAVs continued to stutter. And civilian preferences by and large reflected these military preferences. By January of 1993, even as multiple technologies were improving the performance of unmanned aerial systems and allowing for a wider range of applications for them, the Secretary of Defense still devoted little space to UAVs in his annual report to Congress and was still expressing the same preference to use them for RSTA first laid out by the Air Force and the Army in 1952: “Unmanned aerial vehicles . . . will be used to acquire targets, survey the battlefield, and initiate electronic countermeasures.”369

367 U.S. Department of Defense, Conduct of the Persian Gulf War: Final Report to Congress (Washington, DC: Department of Defense, April 1992), 6-8, https://apps.dtic.mil/dtic/tr/fulltext/u2/a249270.pdf. 368 David Ochmanek, interview with the author, May 20, 2019. The report to which Mr. Ochmanek referred was Christopher Bowie et al., The New Calculus: Analyzing Airpower’s Changing Role in Joint Theater Campaigns (Washington, DC: RAND Corporation, 1993), https://www.rand.org/pubs/monograph_reports/MR149.html. 369 U.S. Department of Defense, Annual Report to the President and the Congress, FY1993 (Washington, DC: January 1993), 79, https://history.defense.gov/Portals/70/Documents/annual_reports/1993_DoD_AR.pdf?ver=2014- 06-24-152422-603. 158

Phase 2: Accelerant (1993-2000)

At the start of 1993, a majority within the military, largely represented by the services, continued to prefer not to invest in UAVs. Those few military advocates still dedicated to developing UAV capabilities continued to focus on reconnaissance purposes. Civilians in general still focused on UAVs’ costs and tended to follow military organizations’ lead on what UAVs could be used for—which is to say, reconnaissance.

But then 1993 became a turning point for UAV development. None of the major preferences had changed in their content. The services’ dominant preference was still to eschew

UAVs; a smattering of military officers continued to be interested in developing and using them for reconnaissance; and civilians broke one way or the other. But the introduction of a few key actors on both the civilian and military sides, coupled with international security events, gave development of UAVs a momentum it had not enjoyed since the late 1950s. By 2000, the number of advocates for UAV use on the battlefield and for strategic reconnaissance was much larger, and a new preference to use UAVs to conduct strikes in unconventional warfare contexts was emerging in the Air Force.

The short accelerant phase shows a vanguard of civilians and military officers breaking away from the long-standing preference to suppress investment in UAVs and instead invest in them as a serious reconnaissance capability. It also shows how key military figures began to think of UAVs as weapons and took steps to shape UAVs for that purpose.

Civilian preferences during this period may appear independent from military preferences to the casual observer because civilians were often seen pushing the services to embrace UAVs.

But this deeper look into the dynamics between preferences reveals that civilians were simply

159

aligning with the pre-existing pro-UAV camp in an old intra-military debate over the uses and usefulness of the capability.

Details

President Clinton was sworn into office at the end of January in 1993 and injected DoD with new civilian leadership. Most important to the UAV story was Bill Perry as Deputy

Secretary, John Deutch as Under Secretary for Acquisition and Technology, and Larry Lynn as

Deutch’s Deputy Under Secretary for Advanced Systems and Concepts. Each man had, in previous positions in government, on government study groups, or in academia, developed expertise in reconnaissance needs, the precision strike and information revolution in warfare exemplified by the first Gulf War, and UAVs. Lynn, for example, contributed to a Defense

Science Board study on global surveillance that encouraged the development of endurance

UAVs.370

Deutch arrived for his second tour in the Pentagon already believing the military needed more drones. “In his view, drones were a potentially lifesaving reconnaissance technology that would have been operational by now if the armed services weren’t so myopic or the acquisition system such a mire.”371 Deutch was greeted at the Pentagon by Vice Chairman of the Joint

Chiefs of Staff Admiral David Jeremiah. By virtue of their positions, the two worked closely together, including as co-chairs of the Defense Acquisition Board. Jeremiah was already an advocate for UAVs as well, and pressed Deutch for assistance.372

370 Richard H. Van Atta, et al, “Transformation and Transition: DARPA’s Role in Fostering an Emerging Revolution in Military Affairs, Volume 1—Overall Assessment.” Alexandria: Institute for Defense Analyses, 2003, 45. 371 Whittle, Predator, 76-77. 372 Deutch characterized his collaboration with Jeremiah as follows: he “worked closely with Admiral Jeremiah . . . from [Deutch’s] first days in the Pentagon. Jeremiah was a strong advocate and the moving force for placing greater emphasis on joint unmanned aerial vehicles programs and a strong proponent for the creation of DARO. Deutch, 160

Also important was the new Director of Central Intelligence James Woolsey—an old friend of Deutch’s—who, like Navy Secretary John Lehman and General P.X. Kelley, had personally witnessed Israeli UAV operations in the 1980s. In a bit of hyperbole that nonetheless conveys his enthusiasm, Woolsey recalled the first few months of the Clinton administration as

“jump-starting, in what has turned out to be a very positive direction, the government’s

Unmanned Aerial Vehicle program.”373

The war in Bosnia provided both the CIA and the Pentagon with the impetus to take a fresh look at UAVs. President Clinton wanted to take action to stop the humanitarian atrocities reportedly underway in the country but “was both shocked and chagrined to find out how little his military and intelligence agencies could tell him about what was actually happening on the ground around Sarajevo.”374 The problem was that Bosnia was frequently overcast, and none of the U.S.’s reconnaissance capabilities could see through the clouds. Woolsey took it upon himself to find an intelligence solution, tasking the CIA to come up with something. One of his analysts then presented him with a picture of a UAV design, one that he recognized as belonging to a UAV developer he had met through his Israeli contacts, Israeli-American Abraham

Karem.375 Karem was working for General Atomics, which had purchased the UAV the CIA staffer had identified and named it the Predator, as well as a less-versatile version called the

Gnat.376 Both the Gnat and the Predator could transmit real-time video of surveillance targets. At

whose experience working with defense industry and as a member of the Defense Science Board, recognized the importance of the technological advance of UAVs and was in the happy position of supporting Admiral Jeremiah’s efforts in the Pentagon and related interagency activities.” John Deutch, email to the author, August 8, 2019. 373 Jim Woolsey et al., interview by Miller Center, University of Virginia, January 13, 2010, https://millercenter.org/the-presidency/presidential-oral-histories/r-james-woolsey-oral-history-director-central. 374 Whittle, Predator, 71. 375 Woolsey et al., interview. 376 Newcome, Unmanned Aviation, 106; and Whittle, Predator, 57. 161

Woolsey’s direction, the CIA purchased the Gnat from General Atomics, and by the winter of

1994, the Agency had sent two systems to Albania to conduct surveillance over Bosnia.377

In parallel, Chairman of the Joint Chiefs of Staff General Colin Powell wanted to solve the same Bosnia intelligence problem, and told his J2 or Director for Intelligence, Rear Admiral

Michael W. Cramer to find a system that would solve the issue. Cramer then met with

Commander Steve Jayjock, who had worked on UAVs at DARPA. It is not clear whether Powell directed Cramer to focus on UAVs or Jayjock recommended it. Regardless, together Cramer and

Jayjock produced a list of capabilities an unmanned reconnaissance and surveillance aircraft would need to provide, a list that was a clear combination of Congress’ long preference for cheaper systems with the Pentagon’s efforts to develop a reconnaissance capability. First, the

UAV would need to use commercial, off-the-shelf technology and cost no more than $2.5 million per copy—this was the legacy of Congressional impatience with expensive UAVs. Next, it needed sensors for ground reconnaissance, and to be able to conduct surveillance by day and night. Third, all equipment should be unclassified in case the system crashed and was recovered by the Serbs. And finally, it needed range and endurance, hallmarks of reconnaissance aircraft.378

Navy Captain Allan Rutherford, working in the still-extant JPO, was tagged as the lead on the Pentagon’s effort. The JPO was still housed under OSD A&T, giving Deutch and Lynn direct influence over its activities. Deutch directed Rutherford to coordinate with the CIA to acquire a system for DoD that was similar to the Gnat. On July 12, 1993, Deutch signed out a memo to the Navy that Rutherford wrote called “Endurance Unmanned Aerial Vehicle (UAV)

377 Reuters, “CIA Sending 2 Drone Aircraft to Albania to Observe Bosnia,” Washington Post, January 29, 1994, A14, https://www.washingtonpost.com/archive/politics/1994/01/29/cia-sending-2-drone-aircraft-to-albania-to- observe-bosnia/734a7e15-fd06-4122-8a0b-516a0c30b6d5/. 378 Whittle, Predator, 73-76. 162

Program.”379 It called for urgent contracting, using the capabilities list Cramer and Jayjock came up with.

In the entire process to this point, it was nearly impossible to see where military preferences ended and civilian preferences began. The President generated the demand for answers on Bosnia, and the CIA director had found one in the form of an off-the-shelf UAV even as the Chairman of the Joint Chiefs of Staff, the Vice Chairman, and the Under Secretary for

Acquisition and Technology all worked in parallel to do the same. On the DoD side, Deutch’s general interest in promoting UAVs was partly due to Admiral Jeremiah’s interest, and the UAV program description he signed out was generated by military officers responding to direction from General Powell. As far as civilians were concerned, the purpose of this UAV was intelligence-gathering on the Serbs in Bosnia; the purpose of UAV capabilities more broadly they left to the military.

Regardless of UAVs’ immediate and strategic uses, OSD-A&T was laying groundwork for the military services to get more. Lynn, newly arrived in the Pentagon and keying off the

Defense Science Board studies and “unfilled requests of theater commanders-in-chief,” conducted an internal study that concluded DoD needed to focus more on aerial reconnaissance by purchasing medium- and long-range UAV systems.380 The study also recommended the creation of a Defense Airborne Reconnaissance Office (DARO). Perry, Deutch, and Jeremiah then worked with Congress to get OSD-written language into the 1994 Defense Authorization

Act.381 Perry lost no time standing up the office, and issued a memo directing its creation in

379 Under Secretary of Defense for Acquisition, “Endurance Unmanned Aerial Vehicle Program,” (official memorandum, Washington, DC, Department of Defense, July 1993), https://nsarchive2.gwu.edu/NSAEBB/NSAEBB484/docs/Predator-Whittle%20Document%201%20- %20Deutch%20Endurance%20UAV%20Memo%2012%20July%201993.pdf. 380 Ehrhard, “Unmanned Aerial Vehicles,” 509, footnote 1298; and Van Atta, 46. 381 Ehrhard, “Unmanned Aerial Vehicles,” footnote 1298; U.S. Congress, House, National Defense Authorization Act for Fiscal Year 1994, 103rd Cong., 1st sess., 1993, H. Rep. 103-357, 599. 163

November of 1993. The JPO was not out yet, but it was down, with DARO now the main overseer and JPO reduced to a programming office. The DARO’s first report to Congress in

1994 demonstrated how the Pentagon’s big ideas about what UAVs should do were clearly products of the services’ experiences in UAV development over the preceding 40 years. What was different was the DoD’s sense of urgency. The report noted that “reconnaissance, surveillance, and target acquisition (RSTA) are the premier missions for UAVs” and that

“allotting these ‘dirty’ and dangerous missions to UAVs decreases the risks to manned aircraft and frees pilots to perform missions that require the flexibility of the manned system.”382

At the same time, an even more important UAV accelerant was established. As he was pitching the DARO idea, Lynn proposed, and Deutch approved, the creation of a rapid acquisition process called “Advanced Concept Technology Demonstrators” or ACTDs, which

“provided for streamlined management and oversight, early participation of the user community, and a tight schedule.”383 It was this “early participation of the user community” that mattered the most to the course of civilian preferences for UAVs, because it brought civilians into direct and systematic contact with those using the capability in operations for the first time. It also gave the operating force a large role in justifying the purpose of UAVs. As the 1997 DoD Annual Report to Congress explained, ACTDs provided “the users with a basis for evaluating and refining their operational requirements, for developing a corresponding concept of operations, and ultimately for developing a sound understanding of the military utility of the proposed solution, before a

382 Defense Airborne Reconnaissance Office, UAV Program Plan (Washington, DC: Department of Defense, 1994), 2-1, https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=2ahUKEwiZosiyxI7oAhVPjp4KHb emDG8QFjAAegQIARAB&url=https%3A%2F%2Fwww.hsdl.org%2F%3Fview%26did%3D745074&usg=AOvVa w3SYideOLNM5og-GT_5wMqD. 383 Van Atta, 47. 164

formal acquisition decision” was approved.384 Deutch and Lynn chose to make the first ACTD a

UAV: General Atomics’ Predator.385

Again, OSD went to the Hill, this time to get buy-in for their rapid acquisition model and to generate general support for UAV development. The first stop was the office of California

Representative Jerry Lewis of the HAC-D. General Atomics was headquartered in his district and he therefore had an interest in defense contracts going to his constituents. Lewis put $20 million for an endurance UAV in the next year’s appropriations bill.386 Others, such as Chairman of the House Permanent Select Committee on Intelligence Representative Norm Dicks of

Washington State, where Boeing is headquartered, were not in committee positions to directly oversee OSD, but were in positions to sustain funding for UAV programing that began in the

1980s. Dicks’ interest was in the so-called Tier III UAVs, which under Deutch and Lynn evolved into the ultimately successful Global Hawk and ultimately unsuccessful Dark Star programs.387

This congressional support for existing programs and uses of UAVs—for reconnaissance— demonstrating that civilians on the Hill were very open to the Pentagon’s rationale for UAVs’ strategic uses, so long as the budgetary politics conformed to their political interests.

To build and sustain momentum, the Pentagon’s civil-military coalition of UAV advocates needed to finally get a system to change minds in the services and transition to a regular acquisition and fielding process. Racing through its ACTD timeline, the Predator, under the management of JPO’s Captain Rutherford, had been tested by the Army’s Military

Intelligence Battalion and set a new endurance record for a UAV at over 40 hours. Its live video

384 U.S. Department of Defense, Annual Report to the President and the Congress, FY1997 (Washington, DC: Department of Defense, April 1997), 71-72, https://history.defense.gov/Portals/70/Documents/annual_reports/1997_DoD_AR.pdf?ver=2014-06-24-153256-173. 385 Whittle, Predator, 79-80. 386 Ibid., 80. 387 Ehrhard, “Unmanned Aerial Vehicles,” 156-157. 165

feed was then piped into the Pentagon for senior military audiences, including now-Vice

Chairman of the Joint Chiefs of Staff Admiral William A. Owens, who became a proponent of the system as a result.388

Rutherford then lobbied European Command to use the Predator over Bosnia. By the end of 1994, the Gnat was well on its way to changing minds skeptical of the notion that UAVs could reliably contribute to military operations.389 Operated by the Army’s intelligence battalion, the

Predator lived up to the reputation for UAVs built by the Gnat, as real-time color video of scenes from Sarajevo flooded into the Combined Air Operations Center (CAOC) in Italy. “Soon the phone in the Predator operations center was ringing constantly, as officers in Naples and Vicenza called with requests from commanders, who had long lists of ‘targets’ they wanted the Predator to fly over.”390

The Army was flying the Predator over Bosnia, but it was the Air Force’s senior leaders who realized a revolution was underway. Back in Virginia at the Air Combat Command (ACC), the commander, General Richard E. Hawley, declared, “the operational capabilities embodied in the Predator UAV system are a significant first step toward the continuous, real-time

Reconnaissance, Surveillance and Target Acquisition required by 21st century joint warfighters.

ACC is committed to developing our ability to employ the family of UAVs in that role.”391 The

Air Force Chief of Staff General Ronald Fogleman gave a speech at the National Defense

University in June of 1995 declaring “the U.S. Air Force on my watch is going to aggressively embrace the UAV concept.”392

388 Whittle, Predator, 99. 389 Newcome, Unmanned Aviation, 109. 390 Whittle, Predator, 104. 391 Quoted in Arthur Holland Michel, “Drones in Bosnia,” Center for the Study of the Drone, June 7, 2013, https://dronecenter.bard.edu/drones-in-bosnia/. 392 Whittle, Predator, 108. 166

Fogleman had come into office with a background in flying high-risk reconnaissance missions during Vietnam, and he was convinced that UAVs were the future of that capability. As

CSAF, Fogleman established the 11th Reconnaissance Squadron at the end of July in 1995, just as the Predator was showing its potential in Bosnia.393 He then went about seizing operational control over the system from the Army, which he achieved in April of 1996 when now-Secretary of Defense Bill Perry signed a memorandum to that effect—although the decision left procurement to the Navy in an effort to maintain the jointness Congress so valued.394

Meanwhile, Big Safari was getting back in the UAV game. Its director, a retired Air

Force colonel named Bill Grimes, appreciated Fogleman’s instinct to get control over the

Predator but had much less faith in the corporate Air Force’s ability to manage the ongoing development and eventual transition of the bird into normal acquisition and operations. On the

Hill Mike Meermans, an old Air Force colleague of Grimes’, was now on the staff of the HPSCI and he agreed with Grimes’ assessment. Meermans arranged an unofficial Big Safari briefing for

Representative Jerry Lewis and freshman Representative James A. Gibbons of Nevada, another retired Air Force officer. Impressed, the Congressmen put language into the HPSCI authorization report that year that “strongly urge[d]” the Air Force to give the Predator program to Big

Safari.395 In 1997, the Air Force did just that.

Two years later, the Predator was sent back to the Balkans, this time to contribute to the air campaign over Kosovo as a target designator for fighter planes that could not fly low enough long enough to spot mobile targets lest they be hit by anti-aircraft fire.396 Once again, the system

393 Ibid., 110. 394 Ibid., 111. 395 Ibid., 125. 396 Ibid., 130-131. 167

pleased warfighters. “We’re getting fabulous imagery of Serb forces,” a NATO officer told the

New York Times.397

The experience in the Balkans was proving Larry Lynn’s theory of ACTDs correct: putting UAVs in the hands of warfighters helped expand enthusiasm for the systems and accelerate their acquisition in DoD. Years later, Lynn recalled, “I can remember standing in the

Pentagon, looking at the real-time Predator images in Bosnia, and that began to capture people’s attention.” It was not just the capability itself, but the context it was used in, that began to convince people in Washington that UAVs might be worth more investment. Lynn said, “We were lucky to have a war come along where they began to be useful. It was the combat experience that really carried them over the top.”398

Beyond being the context for validating the Predator’s utility, the war in Kosovo served as a focal point for a nascent military preference for UAVs: the idea that they could not only gather information on the battlefield but could also initiate lethal action. General Fogleman had been an early proponent of this renewed idea, telling one reporter in 1997, “I think the next area that starts to makes sense for UAVs is some sort of unmanned attack airplane.”399 In Kosovo, the proof of the concept’s battlefield utility started with the “TA” or target acquisition part of the

RSTA mission. Predator video feeds were being used to “talk” fighter and bomber pilots “onto” their targets. But the system was haphazard at best. The new Air Force Chief of Staff directed that the Predator be outfitted with technology to designate targets for manned fighters.400 Big

Safari worked on the project swiftly, outfitting multiple Predators with laser designators that

397 Elizabeth Becker, “They’re Unmanned, They Fly Low, and They Get the Picture,” New York Times, June 3, 1999, A15, https://archive.nytimes.com/www.nytimes.com/library/world/europe/060399kosovo-drones.html. 398 U.S. Department of Defense Executive Secretariat. Interview with Larry Lynn, December 8, 2006, 351. 399 John A. Tirpak, “The Robotic Air Force,” Air Force Magazine, September 1997, 70-74, https://www.airforcemag.com/article/0997robotic/. 400 Whittle, Predator, 142. 168

could “paint” a target for strike aircraft, but only succeeded in demonstrating the capability during one test in Kosovo the day before the Serbians surrendered.401 Nevertheless, General John

Jumper, who had been working in the CAOC during the Kosovo campaign, got the idea that

UAVs could do more than just watch the battlefield. Jumper had flown the same high-risk reconnaissance missions in Vietnam as the now-erstwhile Chief of Staff, General Fogleman.402

He realized how enabling unmanned aircraft could be.

From 1999 to 2000, three related but apparently unconnected developments set two future courses for UAVs. The first two developments—the arming of the Predator and the hunt for Osama bin Laden—would directly shape the uses of UAVs for the immediate decade and a half following them. The third—the rise of the transformation concept and its inclusion of UAVs in conventional combat operations—was a continuation of the civilian alliance with UAV proponents and a trend that would be interrupted by the 9/11 attacks and then taken up again twelve years later as the country tired of the counterterrorism wars of the early twenty-first century.403

In 2000, General Jumper took command of the ACC and discovered that the Air Force had removed the laser designators from the Predators the moment the war in Kosovo was over. It demonstrated the enduring bias in the Air Force against using UAVs for broader purposes. Then,

Jumper received a briefing on the margins of the Air Armament Summit that March called

401 Arthur Holland Michel, “How Rogue Techies Armed the Predator, Almost Stopped 9/11, and Accidentally Invented Remote War,” Wired Magazine, December 17, 2015, https://www.wired.com/2015/12/how-rogue-techies- armed-the-predator-almost-stopped-911-and-accidentally-invented-remote-war/. 402 U.S. Air Force, “General John P. Jumper,” November 1, 2005, https://www.af.mil/About- Us/Biographies/Display/Article/104697/general-john-p-jumper/. 403 In building their theory about implementing military innovations, Peter Dombrowski and Eugene Gholz point out that “advocates of military transformation consciously used . . . civilian transformation [in the technology sector] as a model.” Peter Dombrowski and Eugene Gholz, Buying Military Transformation: Technological Innovation and the Defense Industry, (New York, NY: Columbia University Press, 2006), 2. 169

“Weaponized UAV Demonstration.”404 Jumper was inspired. “Putting a weapon on the Predator wasn’t just a good idea, Jumper reflected—it was a great idea.”405 That May, in a gesture that was surprisingly banal and bureaucratic for such a momentous decision, Jumper issued a notification to the senior leaders of the Air Force that he had “internalized the Predator lessons learned from Operation Allied Force” and had decided to arm it “based upon Predator’s proven track record of warfighting support to the CINCs [Combatant Commanders] since 1995.”406

Of course, arming a light UAV not originally designed to carry ordnance was no small task. Jumper gave the project to Big Safari and Bill Grimes and his team were still working on the problem in October of 2000 when al Qaeda mounted an attack on the USS Cole, an American

Navy destroyer on a port call in Yemen, blasting a hole in the side of the ship and killing seventeen sailors.407 The CIA had been searching for the leader of al Qaeda, Osama bin Laden, ever since his organization had launched simultaneous deadly attacks on American embassies in

East Africa in 1998.408 In the summer and fall of 2000, the Agency and the Air Force had jointly operated the Predator searching for bin Laden during the “Afghan Eyes” program, also known as the “Summer Project.”409 The team believed they had located him. Now the Clinton team, with just a few months left in office were more willing to try to kill the man and to do so with an armed UAV. Big Safari kept up its work.

Unaware of this flurry of Predator activity, another proponent of UAVs had been trying to get the military services into an entirely different concept for UAVs. Senator John Warner, a

404 Whittle, Predator, 167-168. 405 Ibid., 168. 406 Air Combat Command, “RQ-1, Predator, Program Direction,” (cable message to USAF headquarters, unclassified, May 1, 2000). 407 Michel, “How Rogue Techies Armed the Predator.” 408 Barton Gellman, “Broad Effort Launched After ’98 Attacks,” Washington Post, December 19, 2001, A1; Whittle, Predator, 189; and Clarke, Against All Enemies. 409 Whittle, Predator, 147-150. 170

Republican from Virginia and the powerful chairman of the Armed Services Committee, had been encouraging the services to put combat UAVs into their budget proposals. The former

Secretary of the Navy had been a sailor during World War II and a Marine in Korea and attributed much of his perspective on defense matters to his military service.410 Warner’s SASC staff director, Les Brownlee, was a retired Army colonel, and it was he who first proposed pushing the services to be more aggressive in developing unmanned systems, particularly for conventional attack missions in ways that would replace manned aircraft. In a memo he wrote later to summarize his thinking in proposing the legislation, Brownlee explained, “While unmanned air vehicles had been around for several years, they were used almost exclusively for reconnaissance and intelligence gathering. The intent was to broaden the thinking for exploiting their use.”411 Both men explained that they were motivated by the idea that UAVs removed pilots from harm and, with pilots out of the cockpit, allowed aircraft to do things physically that the human body could not withstand.412

It all seemed eminently reasonable to Brownlee and Warner. But Warner was surprised to discover that the services were not interested in conventional applications of UAVs, nor in replacing manned fighters and bombers with unmanned aircraft. Years later, he explained:

I told the services give me your budget submissions for UAVs and they didn’t come to me and they didn’t put it in [budget proposals] and they fought it, and this went on for over 2 years. ‘Fellows I’m offering you money for what you want, tell me what you want!’ Finally, one night I was in a bar with some of these guys from the services and we’re joking, and this guy had a little bit too much to drink. He was in the Air Force and he said “what you don’t understand Senator is they look upon these things as taking our pilot seats away and we’re all about pilots. Since the beginning of the Air Force that’s what we’ve been about, pilots and pilot seats . . .

410 Senator John Warner, interview with the author, August 13, 2019. 411 Les Brownlee, (memorandum to Senator Warner, February 13, 2015), copy retained by the author. 412 George C. Wilson, “Senate Chairman Pushes Unmanned Warfare,” Government Executive, March 6, 2000, https://www.govexec.com/federal-news/2000/03/senate-chairman-pushes-unmanned-warfare/1977/. 171

I say, jeez, you’re kidding me. I said enough’s enough I’m going to require it by statute.413

Warner then, as he put it, took out “his shotgun and fired it into the heavens,” putting language into the 2001 NDAA requiring that as of 2010, one-third of all “deep strike” aircraft across the

Department of Defense would be unmanned.414 But Warner himself in the above statement reveals an interest in following the services’ preferences for UAV development and uses. It was only the services’ refusal to do so that exasperated Warner into taking legislative action without them.

What is also remarkable about Warner’s story is how completely disconnected it is from other UAV developments at DoD, and yet how important the military backgrounds of Warner and Brownlee are to the story—in that they felt their experience in uniform qualified them to challenge the services on this matter. It is an instance when civilians developed a preference for

UAV uses independently from the military, and yet felt empowered to do so because they were of the military, not outside it. Asked how he got the provision on UAVs into the final law,

Warner replied, “In those 30 years [I was a Senator] 70% of U.S. Senators were veterans and they were raised on duty, honor, country . . . We got a lot of things going.”415 In other words, his fellow Senators trusted him because he and they had worn a uniform, imbuing them with mutual confidence in their judgments about military matters. There is no way to know whether Warner could have or would have enacted the UAV provision in law if his background was in a purely civilian career, but his comment suggests that doing so would have been far more difficult.

413 Ibid. 414 Ibid.; The law also stipulated that one-third of ground combat vehicles would have no drivers by 2015. Floyd D. Spence National Defense Authorization Act for Fiscal Year 2001, Pub. L. 106-398, 114 Stat. 1654 (2000), https://www.govinfo.gov/app/details/PLAW-106publ398. 415 Warner, interview with the author, August 13, 2019. 172

Neither Warner nor Brownlee attributed their legislative advocacy for conventional strike-capable UAVs to the ongoing transformation debate, but they echoed it nonetheless. By

2000, parts of DoD were considering uses of UAVs in entirely novel ways, including

“swarming” tactics that overwhelmed adversaries with great numbers of small drone craft. But the technology to execute such missions reliably seemed far in the future.416 At best, the ideas about using UAVs for combat and other non-reconnaissance purposes contained in the transformation concept were finding their way into the Defense budget and the minds of decision-makers sporadically. And what snowballing was happening to add momentum to preferences for new uses of UAVs was about to be halted by the U.S. government’s response to the terrorist attacks on September 11, 2001.

Phase 3: Deployment (2001-2011)

In the fall of 2001, most military and civilian institutions still thought of UAVs as an intelligence and reconnaissance capability, with a vanguard of military veterans on the Hill starting to consider them a conventional combat capability as well. An expanded group of military advocates, especially in the Army and Air Force as well as those who had seen the

Predator in action in the Balkans, supported more investment in UAVs but were not shifting or expanding their preferred purpose for the systems, which was for ISR. It was still not occurring to most military actors that UAVs could do more than conduct surveillance missions, and plenty of military organizations still ignored or resented what they considered a niche, overly delicate technology. Civilians were similarly disposed and divided, with most civilian officials unaware of the hunt for bin Laden and the Predator’s track record in the Balkans, and therefore ignorant

416 Dombrowski and Gholz, Buying Military Transformation, 61. 173

of current UAV capabilities, let alone the potential alternative uses across the spectrum of operations. A few exceptions were those in the White House aware of the bin Laden search, and some advocates of the so-called transformation paradigm on the Hill.

By 2011, most military personnel who had been deployed to the wars in Afghanistan and

Iraq still preferred to use UAVs for information-gathering and the ISR preference was not new so much as stronger and more prevalent. Moreover, a new military preference for the use of

UAVs—lethal strikes—had emerged. But the strikes were not in a conventional military/ peer- to-peer context. Strike UAV capabilities in the post-9/11 wars were for unconventional warfare, meaning they targeted small numbers of people in environments where UAVs could operate without being shot down or even shot at.

As of 2011 civilians had taken these warfighter preferences as their own, relaxing their previous preferences about costs and jointness—and suppressing transformation-associated notions about UAVs’ uses in conventional combat. What civilians preferred when it came to

UAVs was to get them to the warfighter as quickly as possible in large numbers. This fact pitted civilian leaders against the service headquarters back in the Pentagon, which, on average, continued to prefer not to invest in UAVs. Civilians had hence taken sides with warfighters against the services, constrained by military thinking because they privileged military sources of information and judgment.

Details

In April of 2001, some four months into George W. Bush’s first presidential administration, the Washington Post noted the new civilian team’s pre-disposition toward UAVs.

“With a Bush administration looking to redirect military spending into more futuristic weapons

174

systems, unmanned combat aircraft are expected to be one of the big winners.”417 Indeed, as a candidate Bush declared that his military policy would be defined “not by mass or size but by mobility and swiftness” and project force “on the long arc of precision guided weapons.”418

Referred to as transformation, substantively it was not so different from the Revolution in

Military Affairs popularized among defense intellectuals in the 1990s. Regardless, the administration came into office interested in supplanting personnel with technology and in responding more rapidly to conditions around the world.

UAVs fit comfortably—indeed, centrally—into the ideas behind transformation. In a series of wargames sponsored by DoD’s Office of Net Assessment, unmanned combat aerial vehicles “featured prominently.” “In addition to conducting deep strike in denied areas,” the authors of the lessons learned report wrote, UAVs “were also used for counter-air missions, close air support of ground troops,” and a variety of other combat-oriented missions.419 It echoed

Senator Warner’s edict to shift the stable of “deep strike” aircraft toward unmanned systems. The

Navy and Air Force were duly issuing contracts to explore unmanned combat air vehicle

(UCAV) designs. UAVs were on a path to move beyond reconnaissance uses into a wider range of combat applications.

Meanwhile, other elements of the U.S. defense community were pushing the effectiveness of reconnaissance UAVs. Big Safari had figured out not only how to arm the

Predator with Hellfire missiles but also how to fly it remotely from control rooms in the United

States.420 Over the summer, cabinet-level national security council meetings at the White House

417 Greg Schneider, “Combat Planes that Leave Top Guns Home,” Washington Post, April 15, 2001, A1A. 418 George W. Bush, “Bush Campaign Speech,” C-SPAN video, 40:29, September 23, 1999, https://www.c- span.org/video/?152320-1/bush-campaign-speech&start=1504. 419 Michael Vickers and Robert Martinage, Future Warfare 20XX Wargame Series: Lessons Learned Report (Washington, DC: Center for Strategic and Budgetary Analysis, December 2001), 42, http://indianstrategicknowledgeonline.com/web/R.20011201.FutureWarXX.pdf. 420 Michel, “How Rogue Techies Armed the Predator”; and Whittle, Predator, 196-199. 175

considered whether to deploy the now-lethal Predator to Afghanistan to attempt to kill Osama bin Laden. The decision was deferred.

Then came the attacks on September 11, 2001. Just like that, policymakers were enthusiastic about deploying an armed UAV to Afghanistan to hunt terrorists. The first armed

Predator flew over Afghanistan on September 18, 2001.421 A little less than two months later, the

Washington Post reported “the United States is for the first time flying armed, unmanned aircraft into combat and controlling them with operators in the United States thousands of miles from the battlefield in Afghanistan.”422 Civilian preferences for UAVs’ uses became defined by the urgent requirements of war.

Civilian officials were pleased with what they saw. President Bush himself reportedly said of the Predator in an NSC meeting on October 10, 2001, “We ought to have 50 of these things.”423 In a speech to the Citadel later that year, Bush told the audience of cadets that the war in Afghanistan had made it “clear the military does not have enough unmanned vehicles. We’re entering an era in which unmanned vehicles of all kinds will take on greater importance.”424

Positive reviews came from lower levels of analysis as well. “In the first half of January [2002] an official ‘Lessons Learned’ panel traveled to Afghanistan” and concluded that “the star in the air campaign has been the lethal drone aircraft, or Unmanned Combat Aerial Vehicle

(UCAV).”425

421 The first Predator “entered Afghan airspace on Tuesday, September 18,” just seven days after the terrorist attacks in the United States. Whittle, Predator, 244. 422 Thomas E. Ricks, “U.S. Arms Unmanned Aircraft: ‘Revolution’ in Sky Above Afghanistan,” Washington Post, October 18, 2001, A1. 423 Bob Woodward, Bush at War (New York, NY: Simon and Schuster, 2003), 223. 424 “Bush’s Speech at the Citadel,” New York Times, December 11, 2001, https://www.nytimes.com/2001/12/11/international/bushs-speech-at-the-citadel.html. 425 Vernon Loeb and Thomas E. Ricks, “1’s and 0’s Replacing Bullets in U.S. Arsenal,” Washington Post, February 2, 2002, A1. 176

Yet the Predator was not the kind of UCAV the SASC and others in the transformation community had been imagining just a few months prior. The Predator was a medium altitude, long-endurance, light, vulnerable seeing and shooting machine that worked best against enemies who did not have sophisticated air defenses. It was an unconventional warfare ISR capability that could shoot a couple of missiles. It was not a combat system that could survive a sophisticated attack, certainly not a “deep strike” aircraft envisioned just months earlier.

But civilians in high level policy decisions were not thinking about such distinctions in the midst of the response to al Qaeda and the Taliban in Afghanistan. They just knew that warfighters wanted more of the capability. In the 2002 DoD Annual Report, Secretary Rumsfeld told Congress, “Unmanned surveillance and attack aircraft like Global Hawk and Predator offered a glimpse of their potential in Afghanistan. The 2003 budget increases the number of unmanned aircraft being procured and accelerates the development of new unmanned combat aerial vehicles capable of striking targets in denied areas without putting pilots at risk.”426 Many years later, Senator Warner’s SASC director observed that he did not know whether the services actually followed the requirement to make one-third of deep strike aircraft unmanned, “but clearly the emphasis of the legislation heightened the interest in these systems . . . and advanced more rapidly the critical technologies necessary for the development of the sophisticated systems in use today.”427 In the end, civilian enthusiasts were less concerned with what the military used

UAVs for than they were in simply getting UAVs into routine military use. Civilian enthusiasm may therefore have rushed UAVs to warfighters, but then warfighters used UAVs as they saw fit, and civilians liked the results—and liked that warfighters liked the results.

426 U.S. Department of Defense, Annual Report to the President and the Congress, FY2002 (Washington, DC: Department of Defense, 2002), 79, https://history.defense.gov/Portals/70/Documents/annual_reports/2002_DoD_AR.pdf?ver=2014-06-24-153732-117. 427 Brownlee, (memo to Senator John Warner). 177

For example, a November 2001 article in the New York Times reported “the aircraft, known as the RQ-1 Predator, has been widely praised by military commanders.”428 The cover letter for OSD’s 2005 Unmanned Aerial Systems Roadmap observed that the COCOMs “are requesting UAS in even greater numbers. Our challenge is the rapid and coordinated integration of this technology to support the joint fight.”429 The report itself, following the convention used by previous iterations of the roadmap, used capability shortfalls as defined by the combatant commands to determine what kind of UAVs to buy.430

It was a wartime demonstration of the same dynamic engineered into the ACTDs a decade earlier. Once the warfighter got a hold of UAVs, they began to dictate their uses and larger DoD acquisition plans. The Defense Science Board observed in its 2004 Summer Study on

UAVs that “once employed by warfighters, the value of UAVs becomes immediately evident, ideas for new operational concepts are spawned, a constituency is formed, and strong advocacy begins to build.”431 The same report observed the alignment between OSD and the Combatant

Commands, and suggested to the services that they should go along with the majority: “The Task

Force also notes that strong support for UAVs exists in the Office of the Secretary of Defense, with the regional Combatant Commanders, and in the Congress. Therefore, it makes sense for the Military Services to capitalize on the current positive climate and move out with dispatch.”432

Such language revealed that the services remained reluctant to invest in UAVs and civilians broadly sided against them. It also reveals that civilians could not prevail upon the services without the endorsement of warfighters—to the point where civilians let the warfighters

428 James Dao, “U.S. is Using More Drones, Despite Concern over Flaws,” New York Times, November 3, 2001, B2. 429 Office of the Secretary of Defense, Unmanned Aerial Systems Roadmap 2005—2030 (Washington, DC: August 2005), i, https://fas.org/irp/program/collect/uav_roadmap2005.pdf. 430 Ibid., 42. 431 Defense Science Board, Unmanned Aerial Vehicles and Uninhabited Combat Aerial Vehicles (Washington, DC: Department of Defense, February 2004), 9, https://apps.dtic.mil/dtic/tr/fulltext/u2/a423585.pdf. 432 Ibid., 9. 178

shape the strategic uses of capabilities for the sake of getting technology adopted. In the executive summary of their 2002 UAV Roadmap, the civilians in OSD had acknowledged service skepticism and again fell back on their faith in battlefield use of UAVs to win over the late adopters:

Taken as a whole, this technology area offers profound opportunities to transform the manner in which this country conducts a wide array of military and military support operations. As with any new technology, there is naturally some reluctance to transition to a radically new capability. The need to fully demonstrate UAVs in combat and realistic training environments is critical to the migration of this technology.433

By 2006, unprecedented warfighter demand for UAVs shaped the Quadrennial Defense

Review. The report writers highlighted successes such as investments in “new equipment, technology and platforms for the forces, including . . . unmanned vehicles”; the new Unmanned

Aerial Vehicles Squadron for SOCOM; a commitment to “increase procurement of unmanned aerial vehicles to increase persistent surveillance”; that the Army was going “to reinvigorate its aviation capabilities, including unmanned aerial vehicles”; and that the U.S. would “increase investment in unmanned aerial vehicles to provide more flexible capabilities to identify and track moving targets in denied areas.”434 To a great extent, this trend reflected the preferences of the ground forces more than the Navy or Air Force, the latter of which continued, as ever, to be of two minds about UAVs. As Robert Gates, Secretary of Defense from late 2006 to mid-2011 relayed in his memoir, the Army pushed hard for tactical UAVs even as the Air Force was only slowly building on its few Predator “combat air patrols” or CAPs.435

433 Office of the Secretary of Defense, Unmanned Aerial Vehicles Roadmap, 2002-2027 (Washington, DC: Department of Defense, December 2002), iii, https://www.hsdl.org/?abstract&did=446254. 434 U.S. Department of Defense, Quadrennial Defense Review Report (Washington, DC: February 2006), viii, 5-6, 42, 57, https://archive.defense.gov/pubs/pdfs/QDR20060203.pdf. 435 Gates, Duty, 128-130. 179

Overall, the pattern appeared to be that warfighter enthusiasm about UAVs formed the preferences of some civilians and simply enabled others’. Rumsfeld’s transformation agenda was certainly consistent with warfighter demand and vice versa. But Gates was not initially enamored of a more technologically dependent battlefield; it took conversations with troops who had deployed to Iraq and Afghanistan to change his mind.436 But thereafter, he was the major proponent for ISR (and therefore UAVs) as demanded by forces in the field. In his memoirs,

Gates recalls his conscious decision to give greater weight to the requests of the deployed force:

I directed [Chairman of the Joint Chiefs of Staff Admiral Michael Mullen] to develop and implement a process by which I would be informed immediately of any request specifically addressed to me by our commanders in Iraq and Afghanistan. The immediate problem that provoked those expressions of impatience was the difficulty we were having in meeting our field commanders’ need for intelligence, surveillance, and reconnaissance (ISR) capabilities.437

Gates expresses an interest in giving warfighters and their preferences direct access to him as Secretary of Defense. Importantly, and as Gates’ writing also suggests, the wars in Iraq and Afghanistan (and operations elsewhere) did not generate consensus among military organizations; rather, it simply rebalanced the numbers between detractors from and advocates for UAVs and empowered the latter with civilians. Civilian leaders found themselves constrained by the two major and long-running military preferences: to develop UAVs for ISR (née RSTA) purposes or to eschew them in favor of traditional manned capabilities. Although civilians now sided almost wholly with the pro-UAV camp, they still had to navigate entrenched service resistance to the capability.

The fight over whether to buy UAVs may also have been a reason for civilians to defer their preferences over what they should be for—ceding that ground to the side of the military that

436 General Jim Cartwright, interview with the author, July 16, 2019. 437 Gates, Duty, 127. 180

would advocate in favor of UAVs meant not having two simultaneous fights. Such an approach essentially attempted to give everyone what they wanted rather than pick winners and losers.

Rumsfeld’s first major budget pursued “relatively little shake-up in the big-ticket weapons programs and conventional-force structure that the administration came into office vowing to streamline.”438 At the same time, however, the Defense Secretary proposed “the development of unmanned long-range bombers as part of a spending plan that invests serious money in what he hopes will be the star weapon systems of future wars,” reconciling his priorities with the Air

Force by giving it “the largest budget of the services—$107 billion” along with “most of the

UAV and UCAV increases.”439

The fate of the Unmanned Combat Aerial System (UCAS) provides another piece of evidence that warfighter-civilian enthusiasm for UAVs was often countermanded by the military services, all of which worked hard to keep UAVs contained in their heavily reconnaissance- oriented mission set. Developed by DARPA and the Air Force, the UCAS design concept was first competed among defense industry companies in 1998.440 DARPA and the Navy also competed a design for a variant that could launch from an aircraft carrier.441 Both systems were to be used primarily for the suppression of enemy air defenses (SEAD) mission—essentially, to conduct the casualty-heavy job of destroying anti-aircraft missiles. Development proceeded after

9/11, and the UCAS appeared in the post-9/11 Defense Planning Guidance, the major document through which OSD shapes near- and mid-term military planning and budgeting decisions.442

438 Bradley Graham, “Bush to Propose Sustained Rises in Military Spending,” Washington Post, February 3, 2002, A6. 439 Vernon Loeb, “New Weapon Systems Are Budget Winners,” Washington Post, February 8, 2002, A29. 440 David Axe, “The Secret History of Boeing’s Killer Drone,” Wired Magazine, June 6, 2011, https://www.wired.com/2011/06/killer-drone-secret-history/. 441 Ibid. 442 “In some cases, goals addressed in this document have been directly cited in the DPG, such as the direction for development and demonstration of Unmanned Combat Air Vehicles.” See Office of the Secretary of Defense, Unmanned Aerial Vehicles Roadmap, 2002-2027, 61. 181

DARPA opened a Joint UCAS office in 2004 and an ACTD program was started the same year.443

Of course, the program was not the U.S. military’s first attempt to build a UAV expressly for combat purposes.444 But as the preceding evidence has shown, the armed forces did not sustain an interest in strike-capable UAVs until the Predator came along, were only comfortable with the Predator because its main purpose was reconnaissance, and had indeed remained committed to the RSTA/ISR purpose for UAVs overall. This commitment, combined with generalized service resistance to UAVs, doomed the J-UCAS. The 2006 QDR, despite its outpouring of endorsements for UAVs generally, announced the “restructuring” of the J-UCAS project.445 This evolved into the Air Force’s abandonment of the program, and eventually the

Navy’s as well. As Paul Scharre explained in his book about military robotics and autonomy, both the Navy and the Air Force had a “cultural resistance to combat drones.”446

The ultimate failure of the J-UCAS project compared with success in developing UAVs for ISR revealed that civilians could push UAVs into adoption only when the missions associated with those UAVs conformed with military preferences. As soon as civilians attempted to expand

UAVs into more conventional military missions (i.e., not reconnaissance nor strike-in- uncontested-airspace missions), the services’ resistance was too great.

Then came the transition from Rumsfeld to Gates, marking the final shift away from the transformation agenda and toward civilian preferences’ reliance on the military (warfighters) for the information that would shape civilians’ preferences for UAVs. It happened as part of the

443 Henry S. Kenyon, “Unmanned Combat Aircraft Program Takes Off,” SIGNAL Magazine, July 2004, https://www.afcea.org/content/unmanned-combat-aircraft-program-takes. 444 The WWI-era Kettering Bug was the first and most prominent U.S. effort at developing a UAV whose primary mission was striking enemy targets. Richard M. Clark, Uninhabited Combat Aerial Vehicles: Airpower by the People, For the People, But Not with the People (Maxwell Air Force Base, AL: Air University Press, August 2000). 445 DoD, Quadrennial Defense Review Report, 2006, 46. 446 Scharre, Army of None, 62. 182

administration’s approach to the unraveling war in Iraq. Encouraged by a civil-military cadre of counterinsurgency experts to dramatically adjust the military strategy to focus on protection of the population, President Bush sent a “surge” of forces to the country.447 At the same time, in a less well-publicized move, the Joint Special Operations Command began conducting systematic and lethal counterterrorism missions across Iraq. The counterinsurgency-counterterrorism combination was an information-thirsty campaign. “It was only after the military turned to new counterinsurgency techniques in 2007,” the New York Times reported, “that demand for drones became almost insatiable.”448 The deployed forces were discovering that UAVs—a capability long considered by the military to be best in reconnaissance situations—really were great reconnaissance assets, but also were handy for conducting lethal operations against terrorist suspects.

The services, however, lagged the warfighters, often clinging to their preference that

UAVs were a low investment priority. To get the Air Force in particular to step up its UAV game, Gates directed the uniformed chief of staff and the civilian secretary to give him a plan for increasing their UAV production. They responded with a proposal to stop funding the U-2 but keep up support for the manned F-22 fighter jet, an approach that Gates found defiant and frustrating.449 Gates eventually fired both men for an unrelated series of failures, and selected

General Norton (Norty) Schwartz to be the next Chief of Staff of the Air Force in part because

Gates believed Schwartz would be more responsive to his UAV direction. In his memoirs,

Schwartz blames the Air Force’s resistance to Gates on “the traditional ‘fighter/bomber’ culture

447 Max Fisher, “The Iraq Success Story that Propelled David Petraeus to the Top,” Washington Post, November 9, 2012, https://www.washingtonpost.com/news/worldviews/wp/2012/11/09/the-iraq-success-story-that-propelled- david-petraeus-to-the-top/?utm_term=.e9a0242d793f. 448 Christopher Drew, “For U.S., Drones are Weapons of Choice in Fighting Qaeda,” New York Times, March 17, 2009, A1, https://www.nytimes.com/2009/03/17/business/17uav.html. 449 Gates, Duty. 183

[that] seemed to cloud the leaders’ abilities to consider new paradigms” and recalls his own role as a bridge between a reluctant service and a demanding deployed force backed up by a fed-up

Secretary.450

Schwartz took a series of steps not only to increase UAV production, but also to institutionalize UAVs in Air Force culture, including by rewarding pilots of UAVs. He later described the internal tensions in the Air Force as the familiar split between warfighters using

UAVs and those with a bureaucratic service perspective. “People that were doing this [flying remotely piloted vehicles] got it. It was people who were on the outside [non-deployed/ non-

UAV crews in USAF and retired personnel] who were not privy to the events and their impact on the battlefield who had problems with it.”451

The episode with Secretary Gates and the Air Force appears, on one level of analysis, to be a case of a civilian forcing his preference on a reluctant service. But the sources of Gates’ preference to procure more UAVs for ISR purposes came from the warfighters. The struggle itself was between two preferences with origins in military institutions: the warfighters wanted

UAVs for ISR, and the Air Force did not want UAVs for much of anything. Gates made himself a champion of the former over the latter. Neither he nor the warfighters waging an unconventional war nor the Air Force leadership had any interest at this point in using UAVs as conventional combat platforms. The military did not raise alternatives, and neither did Gates.

Gates, of course, was explicitly focused on current wars, not future ones. The structure of the situation drove his selection of warfighter preferences.

450 Norty Schwartz and Suzie Schwartz with Ronald Levinson, Journey: Memoirs of an Air Force Chief of Staff (New York, NY: Skyhorse Publishing, 2018), 297, 255. 451 Norty Schwartz, interview with the author, August 8, 2019. 184

Importantly, the civilian preference to expand UAV capabilities into combat did not disappear under Gates, but it became latent compared to the years prior to his occupation of the office of Secretary of Defense. Indeed, then-Secretary of the Air Force Don Rice, sent a memo to then-Secretary Rumsfeld in early 2001 saying “a long-range, dwell capability for electronic combat—manned or unmanned—would be valuable in many circumstances.452 The 2006 QDR contained a declaration that “approximately 45% of the future long-range strike force” (aka bombers) would be transitioned to unmanned capabilities.453 Both moves echoed Senator

Warner’s demand in the national defense bill several years prior, suggesting that civilians had certainly developed a preference for making unmanned aerial capabilities part of the concept of operations for major conflicts.

As will be explored more thoroughly in the next section, a concern about the rise of

China was percolating among mid-and-upper levels of civilian management at the Pentagon long before the Obama administration’s so-called “pivot” or “rebalance to Asia” strategy emerged in

2011. Kathleen Hicks, who served as a civilian at the Pentagon under the Clinton, Bush, and

Obama Administrations, stated that concerns about China drove interest in using UAVs to replace or augment existing conventional strike capabilities. According to Hicks, OSD was

“worrying about China already—back in the Bush administration we were worrying about

China.”454 Other civilians outside OSD, particularly those at the Center for Strategic and

Budgetary Assessments (CSBA), which supplied many political appointees to the Pentagon under both the Bush and Obama administrations, made the connection between China and

452 Quoted in Kelsey D. Atherton, “How Rumsfeld Viewed the Most Iconic Aircraft of the War on Terror,” C4ISRNet, January 25, 2018, https://www.c4isrnet.com/unmanned/2018/01/25/how-rumsfeld-viewed-the-most- iconic-aircraft-of-the-war-on-terror/. 453 DoD, Quadrennial Defense Review Report, 2006. 454 Kathleen H. Hicks, interview with the author, September 6, 2019. 185

unmanned bombers.455 This line of thinking seems to have been near-exclusively a civilian one, although there are hints that a minority of uniformed and former military allies were thinking in parallel with civilians. General John Jumper—the same man who directed the Predator be armed—told Air Force Magazine in 2005 that he imagined an unmanned bomber could “loiter” over targets.456

Regardless by 2007 the Pentagon’s civilian leadership, at the behest of Robert Gates, had generally adopted warfighter preferences for UAVs in a way that almost completely crowded out alternative UAV uses. Although Gates relented to those in OSD still concerned about future battles with China and directed that the new bomber be “capable of manned and unmanned operations,” such efforts acted as placeholders for a time when the Department’s priorities were not focused on battles in the Middle East and South Asia.457

The warfighters’ demands for ISR-focused UAVs had not abated by 2009 when the new

Obama administration entered office. One former senior OSD official under Obama confirmed that upon entering office, the Obama team learned first-hand that “the demand from CENTCOM

[the U.S. Central Command, covering the Middle East and parts of South Asia] for ISR in

Afghanistan was insatiable.”458 In the preceding administration, the effort to meet that demand was herculean. The Army alone “had gone from eight unmanned aerial vehicles (UAVs) in 2003

455 See for example, former Bush administration Deputy Assistant Secretary of Defense for Force Transformation Mark Gunzinger’s report, Sustaining America’s Strategic Advantage in Long-Range Strike (Washington, DC: Center for Strategic and Budgetary Assessments, 2010), https://csbaonline.org/research/publications/americas-strategic- advantage-long-range-strike. 456 Quoted in John A. Tirpak, “Toward an Unmanned Bomber,” Air Force Magazine, June 2005, https://www.airforcemag.com/article/0605bomber/. 457 Office of the Secretary of Defense, “Long Range Strike and Penetrating Bomber Capabilities,” in Audit of the Acquisition of the Long-Range Strike Bomber (Washington, DC: Department of Defense, September 2015), https://www.documentcloud.org/documents/4177863-DOD-OIG-Report-2015-170-Audit-of-the-Acquisition.html. 458 Senior OSD Official, interview with the author, June 3, 2019. 186

to 1,700 in Iraq in 2008.”459 An Air Force briefing documented a 660% increase in Predator and

Reaper combat air patrols between 2004 and 2009 (the result of Schwartz’s strong-arming).460

Now, the Obama team—still led by Gates at the Pentagon—assumed the mantle of getting UAVs to the warfighter with no less conviction. In May of 2009, David Ahern of OSD-

AT&L told the House Armed Services Committee’s Subcommittee on Air and Land Forces that

“by 2016 the Air Force will have procured sufficient MQ-9s to provide at least 50 continuous

Combat Air Patrols. These unmanned, high-endurance platforms are well suited in important and unique ways for irregular warfare operations.”461 The New York Times reported in January of

2010 that “the number of mission hours flown by drones in Afghanistan and Iraq” had risen to over 200,000 hours in 2009, up from approximately 75,000 hours in 2007.462 And it was not just that the numbers were still climbing; wartime strategy had come to rely on UAVs as well. When

Obama announced his own troop surge to Afghanistan at the end of 2009, “military officers said that they could maintain pressure on insurgents in remote regions by using surveillance drones.”463 By 2012, the Congressional Research Service reported that “demand for UAS capabilities in the field is essentially unconstrained.”464

Preferences for the uses of UAVs remained the same: warfighters wanted UAVs primarily for ISR and support functions—such as detecting improvised explosive devices—and so did civilians. UAVs became synonymous with ISR. “Navy Outlines Future UAV Strategy,

459 Gates, Duty, 233. 460 David Deptula, “Air Force Unmanned Aerial Systems Flight Plan, 2009-2047,” (presentation, Headquarters, U.S. Air Force, Washington, DC, n.d.), https://apps.dtic.mil/dtic/tr/fulltext/u2/a504096.pdf. 461 David Ahern, Air Force Modernization Programs, Statement before the Subcommittee on Air and Land Forces of the House Armed Services Committee, Hearing on Budget Request for Air Force Modernization Programs, Hearings on National Defense Authorization Act for Fiscal Year 2010 and Oversight of Previously Authorized Programs, 111th Cong., 1st sess., May 20, 2009. 462 Christopher Drew, “Drone Flights Leave Military Awash in Data,” New York Times, January 11, 2010, A1. 463 Sheryl Gay Stolberg and Helene Cooper, “Obama Speeds Troops to Afghanistan and Vows to Start Pullout by 2011,” New York Times, December 2, 2009, A1. 464 Gertler, U.S. Unmanned Aerial Systems, 6. 187

First Focus is on Long Dwell ISR” one industry paper’s headline read.465 Representative Adam

Smith remarked to administration officials in a hearing on “the expanding number of UAVs and other ISR platforms.”466

But the deployed military—and, importantly, the CIA—had made an important addition to the reconnaissance preference: the strike or “hunter-killer” missions of the kind the Predator first introduced also became a foremost feature of how civilians and the military thought about

UAVs. The CIA’s involvement meant the civilian policymaker preference to use UAVs to kill terrorists was not a closed loop with military preferences as was the case with ISR uses. Instead, a division of labor for targeted killings using UAVs between the CIA and DoD meant that two institutions helped to establish and reinforce a broad foreign policy community preference to use

UAVs to strike terrorist targets.467 This created a norm for what UAVs should do even as it diverted civilian policymaker attention away from alternative uses of UAVs.

The diversion of attention that might have been directed at broader uses of UAVs occurred because the involvement of the CIA shifted the focus of preferences about UAVs away from the strategic and military appropriateness of their applications to the legal, accountability, and transparency implications of doing so. Although the claim that lawyers were involved in target selection is likely apocryphal, civilian officials spent far more time determining the legality of strikes than their strategic uses and alternatives to them.468 As one former Obama official explained, “by framing these questions as legal ones (or to be arbitrated by lawyers), it

465 “Navy Outlines Future UAV Strategy, First Focus is on Long Dwell ISR,” Defense Daily International 3, no. 16 (February 2002). 466 Adam Smith, Hearing on National Defense Authorization Act Fiscal Year 2012 and Oversight of Previously Authorized Programs, Budget Request from the Department of the Air Force before the House Armed Services Committee, 112th Cong., 1st sess., February 17, 2011. 467 Jonathan Masters, “Targeted Killings,” Council on Foreign Relations, May 23, 2013, https://www.cfr.org/backgrounder/targeted-killings. 468 General (ret.) Norty Schwartz, former Chief of Staff of the Air Force, interview with the author, August 8, 2019. 188

created the possibility that policymakers might tend to turn to legal analysis first and judgment second.”469

In public, justifications for drone use usually came back to wartime legal authorities.

Harold Koh, General Counsel to Secretary of State Hillary Clinton, explained in a speech in

March of 2010 that UAV strikes on suspected terrorists were authorized by Congress by the

2001 Authorization for the Use of Military Force (AUMF). He stated, without referring to non- military uses of UAVs, “the United States is in an armed conflict . . . in response to the horrific

9/11 attacks, and may use force consistent with its inherent right to self-defense under international law.”470 He argued that unmanned capabilities made no difference to the legality of counterterrorism strikes. The DoD’s General Counsel Jeh Johnson echoed Koh in a speech at

Yale University two years later. “The rules that govern targeting do not turn on the type of weapon system used, and there is no prohibition under the law of war on the use of technologically advanced weapons systems in armed conflict.”471

Legal questions were also wrapped up in whether a particular strike was to take place in or outside of “areas of active hostilities,” the legalese term for warzones. The division of labor between the CIA and DoD was, in part, based on the different legal authorities that these institutions had and how they matched up with the diplomatic context of the strikes, with the

DoD responsible for warzone uses of UAVs. Although civilians at the White House understood that the CIA and the military were very different institutions with different purposes, when it

469 Loren DeJonge Schulman, Behind the Magical Thinking: Lessons from Policymaker Relationships with Drones (Washington, DC: Center for a New American Security, July 2018), 16, https://www.cnas.org/publications/reports/behind-the-magical-thinking. 470 Harold Hongju Koh, “The Obama Administration and International Law,” (speech, American Society of International Law, Washington, DC, March 25, 2010), https://2009-2017.state.gov/s/l/releases/remarks/139119.htm. 471 Jeh Charles Johnson, “National Security Law, Lawyers and Lawyering in the Obama Administration,” (Dean’s lecture, Yale Law School, New Haven, CT, February 22, 2012), https://ylpr.yale.edu/sites/default/files/YLPR/johnson_national_security_law_lawyers_and_lawyering_in_the_obam a_administration.pdf. 189

came to UAVs the purpose was assumed while the justification for each individual use was the germane topic. President Obama revealed this framework for thinking about UAV strikes in his

2013 speech announcing his new Presidential Policy Guidance, “Procedures for Approving

Direct Action Against Terrorist Targets Located Outside the United States and Areas of Active

Hostilities.”472 He succinctly asserted the effectiveness of drone strikes, while spending he balance of the speech assuring the public that “the use of drones is heavily constrained” and conducted within legal parameters.473

In fact, as much as the Obama White House worked to control the use of armed UAVs, civilian officials there accepted the basic premise that UAVs were meant for HVT strike missions and ISR, precisely the framing the military supplied. In the end, the formal institutional responsibility for a particular strike may not have much affected what White House civilians thought UAVs were primarily for: protecting American security by killing terrorists. Rahm

Emanuel, the President’s first chief of staff, “showed an intense interest in the drone strikes and called CIA Director Leon Panetta regularly with one question: ‘Who did we get today?’”474 This idea simply crowded out other possible uses for UAVs, and the fixation on HVT strikes narrowed the idea of UAV technology into the drone category. In her analysis of civilians’ views of drone policy, Loren Schulman found, “demand rather than effectiveness and appropriateness often set the pace for drone allocation under the Obama administration, and the efficacy of strategies associated with drones is inconclusive.”475

472 “Procedures for Approving Direct Action Against Terrorist Targets Located Outside the United States and Areas of Active Hostilities,” May 22, 2013, https://www.justice.gov/oip/foia- library/procedures_for_approving_direct_action_against_terrorist_targets/download. 473 Barack H. Obama, “Remarks by the President at National Defense University,” (remarks, Fort McNair, Washington, DC, May 23, 2013), https://obamawhitehouse.archives.gov/the-press-office/2013/05/23/remarks- president-national-defense-university. 474 Woodward, Obama’s Wars, 284. 475 Schulman, Behind the Magical Thinking, 1. 190

Hence, as demand from the warfighters for ISR and hunter-killer UAV capabilities remained high, civilian preferences matched it. OSD AT&L’s 2012 Report to Congress on UAS

Training and Sustainment noted the continued reliance of civilian preferences on warfighter preferences for the uses of UAV capabilities:

The Department of Defense (DoD) continues to increase its investment in unmanned aircraft systems (UAS) to meet battlefield commanders’ demand for their unique capabilities. The emphasis on long-endurance, unmanned intelligence, surveillance and reconnaissance (ISR) assets - many with strike capabilities - is a direct reflection of recent operational experience and further Combatant Commander demands.476

It was an explicit expression of civilians basing their preferences on information derived from the military, drawing civilian and military preferences for the uses of UAV capabilities into overlapping positions. Meanwhile, civilians at the White House were more concerned by whether drone strikes were legal—and the number of HVTs killed—than the broader military uses of UAV capabilities.

Phase 4: Offset (2013-2016)

With warfighter demand still high, the dedication of UAVs to unconventional operations and counterterrorism seemed to have an endless inertia. And then, a puzzling thing happened: civilian preferences began to shift back toward using UAVs as a combat capability for conventional war. Whereas in 2012 civilian preferences for UAV uses were still largely conditioned on warfighter preferences, by 2016 a significant cadre of civilians preferred to think of UAVs much more expansively, including as a capability freed from its mimicking of manned

476 Under Secretary of Defense for Acquisition, Technology, and Logistics. Department of Defense Report to Congress on Future Unmanned Aircraft Systems Training, Operations, and Sustainability. Washington, DC: Department of Defense, April 2012. https://fas.org/irp/program/collect/uas-future.pdf. Unified Combatant Command for Special Operations Forces, U.S. Code 10 (2011), § 167., ii 191

platforms and its exclusion from more conventional combat scenarios. Much of the civilian thinking picked up where transformation had left off. Military institutions, meanwhile, largely maintained their preference to use UAVs for reconnaissance/ISR and strike, and other support functions.

The retirement of Robert Gates as Secretary of Defense on June 30, 2011 provides a first explanation for the shift. As explored above, Gates personally suppressed growing civilian interest in pushing the uses of UAVs beyond reconnaissance and unconventional warfare. A drop in the number of deployed forces also freed civilians to think beyond Iraq and Afghanistan. Most

American troops were withdrawn from Iraq by the end of 2011, and between 2012 and 2014 force levels in Afghanistan also dropped significantly.477 Although the U.S. military was not really “out” of either operational theatre, the steep force reductions meant discussions about the demand for UAVs could and did shift. Moreover, both the domestic and broader international contexts were shifting as far as many senior civilian and military officials in the Pentagon and the White House were concerned, with renewed focus on the domestic economy and the U.S. role in the Asia-Pacific resetting the terms of defense strategy.

This final phase of the case shows how civilian preferences changed, highlighting especially the variability of warfighter political power and the power of their preferences over civilian preferences. Yet when wartime concerns abated and the armed services and civilian acquisition officials again formed the key civil-military relationship, with technical expertise and debates over the future of warfare overtaking contemporary combat expertise in importance, civilians formed preferences that were independent of the military.

477 Heidi M. Peters, Department of Defense Contractor and Troop Levels in Afghanistan and Iraq: 2007-2018, CRS Report No. R44116 (Washington, DC: Congressional Research Service, May 10, 2019), https://fas.org/sgp/crs/natsec/R44116.pdf. 192

Details

In 2011, two important events affected the course of preferences about UAV capabilities.

First, the U.S. failed to negotiate a new status-of-forces agreement with Iraq, precipitating the

U.S. withdrawal from the country by the end of December.478 Second, the fiscal consequences of the 2008 economic crash finally began to impact the Department of Defense.

The latter event happened circuitously. As the nation approached the so-called debt ceiling, the maximum allowable national debt by law, Democrats on the Hill wanted to pass legislation to raise the ceiling while Republicans preferred to cut government spending. A compromise brokered in the Senate put upper limits on the discretionary portion of the federal budget—those departments and programs whose budgets are determined on an annual basis rather than codified in law. The compromise would cut $1 trillion from the DoD budget over 10 years.479

The Department responded to the budgetary pressure, in part, by adjusting its strategy. In

January of 2012, it issued its Defense Strategic Guidance, a Presidential and Secretary of

Defense-level document designed to grapple with the twin challenges of sequestration and the instability in the international environment—what President Obama called “a moment of transition” in his cover letter for the strategy. UAVs remained a priority, as Obama highlighted

“the capabilities critical to future success, including intelligence, surveillance, and

478 Tim Arango and Michael S. Schmidt, “Last Convoy of American Troops Leaves Iraq,” New York Times, December 18, 2011, A6, https://www.nytimes.com/2011/12/19/world/middleeast/last-convoy-of-american-troops- leaves-iraq.html. 479 Brendan W. McGarry, The Defense Budget and Budget Control Act: Frequently Asked Questions, CRS Report No. R44039 (Washington, DC: The Congressional Research Service, July 2018), https://fas.org/sgp/crs/natsec/R44039.pdf; Todd Harrison, “What Has the Budget Control Act of 2011 Meant for Defense?,” Center for Strategic and International Studies, Critical Questions, August 1, 2016, https://www.csis.org/analysis/what-has-budget-control-act-2011-meant-defense. 193

reconnaissance.”480 But the uses of ISR were no longer inexorably tethered to operations in Iraq and Afghanistan. The document’s drafting had been a close collaboration between the Joint Staff and civilians in OSD-Policy. This meant that although the demand for more economy at DoD came from the civilian political process, the military was heavily involved in the internal

Pentagon debates about how best to engineer budget cuts.481

As senior leaders’ attention began to shift to rebalancing the investment portfolios in

DoD, it became increasingly clear that the rapid wartime acquisition process had precluded broader thinking about how defense capabilities and policy priorities matched up. Writing in their 2013 UAV Roadmap, OSD authors observed the tendency of wartime contracting to outpace longer-term strategic thinking, noting although “unmanned systems were rapidly developed to meet the immediate needs of the warfighter in the near term, they have not undergone rigorous requirements review.”482

At the same time, the Obama administration was in the midst of implementing its “Pivot” or “Rebalance” to Asia policy. Launched by Secretary of State Hillary Clinton, the rebalance was designed to recognize the importance of Asia to the American economy and security.483 To DoD, the shift in emphasis from the Middle East to Asia had a major implication: the department had to think about how to fight a war with an increasingly militarily capable China. As explored earlier, this was a concern that had been simmering among civilian defense intellectuals in and out of the Pentagon since the George W. Bush administration. Now that the pressures of the wars

480 U.S. Department of Defense, Sustaining U.S. Global Leadership: Priorities for 21st Century Defense (Washington, DC: Department of Defense, January 2012), https://archive.defense.gov/news/Defense_Strategic_Guidance.pdf. 481 Kathleen H. Hicks, interview with the author, September 6, 2019. 482 U.S. Department of Defense, Unmanned Systems Integrated Roadmap FY2013-2038 (Washington, DC: Department of Defense, January 2014), 20, https://apps.dtic.mil/dtic/tr/fulltext/u2/a592015.pdf. 483 Hillary Clinton, “America’s Pacific Century,” Foreign Policy, October 11, 2011, https://foreignpolicy.com/2011/10/11/americas-pacific-century/. 194

in Iraq and Afghanistan at least appeared to be abating, and the pressures on the Defense budget appeared to require genuine prioritizing, the potential for Chinese aggression rose to the top of

Defense officials’ minds.

But after over a decade of unconventional war and several years of budgetary instability, the U.S. defense community was not confident in the military’s readiness to defeat nor deter a

“peer competitor.” Obama’s then-Deputy Secretary of Defense Ashton Carter was among those concerned. One day he called Dr. Will Roper, a civilian analyst with the Missile Defense

Agency, and asked him to build an office at the Pentagon that would focus on technological competition with China.484 That office became the Strategic Capabilities Office, or SCO.

Coming from the perspective that outpacing China in technological development would be key to competing with Beijing, the office’s approach was to take existing military capabilities and use them in novel ways. As then-Deputy Secretary of Defense Robert Work explained it in 2016,

SCO “takes systems the Defense Department already has tremendous investments in and transforms, or repurposes, them for use in ways the world has never seen.”485 One of these capabilities was UAVs. Because much of the American problem in fighting a war with China was vulnerability, it struck Roper as most rational to rely on systems that could be produced and lost in large numbers.486 The people at SCO set about exploring novel ways to put UAVs into a conventional war fight, including by making them small and teaching them to “swarm” using artificial intelligence (AI).487 By October of 2016, SCO demonstrated both swarming technology

484 Will Roper, assistant secretary of the Air Force for Acquisition, former director, Strategic Capabilities Office, interview with the author, August 14, 2019. 485 Cheryl Pellerin, “Deputy Secretary: Third Offset Strategy Bolsters America’s Military Deterrence,” Department of Defense, October 31, 2016, https://www.defense.gov/Explore/News/Article/Article/991434/deputy-secretary- third-offset-strategy-bolsters-americas-military-deterrence/. 486 Ibid. 487 Dan Lamothe, “Veil of Secrecy Lifted on Pentagon Office Planning ‘Avatar’ Fighters and Drone Swarms,” Washington Post, March 8, 2016, https://www.washingtonpost.com/news/checkpoint/wp/2016/03/08/inside-the- secretive-pentagon-office-planning-skyborg-fighters-and-drone-swarms/. 195

and the manned-unmanned teaming concept by launching 103 small UAVs from three manned fighter jets flying over the Naval Air Weapons Station in China Lake, California.488

Together, the DSG and the SCO reflected and contributed to a process of reorienting

DoD toward large-scale conventional warfare, even as it tried to retain its counterterrorism capabilities. In 2014, outgoing Secretary of Defense Chuck Hagel explained that despite the fatigue of thirteen years of war and the fiscal pressures on DoD, “we face the rise of new technologies, national powers, and non-state actors; sophisticated, deadly and often asymmetric emerging threats, ranging from cyberattacks to transnational criminal networks; as well as persistent, volatile threats we have faced for years.”489

The name for the innovation framework that Deputy Secretary Work attempted to implement was called the Third Offset.490 Work explained to a European audience in 2016, “we seek ways in which to offset our potential adversary's advantages.” It was, he argued, the third in a series of technological offsets the U.S. had successfully pursued in the twentieth century. Work went on, “the technological sauce of the Third Offset is going to be advances in Artificial

Intelligence (AI) and autonomy.”491 AI and autonomy were, in many ways, evolutions in unmanned systems. Autonomy had long been a feature of unmanned technology but had usually involved preset commands and pre-determined flight paths. Now, the civilian leaders of the

Department of Defense were seeing a future where robots could make their own choices to speed

488 U.S. Department of Defense, “Department of Defense Announces Successful Micro-Drone Demonstration,” press release, January 9, 2017, https://www.defense.gov/Newsroom/Releases/Release/Article/1044811/department- of-defense-announcessuccessful-micro-drone-demonstration/. 489 Chuck Hagel, “Reagan National Defense Forum Keynote Address,” (speech, Ronald Reagan Presidential Library, Simi Valley CA, November 15, 2014), https://dod.defense.gov/News/Speeches/Speech- View/Article/606635/. 490 Secretary of Defense Ash Carter preferred the term “Defense Innovation Initiative,” although many of the ideas overlapped. 491 Robert O. Work, “Remarks by Deputy Secretary Work on Third Offset Strategy,” (speech, Brussels, Belgium, April 28, 2016), https://www.defense.gov/Newsroom/Speeches/Speech/Article/753482/remarks-by-deputy- secretary-work-on-third-offset-strategy/. 196

reactions to adversary offensive moves.492 As Shawn Brimley, a former Pentagon and NSC staffer during the Obama administration and close advisor to Work put it, “a critical component of addressing the challenges our forces will face . . . will revolve around fully harnessing the possibilities inherent in unmanned and increasingly autonomous (robotic) systems.”493 Prior to

Work’s return to government as Deputy Secretary of Defense, he and Brimley also published a think tank report that expressed their own preference for using UAVs not just for reconnaissance, but to engage and destroy enemy targets, predicting “a new war-fighting regime in which . . . unmanned and autonomous systems have become central to combat.”494

The key actors in all of these developments were civilians. Inside the Pentagon, “it really was a small group of civilians in OSD who pushed it,” said Work’s Special Assistant, Greg

Grant, citing Frank Kendall as the AT&L head, Kendall’s Director for Research and Engineering

Steve Welby, and Will Roper at SCO.495 Work himself confirmed that he consulted former

Secretaries of Defense and other high-level civilians to develop his ideas about the changing technology of warfare—particularly Bill Perry and Harold Brown, who introduced him to the term “offset”—and that he could not recall anyone in uniform playing a role in his thinking or working to persuade him to push for different uses of UAVs.496 More broadly, the rebalance to

Asia was a State Department initiative, the policy toward Iraq came from bilateral diplomacy and from the White House, the budgetary pressures were a product of domestic and congressional

492 Scharre, Army of None. 493 Shawn W. Brimley, The Third-Offset Strategy: Securing America’s Military-Technical Advantage, Testimony before the House Armed Services Committee Subcommittee on Seapower and Projection Forces, 113th Cong., 2nd sess., December 2, 2014, https://s3.amazonaws.com/files.cnas.org/documents/Brimley_HASC_PreparedStatement_12022014.pdf?mtime=201 60906080438. 494 Robert O. Work and Shawn Brimley, 20YY: Preparing for War in the Robotic Age (Washington, DC: Center for a New American Security, January 2014), 6, https://www.cnas.org/publications/reports/20yy-preparing-for-war-in- the-robotic-age. 495 Greg Grant, interview with the author, September 9, 2019. 496 Robert O. Work, interview with the author, September 16, 2019. 197

politics. This civilian mobilization around broader uses of UAVs had been building for over a decade, and marked, finally, a break in civilian dependence on military preferences for the uses of UAVs. As one former senior Pentagon official explained: “[With] the rise of China [and the] pivot toward Asia, the Air Force attitude changed—all of the services’ attitudes changed—from, we’re paying for equipment for the Agency to use to, oh, these are critical to our operations.”497

But are the services truly embracing these new uses of UAVs and putting service budgets and operational concepts behind them, or is there still meaningful resistance to civilian preferences? Work lamented that they are not.498 As of this writing, military doctrine for UAVs was still underdeveloped. The Joint Staff has no stand-alone doctrinal document on either unmanned combat or ISR. The Air Force has issued an annex to its “core doctrine” entitled

“Global Integrated ISR Ops” which focus on the intelligence functions of unmanned systems.499

Combat functions are not highlighted, and one has to search thoroughly to discover them at all.

The Navy still does not have doctrine for UAVs, and “has affirmed its preference for traditional manned assets within the carrier air wing” by “relegating the Unmanned Carrier-Launched

Airborne Surveillance and Strike (UCLASS) program to a refueling role.”500 Roper and Grant both stated that the services have still needed convincing. “We did that swarming demonstration out at China Lake to make the services believe,” Roper said.501 “The real obstacle” to Third

Offset thinking, Grant added, “was the services.”502 Military institutions seemed to be lagging civilian preferences on the strategic purposes of UAVs.

497 Former senior Pentagon official, interview with the author, June 3, 2019. 498 Robert O. Work, interview with the author, September 16, 2019. 499 “Global Integrated ISR Operations,” U.S. Air Force, January 29, 2015, https://www.doctrine.af.mil/Doctrine- Annexes/Annex-2-0-Global-Integrated-ISR-Ops/. 500 Daniel M. Marzluff, “Flying Trucks? NO. Strike Assets? YES!” U.S. Naval Institute, Proceedings 142, no. 9 (September 2016), https://www.usni.org/magazines/proceedings/2016/september/flying-trucks-no-strike-assets-yes. 501 Will Roper, interview with the author, August 14, 2019. 502 Greg Grant, interview with the author, September 9, 2019. 198

And yet Roper, Grant, and others see more complexity in the exchange of preferences between civilian and military organizations. Roper noted that the current Chief of Staff of the Air

Force, General Goldfein, specifically recruited him into the Air Force acquisition role to help the service embrace UAVs along with other new technologies. Grant observed that Deputy Secretary of Defense Work “found a fellow-traveler in [Vice Chairman of the Joint Chiefs of Staff]

Admiral Winnefeld.”503 When it comes to civilians working with the military, Roper said, “I do feel like the innovators find each other . . . Everything feels like a collaborative decision.”504

Findings and Analysis

Military preferences for the uses of UAVs conditioned civilian preferences for sixty years. During this overwhelming majority of the study period, the military’s two major preferences regarding UAV capabilities were to use them for reconnaissance purposes or not to use them at all. For decades, civilians ceded development of and thinking about military UAVs to the services—and to DARPA under highly classified and therefore barely acknowledged experimental programs. Except for John Warner’s “shot into the heavens” over UAVs in 2000, civilians did not press the services to dedicate UAVs for purposes other than reconnaissance— and even Warner and his staff showed little concern that the focus on “deep strike” platforms in the legislative language went ignored. Until, first, Secretary Gates’ demand for an unmanned bomber design in 2012 and, more prominently, the Third Offset arose in 2014, UAV capabilities only extended into strike missions when the services decided to order or modify existing aircraft for that purpose.

Why did civilians adopt the military preference to limit UAV uses to reconnaissance?

503 Ibid. 504 Will Roper, interview with the author, August 14, 2019. 199

Civilian inattention or different attention is one major reason that military preferences for UAV purposes dominated civilian preferences. For many years, the civilians involved in UAV development, coming from the acquisition and budgeting communities, were motivated by their positions in the bureaucracy to focus on the feasibility of technology and the politics of the financial investments required rather than the strategic demands for or uses of UAVs. At the same time, UAVs simply were not major enough programs to attract civilian attention, demonstrated by their absence from national strategic documents until the 1980s. Even periodic crises simply called civilian attention to intelligence uses for UAVs rather than battlefield purposes, reinforcing the association between UAVs and reconnaissance missions. In sum, the services paid attention, albeit episodic, to UAVs while civilians simply did not.

The two major phases of UAV evolution—development and deployment—also help explain the ways civilian preferences became more and more strongly conditioned on military preferences. During the development phase, civilians had more independence from military preferences relative to deployment, at least enough to incorporate other factors into their preferences. During deployment, civilian preferences became more explicitly bound to military preferences. Both outcomes had to do with which civilians were involved in development and deployment, the conditions under which they were involved, and the internal divisions within and relative power and salience among the services.

First, the divisions among the military activated civilians’ adjudication function, confining civilians to information supplied by military groups and choices among military preferences by seeming to present the reasonable range of uses according to the expert judgment of the military. Although these divisions limited civilians to military preferences, the fact of a bargaining range between sub-groups of the military gave civilians the flexibility to incorporate

200

other factors into their own preferences, including costs, constituent interests, and competition with other DoD programs.

However, civilians struggled to impose these non-military preferences for UAV development on the military services, because the services still exercised more direct control over deployed forces and thus could sustain institutional positions. The prime example of this dynamic was Secretary of the Navy Lehman’s effort to get his service to adopt UAVs in 1984.

Then came the Goldwater-Nichols Act, which gave the combatant commanders direct lines to the Secretary of Defense and the President and removed the services from operational command. Seemingly by accident of history, civilians turned to models that deployed capabilities more rapidly in order to gain leverage over the chiefs just as the combatant commands were consolidating power inside the DoD and confronting the challenges of Military Operations Other than War. The demands of MOOTW led the COCOMs to articulate requirements that continued to limit UAVs to reconnaissance roles, but civilians—again, still those in the acquisition realms— were less interested in UAVs’ uses than the fact of their procurement. Civilians essentially traded the independence of their preferences for the uses of UAVs in order to force the services to buy them at all.

Next, the UAV case reveals the different dynamics between military and civilian preferences under conditions of relative peacetime and wartime. Because peacetime development involved the acquisition and budgeting sides of the Pentagon and the Hill, civilians paid more attention to costs and technological feasibility, working to force the services to adopt UAVs regardless of their operational and strategic uses. Thus, the uses of UAVs evolved out of service preferences and how the services interpreted operational experiences in the Balkans. Civilians did seem to be paying more attention to the broader uses of UAVs in debates over

201

transformation, but the attacks on 9/11 interrupted that process. Warfighters quickly became enamored with UAVs for reconnaissance, close air support, and strikes on individuals. The domestic politics of wartime in the U.S. coupled with the overwhelming advocacy for reconnaissance UAVs from the warfighters short-circuited civilian preference formation to make military preferences a sufficient source of information to shape civilians’ own UAV-related desires.

Civilians, eager to convince the services of the general merits of UAV technology, were only too happy to provide warfighters with more UAVs. Meanwhile, the services pushed back against other uses of UAVs that civilians had been pursuing prior to 9/11. Because civilians were at least getting more UAVs and were meeting warfighter demands, they relented (for a while) in their preference to use UAVs for combat purposes as well.

Importantly, even without a binary shift from peace to war, whether the U.S. had deployed forces using UAVs at all also strengthened the association between military and civilian preferences. Even prior to the war on terrorism, civilians tended to defer more to combatant commands’ preferences for UAVs than to the services—and civilian deference only increased during the GWOT. Given that many of the COCOMs have forces deployed to their theaters of operations on a steady rotational basis, there is a standing potential for forces in the field to seize the initiative and begin to shape civilian preferences even in the absence of combat.

This finding suggests the inclusion of the COCOMs in UAV capabilities development gave military preferences more influence over civilian preferences. This does not appear to be a direct result of the Goldwater-Nichols DoD reforms, but reflects the conscious civilian choice to outsource the justifications for capabilities to warfighters and is consistent with the general increase in COCOM power and relevance in DoD budgeting and operations over time. In other

202

words, Goldwater-Nichols made COCOM power a possibility, and civilian choices made it a reality.

The findings thus far may erroneously suggest that UAV advocacy was a widespread phenomenon. This is not the case. Instead, powerful individuals, civilian and military, acted as champions for UAV capabilities over the course of the study. Those few advocates for using

UAVs at all (which meant using them for reconnaissance) who succeeded did so by building small but powerful civil-military, interagency, and inter-branch coalitions. Related to but adjusting the concept of civilians finding military mavericks, it appears more accurate that, as

Will Roper observed, innovators on the civilian and military side found each other. Whether the innovator was Lt Col Leghorn making common cause with civilian scientists to push the Air

Force toward strategic reconnaissance, or Secretary of the Navy John Lehman and General P.X.

Kelley drawing similar conclusions from tragedy, or USD A&T John Deutch and VCJCS

Admiral David Jeremiah in one generation and Deputy Secretary of Defense Bob Work and

VCJCS Admiral Sandy Winnefeld in another pushing the DoD to embrace UAV technology, civil-military collaboration was a persistent pattern in the case. What was important was that, until the Third Offset took hold, the civilians volunteered or were recruited to defend one military preference or another. It was not until the second decade of the twenty-first century that civilians chose to counter military preferences with their own.

Some coalition building was more organic than intentional. The DoD has a practice— strengthened by joint assignment requirements for officer promotion mandated by the

Goldwater-Nichols Act—of cross-pollinating management structures with military and civilian officials, as well as funding civilian functions with service budgets. The JPO and DARO offices were cases in point. JPO’s executive agent was the Navy, even though it was an OSD office with

203

OSD budget authority. DARO had a similar structure—it did not have a service executive agent, but it was run by an Air Force two-star general. At the DoD level, institutions are more-civilian and more-military; some are in the operational chain of command and some are not; but the overlap between civilian and military functions, even in OSD, is significant enough to eliminate the possibility that there are purely civilian or military preferences within the Department.

This suggests that on average, only civilians outside DoD have some hope of having preferences that are truly independent from the military. Even then, civilian scientists examined here relied on collaborative relationships with military institutions in order to perform their analysis and gain credibility with the very military institutions they were advising. From the

BEACON HILL and TEOTA studies to Larry Lynn’s summer study on reconnaissance and Jim

Cartwright’s Deep Attack Weapons Mix study, senior decision-makers often came to their preferences through mixed civil-military analysis. Both the BEACON HILL and TEOTA studies explicitly referred to military advisors as their main sources. That scientists outside government as well as acquisition officials and even the Secretary of Defense acknowledged the influence of military preferences on their own preferences revealed the extent to which military preferences permeated civilian preferences for the uses of UAVs.

But these are all domestic factors. What about international conditions? Changes in the international security environment obviously mattered, but not necessarily in uniform ways, suggesting that different combinations of domestic and international factors produced different dynamics between military and civilian preferences for the uses of UAVs. Civilians showed far more capacity to exercise independence from military preferences about UAVs under peacetime development, but civilians still drew their preferences for UAV uses from military preferences throughout the Cold War and into the post-Cold War decade—civilians never seriously

204

challenged the idea of limiting UAV uses to reconnaissance missions until at least the mid-

2000s. In fact, the first Gulf War may provide more evidence of civilian deference to warfighters, in that those military personnel who deployed for Desert Storm and Desert Shield hardly noticed the existence of UAVs, let alone their contributions to major combat. The combat veterans of the 1990s emphasized airpower dominated by manned strike platforms, not UAVs.

However, the relative peacetime of the post-Iraq drawdown years formed the context for the greatest degree of civilian preference independence from the military for UAVs in sixty years. It is unclear why this was so, but the answer likely has to do with the shift away from current wars to the consensus about the likely future war with China. As the Pentagon’s civilian

Director of Research and Engineering Steve Welby explained to a reporter, “this focus on speed of innovation” associated with the Third Offset was driven by “an external reality” where “other countries now have significant resources supplied to military modernization.”505 Writing in the

Washington Post in 2018, now-former Deputy Secretary of Defense Robert Work asserted that the U.S. had long ignored adversary modernization in favor of near-term warfighting goals, lamenting, “we have robbed the Peter of advanced capabilities to pay the Paul of day-to-day force employment.”506 In other words, civilian managers had adopted warfighter’s preferences at the expense of their own.

What is intriguing is that civilians and the military were both deriving their preferences from external, international structural factors and threats, but along completely different timelines. That the small cadre of civilians fixated on competition with China finally broke out

505 Quoted in Aaron Mehta, “Interview: Stephen Welby, Assistant US Secretary of Defense for Research and Engineering,” Defense News, April 4, 2016, https://www.defensenews.com/industry/techwatch/2016/04/04/interview-stephen-welby-assistant-us-secretary-of- defense-for-research-and-engineering/. 506 Robert O. Work and Elbridge Colby, “The Pentagon Must Modernize Before It’s Too Late,” Washington Post, September 17, 2018, https://www.washingtonpost.com/opinions/the-pentagon-must-modernize-before-its-too- late/2018/09/17/8a9ab1e8-b1fe-11e8-aed9-001309990777_story.html. 205

and shifted civilian preferences in a direction relatively free of the influence of military preferences appears to have been a function of civilians’ longer-term focus than much of the military, and the reduced political power of short-term-thinking warfighters. As Welby’s and

Work’s statements also suggest, the shift in long-term thinking also meant that international factors different from current operations and threats were motivating vanguard civilian thinking and giving an existing preference (using UAVs as combat vehicles) a new context in which to spread to more civilians and to become a priority.

Over the long course of the case of UAV capabilities, then, civilians were always susceptible to military preferences, but the conditions of wartime coupled with strong, consolidated military preferences magnified that susceptibility. In peacetime, civilians’ choices about the allocation of their attention appear to have given them much greater control over whether they adopted military preferences as their own—although it did not guarantee that civilians could transfer their preferences over to critical masses of or pivotally powerful military officers. And in wartime, civilian and military preferences often merged, a combination of strong domestic political incentives and another set of calculated choices that siding with one military faction would allow them to achieve other, non-military, fiscal and political goals.

Testing the Analysis: Alternative Explanations and Counterfactuals

It would be easy to assume that technology determined the plausible uses of UAVs and therefore was responsible for the range of possible preferences for their use. But the microprocessor and GPS were not required to make UAVs operationally useful—and in fact, lagged behind UAVs’ major operational proof-of-concept by at least 20 years. Nor was the technology to improve UAV performance on some inevitably plodding timeline. Instead, as

206

Harvey Sapolsky has shown elsewhere, service parochialism, national politics and defense strategies drove technological investment rather than vice versa.507 Indecision at the service level and, later on in the development phase, Congressional interventions to suppress costs generated on-again off-again investment cycles that doomed programs into cost overruns and schedule delays. This pattern of half-hearted investment was both a consequence and a cause of skepticism among military and civilian groups alike. It is thus true that enabling technologies made great leaps in the 1980s and 1990s, but it was also true that the UAV technology overall was not given sufficient support by the U.S. until the 1990s, concurrent with the accelerant stage of the case.

A large part of the delay was in fact that the services believed they already had technology that performed the kinds of missions they envisioned for UAVs: manned platforms.

That manned capabilities were more appealing to service leaders is not surprising from the perspective of the scholarly literature on organizational culture. For one, manned aircraft already worked, and the notoriously conservative services are reluctant to abandon something that works in favor of an experimental capability.508 Moreover, many sub-service cultures placed manned aircraft capabilities at the core of their identities and existence and manned aircraft supported concepts of operation and doctrine upon which other branches and services depended. Long- established career constituencies in the services and private constituencies in industry together sustained constituencies on the Hill.509 Technologically and organizationally, UAVs needed too much investment and change. As an institutional matter, the services had no interest in

507 Harvey M. Sapolsky, The Polaris System Development: Bureaucratic and Programmatic Success in Government (Cambridge, MA: Harvard University Press, 1972). 508 Edward L. Katzenbach, Jr., “The Horse Cavalry in the Twentieth Century,” in The Use of Force: International Politics and Foreign Policy, ed. Robert J. Art and Kenneth Waltz (Boston: Little Brown and Company, 1971). 509 Gordon Adams, The Iron Triangle: The Politics of Defense Contracting (New York, NY: Council on Economic Priorities, 1981). 207

overcoming such hurdles. And they therefore had an interest in convincing civilians to share their preference to avoid UAVs.

It would be equally easy to assume that UAVs were dedicated to unconventional warfare operations because they were more naturally suited to them. But again, it was instead the case that military institutions preferred to use them that way. The U.S. actually used UAVs for SEAD

—a precursor to using them for wider conventional strike operations—almost 20 years before the

Israelis, during the war in Vietnam. A UAV system was again helpful in directing fires during the Gulf War, and there was no logical reason for operators on the ground not to come to the conclusion that UAVs could act as conventional strike aircraft, the same idea that the Air Force came to a decade later in the Balkans. Yet the services continued to shunt UAV capabilities into reconnaissance missions after both wars ended. The services simply did not want to see UAVs used for non-reconnaissance purposes if they were used at all. Until 2012, civilians, as a bloc, followed suit.

A counterfactual also tests the validity of the explanation that information asymmetries led civilians to adopt military preferences. First, suppose that the attacks of 9/11 had never happened. It is likely that the Bush administration would have continued to push for transformation as a means to secure military primacy—especially given that Rumsfeld established his Office of Force Transformation several weeks after 9/11.510 It is equally likely that service resistance to the major changes to force structure demanded by transformation would have also persisted. What would not necessarily have happened would be the major operations in

Afghanistan and Iraq, and the demand for UAVs for counterterrorism operations would also likely not have materialized. This would remove the warfighter advocacy for ISR- and irregular

510 Dombrowski and Gholz, Buying Military Transformation, 72. 208

warfare-centric UAV uses. Indeed, the dynamics of the Third Offset, where civilians were able to develop preferences independently from the military but struggled to impose policies guided by those preferences on the military services, would likely have developed sooner. The question then is whether it would be information symmetry or the structure of the international system that better explained the source of civilian preferences for UAV uses and their independence from the military. The answer is likely that both would play a role, but informational symmetry, or even asymmetry in civilians’ favor, would be necessary for civilians to develop independent preferences. Without their own sources of information and expertise, and their own sustained attention to the uses of UAVs, civilians would not likely determine those uses given the international environment. Put another way, information and expertise are necessary to do the interpretive work of understanding what the environment means in term of the capabilities needed for national defense and how they should be used to achieve strategic ends. When civilians could perform such interpretation on their own, they developed their own preferences.

What of the cultural-political explanation for civilian preferences? Culture dictates attention, directs information searches, and shapes how civilians privilege information. However, it is the differential in information that determines the final preference adopted. We know this is true because relative civilian and military information matter in both the counterfactual and the actual case—without 9/11, civilians continue to develop sources of information independent from the military and therefore have independent preferences, but with 9/11 informational asymmetries lead to civilian deference to military preferences.

209

Conclusion

Information asymmetries caused civilian preferences to reflect military preferences for sixty years, through fluctuations in the DoD budgets, changes in strategy and adversaries, technological innovations, presidential administrations and other domestic political upheavals, and a variety of international crises and military operations. After fifty years of military preferences being the core of civilian preferences, the two overlapped entirely after the start of operations in Afghanistan and Iraq. During the development phase, although military preferences always shaped civilian preferences, civilians did vary in which military preference they adopted based on their other institutional preferences. For decades, civilians generally adopted the majority of the military’s preference to eschew UAVs. Then, in the 1980s, a vanguard of civilians began to shift toward embracing UAVs, although only for the reconnaissance missions for which they had always been imagined. It was not until the second decade of the twenty-first century when senior civilians managed to develop preferences for UAV uses that were independent from military preferences.

The drivers of this seemingly simple story were anything but. The array of civilian and military institutions and actors holding salient preferences shifted over time, as different services took very different approaches to UAV development—the Air Force focusing on strategic reconnaissance, the Army on RSTA/ISR, and the Navy avoiding UAV capabilities as much as possible. Decades of failed service efforts to sustainably field a UAV system drove the Congress to attempt a variety of mechanisms and incentives to get a joint system through the acquisition process. In doing so, Congress acted on its interest in reducing resources devoted to UAVs but deferred the strategic justification for the capability to the services. It took civilian acquisition officials in DoD transferring key elements of the acquisition process to deployed forces to gain

210

leverage over the services—but also to lose leverage to those deployed forces themselves.

Meanwhile civilians in Congress and those in OSD focused on technological innovation were not interested in alternatives to military preferences for the uses of UAVs. They simply wanted the capability to be adopted by the services. Consistent with the literature on military innovation, they needed military partners to overcome service inertia.

In the end, the civilian advocates of the first offset strategy, the RMA, and transformation all eventually acquiesced in accepting service and later warfighter preferences for UAV purposes, the latter because wartime changes in the distribution of information between civilians and the military triggered civilian deference. Only the Third Offset leaders broke away from military preferences because they relied on information that was more independent of the military. They also followed the pattern of greater civilian preference independence from the military during (relative) peacetime.

The UAV case reveals that civilians struggled to avoid basing their preferences for the uses of UAVs on military preferences, especially the more they wanted UAV capabilities to be adopted by the services or when they faced the domestic politics and international risk inherent to military operations and war. Perversely, the more civilians wanted the military to use UAVs, the more constrained by military preferences for the uses of UAVs they became.

211

CHAPTER 4

CYBER CAPABILITIES: 1984-2019

This shadow case study of cyber capabilities from 1984 to 2019 demonstrates that the more civilians pay attention, build expertise, and have access to information about cyber issues, the more civilians form their own preferences independent from the military. At the start of the case, civilians were far better informed about cyber issues than military actors and institutions and publicly defined the purpose of military cyber capabilities as primarily defensive.511 By

2019, a significant sub-group of the military, in the form of Cyber Command, was expanding the military’s role in national cybersecurity and pushing civilians to publicly endorse offensive uses of cyber capabilities, working to change not only civilian preferences but cyber strategy itself.

Civilian preferences regarding cyber have shifted to overlap more with military preferences for three likely reasons. First, the nature of cyber threats has itself evolved to suggest that changes in the U.S. approach were warranted. But second, military institutions that focus on cyber have interpreted these changes through a military cultural lens for civilian legislators and policymakers. Third, the civil-military balance of information resources is shifting away from civilians and toward the military. As a result, Cyber Command and other military groups may have been able to increase their influence over civilian perceptions of military cyber capabilities from primarily defensive in nature toward an increasing emphasis on offensive activities.

511 In conducting the research for this case, numerous practitioners told me that offensive uses of cyber capabilities have, until very recently, been classified and separated from defensive uses in organizational and information-access terms. Here, I trace the change in civilians’ publicly stated preferences about cyber capabilities, which evolved from prioritizing defensive uses to embracing offensive uses as well. I judge that this new willingness to endorse offensive uses publicly is a meaningful change in preferences, in that it signals civilians’ increasing comfort with regularly using cyber capabilities in offensive ways, as opposed to reserving offensive uses of cyber tools for rare exigencies. 212

This is a shadow case. I therefore conducted it narrowly to test the validity of my argument. Moreover, because cyber events are still unfolding, the conclusions drawn regarding emerging cyber capabilities are more tentative than those for the SOF and UAV cases. Existing data may not be fully representative of reality due to the high level of classification on U.S. cyber activities. Nevertheless, the empirical evidence I present here is consistent with the patterns revealed in the prior two case studies, and what is available publicly still tells a story relevant to civil-military relations and civilian preference formation.

Why Cyber? Definitions and Case Justification

The definition of cyber capabilities is subject to ongoing debate, and what is today called

“cyber” has existed under different names over time. As computer networking expanded and evolved in the 1980s and 1990s, and the Department of Defense began to explore the role played by computers and communications networks in military activities, concepts such as information warfare and network centric warfare took root. Information warfare comprised not just what we would today think of as cyber war, but manipulation of the range of media that carries news, ideas, and communications.512 Network centric or just “netcentric” warfare described mostly how the U.S. military would use communication technologies to increase the speed of decision- making and targeting precision.513

512 Catherine A. Theohary, Information Warfare: Issues for Congress, CRS Report No. R45142 (Washington, DC: Congressional Research Service, March 5, 2018), https://fas.org/sgp/crs/natsec/R45142.pdf. For an exploration of information warfare concentrated in the cyber domain, see Roger C. Molander, Andrew S. Riddile, and Peter A. Wilson, Strategic Information Warfare: A New Face of War (Arlington, VA: RAND Corporation, 1996), https://www.rand.org/pubs/monograph_reports/MR661.html. 513 Clay Wilson, Network Centric Operations: Background and Oversight Issues for Congress, CRS Report No. RL32411 (Washington, DC: Congressional Research Service, March 15, 2007), https://fas.org/sgp/crs/natsec/RL32411.pdf. 213

Cyberspace and cyber warfare reflect elements of both information warfare and netcentric operations but have become distinct from them. Cyberspace is now conceived as its own domain or arena of activity.514 Cyber warfare takes place in and through cyberspace, affecting computer processing as well as the physical world. Cyber capabilities refer to the exploitation of software and hardware vulnerabilities, be they mistakes in computer code or exposure to interception during communications, and to the defense of such vulnerabilities.515

Cyber capabilities can be used as tools for intelligence collection, sabotage, or outright attack on computer networks and physical destruction of computers or computer-dependent machines.516

To date, practitioners and analysts have come to agree on two important ideas: that the line between defensive and offensive cyber capabilities is not only ill-defined but unavoidably dynamic, and that the nature of nation-state competition within cyberspace favors offensive action.517 But categorizing actual cyber actions into these two ideal-types is not always straightforward, and a range of “active defense” and espionage uses of cyber tools are neither offensive in a destructive or harmful sense nor do they leave adversary systems undisturbed in the classic sense of defense.518 This suggests that the “logic” of cyber capabilities and who

514 I explore below the ways in which the U.S. military shaped national thinking about cyberspace as a domain of warfare. 515 Gregory Konti and David Raymond, On Cyber: Towards an Operational Art for Cyber Conflict (Middletown, DE: Kopidion Press, 2017). 516 I do not include autonomous systems or artificial intelligence in my definition of cyber because they imply capabilities beyond cyberspace often used in the physical world. U.S. government cyber strategies also generally do not cover AI or autonomous/robotic systems, which are treated separately. 517 Lucas Kello, The Virtual Weapon and International Order (New Haven, CT: Yale University Press, 2017). 518 Defensive capabilities and activities, often conflated with the notion of cybersecurity, protect the integrity of data and user privacy and counter attempts to interrupt communications, steal data, or cause physical destruction. Offensive capabilities and activities involve uses of cyberspace or cyber tools to cause destruction of both virtual and real-world targets. Those who study the military applications of cyber capabilities often compare them to capabilities and activities in the physical world. For example, one group of authors explains that “hunting within one’s own network for malware” is “analogous to launching intercept aircraft in response to an intruding enemy aircraft.” C. Robert Kehler, Herbert Lin, and Michael Sulmeyer, “Rules of Engagement for Cyberspace Operations,” in Bytes Bombs and Spies: The Strategic Dimensions of Offensive Cyber Operations, ed. Herbert Lin and Amy Zergart (Washington, DC: Brookings Institution Press, 2018), 293-294. 214

controls the definitions of those capabilities and activities is an important part of civilian and military preference formation, perhaps even more so than for the other two case studies in this volume.

The military shares responsibilities for government activity in and defense of cyberspace with several civilian agencies of government, most prominently the Department of Homeland

Security. Where the military fits into national cyber strategies, therefore, has been central to debates regarding the uses of cyber tools. This case study will focus on civilian and military preferences for the military uses of cyber capabilities, including which capabilities are or should become military in nature.

The emergence of cyber capabilities offers a test of the hypothesis unique among the cases because it begins with information asymmetries in civilians’ favor. Moreover, because the military became more informed over time, by some measures surpassing the average civilians’ knowledge regarding cyber issues, the cyber case has the most dramatic measure of change in the independent variable and therefore poses an excellent test for covariance with civilian preferences.

Cyber Capabilities 1984-2018: Key Events, Actors, Preferences

Case Organization and Key Events

For this case I have borrowed from Jason Healey’s periodization of U.S. cyber capabilities, but have reduced it from three to just two periods with dates adjusted accordingly: realization (1984-2005) and militarization (2006-2018).519 The case begins well after the

519 Healey’s original periodization includes three periods: realization, takeoff, and militarization. For the purposes of a parsimonious shadow case, I found that the realization and takeoff periods were not sufficiently different from a preferences perspective. Jason Healey, ed., A Fierce Domain: Conflict in Cyberspace, 1986 to 2012 (Washington, DC: Atlantic Council, 2013). 215

development of modern computers, and 15 years after the invention of the proto-Internet, but at the start of the realization that computers had major military utility even as they generated major military vulnerabilities. The periods are punctuated by what Healey refers to as “wake-up calls,” or “shocks” that attracted leader’s attention and generated new and broader understandings of the risks inherent to national reliance on cyberspace.520

Prior to the beginning of the case in 1984, it was Department of Defense civilians who were deeply (if not broadly across DoD’s various bureaucracies) involved in what we would today call cyber issues. The Advanced Research Projects Administration contracted for development of the precursor to the modern internet, the so-called ARPANET.521 The civilian in charge of ARPA’s computer communications efforts, J.C.R. Licklider, was especially influential in early DoD experiments with computers.522 DoD’s sponsorship of such efforts was prompted by military-related interest in command and control issues, but the efforts themselves were products more of the academic computer science community (including MIT, the University of

California at Berkeley and at Santa Barbara, the Stanford Research Institute, and the University of Utah) than the military.523 In 1970 the Defense Science Board directed, and the RAND

520 Contextually, cyber issues fall within several other major national debates: about civil liberties and government overreach; the relationship between the public and private sectors, to include government regulation; the speed of government acquisitions; broader decision-making processes and whose voices are included; the idea of “net neutrality”; and so on. Not all of these can be covered in a shadow case study, but they will be discussed briefly when necessary for understanding the political and security sources of preferences. 521 Barry M. Leiner et al., “The Past and Future History of the Internet,” Communications of the ACM 40, no. 2 (February 1997): 102-108, https://groups.csail.mit.edu/ana/Publications/PubPDFs/The%20past%20and%20future%20history%20of%20the%20 internet.pdf; ARPA (today DARPA, the D standing for Defense) itself was not and is not attached to the military services. Instead, its model is to draw staff from the academic and “entrepreneurial” research and development communities in the private sector, making DARPA a largely civilian organization. See: Congressional Research Service, Defense Advanced Research Projects Agency: Overview and Issues for Congress, CRS Report No. R45088 (Washington, DC: Congressional Research Service, July 24, 2018), https://fas.org/sgp/crs/natsec/R45088.pdf. 522 Glenn Fowler, “Joseph C.R. Licklider Dies at 75; Foresaw New Uses for Computers,” New York Times, July 3, 1990, B6, https://www.nytimes.com/1990/07/03/obituaries/joseph-cr-licklider-dies-at-75-foresaw-new-uses-for- computers.html. 523 Ronda Hauben and Michael Hauben, Netizens: On the History and Impact of Usenet and the Internet (London: Wiley, 1997). 216

Corporation published, the first report on computer network security, which was classified and therefore not widely disseminated.524 At the helm of crafting both that report and an earlier exploration of computer security was a man named Willis Ware, an engineer and one of the earliest computer scientists, who would go on to be a long-time expert consultant for the government.525 ARPANET development continued through the 1970s; by 1983 a separate network for the military was carved off from the original ARPA system.526

Meanwhile, by the late 1970s advances in intelligence collection on other communications technologies (called signals intelligence or SIGINT) had inspired a man named

Bill Perry to develop the concept of “counter-command-and-control warfare” or “counter-C2 warfare.” The idea was that interrupting adversaries’ ability to transmit information, including orders, between commanders and forces in the field would undermine an enemy’s military effectiveness. Bill Perry was President Carter’s Assistant Secretary of Defense for Research and

Development and had run a company in the private sector involved in computing.527 Perry thus brought expertise in computer networking with him into his later government roles.

These developments set the stage for the modern emergence of cyber warfare and military uses of cyber capabilities. The first phase of the case study, realization, was marked by a dramatically expanded use of computers and the internet in the U.S. and around the world as well as several significant hacks of government computer and communications systems. The period includes the first combat use of computer-based networks during the first Gulf War and is punctuated by two White House national security decision directives (NSDDs) regarding

524 Willis W. Ware, ed., Security Controls for Computer Systems: Report of Defense Science Board Task Force on Computer Security (Santa Monica, CA: RAND Corporation, February 1970/ reissued October 1979), https://www.rand.org/pubs/reports/R609-1/index2.html; and Healey, A Fierce Domain, 28. 525 Fred Kaplan, Dark Territory: The Secret History of Cyber War (New York, NY: Simon & Schuster, 2016). 526 Healey, A Fierce Domain, 28. 527 Kaplan, Dark Territory, 15. 217

computer network vulnerabilities, the first in 1984 and the second in 1998. 1998 also marked the establishment of the first military joint task force focused on computer network activities. By the end of the 1990s, DoD had used cyber tools offensively (as opposed to simply relying on computing for U.S. communication) during major operations for the first time in Kosovo. The period also covers the production of the first national cyber strategy in 2003.

During realization American elites grappling with the implications of the pervasive reliance on cyberspace for communications and market activities and the increasing number of actors and adversaries contesting U.S. use of and security in cyberspace, including private hacking groups and criminal activity. In the year 2000, 50% of American adults were using the internet; by 2006, 71% of American adults were online.528 The emergence of new actors and new activities in cyberspace prompted the creation of new military and civilian organizations, doctrine, and policy. In particular, analysts began to appreciate the degree to which cyberspace relied on private companies for its existence and maintenance.

The militarization stage covers the rapid build-up in cyber expertise, capabilities, and headquarters and force structure in the armed services, as well as the military’s transition from accepting civilian-directed defensive uses for cyber capabilities to greater emphasis on their own preferred offensive uses. It begins with the 2006 National Military Strategy declaring cyberspace a military domain, followed by the establishment of Cyber Command in 2009, the first publicly documented offensive use of cyber capabilities against Iranian nuclear facilities in 2010, and the first publicly acknowledged use of offensive cyber operations against the Islamic State in 2016.

The militarization period includes the rise of Russian and Chinese cyber operations against the

528 “Internet/Broadband Fact Sheet,” Pew Research Center, June 12, 2019, https://www.pewinternet.org/fact- sheet/internet-broadband/. Underlying these numbers are wide but steadily closing gaps in usage based on education and income levels, suggesting relationships between cyber usage and wealth and economic activity in the U.S. 218

U.S. and its partners, and Iranian and North Korean entry into cyber competition as well.

Meanwhile, the U.S. became ever-more dependent on cyberspace for financial transactions, data storage, and communications. By 2018, 89% of American adults were using the internet.529

Actors and Preferences

Much as in the other two cases, the key civilian and military actors broke into sub- groups. In general, both civilians and the military had two key sub-groups of actors: one for whom cyber expertise was in their explicit job portfolio, and a second, much larger, class for whom cyber information was not only irrelevant on a daily basis but often alien. While there is always a distance between specialists and non-specialists, the technical and often classified nature of cyber issues meant that the basic level of knowledge and awareness usually present among national security practitioners on other matters did not exist in the cyber case.

The two major preferences for the uses of military cyber capabilities traced here are for defensive and offensive purposes. For most of the case, civilians focused largely on the vulnerabilities created by government and private sector reliance on computer networking, and articulated preferences to develop and focus cyber capabilities to be defensive in nature. Within the past two years, civilians in the Executive branch appear to be shifting their public preference toward offensive uses of cyber capabilities as well.

The sub-groups of the military services and joint commands evolved from thinking about cyber capabilities as communications tools to considering them tactical offensive weapons in support of real-world military operations. By the end of the case, Cyber Command preferred to use cyber capabilities offensively within cyberspace itself.

529 Ibid. 219

The White House formed the earliest civilian preferences on computer networking- related capabilities during the case study period, with President Reagan’s National Security

Decision Directive 145, “National Policy on Telecommunications and Automated Information

Systems Security.” This work was, in turn, based on the knowledge gathered by civilians in the

Department of Defense, including those at the Advanced Research Projects Agency who had invented the internet and Assistant Secretary of Defense for Command, Control,

Communications and Intelligence or ASD(C3I). For the most part, then, civilians paying enough attention to cyber issues to form consolidated preferences were at high levels of government in the Executive Branch. In contrast, the first military cyber unit, the 609th Information Warfare

Squadron in the Air Force, did not stand up until 1995, and a joint military organization did not exist until 1998, fourteen years after the Reagan Administration issued national security guidance at the tactical level. The unified combatant command for military-related cyber operations, Cyber Command, was not established until 2009, and was not a working bureaucracy until 2010. The contrast in information was thus complemented by the contrast in levels of institutional power. The sub-groups in the military focused on cyber issues remained at the tactical level for two decades, whereas the civilians who focused on it were at the strategic level in the White House and the Office of the Secretary of Defense (OSD).

However, once Cyber Command was established and the commander became the simultaneous director of the National Security Agency (NSA), the intelligence organization traditionally in charge of code-breaking and housing much of the country’s cyber-related human capital, the consolidation and elevation of the military’s institutional role in developing and deploying cyber capabilities gave it increasing control over cyber information.530 At the same

530 It is worth noting that there is some ambiguity over the appropriate classification of the NSA—is it a military organization or a civilian one? It is organized under the Defense Department and reports to the secretary of defense 220

time, government civilian informational assets remained relatively static, and may even have been reduced to some degree because classification limited access to so few people overall and because military organizations had grown in number and size.531 In 2018, the White House eliminated the position of cybersecurity coordinator at the National Security Council.532 Relative to the military, civilians began to lose control over sources of information and there are now signs that civilian preferences have begun to align more closely with the preferences of Cyber

Command.

The Case

Realization: 1984-2005

At the beginning of the realization period, civilians preferred to use tools related to computer networks (what evolved into cyberspace) for communications and defensive purposes.

and the director of national intelligence rather than a military service, although the Goldwater-Nichols legislation established it as “a combat support agency.” Its workforce consists of a mix of civilians and military personnel, but its director has always been an active duty senior military officer. Now that the NSA and Cyber Command are co- located and its director a dual-hatted military officer, the agency’s civilian or military identity is even harder to distinguish. Below, I categorize Cyber Command and the commander/director of NSA as military and treat the other components of NSA as civilian, given its personnel’s distinction from the armed services. In some ways, a detailed and subtle examination and classification of the NSA is a microcosm for the entire study, one of the purposes of which is to explore how distinct civilian and military identities really are. “National Security Agency/Central Security Service Transition 2001,” in The National Security Agency Declassified, ed. Jeffrey Richelson (Washington, DC: George Washington University National Security Archives, December 2000), https://nsarchive2.gwu.edu//NSAEBB/NSAEBB24/nsa25.pdf; The NSA keeps a list of all of its directors, with their military ranks, on its website: https://www.nsa.gov/about/leadership/former-directors/. 531 Although the civilian private sector by this point was robust and sprawled far beyond military capabilities, government relations with the computer technology sector were more strained than they had been earlier in history. For a discussion of the acrimony between Silicon Valley and DC at the policy level, see Amy Zegart and Kevin Childs, “The Divide Between Silicon Valley and Washington is a National Security Threat,” The Atlantic, December 13, 2018, https://www.theatlantic.com/ideas/archive/2018/12/growing-gulf-between-silicon-valley-and- washington/577963/. At the same time, retaining personnel in government cyber jobs given the competition with the private sector is a challenge for civilian agencies. A Senate report found in 2019 that the rate of turnover in chief information officers at DHS, the State Department, and other civilian agencies was very high and resulting in grave management challenges. See United States Senate Committee on Homeland Security and Governmental Affairs Permanent Subcommittee on Investigations, “Federal Cybersecurity: America’s Data at Risk,” June 25, 2019, https://www.hsgac.senate.gov/imo/media/doc/2019-06-25%20PSI%20Staff%20Report%20- %20Federal%20Cybersecurity%20Updated.pdf. 532 Nicole Perlroth and David E. Sanger, “White House Eliminates Cybersecurity Coordinator Role,” New York Times, May 15, 2018, B2, https://www.nytimes.com/2018/05/15/technology/white-house-cybersecurity.html. 221

By the end of the period, the civilian preference to focus on cybersecurity and secure communications, and to limit DoD’s and the military’s role in cyberspace accordingly, had not changed. Civilians maintained their preference for defense despite a series of DoD exercises and real-world hacks on DoD networks. In fact, such events made more civilians conclude that cyber defenses should be a priority. But the same series of wake-up calls generated different preferences among the military services. Military groups generally (although not universally) began to conclude that computer networks would need to be incorporated into traditional warfighting—meaning the military saw the importance of cyber as tools that could support or even wage war but as support to conventional operations on land, sea, and in the air. That civilian preferences did not co-vary with military preferences suggests that civilians did not depend on the military or were not constrained by military actors and institutions to form their preferences for the uses of cyberspace and cyber capabilities.

Details

On a seemingly random day in 1984, President Reagan asked the Chairman of the Joint

Chiefs of Staff about the security of military computing networks. But it was not random.

Reagan had just seen the movie War Games, in which a teenager hacks into the computers controlling America’s nuclear weapons. General John William Vessey, the Chairman of the Joint

Chiefs of Staff, knew little about computing but swiftly learned from the ASD(C3I) Donald

Latham that the War Games scenario was not all that far-fetched.533 That the President turned to

Vessey for the answer, but Vessey had to tap the civilian expertise in the department, is a

533 Kaplan, Dark Territory, 19. 222

microcosm of the realization period: a civil-military partnership where the military was distinctly less informed than civilians.

Latham, whose job description included coordination with and OSD oversight of the

NSA, was assigned to be a lead drafter for President Reagan’s new National Security Decision

Directive-145, “National Policy on Telecommunications and Automated Information Systems

Security.”534 That document, issued in the fall of 1984, put the NSA in charge of communications security for the U.S. government and was remarkably prescient about the emergence of what today we call cyberspace, and the vulnerabilities inherent to it. It is worth quoting at length:

As new technologies have been applied, traditional distinctions between telecommunications and automated information systems have begun to disappear. Although this trend promises greatly improved efficiency and effectiveness, it also poses significant security challenges. Telecommunications and automated information processing systems are highly susceptible to interception, unauthorized electronic access, and related forms of technical exploitation, as well as other dimensions of the hostile intelligence threat. The technology to exploit these electronic systems is widespread and is used extensively by foreign nations and can be employed, as well, by terrorist groups and criminal elements. Government systems as well as those which process the private or proprietary information of US persons and businesses can become targets for foreign exploitation.535

The directive identified the need for defensive measures against foreign exploitation and couched emerging telecommunications technology in counterintelligence and defensive terms. It also noted the democratization of networked communications and the role played by private actors both as spoilers (as criminals and violent non-state actors) and as partners (in the private sector).

Because the focus was on securing government networks and certain critical private ones, there is no mention of using computer network-based tools for military purposes beyond

534 The White House, National Security Decision Directive Number 145, National Policy on Telecommunications and Automated Information Systems Security (NSDD-145), September 17, 1984. 535 Ibid., 1. 223

securing internal DoD communications. Moreover, the directive assigned information security responsibilities to multiple civilian government agencies, and in fact set up a multi-agency

National Telecommunications and Information Systems Security Committee to coordinate among the many government stakeholders.536 Such arrangements both signaled and codified the diverse participation of civilian institutions and experts in early cyber policy matters. It also demonstrated that senior policymakers—despite the military setting of Wargames and Reagan initially turning to his JCS chairman—did not think of these proto-cyber matters as having military implications.

Nevertheless, the Department of Defense was in a strong bureaucratic position over nascent cyber policy as a result of NSDD-145. The ASD(C3I) surely played a role in where future responsibilities would fall (stakeholders listed in the document also included all four military services and the Chairman of the Joint Chiefs of Staff, and the Secretary of Defense was made the “executive agent” for telecommunications security).537 DoD was not the lead player as a result of NSDD-145, but it was an important one, and it gained an official foothold in national cyber policy.

Still, in 1984 more of the military services’ influence over cyber matters was on paper than in action. To some extent, the military was still in the same general boat as the rest of the government, newly and increasingly reliant on computer networks for communications, but seeing computers mostly as items of logistical value. It would be years before the services

536 This included the Departments of Commerce, Transportation, and Energy, as well as the FBI. 537 Although, as referred to in directive itself, this status for the secretary of defense was built on an earlier executive order establishing DoD as the executive agent for communications security. NSDD-145, 6. 224

established anything close to the kind of expertise and organizational seniority that the

ASD(C3I) team had.538

But real-world experiences were about to demonstrate to the U.S. military in particular that computer technologies had introduced new types of targets into modern conflict. The first attack targeting DoD systems and the defense industry was called “Cuckoo’s Egg,” and involved

German hackers looking for information on the Strategic Defense Initiative, a missile defense system.539 Then in 1988 a far more damaging intrusion called “the Morris Worm” disabled some ten percent of the internet. Created by a graduate student in computer science at Cornell

University, diagnosed and dissected by a professor at Purdue University and more graduate students at the University of California at Berkeley, the U.S. government was mostly a helpless bystander.540 These two events demonstrated to DoD that hostile actors could invade American networks undetected to collect information and possibly induce chaos in communications.

Enemies could get behind friendly lines.

Military communities set about processing the implications of these attacks as well as the

U.S.’ own networking capabilities through familiar, and seemingly-related, ideas of warfare.

Concepts around information warfare made up a major part of the Revolution in Military Affairs, an approach reinforced by the U.S. success in the 1991 Gulf War. That war was the first time the

538 The ASD(3CI) was first established in 1970 as the assistant secretary of defense for Telecommunications, and later became ASD(3CI) in 1977. Joshua Boehm with Craig Baker, Stanley Chan, and Mel Sakazaki, A History of United States National Security Space Management and Organization (Washington, DC: Commission to Assess United States National Security Space Management and Organization, January 2001). 539 Healey, A Fierce Domain, 29. The name Cuckoo’s Egg was based on a novel about Soviet computer spying. The hack was discovered by the systems administrator at Lawrence Berkeley National Laboratory, Clifford Stoll, who later chronicled his detective work to track down and identify the hacker. See Clifford Stoll, “Stalking the Wily Hacker,” Communications of the ACM 31, no. 5 (May 1988), https://dl.acm.org/doi/10.1145/42411.42412. 540 Timothy B. Lee, “How a Grad Student Trying to Build the First Botnet Brought the Internet to Its Knees,” Washington Post, November 1, 2013, https://www.washingtonpost.com/news/the-switch/wp/2013/11/01/how-a- grad-student-trying-to-build-the-first-botnet-brought-the-internet-to-its-knees/; At the same time, a flurry of legislation on the Hill raced to help the law keep up with computing. Many of these laws focused on civil liberties and privacy issues and included the Computer Fraud and Abuse Act of 1986, the Electronic Communications Privacy Act of 1986, and the Computer Security Act of 1987. Healey, A Fierce Domain, 32. 225

U.S. military both relied on rapid computer networked communications and executed Perry’s ideas about counter-C2 warfare. Teaming up, the Joint Staff’s intelligence directorate and the

NSA went about exploiting and sabotaging Saddam Hussein’s command and control network.541

The information components of Operation Desert Storm generated decisive outcomes in terms of combat and intelligence collection.542

The man who ran the Gulf War Joint Intelligence Center, Rear Admiral Mike McConnell, went on to become the Director of the NSA.543 In direct response to his experiences in Iraq,

McConnell created a new role at NSA: Director of Information Warfare. He asked one of his

Gulf War deputies, a Naval Intelligence officer, to fill the job.544 McConnell represented the

Navy’s re-capture of NSA, in fact. The job had previously, and pivotally, belonged to Admiral

Bobby Ray Inman, who had conducted various bureaucratic maneuvers to ensure that NSA and not the CIA (for which Inman became Deputy Director) got the job of intelligence collection related to computers.545 Now the Naval Intelligence community was again reigning over the national signals intelligence collector and laying the groundwork for what would become cyber capabilities.

Meanwhile at the Pentagon, OSD and the Joint Staff began a dialogue over information warfare that would go on for most of the 1990s. In general, there was greater specificity in the

Joint Staff documents. The Joint Staff worked to incorporate modern communications into

541 Amusingly (in retrospect), the commanding general of the war, Norman Schwarzkopf, the same officer who had thwarted special operators and UAV enthusiasts also tried to reject the Joint Staff-NSA SIGINT and counter-C2 operations; he had to be overruled by Chairman of Joint Chiefs Colin Powell. Kaplan, Dark Territory, 23. 542 Alan D. Campen, ed. The First Information War: The Story of Communications, Computers, and Intelligence Systems in the Persian Gulf War (Fairfax, VA: AFCEA International Press, 1992); see also Robert David Steele. “C4I: The New Linchpin,” U.S. Naval Institute, Proceedings (July 1992). 543 Scott Shane, “John Michael McConnell, A Member of the Club,” New York Times, January 5, 2007, A3, https://www.nytimes.com/2007/01/05/washington/05mcconnell.html. 544 Kaplan, Dark Territory, 32. 545 Ibid., 27. 226

existing military concepts, even re-working Bill Perry’s ideas about counter-command and control warfare to minimize the role of computers and networks. The first directives on

“information warfare” and “command and control warfare” illustrate this dialogue.546 Issued by the Deputy Secretary of Defense in 1990 the OSD directive generally established information warfare as a key component of modern warfare and exhorted every part of DoD to take it seriously without further specification. Issued by then-CJCS Powell in 1993, the Joint Staff policy guidance described command and control warfare as “the integrated use of operations security (OPSEC), military deception, psychological operations (PSYOP), electronic warfare

(EW) and physical destruction, mutually supported by intelligence.” The Joint Staff’s idea of C2 warfare was thus an amalgam of traditional military activities, whereas Perry’s notion focused more squarely on new communications technology.

Both documents also talked about control over information and communications as a decisive element of modern combat. The OSD document stated: “The objective of information warfare is to attain a significant enough information advantage to enable the force overall to predominate and to do so quickly.” The Joint Staff document declared that C2 warfare (or C2W) could allow U.S. forces to defeat enemies decisively before a war even fully began: “C2W offers the commander the potential to deliver a KNOCKOUT PUNCH [sic] before the outbreak of traditional hostilities.”547 Critically, the Joint Staff document described command and control warfare as “both offensive and defensive,” a conceptual framing that implied the dual purposes of what would in the future be called cyber capabilities.548

546 U.S. Department of Defense, Information Warfare, DoD Directive TS-3600.1, Washington, DC: Department of Defense, December 1992), https://www.esd.whs.mil/Portals/54/Documents/FOID/Reading%20Room/Other/14-F- 0492_doc_02_Directive_S-3600-1.pdf; Chairman of the Joint Chiefs of Staff, “Memorandum of Policy No. 30 (CMOP); Command and Control Warfare,” (memorandum, Washington, DC, March 8, 1993 as revised; originally issued July 17, 1990), https://apps.dtic.mil/dtic/tr/fulltext/u2/a389344.pdf. 547 Ibid., 8. 548 Ibid., 5. 227

This Pentagon-level dialogue, alongside the wake-up calls of the late 1980s, spurred some cyber-specific reorganization among the military services but not in OSD. However, service reorganizations also tended to be at lower levels. Information Warfare centers sprang up in the Air Force, Navy and the Army. The Air Force also responded at the tactical level, standing up the “first true combat cyber unit” in the form of the 609th Information Warfare Squadron in

1995.549

Throughout the 1990s, civilian and military-penned documents leap-frogged each other, adding concepts of cyber war in turn and making their way toward the weaponization of cyber tools and greater emphasis on offensive uses of those tools. By 1996, the Joint Staff issued its first proper doctrine on command and control warfare, claiming that “the full dimensions of IW policy and its implementation are still emerging.”550 By the end of the year, the deputy secretary’s office again issued a directive on the topic, this time calling it “information operations” and defining such operations as “actions taken to affect adversary information and information systems while defending one’s own information, and information systems.”551 This framing echoed the Joint Staff’s earlier assertion that cyber activities were often offensive and defensive at the same time. But it also took another step toward offensive uses of cyber capabilities as the first high level DoD document to use the term “computer network attack.” In

1998 General Hugh Shelton, now the CJCS, issued updated doctrine on information operations

549 Healey, A Fierce Domain, 35 550 U.S. Joint Chiefs of Staff, Joint Publication 3-13.1, Joint Doctrine for Command and Control Warfare (C2W) (Washington, DC: Department of Defense, 1996), 1, https://www.hsdl.org/?abstract&did=3760. As one observer has noted, “This implied, a little disingenuously in the light of [the information warfare directive], that the Pentagon had not made up its collective mind about the larger field of information warfare.” Michael Warner, “Notes on Military Doctrine for Cyberspace Operations in the United States, 1992-2004,” Cyber Defense Review, August 27, 2015, https://cyberdefensereview.army.mil/CDR-Content/Articles/Article-View/Article/1136012/notes-on-military- doctrine-for-cyberspace-operations-in-the-united-states-1992/. 551 U.S. Department of Defense, Information Operations (IO), DoD Directive S-3600.1 (Washington, DC: Department of Defense, December 9, 1996), https://fas.org/irp/doddir/dod/info_ops.pdf. 228

that reiterated the concepts of simultaneously offensive and defensive actions and computer network attack.552 But in another new step, Shelton’s document introduced the idea of “the information environment,” a precursor to the notion of cyberspace as a domain.

Despite the building policy and doctrine at the Pentagon, OSD’s, the JS’, and the services’ expertise in and development of cyber capabilities still lagged considerably behind the private sector and even other parts of government, most notably the NSA. Practitioners there were already using the term “cyberspace” by the mid-1990s, defining it as an operating environment, and pondering the strategic implications of this new intelligence domain. On one level, NSA’s approach seemed a natural fit with the rest of the DoD and the military services in particular. “The future of warfare” declared one in-house publication—authored by a special assistant to the Director of the NSA, implying close affiliation with the institution’s broader direction—“is warfare in cyberspace.”553 But NSA was also consolidating power over government activity in cyberspace, having been assigned the task of developing computer network attack capabilities by the Secretary of Defense in 1997. The same article argued that cyber capabilities’ “central activities” were defense, espionage, and offense, and these would be

“worked together.”554 NSA’s position was that the government should adapt to the Information

Age because it was folly to try to force the Information Age to adapt to the government. This, of course, was exactly what the uniformed side of the Pentagon was trying to do: force cyber issues to complement existing operational concepts based on the physical dominance of operational domains and kinetic destruction of targets.

552 U.S. Joint Staff, Joint Publication 3-13. Joint Doctrine for Information Operations (Washington, DC: Department of Defense, October 9, 1998), https://www.jcs.mil/Portals/36/Documents/Doctrine/pubs/jp3_13.pdf. 553 William B. Black, Jr., “Thinking Out Loud about Cyberspace,” Cryptolog 8, no. 1 (Spring, 1997): 3, https://nsarchive2.gwu.edu//dc.html?doc=2700088-Document-11. 554 Ibid., 2. The article went on to explain that the nature of cyberspace was revolutionary, observing “a great deal of confusion and turmoil as human nature attempts to force the “new” of the Information Age into the “known” of the Industrial Age.” 229

This essential tension between the NSA’s technological and intelligence-based interpretation of the uses of cyber capabilities and the uniformed military’s desire to use cyber tools in ways analogous to existing defense and offensive weapons formed the basis of a contentious NSA-military relationship that continues to this day. This tension likely increased when the NSA conducted, at its own initiative although with a “no notice” espionage exercise on

DoD systems code-named “Eligible Receiver.” Authorized by the Chairman of the Joint Chiefs of Staff General John Shalikashvili, it took four days for eight hackers to get inside the entire

DoD network, including the National Military Command Center.555

Eligible Receiver was another major wake-up call for the Pentagon. But, as Congress increasingly understood, DoD had been hitting the snooze button for several years already. By

1995, for example, Senators were asking Defense officials questions about DoD computer network vulnerabilities during confirmation hearings.556 In 1995, in response to a request from

Congressional leaders, including Senator Sam Nunn (R-GA), the GAO released a report titled,

“Information Security: Computer Attacks at Department of Defense Pose Increasing Risks.” The report found that “[a]ttacks on Defense computer systems are a serious and growing threat” and that “[t]he potential for catastrophic damage is great.” In GAO’s estimation, DoD was being reactive to attacks as it identified them rather than proactive in its own defense.557

Compared to the NSA and the public sector, and even some quarters in Congress, the corporate military was behind the cyber curve. Part of the armed services’ delay in creative

555 Kaplan, Dark Territory, 65-70. 556 During then-Deputy Secretary of Defense John Deutch’s hearing to become the next director of Central Intelligence, Senator Mack asked Deutch whether DoD was “moving too quickly to increase its reliance on modern computer systems, which in wartime could be vulnerable to disruptions or alteration.” United States, Senate, Hearing before the Select Committee on Intelligence, Nomination of John M. Deutch to be Director of Central Intelligence, 104th Congress, 1st sess., April 26 and May 3, 1995, https://www.intelligence.senate.gov/sites/default/files/hearings/nominationofjohn1995unit.pdf. 557 General Accounting Office, Information Security: Computer Attacks at Department of Defense Pose Increasing Risks, GAO/AIMD-96-84 (Washington, DC: GAO, May 22, 1996), https://fas.org/irp/gao/aim96084.htm. 230

thinking about cyber issues was their effort to shoe-horn communications technology into existing military concepts. But part of the problem was also that the few people who worked on cyber inside DoD kept the details of cyber issues, including U.S. capabilities, classified.

According to the first commander of the 609th Information Warfare Squadron, in the mid-1990s

“nobody [within the Air Force] really understood what information warfare was” and “the offensive piece was still classified. You couldn’t even discuss it in an open forum,” making it difficult to socialize cyber capabilities with the wider force.558 This turned out to be true more broadly, as secondary sources frequently hint.559

There was also a tendency across the government to see computer network vulnerabilities as an element of other threats rather than its own stand-alone issue. Terrorism in particular was of growing concern, and in fact prompted some of the Clinton administration’s most extensive policy explorations of cyber issues, and almost by accident. Grappling with the aftermath of the domestic terrorist attack in Oklahoma City that killed 168 people, the White House ordered a study run by the Department of Justice to examine ways to secure so-called critical infrastructure. Many of the participants in the Critical Infrastructure Working Group study were from the NSA and Naval intelligence communities, and therefore dedicated substantial parts of the final report to recommendations for “cybersecurity.” The director of the study, deputy attorney general Jamie Gorelick, also had a background as the general counsel at DoD and was alarmed by DoD briefings on the ever-growing private sector reliance on computer

558 “Transcript: Lessons from our Cyber Past—the First Military Cyber Units,” Atlantic Council, March 5, 2012, https://www.atlanticcouncil.org/commentary/transcript/transcript-lessons-from-our-cyber-past-the-first-military- cyber-units/. James Lewis, a former cyber expert with the State Department, corroborated that offensive capabilities were discussed separately from defensive capabilities. James A. Lewis, interview with the author, November 15, 2019. 559 For example, Fred Kaplan writes about the concern among NSA officials that growing awareness of cyber issues within the government could lead to leaks about U.S. activities in cyberspace, some of them already offensive and apparently highly classified. Kaplan, Dark Territory, 49. 231

networking.560 By 1997, Clinton had directed that a national commission study these issues further.561 Chosen as chairman for the commission was a retired four-star Air Force general,

Robert Marsh, who in turn called upon the RAND computer scientist who had first highlighted the vulnerabilities inherent to computer networking, Willis Ware.562 The effort eventually resulted in Presidential Decision Directive-63, “Critical Infrastructure Protection,” issued in

1998.563 As with the original study led by Jamie Gorelick, PDD-63 emphasized the vulnerabilities introduced by national reliance on computer networking and thus focused on the defensive purposes of cyber capabilities. It treated cyber infrastructure as equal in importance to physical infrastructure and spread responsibility for cyber defense across government, including naming agencies as sector-specific liaisons with private companies and industries. DoD was referred to as a kind of consulting firm for other agencies and retained its (specifically NSA’s) executive agent role regarding national telecommunications security. The word “attack” was always used in reference to the U.S. as victim, never as attacker. Although this did not mean the

U.S. had not developed or used offensive cyber tools per se, it did mean the balance of emphasis on cyberspace and cyber capabilities were about protection and security rather than weaponization.

For the time being, mounting better defenses seemed like a first rational response to real- world events. Shortly after Eligible Receiver demonstrated the vulnerabilities of DoD’s networks and the White House issued its directive on critical infrastructure, DoD suffered another cyber- attack on its unclassified networks, this one at the hands of teenagers living in California,

560 Ibid., 40. 561 Commission on Critical Infrastructure Protection, Critical Foundations: Protecting America’s Infrastructures (Washington, DC: October 1997), https://fas.org/sgp/library/pccip.pdf. 562 Kaplan, Dark Territory, 50. 563 The White House, Presidential Decision Directive-63: Critical Infrastructure Protection. (Washington, DC: The White House, May 1998), https://fas.org/irp/offdocs/pdd/pdd-63.htm. 232

(although for days no one could attribute the intrusions to their source).564 Deputy Secretary of

Defense John Hamre, disturbed by the Eligible Receiver exercise and this fresh attack, instructed the Joint Staff to create a central organization responsible for defending DoD networks. The result was “the first joint cyber war-fighting organization,” the Joint Task Force-Computer

Network Defense (JTF-CND).565 The resistance to change first expressed in the early command and control doctrine a few years earlier now expressed itself in leeriness among military communities about a new cyber-focused military organization. This meant that the first joint operational cyber unit had to survive and fit into existing military culture before it could even begin to think about challenges outside the Pentagon. The first JTF-CND commander, then-

Major General John Campbell, explained that it took “a lot of work in the tank [where the Joint

Chiefs of Staff meet]” to establish the task force. “If it had not been for the top cover that John

Hamre provided as the DepSecDef [sic] and – Denny Blair I think was the – was the director of the joint staff who had a big piece of this – we probably wouldn’t have made it.”566 Moreover,

Campbell revealed that his staffing “philosophy was that if you’re going to have any credibility with the war fighters, you had to have operational people” instead of computer experts. The problem with that was a gap in technical expertise, so the task force wound up sending personnel out for supplemental training.567 Technical expertise was more important than combat experience.

But the JTF-CND soon had the chance to prove its utility. By the time it set up shop at the end of 1997, a new intrusion had been going on for months, and the task force went to work

564 Kaplan, Dark Territory, 73-77. 565 Healey, A Fierce Domain, 44. 566 Atlantic Council, “Transcript: Lessons from our Cyber Past—the First Military Cyber Units.” 567 Ibid. 233

on it alongside the FBI, NSA, and other intelligence agencies.568 Nicknamed Moonlight Maze, the various agencies tracking the hackers eventually determined that they were Russian in origin, and were looking for information about weapons development.569 In the months that followed,

American officials got a crash course in the many facets of cyber competition. While the FBI opened a criminal investigation—the hack being the equivalent of trespassing and theft—DoD continued to collect intelligence on the hackers and the White House pursued quiet diplomacy with Russia. At first the Russians were friendly and cooperative, but swiftly changed their approach to intransigence.570 The turnaround was ominous. The U.S. was not just dealing with teenage hackers or insider red teams now. It was dealing with a former and future state adversary.

Meanwhile, excitement about new computer technologies and the role they could play in the Revolution in Military Affairs was rising, but unevenly. One Navy ensign and military innovation scholar, Thomas Mahnken, crowed in the pages of Joint Forces Quarterly, “decisive outcomes are likely where one side has a marked information advantage.”571 The “Secretary’s

Message” in the 1997 QDR likewise promoted the idea of “information superiority”—an idea explicitly derived from the Chairman’s Joint Vision 2010 document released in 1996.572

However, the 1997 DoD annual report relegated cyber issues to a single paragraph, and only as

568 Kaplan, Dark Territory, 82-85. 569 Michael A. Vatis, Director, National Infrastructure Protection Center, Federal Bureau of Investigation, Statement for the Record before the Senate Armed Services Committee Subcommittee on Emerging Threats and Capabilities, March 1, 2000, 11; In a sign of asymmetry of personnel between DoD and other agencies, one news report issued a year before Vatis testified stated that the FBI’s goal was to have 243 agents working on cyber defense issues out of a total number of 11,639 agents—and that the bureau was well short of its own target, employing only “a handful” of cyber specialists at that time. Robert Suro, “FBI Lagging Behind on Cyber Crime,” Washington Post, October 7, 1999, A2, https://www.washingtonpost.com/wp-srv/national/daily/oct99/cyber7.htm. 570 Kaplan, Dark Territory, 87. 571 Thomas G. Mahnken, “War in the Information Age,” Joint Forces Quarterly (Winter 1995-1996): 39-43. 572 Department of Defense, Report of the Quadrennial Defense Review (Washington, DC: May 1997), vi, https://www.files.ethz.ch/isn/32542/qdr97.pdf. 234

an element of battlefield information-gathering and warfighter connectivity in a still- experimental context.573 Cyber was becoming divorced from information operations but had not yet achieved status a separate area for policy.

Regardless, by 1999 the internet was a part of daily life in the U.S., and so were viruses, worms, hacks, and the dread over the supposed (and ultimately minimal) “Y2K” or millennium bugs in most computer operating systems.574 The proliferation in internet-based menaces reflected a general explosion in the number of actors in cyberspace—governments, companies, and private citizens of the innocent and criminal varieties.575 From its origins at DARPA, networked communications had been democratized and made widely available to the national and global public.576 By then, the entire U.S. government understood the vulnerabilities inherent to dependence on computer networking and the internet. Its answer to those challenges was a federated, every-agency-for-itself approach that also recognized the autonomy of the private sector. The military—particularly the Air Force and the Navy—had processed the lessons learned in the Persian Gulf War about both the utility and vulnerability of networked operations but was still scrambling to reconcile cyber capabilities and information operations with doctrine, force structure, training, and organization even as technologies and techniques evolved ever- more-rapidly.

573 U.S. Department of Defense, Annual Report to the Congress, FY1988 (Washington, DC, January 1987), 144, https://history.defense.gov/Portals/70/Documents/annual_reports/1988_DoD_AR.pdf?ver=2014-06-24-151204-840. The language was also quite technical: “The Federated Battle Lab represents a major step towards the creation of a virtual joint experimentation environment. The cyberspace linking of each Service's battle labs will enable more rapid development of joint concepts and equally rapid initial testing and experimentation utilizing state-of-the-art models and simulations, without duplicating efforts.” 574 “Watching for the Y2K Bug,” New York Times, December 30, 1999, A26, https://www.nytimes.com/1999/12/30/opinion/watching-for-the-y2k-bug.html. 575 Healey, A Fierce Domain, 50-52. 576 For more on the consequences of open innovation systems and emerging technologies, see Audrey Kurth Cronin, Power to the People: How Open Technological Innovation is Armed Tomorrow’s Terrorists (Oxford, UK: Oxford University Press, 2020). 235

Once again, real-world military experience prompted an evolution in how the military thought about cyber capabilities. If the first Gulf War was a dress rehearsal for the future of cyber operations, the NATO campaign in Kosovo was opening night. And like most opening nights, it revealed strengths and weaknesses in the production. Lasting from March to July of

1999, Operation Allied force involved cyber-attacks on Serbian networks, including on its air defense systems.577 The U.S. and NATO also suffered attacks from the Serbs with suspected aid from Russia, although none of the attacks were deemed to have had a strategic effect on the conflict.578 Kosovo was a kind of cyber experiment for all sides, the most important lesson of which may have been how unconcerned actors were about escalating the violence through cyber activities, either in terms of physical damage or diplomatic blowback.

Operations in Kosovo exposed the uneven investment in military personnel trained to execute cyber missions and the hesitation to use cyber tools at unspecified, but higher, levels in the chain of command. As ever, classification and the isolation of information and skills blocked knowledge and use of cyber capabilities. Altogether, the picture was of military disintegration regarding cyber tools, even though a newfound operational sophistication was also evident. One news article summed up the successes and failures neatly:

The Air Force Office of Information, in a 28-page report on Operation Allied Force released last week, all but confirmed that IO [Information Operations] played a significant role in the Serbian campaign. “The secret new arts of disrupting enemy capabilities through cyber-space attacks appeared to have been a big part of the campaign,” according to the report.

[A separate] naval briefing called the IO campaign "perhaps the greatest failure of the war." Although "all the tools were in place . . . only a few were used," according to the briefing. The briefing praised the IO staff as "great people" but said they "were too junior and from the wrong communities to have the required impact on planning and execution."

577 Kaplan, Dark Territory, 114. 578 Kenneth Geers, “Cyberspace and the Changing Nature of Warfare,” NATO Cooperative Cyber Defense Center of Excellence, n.d., https://ccdcoe.org/uploads/2018/10/Geers2008_CyberspaceAndTheChangingNatureOfWarfare.pdf. 236

IO is "not yet understood by warfighters . . . and classified beyond their access," according to the briefing.579

The use of cyber capabilities in Kosovo also prompted DoD’s first comprehensive examination of how the laws of armed conflict applied in cyberspace, a process that may have constrained military actions in real time.580 In May of 1999, midway through Operation Allied

Force, the DoD General Counsel’s office—a component of the civilian-led OSD—released a report entitled, “Assessment of International Legal Issues in Information Operations.” According to a preface note, the document was prepared by a large civil-military team, and its principal drafter was a retired Air Force colonel. The report discussed what constituted an armed attack in or via cyberspace, concluding that the “consequences” rather than the “means” were the important factor.581 It also may have been the first codified introduction of the notion of “active defense” in cyberspace, an idea that amounted to retaliatory punishment for intrusions and attacks.582

However, few military and civilian officials seemed to pay attention to such cyber lessons. Information warfare was still considered a kind of combat support function, of less interest to senior officers. Despite the example set by Kosovo, then-ASD(C3I) Art Money struggled to connect his office and the cyber team on the Joint Staff with combatant command planners to further integrate information operations with broader DoD planning.583 So Money worked with the new NSA Director, General Michael Hayden, to get the NSA “into cyber

579 Bob Brewin, “Kosovo Ushered in Cyberwar,” Federal Computer Week, September 27, 1999, https://fcw.com/articles/1999/09/27/kosovo-ushered-in-cyberwar.aspx. 580 Julian Borger, “Pentagon Kept the Lid on Cyberwar in Kosovo,” The Guardian, November 8, 1999, https://www.theguardian.com/world/1999/nov/09/balkans. 581 Office of General Counsel, Assessment of International Legal Issues in Information Operations (Washington, DC: Department of Defense, May 1999), 18, https://fas.org/irp/eprint/io-legal.pdf. 582 Ibid., 22-23. 583 Kaplan, Dark Territory, 121. 237

offensive operations.”584 But the small civil-military coalition of cyber warriors was unable to convince broader audiences in DoD to break cyber out of its defensive and support purposes, and unable to convince intelligence analysts to risk exposing their access to foreign networks by using them to undermine adversaries.585

The information warriors in DoD were themselves still digesting what had happened during operations in the Balkans when al Qaeda conducted its terrorist attacks on American soil on September 11, 2001. Because the attacks exposed physical vulnerabilities rather than poor computer network defenses, cyber experts immediately queried their contributions to the intelligence failure rather than focusing on the implications for cyberwarfare.586 Popular questions followed for years about the possibility of cyber-terrorism or a “cyber 9/11” and of the ways terrorists used the internet.587

But by and large, and unlike the other two case studies in this volume, the 9/11 attacks did not have a great effect on how military and civilian actors viewed the military purposes of

U.S. cyber capabilities. Rather, 9/11 was one more data point about the need for defenses against vulnerabilities and computer network exploitation tools for intelligence-gathering efforts.

Moreover, the new Department of Homeland Security (DHS) became the focal point in the government for cybersecurity—the mission assignment itself another expression of civilians’

584 Ibid., 125. 585 A big part of the hesitation on the part of those who were not zealous cyber warfare proponents was legal. Cyber war often operates in the gray area between DoD’s Title 10 authorities for military activities and the intelligence community’s Title 50 authorities over covert operations. For a thorough exploration of these issues, see: Robert Chesney, “Military-Intelligence Convergence and the Law of the Title 10/Title 50 Debate,” Journal of National Security Law and Policy 5 (2012): 539-629, https://jnslp.com/wp-content/uploads/2012/01/Military-Intelligence- Convergence-and-the-Law-of-the-Title-10Title-50-Debate.pdf. 586 The 9/11 Commission report itself noted that, “An almost obsessive protection of sources and methods by the NSA, and its focus on foreign intelligence, and its avoidance of anything domestic would, as will be seen, be important elements in the story of 9/11.” National Commission on Terrorist Attacks upon the United States, The 9/11 Commission Report (Washington, DC: National Commission on Terrorist Attacks upon the United States, 2004), 88, http://govinfo.library.unt.edu/911/report/911Report.pdf. 587 Clifford S. Magee, “Awaiting Cyber 9/11,” Joint Forces Quarterly 70 (3rd quarter, 2013), https://ndupress.ndu.edu/Portals/68/Documents/jfq/jfq-70/JFQ-70_76-82_Magee.pdf. 238

concept of cyberspace as something requiring defense rather than prompting offense.588 It became conventional wisdom that terrorists used the internet and other cyber networking platforms mostly for communication and propaganda dissemination, not as a medium of attack itself.589 If anything, the unease caused by uncertainty of random attacks only bolstered desires to shore up vulnerabilities in critical infrastructure, from transportation to computer networks.

Indeed, the Federal Information Security Management Act of 2002 was one of the legislative responses to 9/11 and was defense minded.590

The cyber field thus took note of terrorism’s new salience to national security but was not especially shaped by it. Nevertheless, it is possible that civilian officials and military personnel alike shifted some of the attention they might have paid to cyber issues toward the threat of

(physical) terrorism instead. From 2002-2009, the Senate Armed Services Committee did not hold any hearings on cyber issues specifically (although cyber came up in nominations and the annual worldwide threat assessment from the intelligence community).591 The first hearing to break the trend occurred after DoD established Cyber Command with a January, 2010 hearing on

“Cybersecurity.” 2012 was the first year with multiple SASC hearings on cyber issues.

Moreover, the intellectual trends in cyber were unperturbed by the attacks on September

11. For the broader government, cyber issues remained focused on critical infrastructure protection, on network and data vulnerabilities. For the military, cyber stayed within a context of

588 Healey, A Fierce Domain, 64. 589 According to Jim Lewis, “the Internet is not a weapon that appeals to terrorists” because “cyberattacks would not have the dramatic and political effect that meets the psychological needs of terrorists to commit violent attacks.” James A. Lewis, “The Internet and Terrorism,” Proceedings of the Annual Meeting, American Society of International Law 99 (March 30-April 2, 2005): 112-115, doi:10.1017/S0272503700071196. There was also an appreciation of cyber conduits for terrorist financing. See: Michael Jacobsen, “Terrorist Financing on the Internet.” CTC Sentinel, Volume 2, Issue 6, June 2009, https://ctc.usma.edu/terrorist-financing-on-the-internet/. 590 Daniel M. White, “The Federal Information Security Management Act of 2002: A Potemkin Village., Fordham University Law Review 79, no. 1 (2011): 369-405, https://ir.lawnet.fordham.edu/flr/vol79/iss1/12/. 591 Search of full committee and subcommittee hearings conducted by the author November 11, 2019 on the Senate Armed Services Committee public website: https://www.armed-services.senate.gov/hearings?year=2007&c=all. 239

technological change on the conventional battlefield. For example, DoD’s 2005 annual report to

Congress couched information operations in terms of the transformation agenda. “One trend is clear: the Department’s transformation will be shaped by the emerging realities of the information age.”592 The report integrated information and communications technology into service-level agendas for improving battlefield operations. This was less about turning cyber into weapons, however, and more about making information and communication technologies enable the effectiveness of other weapons.

Nevertheless, and separate from broader concerns about countering terrorism and the role of information warfare in defense transformation, an evolution in the approach to cyber capabilities was underway. NSA Director Hayden, encouraged by civilian ASD(C3I) Money, had continued the trend, first evident in Joint Staff documents in the 1990s highlighted earlier, toward marrying the defensive and offensive sides of cyber, seeing them as aspects of each other.

Moreover, the concept of computer network exploitation (CNE) took hold, a kind of middle ground between purely defensive measures and destructively offensive ones because it could serve either goal, by gathering information useful to defending U.S. networks or sabotaging adversary networks.593

But the conflation of offense and defense was not generally welcome among intelligence analysts who wanted to preserve hard-earned cyber access and feared that offensive operations would only, and sometimes literally, blow up their efforts. But as the wars in Afghanistan and, especially, Iraq proceeded, tactical commanders placed increasing demand on NSA for the ability to hack terrorists’ internet- and cellphone-based communications. At the same time,

592 U.S. Department of Defense, Annual Report to the President and the Congress, FY2005 (Washington, DC: Department of Defense, 2005), https://history.defense.gov/Portals/70/Documents/annual_reports/2005_DoD_AR.pdf?ver=2014-06-24-153944-813. 593 Kaplan, Dark Territory, 138-139. 240

General Keith Alexander moved from running intelligence operations at Central Command to the

Army’s Intelligence and Security Command, and worked with General Stanley McChyrstal, commander of the Joint Special Operations Command, to get such tools in the hands of operators in Iraq.594 When Alexander, who had invested time in persuading Secretary of Defense Rumsfeld of the value of intelligence-driven operations aided by computer network exploitation tools, was tapped to become the next Director of the NSA in 2005, he was poised to bring cyber capabilities squarely into military operations.595

By 2006, then, wider circles of the U.S. military had shifted toward preferring to use cyber capabilities in offensive ways, or at the very least in ways that enabled other offensive activities. Meanwhile, a handful of senior civilians had begun to embrace the more routine uses of computer network attack and the marriage of offensive and defensive approaches to military cyber activities. But on the whole, strategic documents put out by the White House and even

OSD still emphasized defensive uses of cyber tools, and offensive cyber activities were still highly constrained by lawyers and policymakers alike. But the ranks of advocates of offensive cyber operations were growing.

Militarization: 2006-2018

The militarization phase in the cyber case study saw an increase in the number of civilian actors who preferred offensive uses of cyber capabilities in addition to traditional defensive uses.

Using cyberspace and cyber capabilities as part of regular war plans had long been military policy, but the publication of the 2006 National Military Strategy for Cyberwarfare seemed to galvanize the weaponization of cyber. It is clear that military actors had long wished to use cyber

594 Ibid., 150. 595 Ibid., 156-160. 241

tools offensively—and were doing so effectively in Iraq—but were often limited in conducting broader offensive cyber activities by law and policy. The tensions between military preferences and civilian-imposed constraints came to a head during the Obama administration and seem to have largely resolved in favor of Cyber Command during the Trump administration, as of this writing.

At the same time, the proliferation of so-called bad actors in cyberspace produced relentless attacks, disruptions, and threats. The sources of these threats continued to be a mix of government, government-sponsored, and private entities, even as attribution continued to bedevil

U.S. officials. However, actions by Russia, China, North Korea, and Iran were identifiable enough to emerge into the public record.

The signal event of the period’s start was the then-classified publication of the National

Military Strategy (NMS) for Cyberspace Operations in 2006.596 As a DoD cyber strategy, the document was unprecedented in its scope and detail, merging previously disjointed doctrinal terminology into the concept of cyber. Most important to this inquiry, it superimposed a warfighting vocabulary over the range of concepts related to cyber that had evolved over the previous two decades, stating that DoD needed to develop joint capabilities for “battlespace awareness,” command and control, critical infrastructure protection, information operations, and net-centric operations among others. None of that was new. But the document also began a shift from framing cyber warfare as defensive to offensive. Terms such as “dominance,” and

“superiority” and phrases such as “engage adversaries decisively” conveyed cyber as means to defeat enemies in battle. Most important, the document highlighted the benefits of offensive approaches: “Offensive capabilities in cyberspace offer the United States and our adversaries an

596 The document has since been partially released via the Freedom of Information Act. 242

opportunity to gain and maintain the initiative.”597 Such framing was an evolution from long- stated cyber policies that focused on shoring up vulnerabilities and using cyber capabilities solely to defend networks, data, and communications. Now, the Joint Staff was forging the conceptual space to take cyber in a more proactive, aggressive direction.598

Indeed, the NMS was not just a strategy for conducting cyber operations. It was also a strategy for building DoD’s cyber capacity. It defined cyberspace as an operational domain co- equal with land, sea, air, and space, a designation that implied a dramatic increase or shift in force structure and other military organizations to develop “superiority” over that new domain.

Appendix F detailed the four strategic priorities that military departments “should consider as they allocate priorities,” made joint doctrine development one of the priorities, and called on the joint force to “change and implement the appropriate rules of engagement (ROE) to facilitate cyberspace operations.”599

Despite the fact that at least one Secretary of Defense and multiple ASDs for C3I had directed the military to build up its computer network attack capabilities, the forward-leaning approach to cyber displayed in the 2006 NMS amplified the military’s own commitment to weaponizing cyber tools in a way that exceeded the preferences expressed by most civilian officials. Even within the NMS itself, the contrast between the Secretary’s and the Chairman’s forwarding notes showed the military’s greater enthusiasm for offensive cyber capabilities.

597 U.S. Joint Chiefs of Staff, National Military Strategy-Cyber Operations (Washington, DC: Department of Defense, 2006), 10, https://www.hsdl.org/?abstract&did=35693. 598 Of course, offensive uses of cyber capabilities preceded this high-level reframing but had traditionally been only about 30 percent of the work of the Joint Staff’s Joint Task Force-Computer Network Defense. But three years prior to the NMS publication, offensive capabilities were shifted to a new organization at the NSA under the control of Strategic Command. “NSA got this offensive mission for several reasons, not least because of the depth of their expertise in cyber issues, but also due to their unique ‘accesses’ into foreign cyber systems. NSA was also helped because its Director has regular access to the White House as a senior intelligence official, a privilege not normally extended to generals that ran military networks.” Healey, A Fierce Domain, 65. 599 U.S. Joint Chiefs of Staff, National Military Strategy-Cyber Operations, F-1. 243

Whereas the Secretary emphasized U.S. military vulnerabilities due to its reliance on computing and networks, the Chairman highlighted integrating offensive capabilities with defensive activities.600

The 2006 QDR, approved and released several months later, conveyed confusion on the civilian side about cyberspace and the purposes of cyber capabilities. This may have been a product of the document’s unclassified status, in contrast to the NMS. Nevertheless, in contrast both to the NMS and to other topics addressed in its own pages, the 2006 QDR’s references to cyber were so general that little meaning could be gleaned from them. For example, it promised the DoD would build “capabilities to shape and defend cyberspace” along with “joint command and control capabilities that are survivable” after a cyber-attack. The section on decisions affecting Strategic Command (which managed both nuclear and cyber capabilities at the time) listed specific decisions such as the retirement of four E4-B aircraft and the Mobile Consolidated

Command Center by 2007 but for more directly cyber-related matters vaguely referred to

“additional investments,” and said that the Department would “strengthen coordination” and

“leverage lessons learned.”601 A National Security Presidential Directive on Cybersecurity issued two years later (and since declassified) also couched cyber in defensive terms. It starts with description of vulnerabilities and places DHS as the lead agency, coordinator, and defender of

U.S. cyberspace. DoD simply shares “indications and warning” responsibilities with the Director of National Intelligence (DNI) and is directed to defend its own networks. Even the purpose of offensive capabilities is “to defend U.S. information systems” and DoD shares that responsibility

600 Ibid. 601 U.S. Department of Defense, Quadrennial Defense Review Report (Washington, DC: February 2006), 50-51, https://archive.defense.gov/pubs/pdfs/QDR20060203.pdf. The document did raise cyber more than previous public defense strategies. It observed the specter of cyber-terrorism, calling the internet a “cyber-sanctuary” for terrorist groups. It highlighted Americans’ vulnerability to cyber-attacks, pledging DoD efforts toward deterrence of such attacks, and also observed increasing Chinese willingness to conduct cyber warfare. 244

with the Secretary of State, the DNI, the Attorney General, and DHS.602 The third period of the cyber case thus began with civilians still thinking of cyber capabilities in largely defensive terms—or, at least, still separating the defense and the offense for the purposes of tightly controlling the latter—while the military was transitioning into imagining them for offensive purposes and more persistently advocating such uses.

At the same time, cyber-related threats were multiplying, especially among state actors.

Russia was a breakaway player in the cyber arena in terms of sheer boldness, mixing real-word operations with political warfare in attacks on Estonian and Georgian internet systems in 2007 and 2008 respectively.603 Attacks originating from China were both far more numerous and lower profile, involving “technology theft, intelligence gathering, exfiltration, research on DOD operations and the creation of dormant presences in DoD networks for future action.”604

The increase in state-based cyber-attacks and higher tempo of activity raised awareness among U.S. officials, but it was a major hack on Central Command in 2008 that spurred DoD thinking and generated momentum toward creating a Cyber Command. The hack was different because it breached classified networks, reportedly giving the intruder—suspected to be a

Russian intelligence agency—access to war plans.605 Deputy Secretary of Defense Bill Lynn revealed the incident in a Foreign Affairs essay, calling it “the most significant breach of U.S.

602 The White House, National Security Presidential Directive-54/Homeland Security Presidential Directive-23 (Washington, DC: The White House, January 8, 2008), https://fas.org/irp/offdocs/nspd/nspd-54.pdf. 603 “A Cyber-Riot,” The Economist, May 10, 2007, https://www.economist.com/europe/2007/05/10/a-cyber-riot; John Markoff, “Before the Gunfire, Cyber Attacks,” New York Times, August 12, 2008, A1, https://www.nytimes.com/2008/08/13/technology/13cyber.html; and William C. Ashmore, “Impact of Alleged Russian Cyber Attacks,” Baltic Security and Defense Review 11 (2009): 4-40, https://nsarchive2.gwu.edu/NSAEBB/NSAEBB424/docs/Cyber-027.pdf. 604 Josh Rogin, “Cyber Officials: Chinese Hackers Attack ‘Anything’ and ‘Everything’,” Federal Computer Week, February 13, 2007, https://fcw.com/articles/2007/02/13/cyber-officials-chinese-hackers-attack-anything-and- everything.aspx?sc_lang=en. 605 Brian Knowlton, “Military Computer Attack Confirmed,” New York Times, August 25, 2010, https://www.nytimes.com/2010/08/26/technology/26cyber.html; and Ellen Nakashima, “Cyber-Intruder Sparks Response, Debate,” Washington Post, December 8, 2011, https://www.washingtonpost.com/national/national- security/cyber-intruder-sparks-response-debate/2011/12/06/gIQAxLuFgO_story.html. 245

military computers ever, and it served as an important wake-up call. The Pentagon’s operation to counter the attack, known as Operation Buckshot Yankee, marked a turning point in U.S. cyberdefense strategy.”606 But that turning point, according to the civilian deputy Defense secretary, did not change the ultimately defensive nature of cyber capabilities. “The Pentagon has built layered and robust defenses around military networks and inaugurated the new U.S.

Cyber Command to integrate cyberdefense operations across the military,” Lynn wrote.

Nevertheless, the event allowed the military to consolidate more authority around cyber issues, especially in the form of Cyber Command. The idea for Cyber Command came from military actors. Former Director of the NSA and Director of National Intelligence, and retired

Vice Admiral Mike McConnell, for one, had long been advocating for DoD to create a cyber combatant command, including to Secretary of Defense Robert Gates himself.607 The proposal for a cyber COCOM was a point of contention between civilian intelligence and military officials, with NSA partisans arguing that a new command would be redundant to the capabilities housed at Fort Meade. With a career background in intelligence, Gates was inclined to broker a compromise.608

In 2009, in response to the hack on CENTCOM, Gates created the new Cyber Command as a sub-component of Strategic Command, although the new entity did not have a functioning staff until the fall of 2010. In his letter directing the command’s creation, Gates disestablished the two main cyber entities in DoD, one that had focused on offensive activities and one that had specialized in defense and transitioned them both into the new command. He also made the

606 William J. Lynn, “Defending a New Domain: The Pentagon’s Cyberstrategy,” Foreign Affairs 89, no. 50 (September/October 2010): 97-108, https://www.law.upenn.edu/live/files/6465-12-lynn-defending-a-new- domainpdf. 607 Kaplan, Dark Territory, 185. 608 Richard Clarke and Robert K. Knake, Cyber War: The Next Threat to National Security and What to Do About It (New York, NY: HarperCollins, 2010). 246

Director of the NSA the simultaneous commander of Cyber Command, and stipulated that the commander be a four-star general, a rank important to establish co-equal status with other combatant commanders (although, because Cyber Command was still subordinate to Strategic

Command, it would not be co-equal with other stand-alone combatant commands and would still have to route its advice to the Secretary of Defense through the Strategic Command commander). Finally, Gates directed that the Joint Staff manage the process that established

Cyber Command’s “mission, roles, and responsibilities,” effectively giving military institutions control over what the new organization would do.609 The command went to work building its force structure, a Cyber Mission Force with 133 teams fulfilling combat and defensive functions.610 Within nine years (and after the Trump Administration made it a stand-alone unified command), the Command’s budget went from $120 million a year to $600 million, and its staffing eventually swelled to over 6,000 military and civilian personnel.611 In response to the creation of Cyber Command, the Air Force and the Navy reorganized their cyber forces, and the

Navy even stood up a new “war-fighting command.”612

The creation of Cyber Command increased the military’s influence over cyber capabilities and their uses, but the new organization was not stepping into a vacuum of

609 Office of the Secretary of Defense, “Establishment of a Subordinate Unified U.S. Cyber Command Under U.S. Strategic Command for U.S. Military Cyberspace Operations,” June 23, 2009, https://www.hsdl.org/?abstract&did=35655. 610 Michael Sulmeyer, “Military Cyber Issues,” Center for Strategic and International Studies, Cyber Policy Task Force Working Group Discussion Papers, 2017. “. . . prior to the standup of US Cyber Command, US Strategic Command employed a stand-alone Joint Functional Component Command for Network Warfare (JFCC-NW) for Title 10 offense and a Joint Task Force for Global Network Operations (JTF-GNO) for Title 10 defense. Under this pre-US Cyber Command construct, the Director of NSA was dual hatted as the commander of JFCC-NW and the Director of the Defense Information Systems Agency (DISA) was dual hatted as the commander of JTF-GNO. Both of these Title 10 constructs were rolled into US Cyber Command, preserving subordination of each of these functions to U.S. Strategic Command and the firewalling of Title 10 military operations from the intelligence support provided by NSA.” 611 Max. W. E. Smeets and Herbert Lin, “A Strategic Assessment of the U.S. Cyber Command Vision,” in Bytes, Bombs, and Spies: The Strategic Dimension of Offensive Cyber Operations, ed. Herbert Lin and Amy Zegart (Washington, DC: Brookings Institution Press, 2018). 612 Clarke and Knake, Cyber War, 41-42. 247

governance. Around the same time, the new Obama administration was working to develop rules of engagement for responding to foreign cyber-attacks, and the interagency debate that ensued demonstrated the range of civilian voices and interests involved, even just within the Executive branch. According to one report in the Washington Post:

The Justice Department feared setting a legal precedent for military action in domestic networks. The CIA resisted letting the military infringe on its foreign turf. The State Department worried the military would accidentally disrupt a server in a friendly country without seeking consent, undermining future cooperation. The Department of Homeland Security, meanwhile, worked to keep its lead role in securing the nation against cyberthreats.613

DHS in particular worked to maintain its jurisdiction over domestic cybersecurity operations. In

2011, Janet Napolitano gave a speech at the University of California’s engineering school and stated, in an allusion to Cyber Command, that “at DHS, we believe that cyberspace is fundamentally a civilian space.”614

These debates, however, were as much about the boundaries that should be placed around cyber activities as the nature of the activities themselves. And in the years that followed, evidence emerged that the U.S. government was increasingly using cyber capabilities for offensive purposes.

The signal event in this regard was the cyber worm dubbed “Stuxnet.” Discovered in

2010, the malware caused centrifuges in an Iranian nuclear facility to break down.615 Although it has never been claimed by the U.S., overwhelming evidence gathered by journalists and cyber specialists suggest it was an American program aimed at destroying or delaying Iranian efforts to

613 Nakashima, “Cyber-Intruder Sparks Debate, Response.” 614 William Jackson, “DHS Secretary: ‘Cyberspace is Civilian Space’,” GCN.com, April 27, 2011, https://gcn.com/articles/2011/04/27/napolitano-dhs-role-cybersecurity.aspx. 615 Kim Zetter, “An Unprecedented Look at Stuxnet, the World’s First Digital Weapon,” Wired Magazine, November 3, 2014, https://www.wired.com/2014/11/countdown-to-zero-day-stuxnet/. 248

build nuclear weapons.616 What is important here is that the program must have been created prior to the establishment of Cyber Command, and the decision to use this new type of cyber weapon, capable of physical destruction, made before the military had re-consolidated its offensive and defense cyber task forces.617

Regardless of Stuxnet’s origins, it appeared to break a seal on the use of cyber tools as weapons against adversary states. Much the way the 2008 attack on Central Command ushered in a new institutionalization of military cyber capabilities, Stuxnet opened the door to offensive uses of those same capabilities. In 2016, President Obama’s last Secretary of Defense, Ash

Carter, and the Chairman of the Joint Chiefs of Staff acknowledged before Congress and in a later press briefing that DoD had used offensive cyber tools against the Islamic State in Syria.618

According to news reports, Obama also appeared to have bequeathed to the Trump White House a cyber campaign against North Korean nuclear launch systems.619

616 Journalistic investigations included David Sanger’s series in the New York Times, which he turned into a book: David E. Sanger, The Perfect Weapon: War, Sabotage, and Fear in the Cyber Age (New York, NY: Random House, 2018). Companies that examined the malware included Symantec, see: https://www.symantec.com/security- center/writeup/2010-071400-3123-99. 617 The malware also had consequences that reached far beyond the Iranian nuclear program. According to the virus protection company McAfee, despite the fact that the designers of Stuxnet programmed it to stop working in 2012, imitations of Stuxnet have plagued the global energy sector ever since, with upwards of 80% of Mexican energy companies and 60% of Indian energy companies affected by these “sons of Stuxnet.” “What is Stuxnet?,” McAfee, accessed November 13, 2019, https://www.mcafee.com/enterprise/en-us/security-awareness/ransomware/what-is- stuxnet.html. 618 Sean Lyngaas, “Carter: U.S. Disrupting Islamic State Computer Networks,” Federal Computer Week, February 29, 2016, https://fcw.com/articles/2016/02/29/carter-isis-networks.aspx; In a later report for Harvard’s Belfer Center, Carter wrote, “I acknowledged in congressional testimony in May 2016 that I had directed U.S. Cyber Command to initiate offensive cyber operations against ISIS—a first for the U.S. military. Our goal, I said, was to disrupt ISIS finances, recruiting and propaganda.” Carter also revealed his civilian alternative source of expertise on cyber matters: “My chief of staff, Eric Rosenbach, was an invaluable adviser on everything having to do with ISIS, but his expertise on cyber and homeland security issues was particularly useful on this thorny topic.” Ash Carter, A Lasting Defeat: The Campaign to Destroy ISIS (Boston, MA: Harvard Kennedy School, Belfer Center for Science and International Affairs, October 2017), https://www.belfercenter.org/LastingDefeat. 619 David E. Sanger and William J. Broad, “Trump Inherits a Secret Cyberwar Against North Korean Missiles,” New York Times, March 4, 2017, https://www.nytimes.com/2017/03/04/world/asia/north-korea-missile-program- sabotage.html. 249

Formal expressions of military preference continued to emphasize offensive cyber tools.

In 2013, the Joint Staff released its Joint Publication 3-12: “Cyberspace Operations,” which further normalized the use of cyber weapons. JP 3-12 defined offensive cyberspace operations as

“intended to project power by the application of force in and through cyberspace” and declared that cyber actions “will be authorized like offensive operations in the physical domains, via an execute order (EXORD).620 As in the NMS before it, JP 3-12 emphasized the marriage of offense and defense in cyberspace—an element of cyber operations that many experts assert, but that also means that any desire to maintain adequate defenses condones continued DoD-led offensive cyber action.

But it was with the transition to the Trump Administration that civilian preferences more publicly and formally adopted offensive purposes for cyber weapons. In 2018, Cyber Command released its own “vision,” presaging to its policy audience that the new cyber operating procedures would be more assertive than in prior approaches:

We have learned we must stop attacks before they penetrate our cyber defenses or impair our military forces; and through persistent, integrated operations, we can influence adversary behavior and introduce uncertainty into their calculations. Our forces must be agile, our partnerships operational, and our operations continuous. Policies, doctrine, and processes should keep pace with the speed of events in cyberspace to maintain decisive advantage.621

Later in 2018, the administration released two cyber strategies, one at the national level and one for DoD. The DoD strategy (of which only a summary was released publicly) echoes the

Cyber Command vision, saying the DoD must “take action in cyberspace during day-to-day

620 U.S. Joint Chiefs of Staff, Joint Publication 3-12: Cyberspace Operations (Washington, DC: Department of Defense, February 5, 2013), II-2, https://www.jcs.mil/Portals/36/Documents/Doctrine/pubs/jp3_12.pdf. 621 U.S. Cyber Command Achieve and Maintain Cyberspace Superiority: Command Vision for U.S. Cyber Command (Fort Meade: Department of Defense, April 2018), https://www.cybercom.mil/Portals/56/Documents/USCYBERCOM%20Vision%20April%202018.pdf?ver=2018- 06-14-152556-010. 250

competition,” and “defend forward to disrupt or halt malicious cyber activity at its source,” a seeming-conclusion to years of debate about whether sovereignty could be violated in the cyber domain and if the U.S. would embrace doing so.622

Cyber Command is now promoting this new, aggressive, approach with broader audiences. Commander of Cyber Command General Paul Nakasone said in a recent Joint Forces

Quarterly interview, “we’re not waiting for adversaries to come to us.” He went on:

Unlike the nuclear realm, where our strategic advantage or power comes from possessing a capability or weapons system, in cyberspace it’s the use of cyber capabilities that is strategically consequential. The threat of using something in cyberspace is not as powerful as actually using it because that’s what our adversaries are doing to us.623

In the online materials advertising DoD’s 2019 Cyber Conflict conference, “defending forward” is the conceptual organizing principle of the gathering and explained as a “preemptive strategy” to disrupt harmful adversary actions.624

Although the Cyber Command Vision and the DoD Cyber Strategy lean into offensive uses of cyber capabilities, the National Cyber Strategy does not. Rather, it refers obliquely to imposing “consequences” in response to “malicious activities.”625 Elsewhere, however, the administration has refused to rule out the U.S. government’s use of cyber-attacks and the president has reportedly ordered cyber-attacks on Iranian targets.626 As the Washington Post reported:

622 U.S. Department of Defense, Summary: Department of Defense Cyber Strategy 2018 (Washington, DC: Department of Defense, 2018), https://media.defense.gov/2018/Sep/18/2002041658/-1/- 1/1/CYBER_STRATEGY_SUMMARY_FINAL.PDF. 623 “An Interview with Paul M. Nakasone,” Joint Forces Quarterly 92 (1st quarter 2019), https://ndupress.ndu.edu/Portals/68/Documents/jfq/jfq-92/jfq-92_4-9_Nakasone-Interview.pdf. Emphasis original. 624 http://aci.cvent.com/events/2019-international-conference-on-cyber-conflict-cycon-u-s-/event-summary- 6fa2dad0721d43249d40ca05d3a5e94a.aspx; accessed in November 2019. 625 The White House, National Cyber Strategy of the United States of America (Washington, DC: The White House, September 2018), 21, https://www.whitehouse.gov/wp-content/uploads/2018/09/National-Cyber-Strategy.pdf. 626 David E. Sanger, “U.S. Declines to Sign Declaration Discouraging Use of Cyberattacks,” New York Times, November 12, 2018, A6, https://www.nytimes.com/2018/11/12/us/politics/us-cyberattacks-declaration.html. 251

[the] strikes against the Revolutionary Guard represented the first offensive show of force since Cyber Command was elevated to a full combatant command in May. It leveraged new authorities, granted by the president, that have streamlined the approval process for such measures. It is also a reflection of a new Cyber Command strategy — called “defending forward” — that its leader, Gen. Paul Nakasone, has defined as operating “against our enemies on their virtual territory.”627

It has also been reported that the president signed a National Security Presidential Memorandum on the military’s use of offensive cyber tools, although the content of the memorandum is unknown, and members of Congress have demanded access to it.628

Has the administration begun to embrace offensive purposes for cyber capabilities in a new way that aligns with Cyber Command’s preferences, as the above quote suggests? Has

Cyber Command expanded enough to control enough information that it can frame and influence civilian thinking? Have alternative civilian voices in DHS and elsewhere been overshadowed by the command? The evidence does not answer these questions conclusively, but it does show the military’s preference to use offensive cyber capabilities, and a growing willingness on the part of civilian officials to defer to the justifications Cyber Command makes for offensive uses of cyber capabilities.

Analysis and Conclusions

The cyber case shows how diverse sources of information for civilians can help them maintain preferences that are independent from military institutional preferences. For about two

627 Ellen Nakashima, “Trump Approved Cyber-Strikes Against Iranian Computer Database Used to Plan Attacks on Oil Tankers,” Washington Post, June 22, 2019, https://www.washingtonpost.com/world/national-security/with- trumps-approval-pentagon-launched-cyber-strikes-against-iran/2019/06/22/250d3740-950d-11e9-b570- 6416efdc0803_story.html. 628 Ellen Nakashima, “Trump Gives the Military more Latitude to Use Offensive Cyber Tools Against Adversaries,” Washington Post, August 16, 2018, https://www.washingtonpost.com/world/national-security/trump-gives-the- military-more-latitude-to-use-offensive-cyber-tools-against-adversaries/2018/08/16/75f7a100-a160-11e8-8e87- c869fe70a721_story.html. Representative Jim Langevin (D-RI) along with three other colleagues wrote a letter to the president in February of 2019 requesting a copy of the directive. It is accessible at Langevin’s website: https://langevin.house.gov/sites/langevin.house.gov/files/documents/NSPM_13_signed.pdf. 252

decades, civilian policymakers preferred that military cyber capabilities serve defensive purposes, and for much of that time civilians further constrained military activities to DoD networks. Over those two decades, key civilian officials came into government with their own expertise in computing and computer networking. Moreover, the private sector played a large role in developing and maintaining what we now called cyberspace as well as the tools used within it. For years, the military lagged even another part of DoD—the NSA—when it came to developing ideas about the uses of cyber tools beyond tactical communications on the battlefield.

Meanwhile, civilians in the intelligence community developed an opposition to computer network attack because it could compromise ongoing data-gathering efforts—the intelligence community saw weaponization of cyber as counterproductive to intelligence missions.

But after the cyber operations in Kosovo at the end of the 1990s, the military began to promote the merits of offensive uses of cyber capabilities, first in support of physical operations, and then as a complement to deterrence and defense activities in cyberspace. The Joint Staff and the services, as well as uniformed Directors of the NSA were centers of activity in this regard.

Once Cyber Command was established in 2009, it took on the role of the military’s lead advocate for using cyber capabilities offensively. Given reporting about cyber campaigns with the goal of destroying or holding at risk physical assets and networks in other countries, it can be inferred that civilian officials were also increasingly embracing offensive uses of cyber tools. As of this writing, civilian officials in the Trump administration are expressing support for increasingly using cyber capabilities for harm and the threat of harm in preemptive ways to maintain American dominance in cyberspace or to punish or deter actions outside cyberspace. A civil-military congruence of preferences, therefore, appears to be emerging, with civilian preferences changing to more closely resemble Cyber Command’s.

253

There is room for doubt, however, that civilian officials adopted the military’s preferences. It is also reasonable to infer that threats in cyberspace have prompted civilians to determine that cyberattacks cannot merely be defended against, but that attackers and their capabilities must be proactively disabled. This is Cyber Command’s essential argument right now, but it is based on empirical observations that civilians are equally capable of making—and indeed, that some civilians helped to make as far back as the 1990s. Indeed, the role of civil- military coalitions in changing cyber policy, and the role of civil-military dialogue in building preferences for the uses of cyber capabilities, is evident in the case. The exchange between OSD directives and Joint Staff doctrine in the 1990s, with each camp taking the other’s concepts and refining and expanding upon them, shows an iterative development of preferences for the uses of cyber capabilities that makes disentangling “civilian” and “military” preferences at senior levels almost impossible.

There are also at least two reasons that offensive cyber capabilities may be politically attractive to civilians. First, the offensive use of cyber capabilities may be preferable not so much to their defensive uses but to other types of offensive tools available to implement national policy. Offensive actions in cyberspace involve things that might be considered violations of sovereignty, destruction of national assets, or generally threatening, but in ways that are far less escalatory than committing such acts in the “real” world.629 Thus, using cyber capabilities offensively may be politically desirable to civilians who want to elevate great power competition without succumbing to accusations of being needlessly provocative, starting another war, or risking casualties. Second, the change in civilian preferences accompanied a change in the party

629 For more on the limited-to-ambiguous escalatory potential of cyber-attacks, see Jacquelyn Schneider, “Cyber and Crisis Escalation: Insights from Wargaming,” Naval War College, n.d., https://pacs.einaudi.cornell.edu/sites/pacs/files/Schneider.Cyber%20and%20Crisis%20Escalation%20Insights%20fr om%20Wargaming%20Schneider%20for%20Cornell.10-12-17.pdf. 254

controlling the Executive branch, as well as a shift to a particular administration that eschews military restraint. It could be that officials in the Trump administration are generally more disposed to using cyber capabilities for offensive purposes, and the cyber case could present the kind of convergence of independent views described in the civil-military relations literature.

The evidence here does demonstrate an emerging congruence between civilian and military preferences, with some evidence that military groups coalesced around, and began to advocate, offensive uses for cyber capabilities earlier than civilians. Civilians have maintained access to information and to alternative sources of information, although true expertise on cyber in both the Executive and Legislative branches of the government is still not widespread, was always housed largely at the NSA, and is being grown at Cyber Command on a daily basis.

What the case does show is that civilians generate preferences more independently from the military the more information they possess and the more proactive they are about seeking alternatives to military judgment. Although there is not enough evidence here to suggest whether military institutions managed to change civilian preferences—with or without information asymmetries—the cyber case does provide evidence that civilian preferences for the uses of a capability can develop independently from the military when civilians have their own sources of information.

One more important aspect of the case evidence, in fact, reinforces the likelihood that information asymmetries in civilians’ favor allowed them to maintain independent preferences: for much of the case, the military services were the ones who were inattentive relative to civilians. Moreover, the military actors who did pay attention were generally not forming preferences in their service capacity, but in their joint capacity. Especially during the realization phase, it was the Joint Staff that wrestled most with cyber issues, while the services (with the

255

exception of parts of the Air Force) were largely disengaged. Because development of cyber capabilities was dominated by DARPA and the NSA, and possibly because such capabilities did not threaten core services missions or identities, what remained were doctrinal considerations at the joint level—after all, cyber threats were a blanket problem across all DoD and U.S. government networks. At the same time, civilians were drawing clear boundaries around the military’s cyber activities, limiting DoD cyber actions to defense of its own data and communications. Military institutions felt neither threatened by cyber issues nor particularly compelled to take them on until Russia and China began to demonstrate their own interests in mixing cyber tools into military operations.

Military institutional actors were not advocating strong views with civilian leaders.

Although successive directors of the NSA played influential roles, their influence was not systematic or widespread. Civilians were free therefore of organized military advocacy and framing—the military did not wield information in the same ways it did in the other two cases.

But the creation of Cyber Command increased military attention on cyber capabilities and their applications. The command is forcefully communicating its own evaluations of the cyber threat environment and the military’s role in that environment to policymakers and legislators. At the same time, a correlating change in civilian views to adopt military preferences about offensive uses of cyber suggests that the change in the information balance—especially in terms of attention—plays a key role in shaping civilian preferences.

256

CHAPTER 5

CONCLUSION

In tracing the process of civilian preference formation over the uses of special operations forces, unmanned aerial vehicles, and cyber capabilities, I found that civilians frequently adopted military preferences. At many points in these cases, military preferences did not prevail over civilian preferences; military preferences shaped civilian preferences in the first place. The most consistent explanation across the three cases of both civilian adoption of military preferences and civilian independence from the military in forming preferences was the level of information civilians possessed relative to the military.

I began by identifying congruences between civilian and military preferences across the three cases, finding the military often had developed a particular preference prior to civilians expressing that same preference. I then traced the development of civilian preferences to determine whether the military, and in particular the military’s command of information, generated each convergence. I found that the military, and in particular the information the military wielded, frequently did shape civilian preferences.

Consistent with my hypothesis, information asymmetries affected whether civilians adopted military preferences or not. But the cases suggested a refinement of the hypothesis to focus on how civilians allocated their attention. Although expertise and access to information, including alternatives to the military for sources of information, did play important roles, civilian attention generated civilian expertise, spurred searches for views alternative to the military’s, and correlated with civilian preferences that were independent from the military. Civilians were not powerless to the persuasion or informational dominance of military groups and their preferences; they could exercise agency by allocating attention.

257

What made civilians pay more or less attention to an emerging military capability? There were at least three reasons. First were the intentional reasons. Civilians sometimes chose not to allocate attention to capabilities at all, or to pay more attention to aspects of them other than their strategic utility. Civilians’ choice to direct their attention elsewhere was typically a result of their own organizational and political interests. Civilians’ interests could increase their deference to parts of the military or could be a condition that facilitated civil-military alliance-building, which also resulted in civilians adopting the preferences of their military allies.

The information and communications costs of bureaucracy for politicians and the difficulties of monitoring the implementation of preferences are well-explored elsewhere, including in principal-agent models that have been applied to civil-military relations. But the under-explored consequence for civilian control of the military is that civilians often cede their preference formation to the military.

In this chapter I will review the major findings of the cases, examine the strength of information asymmetries as an explanation for the congruence between civilian and military preferences, and explore alternative explanations for the sources of civilian preferences and why they either do not outperform or complement my hypothesis. I conclude with a discussion of the implications of my findings for the scholarly literature on civil-military relations and for the practice of civilian control of the military, as well as an agenda for future research.

Major Findings

In this section I summarize and synthesize the findings from the three empirical case studies. Overall, I found that civilians often adopted military preferences for the uses of emerging military capabilities but did not always do so. Consistent with my hypothesis, the

258

difference in whether civilians adopted military preferences or not was the level of information civilians possessed about capabilities relative to the military. In particular, the level and focus of civilian attention determined whether civilians simply adopted a military preference in order to focus on another aspect of a capability or on other matters entirely. I also found that preferences often were not held exclusively by “civilian” and “military” groups but instead by coalitions or alliances comprised of a mix of civilian and military actors. These coalitions were often an outgrowth of long-standing intra-military disputes over the uses of capabilities, and sites where civilians adopted military preferences.

Civilians Adopting Military Preferences

Two of the three cases revealed a strong pattern of congruence between military and civilian preferences, while the third case suggested a developing congruence. In the SOF and

UAV cases, there were consistent patterns of civilians adopting military preferences. Over the three periods in the SOF case, civilians adopted either the services’ preference to keep SOF as

“adjuncts” to the conventional force or the SOF community’s preference to keep SOF for unconventional missions separate from those performed by the general-purpose force. The pattern of civilian adoption of military preferences was also evident because the dominant preference among the SOF community shifted over time from low-intensity conflict to counterterrorism missions, and civilians’ preferences shifted thereafter to converge with SOF preferences. Civilians came to share, in particular, JSOC’s preference that special operators become the lead counterterrorism force and that counterterrorism become synonymous with direct action missions despite the availability of alternative capabilities among both military and paramilitary (CIA) institutions.

259

In the UAV case, the services arrived at a general agreement in the 1950s and 1960s that

UAVs should be used for reconnaissance purposes. Civilian policymakers generally adopted that preference until the first years of the twenty-first century, and thereafter struggled to convince both the services and fellow civilians to broaden the uses of UAVs. In particular during the wars in Afghanistan and Iraq warfighters demanded UAVs for ISR purposes and civilians largely deferred to them. By the end of the case, civilians had achieved some measure of independence from the enduring military preference to use UAVs for reconnaissance and had begun to advocate more consistently and forcefully that UAVs be used in peer-to-peer combat applications.

It is unclear as of this writing whether civilians at DoD and the White House are adopting the military preference to use cyber capabilities offensively or if civilian preferences are converging with the military’s because the threat environment has inspired both parties to draw the same conclusions (more below). Regardless, the evidence suggests that civilian preferences are converging with the military’s—or, more precisely, Cyber Command’s.

Coalitions between civilians and the military were a major site of preference transmission. Such coalitions formed because of a divide in the military community itself, with each side recruiting civilians to defend their cause. The so-called SOF Liberation Front was made of active duty and retired special operations officers allied with civilian officials. The officers had persuaded these civilians to share their preference to resource SOF by making arguments about the utility of SOF in the burgeoning strategic environment. Meanwhile, service leaders and the Chairman of the Joint Chiefs of Staff allied with senior DoD civilians to advance an opposing preference to keep the SOF organizational status quo. Later in the case, JSOC aligned itself with the Secretary of Defense to override SOCOM and service resistance to putting

260

SOF and direct action in the lead of the War on the Terror. In such coalitions, civilians adopted military preferences and adjudicated among preferences held by military sub-groups.

Civil-military coalitions played a key role in the UAV case as well. First, civil-military coalitions between the Joint Staff and interested service communities on the one hand and civilians in the Office of the Under Secretary for Acquisition on the other were what finally helped UAV development overcome internal DoD resistance in a lasting way. Those coalitions married a civilian desire for investment in advanced technologies and the military preference to focus UAVs on reconnaissance. Later, as warfighters demanded UAVs for ISR, most civilian leaders aligned with their preferences. By then, the services had given up on abolishing UAVs and had adopted the preference for reconnaissance uses; now they opposed a vanguard of civilians who had pressed for combat UAVs as part of the transformation agenda. A little over ten years later, a subset of civilians attempted to recruit military mavericks, much as Posen would expect, to advance combat uses of UAVs as part of the Third Offset/Defense Innovation vision.

Civil-military coalitions formed to do battle over preferences rooted in military parochial interests and disagreements. The coalitions that formed over the purposes of SOF and UAVs were attempting to maintain or overturn the services’ status quo preferences. Although civilians did sometimes maintain their own parochial preferences over capabilities, those preferences were often not about the strategic uses of capabilities. Civilians thus tended to adopt military preferences for SOF and UAVs’ uses because their attention was on other aspects or implications of those capabilities. And in any case, civilians relied on military consent to translate their preferences into action. The purpose of civil-military coalitions initiated by the military was to overcome military opponents, and the purpose of civil-military coalitions created by civilians

261

was also to overcome military opponents. In both situations, the military and its preferences drove debate; civilians and their preferences were often hitchhikers. Civilians wound up adopting military preferences for the uses of military capabilities for the sake of convenience or because they seemed to have little choice otherwise.

Civilian Independence from Military Preferences

Civilians did not always adopt military preferences. They maintained preferences about emerging military capabilities’ development, costs, and other attributes rooted in their own organizational interests and even developed preferences for the uses of capabilities, both independent of military preferences. In the case of UAVs, civilians developed an independent preference for their uses after having adopted and then shared for decades the majority military preferences to use UAVs for reconnaissance. But they had also long held preferences regarding

UAVs’ costs and the efficiency of their development that were derived from congressional and acquisition executives’ interests in defense budgeting, industry health, and constituent interests.

In the cyber case, civilians maintained a preference for defensive uses of cyber capabilities from the beginning of the case, and only appeared to adopt or share military preferences for offensive uses toward the case’s end. In the SOF case, civilians never stopped deriving their preferences for SOF’s uses from military groups, but both the Clinton and George W. Bush administrations began with alternative uses of SOF in mind, if not strongly in mind, and the Clinton administration based its preferences about SOF at least as much on their political experiences of

SOF operations in Somalia, Haiti, and the Balkans as on factors specific to the SOF community’s own preferences about SOF capabilities. Civilians, then, demonstrated that they are not so constrained by military institutions and their preferences that they must adopt them.

262

The room for civilians and the military to hold complementary preferences is especially important. Complementary preferences refer to civilians and the military having and focusing on preferences for different aspects of an emerging capability. That civilians and the military might focus on different aspects or consequences of a military capability sheds light on the conjecture that civilians and the military have “irreducible” differences. They may not have differences over a capability but different interests in and thus different preferences regarding it. These differences are not the same as incompatibilities, and in fact civilian and military actors often looked for ways to make them compatible. For much of the UAV case, civilians did not care what the services used UAVs for as much as they cared about their costs and whether they could be developed jointly—which, really, they cared about because they imagined it would bring down costs. When civilians intervened in UAV development, they did so because of runaway spending, with the billion-dollar Aquila being the most prominent example. They did not intervene because they made a judgment about the Aquila’s imagined operational and strategic uses. When UAVs were developed in ways that did not challenge civilian parochial interests or even offset them—expensive perhaps but demanded by highly credible warfighters for the war effort—then civilians were able to adopt military preferences even as they maintained their own.

It was also a window into what civilians chose to pay attention to and what they did not, a finding explored more below.

Explanatory Power of Information Asymmetries

Did information play a systematic role in civilians’ adoption of military preferences, or in their rarer ability to maintain preferences independent from the military? By process tracing the formation of civilian preferences for the uses of emerging military capabilities, I found

263

substantial evidence that the level of information civilians had relative to military groups generally determined whether civilians would adopt military preferences as their own. Most important, I found that the attention aspect of information played the greatest role in whether civilians derived their preferences from the military.

As defined in the introduction, the concept of information in the civil-military relationship has three key dimensions: expertise (which connotes experience as well as technical knowledge), access to information (related but not limited to the idea of private information), and attention. Below, I explain how each dimension of information affected civilian preference formation in the context of the civil-military relationship, and why attention had the greatest explanatory power.

Expertise

Expertise was the result of, and played an important role in, civilian education. When expertise over capabilities was housed mainly in military channels, civilians were naturally dependent on military groups for their foundational learning. This was most evident in the SOF case, but also came up in the first several decades of the UAV case. Developing expertise about special operations was very difficult to do outside of the SOF community, as evidenced by civilian elites stating in interviews that much of their initial education about SOF capabilities came from special operations officers. It was more possible for civilian officials to learn about

UAV capabilities from the private sector, although such information pertained more to the systems’ technical specifications and less to their potential roles in warfare. In both the SOF and

UAV case, civilians were highly dependent on military expertise and judgment about the capabilities’ strategic and operational utility.

264

Expertise in cyber matters was the most organically civilian knowledge among the three cases as civilian academic institutions and the private sector developed, defined, and dominated computing and computer science from the start. This expertise was not only in how hardware, software, and computer networking worked, but also in the creation of cyberspace and its implications for communications, markets, and the global economy, including the vulnerabilities created by reliance on cyberspace.

Finally, I found evidence for an expertise-related phenomenon discussed in the introduction: the unique legitimacy conferred on expertise developed as a result of participating in combat. In the SOF case and even more in the UAV case I found that arguments supporting a capability’s combat necessity made by those deemed “warfighters” had a particularly persuasive power with civilian leaders. In the UAV case, many civilians used combat-derived expertise almost like a heuristic, assuming the superiority of the warfighter’s evaluation to others’ interpretation of events. Thus, conditions of wartime actually tended to make civilian preferences more dependent on the military, in part because civilians sought information directly from those fighting in combat. Once again, the cyber case is useful because it removed combat as both an actual and possible condition for generating expertise, making debate over the military uses of a capability take place without the authoritative input of those with first-hand knowledge of combat.

The power of combat experience to increase the credibility of military expertise suggests that civilians are more able to question and challenge military-derived information if they themselves have served in uniform under combat conditions. If they did not have veteran status, civilians clearly had an incentive to use veterans as proxies, bolstering their arguments by aligning with credible spokespeople. The prevalence of veterans in the ASD(SOLIC) role, Noel

265

Koch’s use of his retired-senior-officer-filled Special Operations Policy Advisory Group, the recruitment of “military mavericks,” and civilians joining civil-military coalitions in order to win arguments over the development and uses of the capabilities explored here all build a case that civilians need to get military experience on their side if it is not on their CV. Whether the superiority of combat (or even just service experience) to other forms of expertise is rational or fair is immaterial to its social capital. And the utility of expertise based on experience in the military was critical in the SOF and UAV cases.

Civilian expertise thus ranged from highly dependent on the military in the SOF case to variably dependent on the military in the UAV case to nearly independent in the cyber case. At the same time, civilians showed the least independence in their preferences when it came to SOF, eventually developed independence in preferences over the purposes of UAV capabilities and were almost wholly independent form the military in forming preferences over cyber capabilities until, perhaps, the final years of that case.

Access

Access to information contains two elements: intent and alternatives. Intent pertains to whether access was purposely limited by those in possession of information. Did the military have private information they refused to give to civilians, in whole or in critical part?

Alternatives means that those who seek information can find it via multiple sources. Did civilians have readily available sources of information other than the military? In other words, was the information marketplace diverse or limited to military sources?

As with expertise, the cases showed variance in access. In the SOF case, civilians had low access to information except when military sub-groups provided it. This was partly a

266

function of SOF capabilities being based in doctrine and training more than technology, and so the personnel in the community itself were an almost wholly isolated source of information. One needed access to special operators (or access to others who had access to them) to learn about special operations. This allowed special operators to pick and choose the information they provided to civilians, including after internal battles over what to emphasize. The lobbying done by special operators to generate congressional support for SOCOM and to get Secretary of

Defense Rumsfeld to see SOF as a counterterrorism tool revealed the SOF community’s control over information. UAVs, because of their technological features, meant information was more widely dispersed among military and private sector sources.

Classification is one formalized means of keeping information private, but this was only evidently at work in the cyber case, and its effect on the civil-military information balance and on civilian preference formation was unclear. The case suggests that compartmenting information can stunt both the spread of and debates among military preferences and limit intra- military information sharing as much as it limits civil-military information sharing. Classification may also have mattered in the UAV case, as many of the more advanced experiments in unmanned aircraft were highly classified. But these were generally used for intelligence purposes only, and by the time UAVs were being routinely discussed for military uses information about them was available to a broader set of stakeholders inside government. Finally, there is little revealed about classification in the evidence compiled for the SOF case. That a separate process for approvals of counterterrorism missions existed at the NSC (known as the Counterterrorism

Security Group) is widely believed to have limited access to information about ongoing operations.630 Whether such limited access helped shape civilian preferences or prevented

630 Christopher Lamb and Erin Staine-Pyne, 9/11, Counterterrorism and the Senior Interagency Strategy Team: Interagency Small Group Performance in Strategy Formulation and Implementation (Fort Leavenworth, KS: The 267

civilians from forming preferences independent from the military is not clear, although the access limitations may have helped sustain the path dependency of the preference to use SOF for

CT once it was in place.

Alternative sources of information also varied among the cases and was partly a function of whether and how much information was privately held by the military. But alternative sources of information were also determined by the degree to which the origins of the capability lay outside the military. SOF capabilities, again, were proprietary to SOF and thus alternative sources of information were hard to come by. To some extent, the intelligence community could act as another place for civilians to learn about SOF capabilities, especially as the intelligence and special operations communities worked together more closely after 9/11. In particular, the

CIA’s role as an alternative source of information exposes the opportunity for bureaucratic competition to provide civilians with additional sources of information, but also that the information provided is likely filtered through the interests of the competing organization.631

In fact, operational performance was an important source of information about SOF for civilians overall, including information about whether or not SOF capabilities were really as useful for the purposes SOF claimed they were. UAVs also revealed their strategic and operational utility and risks during deployments, but the fact that much of their development happened in private companies gave civilians managing acquisitions and members of Congress

Alfred D. Simons Center for Interagency Cooperation, 2014), http://thesimonscenter.org/wp- content/uploads/2014/04/IAS-003-APRIL14.pdf. 631 An example that is outside the scope of the SOF case study but relevant to the argument is the CIA’s evaluations of the SOF-led counterterrorism operations in Yemen. Members of Congress came to believe that the CIA’s missions reflected “meticulous care,” whereas the SOF operations did not. Ali Watkins, “Two Years Later, White House Still Hitting Roadblocks in Effort to Move Drone Program Out of CIA Control,” Huffington Post, March 10, 2015, https://www.huffpost.com/entry/obama-drone-cia_n_6817858?guccounter=1. A DoD official commented to me that CIA briefings, “almost certainly helped folks on the Hill along to that conclusion.” Anonymous DoD Official, text message exchange with the author, December 3, 2019. 268

an important alternative source of information.632 Cyber capabilities of course were largely developed in the academic and private sectors, as well as at the NSA and to a far more limited extent, DHS. Civilians thus had a wealth of alternative sources of information about cyber capabilities. However, as mentioned above, classification in the cyber realm was an oft-cited reason for some information, especially about actual government activities, to be extremely limited. Alternative sources of information did not always match the military or other parts of government in topic, scope, or quality. That members of Congress are trying to get access even to the Trump administration’s policy about the offensive uses of cyber capabilities offers a window into the unequal nature of information and the importance of access not just to information generally, but some information in particular, in order for civilians to form preferences.

As important, the non-military alternative sources of information that civilians had were often at least partly dependent on military groups. The Congressional Budget Office and the

General Accounting Office each routinely conducted interviews of military officials as part of their data-gathering. Both government and independent commissions also relied on information from military services and commands. This was true of even the earliest studies on the potential battlefield uses of UAVs and congressional investigations of special operations. The press could both reveal private information and serve as a source, but again, they were often depending on what the military had to say about itself and its capabilities. This revealed that the military maintained enough private information that it remained a vital source for civilian education.

632 Companies, of course, have their own interests and therefore motivations to present the information they possess in certain ways. Such motivation can be almost as informed by military interests as military information. The point here is simply that such information comes from a non-military source. However, further study of the degree to which companies adopt military preferences would be revealing. 269

Attention

All three cases revealed that attention was the linchpin of information, and therefore of civilian preference formation. When civilians were attentive, as during the cyber case, they developed expertise and ensured they could access data about emerging capabilities. They also understood the history of a capability: which organizations had championed and opposed its development, how it had performed in tests, exercises, and operations, and how theories of its strategic and operational uses had unfolded over time. When civilians were inattentive, as in the early part of the SOF case and to a moderate extent thereafter until 9/11, their resulting low expertise and unchallenged (or, in the case of SOF, military managed) access to information suppressed independent preference formation and led them to adopt military preferences.

Civilians were only partially attentive to UAVs’ military uses over their long development period, and by the time a wide range of policymakers and legislators payed close attention to

UAVs, the services had a strong commitment to using UAVs for reconnaissance, and warfighters echoed that commitment in a context of great strategic and political urgency.

Attention included the element of intent, to an even more important degree than access.

The attention civilians paid to emerging military capabilities was almost always a choice, although it likely sometimes seemed to civilians that constraints such as finite time and other urgent matters forced them to outsource their attention to many matters. The necessary division of labor between principals and agents of any variety means that a certain amount of inattentiveness on the principal’s part is built into the relationship. As discussed in the first chapter, civilians may choose from different models of monitoring, but those models assume a preference exists up front with which civilians are monitoring military compliance. But if civilians do not begin with a preference—in this case, a perfectly reasonable expectation given

270

the still-maturing quality of emerging capabilities—then those paying close attention will always develop more expertise and access than those who are not.

But civilians also were inattentive because their resources in terms of time and personnel were finite. Recall that information here is a relative measure. Even when civilians paid some attention, military groups frequently allocated more. The civilian span of control, even deep within the DoD bureaucracy, is typically wider than their military counterparts, giving them less time to focus on each issue. This problem only worsens with seniority, because the higher up an official is in the civilian chain of command, the broader her portfolio. Secretaries of Defense and

Presidents simply cannot pay full and persistent attention to every issue in the vast military bureaucracies. They must rely on subordinates to inform their preferences. Those subordinates include civilian and military advisors. Moreover, civilian and military staffs are often greatly imbalanced.633 With more capacity comes more ability to generate as well as acquire information and to disseminate it. Exacerbating this issue, I found that retired military officers often continued to play a role in civilian preference formation, advancing military institutional interests by using or channeling military expertise and access to military information that ensured military preferences were always part of the learning process for civilian officials.

However, the imbalance in capacity between civilian and military organizations is not an act of

God; it is an act of Congress and of executive implementation. Once again, civilians have chosen where to allocate resources. They have not always chosen to allocate them to civilian organizations.

633 For example, as detailed in chapter four, Cyber Command now has 6,000 military personnel, and the commander also has access to several thousand experts in the NSA. In contrast, the Cyber Policy office in OSD consists of some 2-3 dozen personnel. 271

The attention aspect of information also highlights the importance of timing, in at least two senses: the delay between first attention paid and the development of access and expertise sufficient to form an independent preference, and the immediate circumstances surrounding the moment that attention is first or newly paid that informs interpretation of the fit between capabilities and strategy. There is a clear pattern in the cases between when civilians first allocated attention to emerging capabilities and when they developed preferences independent from the military. In the UAV case, civilians were prompted by the emergence of the Predator drone and its uses in the Balkans and Afghanistan, concurrent with the emergence of communications technologies and ideas about transformation, to begin examining UAVs’ broader uses. Although it took over a decade to emerge as a strong and persistent preference, sustained civilian attention to UAVs eventually resulted in the development of a civilian preference independent of the military. In the cyber case, civilian attention was far greater than the military’s for years, and civilian preferences for cyber uses were therefore independent of the military from the beginning.

The SOF case is the only one where access and attention appeared to be nearly equally important, although in that case where and how civilians allocated attention to SOF capabilities did matter. The Clinton administration’s negative experience with kinetic uses of SOF in

Mogadishu and positive experience with foreign internal defense uses of SOF in Haiti did influence the administration’s preference for emphasizing low intensity conflict uses rather than for counterterrorism uses.

In short, where civilians paid close and sustained attention to an emerging capability, they were more likely to develop their own expertise, seek alternative sources of information,

272

and develop preferences for the uses of capabilities apart from the preferences held and advanced by military groups. Where civilians were inattentive, they adopted military preferences.

This finding is somewhat complicated, however by the evidence in the UAV case that it also mattered which civilians were paying attention, because their functions directed them to care about different implications of capabilities development. In the UAV case, civilians in the acquisition and budgeting realm were the ones to pay more sustained attention to UAVs than civilians in broader policymaking and oversight positions. This left consideration of UAVs’ strategic utility up to the services and joint military organizations. Moreover, those same civilians interested in adopting the technology of UAVs actually used military operational and strategic demands for them as leverage to shepherd them through program development. But it demonstrated the possible compatibility of different civilian and military preferences regarding the same capability. It showed how civilians came to adopt military preferences not because they were not paying any attention, but because they were paying attention to different aspects of

UAVs.

Civilian Agency

Research finding that civilians frequently adopt military preferences risks inferring that civilians are passive participants in military policymaking. The purpose of the following discussion is to refute such suspicions. The explanatory power of the level of information civilians had relative to the military, especially the power of attention, illustrates that civilians had and exercised agency over their own preferences. When civilians chose to gather more information, they tended to develop and then hold preferences that were different from military preferences. Moreover, they were able to identify alternatives to the military: alternative sources

273

of information; alternative uses of the capabilities studied here; and alternative organizations to perform the same missions or functions that military sub-groups preferred to perform with the capabilities in question.

In some of the cases (and some individual periods of the cases), such alternatives were apparent to civilians. Sometimes there was little evidence that civilians saw or considered alternatives, but they were logically available. In all three cases, civilians had alternative organizations to turn to. In the case of cyber, civilians did turn to alternative agencies to the military because civilians’ preference to use cyber tools defensively prompted them to lean on civilian agencies such as the FBI and DHS more than DoD. And it may be the case that the NSA or other intelligence agencies could handle offensive cyber operations as well as the military.

Throughout much of the UAV case civilians could and did use the CIA to operate them for the very reconnaissance purposes the military claimed them for. And for special operations’ counterterrorism mission, both the CIA and the general-purpose force were viable alternatives. In none of these cases, therefore, was the performance of the mission civilians eventually adopted

(or, in the case of cyber, appeared to be in the course of adopting) inevitably married to the military institution wielding the capabilities examined. SOF did not have to wind up as the counterterrorism force, UAVs were not the only way to conduct to reconnaissance, and cyber capabilities have been wielded by non-military agencies for much of their history.

Alternative uses of the capabilities also abounded. In the cyber case, offensive uses were the main alternatives. For UAVs, everything from replacing traditionally manned aircraft such as bombers, strike, and lift planes to new concepts of operations associated with transformation and the Third Offset, such as swarming and manned-unmanned (droid) teaming, were all alternatives.

SOF could be “adjuncts” to the conventional force, conduct civil affairs missions, hunt war

274

criminals and weapons of mass destruction, or do a variety of other things. In every case, at least some civilians were aware of these alternatives.

However, civilians’ agency was often constrained by virtue of the ease or difficulty of gathering information, either from the military itself or via alternative sources as discussed above. Moreover, who articulated the alternatives is germane to this inquiry, because the point is to identify whether and how dependent on military preferences civilians were. The tendency for civilians to defer paying attention often meant that by the time they did, intra-military positions had been staked out and many information and preference alternatives were within military circles themselves. In such cases, especially for the SOF and (most of the) UAV case, civilians found themselves deciding between preferences articulated by military sub-groups using information provided by those groups.

Civilians had choices for which preferences they adopted. Why did they not more frequently depart from those espoused by the military? They were often constrained by a kind of first-mover advantage that ceded the power to define issues to whomever staked out positions first. This is most evident in the SOF and UAV cases. In the former, civilian attention lagged military attention by several years, and then turned away from SOF again after the Battle for

Mogadishu. This meant that the divides between SOF and the general-purpose force and then between SOCOM and JSOC defined preferences before civilians applied attention to SOF capabilities systematically at the strategic level. Where UAVs were concerned, civilians tended to pay attention to technological innovation and budgetary matters but less to strategic ones, leaving military communities to debate the purposes (and limits) of UAVs with only episodic civilian intervention for decades.

275

Which civilians paid attention also mattered. The contrast between the UAV case and the cyber case is of interest here because for UAVs, the civilians who did pay attention from the earliest days did so with technology development or budgetary interests in mind rather than strategic ones, while the services focused on all three from the beginning. Meanwhile, cyber issues emerged from civilian government agencies, academia, and the private sector, and informed civilians at senior policy levels concerned with the strategic uses of capabilities, making the military the entity that had to catch up to already-formed civilian preferences. The opportunity today for Cyber Command is not to form civilian preferences in the first place but to change them—at least, unless and until Cyber Command is able to influence a new, uninformed generation of civilian political leaders with more success than other informed civilians.

The attention lag also meant in the UAV case that even when civilians finally developed preferences independent of the military, they struggled to defeat entrenched military preferences.

This is evident in the difficulty civilians had implementing the ideas about and uses of UAVs in the Third Offset. The cyber case shows how the opposite is true when civilians develop their own preferences first: the military must work hard to overcome civilian biases against offensive uses of cyber capabilities—and the path dependence of existing law, directives, and agency jurisdictions marshalled against the military preference.

The most critical question about civilian agency then is why they so frequently chose not to pay attention to capabilities’ and their potential strategic applications until well after the military community did? One answer, provided by Huntington (1957) and much-beloved by senior military officers, is that the civil-military division of labor is such that developing the means of military action is the job of the armed forces, and civilians’ role is simply to supply the ends. This normative expectation, however, has never been exercised in practice, in part because

276

a substantial part of civilian control of the military happens through the Congress and its role in authorizing the military budget—to include capabilities-development.

The next most obvious answer to why civilians often deferred attention has to do with civilians’ incentive structures. Again, the civilian span of control necessitates prioritization, and capabilities still in development often do not outcompete more urgent or politically sensitive matters, or they only do so sporadically. This is borne out in the cases, where crises and operations called civilians’ attention to SOF, UAVs, and cyber, spurring them to learn and even make changes based on what they learned. From the U2 crash in the Soviet Union to the casualties in Grenada to the Eligible Receiver exercise, crises had a tendency to focus the civilian mind. But crises are not systematic, can soon be subsumed in the flood of other events and issues, and can bias civilians toward thinking of a capability in response to the particular events around the crisis. They can also make civilian leaders unusually open to existing interpretations, proposals, and solutions. Again, to the extent that the military had already prepared such interpretations, proposals, and solutions (as in the SOF and UAV cases), they were able to shape civilian preferences. To the extent that civilians had already built up expertise relative to the military, the military had less influence and civilians exercised more agency (as in the cyber case).

Alternative Explanations

As discussed in the introduction, much of the existing civil-military relations literature assumes civilian preferences are derived from sources other than the military. Some authors contend that civilians are most affected by external threats to the state, whereas others argue that civilians have their own parochial motivations for their preferences.

277

Below, I explore the two major alternative explanations of civilian preferences. I discuss why external threats do not explain civilian preferences adequately, and how explanations based on organizational culture and interests complement my informational explanation.

External Threats

Although threats and crises played important roles in the formation of civilian preferences in all three cases, they were almost always mediated by other information sources, and frequently military sources. As Kier (1997) and others have argued, there is no direct determinative relationship between external threats and either civilian or military preferences.

Something must tell actors in government what threats mean and how to respond to them. The key was which institutions, and what balance or mix of institutions, civilians turned to.

Sometimes civilians sought out civilian sources of interpretation and meaning making, but more often they turned to the military.

Threats and crises often displayed civilian reliance on military expertise and judgment starkly, as when Deputy Secretary of Defense John Hamre turned to the Joint Staff and asked,

“who’s in charge?” or in the many commissions and reports on the lessons of such crises chaired by retired general and flag officers. In fact, threats and crises often formed conditions where civilians deferred to the military more, as when Secretary of Defense Rumsfeld “fell for” the

Joint Special Operations Command’s arguments about how SOF was suited to be the lead military element in the War on Terrorism, or when intelligence operations in Iraq spurred the marriage of direct action and cyber capabilities, or when warfighter advocacy for ISR uses of

UAVs overwhelmed nascent ideas about expanding the systems into conventional combat.

278

The cyber case was an exception. The threat of external (and internal, criminal) breach of

U.S. government networks, including DoD networks, inspired civilians’ focus on defensive cyber capabilities. However, that was also a case where the military was at both an informational and organizational disadvantage relative to civilians. Now that we are ten years into the lifespan of Cyber Command civilian strategic documents increasingly reflect the command’s interpretation of cyber threats and judgment about the appropriate uses of cyber capabilities in response.

In all three cases, wartime was actually a condition that prompted civilians to defer to military preferences more, and even to go so far as to adopt them while rejecting other related preferences. Civilians, having long ago adopted the military preference to use UAVs for reconnaissance, during the wars in Afghanistan and Iraq dropped their additional preference to minimize spending on the systems. In the SOF case, isolated operations prompted reevaluations of the level of investment in special operations capabilities led by SOF voices. And as the international system has experimented more, and more boldly, with cyberwar, military preferences for offensive uses of cyber capabilities have begun to compete with, although not yet supplant, civilian preferences to keep capabilities focused on defense.

However, actual wartime should not be conflated with all international threats, which often unfold over a long period of time. When it comes to long-term planning, civilians do rely more on non-military counsel. In the UAV case, both the preponderance of American technological might at the turn of the twenty-first century and concern over the rise of China some ten years later and the long shadow it cast over the future of U.S. hegemony in the world both prompted civilians to review how UAVs were being used. In the cyber case, it was civilians who first foresaw the potential risks of computer networking, a farsightedness that also prompted

279

them to focus on defensive cyber capabilities. Only SOF somewhat broke this pattern with special operators’ articulation of the rise of both terrorism and low intensity conflict rather than civilians. This was perhaps a function of SOF capabilities and the SOF community itself being one and the same, the need for SOF’s organizational survival inspiring long-term justifications for their services.

Moreover, civilians did use wartime and operational experience as a source of information about capabilities. Sometimes military actors’ interpretation of such operations framed information for civilians, as was the case with Senators Cohen and Nunn listening to retired Major General Dick Scholtes’ closed-door testimony about the invasion of Grenada. But at other times, such as after the battle of Mogadishu, civilians drew conclusions informed by their own parochial interests. However, it is important to consider that the risks of special operations revealed by the Black Hawk Down incident reinforced the bias of many service leaders against SOF, and it is likely that civilians in the Clinton administration heard some of these conventional force and service interpretations of the battle as well.

External threats thus had a mixed record in influencing civilian preferences for the uses of SOF, UAVs, and cyber capabilities, and there is little evidence they did so directly. Threats and crises prompted new thinking, but the course of that thinking was shaped by additional, internal factors. And almost always, at least to some extent, by military voices.

Organizational Culture and Interests

Perhaps, then, parochial motives really do account for civilian preferences. The immediate challenge to this alternative explanation is the empirical record here shows that military organizational culture and interests were frequently more influential than civilian

280

organizational interests. The tendency of even civil-military coalitions to revolve around debates over military preferences bears the influence of military organizations out. And as discussed in the preceding section, military organizations consistently interpreted the outside world for civilians, sometimes even at civilians’ behest.

But organizational interests do explain some of the preferences civilians held. In the

UAV case, civilian preferences to reduce the costs of UAVs and encourage joint development were rooted in the organizational interests of Congress and the acquisition workforce at DoD. In the cyber case, initial and enduring civilian dominance over the development of cyberspace and cyber capabilities meant much of the responsibility for cyber activities were assigned to civilian organizations whose organizational interests involved domestic rule of law and economic stability that lent themselves to cybersecurity rather than cyberwar.

However, there was no evidence in the SOF case that civilians ever developed a strong or sustained preference for the uses of SOF that did not originate with either the special operations community or the general-purpose force. This may have been, perversely enough, because in general, civilian interests in re-election and other organizational matter were not implicated one way or the other by uses of SOF—or, when they were, such as in the Iranian desert and the urban area of Somalia, the political implication was how not to use SOF, rather than how to use it.

The important insight here is that interests do not equal preferences. This explains the problem with Feaver’s assertion that “perfect overlap” between civilian and military preferences is not possible. First, I have shown empirically that civilian and military preferences often do overlap. Second, while it may be true that civilian and military interests may focus on different things, they may still result in the same preference. This is because a preference may serve

281

multiple interests at once, and because while interests are in ideal states of the world, preferences deal with options in real circumstances.

Implications of the Research

Does it matter that civilians often get their preferences for the uses of military capabilities from military institutions and actors? Neither the civil-military relations literature nor the scholarship on military innovation consider the answer to this question because they rarely consider its possibility. Rather, civil-military relations theory considers the normative consequences of military preferences prevailing over civilian preferences in times of disagreement, while military innovation studies claim that civilian preferences are, at worst, irrelevant to a process dominated by the military. I find in my cases a pattern of military cooptation of civilian preferences long before civil-military contestation ever occurs—and that such contestation itself is often between coalitions of civilian and military actors. Moreover, I find the military dominance some scholars perceive in military innovation processes is one of the mechanisms for military influence over civilian preferences.

It is important to note the range of empirical evidence presented here does not prove that civilians have a general tendency to adopt military preferences across a wide range of issues.

Moreover, those who are comfortable with the division of labor espoused by Huntington’s objective control model may not agree that civilian adoption of military preferences for the uses of military capabilities is problematic, even from a democratic perspective. And civilians might still retain control over the most important decisions, including whether to go to war.

But preferences for the uses of military capabilities have implications for how war is conducted and how national strategy and budgetary decisions are made. If it is military

282

institutions, wielding proprietary information, that determine which capabilities civilians think are most germane to strategy, affecting which military capabilities the country does and does not resource, then the military also has a great deal of influence over the range actions the country can take, and therefore the foreign policies it can pursue. Huntington (1961) himself observed that decisions over military capabilities can shape strategy. If civilians are only free to develop preferences for the uses of capabilities independent of military preferences under certain narrow conditions, then civilian control of the military is tenuous at best.

Implications for Practitioners

The clearest implication is the need for strong mechanisms that allow civilians and civilian agencies to allocate attention. However, that recommendation will not often be practical.

Civilian attention allocation, although a choice, is a choice made within the constraints of finite time and many competing demands. Urgency will always prompt some issues to crowd out others. This does not mean that officials cannot do better. But civilian practitioners can also use the knowledge produced by this study of the conditions under which they are likely to adopt military preferences—namely, when they have not been paying attention, have little expertise and little access to alternative information sources to the military, and in wartime. That knowledge can help them audit their own preferences for rational agreement with military preferences (what the literature calls convergence) or agreement generated by mere convenience or default.

In the course of such preference audits, civilian practitioners should especially consider the role played in their preference formation and in the resulting formation of policy by retired military personnel. To the extent that retired personnel continue to advance their former military

283

institution’s preferences, they should be understood as performing that role and amplifying the informational reach of a service or sub-group and its framing of issues.

These audits should be the most thorough for civilians serving in the Department of

Defense. The close collaboration of civilian and military officials within the Department, not to mention their bureaucratic incentives to forge unified positions when confronted with competition from other agencies or intervention from the White House and Congress, resulted in a preference formation process over capabilities that are civil-military in nature. It is therefore especially difficult for civilians in DoD to develop and maintain preferences that are independent from military preferences. Moreover, civilians outside DoD who engage with civilians inside

DoD should also be aware of this tendency.

Implications for Scholarship and Future Research Agenda

This study took observations about the role of information and expertise from the literature on bureaucracy and military innovation and used them to examine the formation of civilian preferences and the consequences for civilian control of the military. There are four major implications for the literature on civil-military relations as a result of this research and analysis. First, the results should undermine scholarly confidence in the assumption that civilian preferences are “irreducibly” different and independent from military institutions and actors in the U.S. national security system. Over decades of strategy-formulation, civilian leaders adopted military preferences for the uses of emerging military capabilities. They had several reasons to do so. Close examination of the time between the articulation of military preferences and when civilians expressed congruent preferences revealed that civilians often chose sides in ongoing military debates about the purposes of capabilities. Civilians sometimes held preferences

284

regarding force structure or joint or interagency collaboration even as they adopted military preferences for the strategic purpose of a capability. Civilians almost always had alternative preferences available, but under conditions of low information systematically adopted military preferences.

Second, and consequently but more importantly, this dissertation provides evidence that civilian control of the military in the United States is not as firm as scholars suppose. Although formal methods of control such as budgeting and legal authorities go unchallenged here, the ability to control military policy and what the country actually does and does not do with its armed forces are not processes that civilians reliably dominate. Civilians do not reliably dominate these processes because civilians do not systematically determine their own preferences for the uses of military capabilities. Instead, the process of civilian preference formation is permeated with military actors, information, framing, interests, and preferences.

Third, this work provides empirical evidence of the contention that preferences are not fixed according to external threats or organizational interests but can be socially constructed and reconstructed within the confines of the civil-military relationship. The active debate between organizations and between issue networks over the valid interpretation of the strategic environment and capabilities’ uses within that environment revealed the malleability of preferences to information and persuasion.

However, it also showed that civilian preferences were more malleable than military preferences. Military organizational inertia as reviewed in the introductory chapter was evident in all three cases, as was the value of military information, especially expertise derived from warfighting. The legitimacy of military expertise meant that civilians needed other military

285

experts to contest military views. Civilians had agency over their preferences but had to work harder than the military to gather the information critical to maintaining that agency.

The empirical record underpinning evaluations of the health of civil-military relations thus needs to be reexamined for a bias toward civilian independence from military preferences.

Instances of the military “working” according to civilian preferences should be examined to verify that the military was not simply complying with a preference that a military institution or institutions persuaded civilians to adopt.

Fourth, the very categories of “civilian” and “military” deserve review and clearer articulation in the civil-military relations literature for three reasons: to accommodate the role of civil-military coalitions in preference formation as well as policy making; to develop a more robust understanding of the roles played by retired career military personnel in policymaking; and to perceive the boundaries of warfare itself more accurately.

The civil-military coalitions highlighted here ranged from large pseudo-advocacy groups to pairings of powerful senior officials. These coalitions would then square off against other coalitions advocating for an alternative preference. Although I found that the division of civilian and military institutional identities was still useful and measurable, controversies over policy preferences frequently could not, themselves, be characterized on the whole as civil-military debates. In fact, they were often characterized as military-military debates. Being clear, therefore, about what is military and what is not is critical to the analysis of whose preferences prevailed.

The evidence presented in the cases also shows an extensive entanglement between civilians and the military because of the active role played by retired military personnel who continue to educate civilians and advocate for things in the interests of their former military

286

institution. In the cases, I frequently coded retired officers as military rather than civilian institutional actors, even though they had become civilians in a legal-formal sense. Many existing studies in civil-military relations might add to the bias toward civilian independence from military preferences by ignoring the ongoing military institutional affinity of retired military personnel.

Finally, a review and revision of how researchers conceptualize “civilians” and “the military” may assist in understanding how warfare itself is evolving and whether and how civil- military relations theory should evolve with it. The continued reliance in the literature on sharp sociological and functional distinctions between civilians and the military leaves researchers without the tools needed to analyze the militarization of civilian activities and the civilianization of the military. The exponential post-9/11 use of the CIA’s paramilitary and UAV capabilities for capture and kill missions alongside the increasing scope of SOF’s intelligence functions was just one aspect of this trend toward ambiguity. The cyber case revealed a realm where elements of the U.S. military are conducting something like operations against adversaries on a persistent basis within a “domain” largely built and maintained by civilian companies for non-military purposes. Fundamental questions about what makes a war and who is a warrior will be unanswered into the foreseeable future. Given the tendency revealed here for civilians to cede preference formation to military institutions, militarization of broader swathes of government activity is not impossible, and the necessity of articulating why the boundaries of civilian identity and activity are not natural but a normative good is increasingly urgent. Antiquated ideas about the solidity of civilian and military identities, institutions, and activities will make civil-military relations theories irrelevant. These categories should be updated with new theorizing and recent empirical evidence of broad scope.

287

The first item on a future research agenda is to expand the number of cases of emerging military capabilities to create a more robust empirical check on the hypothesis. The best starting point would be a re-look at capabilities (more) associated with conventional warfare, such as submarine warfare, stealth capabilities, or missile defenses. Thereafter, an expansion to a wider range of defense issues, including traditional use of force and crisis decision making, would deepen understanding of the military’s influence over civilian preferences and the role the military plays in civilians’ information resources.

Next, given that the research highlighted the particular importance of attention to civilian preference formation, further study of that element of information would deepen understanding of the civil-military dynamics involved in civilian preferences. Among the questions that need further empirical study are: Why did civilians defer attention to emerging military capabilities?

Was deferred attention at all based on preconceived ideas about the capabilities? When civilians chose to minimize the attention they paid to emerging capabilities was it because they thought forming preferences for those things was unnecessary at the time? Did civilians believe military groups whose preferences civilians adopted were credible, trustworthy, or exercising rightful autonomy over tactical and operational military matters?

It would also be valuable to probe more at the different types of positions civilians occupy, the proximity of those roles to military institutions, and the susceptibility of different civilian functions to adopting military preferences. For example, a thorough study of service secretaries and their pattern of adopting their service’s preferences versus adopting administration objectives would shed further light on the conditions for and mechanisms of civilian preference formation.

288

The frequency with which civil-military coalitions were key actors in the three cases here

(although to a far less extent in the cyber case) also deserves further scrutiny and theorizing. For actors to somewhat abandon and even betray the core interests of their home institutions is not an expectation of the literature on bureaucracy, although studies of issue networks have long made similar observations.

The prevalence of retired career military personnel in advisory positions or positions of civilian control and their influence over civilian preferences is another avenue of research this study suggests would be fruitful. There is research on the effects of military veterans in

Congress, and a small cottage industry of commentary on the presence of recently retired general and flag officers in senior political appointments in the Executive branch. But there is a paucity of systematic empirical data on the role retired military personnel play in forming military- related legislation and policy over time. In particular, civilian organizations responsible for oversight of the military that have a marked plurality of retired military personnel are worth studying, with implications for veterans’ preference hiring practices among others.634

Finally, the roles that difference forms of military organizations play in civilian preference formation is another area ripe for research. The role that combatant commands played in determining preferences over the uses of UAVs and the power of Special Operations

Command and the Joint Special Operations Command over civilian evaluations of the match between strategy and capabilities is too important not to understand more deeply.

The hidden role of the military and information asymmetries in civilian preference formation may be a source of bias in studies of civil-military relations. To the extent that civilian officials rely on the military as a source of information or even need the military’s consent to

634 Throughout the course of this research, multiple officials noted the imbalance of civilian and career military personnel in the Office of the ASD(SOLIC), for example. 289

advance preferences with non-military origins, assurances of civilian control over the military in the United States are overstated. The initial empirical analysis offered here suggests that practitioners and scholars alike should reexamine the sources of civilian preferences to reevaluate the quality of civilian control over the military.

290

REFERENCES

Aberbach, Joel D., and Bert A. Rockman. “Conducting and Coding Elite Interviews.” PS: Political Science and Politics 35, no. 4 (December 2002): 673-676. https://www.jstor.org/stable/1554807.

Adams, Gordon. The Iron Triangle: The Politics of Defense Contracting. New York, NY: Council on Economic Priorities, 1981.

Adams, Gordon, and Shoon Murray, eds. Mission Creep: The Militarization of U.S. Foreign Policy? Washington, DC: Georgetown University Press, 2014.

Ahern, David. Air Force Modernization Programs. Statement before the Subcommittee on Air and Land Forces of the House Armed Services Committee, Hearing on Budget Request for Air Force Modernization Programs, Hearings on National Defense Authorization Act for Fiscal Year 2010 and Oversight of Previously Authorized Programs. 111th Cong., 1st sess., May 20, 2009.

Air Combat Command. “RQ-1, Predator, Program Direction.” Cable message to USAF headquarters, unclassified, May 1, 2000.

Anderson, Jack. “Remote Control Superspy.” Washington Post, September 18, 1983.

Arango, Tim, and Michael S. Schmidt. “Last Convoy of American Troops Leaves Iraq.” New York Times, December 18, 2011, A6. https://www.nytimes.com/2011/12/19/world/middleeast/last-convoy-of-american-troops-leaves- iraq.html.

Arquilla, John. The Reagan Imprint: Ideas in American Foreign Policy from the Collapse of Communism to the War on Terror. Chicago: Ivan R. Dee, 2006.

Ashmore, William C. “Impact of Alleged Russian Cyber Attacks.” Baltic Security and Defense Review 11 (2009): 4-40. https://nsarchive2.gwu.edu/NSAEBB/NSAEBB424/docs/Cyber- 027.pdf.

Associated Press. “Military Relying More on Drones, Mostly in Iraq.” NBC News, last updated January 1, 2008. http://www.nbcnews.com/id/22463596/ns/world_news- mideast_n_africa/t/military-relying-more-drones-mostly-iraq/#.XmZ9aR7Qh7M.

Associated Press. “Premier is Bitter.” New York Times, May 12, 1960.

Association for Diplomatic Studies and Training. “The World Was Tired of Haiti: The 1994 U.S. Intervention.” July 11, 2016. https://adst.org/2016/07/world-tired-haiti-1994-u-s-intervention/.

Atherton, Kelsey D. “How Rumsfeld Viewed the Most Iconic Aircraft of the War on Terror.” C4ISRNet, January 25, 2018. https://www.c4isrnet.com/unmanned/2018/01/25/how-rumsfeld- viewed-the-most-iconic-aircraft-of-the-war-on-terror/.

291

Atlantic Council. “Transcript: Lessons from our Cyber Past—the First Military Cyber Units.” March 5, 2012. https://www.atlanticcouncil.org/commentary/transcript/transcript-lessons-from- our-cyber-past-the-first-military-cyber-units/.

Avant, Deborah. Political Institutions and Military Change: Lessons from Peripheral Wars. Ithaca, NY: Cornell University Press, 1994.

Axe, David. “The Secret History of Boeing’s Killer Drone.” Wired Magazine, June 6, 2011. https://www.wired.com/2011/06/killer-drone-secret-history/.

Barnett, Frank R., B. Hugh Tovar, and Richard H. Schultz, eds. Special Operations in U.S. Strategy. Washington, DC: National Defense University Press, 1984.

Beach, Derek, and Rasmus Brun Pedersen. Process-Tracing Methods: Foundations and Guidelines. Ann Arbor: The University of Michigan Press, 2013.

Becker, Elizabeth. “They’re Unmanned, They Fly Low, and They Get the Picture.” New York Times, June 3, 1999, A15. https://archive.nytimes.com/www.nytimes.com/library/world/europe/060399kosovo-drones.html.

Beckwith, Charlie A., and Donald Knox. Delta Force. New York, NY: Harcourt Brace Jovanovich Publishers, 1983.

Berkowitz, Bruce, with Michael Suk. The National Reconnaissance Office at 50 Years: A Brief History, 2nd ed. Chantilly, VA: Center for the Study of National Reconnaissance, July 2018. https://www.nro.gov/Portals/65/documents/about/50thanniv/The%20NRO%20at%2050%20Year s%20-%20A%20Brief%20History%20-%20Second%20Edition.pdf.

Berry, Jeffrey M. “Validity and Reliability Issues in Elite Interviewing.” PS: Political Science and Politics 35, no. 4 (December 2002): 679-682. https://www.jstor.org/stable/1554809.

Berry, Jeffrey M., and Clyde Wilcox. The Interest Group Society. New York, NY: Routledge, 2018.

Best, Jr., Richard A. Intelligence Technology in the Post-Cold War Era: The Role of Unmanned Aerial Vehicles (UAVs). CRS Report No. 93-686 F. Washington, DC: Congressional Research Service, 1993. https://fas.org/sgp/crs/intel/93-686.pdf.

Betts, Richard K. Soldiers, Statesmen, and Cold War Crises. New York, NY: Columbia University Press, 1991 [1977].

Black, Jr., William B. “Thinking Out Loud about Cyberspace.” Cryptolog 8, no. 1 (Spring, 1997). https://nsarchive2.gwu.edu//dc.html?doc=2700088-Document-11.

Blankshain, Jessica Deighan. “Essays on Interservice Rivalry and American Civil-Military Relations.” PhD diss., Harvard University, 2014. https://dash.harvard.edu/handle/1/12274304.

292

Blom, John David. Unmanned Aerial Systems: A Historical Perspective, Occasional Paper 37. Fort Leavenworth, KS: Combat Studies Institute Press, 2010. https://pdfs.semanticscholar.org/94af/8293b07716f2b2052dd0a8315a3da934fb63.pdf.

Boehm, Joshua, with Craig Baker, Stanley Chan, and Mel Sakazaki. A History of United States National Security Space Management and Organization. Washington, DC: Commission to Assess United States National Security Space Management and Organization, January 2001.

Borger, Julian. “Pentagon Kept the Lid on Cyberwar in Kosovo.” The Guardian, November 8, 1999. https://www.theguardian.com/world/1999/nov/09/balkans.

Bowden, Mark. Black Hawk Down: A Story of Modern War. New York, NY: Penguin Books, 1999.

Bowie, Christopher, Fred L. Frostic, Kevin N. Lewis, John Lund, David Ochmanek, and Philip Propper. The New Calculus: Analyzing Airpower’s Changing Role in Joint Theater Campaigns. Washington, DC: RAND Corporation, 1993. https://www.rand.org/pubs/monograph_reports/MR149.html.

Boykin, William G. Special Operations and Low-Intensity Conflict Legislation: Why It Was Passed and Have the Voids Been Filled. Carlisle, PA: U.S. Army War College, 1991.

Brewin, Bob. “Kosovo Ushered in Cyberwar.” Federal Computer Week, September 27, 1999. https://fcw.com/articles/1999/09/27/kosovo-ushered-in-cyberwar.aspx.

Brimley, Shawn W. The Third-Offset Strategy: Securing America’s Military-Technical Advantage. Testimony before the House Armed Services Committee Subcommittee on Seapower and Projection Forces, 113th Cong., 2nd sess., December 2, 2014. https://s3.amazonaws.com/files.cnas.org/documents/Brimley_HASC_PreparedStatement_12022 014.pdf?mtime=20160906080438.

Brooks, Risa. Shaping Strategy: The Civil-Military Politics of Strategic Assessment. Princeton, NJ: Princeton University Press, 2008.

Brown, Harold, with Joyce Winslow. Star Spangled Security: Applying Lessons Learned Over Six Decades Safeguarding America. Washington, DC: Brookings Institution Press, 2012.

Brown, J. H., and R. G. Ollila, and R. D. Minckler. A Survey and Technical Systems Assessment of Drone Aircraft for Tactical Reconnaissance and Surveillance, Volume 1. Columbus, OH: Advanced Research Projects Agency, February 1972. https://apps.dtic.mil/dtic/tr/fulltext/u2/595802.pdf.

Bruneau, Thomas C., and Richard B. Goetze, Jr. “Ministries of Defense and Democratic Control,” in Who Guards the Guardians and How: Democratic Civil-Military Relations, edited by Thomas C. Bruneau and Schott D. Tollefson. Austin, TX: University of Texas Press, 2006.

293

Builder, Carl. The Masks of War: American Military Styles in Strategy and Analysis. Baltimore, MD: Johns Hopkins University Press 1989.

Bush, George H.W. Public Report of the Vice President’s Task Force on Combatting Terrorism. Washington, DC, February 1986. https://www.ncjrs.gov/pdffiles1/Digitization/138789NCJRS.pdf.

Bush, George W. “A Period of Consequences.” Speech, The Citadel, SC, September 23, 1999.

Bush, George W. “Bush Campaign Speech.” C-SPAN video, 40:29, September 23, 1999. https://www.c-span.org/video/?152320-1/bush-campaign-speech&start=1504.

Bush, George W. Decision Points. New York, NY: Random House, 2010.

Butler, Michael J. Selling a ‘Just’ War: Framing, Legitimacy, and U.S. Military Intervention. London: Palgrave Macmillan, 2012.

Campbell, Colton C., and David Auerswald, eds. Congress and Civil-Military Relations. Washington, DC: Georgetown University Press, 2015.

Campen, Alan D., ed. The First Information War: The Story of Communications, Computers, and Intelligence Systems in the Persian Gulf War. Fairfax, VA: AFCEA International Press, 1992.

Carter, Ash. A Lasting Defeat: The Campaign to Destroy ISIS. Boston, MA: Harvard Kennedy School, Belfer Center for Science and International Affairs, October 2017. https://www.belfercenter.org/LastingDefeat.

Carter, Ash. Inside the Five-Sided Box: Lessons from a Lifetime of Leadership in the Pentagon. New York, NY: Penguin Random House, 2019.

Chairman of the Joint Chiefs of Staff. “Memorandum of Policy 30; Command and Control Warfare.” Memorandum, Washington, DC, March 8, 1993 (as revised; originally issued July 17, 1990). https://apps.dtic.mil/dtic/tr/fulltext/u2/a389344.pdf.

Chamayou, Gregoire. A Theory of the Drone. New York, NY: The New Press, 2013.

Cheney, Dick, with Liz Cheney. In My Time: A Personal and Political Memoir. New York, NY: Threshold Editions, 2011.

Cheney, Richard. Conflicting Trends and Long Term Defense Needs, Statement before the Senate Armed Services Committee, February 17, 1991.

294

Cheney, Richard. Interview by Philip Zellikow. “Richard B. Cheney Oral History.” The Miller Center, University of Virginia. March 16-17, 2000. https://millercenter.org/the- presidency/presidential-oral-histories/richard-b-cheney-oral-history-secretary-defense.

Chesney, Robert. “Military-Intelligence Convergence and the Law of the Title 10/Title 50 Debate.” Journal of National Security Law and Policy 5 (2012): 539-629. https://jnslp.com/wp- content/uploads/2012/01/Military-Intelligence-Convergence-and-the-Law-of-the-Title-10Title- 50-Debate.pdf.

Chollet, Derek, and James Goldgeier. America Between the Wars: From 11/9 to 9/11. New York, NY: Public Affairs, 2009.

Chong, Dennis, and James N. Druckman. “Framing Theory.” Annual Review of Political Science 10, no. 1 (2007): 103-126. doi:10.1146/annurev.polisci.10.072805.103054.

Chung, Timothy. “OFFensive Swarm-Enabled Tactics (OFFSET).” Defense Advanced Research Projects Agency, n.d. https://www.darpa.mil/program/offensive-swarm-enabled-tactics.

Cimbala, Stephen J. “The Role of Military Advice: Civil-Military Relations and Bush Security Strategy,” in U.S. Civil-Military Relations: In Crisis or Transition?, edited by Don M. Snider and Miranda A. Carlton-Carew. Washington, DC: Center for Strategic and International Studies, 1995.

Clancy, Tom, with Carl Steiner and Tony Koltz. Shadow Warriors: Inside the Special Forces. New York, NY: Berkley Books, 2002.

Clarke, Richard A. Against All Enemies: Inside America’s War on Terror. New York, NY: Free Press, 2004.

Clarke, Richard A., and Robert K. Knake. Cyber War: The Next Threat to National Security and What to Do About It. New York, NY: HarperCollins, 2010.

Clark, Richard M. Uninhabited Combat Aerial Vehicles: Airpower by the People, For the People, But Not with the People. Maxwell Air Force Base, AL: Air University Press, August 2000.

Clements, S.E., S. Seely, and F.F. Urhane, eds. Problems of Battlefield Surveillance. Washington, DC: United States Army, Project TEOTA, Office of the Chief Signal Officer, 1953. https://apps.dtic.mil/dtic/tr/fulltext/u2/032532.pdf.

Clemmer, Wayne A. Low-Intensity Conflicts and the United States Marine Corps. Carlisle, PA: U.S. Army War College, 1991.

Clinton, Bill. My Life. New York, NY: Knopf, 2004.

295

Clinton, Hillary. “America’s Pacific Century.” Foreign Policy, October 11, 2011. https://foreignpolicy.com/2011/10/11/americas-pacific-century/.

Cohen, Eliot A. Supreme Command: Soldiers, Statesmen, and Leadership in Wartime. New York, NY: Simon and Schuster, 2002.

Cohen, Michael, James March, and John Olsen. “A Garbage Can Model of Organizational Choice.” Administrative Science Quarterly 17 (March 1972): 1-25. doi:10.2307/2392088.

Collins, Jeffrey, and Andrew Futter, eds. Reassessing the Revolution in Military Affairs: Transformation, Evolution and Lessons Learnt. London: Palgrave Macmillan, 2015.

Collins, John. Special Operations Forces: An Assessment. Washington, DC: National Defense University Press, 1994.

Commission on Critical Infrastructure Protection. Critical Foundations: Protecting America’s Infrastructures. Washington, DC: October 1997. https://fas.org/sgp/library/pccip.pdf.

Comptroller General. DOD’s Use of Remotely Piloted Vehicle Technology Offers Opportunities for Saving Lives and Dollars. MASAD-81-20. Washington, DC: General Accounting Office, April 3, 1981. https://books.google.com/books?id=x0ssAAAAIAAJ&printsec=frontcover&source=gbs_ViewA PI#v=onepage&q&f=false.

Congressional Research Service. Defense Advanced Research Projects Agency: Overview and Issues for Congress. CRS Report No. R45088. Washington, DC: CRS, July 24, 2018. https://fas.org/sgp/crs/natsec/R45088.pdf.

Conti, Gregory, and David Raymond. On Cyber: Towards an Operational Art for Cyber Conflict. Middletown, CA: Kopidion Press, 2017.

Corr, Edwin G., and David C. Miller. “Organizing for Operations Other Than War in the Post- Cold War Era,” in Managing Contemporary Conflict: Pillars of Success, edited by Max G. Manwaring and Wm. J. Olson. Boulder, CO: Westview Press, 1996.

Crile, George. Charlie Wilson’s War: The Extraordinary Story of how the Wildest Man in Congress and a Rogue CIA Agent Changed the History of Our Times. New York, NY: Grove Press, 2003.

Cronin, Audrey Kurth. Power to the People: How Open Technological Innovation is Armed Tomorrow’s Terrorists. Oxford: Oxford University Press, 2020.

Crowe, William J. “Implications of Low-Intensity Conflict for U.S. Policy and Strategy,” in Low-Intensity Conflict: Old Threats in a New World, edited by Edwin G. Corr and Stephen Sloan. Boulder, CO: Westview Press, 1992.

296

Cushman, Jr., John H. “Pentagon Plans Unit to Build Up Special Forces.” New York Times, July 19, 1986, A8. https://www.nytimes.com/1986/07/19/us/pentagon-plans-unit-to-build-up-special- forces.html.

Dahm, Matthias, and Nicolas Porteiro. “Informational Lobbying Under the Shadow of Political Pressure.” Social Choice and Welfare 30, no. 4 (May 2008): 531-559. https://link.springer.com/article/10.1007/s00355-007-0264-x.

Dao, James. “U.S. is Using More Drones, Despite Concern over Flaws.” New York Times, November 3, 2001, B2.

Dauber, Cori. “The Practice of Argument: Reading the Conditions of Civil-Military Relations.” Armed Forces and Society 25, no. 3 (1998): 435-446. doi:10.1177/0095327X9802400306.

Davidson, Janine. “Civil-Military Friction and Presidential Decision-Making: Explaining the Broken Dialogue.” Presidential Studies Quarterly 43, no. 1 (March 2013): 129-145. https://www.cfr.org/content/member/PSQ2013_Broken_Dialogue_Davidson.pdf.

Davis, Vincent. The Politics of Innovation: Patterns in Navy Cases. Denver, CO: University of Denver Press, 1967.

Defense Airborne Reconnaissance Office. UAV Program Plan. Washington, DC: Department of Defense, 1994. https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=2ahUKEwiZosi yxI7oAhVPjp4KHbemDG8QFjAAegQIARAB&url=https%3A%2F%2Fwww.hsdl.org%2F%3F view%26did%3D745074&usg=AOvVaw3SYideOLNM5og-GT_5wMqD.

Defense Daily International. “Navy Outlines Future UAV Strategy, First Focus is on Long Dwell ISR.” Vol. 3, no. 16 (February 2002).

Defense Science Board. Unmanned Aerial Vehicles and Uninhabited Combat Aerial Vehicles. Washington, DC: Department of Defense, February 2004. https://apps.dtic.mil/dtic/tr/fulltext/u2/a423585.pdf.

Deptula, David. “Air Force Unmanned Aerial Systems Flight Plan, 2009-2047.” Presentation, Headquarters, U.S. Air Force, Washington, DC, n.d. https://apps.dtic.mil/dtic/tr/fulltext/u2/a504096.pdf.

Desch, Michael C. Civilian Control of the Military: The Changing Security Environment. Baltimore, MD: Johns Hopkins University Press, 1999.

Dombrowski, Peter, and Eugene Gholz. Buying Military Transformation: Technological Innovation and the Defense Industry. New York, NY: Columbia University Press, 2006.

Donnithorne, Jeff. “Tinted Blue: Air Force Culture and American Civil-Military Relations.” Strategic Studies Quarterly (Winter 2010).

297

https://www.airuniversity.af.edu/Portals/10/SSQ/documents/Volume-04_Issue- 4/Donnithorne.pdf.

Downing, General Wayne A. “New National Security Challenges,” in Managing Contemporary Conflict: Pillars of Success, edited by Max G. Manwaring and Wm. J. Olson. Boulder, CO: Westview Press, 1996.

Drew, Christopher. “Drone Flights Leave Military Awash in Data.” New York Times, January 11, 2010, A1.

Drew, Christopher. “For U.S., Drones are Weapons of Choice in Fighting Qaeda.” New York Times, March 17, 2009, A1. https://www.nytimes.com/2009/03/17/business/17uav.html.

Druckman, James N. “Political Preference Formation: Competition, Deliberation, and the (Ir)relevance of Framing Effects.” American Political Science Review 98, no. 4 (November 2004): 671-686. doi:10.1017/S0003055404041413.

Druckman, James N., and Arthur Lupia. “Preference Formation.” Annual Review of Political Science 3, no. 1 (2000): 1-24. doi:10.1146/annurev.polisci.3.1.1.

The Economist. “A Cyber-Riot.” May 10, 2007. https://www.economist.com/europe/2007/05/10/a-cyber-riot.

Ehrhard, Thomas P. “Seeds of a Revolution: Maritime UAVs in the 1960s,” in Providing the Means of War: Perspectives on Defense Acquisition, 1945-2000, edited by Shannon A. Brown. Washington, DC: United States Army Center of Military History, 2015.

Ehrhard, Thomas P. “Unmanned Aerial Vehicles in the United States Armed Services: A Comparative Study of Weapon System Innovation.” PhD diss., Johns Hopkins School of Advanced International Studies, 2000. https://catalyst.library.jhu.edu/catalog/bib_2239049.

Emerson, Steven. Secret Warriors: Inside the Covert Military Operations of the Reagan Era. New York, NY: G.P. Putnam’s Sons, 1988.

Emerson, Steven. “Stymied Warriors.” New York Times Magazine, November 13, 1988.

Engel, Gregory A. “Cruise Missiles and the Tomahawk,” in The Politics of Naval Innovation, edited by Bradd C. Hayes and Douglas C. Smith. Newport, RI: U.S. Naval War College, 1994.

Farrell, Theo. “Figuring Out Fighting Organizations: The New Organizational Analysis in Strategic Studies.” Journal of Strategic Studies 19, no. 1 (Spring 1996). https://www.tandfonline.com/doi/abs/10.1080/01402399608437629.

Farrell, Theo, and Terry Terriff, eds. The Sources of Military Change: Culture, Politics, Technology. Boulder, CO: Lynne Rienner, 2002.

298

Feaver, Peter D. Armed Servants: Agency, Oversight, and Civil-Military Relations. Cambridge, MA: Harvard University Press, 2003.

Feith, Douglas J. War and Decision: Inside the Pentagon at the Dawn of the War on Terrorism. New York, NY: HarperCollins, 2008. de Figueiredo, John M. “Lobbying and Information in Politics.” Business and Politics 4, no. 2 (2002): 125-29. doi:10.2202/1469-3569.1033.

Fisher, Max. “The Iraq Success Story that Propelled David Petraeus to the Top.” Washington Post, November 9, 2012. https://www.washingtonpost.com/news/worldviews/wp/2012/11/09/the-iraq-success-story-that- propelled-david-petraeus-to-the-top/.

Floyd D. Spence National Defense Authorization Act for Fiscal Year 2001, Pub. L. 106-398, 114 Stat. 1654 (2000). https://www.govinfo.gov/app/details/PLAW-106publ398.

Fowler, Glenn. “Joseph C.R. Licklider Dies at 75; Foresaw New Uses for Computers.” New York Times, July 3, 1990, B6. https://www.nytimes.com/1990/07/03/obituaries/joseph-cr-licklider- dies-at-75-foresaw-new-uses-for-computers.html.

Fuller, Christopher J. See It/Shoot It: The Secret History of the CIA’s Lethal Drone Program. New Haven: Yale University Press, 2017.

Galvin, John R. “Conflict in the Post-Cold War Era,” in Low-Intensity Conflict: Old Threats in a New World, edited by Edwin G. Corr and Stephen Sloan. Boulder, CO: Westview Press, 1992.

Gates, Robert M. Duty: Memoirs of a Secretary at War. New York, NY: Random House, 2014.

Geers, Kenneth. “Cyberspace and the Changing Nature of Warfare.” NATO Cooperative Cyber Defense Center of Excellence, n.d. https://ccdcoe.org/uploads/2018/10/Geers2008_CyberspaceAndTheChangingNatureOfWarfare.p df.

Gellman, Barton. “Broad Effort Launched After ’98 Attacks.” Washington Post, December 19, 2001, A1. https://www.washingtonpost.com/archive/politics/2001/12/19/broad-effort-launched- after-98-attacks/888eedf6-3ed1-4d13-8391-449b8a97cd06/.

General Accounting Office. Aquila Remotely Piloted Vehicle, Its Potential Battlefield Contribution Still in Doubt. GAO/NSAID-88-19. Washington, DC: Government Accountability Office, October 1987. https://www.gao.gov/assets/210/209657.pdf.

General Accounting Office. Aquila Remotely Piloted Vehicle: Recent Developments and Alternatives. NSIAD-86-41BR. Washington, DC: Government Accountability Office, January 1986. https://www.gao.gov/products/nsiad-86-41br.

299

General Accounting Office. The Army’s Remotely Piloted Vehicle Shows Good Potential but Faces a Lengthy Development Program. C-MASAD-82-8. Washington, DC: Government Accountability Office, February 1982. https://www.gao.gov/products/C-MASAD-82-8.

George, Alexander L., and Andrew Bennett. Case Studies and Theory Development in the Social Sciences. Cambridge: The MIT Press, 2005.

Gerring, John. “What Makes a Concept Good? A Critical Framework for Understanding Concept Formation in the Social Sciences.” Polity 31, no. 3 (Spring 1999): 357-393. doi:10.2307/3235246.

Gertler, Jeremiah. U.S. Unmanned Aerial Systems. CRS Report No. 7-5700. Washington, DC: Congressional Research Service, January 3, 2012. https://fas.org/sgp/crs/natsec/R42136.pdf.

Gettinger, Dan. Drone Spending in the Fiscal Year 2017 Defense Budget. New York, NY: Bard College, February 2016. https://dronecenter.bard.edu/files/2016/03/DroneSpendingFy17_CSD_3-1.pdf.

Gibson, Christopher P., and Don M. Snider. “Civil-Military Relations and the Potential to Influence: A Look at the National Security Decision-Making Process.” Armed Forces and Society 25, no. 2 (Winter 1999). doi:10.1177/0095327X9902500202.

Girard, Philippe R. “Peacekeeping, Politics, and the 1994 U.S. Intervention in Haiti.” Journal of Conflict Studies 24, no. 1 (2004). https://journals.lib.unb.ca/index.php/jcs/article/view/290/46.

Goertz, Gary, and James Mahoney. “A Tale of Two Cultures: Causal Mechanisms and Process Tracing,” in “Symposium: Causal Mechanisms, Process Tracing, and Causal Inference.” Qualitative and Multi-Method Research 8, no. 2 (2010).

Golby, James, and Mara Karlin. “Why ‘Best Military Advice’ is Bad for the Military—and Worse for Civilians.” Orbis 62, no. 1 (Winter 2018): 137-153. https://www.fpri.org/article/2018/01/best-military-advice-bad-military-worse-civilians/.

Graham, Bradley. “Bush to Propose Sustained Rises in Military Spending.” Washington Post, February 3, 2002, A6.

Grissom, Adam. “The Future of Military Innovation Studies.” Journal of Strategic Studies 29, no. 5 (20016): 905-934. doi:10.1080/01402390600901067.

Gunzinger, Mark. Sustaining America’s Strategic Advantage in Long-Range Strike. Washington, DC: Center for Strategic and Budgetary Assessments, 2010. https://csbaonline.org/research/publications/americas-strategic-advantage-long-range-strike.

Gwertzman, Bernard. “Bush Challenges Shultz’s Position on Terror Policy.” New York Times, October 27, 1984, A1.

300

Hagan, Joe D., Philip P. Everts, Haruhiro Fukui, and John D. Stempel. “Foreign Policy by Coalition: Deadlock, Compromise, and Anarchy.” International Studies Review 3, no. 2 (Summer 2001): 169-216. https://www.jstor.org/stable/3186568.

Hagel, Chuck. “Reagan National Defense Forum Keynote Address.” Speech, Ronald Reagan Presidential Library, Simi Valley CA, November 15, 2014. https://dod.defense.gov/News/Speeches/Speech-View/Article/606635/.

Halberstam, David. War in a Time of Peace: Bush, Clinton, and the Generals. New York, NY: Simon and Schuster, 2001.

Halloran, Richard. “Military is Quietly Rebuilding its Special Operations Forces.” New York Times, July 19, 1982, A1. https://www.nytimes.com/1982/07/19/us/military-is-quietly- rebuilding-its-special-operations-forces.html.

Halloran, Richard. “U.S. Buys Israeli Pilotless Planes.” New York Times, May 24, 1984, A8. https://www.nytimes.com/1984/05/24/world/us-buys-israeli-pilotless-planes.html.

Harrison, Todd. “What Has the Budget Control Act of 2011 Meant for Defense?” Center for Strategic and International Studies, Critical Questions, August 1, 2016. https://www.csis.org/analysis/what-has-budget-control-act-2011-meant-defense.

Hauben, Ronda, and Michael Hauben. Netizens: On the History and Impact of Usenet and the Internet. London: Wiley, 1997.

Hayes, Bradd C., and Douglas C. Smith, eds. The Politics of Naval Innovation. Newport, RI: U.S. Naval War College, 1994.

Healey, Jason, ed. A Fierce Domain: Conflict in Cyberspace, 1986 to 2012. Washington, DC: Atlantic Council, 2013.

Hiatt, Fred. “Weinberger Says Navy Has Bought Israeli Drones.” Washington Post, May 24, 1984.

Hunt, Alice. “Civil-Military Relations and Humanitarian Intervention.” Unpublished master’s thesis, American University, 2006.

Huntington, Samuel P. “Interservice Competition and the Political Roles of the Armed Services.” American Political Science Review 55, no. 1, (March 1961): 40-52. doi:10.1017/S0003055400124165.

Huntington, Samuel P. The Common Defense: Strategic Programs in National Politics. New York, NY: Columbia University Press, 1961.

Huntington, Samuel P. The Soldier and the State. Cambridge, MA: Belknap Press, 1957.

301

Jackson, William. “DHS Secretary: ‘Cyberspace is Civilian Space’.” GCN.com, April 27, 2011. https://gcn.com/articles/2011/04/27/napolitano-dhs-role-cybersecurity.aspx.

Jacobsen, Annie. Surprise, Kill, Vanish: The Secret History of CIA Paramilitary Armies, Operators, and Assassins. New York: Little, Brown, and Company, 2019.

Jacobsen, Annie. The Pentagon’s Brain: An Uncensored History of DARPA, America’s Top Secret Military Research Agency. New York, NY: Little, Brown, and Company, 2015.

Jacobsen, Michael. “Terrorist Financing on the Internet.” CTC Sentinel 2, no. 6 (June 2009). https://ctc.usma.edu/terrorist-financing-on-the-internet/.

Janowitz, Morris. The Professional Soldier: A Social and Political Portrait. New York, NY: Free Press, 1960.

Johnson, David E. “Modern U.S. Civil-Military Relations: Wielding the Terrible Swift Sword.” Institute for National Strategic Studies, National Defense University, McNair Paper 57,1997. https://apps.dtic.mil/dtic/tr/fulltext/u2/a421922.pdf.

Johnson, Jeh Charles. “National Security Law, Lawyers and Lawyering in the Obama Administration.” Dean’s Lecture, Yale Law School, New Haven, CT, February 22, 2012. https://ylpr.yale.edu/sites/default/files/YLPR/johnson_national_security_law_lawyers_and_lawy ering_in_the_obama_administration.pdf.

Joint Forces Quarterly. “An Interview with Paul M. Nakasone.” No. 92 (1st Quarter 2019). https://ndupress.ndu.edu/Portals/68/Documents/jfq/jfq-92/jfq-92_4-9_Nakasone-Interview.pdf.

Kanter, Arnold. Defense Politics: A Budgetary Perspective. Chicago, IL: University of Chicago Press, 1979.

Kaplan, Fred. Dark Territory: The Secret History of Cyber War. New York, NY: Simon & Schuster, 2016.

Katzenbach, Edward L. Jr. “The Horse Cavalry in the Twentieth Century,” in The Use of Force: International Politics and Foreign Policy, edited by Robert J. Art and Kenneth N. Waltz. Boston, MA: Little, Brown and Company, 1971.

Kehler, C. Robert, Herbert Lin, and Michael Sulmeyer. “Rules of Engagement for Cyberspace Operations,” in Bytes Bombs and Spies: The Strategic Dimensions of Offensive Cyber Operations, edited by Herbert Lin and Amy Zergart. Washington, DC: Brookings Institution Press, 2018.

Keller, Bill. “Conflict in Pentagon is Seen Hurting Elite Units’ Buildup.” New York Times, January 6, 1986, A1. https://www.nytimes.com/1986/01/06/us/conflict-in-pentagon-is-seen- hurting-elite-units-buildup.html.

302

Kello, Lucas. The Virtual Weapon and International Order. New Haven, CT: Yale University Press, 2017.

Kenyon, Henry S. “Unmanned Combat Aircraft Program Takes Off.” SIGNAL Magazine, July 2004. https://www.afcea.org/content/unmanned-combat-aircraft-program-takes.

Kier, Elizabeth. Imagining War: French and British Military Doctrine Between the Wars. Princeton, NJ: Princeton University Press, 1997.

King, Nigel, and Christine Horrocks. Interviews in Qualitative Research. London: Sage Publications, 2010.

Kingdon, John W. Agendas, Alternatives, and Public Policies, 2nd ed. New York, NY: Addison- Wesley Educational Publishers, Inc., 2003.

Knowlton, Brian. “Military Computer Attack Confirmed.” New York Times, August 25, 2010. https://www.nytimes.com/2010/08/26/technology/26cyber.html.

Koh, Harold Hongju. “The Obama Administration and International Law.” Speech, American Society of International Law, Washington, DC, March 25, 2010. https://2009- 2017.state.gov/s/l/releases/remarks/139119.htm.

Kohn, Richard. “Coming Soon: A Crisis in Civil-Military Relations.” World Affairs 170, no. 3 (Winter 2008): 69-80. https://www.jstor.org/stable/20672810.

Kohn, Richard. “How Democracies Control the Military.” Journal of Democracy 8, no. 4 (1997): 140-153. doi:10.1353/jod.1997.0060.

Kohn, Richard. “Out of Control: The Crisis in Civil-Military Relations.” National Interest 35 (Spring 1994). https://nationalinterest.org/print/article/out-of-control-the-crisis-in-civil-military- relations-343.

Konti, Gregory, and David Raymond. On Cyber: Towards an Operational Art for Cyber Conflict. Middletown, DE: Kopidion Press, 2017.

Koren, Jr., Henry L.T. “Congress Wades into Special Operations.” Parameters 122 (December 1988). https://apps.dtic.mil/dtic/tr/fulltext/u2/a528146.pdf.

Krepinevich, Andrew, Barry Watts, and Robert Work. Meeting the Anti-Access and Area-Denial Challenge. Washington, DC: Center for Strategic and Budgetary Analysis, May 2003. https://csbaonline.org/research/publications/a2ad-anti-access-area-denial.

Krulak, Charles A. “The Strategic Corporal: Leadership in the Three Block War.” Marines Magazine, January 1999. https://apps.dtic.mil/dtic/tr/fulltext/u2/a399413.pdf.

303

Lamb, Christopher, and Erin Staine-Pyne. 9/11, Counterterrorism and the Senior Interagency Strategy Team: Interagency Small Group Performance in Strategy Formulation and Implementation. Fort Leavenworth, KS: The Alfred D. Simons Center for Interagency Cooperation, 2014. http://thesimonscenter.org/wp-content/uploads/2014/04/IAS-003- APRIL14.pdf.

Lamb, Christopher J., and Evan Munsing. Secret Weapon: High-Value Target Teams as Organizational Innovation. Washington, DC: Institute for National Strategic Studies, National Defense University Press, 2011. https://inss.ndu.edu/Portals/68/Documents/stratperspective/inss/Strategic-Perspectives-4.pdf.

Lamothe, Dan. “Veil of Secrecy Lifted on Pentagon Office Planning ‘Avatar’ Fighters and Drone Swarms.” Washington Post, March 8, 2016. https://www.washingtonpost.com/news/checkpoint/wp/2016/03/08/inside-the-secretive- pentagon-office-planning-skyborg-fighters-and-drone-swarms.

Landler, Mark. “The Afghan War and the Evolution of Obama.” New York Times, January 1, 2017. https://www.nytimes.com/2017/01/01/world/asia/obama-afghanistan-war.html.

Lee, Timothy B. “How a Grad Student Trying to Build the First Botnet Brought the Internet to Its Knees.” Washington Post, November 1, 2013. https://www.washingtonpost.com/news/the- switch/wp/2013/11/01/how-a-grad-student-trying-to-build-the-first-botnet-brought-the-internet- to-its-knees/.

Legro, Jeff. “Culture and Preferences in the International Cooperation Two-Step.” American Political Science Review 90, no. 1 (1996): 118-137. doi:10.2307/2082802.

Lehman, Jr., John F. Command of the Seas. Annapolis, MA: Naval Institute Press, 1988.

Leiner, Barry M., Vinton G. Cerf, David D. Clerk, Robert E. Kahn, Leonard Kleinrock, Daniel C. Lynch, Jon Postel, Lawrence G. Roberts, and Stephen S. Wolff. “The Past and Future History of the Internet.” Communications of the ACM 40, no. 2 (February 1997): 102-108. https://groups.csail.mit.edu/ana/Publications/PubPDFs/The%20past%20and%20future%20histor y%20of%20the%20internet.pdf.

Lenahan, Rod. Crippled Eagle: A Historical Perspective of U.S. Special Operations, 1976-1996. Charleston, SC: Narwhal Press, 1998.

Lewis, James A. “The Internet and Terrorism.” Proceedings of the Annual Meeting, American Society of International Law 99 (March 30-April 2, 2005): 112-115. doi:10.1017/S0272503700071196.

Lin, Herbert, and Amy Zegart, eds. Bytes, Bombs, and Spies: The Strategic Dimension of Offensive Cyber Operations. Washington, DC: Brookings Institution Press, 2018.

304

Locher, III, James R. “Congress to the Rescue: Statutory Creation of USSOCOM.” Air Commando Journal 1, no. 3 (Spring 2012).

Lock-Pullan, Richard. “How to Rethink War: Conceptual Innovation and AirLand Battle Doctrine.” Journal of Strategic Studies 28, no. 4, (August 2005): 679-702. doi:10.1080/01402390500301087.

Loeb, Vernon. “New Weapon Systems Are Budget Winners.” Washington Post, February 8, 2002, A29.

Loeb, Vernon, and Thomas E. Ricks. “1’s and 0’s Replacing Bullets in U.S. Arsenal.” Washington Post, February 2, 2002, A1.

Lovelace, Douglas C., ed. Terrorism: Documents of International and Local Control. Volume 86. Oxford, UK: Oxford University Press, April 2008.

Lupia, Arthur, and Matthew D. McCubbins. The Democratic Dilemma: Can Citizens Learn What They Need to Know? New York, NY: Cambridge University Press, 1998.

Lyngaas, Sean. “Carter: U.S. Disrupting Islamic State Computer Networks.” Federal Computer Week, February 29. 2016. https://fcw.com/articles/2016/02/29/carter-isis-networks.aspx.

Lynn, William J. “Defending a New Domain: The Pentagon’s Cyberstrategy.” Foreign Affairs 89, no. 50 (September/October 2010). https://www.law.upenn.edu/live/files/6465-12-lynn- defending-a-new-domainpdf.

Magee, Clifford S. “Awaiting Cyber 9/11,” Joint Forces Quarterly 70 (3rd Quarter 2013). https://ndupress.ndu.edu/Portals/68/Documents/jfq/jfq-70/JFQ-70_76-82_Magee.pdf.

Mahnken, Thomas G. “War in the Information Age.” Joint Forces Quarterly, Winter 1995-1996: 39-43.

Maksel, Rebecca. “D.A.S.H. Goes to War.” Air and Space Magazine, March 2012. https://www.airspacemag.com/military-aviation/dash-goes-to-war-23369442/?page=1.

March, James G., with Chip Heath. A Primer on Decision Making: How Decisions Happen. New York, NY: The Free Press, 1994.

Marcinko, Richard. Rogue Warrior. New York, NY: Pocket Books, 1992.

Markoff, John. “Before the Gunfire, Cyber Attacks.” New York Times, August 12, 2008, A1. https://www.nytimes.com/2008/08/13/technology/13cyber.html.

Marquez, Ricardo A. “Transformation Achieved? Revisiting the 1997 National Defense Panel.” Joint Forces Quarterly 61 (2nd Quarter 2011).

305

Marquis, Susan. Unconventional Warfare: Rebuilding U.S. Special Operations. Washington, DC: Brookings Institution Press, 1997.

Marsh, John O. “Keynote Address,” in Special Operations in U.S. Strategy, edited by Frank R. Barnett, B. Hugh Tovar, and Richard H. Schultz. (Washington, DC: National Defense University Press, 1984).

Martin, David C., and John Walcott. Best Laid Plans: The Inside Story of America’s War Against Terrorism. New York, NY: Harper and Row, 1988.

Marzluff, Daniel M. “Flying Trucks? NO. Strike Assets? YES!” Proceedings 142, no. 9 (September 2016). https://www.usni.org/magazines/proceedings/2016/september/flying-trucks- no-strike-assets-yes.

Massachusetts Institute of Technology. BEACON HILL Report: Problems of Air Force Intelligence and Reconnaissance, Project Lincoln, Massachusetts Institute of Technology (MIT), 1952. Cambridge, MA: Massachusetts Institute of Technology, June 1952. https://www.governmentattic.org/12docs/USAF-BeaconHillReport_1952.pdf.

Masters, Jonathan. “Targeted Killings.” Council on Foreign Relations, May 23, 2013. https://www.cfr.org/backgrounder/targeted-killings.

Mattis, James, and Kori Schake, eds. Warriors and Citizens: American View of Our Military. Stanford, CA: Hoover Institution Press, 2016.

McAfee. “What is Stuxnet?” Accessed November 13, 2019. https://www.mcafee.com/enterprise/en-us/security-awareness/ransomware/what-is-stuxnet.html.

McCarty, Nolan M., and Adam Meirowitz. Political Game Theory: An Introduction. New York, NY: Cambridge University Press, 2007.

McChrystal, Stanley. My Share of the Task: A Memoir. New York, NY: Random House, 2013.

McCubbins, Matthew D., and Thomas Schwartz. “Congressional Oversight Overlooked: Police Patrols versus Fire Alarms.” American Journal of Political Science 28, no. 1 (February 1984): 165-179. doi:10.2307/2110792.

McGarry, Brendan W. The Defense Budget and Budget Control Act: Frequently Asked Questions, CRS Report No. R44039. Washington, DC: Congressional Research Service, July 2018. https://fas.org/sgp/crs/natsec/R44039.pdf.

McMaster, H.R. Dereliction of Duty: Lyndon Johnson, Robert McNamara, the Joint Chiefs of Staff, and the Lies that Led to Vietnam. New York, NY: HarperCollins, 1997.

McRaven, William H. Spec Ops: Case Studies in Special Operations Warfare: Theory and Practice. New York, NY: Ballantine Books, 1995.

306

Mehta, Aaron. “Interview: Stephen Welby, Assistant US Secretary of Defense for Research and Engineering.” Defense News, April 4, 2016. https://www.defensenews.com/industry/techwatch/2016/04/04/interview-stephen-welby- assistant-us-secretary-of-defense-for-research-and-engineering/.

Michel, Arthur Holland. “Drones in Bosnia.” Center for the Study of the Drone, June 7, 2013. https://dronecenter.bard.edu/drones-in-bosnia/.

Michel, Arthur Holland. “How Rogue Techies Armed the Predator, Almost Stopped 9/11, and Accidentally Invented Remote War.” Wired Magazine, December 17, 2015. https://www.wired.com/2015/12/how-rogue-techies-armed-the-predator-almost-stopped-911- and-accidentally-invented-remote-war/.

Millet, Allan R., Williamson Murray, and Kenneth H. Watman. “The Effectiveness of Military Organizations.” International Security 11, no. 1 (Summer 1986): 37-17. doi:10.2307/2538875.

Mitchell, Charlie. Hacked: The Inside Story of America’s Struggle to Secure Cyberspace. Lanham, MA: Rowman and Littlefield, 2016.

Molander, Roger C., Andrew S. Riddile, and Peter A. Wilson. Strategic Information Warfare: A New Face of War. Arlington, VA: RAND Corporation, 1996. https://www.rand.org/pubs/monograph_reports/MR661.html.

Moore, Jonathan, ed. The Campaign for President: 1980 in Retrospect. Cambridge, MA: Ballinger Publishing Co., 1981.

Moyar, Mark. Oppose Any Foe: The Rise of America’s Special Operations Forces. New York, NY: Basic Books, 2017.

Murphy, Caryle, and Charles R. Babcock. “Army’s Covert Role Scrutinized.” Washington Post. November 29, 1985, A1. https://www.washingtonpost.com/archive/politics/1985/11/29/armys- covert-role-scrutinized/5d3ba408-3469-4742-9f79-515ad97bd26e/.

Nakashima, Ellen. “Cyber-Intruder Sparks Response, Debate.” Washington Post, December 8, 2011. https://www.washingtonpost.com/national/national-security/cyber-intruder-sparks- response-debate/2011/12/06/gIQAxLuFgO_story.html.

Nakashima, Ellen. “Trump Approved Cyber-Strikes Against Iranian Computer Database Used to Plan Attacks on Oil Tankers.” Washington Post, June 22, 2019. https://www.washingtonpost.com/world/national-security/with-trumps-approval-pentagon- launched-cyber-strikes-against-iran/2019/06/22/250d3740-950d-11e9-b570- 6416efdc0803_story.html.

Nakashima, Ellen. “Trump Gives the Military more Latitude to Use Offensive Cyber Tools Against Adversaries.” Washington Post, August 16, 2018.

307

https://www.washingtonpost.com/world/national-security/trump-gives-the-military-more- latitude-to-use-offensive-cyber-tools-against-adversaries/2018/08/16/75f7a100-a160-11e8-8e87- c869fe70a721_story.html.

National Commission on Terrorist Attacks upon the United States. The 9/11 Commission Report. Washington, DC: National Commission on Terrorist Attacks upon the United States, 2004. http://govinfo.library.unt.edu/911/report/911Report.pdf.

Naylor, Sean. Relentless Strike: The Secret History of Joint Special Operations Command. New York, NY: St. Martin’s Press, 2015.

Newcome, Laurence R. Unmanned Aviation: A Brief History of Unmanned Aerial Vehicles. Barnsley, UK: Pen and Sword Books, Ltd., 2004.

New York Times. “Bush’s Speech at the Citadel.” December 11, 2001. https://www.nytimes.com/2001/12/11/international/bushs-speech-at-the-citadel.html.

New York Times. “Clinton’s Speech Accepting the Democratic Nomination for President.” August 30, 1996, https://www.nytimes.com/1996/08/30/us/clinton-s-speech-accepting-the- democratic-nomination-for-president.html.

New York Times. “Watching for the Y2K Bug.” December 30, 1999. https://www.nytimes.com/1999/12/30/opinion/watching-for-the-y2k-bug.html.

Obama, Barack H. “Remarks by the President at National Defense University.” Remarks, Fort McNair, Washington, DC, May 23, 2013. https://obamawhitehouse.archives.gov/the-press- office/2013/05/23/remarks-president-national-defense-university.

Office of the General Counsel. Assessment of International Legal Issues in Information Operations. Washington, DC: Department of Defense, May 1999. https://fas.org/irp/eprint/io- legal.pdf.

Office of the Secretary of Defense. 2008 National Defense Strategy. Washington, DC: Department of Defense, 2008. https://archive.defense.gov/pubs/2008NationalDefenseStrategy.pdf.

Office of the Secretary of Defense. “Establishment of a Subordinate Unified U.S. Cyber Command Under U.S. Strategic Command for U.S. Military Cyberspace Operations.” June 23, 2009. https://www.hsdl.org/?abstract&did=35655.

Office of the Secretary of Defense. “Long Range Strike and Penetrating Bomber Capabilities,” in Audit of the Acquisition of the Long-Range Strike Bomber. Washington, DC: Department of Defense, September 2015. https://www.documentcloud.org/documents/4177863-DOD-OIG- Report-2015-170-Audit-of-the-Acquisition.html.

308

Office of the Secretary of Defense. Unmanned Aerial Systems Roadmap 2005—2030. Washington, DC: August 2005. https://fas.org/irp/program/collect/uav_roadmap2005.pdf.

Office of the Secretary of Defense. Unmanned Aerial Vehicles Roadmap, 2002-2027. Washington, DC: Department of Defense, December 2002. https://www.hsdl.org/?abstract&did=446254. Owens, Mackubin Thomas. U.S. Civil-Military Relations After 9/11: Renegotiating the Civil- Military Bargain. New York, NY: Continuum International Publishing Group, 2011.

Owens, Mackubin Thomas. “What Military Officers Need to Know About Civil-Military Relations.” Naval War College Review 65, no. 2 (Spring, 2010): 67-87. https://digital- commons.usnwc.edu/cgi/viewcontent.cgi?article=1430&context=nwc-review.

Paul, Christopher, Isaac R. Porche, and Eliot Axelband. The Other Quiet Professionals: Lessons for Future Cyber Forces from the Evolution of Special Forces. Washington, DC: RAND Corporation, 2014.

PBS Frontline. “Timeline: Al Qaeda’s Global Context.” October 3, 2002. https://www.pbs.org/wgbh/frontline/article/timeline-al-qaedas-global-context/.

Pedlow, Gregory W., and Donald E. Welzenbach. The Central Intelligence Agency and Overhead Reconnaissance: The U-2 and OXCART Programs, 1954-1974. New York, NY: Skyhorse Publishing, 2016.

Pellerin, Cheryl. “Deputy Secretary: Third Offset Strategy Bolsters America’s Military Deterrence.” Department of Defense, October 31, 2016. https://www.defense.gov/Explore/News/Article/Article/991434/deputy-secretary-third-offset- strategy-bolsters-americas-military-deterrence.

Perl, Raphael F. Terrorism: U.S. Response to Bombings in Kenya and Tanzania: A New Policy Direction? CRS-98-733 F. Washington, DC: Congressional Research Service, 1998. https://digital.library.unt.edu/ark:/67531/metacrs603/.

Perlroth, Nicole, and David E. Sanger. “White House Eliminates Cybersecurity Coordinator Role.” New York Times, May 15, 2018, B2. https://www.nytimes.com/2018/05/15/technology/white-house-cybersecurity.html.

Perry, William. Interview by Russell L. Riley, Darby Morrisoe, and Andrew Ross. “William Perry Oral History.” The Miller Center, University of Virginia. February 21, 2006. https://millercenter.org/the-presidency/presidential-oral-histories/william-perry-oral-history- deputy-secretary-defense.

Peters, Heidi M. Department of Defense Contractor and Troop Levels in Afghanistan and Iraq: 2007-2018. CRS Report No. R44116. Washington, DC: Congressional Research Service, May 10, 2019. https://fas.org/sgp/crs/natsec/R44116.pdf.

309

Peterson, Bill. “Coming in from the Cold, Going out to the Bush Campaign.” Washington Post, March 1, 1980, A2. https://www.washingtonpost.com/archive/politics/1980/03/01/coming-in- from-the-cold-going-out-to-the-bush-campaign/3758ff60-0d13-43a6-9a6f-c692e20d5378/.

Pew Research Center. “Internet/Broadband Fact Sheet.” June 12, 2019. https://www.pewinternet.org/fact-sheet/internet-broadband/.

Pierson, Paul. “Power and Path Dependence,” in Advances in Comparative Historical Analysis, edited by James Mahoney and Kathleen Thelen. London: Cambridge University Press, 2015.

Plaw, Avery, Matthew S. Fricker, and Carlos R. Colon. The Drone Debate: A Primer on the U.S. Use of Unmanned Aircraft Outside Conventional Battlefields. New York, NY: Rowman and Littlefield, 2016.

Polmar, Norman. “The Pioneering Pioneer.” Naval History 27, no. 5 (October 2013). https://www.usni.org/magazines/naval-history-magazine/2013/september/historic-aircraft- pioneering-pioneer.

Posen, Barry R. The Sources of Military Doctrine: France, Britain, and Germany Between the World Wars. Ithaca, NY: Cornell University Press, 1984.

Potters, Jan, and Frans Van Winden. “Lobbying and Asymmetric Information.” Public Choice, 74 (1992): 269-292. https://www.jstor.org/stable/30025608.

Powell, Colin, with Joseph E. Perisco. My American Journey. New York, NY: Random House, 2010.

Priest, Dana. The Mission: Waging War and Keeping Peace with America’s Military. New York, NY: W.W. Norton and Company, 2003.

“Procedures for Approving Direct Action Against Terrorist Targets Located Outside the United States and Areas of Active Hostilities.” May 22, 2013. https://www.justice.gov/oip/foia- library/procedures_for_approving_direct_action_against_terrorist_targets/download.

Ramirez, Armando J. “From Bosnia to Baghdad: The Evolution of U.S. Army Special Forces 1994-2004.” Master’s Thesis, U.S. Naval Post Graduate School, 2004. https://fas.org/man/eprint/ramirez.pdf.

Reagan, Ronald W. “A Strategy for Peace in the 80s.” Televised Address, Washington, DC, October 19, 1980. National Archives. https://www.reaganlibrary.gov/10-19-80.

Record, Jeffrey. “Back to the Weinberger-Powell Doctrine?” Strategic Studies Quarterly 1, no. 1 (Fall 2007): 79-95. www.jstor.org/stable/26268385.

Reid, Steven E. “Operational Use of the Pioneer Unmanned Aerial Vehicle System.” Society for Photo-Optical Instrumentation Engineers, SPIE Proceedings 2829 (November 1996).

310

https://www.spiedigitallibrary.org/conference-proceedings-of-spie/2829/1/Operational-use-of- the-Pioneer-unmanned-aerial-vehicle-UAV-system/10.1117/12.259731.short.

Reuters. “CIA Sending 2 Drone Aircraft to Albania to Observe Bosnia.” Washington Post, January 29, 1994. https://www.washingtonpost.com/archive/politics/1994/01/29/cia-sending-2- drone-aircraft-to-albania-to-observe-bosnia/734a7e15-fd06-4122-8a0b-516a0c30b6d5/.

Rich, Ben R., and Leo Janos. Skunk Works: A Personal Memoir of My Years at Lockheed. New York, NY: Little, Brown, and Company, 1994.

Richelson, Jeffrey, ed. “The National Security Agency Declassified.” Washington, DC: George Washington University National Security Archives, December 2000. https://nsarchive2.gwu.edu//NSAEBB/NSAEBB24/nsa25.pdf.

Ricks, Thomas E. Fiasco: The American Military Adventure in Iraq. New York, NY: The Penguin Press, 2006.

Ricks, Thomas E. “U.S. Arms Unmanned Aircraft: ‘Revolution’ in Sky Above Afghanistan.” Washington Post, October 18, 2001, A1.

Robinson, Linda, Austin Long, Kimberly Jackson, and Rebeca Orrie. Improving the Understanding of Special Operations. Santa Monica, CA: RAND Corporation, 2018. https://www.rand.org/pubs/research_reports/RR2026.html.

Rogin, Josh. “Cyber Officials: Chinese Hackers Attack ‘Anything’ and ‘Everything’.” Federal Computer Week, February 13, 2007. https://fcw.com/articles/2007/02/13/cyber-officials-chinese- hackers-attack-anything-and-everything.aspx?sc_lang=en.

Rosen, Stephen Peter. Winning the Next War: Innovation and the Modern Military. Ithaca, NY: Cornell University Press, 1991.

Rumsfeld, Donald H. Known and Unknown: A Memoir. New York, NY: Penguin Group, 2011.

Rumsfeld, Donald H. “Memorandum for the President: Strategic Thoughts.” Office of the Secretary of Defense, September 30, 2001. http://library.rumsfeld.com/doclib/sp/272/2001-09- 30%20to%20President%20Bush%20re%20Strategic%20Thoughts.pdf.

Rumsfeld, Donald H. “Transforming the Military.” Foreign Affairs, May/June 2002. https://www.foreignaffairs.com/articles/2002-05-01/transforming-military.

Sale, Richard. Clinton’s Secret Wars: The Evolution of a Commander in Chief. New York, NY: St. Martin’s Press, 2009.

Sanger, David E. The Perfect Weapon: War, Sabotage, and Fear in the Cyber Age. New York, NY: Random House, 2018.

311

Sanger, David E. “U.S. Declines to Sign Declaration Discouraging Use of Cyberattacks.” New York Times, November 12, 2018, A6. https://www.nytimes.com/2018/11/12/us/politics/us- cyberattacks-declaration.html.

Sanger, David E., and William J. Broad. “Trump Inherits a Secret Cyberwar Against North Korean Missiles.” New York Times, March 4, 2017. https://www.nytimes.com/2017/03/04/world/asia/north-korea-missile-program-sabotage.html.

Sapolsky, Harvey M. Polaris System Development: Bureaucratic and Programmatic Success in Government. Cambridge, MA: Harvard University Press 1972.

Sarkesian, Sam C. “Special Operations, Low Intensity Conflict (Unconventional Conflicts), and the Clinton Defense Strategy,” in Clinton and Post-Cold War Defense, edited by Stephen J. Cimbala. Westport, CT: Praeger, 1996.

Saunders, Elizabeth N. Leaders at War: How Presidents Shape Military Interventions. Ithaca, NY: Cornell University Press, 2011.

Sayler, Kelly. A World of Proliferated Drones: A Technology Primer. Washington, DC: Center for a New American Security, 2015. https://www.jstor.org/stable/resrep06394.

Scharre, Paul. Army of None: Autonomous Weapons and the Future of War. New York, NY: W.W. Norton and Company, 2018.

Schneider, Greg. “Combat Planes that Leave Top Guns Home.” Washington Post, April 15, 2001, A1A.

Schneider, Jacquelyn. “Cyber and Crisis Escalation: Insights from Wargaming.” Naval War College, n.d. https://pacs.einaudi.cornell.edu/sites/pacs/files/Schneider.Cyber%20and%20Crisis%20Escalation %20Insights%20from%20Wargaming%20Schneider%20for%20Cornell.10-12-17.pdf.

Schulman, Loren DeJonge. Behind the Magical Thinking: Lessons from Policymaker Relationships with Drones. Washington, DC: Center for a New American Security, July 2018. https://www.cnas.org/publications/reports/behind-the-magical-thinking.

Schultz, George P. Interview by Stephen Knott, Marc Selverstone, and James Sterling Young. “George P. Schultz Oral History.” Miller Center of Public Affairs, University of Virginia, 2005. https://millercenter.org/the-presidency/presidential-oral-histories/george-p-shultz-oral-history- secretary-state.

Schwartz, Norty, and Suzie Schwartz with Ronald Levinson. Journey: Memoirs of an Air Force Chief of Staff. New York, NY: Skyhorse Publishing, 2018.

312

Shamir, Eitan, and Eyal Ben-Ari. “The Rise of Special Operations Forces: Generalized Specialization, Boundary Spanning and Military Autonomy.” Journal of Strategic Studies, 41, no. 3 (2018): 335-371. doi:10.1080/01402390.2016.1209656.

Shane, Scott. “John Michael McConnell, A Member of the Club.” New York Times, January 5, 2007, A3. https://www.nytimes.com/2007/01/05/washington/05mcconnell.html.

Shanker, Thom. “A Secret Warrior Leaves the Pentagon as Quietly as He Entered.” New York Times, May 1, 2015. https://www.nytimes.com/2015/05/02/us/a-secret-warrior-leaves-the- pentagon-as-quietly-as-he-entered.html.

Shelton, Hugh, with Ronal Levinson and Malcolm McConnell. Without Hesitation: The Odyssey of an American Warrior. New York, NY: St. Martin’s Press, 2010.

Shultz, Jr., Richard H. “Showstoppers: Nine Reasons Why We Never Sent Our Special Operations Forces after al Qaeda before 9/11.” The Weekly Standard, January 26, 2004. https://www.washingtonexaminer.com/weekly-standard/showstoppers.

Simon, Herbert A. Administrative Behavior: A Study of Decision-Making Processes in Administrative Organizations, 4th ed. New York, NY: The Free Press, 1994 [1945].

Singer, P.W., and Allan Friedman. Cybersecurity and Cyberwar: What Everyone Needs to Know. New York, NY: Oxford University Press, 2014.

Smeets, Max W. E., and Herbert Lin. “A Strategic Assessment of the U.S. Cyber Command Vision,” in Bytes, Bombs, and Spies: The Strategic Dimension of Offensive Cyber Operations, edited by Herbert Lin and Amy Zegart. Washington, DC: Brookings Institution Press, 2018.

Smith, Adam. Hearing on National Defense Authorization Act Fiscal Year 2012 and Oversight of Previously Authorized Programs, Budget Request from the Department of the Air Force before the House Armed Services Committee. 112th Cong., 1st sess., February 17, 2011.

Snyder, Jack. Ideology of the Offensive: Military Decision Making and the Disasters of 1914. Ithaca, NY: Cornell University Press, 1984.

Special Operations Review Group. Rescue Mission Report. Washington, DC: Joint Chiefs of Staff, 1980.

Steele, Robert David. “C4I: The New Linchpin.” U.S. Naval Institute, Proceedings, July 1992. https://dochot.net/document/1992-c4i-the-new-linchpin-usni-proceedings.

Stiner, Carl W. “The Strategic Employment of Special Operations Forces.” Military Review 71, no. 6 (June 1991).

313

Stirling, Wynn C., and Teppo Felin. “Game Theory, Conditional Preferences, and Social Influence.” PLoS ONE 8, no. 2 (February 2013). https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0056751.

Stolberg, Sheryl Gay, and Helene Cooper. “Obama Speeds Troops to Afghanistan and Vows to Start Pullout by 2011.” New York Times, December 2, 2009, A1.

Stoll, Clifford. “Stalking the Wily Hacker.” Communications of the ACM 31, no. 5 (May 1988). doi:10.1145/42411.42412.

Sulmeyer, Michael. “Military Cyber Issues.” Center for Strategic and International Studies, Cyber Policy Task Force Working Group Discussion Papers, 2017.

Suro, Robert. “FBI Lagging Behind on Cyber Crime.” Washington Post, October 7, 1999, A2. https://www.washingtonpost.com/wp-srv/national/daily/oct99/cyber7.htm.

Taillon, J. Paul de B. The Evolution of Special Forces in Counter-Terrorism: The British and American Experiences. Westport, CT: Praeger, 2001.

Theohary, Catherine A. Information Warfare: Issues for Congress. CRS Report No. R45142. Washington, DC: Congressional Research Service, March 5, 2018. https://fas.org/sgp/crs/natsec/R45142.pdf.

Thompson, Mark. “A Killer Drone.” Time Magazine, November 18, 2001. http://content.time.com/time/magazine/article/0,9171,184971,00.html.

Tirpak, John A. “The Robotic Air Force.” Air Force Magazine, September 1997. https://www.airforcemag.com/article/0997robotic/.

Tirpak, John A. “Toward an Unmanned Bomber.” Air Force Magazine. June 2005. https://www.airforcemag.com/article/0605bomber/.

Tomes, Robert. “The Cold War Offset Strategy: Origins and Relevance.” War on the Rocks, November 6, 2014. https://warontherocks.com/2014/11/the-cold-war-offset-strategy-origins-and- relevance/.

Toumanoff, Vladimir I. Interview by William D. Morgan. “Vadimir I. Toumanoff.” Association for Diplomatic Studies and Training Foreign Affairs Oral History Project. June 18, 1999. https://www.adst.org/OH%20TOCs/Toumanoff,%20Vladimir%20I.toc.pdf.

Trainor, Bernard E. “Special Military Forces: Congress Sees Room for Improvement.” New York Times, September 6, 1986, A9. https://www.nytimes.com/1986/09/06/us/special-military-forces- congress-sees-room-for-improvement.html.

Tversky, Aaron, and Daniel Kahneman. “Availability: A Heuristic for Judging Frequency and Probability.” Cognitive Psychology 4 (1973): 207-232. doi:10.1016/0010-0285(73)90033-9.

314

Tversky, Aaron, and Daniel Kahneman. “Prospect Theory: An Analysis of Decision Under Risk.” Econometrica 47, no. 2 (March 1979): 263-292. doi:10.2307/1914185.

Tyson, Ann Scott. “Sorry Charlie. This is Michael Vickers’s War.” Washington Post, December 28, 2007. https://www.washingtonpost.com/wp- dyn/content/article/2007/12/27/AR2007122702116.html.

Under Secretary of Defense for Acquisition. “Endurance Unmanned Aerial Vehicle Program.” Official memorandum. Washington, DC: Department of Defense, July 1993, https://nsarchive2.gwu.edu/NSAEBB/NSAEBB484/docs/Predator- Whittle%20Document%201%20- %20Deutch%20Endurance%20UAV%20Memo%2012%20July%201993.pdf.

Under Secretary of Defense for Acquisition, Technology, and Logistics. Department of Defense Report to Congress on Future Unmanned Aircraft Systems Training, Operations, and Sustainability. Washington, DC: Department of Defense, April 2012. https://fas.org/irp/program/collect/uas-future.pdf. Unified Combatant Command for Special Operations Forces, U.S. Code 10 (2011), § 167.

U.S. Air Force. “General John P. Jumper.” November 1, 2005. https://www.af.mil/About- Us/Biographies/Display/Article/104697/general-john-p-jumper/.

U.S. Air Force. “Global Integrated ISR Operations.” January 29, 2015. https://www.doctrine.af.mil/Doctrine-Annexes/Annex-2-0-Global-Integrated-ISR-Ops/.

U.S. Army Training and Doctrine Command. Joint Low-Intensity Conflict Project Final Report, Volume I: Analytical Review of Low-Intensity Conflict. Fort Monroe, VA, 1986. https://apps.dtic.mil/dtic/tr/fulltext/u2/a185971.pdf.

U.S. Congress. House. Committee on Appropriations. Department of Defense Appropriations Bill, 1992. 102nd Cong., 2nd sess., 1991. H. Rep. 102-95.

U.S. Congress. House. Committee on Armed Services. Department of Defense Authorization Act 1985 Report. 98th Cong., 2nd sess., April 19, 1984. H. Rep. 98-691.

U.S. Congress. House. National Defense Authorization Act for Fiscal Year 1994. 103rd Cong., 1st sess., 1993. H. Rep. 103-357, 599.

U.S. Congress. Senate. Committee on Appropriations. Department of Defense Appropriations Act 1988. 100th Cong., 2nd sess., 1988. S. Rep. 100-235.

U.S. Congress, Senate. Defense Organization: The Need for Change. Washington, DC: October 1985.

315

U.S. Congress. Senate. Hearings before the Select Committee on Intelligence, Nomination of John M. Deutch to be Director of Central Intelligence. 104th Cong., 1st sess. April 26 and May 3, 1995. https://www.intelligence.senate.gov/sites/default/files/hearings/nominationofjohn1995unit.pdf.

U.S. Congress, Senate, To Combat Terrorism and other Forms of Unconventional Warfare, Hearing before the Subcommittee on Sea Power and Force Projection of the Committee on Armed Services, 99th Cong., 2nd sess., August 5, 1986.

U.S. Cyber Command. Achieve and Maintain Cyberspace Superiority: Command Vision for U.S. Cyber Command. Fort Meade: U.S. Department of Defense, April 2018. https://www.cybercom.mil/Portals/56/Documents/USCYBERCOM%20Vision%20April%20201 8.pdf?ver=2018-06-14-152556-010.

U.S. Department of Defense. Annual Report to the Congress, FY1978. Washington, DC: Department of Defense, January 1977. https://history.defense.gov/Portals/70/Documents/annual_reports/1978_DoD_AR.pdf?ver=2014- 06-24-150750-460.

U.S. Department of Defense. Annual Report to the Congress, FY1984. Washington, DC: Department of Defense, February 1983. https://history.defense.gov/Portals/70/Documents/annual_reports/1984_DoD_AR.pdf?ver=2014- 06-24-151113-310.

U.S. Department of Defense, Annual Report to the Congress, FY1988. Washington, DC: Department of Defense, January 1987. https://history.defense.gov/Portals/70/Documents/annual_reports/1988_DoD_AR.pdf?ver=2014- 06-24-151204-840.

U.S. Department of Defense. Annual Report to the Congress, FY1989. Washington, DC: Department of Defense, February 1989. https://history.defense.gov/Portals/70/Documents/annual_reports/1989_DoD_AR.pdf?ver=2014- 06-24-151548-313.

U.S. Department of Defense. Annual Report to the President and the Congress, FY1992. Washington, DC: Department of Defense, February 1992. https://history.defense.gov/Portals/70/Documents/annual_reports/1992_DoD_AR.pdf?ver=2014- 06-24-152154-673.

U.S. Department of Defense. Annual Report to the President and the Congress, FY1993. Washington, DC: Department of Defense, January 1993. https://history.defense.gov/Portals/70/Documents/annual_reports/1993_DoD_AR.pdf?ver=2014- 06-24-152422-603.

U.S. Department of Defense. Annual Report to the President and the Congress, FY1994. Washington, DC: Department of Defense, January 1994.

316

https://history.defense.gov/Portals/70/Documents/annual_reports/1994_DoD_AR.pdf?ver=2014- 06-24-152508-117.

U.S. Department of Defense. Annual Report to the President and the Congress, FY1995. Washington, DC: Department of Defense, February 1995. https://history.defense.gov/Portals/70/Documents/annual_reports/1995_DoD_AR.pdf?ver=2014- 06-24-152712-813.

U.S. Department of Defense. Annual Report to the President and the Congress, FY1996. Washington, DC: Department of Defense, February 1996. https://history.defense.gov/Portals/70/Documents/annual_reports/1996_DoD_AR.pdf?ver=2014- 06-24-153009-697.

U.S. Department of Defense. Annual Report to the President and the Congress, FY1997. Washington, DC: Department of Defense, April 1997. https://history.defense.gov/Portals/70/Documents/annual_reports/1997_DoD_AR.pdf?ver=2014- 06-24-153256-173.

U.S. Department of Defense. Annual Report to the President and the Congress, FY2002. Washington, DC: Department of Defense, 2002. https://history.defense.gov/Portals/70/Documents/annual_reports/2002_DoD_AR.pdf?ver=2014- 06-24-153732-117.

U.S. Department of Defense. Annual Report to the President and the Congress, FY2005. Washington, DC: Department of Defense, 2005. https://history.defense.gov/Portals/70/Documents/annual_reports/2005_DoD_AR.pdf?ver=2014- 06-24-153944-813.

U.S. Department of Defense. Conduct of the Persian Gulf War: Final Report to Congress. Washington, DC: Department of Defense, April 1992. https://apps.dtic.mil/dtic/tr/fulltext/u2/a249270.pdf.

U.S. Department of Defense. “Department of Defense Announces Successful Micro-Drone Demonstration.” Press release, January 9, 2017. https://www.defense.gov/Newsroom/Releases/Release/Article/1044811/department-of-defense- announcessuccessful-micro-drone-demonstration/.

U.S. Department of Defense. DoD Joint UAV Program Master Plan. Washington, DC: Department of Defense, July 1988. https://apps.dtic.mil/dtic/tr/fulltext/u2/a197751.pdf.

U.S. Department of Defense. Information Warfare. DoD Directive TS-3600.1. Washington, DC: Department of Defense, December 1992. https://www.esd.whs.mil/Portals/54/Documents/FOID/Reading%20Room/Other/14-F- 0492_doc_02_Directive_S-3600-1.pdf.

317

U.S. Department of Defense. Information Operations (IO). DoD Directive S-3600.1. Washington, DC: Department of Defense, December 9, 1996. https://fas.org/irp/doddir/dod/info_ops.pdf.

U.S. Department of Defense. Operation of the Adaptive Acquisition Framework. DoD Directive 5000.02. Washington, DC: Department of Defense, January 2020. https://www.esd.whs.mil/Portals/54/Documents/DD/issuances/dodi/500002p.pdf?ver=2020-01- 23-144114-093.

U.S. Department of Defense. Report of the Quadrennial Defense Review. Washington, D.C.: May 1997. https://www.files.ethz.ch/isn/32542/qdr97.pdf.

U.S. Department of Defense. Quadrennial Defense Review Report. Washington, DC: September 2001. https://archive.defense.gov/pubs/qdr2001.pdf.

U.S. Department of Defense. Quadrennial Defense Review Report. Washington, DC: February 2006. https://archive.defense.gov/pubs/pdfs/QDR20060203.pdf.

U.S. Department of Defense. Summary: Department of Defense Cyber Strategy 2018. Washington, DC: Department of Defense, 2018. https://media.defense.gov/2018/Sep/18/2002041658/-1/- 1/1/CYBER_STRATEGY_SUMMARY_FINAL.PDF.

U.S. Department of Defense. Sustaining U.S. Global Leadership: Priorities for 21st Century Defense. Washington, DC: Department of Defense, January 2012. https://archive.defense.gov/news/Defense_Strategic_Guidance.pdf.

U.S. Department of Defense. The Defense Acquisition System. DoD Directive 5000.01. Washington, DC: Department of Defense, May 2003. https://www.esd.whs.mil/Portals/54/Documents/DD/issuances/dodd/500001p.pdf.

U.S. Department of Defense. “The Kennedy Commitments and programs, 1961,” in Report of the Office of the Secretary of Defense Vietnam Task Force (The Pentagon Papers). Washington, DC.

U.S. Department of Defense, Unmanned Systems Integrated Roadmap FY2013-2038. Washington, DC: Department of Defense, January 2014. https://apps.dtic.mil/dtic/tr/fulltext/u2/a592015.pdf.

U.S. Department of State, Office of the Historian. Foreign Relations of the United States. Volumes X/XI/XII. American Republics; Cuba. https://static.history.state.gov/frus/frus1961- 63v10-12mSupp/pdf/d428.pdf.

U.S. General Accounting Office. Information Security: Computer Attacks at Department of Defense Pose Increasing Risks, GAO/AIMD-96-84. Washington, DC: GAO, May 22, 1996. https://fas.org/irp/gao/aim96084.htm.

318

U.S. Government Accountability Office. Special Operations Forces: Opportunities Exist to Improve Transparency of Funding and Assess Potential to Lessen Some Deployments. GAO-15- 571. Washington, DC: GAO, July 2015.

U.S. General Accounting Office. Special Operations Forces: Opportunities to Preclude Overuse and Misuse. NSIAD-97-85. Washington, DC: GAO, May 1997. https://www.gao.gov/products/GAO/NSIAD-97-85.

U.S. Joint Chiefs of Staff. Joint Publication 1, Doctrine for the Armed Forces of the United States. Washington, DC, Department of Defense, 1979. https://www.jcs.mil/Portals/36/Documents/Doctrine/pubs/jp1_ch1.pdf.

U.S. Joint Chiefs of Staff. Joint Publication 3-05: Special Operations. Washington, DC, Department of Defense, July 2014. https://www.jcs.mil/Portals/36/Documents/Doctrine/pubs/jp3_05.pdf.

U.S. Joint Chiefs of Staff. Joint Publication 3-07: Joint Doctrine for Military Operations Other Than War. Washington, DC: Department of Defense, June 1995. https://www.jcs.mil/Portals/36/Documents/Doctrine/pubs/jp3_07.pdf.

U.S. Joint Chiefs of Staff. Joint Publication 3-12: Cyberspace Operations. Washington, DC: The Joint Staff, February 5, 2013. https://www.jcs.mil/Portals/36/Documents/Doctrine/pubs/jp3_12.pdf.

U.S. Joint Chiefs of Staff. Joint Publication 3-13, Joint Doctrine for Information Operations. Washington, DC: Department of Defense, October 9, 1998. https://www.jcs.mil/Portals/36/Documents/Doctrine/pubs/jp3_13.pdf.

U.S. Joint Chiefs of Staff. Joint Publication 3-13.1, Joint Doctrine for Command and Control Warfare (C2W). Washington, DC: Department of Defense, 1996. https://www.hsdl.org/?abstract&did=3760.

U.S. Joint Chiefs of Staff. National Military Strategy-Cyber Operations. Washington, DC: Department of Defense, 2006. https://www.hsdl.org/?abstract&did=35693/.

U.S. National Security Council. A National Security Strategy of Engagement and Enlargement. Washington, DC: The White House, 1994, https://history.defense.gov/Portals/70/Documents/nss/nss1994.pdf?ver=2014-06-25-121219-500.

U.S. National Security Council. A National Security Strategy for a Global Age. Washington, DC: The White House, 2000. https://www.hsdl.org/?abstract&did=460717.

U.S. National Security Council. National Security Strategy of the United States. Washington, DC: The White House, 1988. https://history.defense.gov/Portals/70/Documents/nss/nss1988.pdf?ver=2014-06-25-121129-393.

319

U.S. National Security Council. National Security Strategy Report. Washington, DC, The White House, 1987. https://nssarchive.us/national-security-strategy-1987/.

U.S. National Security Council. National Security Strategy Report. Washington, DC: The White House, 1990. http://nssarchive.us/national-security-strategy-1990/.

U.S. National Security Council. National Security Strategy Report. Washington, DC: The White House, 1991. http://nssarchive.us/national-security-strategy-1991/.

U.S. National Security Council. Report to the Congress on U.S. Capabilities to Engage in Low Intensity Conflict and Conduct Special Operations. Washington, DC: The White House, 1987.

U.S. National Security Council. The National Security Strategy of the United States of America. Washington, DC: The White House, 2006. https://www.comw.org/qdr/fulltext/nss2006.pdf.

U.S. National Security Council. The National Security Strategy of the United States of America. Washington, DC: The White House, 2010. http://nssarchive.us/national-security-strategy-2010/.

U.S. Special Operations Forces. 1992 Posture Statement. Washington, DC: June 1992.

U.S. Special Operations Command. Factbook 2009. MacDill Air Force Base, FL: 2009.

U.S. Special Operations Command. Factbook 2012. MacDill Air Force Base, FL: 2012.

U.S. Special Operations Command. United States Special Operations Command History. MacDill Air Force Base, FL: 2007.

Van Atta, Richard H., et al. “Transformation and Transition: DARPA’s Role in Fostering an Emerging Revolution in Military Affairs, Volume 1—Overall Assessment.” Alexandria: Institute for Defense Analyses, April 2003.

Vatis, Michael A. Director, National Infrastructure Protection Center, Federal Bureau of Investigation. Statement for the Record before the Senate Armed Services Committee Subcommittee on Emerging Threats and Capabilities, March 1, 2000.

Vickers, Michael. “For Guidance on Iraq, Look to Afghanistan: Use Fewer U.S. Troops, Not More.” USA Today, June 27, 2004.

Vickers, Michael, and Robert Martinage. Future Warfare 20XX Wargame Series: Lessons Learned Report. Washington, DC: Center for Strategic and Budgetary Analysis, December 2001. http://indianstrategicknowledgeonline.com/web/R.20011201.FutureWarXX.pdf.

Waller, Douglas C. The Commandos: The Inside Story of America’s Secret Soldiers. New York, NY: Dell Publishing, 1994.

320

Wagner, William. Lightning Bugs and Other Reconnaissance Drones. Fallbrook, CA: Armed Forces Journal International, 1982.

War is Boring. “The U.S. Army Got its First Drones 55 Years Ago.” Medium, July 27, 2014. https://medium.com/war-is-boring/the-u-s-army-got-its-first-drones-55-years-ago-fc92ca1c3b3d.

Ware, Willis W., ed. Security Controls for Computer Systems: Report of Defense Science Board Task Force on Computer Security. Santa Monica, CA: RAND Corporation, February 1970/ reissued October 1979. https://www.rand.org/pubs/reports/R609-1/index2.html.

Warner, Michael. “Notes on Military Doctrine for Cyberspace Operations in the United States, 1992-2004.” Cyber Defense Review, August 27, 2015. https://cyberdefensereview.army.mil/CDR-Content/Articles/Article- View/Article/1136012/notes-on-military-doctrine-for-cyberspace-operations-in-the-united-states- 1992/.

Washington Post. “General Richard G. Stillwell Dies.” December 26, 1991. https://www.washingtonpost.com/archive/local/1991/12/26/gen-richard-g-stilwell- dies/9187423d-d6d8-417f-aaa7-c6da20b8630e/?noredirect=on&utm_term=.8d570dfcef0d.

Watkins, Ali. “Two Years Later, White House Still Hitting Roadblocks in Effort to Move Drone Program Out of CIA Control.” Huffington Post, March 10, 2015. https://www.huffpost.com/entry/obama-drone-cia_n_6817858?guccounter=1.

Weber, Max. Economy and Society. Berkeley, CA: University of California Press, 1922.

Weigley, Russell F. “The American Military and the Principles of Civilian Control from McClellan to Powell.” The Journal of Military History 57, no. 5 (1993): 27-58. doi:10.2307/2951800.

Weinberger, Caspar. Fighting for Peace: Seven Critical Years in the Pentagon. New York, NY: Warner Books, 1990.

Weinberger, Caspar. “U.S. Defense Strategy” in The Reagan Foreign Policy, edited by William G. Hyland. New York, NY: New American Library, 1987.

Weiner, Sharon K. “Organizational Interests versus Battlefield Needs: The U.S. Military and Mine-Resistant Ambush Protected Vehicles in Iraq.” Polity 42, no. 4 (October 2010): 461-482. https://www.jstor.org/stable/40865427.

Weiner, Sharon K. “Military Advice for Political Purposes,” in Mission Creep: The Militarization of U.S. Foreign Policy?, edited by Gordon Adams and Shoon Murray. Washington, DC: Georgetown University Press, 2014.

321

Western, Jon. “Sources of Humanitarian Intervention: Beliefs, Information, and Advocacy in the U.S. Decisions on Somalia and Bosnia.” International Security 26, no. 4 (Spring 2002). https://www.jstor.org/stable/3092104.

White, Daniel M. “The Federal Information Security Management Act of 2002: A Potemkin Village.” Fordham University Law Review 79, no. 1 (2011): 369-405. https://ir.lawnet.fordham.edu/flr/vol79/iss1/12/.

The White House. National Cyber Strategy of the United States of America. Washington, DC: The White House, September 2018. https://www.whitehouse.gov/wp- content/uploads/2018/09/National-Cyber-Strategy.pdf.

The White House. “National Security Decision Directive 138: Combatting Terrorism.” April 3, 1984. https://fas.org/irp/offdocs/nsdd/nsdd-138.pdf.

The White House. National Security Decision Directive Number 145, National Policy on Telecommunications and Automated Information Systems Security. Washington, DC: The White House, September 17, 1984.

The White House. National Security Decision Directive 207: The National Program for Combatting Terrorism. Washington, DC: The White House, January 20, 1986.

The White House. Presidential Decision Directive 39: U.S. Policy on Counterterrorism. Washington, DC: 1995. https://www.hsdl.org/?abstract&did=462942.

The White House. Presidential Decision Directive/NSC-62: Protection Against Unconventional Threats to the Homeland and Americans Overseas. Washington, DC: The White House, 1998. https://www.hsdl.org/?abstract&did=758094.

The White House. Presidential Decision Directive/NSC-63: Critical Infrastructure Protection. Washington, DC: The White House, May 22, 1998. https://fas.org/irp/offdocs/pdd/pdd-63.htm.

The White House. National Security Presidential Directive-54/Homeland Security Presidential Directive-23. Washington, DC: The White House, January 8, 2008. https://fas.org/irp/offdocs/nspd/nspd-54.pdf.

Whittle, Richard. Predator: The Secret Origins of the Drone Revolution. New York, NY: Henry Holt and Company, 2014.

Wills, David C. The First War on Terrorism: Counter-Terrorism Policy During the Reagan Administration. New York, NY: Rowman and Littlefield, 2003.

Wilson, Clay. Network Centric Operations: Background and Oversight Issues for Congress. CRS Report No. RL32411. Washington, DC: Congressional Research Service, March 15, 2007. https://fas.org/sgp/crs/natsec/RL32411.pdf.

322

Wilson, George C. “Senate Chairman Pushes Unmanned Warfare.” Government Executive, March 6, 2000. https://www.govexec.com/federal-news/2000/03/senate-chairman-pushes- unmanned-warfare/1977/.

Wilson, George C. “Unmanned Weapons Gain Backing.” Washington Post, September 5, 1985, A1.

Wilson, George C. “U.S. Drone Crashes in El Salvador.” Washington Post, August 24, 1985, A12. https://www.washingtonpost.com/archive/politics/1985/08/24/us-drone-crashes-in-el- salvador/091e45c4-daaa-4738-a7b1-33172f788eee/.

Wilson, James Q. Bureaucracy: What Government Agencies Do and Why They Do It. New York, NY: Basic Books, 1989.

Wolf, Harrison G. Drones: Safety Risk Management for the Next Evolution of Flight. New York, NY: Routledge, 2017.

Woods, Chris. Sudden Justice: America’s Secret Drone Wars. New York, NY: Oxford University Press, 2015.

Woolsey, Jim, Joyce Pratt, Janet Andres, Jeffrey Harris, and Richard Haver. Interview by Miller Center, University of Virginia, January 13, 2010. https://millercenter.org/the- presidency/presidential-oral-histories/r-james-woolsey-oral-history-director-central.

Work, Robert O. “Remarks by Deputy Secretary Work on Third Offset Strategy.” Speech, Brussels, Belgium, April 28, 2016. https://www.defense.gov/Newsroom/Speeches/Speech/Article/753482/remarks-by-deputy- secretary-work-on-third-offset-strategy/.

Work, Robert O., and Elbridge Colby. “The Pentagon Must Modernize Before It’s Too Late.” Washington Post, September 17, 2018. https://www.washingtonpost.com/opinions/the-pentagon- must-modernize-before-its-too-late/2018/09/17/8a9ab1e8-b1fe-11e8-aed9- 001309990777_story.html.

Work, Robert O., and Shawn Brimley. 20YY: Preparing for War in the Robotic Age. Washington, DC: Center for a New American Security, January 2014. https://www.cnas.org/publications/reports/20yy-preparing-for-war-in-the-robotic-age.

Woodward, Bob. Bush at War. New York, NY: Simon and Schuster, 2002.

Woodward, Bob. Obama’s Wars. New York, NY: Simon and Schuster, 2010.

Yenne, Bill. Birds of Prey: Predators, Reapers and America’s Newest UAVs in Combat. North Branch, MN: Specialty Press, 2010.

323

Zaller, John. The Nature and Origins of Mass Opinion. New York, NY: Cambridge University Press, 1992.

Zaloga, Steven. Unmanned Aerial Vehicles: Robotic Air Warfare 1917-2007. New York, NY: Osprey Press, 2008.

Zegart, Amy. Flawed by Design: The Evolution of the CIA, JCS, and NSC. Stanford, CA: Stanford University Press, 1999.

Zetter, Kim. “An Unprecedented Look at Stuxnet, the World’s First Digital Weapon.” Wired Magazine, November 3, 2014. https://www.wired.com/2014/11/countdown-to-zero-day-stuxnet/.

Zisk, Kimberly M. Engaging the Enemy: Organization Theory and Soviet Military Innovation, 1955-1991. Princeton, NJ: Princeton University Press, 1993.

324