Women in War: Militancy, Legitimacy, and Rebel Success Meredith Loken A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy University of Washington 2018 Reading Committee: Elizabeth Kier, Chair Jon Mercer Geoffrey P.R. Wallace Program Authorized to Offer Degree: Political Science © Copyright 2018 Meredith Loken University of Washington Abstract Women in War: Militancy, Legitimacy, and Rebel Success Meredith Loken Chair of Supervisory Committee: Elizabeth Kier Political Science This project examines the role of women militants in influencing rebel engagement with civilians and improving rebel outcomes. A wide literature points to civilian support as key to insurgent success. I demonstrate that rebels believe that women’s militancy can make civilians more receptive to their grievances and that they recognize the importance of civilian support to tangible conflict successes. I further contend that women-inclusive groups do experience more conflict-related success, operationalized as territorial control during conflict and favorable conflict outcomes. I suggest that women’s militancy in front-line, support, and leadership roles can have ideologically and politically legitimizing effects among civilians. Rebels capitalize on women’s participation in ways that legitimize political violence by invoking gendered narratives, signaling the exigency of struggle and integrating the insurgency into local communities. Further, because women are generally unexpected actors they can increase rebel capabilities on and off the front-lines. Rebels reference women’s successful violence to valorize their campaigns, but also to increase civilian confidence in their strength. My conclusions are based on a multi-method research design. First, I argue that rebels portray women militants and civilians in their political visuals to legitimize grievances, integrate their struggles into local communities, and increase popular support. I adopt a qualitative strategy that explores rebels’ gendered attempts to generate legitimacy in visual propaganda. I conduct a detailed case study of 532 unique Irish Republican Army (IRA) and other republican visuals produced during the Troubles (1968-1998), a secondary, and less systematic, study of Palestinian political posters, and a third micro-comparison of Afghan, Namibian, Lebanese, Angolan, South African, Mozambican, Nicaraguan, Bangladeshi, and transnational jihadist militant groups’ visuals to establish the transportability of gender tropes in conflict propaganda across time and space. I find that rebels of varying affiliation, ideology, and tactics use their visual platforms to shape narratives about gender and women militants in ways intended to build legitimacy for their campaigns Second, I employ quantitative methods to demonstrate that organizations with women militants experience more conflict-related success, specifically territorial control during conflict and certain favorable conflict outcomes. I introduce an original dataset, the Rebel Women Data Project (RWDP), of women’s participation in 146 cross-national insurgencies operating between 1960-2016. Using these data, I find that women-inclusive rebel organizations are significantly and substantively more likely to control territory than other groups. I further find that female combatants, auxiliaries, and leaders make rebel victory significantly more likely and may sometimes increase the odds of peace agreements. I emphasize the role of civilian support in rebel success and I highlight the myriad ways in which women’s participation affects conflict dynamics. Acknowledgements I am indebted to so many people who made this work and my education possible. I am extremely grateful to my dissertation committee and the faculty of the University of Washington. Beth Kier and Jon Mercer provided me with unparalleled support and advice. I am thankful for every draft they read and every edit they gave me. They pushed me, made me a better theorist, a better writer, and a much stronger academic. Geoffrey P.R. Wallace came to the University of Washington in the perfect moment and is a fantastic advisor. I am grateful for his willingness to help me define the parameters of my work, and for reading – and re-reading – on my behalf. Erica Chenoweth graciously contributed to my work from afar, and I have learned so much from her impeccable research record and ethical compass. I am appreciative of Caitlin Ainsley who, despite not being on my dissertation committee, has exceeded all expectations as a mentor. She taught me to keep things in perspective. I am also very thankful for the Department of Political Science administration, including Susanne Recordon, Ann Buscherfeld, and Catherine Quinn. I am grateful that Steve Herbert kindly agreed to be my GSR. This research is the result of two generous funding opportunities from the University of Washington. I am very appreciative of the Graduate School Presidential Dissertation Fellowship and the Department of Political Science’s Peter May Graduate Research Grant, which enabled me to conduct archival research in Northern Ireland. I am grateful for everyone in Northern Ireland who shared their expertise and homes with me, including the Linen Hall Library librarians- especially Bernadette, Jude, and Alistair - the Eileen Hickey Irish Republican History Museum docents, and Tim. Margaret Ward met with me to share her unequaled knowledge of Irish women and rebellion and Mia Bloom graciously weighed in on this aspect of my research. I am lucky to have worked alongside fantastic graduate colleagues. Vanessa Quince is a brilliant and formidable scholar. Anna Zelenz, Stephen Winkler, Emily Christensen, Wesley Zuidema, Paige Sechrest, Grace Reinke, Carolyn Dapper, and Dave Lopez have provided their support and friendship over the years. I am also grateful for Milli Lake and Emily Kalah Gade for their guidance and for being the finest examples of smart, strong women in academia. Most importantly, I owe every success to my family. My parents and brother have supported me in every endeavor, and none of this would be possible without them. My mother and father are strength and kindness embodied, I stand on the shoulders of giants. To Amy, who can now recite the first half of my job talk, your patience, support, love, and encouragement motivates me. Thank you, iceberg. Table of Contents Abstract Acknowledgements Chapter 1: Introduction 1 Women militants: a typology 4 Legitimacy and rebel propaganda 8 Outline of the dissertation 12 Chapter 2: Theorizing Gender and Rebel Legitimacy 15 The role of civilian support in rebel success 16 Women, legitimacy, and political violence 18 Strategic effects of women’s participation: securitization in rebel propaganda 28 Outcome effects of women’s participation: territorial control and conflict outcomes 32 Territorial control 32 Conflict outcomes 41 Conclusion 48 Chapter 3: The Visual War 49 The Irish Republican Army’s visual campaign 58 Militant republicanism: a history 59 Women in the IRA 62 Visualizing the Troubles 67 Women in republican propaganda 74 Fear of and victimization by the Other 75 Reverence 83 Conclusion 88 Chapter 4: Gender in Rebel Images: Cross-National Evidence 105 Women in Palestinian resistance 106 Visualizing the Palestinian conflict 108 Fear of and victimization by the Other 110 Reverence 114 Visualizing women cross-nationally 118 Conclusion 122 Chapter 5: Assessing Women’s Participation in Rebel Organizations 142 Existing systematic efforts 142 Original dataset 145 Coding women’s participation 147 Ethical challenges 152 Descriptive findings: where and how women rebel 154 Variation by group type and ideology 156 Variation over time 157 Regional variation 158 Variation within conflicts: evidence from Syria 161 Variation within conflicts: evidence from South Yemen 162 Looking beyond ‘high-intensity’ conflicts 163 Conclusion 164 Chapter 6: Rebel Women and Territorial Control 165 Method 165 Transforming women’s militancy variables 165 Conceptualizing and measuring territorial control 166 Alternative explanations 168 Empirical results 170 Assessing the effects of women’s participation across group types 179 Importance of role disaggregation 181 Case illustration: Leftist rebels in Peru 183 Conclusion 186 Chapter 7: Rebel Women and Conflict Outcomes 188 Method 188 Conceptualizing and measuring conflict outcomes 189 Alternative explanations 192 Empirical results 193 Case illustration: Victorious rebels in Uganda 199 Conclusion 202 Chapter 8: Conclusion 203 Appendix 206 Tables and Figures Chapter 2 Table 2.1: Summary of observable implications, explanations for territorial control Table 2.2: Summary of observable implications, explanations for conflict outcomes Chapter 3 Table 3.1: Summary of securitization strategies Table 3.2: Content trends in republican visual propaganda, 1968-1998 Figure 3.1: IRA factions operating during the Troubles Figure 3.2: Distribution of republican propaganda by producer, 1968-1998 Figure 3.3: Distribution of republican propaganda by decade, 1968-1998 Figure 3.4: Distribution of republican propaganda by theme and decade 1968-1998 Figure 3.5: “Stop Strip Searches,” 1984-1985, Sinn Féin Figure 3.6: “Stop the Strip Searches in Armagh,” 1984, Belfast Strip Searching Committee Figure 3.7: Unnamed. 1980s, Sinn Féin Figure 3.8: “Death Bullets,” 1981, Republican Movement Figure 3.9: “Free Our Country,” 1990s, Cumann na mBan Figure 3.10: “The Soldiers of Cumann na mBan," 1970s, Cumann na mBan Figure 3.11: “A Woman’s Place is in the Struggle,” 1999, IRA Figure 3.12: Unnamed. 1980s, Sinn Féin Figure 3.13: “Women in Struggle,” 1990s, Unknown producer Figure 3.14: “Taking Aim,” 1975, Brendan Murphy Figure 3.15: “A is for Armalite…,” 1970s, Republican Movement Figure 3.16: “Release Mary Kennedy!,” 1974-1975, United Front Against Internment Figure 3.17: “Bring All Our Prisoners Home,” 1980s, Sinn Féin Figure 3.18: “Women’s Roll of Honour: The Ultimate Sacrifice,” 1990s, Producer Unknown Figure 3.19: “Mairéad Farrell: Irish Republican Army” 1988, Sinn Féin Women’s Department Chapter 4 Figure 4.1: No Title, 1970, PLO Figure 4.2: “Providence Did,” 1985, PLO UIO Figure 4.3: “Weapons of the Revolutionaries and the Masses,” 1977, Fatah and PLO UIO Figure 4.4: “In Defence of Peace,” 1986, Fatah Figure 4.5: “We Are All For The Resistance,” 1967, PLO Figure 4.6: “The Land,” 1986, PFLP Figure 4.7: “Yom al Mara'a al Alami,” 1980, PFLP Figure 4.8: “Sacred Flame,” 2009, Fatah Figure 4.9: “Mother Palestine,” 1983, Fatah and PLO UIO Figure 4.10: “Dalal Al Mughrabi,” 1978, Fatah Figure 4.11: No Title, 2004, Hamas Figure 4.12: No Title, 1980s, Internal Islamic Front of Afghanistan Figure 4.13: “The Wounds of my Daughter…,” 1970-1980s, SWAPO Figure 4.14: “Join Umkhonto We Sizwe,” 1960s, African National Congress Figure 4.15: “Commemoration of the Mozambican Revolution,” 1973, FRELIMO Figure 4.16: “The blood of martyrs…” 1985, Hizbollah Figure 4.17: “MPLA,” undated, MPLA Figure 4.18: “Bangladesh Mothers and Daughters… ,” 1971, MIB during the War of Independence Figure 4.19: “In the first annual commemoration…,” 1986, LNRF and SSNP Chapter 5 Table 5.1: Recent data collection efforts, women’s rebel participation Table 5.2: Role identification criteria Table 5.3: Rebel Women Data Project coding rules Table 5.4: Rebel Women Data Project data Table 5.6: Distribution of women’s participation by group-type Figure 5.1: Women’s participation in rebel groups by level, 1960-2016 Figure 5.2: Group ideology and women’s leadership participation over time Figure 5.3: Highest recorded level of women’ combat participation by country Figure 5.4: Women’s participation by region, role, and level Chapter 6 Table 6.1: Distribution of the level of women’s participation by role Table 6.2: Cross-tabulation of women’s participation, by role, and rebel territorial control Table 6.3: Cross-tabulation of women’s participation, by role and level, and rebel territorial control Table 6.4: Chapter 6 summary statistics Table 6.5: Models 4 and 5, effects of female combatants on territorial control Table 6.6: Models 6 and 7, effects of female auxiliaries on territorial control Table 6.7: Models 8 and 9, effects of female leaders on rebel territorial control Table 6.8: Models 10, 11, 12, 13, and 14, effects of women’s participation on territorial control, sub- samples Figure 6.1: Model 1, Model 2, and Model 3: effects of women’s participation on territorial control Figure 6.2: Model 4 predicted probabilities Figure 6.3: Model 6 predicted probabilities Figure 6.4: Model 8 predicted probabilities Figure 6.5: Model 5 predicted probabilities Figure 6.6: Model 7 predicted probabilities Figure 6.7: Model 9 predicted probabilities Figure 6.8: Model 15 predicted probabilities Chapter 7 Table 7.1: Distribution of conflict outcomes across rebel organizations Table 7.2: Distribution of conflict outcomes by women’s participation across rebel organizations Table 7.3: Chapter 7 summary statistics Table 7.4: Model 1, Model 2, and Model 3: effects of women’s participation on conflict outcomes Table 7.5: Odds ratios: effect of women’s participation on conflict outcomes Table 7.6: Model 4, Model 5, and Model 6: effects of women’s participation and territorial control on conflict outcomes Table 7.7: Odds ratios: interaction effect between women’s participation and territorial control 1 Chapter 1: Introduction In 1986, Cecilia Magni participated in a Manuel Rodriguez Patriotic Front (FMPR) attempt to assassinate Chilean dictator Augusto Pinochet. Magni, known as ‘Comandante Tamara,’ was also a recruiter, logistician, and was the only woman to become an FMPR commander.1 In 1988, military police captured her and other FMPR members. Her tortured body was found that year in the Tinguiririca River; the government acquitted four police officers of her murder after concluding that her injuries could have resulted from an accidental drowning.2 Magni was not the first woman to fight for an insurgency, but she is emblematic of the contributions that women before and after her have made to rebellion cross-nationally. Women have always participated as front-line fighters, auxiliaries, and leaders in political violence aimed at toppling or reconstituting the state. Motivated by ideology, fear, desperation, or adventure, women partake in rebel organizations of all objectives and platforms. Still, historical accounts of conflict often render them invisible. Indeed, the notion that women are not real militants prompted a group of female ex-political prisoners in Chile to declare, “The militant and resistant women do not accept to be treated in our political commitment, as ‘the women or wives of [militants].’”3 In the last three decades, researchers have begun taking militant women more seriously. Much of this work descends from feminist international scholarship of the 1980s and 1990s, which implored security scholars to consider how gender shapes international politics.4 With progressive fractures in international relations’ traditional, androcentric viewpoint and the increasing visibility of women participants in terrorist groups, as genocidaires, and in demobilization programs, security studies scholars have broached the topic of women’s political violence. Much of this research explores why and how women join rebel organizations, which roles women take on, and how women experience conflict in uniquely gendered ways, or not.5 A burgeoning field examines the effects of women’s 1 Cherie Zalaquett Aquea, 2009. Chilenas en Armas. Testimonios e Historia de Mujeres Militares y Guerrilleras Subversivas, Catalonia. 2 United Press International, 1988. “Bodies of two guerrillas found in Chile;” Senate Committee on Foreign Relations and House Committee on Foreign Affairs, 2009, Country Reports on Human Rights Practices of 2009. 3 Chile: Declaracion De Las Mujeres Ex Presas Politicas Bajo La Dictadura. 2004., n.p. Original Spanish: “Las mujeres militantes y resistentes no aceptamos ser tratadas en nuestro compromiso político, como "las mujeres o esposas de …” 4 See, e.g.: Cynthia Enloe, Bananas, Beaches and Bases: Making Feminist Sense of International Politics, Pandora; J. Ann Tickner, 1992. Gender in International Relations: Feminist Perspectives on Achieving International Security, Columbia University Press. 5 Miranda Alison, 2003. “Cogs in the Wheel? Women in the Liberation Tigers of Tamil Eelam,” Civil Wars 6(4): 37-54; ____, 2004. “Women as Agents of Political Violence: Gendering Security,” Security Dialogue 35(4): 447-463; Laura Sjoberg and, Caron 2007. Gentry, Mothers, Monsters, and Whores, Zed Books; Margaret Gonzalez-Perez, 2008. Women and Terrorism: Female Activity in Domestic and International Terror Groups, Routledge; Jocelyn Viterna, 2013. Women in War: The Micro- Processes of Mobilization in El Salvador. Oxford University Press; Sarah Elizabeth Parkinson, 2013; “Organizing Rebellion: Rethinking High-Risk Mobilization and Social Networks in War,” American Political Science Review 107(3): 418-432; Alexis Henshaw, 2017. Why Women Rebel: Understanding Women's Participation in Armed Rebel Groups, Routledge; Meredith Loken 2 participation on rebel behavior. Most of this work focuses on how women’s combat participation affects rebels’ use of violence, including civilian targeting and the tactical effectiveness of female fighters.6 In this dissertation, I suggest that the field’s over-focus on female combatants and rebel tactics obscures the gendered dynamics of women’s militancy that can shape conflict trajectory. Specifically, I argue that women’s participation in front-line, auxiliary, and leadership roles offer insurgencies a strategic advantage in their engagement with the domestic public. I demonstrate that rebels believe that women’s militancy can make civilians more receptive to their grievances and that they recognize the importance of civilian support to tangible conflict successes. I further contend that women-inclusive groups do experience more conflict-related success, specifically territorial control during conflict and some favorable conflict outcomes. Studies of civil conflict conclude that civilian support is key to rebel success. An important, resultant question is what factors make rebels attractive to civilians. I argue that women’s participation is critical, because women’s militancy can have legitimizing effects. This is for two distinct but interrelated reasons. First, the assumption that women are traditionally non-violent actors influences how observers perceive and respond to militant women. Rebels capitalize on women’s participation in ways that legitimize political violence by invoking gendered narratives, signaling the exigency of struggle and integrating the insurgency into local communities. Second, and relatedly, because women are generally unexpected actors they can increase rebel capabilities on and off the front-lines. Rebels reference women’s successful violence to valorize struggle, but also to increase civilian confidence in their capabilities. Together, increased viability and perceptions of legitimate grievances should make the public more supportive of insurgents. My conclusions are based on a multi-method research design. First, I argue that rebels portray women militants and civilians in their political visuals to legitimize grievances, integrate their struggles into local communities, and increase popular support. I adopt a qualitative strategy that explores rebels’ gendered attempts to generate legitimacy in visual propaganda, and I conduct a detailed case study of Irish Republican Army (IRA) and other republican visuals produced during the Troubles and Anna Zelenz. 2018. “Explaining Extremism: Western Women in Daesh,” European Journal of International Security 3(1): 45-68. 6 Lindsey A. O’Rourke, 2009. “What’s Special About Female Suicide Terrorism?” Security Studies 18(4): 681-718; Dara Kay Cohen, 2013. “Female Combatants and the Perpetration of Violence: Wartime Rape in the Sierra Leone Civil War,” World Politics 65(3): 383-415; Jakana Thomas, 2017. “Wolves in Sheep’s Clothing: The Lethality of Female Suicide Bombers,” unpublished manuscript; Caglayan Baser, 2017. “The Impact of Women Insurgents in Rebel Groups on Rebel Strength and Democratic Sponsors,” paper prepared for the Journeys in World Politics Workshop University of Iowa; Devorah Margolin, 2018. “A Palestinian Woman’s Place in Terrorism: Organized Perpetrators or Individual Actors?: Studies in Conflict & Terrorism 30(10): 912-934. 3 (1968-1998). Through archival research in the Linen Hall Library, the Eileen Hickey Irish Republican History Museum, the International Institute of Social History, and Ulster University’s Conflict Archive on the Internet (CAIN), I collect 532 unique posters and similar materials produced by IRA groups, their political wings, and republican-sympathetic organizations in the United Kingdom and Republic of Ireland. I supplement my primary case study with a secondary, and less systematic, study of Palestinian political posters produced between 1960-2016 and primarily collected from the Palestinian Poster Project Archive (PPPA). I collect posters from Palestinian Liberation Organization (PLO) factions, including the Popular Front for the Liberation of Palestine (PFLP), the Democratic Front for the Liberation of Palestine (DFLP) and Fatah, as well as from Hamas. I also analyze relevant, gendered political visuals from Afghan, Namibian, Lebanese, Angolan, South African, Mozambican, Nicaraguan, Bangladeshi, and transnational jihadist militant groups to establish the transportability of gender tropes in conflict propaganda across time and space. I find that rebels of varying affiliation, ideology, and tactics use their visual platforms to shape narratives about gender and women militants in ways intended to build legitimacy for their campaigns. Second, I employ quantitative methods to demonstrate that organizations with women militants experience more conflict-related success, specifically territorial control during conflict and some favorable conflict outcomes. I introduce an original dataset, the Rebel Women Data Project (RWDP), of women’s participation in 146 cross-national insurgencies operating between 1960-2016. These data include ordinal measures of women’s participation in front-line roles, in auxiliary positions, and as leaders. Using these data, I find that women-inclusive rebel organizations are significantly and substantively more likely to control territory than other groups. I further find that female combatants, auxiliaries, and leaders make rebel victory significantly more likely and, in some cases, can increase the odds of peace agreements. I do not draw a direct, causal line between visual propaganda, legitimacy, and rebel successes, but instead emphasize the myriad ways in which women’s participation affects civilian support for insurgency and the effect of these gendered dynamics on conflict trajectory. In this chapter, I detail the concepts that I engage in this project and offer an overview of the arguments and chapters to come. I first interrogate the definitions of women’s militancy and summarize women’s contributions to political violence in front-line, auxiliary, and leadership roles cross-nationally. I then define and explain rebel legitimacy, differentiating between ideological and political legitimacy, and discuss the theorized relationship between legitimacy and civilian support. Finally, I conceptualize visual propaganda and explore how insurgencies can use images during conflict to shape civilian perceptions of their grievances before providing a roadmap for the dissertation. 4 Women militants: a typology Women’s participation in political violence is frequent and varied. I use the term ‘women militants’ to describe women actively contributing to politically violent efforts, irrespective of role. I also adapt Henshaw’s typology of participation types: front-line, auxiliary, leadership, and founders.7 Here, I define each and discuss how women militants fill these roles across and within varied conflicts. Front-line roles are those where combatants regularly engage in the front-line environment in support of the group. It includes the use of or close proximity to violence.8 Women often participate in front-line roles as we conventionally conceptualize them. For example, female Araguaia Guerrillas fought the Brazilian military in the jungle, suffering ambushes and engaging in intimate combat.9 Women composed between 20-30 percent of Revolutionary Armed Forces of Colombia—People's Army (FARC-EP) combatants, often using guerilla tactics against the government.10 A third of fighters in the Tigrayan People’s Liberation Front (TPLF) were women during the Ethiopian Civil War. 11 Berhe notes that women stepped into combat roles even without officially joining the TPLF: “Parallel to the formation of regular fighting forces, a militia army, as a rearguard, was established throughout rural Tigrai… 5-10 % of the militia were women.”12 Women composed approximately 20 percent of participants in the Liberians United for Reconciliation and Democracy (LURD), and 71 percent report their primary position as combat fighter.13 Many of these women operated in an all-female group called the Women’s Artillery Commandos (WAC), which “was known for its fighting capacity and was integrated into LURD military structures.” 14 Similarly, there are currently multiple all-female brigades fighting with the Free Syrian Army (FSA) against Bashar al-Assad’s forces.15 Women also participate on the front-line as suicide attackers. Palestinian militant groups pioneered female suicide bombing in 1975 when Wafa Idris detonated a 22-pound bomb in Jerusalem, killing two and injuring an estimated 100.16 The Taleban claims that in 2011 and 2012 at least three 7 Henshaw, Why Women Rebel. I add the role of ‘founders’ in this analysis. 8 Ibid. 9 Janaína de Almeida Teles, 2017. “The Araguaia Guerrilla War (1972–1974) Armed Resistance to the Brazilian Dictatorship,” Latin American Perspectives 44(5): 30-52. 10 Gonzalez-Perez, Women and Terrorism. 11 Gemma Burgess, 2013. “A Hidden History: Women 's Activism in Ethiopia,” Journal of International Women’s Studies 14(3): 96-107. 12 Aregawi Berhe, 2008. A Political History of the Tigray People’s Liberation Front (1975-1991): Revolt, Ideology and Mobilisation in Ethiopia, Tsehai Publishers. pg. 136. 13 James Pugel, 2006. What the Fighters Say: A Survey of Ex-combatants in Liberia, February-March 2006, UNDP. 14 Irma Specht 2006, Red Shoes: Experiences of Girl-Combatants in Liberia, International Labour Office Geneva, pg. 23. 15 The New York Times, 2013. “Sisters in Arms Join the Fighting in Syria,” April 10. 16 Idris killed herself and one civilian. 5 women carried out martyrdom attacks in Afghanistan and Pakistan.17 The East Turkestan Islamic Movement (ETIM) in China similarly employed three known female suicide bombers.18 At least seven women carried out suicide missions for the Lebanese National Resistance Front (NRF), and Chechen female suicide bombers are responsible for an estimated 80 percent of attacks.19 Chechen women played an important role in the 2002 Moscow theater siege that killed 170 and the 2004 Belsan school siege that killed 334. Female bombers also accounted for 24.7 percent of all Iraqi suicide bombings between 1968-2012.20 Until recently, the Liberation Tigers of Tamil Eelam (LTTE) held the perverse record for using more female suicide bombers - 44 - than any other group. Boko Haram gruesomely wrangled this title: between April 2011 and June 2017 the group deployed at least 244 female bombers, 56 percent of the group’s total suicide attackers.21 While female combatants are a common feature of modern warfare, women often participate in insurgency as auxiliaries. Auxiliary roles, also called support roles, are those where participants identify with the goals, ideology, or effort of the group and routinely offer general support labor. These roles includes, but are not limited to, nurses, cooks, spies, intelligence officers, fundraisers, smugglers, logisticians, or recruiters. Auxiliaries rarely use or are in close proximity to violence.22 The breadth of supportive participation is wide. As acts of anti-Soviet jihad, women in Afghanistan served as couriers carrying supplies, radios, and weapons to mujahedeen at the front.23 Reports implicate women in transporting small arms and smuggling goods for the Palipehutu - National Liberation Forces (FNL) in Burundi.24 In the 1970s, East Timorese resistance fronts similarly established a female organization to assist as auxiliaries and to smuggle and share information between fighters.25 Women represented 35 percent of healthcare staff in the Ethiopian People’s Liberation Front (EPLF).26 Most nurses in the 17 Bill Roggio, 2011. “Taliban use females in recent suicide attacks in Afghanistan and Pakistan,” Long War Journal, June 26; Emma Graham-Harrison, 2012. “Afghanistan suicide blast kills 12 in attack on Kabul airport contractors,” The Guardian, September 18. 18 J. Todd Reed and Diana Raschke, 2010. The ETIM: China's Islamic Militants and the Global Terrorist Threat, Praeger Security International. 19 V. G. Julie Rajan, 2012. Women Suicide Bombers: Narratives of Violence, Routledge; Sjoberg and Gentry, Mothers, Monsters, and Whores. 20 Jessica Davis, “Evolution of the Global Jihad: Female Suicide Bombers in Iraq,” Studies in Conflict and Terrorism 36:4 (2013), p. 279-291. 21 Jason Warner and Hilary Matfess, 2017. Exploding Stereotypes: The Unexpected Operational And Demographic Characteristics of Boko Haram’s Suicide Bombers. Combating Terrorism Center at West Point. 22 Henshaw, Why Women Rebel. 23 Matthew Dearing, 2010. “Like Red Tulips at Springtime: Understanding the Absence of Female Martyrs in Afghanistan,” Studies in Conflict and Terrorism 33(12): 1079-1103. 24 Alexis Henshaw, 2013. “From ‘Followers’ to Leaders: Re-examining the Roles of Women in Armed Rebellion,” E-IR, August 16. 25 Christine Mason, 2005. “Women, Violence and Nonviolent Resistance in East Timor,” Journal of Peace Research 42(6): 737-749. 26 David Pool, 2001. From Guerrillas to Government: Eritrean People’s Liberation Front. Ohio University Press. 6 Western Somali Liberation Front (WSLF) were women, and women medics participated in groups including the Sudan People's Liberation Movement/Army (SPLM/A), the Mizo National Front (MNF), and the Zimbabwe African National Liberation Army (ZANLA).27 WSLF women also played a crucial role in the danta, a traditional dance intended to lift the spirits of combat fighters.28 Many female Mozambique Liberation Front (FRELIMO) participants were intelligence officers and recruiters.29 Indeed, a key role for women was covert ops, as women were highly effective spies. In the Philippines, the Moro Islamic Liberation Front (MILF) operated women’s auxiliary units responsible for fundraising and education.30 Another participatory role is leadership, both political and command. Women take on leadership roles when they exercise direct control over and provide oversight of other participants and/or exercise direct control over the strategy, policies, and/or ideology of the group.31 In the Zapatista Army of National Liberation (EZLN), Mexican women often serve in high-ranking positions and command their own units.32 An estimated 20 percent of Farabundo Martí National Liberation Front (FMLN) military leadership in El Salvador was female, and all of the front’s member organizations had senior female commanders.33 Some female Liberian commanders led mixed-gender groups, while others, like many in the LURD, led women only units.34 The Nepalese People’s Liberation Army (PLA) included women leaders at high levels, though this waned over time.35 In some Naxal areas of India, female commanders outnumber men.36 Women also controlled much of Sendero Luminoso’s leadership in Peru.37 Still, women’s access to these positions are often limited and few rebel groups have widespread 27 Remco Van Hauwermeiren, 2011. The Ogaden War: Somali Women’s Roles and the Psychological Ramifications. Master’s essay, Ghent University. Annette Weber, 2011. “Women Without Arms: Gendered Fighter Construction in Eritrea and Southern Sudan,” International Journal of Conflict and Violence 5(2): 357-370; Northeast Review, 2015. Memoirs of A Mizo Rebel; Tanya Lyons, 2004. Guns and Guerilla Girls: Women in the Zimbabwean National Liberation Struggle, Africa World Pr. 28 Van Hauwermeiren, The Ogaden War. 29 Chris Coulter, Mariam Persson and Mats Utas, 2008. Young Female Fighters in African Wars Conflict and Its Consequences, Nordic Africa Institute, Uppsala. 30 Rosalie Arcala Hall and Joanna Pares Hoare, 2015. “Philippines,” in Jenny Hedström and Thiyumi Senarathna, eds., Women in Conflict and Peace, International Institute for Democracy and Electoral Assistance: 89-123. 31 Henshaw, Why Women Rebel. 32 Gonzalez-Perez, Women and Terrorism. 33 Tommie Sue Montgomery, 1994. Revolution In El Salvador: From Civil Strife To Civil Peace, Second Edition, Westview Press. 34 Specht, Red Shoes. 35 Sudha Ramachandran, 2005. “Dying to be Equal: Women Militants and Organizational Decision Making,” in Farah Faizal and Swarna Rajagopalan, eds., Women, Security, South Asia: A Clearing in the Thicket, Sage Publications: 154-176. 36 Seema Shekhawat, 2015. “Visible in Conflict, Invisible in Peace Positioning Women in the Militancy in Kashmir,” in Seema Shekhawat, ed., Female Combatants in Conflict and Peace: Challenging Gender in Violence and Post-Conflict Reintegration, Palgrave MacMillan: 100-116. 37 Margaret Gonzelez-Perez, 2006. “Guerrilleras in Latin America: Domestic and International Roles,” Journal of Peace Research 43(3): 313-329. 7 female leadership. For example, while women regularly advanced as unit officers in the FARC-EP, but no woman reached the Secretariat, the main decision-making body. Women composed around 30 percent of Argentina’s Montoneros, but no women achieved leadership positions. 38 In India’s People’s War Group (PWG), women make up nearly 35 percent of participants, but leadership roles are highly segregated by caste. Dalit and tribal women, the lowest in India’s caste system, compose rank-and-file cadre but have not yet achieved leadership positions.39 Finally, some women take part as founders of insurgent groups. Founders participate as core members of the group’s inception and initial publicization. In Colombia, women co-founded the 19th of April (M-19) guerrilla movement, which was the second largest insurgency in the country. Alice Auma, a rebel and spirit-medium, started the Holy Spirit Movement in Uganda.40 She even convinced the Ugandan People’s Democratic Army (UPDA), a co-existing rebel group, to give her some of their troops. Ulrike Meinhof, who is perhaps the most famous female militant, co-founded Germany’s Red Army Faction (also known as the Baader-Meinhof Group) with Gudrun Ensslin and a handful of men in 1970. In many groups, women’s participation fluctuates temporally. In some organizations, like the PLA, women initially participate at high levels but their involvement declines over time. Other insurgencies do not officially include women in their initial years but accept them later. The Eritrean Liberation Front (ELF), for example, did not begin allowing women until at least five years into operations.41 Women’s participation may also shift as rebel groups evolve into new partnerships, fracture, or transform. While Myanmar’s Shan United Revolutionary Army (SURA) actively recruited women as rank-and-file soldiers, it lost significant autonomy when it merged to become the Mong Tai Army (MRA). The MTA barred female combatants, and SURA fighters found themselves repositioned as auxiliaries.42 One aim of this project is to better articulate women’s participation in civil conflict by recognizing women militants not only as front-line fighters, but as auxiliaries, leaders, and founders. Rebel women participate for many reasons, and often join voluntarily. Many women in all group-types 38 Karin Grammático, 2010. “Populist Continuities in “Revolutionary” Peronism? A Comparative Analysis of the Gender Discourses of the First Peronism (1946–1955) and the Montoneros,” in Karen Kampwirth, ed. Gender and Populism in Latin America: Passionate Politics, Penn State University Press. 39 Shruba Mukherjee, 2003. Women Cadres of the PWG: Conflict Mongers or Managers, WISCOMP Discussion Paper. 40 Tim Allen, 1991. “Understanding Alice: Uganda’s Holy Spirit Movement in Context,” Africa: Journal of the International African Institute 61(3): 370–99. 41 Jakana Thomas and Kanisha Bond, 2015. “Women’s Participation in Violent Political Organizations,” American Political Science Review 109(3): 488-506. 42 Jane Ferguson, 2013. “Is the Pen Mightier than the AK-47? Tracking Shan Women's Militancy Within and Beyond, Intersections,” Gender and Sexuality in Asia and the Pacific 33: 1-12. 8 are proud of their participation in rebellion and resistance or view the insurgent group as a vehicle to escape poverty, violence, oppression, or social stagnation. For example, FARC-EP women often report a “sense of accomplishment and an opportunity to play a successful role in… a successful enterprise.”43 FRELIMO women perceived themselves as freedom fighters, “motivated by a sense of freedom and camaraderie.”44 Many women who voluntarily joined Boko Haram in support roles view life in the terrorist organization as better than life living under the government. One motivating factor is that in Boko Haram, bride price is paid directly to a woman instead of to her family.45 Women’s militancy, and the narratives surrounding it, may inspire other women to join or support rebel organizations. Consequently, women often dictate the terms of their political violence. However, I am also explicitly interested in how rebels perceive women militants and how they leverage their expectations of how civilians will perceive women militants to build legitimacy among their domestic populations.46 I suggest that rebels believe women’s participation shapes civilian perception of insurgent grievances and consequently exploit female militancy in gendered ways that resonate with their local publics. Demonstrating how rebels cross-nationally capitalize on women’s participation to legitimize their struggle does not mean that rebels do not truly want, appreciate, or benefit from female militancy. Women’s participation and the representation of women’s participation are separate, though related, phenomena. I analyze this representation in rebel-produced posters and similar materials, which I refer to collectively as ‘visual propaganda.’ The following section interrogates the concept of rebel legitimacy, discusses the relationship between gender and legitimacy-building during conflict, offers a brief introduction to rebel propaganda, and defines the boundaries of this terminology. Legitimacy and rebel propaganda Legitimacy is a multifarious concept that no single definition can fully capture. States, rebels, and other conflict actors experience varying degrees of legitimacy that shift throughout periods of unrest. I adopt broad definitions of legitimacy that leave room for these intricacies. Specifically, I identify two types 43 Natalia Hererra and Douglas Porch, 2008. “‘Like Going to a Fiesta-’ the Role of Female Fighters in Colombia’s FARC-EP,” Small Wars & Insurgencies 19(4): pg. 612. 44 Coulter, Persson and Utas, Young Female Fighters in African Wars Conflict and Its Consequences, pg. 12. 45 Hilary Matfess, 2017. Women and the War on Boko Haram: Wives, Weapons, Witnesses, Zed Books. 46 Rebels are also concerned with international perception and often launch propaganda initiatives aimed at improving their reputation among other organizations with similar aims (such as anti-colonial groups) or sympathetic state governments. The influence of these external actors may certainly play a role in conflict outcomes. As I discuss in Chapter 5, I do explore the possibility of external support as an explanation for conflict successes, but I do not extend my analysis of legitimacy and propaganda to external actors. The international audience, while theoretically worth exploration, is beyond the scope of this project. 9 of rebel legitimacy: ideological and political. I define ideological legitimacy as, to quote Suchman, the “perception or assumption that something is right, proper, or appropriate within the bounds of a system of norms, values, or beliefs.”47 Ideological legitimacy is concerned with “moral validity,” ‘rightness,’ or salience with existing social, political, and religious narratives.48 I follow Haung in defining political legitimacy as “the idea that a group can be a viable political entity.”49 Political legitimacy is concerned with “belief in the validity of the structure and norms and, thus, the roles of the authorities operating in such structures.”50 These types of legitimacy are not sufficient conditions of one another - civilians need not believe that rebels are right or appropriate to view them as politically viable, and vice versa - but they are also not mutually exclusive. Viewing rebels as ideologically appropriate can certainly make rebels’ governance and structure appear more compelling. Conversely, political legitimacy may enhance ideological legitimacy by signaling the practical effects of rebels’ sociopolitical beliefs and actions. Scholars are most concerned with political legitimacy, as evidenced by the breadth of recent research on rebel governance. For example, Stewart demonstrates that rebels often provide service provisions in the form of food, medical care, religious services, or education to present themselves as a viable alternative to the state.51 Haung calls attention to rebel diplomacy during civil war, emphasizing that rebel groups often contact foreign media, appeal to international organizations, and expand their overseas networks to influence international perceptions and gain political capital.52 Wood concludes that when rebels rely on local support, gains in group capabilities decrease rebel-wielded violence against civilians.53 In contrast, when rebels rely on alternative sources of support, gains in group capabilities increase rebel-wielded violence against civilians. This suggests that when rebels care about civilian perceptions, they try to maintain some aspects of political legitimacy. Less attention is paid to ideological legitimacy, but this type of public perception can be critically important for rebels. As Washburne argues, legitimacy is “promoted through a system designed to persuade the followers of the appropriateness of the regime,” which is done by creating 47 Mark C. Suchman, 1995. “Managing Legitimacy: Strategic and Institutional Approaches,” The Academy of Management Review 20(3): pg. 574. 48 Sarah Lykes Washburne, 2010. Legitimacy, Identity and Conflict: The Struggle for Political Authority in Southern Sudan, 2005- 2010. Ph.D diss. University of Exter, pg. 20. 49 Reyko Huang, 2016. “Rebel Diplomacy in Civil War,” International Security 40(4): pg. 99. 50 Washburne, Legitimacy, Identity and Conflict, pg. 21. 51 Megan A. Stewart, 2018. “Civil War as State-Making: Strategic Governance in Civil War,” International Organization 72(1): 205–226. 52 Huang, “Rebel Diplomacy in Civil War. 53 Reed M. Wood, 2014. “Opportunities to Kill or Incentives for Restraint? Rebel Capabilities, the Origins of Support, and Civilian Victimization in Civil War,” Conflict Management and Peace Science 31(5): 461–80. 10 faith in the regime.54 While I do argue that women’s participation can increase political legitimacy, I am primarily interested in the application of ideological legitimacy frameworks to rebel organizations. Specifically, I argue that rebels exploit gendered beliefs about women and women’s militancy to persuade civilians of the appropriateness and necessity of their grievances. Entrenched gender beliefs about female passivity create the flawed assumption that women are non-traditional conflict actors. Consequently, women’s participation can signal the direness of struggle and symbolize a movement’s broad, emancipatory objectives. I suggest that rebels use women’s inclusion and victimization as evidence of the need for rebellion. Conflict actors use highly politicized gender beliefs to build ideological legitimacy. Because ‘civilian’ is a fundamentally gendered category, states and non-governmental organizations often exploit beliefs about female victimhood to justify militarism.55 For example, I argue elsewhere in work with Lake and Cronin-Furman that the Sri Lankan government used “selective condemnation of wartime rape is a strategic legitimacy-building practice adopted for military ends” during the Sri Lankan Civil War.56 We suggest that the salience of beliefs about female militancy and victimhood – which permeate most societies – make women’s bodies particularly exploitable for legitimacy-gains.57 Viterna similarly finds that the FMLN recruited women on the basis of protecting them from state- perpetrated sexual violence.58 This platform painted the state as monstrous and established moral authority for the FMLN. As I explain in Chapter 2, rebels regularly invoke women in particularly gendered ways that appeal to their publics by signaling the exigency of circumstances. Public outreach is a primary method for violent non-state actors to build legitimacy, and thus in this project I explore rebels’ use of gender as an influential tool through their visual propaganda. ‘Propaganda,’ as a term, is controversial. Maasri highlights a divide in the literature on political visuals that demarcates some strategies as ‘propaganda,’ and others as ‘activism.’ The former is derogatory, signaling manipulation, corruption, and deceit. The latter is participatory, a type of civil protest rallying commitment for a cause. Maasri critiques this binary, noting “I remain skeptical of the bipolar distinction between propaganda and activism. This polarity removes possibilities of envisioning propaganda strategies as activist quests and, conversely, disregards the emancipatory aims 54 Washburne, Legitimacy, Identity and Conflict, pg. 19. 55 See, e.g. Lila Abu-Lughod, 2013. Do Muslim Women Need Saving? Harvard University Press for a discussion of Western foreign interventions in the Middle East justified by the emancipation of Muslim women. 56 Meredith Loken, Milli Lake, and Kate Cronin-Furman, ““Deploying Justice: Strategic Accountability for Wartime Sexual Violence,” forthcoming, International Studies Quarterly, pg. 4. 57 Ibid. 58 Viterna, Women in War. 11 of propaganda. … [it] fails to explain the complex political role political posters play in the case of a civil war.”59 She argues against Jowett and O’Donnell’s widely employed definition of propaganda, as the “deliberate, systematic attempt to shape perceptions, manipulate cognitions, and direct behavior to achieve a response that furthers the desired intent of the propagandist.”60 Writing on the Lebanese Civil War, Maasri contends, instead, that wartime political posters are “sites of hegemonic struggle” on the part of actors attempting to “construct a ‘regime of truth’ … to win the consent of their own communities and maintain dominance.”61 This approach accepts that the communication, relationship, and transfer of ideas from propagandist to audience is not a linear process, but is instead a cumulative endeavor involving parties who exist in the same society and under the same power structures. It “cannot be reduced to a mechanical process of transmission of manipulative messages.”62 Maasri’s approach upends the myth of audience passivity in this type of communication. I agree with Maasri, and I apply this approach explicitly to rebel visuals, which I refer to in this project as ‘propaganda.’63 While I appreciate the shift away from propaganda as a concept in Maasri’s work, I recognize the benefit of this terminology to the discussion of rebel images. There are different types of rebel visuals. Groups produce their own posters that directly reference their struggle, they produce posters in solidarity with other organizations, they publish leaflets with information about rallies and protests, and they circulate materials produced professionally by their allies abroad. Consequently, I find it useful to demarcate certain types of rebel visuals as ‘propaganda’ while reclaiming the term from its derogatory connotation. In Chapter 3, I discuss at length how I identify ‘propaganda’ posters in my individual cases. Here, I briefly explain my underlying approach. Propaganda attempts to influence the observer to benefit its sponsor. This identifiable motive separates propaganda from other forms of communication, like information or advertising.64 Propaganda also attempts to shape the discourses, desires, and fears of political communities during conflict.65 I suggest that visual propaganda uses incomplete information and selected symbols to tap into salient beliefs and influence audience perceptions. Specifically, in this project I argue that rebel propaganda exploits gender beliefs and, often, misrepresents women’s participation to influence its audience. Propagandists need only feign, not show, to be effective. Thus, propaganda is not truly 59 Zeina Maasri, 2009. Off the Wall: Political Posters of the Lebanese Civil War. I.B. Tauris, pgs. 5-6. 60 Garth Jowett and Victoria O’Donnell, 1999. Propaganda and Persuasion, Third Edition. Sage Publications, pg. 6. 61 Maasri, Off the Wall, pg. 15. 62 Ibid, pg. 9. 63 I use the terms ‘visuals’ and ‘images’ interchangeably. 64 Joanne Wright, 1991. Terrorist Propaganda: The Red Army Faction and the Provisional IRA, 1968-86, Palgrave Macmillan. 65 Maasri, Off the Wall pg. 17 uses the ‘discourses, desires, and fears’ terminology but applies it to “hegemonic articulation,” not “propaganda.” 12 ‘informative.’ Images are a particularly compelling form of propaganda because of their visceral, emotive qualities and the potential for circulation. Herfried Münkler argues that “those who have no capacity to attack the conventional forces of states with any chance of success seek to disseminate images in which the consequences of the acts of violence are made directly visible.”66 As I discuss in Chapters 2, 3, and 4, images are highly intentional and aim to provoke emotional investment from both sympathetic and uncommitted audiences. Rebel visuals may aim to increase ideological legitimacy among some audiences, but they may also intend to sow fear, or political legitimacy among others. One example is the photographs of a beheaded James Foley that Daesh circulated online in 2014. In this project, I am concerned primarily with visuals in the former category: those produced by rebel organizations aimed at positively influencing domestic audiences’ perceptions of violent non-state struggle. Because “arenas where insurgents launch attacks comprise neither passive audiences nor even playing-fields,” images offer rebels a ground-up way to improve their outcomes by fighting for ‘hearts and minds.’67 Despite the relevance of visual propaganda in asymmetric warfare, conflict researchers pay surprisingly little attention to insurgent images. I explore this neglect in Chapters 2, 3, and 4 and I focus my research on rebel posters and other one-dimensional propaganda visuals. Before the internet enveloped insurgent messaging, political posters were a dominant vehicle for conveying rebel platforms to diverse audiences. In many cases, they still are. The rise of social media has changed propaganda’s method and circulation, but the intentions remain the same. As the political visual shifts from lampposts to Twitter feeds, it is critical that we understand the salience of these images as a tool of legitimacy-building for violent non-state actors. I suggest that these visuals are, themselves, weapons of war. I contend that political images are unique vehicles for shaping public perceptions, and that rebels regularly visually mobilize gender to meet this objective. Outline of the dissertation The rest of the dissertation proceeds as follows. Chapter 2 introduces my theory that rebels use women’s participation to build ideological and political legitimacy and that female-inclusive groups should experience more success. I draw on gender studies, sociology, political science, and history to build this framework. I argue that rebels view civilian support as important to conflict achievements 66 Herfried Münkler, 2005. The New Wars. Trans. by Patrick Camiller, Cambridge University Press, pg. 111. 67 Neville Bolt, 2012. The Violent Image: Insurgent Propaganda and the New Revolutionaries, Columbia University Press, pg. 40. 13 and believe that women’s participation positively mobilizes public perceptions of their grievances, conditional on how that participation is represented. I further theorize how legitimacy may improve the likelihood of conflict successes, operationalized as territorial control and favorable conflict outcomes. I review alternative explanations for rebel success in these two areas and conclude the chapter with a series of observable implications from my theory and competing explanations. I emphasize that complex processes explain rebel trajectory and that no singular explanation can account for all outcomes. In Chapter 3, I introduce my qualitative, interpretivist research design for analyzing rebel propaganda. I meld art, art history, feminist, and traditional security studies literatures to adopt an iconological approach. I demonstrate that rebels visually securitize certain gender expectations to legitimate political violence. This chapter then presents my case study of Irish Republican Army (IRA) propaganda. I offer a brief history republican insurgency in the Republic of Ireland and Northern Ireland and discuss women’s participation in the IRA. I discuss republicans’ propaganda production during the Troubles and dig descriptively into my dataset of political visuals to identify thematic trends. I argue that republican propaganda securitized gender, and women’s militant participation, to construct narratives of threat and legitimate the IRA’s fight against the British government. Specifically, I find that depictions of female militants and nationalist civilians constitute a substantive part of the rebels’ propaganda campaign. Chapter 4 extends this analysis beyond the IRA to demonstrate that the group is not unique in its exploitation of gender as a security device. I find that the same underlying ideas – or literal visual reproductions of them – function as legitimacy-builders for rebels cross-nationally. To demonstrate the transportability of gender tropes in conflict propaganda across rebel organizations, I offer substantive but less systematic evidence from armed factions in the Palestinian-Israeli conflict and several other insurgent groups. These cases illustrate the strategic value that rebels see in promoting controlled narratives about women militants and civilians in their visual propaganda. I suggest that rebels clearly believe women militants influence observer opinions about political violence, and I demonstrate how these organizations capitalize on the gendered dynamics of women’s wartime experiences to appeal to noncombatants. Chapter 5 moves beyond case studies into cross-national analyses. Here, I introduce an original dataset of women’s front-line, auxiliary, and leadership participation in global insurgencies. I review existing efforts to systematically assess female militancy in internal conflicts and explore the ethical dilemmas of quantifying women’s wartime participation. I explain my data collection process, weigh 14 the benefits and limitations of this approach, and offer descriptive findings concerning the scope, scale, and geography of how women participate in rebel organizations. In Chapter 6, I use my original dataset in a series of statistical analyses to examine the relationship between women’s participation in rebel organizations to success during conflict. Specifically, I explore the effect of female militancy on rebel territorial control. I discuss the conceptualization and measurement of territorial control and evaluate competing hypotheses for why some rebels achieve this feat while others fail. I show that rebel groups with female combatants, auxiliaries, and leaders have higher probabilities of controlling territory than other groups, and I demonstrate that the level of women’s participation substantively affects this relationship. I supplement these analyses with a case illustration of territorial control among Peru’s leftist insurgencies. I use this study to draw out the theoretical dynamics of gender, legitimacy, and rebel territorial control during conflict. I focus on the relationship between women’s participation in rebel organization on success in conflict outcomes in Chapter 7. I assess the likelihood of favorable conflict outcomes, namely peace agreements and rebel victories, relative to the likelihood of government defeat. I find that women’s participation often increases the odds of a peace negotiation, but that this finding is not statistically significant. I further find that regardless of role, women’s participation significantly increases the likelihood of rebel victory. To explore the potential legitimacy mechanisms working in these cases, I employ a qualitative assessment of Uganda’s National Resistance Army (NRA). I conclude in Chapter 8 by discussing my conclusions and the implications of my findings. I articulate a need for further research on women participants in civil conflict, especially women who contribute off the front-line. I imagine future possibilities for research in this field and for practitioners of peace-building in adopting gender-cognizant approaches in their programming. 15 Chapter 2: Theorizing Gender and Rebel Legitimacy What explains variation in rebel successes during conflict? What role does civilian support play? How does women’s participation affect these civil conflict dynamics? Scholars credit factors including shared ethnicity, religion, ideology, material benefits, service provision, and the use of violence as shaping critical civilian support for insurgencies.68 However, despite women participating in over half of global rebellions since 1960, few studies explore how the gender of insurgents helps rebels win civilian support.69 I argue women’s militancy can help legitimize rebel organizations. First, the expectation women are naturally non-violent and peripheral to conflict makes their participation appear anomalous. Rebels represent, and mis-represent, women’s participation in ways that ideologically legitimize political violence by capitalizing on gendered narratives, humanizing insurgents and emphasizing the urgency of their grievances. I contend that rebels exploit gender beliefs to frame women’s inclusion as evidence of the direness of struggle and of the integration of insurgency into local communities. Second, but relatedly, because women are generally unexpected actors they contribute to rebel strength in front-line, auxiliary, and leadership roles. Rebels call attention to women’s successful violence to signal the ‘rightness’ of their campaigns, but also to build public confidence in their capabilities. Women can be highly effective because gender beliefs make them relatively unanticipated attackers, and they demonstrate strategic acumen as auxiliaries integrated into local social networks. This increased capability should make noncombatants view rebels as more politically viable and, consequently, make them more supportive of insurgents. It also shapes the gendered narratives rebels use to defend and legitimize their violence. Thus, women’s participation can be both ideologically and politically legitimizing. I contend that rebels view civilian support as essential to achievements during conflict and believe that women’s participation, and representations of women’s participation, augment this support. Moreover, I conclude that women’s participation can indeed improve the likelihood of rebel successes. The effects of ideological and political legitimacy are not mutually exclusive, and they work 68 See: Laia Balcells and Stathis Kalyvas, 2015. “Revolutionary Rebels and the Marxist Paradox,” manuscript; Mia Bloom, 2004. “Palestinian Suicide Bombing,” Political Science Quarterly 119(1): 61-88; Lindsay Heger and Danielle Jung, 2017. “Negotiating with Rebels: The Effect of Rebel Service Provisions on Conflict Negotiations,” Journal of Conflict Resolution 61(6): 1203-1229; Macartan Humphreys and Jeremy M. Weinstein, 2008. ‘‘Who Fights? The Determinants of Participation in Civil War,’’ American Journal of Political Science 52 (2): 436-55; Kentaro Hirose, Kosuke Imai, and Jason Lyall, “Can Civilian Attitudes Predict Insurgent Violence?” Journal of Peace Research 54(1): 47-63; Janet Lewis, 2016. “How Does Ethnic Rebellion Start?” Comparative Political Studies 50(10): 1420-1450. 69 A notable exception is Reed M. Wood, Jakana Thomas, and Devorah Manekin, 2016. “Female Combatants and the Outcome of Civil Conflict,” manuscript. 16 symbiotically. Their net effect is that female-inclusive organizations are, across multiple measures, more successful during conflict than other groups. This suggests that rebels’ civilian-engagement strategies and the inclusion of women on and off the front-line have tangible benefits. This chapter covers a great deal of theoretical ground, which I present in four parts. I first explain the role of civilian support in civil conflicts, explicating the importance of civilian support for rebel successes and analyzing the relationship between legitimacy and support. Second, I introduce the underlying gendered concepts that ground this work, specifically discussing how gendered expectations shape observer opinion of politically violent women. I theorize the gendered dynamics of women’s participation and rebel-civilian engagement, arguing that gender can fundamentally shape organizational strategies and rebels’ engagement with constituent audiences to advance their cause. Third, I draw out the argument that rebels believe women’s participation helps build legitimacy by exploring how insurgents may represent gender in visual propaganda to encourage civilian support. I generate a series of expectations for how rebels will visualize women militants and gender more broadly in their propaganda based on this theoretical framework. Finally, I explore the tangible effects of women’s participation on rebel organizations by theorizing how women’s participation may increase legitimacy and make insurgent successes more likely. I specifically apply this framework to two areas of conflict-related success - territorial control and favorable conflict outcomes - and explore the observable implications. I review and derive testable hypotheses from alternative explanations for these achievements, and discuss the possible inter- relational effects of my argument with these theories. The role of civilian support in rebel success A widely held premise in research on civil conflicts is that rebels generally intend to mobilize a population against the government. As Wood concludes, “the distribution of civilian loyalty potentially shapes war outcomes.”70 Civilians provide food, shelter, weapons, money, recruitment pools, and information to insurgents. Sarbahi finds in a study of 166 civil wars that rebel groups deeply embedded into local populations “are better able to withstand a disproportionately powerful state.”71 Hazleton argues that the most effective counterinsurgencies break resource flows from noncombatants, often 70 Reed M. Wood, 2010. “Rebel Capability and Strategic Violence Against Civilians,” Journal of Peace Research 47(5): pg. 602. 71 Anoop K. Sarbahi, “Insurgent-Population Ties and the Variation in the Trajectory of Peripheral Civil Wars,” Comparative Political Studies 47(10): pg. 1474. 17 through brute force.72 However, the conceptualization of ‘civilian support’ is complex, and the distribution of such support is quite varied. In civil conflicts “the fight is conducted through the people” but most civilians do not actually ‘choose’ to decisively support one side over another.73 Many civilians do share ideological and political beliefs with insurgents. However, as Kalyvas writes, “[m]ost ordinary people appear to display a combination of weak preferences and opportunism, both of which are subject to survival considerations. Their association with risk-taking minorities tends to be loose and subject to the fortunes of the war and its impact on one’s welfare.”74 The ‘loyalty’ behaviors we commonly subsume under the umbrella of ‘support’ (joining, providing resources, recruiting, sharing information for insurgents) can result from ideological salience but can just as easily result from fear, criminal or greed-based incentives, survival incentives, or adventure-seeking. Civilians may also be ideologically or politically supportive of rebels in some areas while repudiating them in others. For example, Northern Irish Catholics yielded divided opinions about IRA violence against security forces as opposed to civilian targets.75 Thus, ‘civilian support’ and its material benefits are not always or entirely about endorsement. A useful approach to conceptualizing civilian support is to shift the focus away from what civilians want from rebels and onto what rebels need from civilians. Insurgents need civilians to make conflict gains, and thus their primary goals are to avoid what Kalyvas terms ‘defection:’ non- compliance, side-switching, and informing.76 I theorize ‘civilian support’ as rebels successfully preventing defection. Non-compliance jeopardizes recruitment, funding, and material resources. Side- switching clearly strengthens insurgents’ opposition. Informing is a genuine threat to rebel success, and thus information is a critical factor in insurgent-civilian relationships. Rebels rely on noncombatants to gather and share useful intelligence, but they also need civilians to keep information about rebel activity from the government. Insurgents consequently go to great lengths to control the spread of information by engendering civilian support. Insurgents who cannot win civilian support may try to compel it through violence. However, violence is risky. Jones finds that this strategy decreases a group’s odds of victory because civilians come to resent or fear the insurgency more than they dislike or feel threatened by the government. 77 72 Jacqueline Hazelton, 2017. “The ‘Hearts and Minds’ Fallacy: Violence, Coercion, and Success in Counterinsurgency Warfare,” International Security 42(1): 80-113. 73 Stathis N. Kalyvas, 2006. The Logic of Violence in Civil War, Cambridge University Press, pg. 91. 74 Ibid, pg. 103. 75 James Bennet, 1991. “I.R.A. Supporters Fight Over One Man’s Opinion,” New York Times, 13 December. 76 Kalyvas, The Logic of Violence in Civil War. 77 Seth G. Jones, 2016. Waging Insurgent Warfare: Lessons from the Vietcong to the Islamic State. Oxford University Press, pg. 47. 18 Brutalized civilians may be more receptive to counterinsurgent campaigns because they do not feel that rebels serve their interests. For example, Wright and colleagues find that civilians in Afghanistan respond to abuse by providing intelligence to security forces.78 Thus, rebels try to convince civilians to support their objectives through legitimacy-building enterprises. This helps explains why many groups operate such substantive propaganda campaigns and deliver service provisions in target areas. While there is widespread agreement that civilian support matters for insurgency, scholars generally overlook gender as a mechanism of rebel appeal. I address this neglect by explicating the importance of legitimacy in explaining civilian support for rebels. I suggest that civilians may be most supportive of women-inclusive insurgencies because these rebels can generate unique ideological and political legitimacy. As I discuss in Chapter 1, civilians may find rebels politically legitimate because they view them as viable as immediate or long-term alternatives to the states. Civilians may also find rebels ideologically legitimate because they view them as fitting appropriately into social or ideological community norms. I argue that rebels understand this, and that they invoke women strategically to further these objectives. Insurgent groups do this by representing women in explicitly gendered ways in their propaganda and through in-person engagement with local communities. Thinking through the legitimacy aspects of civilian support helps tease apart why civilians fund, protect, contribute to, and join insurgencies. In the following sections, I argue that women’s participation helps situate rebels within existing normative social frameworks and this contributes to their overall ability to contest state authority. These effects legitimize rebels to civilians, resulting in civilian support for insurgents. Women, legitimacy, and political violence Deeply held gender beliefs about female passivity and non-violence facilitate the expectation that women are peripheral to conflict. Scholars of gender and international relations stress the ubiquity of these beliefs: though they may vary in strength, these ideas are nearly universal.79 Gender is “a matter of social processes and structures that have been created and sustained over generations- sometimes coercively.”80 Consequently, gender beliefs structure society at all levels. The belief that women are comparatively passive shapes expectations of their involvement in political violence. As Carpenter 78 Austin L. Wright, Luke N. Condra, Jacob N. Shapiro, and Andrew C. Shaver. 2017. “Civilian Abuse and Wartime Informing,” unpublished manuscript. 79 Viterna, Women in War; Temitope Oriola, 2016. “‘I Acted like a Man’: Exploring Female Ex-Insurgents’ Narratives on Nigeria’s Oil Insurgency,” Review of African Political Economy 43(149): 451–456. 80 Enloe, Bananas, Beaches, and Bases, p. 6. 19 notes, the disproportionate application of non-combatant status to women perpetuates the idea that women are vulnerable and incapable of violent action.81 While little work explores how gender affects civilian perceptions during war, other literatures examine how these beliefs shape attitudes towards women’s political participation. For example, research on criminal courts shows that legal systems often treat women’s violence as aberrant. Domestic courts often hold female perpetrators less responsible for offenses.82 In one experimental study, both men and women were more likely to justify a violent crime, assign more lenient jail time, and have a more positive perception of and sympathy for a suspect when the suspect was female.83 In a 2015 Parliament meeting, the Anti-Terror Chief of London’s Metropolitan Police offered immunity to three women who left for Syria to join Daesh in February 2015. He notes that the police have no evidence that these women are responsible for terrorist offenses despite joining the organization.84 This marks the first time the police have offered immunity from prosecution to a returnee from Daesh and is a stark divergence from increasingly harsh punishments for male recruits from Western countries.85 An immunity deal suggests that these women are not real extremists and thus are less dangerous domestically than their male counterparts. In another example, government records show that only 10 percent of those arrested for their part in the 2011 London riots were women. Interview data suggests that women composed at least 20 percent of the rioters but were less likely to be arrested by police.86 Research on female political leaders similarly demonstrates that observers perceive them to be comparatively peace-prone. Dube and Harish exploit hereditary succession in European politics over the 15th-20th centuries and find that among unmarried monarchs, queens were more likely to be attacked than kings. This is not because these queens were weaker or less prone to violence: among married monarchs, queens were 27 percent more likely to wage wars than their husbands. Instead, they note, “unmarried queens, specifically, may have been perceived as weak and attacked by others” (emphasis added).87 Research on Julia Gillard, Australia’s first Prime Minister, emphasizes how she 81 R. Charli Carpenter, 2005. “‘Women, Children and Other Vulnerable Groups’: Gender, Strategic Frames and the Protection of Civilians as a Transnational Issue,” International Studies Quarterly 49(2): 295–334. 82 Ester E.J Strømmen, 2017. Jihadi Brides or Female Foreign Fighters? Women in Da’esh – from Recruitment to Sentencing. PRIO Centre on Gender, Peace, and Security Policy Brief. 83 Osei Appiah et. al. 2017. “Sugar and Spice, and Everything Nice: Do Female Stereotypes Supersede In-group Favoritism Among Men When Evaluating Female Criminal Suspects in News Stories?” Journalism and Mass Communication 7(4); 165-186. 84 Loken and Zelenz, “Explaining Extremism.” 85 Ibid. 86 James Ball, Matthew Taylor, and Tim Newburn, 2011. “Who Were the Rioters?” The Guardian, 4 December. 87 Oeindrila Dube and S.P Harish, 2017. “Queens,” NBER Working Paper, pg. 3. 20 struggled to maintain enough “femininity,” “warmth,” and “nurturance” to be liked by the Australian public.88 Hall and Donaghue note the public’s dissatisfaction with Gillard’s political calculations, arguing that her ambition was “potentially shocking or disappointing in the face of expectations that voters might have had for a female prime minister.”89 One commenter scolded Gillard, remarking that “nice girls don’t carry knives.”90 An important conclusion is that people fail to update gender beliefs when faced with disconfirming evidence of women’s capacity for violence. Beliefs are deeply engrained, and gender is a principal identity category that structures interpersonal interactions. Mercer contends, “Rationalists suggest that one should accommodate (or update) beliefs by using new information to revise those beliefs… However, if beliefs, expectations, or theories influence interpretations of evidence, then one assimilates data to beliefs. Overstating the power of this assimilation process is difficult.”91 Observers most often interpret women’s violence as anomalous. Importantly, women do commit, on average, fewer violent crimes than men. These beliefs may consequently be founded, but they rely on a stereotype that reduces the perceived probability of female violence to zero. People can fall back on the knowledge that men disproportionately commit violence, which leads to an even stronger expectation that they will assimilate, rather than update, gender beliefs when confronted with violent women. As Sjoberg and Gentry contend, “Women’s [political] violence is often discussed in terms of violent women’s gender: women are not supposed to be violent.”92 I suggest that the obstinacy of these beliefs carry into conflict. To reconcile the belief in women’s passivity with women’s violence, civilians often attribute women’s brutality to a temporary lapse in gender norms and responsibilities. For example, Dr. Ezechiel Sentama of the University of Rwanda Centre for Conflict Management notes of female Hutu genocidaires, “I was able to see it was really bad and women, I could say that they forgot their role as far as the Rwandan culture is, but actually the role, the role of woman we know. A woman is a mother. She is somebody who seems to have much more compassion than men, you know. So they forgot their role of women.”93 Schulze concludes that during the Lebanese Civil War, “society tolerated female ‘warriors’ since they were 88 L.J Hall and N. Donaghue, “‘Nice Girls Don’t Carry Knives: Constructions of Ambition in Media Coverage of Australia's First Female Prime Minister,” Journal of Social Psychology 52(4): 631-647. 89 Ibid, pg. 638. 90 Ibid. 91 Jonathan Mercer, 2010. “Emotional Beliefs,”. International Organization 64(1): pg. 8. 92 Sjoberg and Gentry, Mothers, Monsters, and Whores, pg. 2. 93 Ibid. 21 perceived as temporary.”94 They accepted short-term deviations from ‘appropriate’ gender norms because of the exigency of circumstances. These gender beliefs preserve the expectation that women are non-traditional conflict actors despite disconfirming evidence. They produce a one-dimensional image of women that emphasizes their victimization as a natural, gendered condition. Insurgents can capitalize on the gendered dynamics of women’s participation to build legitimacy in several ways. The underlying thread between them is the use of and construction of narratives around female militants to shape civilians’ threat perception. Rebels can invoke women’s militancy as a sign of righteousness and a signal of the necessity for political violence. Importantly, rebel organizations vary in socialization and the gender expectations of their target public. Thus, what matters is not only women’s militancy, but the roles women take on within rebel organizations. Unsurprisingly, and as I discuss at length in Chapter 5, rebel organizations generally include women in roles and at levels consistent with the gender beliefs of their populations. Still, insurgent groups across group-types mobilize the same underlying norms of victimhood and the expectation that women are peripheral to conflict to legitimize anti-state violence. For example, conservative groups call for a return to traditional gender roles that the state has disrupted, especially by including women in auxiliary positions. Groups like Dukhtarane-Millat (DeM) in India and Daesh recruit Muslim women by stressing the need to return to conservative, patriarchal values. Daesh, like other movements, aims to present an alternative vision of society. These groups link their jihad to the stifling of ‘appropriate’ gender performance by the West, and use women to demonstrate a commitment to piety and proper family roles.95 In Daesh, women’s roles as mothers and wives for future martyrs is indispensable to the organization because women’s primary job is to nurture and grow the caliphate. Daesh is subsequently attractive to women who feel like society has forced them out of traditional roles and to men who lament this perceived emasculation.96 Female Daesh members also recruited other women by invoking the group’s protectionism. One recruit recalls feeling unsafe walking home in Canada, but says, “Now I walk home… Surrounded by Mujahideen [fighters], knowing no-one can harm me. What a difference.”97 Another similarly notes, “Dawla [Daesh] has the best types of men, and the shariah [Islamic] court protects you from all types of 94 Kirsten E. Schulze, 1998. “Communal Violence, Civil War and Foreign Occupation: Women in Lebanon,” in Robert E. Miller and Rick Wilford, Women, Ethnicity and Nationalism: the Politics of Transition. Routledge, pg. 157. 95 Loken and Zelenz, “Explaining Extremism.” 96 Ibid. 97 Bint Muhajid, Twitter (17 July 2015), quoted in Loken and Zelenz, “Explaining Extremism.” 22 violence.”98 Creating space for women, even superficially, demonstrates the insurgency’s commitment to its goals and its desire for a rounded, inclusive movement that reflects and mobilizes entire societies. Daesh and similarly-minded groups allow women to be front-line fighters very occasionally and only when a defeat or threat forces them to change strategy.99 This departure from acceptable gender norms sells the importance of the group’s struggle to its supporters: the state is so oppressive that even women are forced onto the front-lines. Alternatively, leftist groups often advocate for a shift in gender norms that civilians find attractive. The anti-state nationalisms of many rebel groups create space for women to participate that do not exist within the state’s status quo.100 In this sense, opposition to the state doubles as a liberatory practice of shedding some gender norms. Women’s emancipation may be literally and symbolically related to emancipation from the state’s ideology. Some rebels exploit gender beliefs about female victimhood and non-violence to sell themselves as protecting vulnerable women from the state. This involves invoking female victims to justify violence, and rebels often draw directly on this trope in their recruiting and propaganda.101 This narrative enables insurgencies to send two paradoxical but effective messages to civilians: they conform to traditional gendered narratives and protect female members, and they also provide new opportunities for women that the state impedes. For example, The LTTE actively recruited on the basis of protecting women from state sexual violence.102 Emphasizing women’s protection may broaden the appeal of an inclusive movement by contrasting rebel platforms with the state’s violent philosophy. Viterna writes of FMLN recruitment: [C]ultural meanings about women, women’s bodies, and women’s inherent nature were critical to the FMLN’s ability to narrate themselves as the ‘good’ guys during the Salvadoran civil war…they protected vulnerable young women’s sexuality by enlisting them into their guerilla forces, where women’s bodies were respected. By painting a picture of a social order to thoroughly destroyed by the state military that women had no choice but the join the FMLN, the FMLN not only secured women’s active participation, but they also secured their image as righteous protectors with the surrounding civilian population.103 98 Umm Abbas, personal interview conducted by Nabeelah Jaffer over Twitter, quoted in Loken and Zelenz, “Explaining Extremism.” 99 Daesh has recently lifted its prohibition on women combatants in response to its rapid ground loss and transformation away from proto-state into regular insurgency. See: Devorah Margolin and Charlie Winter, 2017. “The Mujahidat Dilemma: Female Combatants and the Islamic State,” Combating Terrorism Center at West Point 10(7). 100 Alison, “Women as Agents of Political Violence.” 101 Cynthia Enloe, 2000. Maneuvers: The International Politics of Militarizing Women’s Lives. University of California Press; Laura Sjoberg, 2013. Gendering Global Conflict: Toward a Feminist Theory of War. Columbia University Press. 102 Alison, “Cogs in the Wheel?” 103 Viterna, Women in War, pgs. 79, 113-114. 23 By exploiting emotional beliefs about women as most vulnerable, rebels demonize the state as particularly monstrous. This can help create ideological salience between civilians and rebels, increasing rebels’ ideological and political legitimacy. This also manifests in the recruitment and visibility of female cadre. For example, the LTTE tied women’s emancipation to broader Tamil liberation to generate civilian support for their separatist cause.104 FARC-EP propaganda actively highlights the leadership positions, education, and self-determination the group offers to women. Pervasive, gendered expectations also make women unexpected and thus uniquely tactical militants, which can strengthen rebel group capabilities. Several studies demonstrate that female attackers are, on average, more lethal than men. O’Rourke finds that female bombers from Lebanese, Palestinian, and Chechen insurgencies, the Kurdistan Worker’s Party (PKK), and the LTTE killed an average of 8 people while male bombers killed an average of 5.105 Women killed at least two more people than men per attack. O’Rourke attributes this discrepancy to security forces’ and civilians’ expectation that women would not be terrorists. Using data on individual suicide attacks from 1989- 2015, Thomas finds that women carried out 8.7% of all attacks and, on average, kill 10.4 people. Men, comparatively, kill about 8.9 people on average.106 Thomas notes that women tend to kill at least 5 more people when attacking political targets. They are also more lethal than men when attacking security forces. Thomas attributes this to “female attackers benefit[ting] from their ability to gain greater access to targets.”107 This effectiveness is not limited to suicide attacks: the LTTE and the Shining Path also used women for non-suicide political assassinations.108 In the IRA, young women laid explosive devices in hopes of avoiding security forces. By 1974 women likely planted most incendiary devices.109 Women hid firebombs in prams or under their clothes to feign pregnancy. There is also evidence that some state forces are reluctant to fire on women because of cultural taboos.110 New research moves beyond individual women’s tactical effectiveness to suggest that women combatants increase rebel capabilities at the group level. Baser finds that the presence of female front- line fighters increases an organization’s military capabilities, assessed using the Non-State Actors in Armed Conflict dataset’s fighting capacity measure.111 She attributes this to women’s ability to avoid suspicion due to gender stereotypes and to the size of female-inclusive groups. Baser suggests that 104 Alison, “Cogs in the Wheel?;” Gonzalez-Perez, Women and Terrorism. 105 O’Rourke, “What’s Special About Female Suicide Terrorism?” 106 Thomas, “Wolves in Sheep’s Clothing.” 107 Ibid, pg. 15 108 Gonzalez-Perez, Women in Terrorism. 109 Belfast Telegraph, 1974. “Women Among IRA Suspects Rounded Up,” 12 April. 110 See: Hererra and Porch, “‘Like Going to a Fiesta,’” pgs. 618-619 for a discussion of this phenomenon in Colombia. 111 Baser, “The Impact of Women Insurgents in Rebel Groups on Rebel Strength and Democratic Sponsors.” 24 women’s front-line participation may also pressure men into joining because they feel emasculated by women performing military roles in their absence. Alternatively, as I discuss below, women’s involvement may mobilize men due to the signals it sends about the need for struggle. Both cases contribute to the overall size of insurgencies, which should contribute to armed group success. These battlefield successes and wider recruitment pools arguably increase rebels’ political legitimacy among civilians by signaling viability and durability. Beyond women’s practical effectiveness, rebel groups use female participation as evidence of the need for violence. Because people assume them to be relatively passive, women’s participation can signal the direness of struggle.112 Rebels capitalize on this narrative in their representation of women militants. Dahal concludes that in Nepal, “the Maoist party mobilized women to give the message that the armed conflict was a demand on all people; such a noble war that even women would fight, supporting and sympathizing with it.”113 Writing on female Palestinian suicide bombers, Schweitzer notes, “The Arab media has not only praised their acts as unmatched heroism, but has called upon men to learn from these women and join the struggle to liberate occupied Arab land and restore Islamic pride.”114 Women’s participation helps demonstrate the importance of struggle, but it also helps integrate rebels into their communities. Because of underlying assumptions about women’s militancy, civilians may take female-inclusive groups more seriously and be less likely to write them off as a fringe movement. Women’s involvement may make rebels seem less extreme, make their vision of society appear more inclusive, and legitimate their use of violence. Teles contends that during the Araguaia Guerilla War in 1970s Brazil, female militants were symbolic figures in mythmaking.115 Some legends about female Araguaia militants concerned one woman, Dina, who was an effective sharpshooter and epitomized bravery. Per Teles, these stories colored interactions between captured female fighters and the Brazilian military.116 State forces feared Dina, and by extension, other guerilla women, in part because of the folklore surrounding them. Teles notes that many of the myths surrounding female guerillas served to both justify rebel actions and to elicit praise and support from local populations.117 In India, one Superintendent of Police suggests that insurgents involve female combatants because it yields credibility for their cause. He contends, 112 Wood et. al. “Female Combatants and the Outcome of Civil Conflict.” 113 Swechchha Dahal, 2015. “Challenging the Boundaries: The Narratives of the Female Ex-Combatants in Nepal,” in Seema Shekhawat, ed. Challenging Gender in Violence and Post-Conflict Reintegration. Palgrave MacMillan, pg. 188. 114 Yoram Schweitzer, 2007. “Palestinian Female Suicide Bombers: Virtuous Heroines or Damaged Goods,” in Cindy Ness, ed. Female Terrorism and Militancy: Agency, Utility, and Organization. Routledge, pg. 132. 115 Teles, “Secrets and Myths about the Araguaia Guerrilla War (1972-1974).” 116 Ibid. 117 Ibid. 25 “It is a general notion that there could not be anything wrong in those activities, in which women are taking part.”118 Temirkulov concludes that women’s participation during Kyrgyzstan’s ‘Tulip Revolution’ “made people more sympathetic to the cause” because their involvement signaled “fair intentions.”119 Indeed, the People’s War Group (PWG) in India consciously increased women’s participation in order to strengthen the movement by making it “more broad-based.”120 This may be a lucrative strategy: the social movements literature recommends diversity for building sustainable movements, concluding that people are most likely to participate when they see themselves and their communities represented. For example, Stephan and Chenoweth contend that campaign success relies on the ability to enhance “domestic and international legitimacy and encourage more broad-based participation in the resistance, which translates into increased pressure being brought to bear on the target.”121 The symbolism of female attackers dying for rebellion is also visceral. In 2002, posters of Wafa Idris, the first female Palestine suicide bomber, were plastered all over Ramallah next to similar posters of male fighters. Wedeen, writing about female martyrs in Lebanon, explores the construction of women as ‘married to’ and dying for the ideological cause.122 She concludes, “Marriage, blood, fecundity, purity, and patriotism are merged in this narrative of female self-sacrifice.”123 When women sacrifice in this violent way, they demonstrate a sense of urgency about the struggle. Their deaths illustrate a fight so noble that even women, the most vulnerable, must die. In 1985, Sana’a Mehaidli of the Syrian Socialist Nationalist (SSNP) drove a Jeep loaded with explosives into an Israeli military site in South Lebanon. She exploded the vehicle, killing herself and two Israeli soldiers while injuring twelve others. Mourners wrote her name on buildings and dedicated processions in her honor, deeming her the ‘Bride of the South’.124 When Israel returned Mehaidli’s body, hundreds of people gathered in Beirut to give her a wedding-funeral. Video of the event shows hundreds of cars escorting her body to her ‘wedding,’ where a stand-in wearing a wedding dress brandished her gun and celebrated with onlookers.125 Receptions followed. 118 Mukherjee, Women Cadres of the PWG, pg. 10. 119 Azamat Temirkulov, 2008. “Informal Actors and Institutions in Mobilization: the Periphery in the ‘Tulip Revolution,’” Central Asian Survey, 27:3-4, pg. 322. 120 Mukherjee, Women Cadres of the PWG, pg. 10. 121 Maria J, Stephan and Erica Chenoweth. 2008. “Why Civil Resistance Work: The Strategic Logic of Nonviolence,” International Security 33(1): pg. 7044. 122 Lisa Wedeen, 1999. Ambiguities of Domination: Politics, Rhetoric, and Symbols in Contemporary Syria. University of Chicago Press. 123 Ibid, pg. 64. 124 Ibid. 125 Al-Jadid/New TV, 2008. July 26. 26 Women’s participation off the front-line can have equally important effects. Female auxiliaries can help rebels establish strong networks required to coordinate and plan attacks without government intervention. Because security forces often overlook them as potential insurgent threats, rebel women in many conflicts excel in coordination and in intelligence collection. Parkinson illustrates the importance of female logistical support to Palestinian insurgents during the Lebanese Civil War, arguing that women used social networks effectively for weapons trade, money transfers, information sharing, and recruitment because of their integration in the community.126 She attributes rebel resilience to women deploying “family, friendship, and community ties and leverag[ing] factors such as gender” to build finance, supply, and information apparatuses.127 Parkinson writes that the arrests and deaths of male fighters led to the destruction of formal command channels and resource pathways. Women, already integrated into their communities, used social networks to fill these gaps. This experience is not unique to Palestine- women build and maintain strong logistical systems in dozens of conflicts spanning from Northern Ireland to Sierra Leone and Sri Lanka. These networks operate informally, as they did in Palestine, or as a primary supply mechanism (e.g, smuggling in Boko Haram). IRA women similarly did much of the legwork for coordinating violence: Alison finds that IRA women had the primary responsibility for concealing and transporting small arms because security forces were unlikely to stop or search them.128 Groups with women auxiliaries, and combatants, may help build ideological and political legitimacy because their members are, or appear, well-integrated into their communities. When women participate in quality of life roles like housing, cooking for, or nursing rebels, it may legitimize rebel operations by grounding them more deeply into the community’s everyday life. Insurgent recruitment and information sharing often takes place in spaces – homes, villages with majority female-headed households, refugee camps – where female activists are most welcome.129 Modern rebels move in and out of civilian life, and female spies, gun runners, recruiters, cooks, nurses, and other support staff often live at home or remain embedded in their communities during conflict. Women are often in a unique position to spread militant ideology and influence community members’ perception of rebels and rebel violence. Their unique position in ‘private’ spaces and importance to the social fabric in most communities make women highly effective recruiters and proselytizers. As one Syrian activist 126 Parkinson, “Organizing Rebellion.” 127 Ibid, pg. 418. 128 Miranda Alison, 2009. “‘That’s Equality for You, Dear’: Gender, Small Arms and the Northern Ireland Conflict,” in Vanessa Farr et. al., eds. Sexed Pistols: The Gendered Impacts of Small Arms and Light Weapons. United Nations University Press. 129 Ibid. 27 notes, “Women are more dangerous than weapons. If you want to spread your ideology, the best way to do it is through women.”130 Viterna finds that the FMLN used women as recruiters to empower potential female cadre and convince the families of young prospects. These insurgents would get close to families who report being more comfortable letting their children join the group when approached by female recruiters. They viewed women as less threatening and more trustworthy.131 Similarly, Cragin and Daily argue that because women logisticians are often the backbone of resistance movements, ‘hearts and minds’ operations should direct more focus onto women.132 In a study of development assistance in post-conflict Liberia, Fearon, Humphreys, and Weinstein emphasize the importance of women’s participation in marshaling community members towards collective action. They note that women “have well-developed, traditional networks and social institutions that could be used to mobilize the women in the community.”133 The authors conclude that mixed-gender groups most effectively mobilize people to cooperate during political activities because of the wide reach that enabled them to solve “mixed-gendered problems.”134 If rebels view civilian support as critical to conflict success and recognize that women’s participation may improve civilian perceptions, then insurgents may advertise gendered aspects of their campaigns. I conclude that rebels often depict women militants, and female civilians, in ways that play on expectations of women’s vulnerability, periphery to conflict, and potential victimization. I identify visual propaganda as a primary vehicle for rebels to convey this messaging. I suggest that rebels represent, and mis-represent, women in their images to generate ideological legitimacy because they believe that this legitimacy encourages critical civilian support. Further, I argue that this strategy offers a pay-off: women’s participation can improve rebel outcomes. Thus, rebels’ representation of militant women and women’s actual participation are equally important pieces in understanding the effects of female participation on rebels’ conflict trajectory. In the sections that follow, I explore the implications of my theoretical framework for rebels’ representation of militant women in their visual propaganda and for women’s actual participation in rebel organizations. I discuss my expectations and highlight the role that ideological and political 130 Adam Heffez, 2013. “Using Women to Win in Syria,” Al-Monitor, September 30. 131 Viterna, Women in War. 132 R. Kim Craigin and Sara A. Daly, 2009. Women as Terrorists: Mothers, Recruiters, and Martyrs. Praeger Security International. 133 James Fearon, Macartan Humphreys, and Jeremy Weinstein, 2015. “How Does Development Assistance Affect Collective Action Capacity? Results from a Field Experiment in Post-Conflict Liberia,” American Political Science Review 109(3): 450-469. 134 Ibid, pg. 450. 28 legitimacy play in the relationships between rebel women and civilian perceptions of insurgent grievances in internal conflicts. Strategic effects of women’s participation: securitization in rebel propaganda Rebels’ securitization of women and gender in political images is a quintessential act of ideological legitimacy-building. Securitization is an approach that legitimates and authorizes political violence by “casting… an issue as one of an existential threat, which calls for ‘extraordinary’ measures beyond the routines and norms of everyday politics.”135 Securitization shifts the political environment into what Williams calls a “politics of emergency,” justifying militarization by capitalizing on public anxieties.136 In conflict images, actors securitize women and gender by deploying gender tropes. Tropes are conventions that remain unaltered in transportation, are easily accessible, are unchallenging in familiarity, and can guarantee their own identity independent of external context.137 Gender tropes can help legitimize violent actors and their militarism by emphasizing familiar notions of gender, or their perversion, to demonstrate the exigency of circumstances. There is surprisingly little research exploring how insurgents leverage political images to legitimate their violence. A small field explores how visuals impact conflict mobilization on social media, namely examining how images convinced people to participate in the Arab Spring and Syrian revolution.138 Some recent studies explore images in jihadist online recruitment, but little is written on how non-state actors deploy political visuals in internal conflicts.139 However, an extensive literature explores state propagandists and gendered images in wartime photojournalism. This scholarship suggests that conflict actors securitize women visually as an act of legitimacy-building to generate support for political violence. I employ this literature to demonstrate how rebels may use gender in their political propaganda to shape narratives about women’s militancy and influence civilian perceptions. 135 Michael C. Williams, 2003. “Words, Images, Enemies: Securitization and International Politics,” International Studies Quarterly 47(4):. pg. 514. 136 Ibid. 137 Marta Zarzycka, 2016. Gendered Tropes in War Photography: Mothers, Mourners, Soldiers, Taylor & Francis. 138 Tamara Kharroub and Ozen Bas, 2016. “Social Media and Protests: An examination of Twitter Images of the 2011 Egyptian Revolution,” New Media & Society 18:9, p. 1973-1992; Hyunjin Seo and Husain Ebrahim, “Visual Propaganda on Facebook: A Comparative Analysis of Syrian Conflicts,” Media, War & Conflict 9, no. 3 (December 1, 2016): 227–51. 139 See: Samantha Mahmood and Halim Rane, 2017. “Islamist Narratives in ISIS Recruitment Propaganda,” The Journal of Interpersonal Communication, 23(1): 15-35; The Combating Terrorism Center, 2006. The Islamic Imagery Project: Visual Motifs in Jihadi Internet Propaganda. Maasri’s research on political posters of the Lebanese Civil War is a notable exception. 29 Gender tropes often elucidate the power relationships underlying conflicts and shape the scope of legitimate security issues. For example, Kearns contends that images of Afghan women in Western media justified U.S. militarization by constructing Afghanistan’s polity as a security threat. He finds two depiction strategies in Western photojournalism- a strategic depiction of suffering where Afghan women were veiled – signaling their victimhood by Islamic oppression – and unveiled – liberated by the Western world after the Taliban’s 2001 defeat.140 Zarzycka similarly proposes that women’s bodies are used as propaganda to mobilize the “loyalty and commitment of the public during wartime.”141 Gender tropes based on generic concepts - victim, mother, sister, life-giver- become objects of security narratives.142 These stereotypes are highly political, and thus easily securitized in pursuit of legitimacy. They are also iconic: easily recognizable, transportable, visceral, and taken as representative of historical memory and normative social structure. Gendered images offer extraordinary symbolic power and the portability of collective emotion, cultural memory, and public discourse. Take, for example, the personification of nations as women and ‘mother lands.’ Baron calls attention the representation of Egypt as a woman in nationalist visuals in the 1800s and argues that this practice likely disseminated from European countries, specifically French republican precedent. She highlights some exceptions – for example, the U.S.’ ‘Uncle Sam’ imagery – but suggests that feminizing the nation is a nearly universal enterprise in nationalist iconography.143 As mothers, feminized nations represent the archetypal gender dichotomy between women requiring protection and men entrusted with their safety. This belief structures social life and is contextually and temporally transportable. It frames threats against the nation as deeply familial. The mother nation genders survival. Scholars also document the successful securitization of gender in state propaganda, particularly during war. These images frequently leverage sexual threats against women as a justification for violence or highlight women’s ‘innate’ victimhood. One famous American recruitment poster, entitled Destroy This Mad Brute—Enlist, features a gorilla (representing the Germans) carrying a distraught, scantily clad woman. The poster references the 1914 Rape of Belgium, when German soldiers sexually assaulted women during their occupation. Gullace argues that this gendered abuse created “irrefutable 140 Matthew Kearns, 2017. “Gender, Visuality and Violence: Visual Securitization and the 2001 War in Afghanistan,” International Feminist Journal of Politics 19(4): pg. 6. 141 Zarzycka, Gendered Tropes in War Photography, pg. xvii. 142 Ibid. 143 Beth Baron, 2004. Egypt as a Woman: Nationalism, Gender, and Politics. Oakland: University of California Press. 30 moral imperatives” to fund the war effort and sacrifice an incredible number of soldiers.144 She concludes that Great Britain similarly articulated its foreign-policy goals at this time around margins “inviolability of a woman's body.”145 Edwards shows that during the 1937-1945 War of Resistance against Japan, Chinese cartoonists launched a propaganda campaign that included drawings of Japanese-perpetrated sexual violence against Chinese women. She argues that the cartoonists believed these images would “build resistance and spur patriotism while equivalent depictions of mutilated male soldiers would sap morale and hamper the war effort.”146 During the 1994 Rwandan genocide, Hutu propagandists characterized Tutsi women as “beautiful, manipulative, highly sexual infiltrators dedicated to humiliating Hutu men and furthering Tutsi dominance.”147 Green suggests that these images, along with radio programming and written text, are largely responsible for widespread Hutu-perpetrated sexual violence during the war. Kangura, a Hutu newspaper, regularly published disparaging and sexual cartoons and visuals.148 These images securitized the Tutsi population, exploiting gender to frame them as a threat to Hutus’ ethnic survival. Gendered tropes are not always so heavy handed. In fact, even those images perceived to be liberating, or even feminist, for women during war often reinforce conservative gender norms. In her study of government depictions of women in World War I propaganda, Shover identifies what she calls gender “role management” – manipulating and reinforcing gender roles - rather than simple role recognition in the images. She concludes, “combatant governments attempted to expand the feminine role to meet the wartime needs of public policy. At the same time, governments attempted to preserve the traditionally passive feminine role.”149 She notes that many in the women’s liberation movement at the time and in the 1970s enthusiastically embraced these posters as progressive, when in fact they were quite conservative. Perhaps the most famous example of propaganda in this vein is Rosie the Riveter and factory-women in U.S. World War II postering. These depictions legitimated U.S. militarization by suggesting that women were being ripped from their ‘normal’ roles as homemakers to fill in for men fighting abroad. The temporary nature of their work was no secret, and the saturation 144 Nicoletta Gullace, 2002. The Blood of Our Sons: Men, Women, and Renegotiation of British Citizenship During the Great War, Palgrave Macmillan US. 145 Sharon Pickering, 1997. “Sexual Violence and Family Honor: British Propaganda and International Law during the First World War,” American Historical Review 102(3): 714-747. 146 Louise Edwards, 2013. “Drawing Sexual Violence in Wartime China: Anti-Japanese Propaganda Cartoons,” The Journal of Asian Studies 72(3): p. 563. 147 Llezlie L. Green, 2002. “Gender Hate Propaganda and Sexual Violence in the Rwandan Genocide: An Argument for Intersectionality in International Law Note,” Columbia Human Rights Law Review 33. pg. 734. 148 Ibid. 149 Michele J. Shover, 1975. “Roles and Images of Women in World War I Propaganda,” Politics & Society 5(4), pg. 469. 31 of these images in U.S. politics illustrates the importance of this narrative to the war efforts. This all- hands-on-deck approach implied that women should be nowhere near the deck, and that military foes could be faulted with this shift in gender roles. Winning the war, in this sense, meant that U.S. gender roles would revert to their ‘natural’ state.’ Rosie and her compatriots identify gender (and power) reconfiguration as a security threat. I suggest that like states, rebels embark on this role management through the securitization of gender in their political visuals. For example, the FARC-EP often used women in photo-ops because women “softened” their image.150 In Chapter 3 I detail my approach to analyzing rebel propaganda, which is based on the iconological study of images. From my theoretical framework and the history of states invoking women in their propaganda, I generate a series of expectations for how rebels will convey gender expectations in their visuals to generate legitimacy among civilians. I expect, foremost, that rebels will invoke gendered tropes of victimhood and social subordination to present female militants as an atypical necessity. I further expect rebels’ images to laud women’s successful use of violence. I expect the depiction of women militants to be highly gendered and very feminine. I also expect that rebels highlight the threat of state-perpetrated sexual violence against female participants in their visuals, selling themselves as protectors. Moving beyond women militants, I expect that rebels, like states, invoke gender tropes about civilian women in their propaganda. Primarily, I expect that rebels visualize civilian women as vulnerable victims of state violence, particularly sexual violence, and that they use these depictions to paint the state as uniquely evil. This paternalistic imagery of both militants and civilians should engage deeply rooted gender expectations to encourage ideological and political legitimacy. Rebels should use gender tropes to sell themselves as morally righteous, effective, and viable alternatives to the state. I describe more concretely the relationship between these expectations and securitization as a strategy in Chapter 3. In this project, I argue that rebels believe women’s participation helps cultivate ideological and political legitimacy and thus they represent women militants, and civilian women, in highly gendered ways in their visual propaganda. I also suggest that there are indeed tangible benefits for including women in rebel organizations, namely in the forms of territorial control and conflict outcomes. In the following section, I explore how women’s militancy should affect these rebel successes, drawing out my expectations for territorial control and favorable conflict outcomes in turn given the theoretical framework presented in this chapter. 150 Hererra and Douglas Porch, “‘Like Going to a Fiesta,’” pg. 614. 32 Outcome effects of women’s participation: territorial control and conflict outcomes How might women’s participation shape rebel legitimacy in ways that lead to tangible conflict successes? There are several possible, and overlapping, mechanisms at work. First, increased ideological legitimacy may make the public more receptive to rebels. With informational loyalty, funds, and recruits, rebels can wage more compelling campaigns for territorial control and for favorable conflict outcomes. Second, increased rebel capabilities stemming from women’s participation may in themselves encourage rebel success. Third, and relatedly, increased rebel capabilities stemming from women’s participation may encourage civilian support for rebels perceived as viable, contributing to rebel success. Fourth, women militants may inspire civilian women and men to join insurgencies in rebel territory or in areas rebels are attempting to control. Thus, gender dynamics can shape the conflict environment in several ways. Here, I explore territorial control and conflict outcomes as key areas of rebel success and discuss the implications of my theory for each. I also discuss alternative explanations, generating a set of testable hypotheses. Territorial control Rebels count on many successes during operations, sometimes even during periods of non- violence, to survive and operate compelling campaigns. It is therefore important to assess not only how insurgencies end, but how rebels succeed and fail along the way. I focus on territorial control as a critical measure of insurgent success. The political violence literature under-appreciates territorial control, which is consequential for planning and executing attacks, avoiding government reconnaissance, recruitment, and negotiating. Indeed, nearly all rebels, including terrorist organizations with seemingly trans-territorial aims, aspire to territorial control in some form.151 It is a tangible show of strength that may shape local and international opinions, as territorial control signals a loss of state sovereignty. Territorial control also offers practical benefits. de la Calle and Sánchez-Garcia conclude that rebels unable to control territory must rely mainly on bombings and other types of violence that do not require extensive logistical planning. In contrast, groups with territory can build infrastructure and fight state forces directly. The authors argue that rebels with territory have greater military power and are most lethal.152 Asal and colleagues conclude that states are more likely to negotiate with rebels 151 Justin V. Hastings, 2010. No Man’s Land: Globalization, Territory, and Clandestine Groups in Southeast Asia, Cornell University Press; Jaume Castan Pinos and Steven M. Radil. 2018, “The Territorial Contours of Terrorism: A Conceptual Model of Territory for Non-State Violence,” Terrorism and Political Violence, OnlineFirst. 152 Luis de la Calle and Ignacio Sánchez-Cuenca, 2015. “How Armed Groups Fight: Territorial Control and Violent Tactics,” Studies in Conflict and Terrorism 38(10): 795-813. 33 that control territory, for several reasons: because these insurgencies already signal legitimacy, the state is not as concerned about legitimizing them through talks; territory provides rebels a strategic bargaining chip; territorial control may yield international legitimacy for rebels; and organizations controlling territory can better overcome “main culprits of bargaining breakdown—information issues, commitment problems, and issue indivisibility.”153 The literature agrees that civilian support is an important factor in maintaining territorial control. For example, Kalyvas suggests that civilians are least likely to collaborate with the state in areas of rebel control because the costs of defection are so high.154 Similarly, when rebels can maintain stronger control, they are less incentivized to use violence against civilians in their territory. Rubin concludes that once they establish territory, rebels work to sustain civilian collaboration through a combination of service provision and coercive violence.155 He notes that rebels rely on civilians to conceal them within their local populations and supply funds, material goods, shelter, and information. But while scholars converge on explaining civilian and rebel behavior once rebels take over territory, there is surprisingly little scholarship concerned with how they achieve control in the first place. The barriers to entry for rebel territorial control are too high for most organizations, and few succeed.156 Rubin acknowledges that to seize territorial control, rebels pay entry costs including “sending personnel and resources to the village necessary to mobilize support, monitor community activities, and deter or defeat counterinsurgent reprisals.”157 I suggest that building ideological and political legitimacy among local communities may be an important part of those entry costs as well crucial to controlling achieved territory. As Siqueria and Sekeris contend, “when seeking to entrench itself in an area, an insurgent group often endeavors… to acquire a system of legitimacy in the eyes of the public.”158 This means that territory-seeking rebels are often reliant on civilian attitudes about them. Territorial control is a process made easier when civilians view rebels as legitimate, because rebels will be more successful if they can avoid defection with minimal use of punishment violence. One way rebels can build this legitimacy is through the strategic deployment of women. For example, the FARC-EP sent front-line women to “interact with civilian populations in areas controlled by the guerillas because they were “more likely to gain the cooperation 153 Victor Asal, Daniel Gustafson, and Peter Krause. 2017. “It Comes with the Territory: Why States Negotiate with Ethno-Political Organizations,” Studies in Conflict & Terrorism OnlineFirst, pg. 3. 154 Kalyvas, The Logic of Violence in Civil Wars. 155 Michael A. Rubin, 2017. “Rebel Territorial Control and Civilian Collective Action in Civil War: Evidence from the Philippines,” unpublished manuscript. 156 Asal et. al. “It Comes with the Territory.” 157 Rubin, “Rebel Territorial Control and Civilian Collective Action in Civil War,” pg. 9. 158 Ibid, pg. 158. 34 of the locals.” Herrera and Porch note that this justified “the FARC claim to be the ‘army of the people,’ legitimize[d] its revolutionary vision, and further[ed] its strategic aims.”159 Women in FRELIMO also worked in mobilization within the group’s territory because civilians accepted them more easily than men.160 Women may further signal the righteousness of violence and rebel strength, shaping civilian opinion in support of rebels and mitigating incentives for defection. Rebels may also achieve and maintain territorial support by providing service provisions. Stewart and Adelstein suggest that female-inclusive insurgents will provide higher levels of social services than other groups because women are likely to be funneled into auxiliary positions that mirror their civilian lives.161 As cooks, educators, health care practitioners, and other support facilitators, women may integrate the insurgency into local communities and demonstrating viability as the territorial authority. By mobilizing social networks, women may be uniquely able to establish intelligence operations, prevent civilian defection, and equip the insurgency. Civilians may be more receptive to rebels’ authority when they view the group as comparatively less threatening, effective, and as advocating norms salient with the local population. Where women participate as leaders, the public may be more supportive of rebel authority. As Henshaw contends, leaders play a significant role in framing insurgent messaging. She notes, “[b]y emphasizing certain politically relevant conditions over others, leaders can influence what issues or factors constituents focus on when constructing opinions about a conflict… a rebel movement can capitalize on grievance in strategic ways by creating communal identities… and by prioritizing some potential grievances over others.”162 Women rebels may recognize their gendered, strategic advantages in civilian engagement, in violence perpetration, and in building strong logistical frameworks. Consequently, women leaders may shape how rebels address and engage with gender in territorially desirable areas. Women leaders may also push for governance structures that politically legitimate insurgent organizations. Stewart and Adelstein conclude that women leaders generally show greater concern for social service provisions, “so women leaders are more likely to advocate for expanded service delivery and a reshaping of service delivery to better adapt to the needs of women.”163 Finally, women’s 159 Hererra and Porch, “‘Like Going to a Fiesta,’” pg. 614 and 621. 160 Marie Jose Arthur, 1998. “Mozambique: Women in the Armed Struggle,” in Patricia McFadden, ed. Southern Africa in Transition: A Gendered Perspective, SAPES Books. 161 Megan A. Stewart and Shirley Adelstein, 2016. “Civil War, Women, and Social Development,” manuscript prepared for the Workshop on Female Combatants, American University. 162 Henshaw, Why Women Rebel, pg. 98. 163 Stewart and Adelstein, “Civil War, Women, and Social Development,” pg. 4. 35 leadership is a bold way for insurgents to convey their ideological platforms. Leadership is highly political and, in many conflicts, is a significant departure from women’s peacetime statuses. As discussed further in Chapter 5, women generally do not participate as leaders in groups where their involvement would generate discord amongst the target population. In conservative groups with restrictive gender norms, female leaders would likely diminish legitimacy because civilians view women in those positions as inappropriate. Consequently, when rebels include women in leadership it is most often a sign of community integration and broad-based participation. This can further civilian support for rebels in their attempts at territorial control. This theory levies a series of observable implications. First, I expect that women’s participation in front-line roles increases the probability of territorial control at the cross-national level. However, I do not necessarily expect that this probability increases linearly as women’s combat participation grows. This is because for many insurgencies, I anticipate a threshold effect. For example, as I explore in Chapter 5, I identify no religious organizations that include women at greater than low levels of front-line participation. Rebels recognize the gender expectations of their publics. Female combatants will shape civilian perceptions differently in a group like Somalia’s Al Ittihad al Islamiya (AIAI) than they will those like the SPLM/A because of pre-existing, deeply held and communal gender beliefs about women, violence, and access to public space. Thus, while I suggest that women’s front-line participation should build ideological and political legitimacy in all cases, the flexibility of these norms and the threshold for signaling exigency and integration likely varies by group. Consequently, I expect that women’s front-line participation increases the probability of civilian support at all levels, but that the substantive effect of unit increases in levels of participation may not be linear. I do expect the linear progression of substantive participation effects in cases of women auxiliaries: as women’s support participation increases, rebels’ probability of territorial control should increase as well. Unlike front-line roles, auxiliary roles in all group-types generally offer women similar positions to those they occupy outside of conflict. When women take on new roles, as intelligence operators, couriers, or recruiters, for example, they often work through existing social networks that integrate the insurgency into prevailing community structures. As a result, women auxiliaries should contribute to rebels’ ideological and political legitimacy across conflicts. There is likely somewhat of a threshold effect here, as a group like the FARC-EP will include women more widely in service provision and educational roles than a group like the Taleban, but overall I expect that women’s auxiliary participation builds civilian support as it increases and subsequently improves the probability 36 of rebel territorial control. Finally, I expect that the effect of female auxiliaries will be the strongest among the three participation types because it is often the most sustained and least temporary. I also expect the effects of women’s leadership on the probability of rebel territorial control to follow a progressive pattern. This is, however, for a different reason. Unlike combat roles, leadership positions signal longer-term flexibility in gender expectations because of the political power these positions innately hold. They are not as temporary as combat roles, which require a much less committed shift in gender roles, and they are not as salient as auxiliary roles, which are generally most reflective of women’s pre-conflict social position. As a result, I discuss in Chapter 5, we generally see female leaders in cases and at levels where they are ideologically expected. Or, at the very least, not ideologically unexpected. Thus, I expect that women’s leadership positions are the most gated during conflict and that when women do participate in these roles, it reflects the gendered expectations of the population. Consequently, I expect leadership roles to behave much auxiliary roles and expect that women’s leadership will substantively increase the probability of territorial control at all levels and in a linear, progressive fashion. Still, there are many explanations for why some rebels control territory and others do not. They are neither mutually exclusive nor singularly applicable. Some of these explanations suggest competing mechanisms for legitimacy and civilian support while others highlight factors in the external environment. Here, I discuss these explanations and generate observative implications from each. I also discuss potential inter-relationships between women’s participation in rebel organizations and these explanations. Table 2.1 below summarizes my expectations from my theory and from competing explanations. First, shared religious or ethnic characteristics between insurgents and the target population may make civilians naturally more supportive of rebels. Jurgensmeyer concludes that “religion can bind a group together and make the members see their struggle as a cosmic battle between good and evil.”164 Religious groups may benefit from wide recruiting pools, and the weight of religious ideology may make militants more willing to fight even in the face of dim prospects or heavy losses. Ethnic homogeneity between rebels and their populations may also make civilians more supportive of insurgents because they identify with them.165 For example, Lewis finds that in Uganda, civilians in ethnically heterogenous areas shared negative information about the insurgents with each other and 164 Mark Juergensmeyer, 2003. Terror in the Mind of God: The Global Rise of Religious Violence. University of California Press. 165 Janet Lewis, 2012. How Rebellion Begins: Insurgent Group Formation and Viability in Uganda, Ph.D diss. Harvard University; Joseph K. Young and Laura Dugan, “Survival of the Fittest: Why Terrorist Groups Endure,” Perspectives on Terrorism 8(2): np. 37 civilian support for rebels waned.166 Like religious groups, ethnic groups may benefit from large recruiting pools and solidarity with – and resources provided by – co-ethnics in other areas. If shared characteristics arguments are correct, we should expect to see that religious and ethnic groups are more likely to territory than other groups because saliences between rebels and their target populations should encourage civilian support. Women are least likely to participate in Islamist insurgencies, but I do not expect this inter-relationship to play a significant role because theoretically, these groups generally do not include women in roles and at levels that are discordant with the gender expectations of their communities. Political ideology may also affect the likelihood of territorial control. Balcells and Kalyvas’ conclude that Marxist rebels are “high quality” insurgents who tend to fight in highly demanding wars that are longer and more lethal than other conflicts. Leftist organizations are most likely to include women, which may contribute to the group’s capabilities and subsequently political legitimacy among civilians. Women likely also improve civilian’s ideological perceptions of leftist rebels, because women’s emancipation is often tied to the larger organizational platform. However, Balcells and Kalyvas argue that states facing Marxist insurgencies endeavor stronger counterinsurgency campaigns because rebels’ “higher quality made the challenge they posed to governments much more credible.”167 As a result, leftist groups may not be more successful than other groups. Consequently, while leftist insurgencies may be able to mobilize the civilian population along political saliences, I expect that governments crack down harder on these insurgencies than on other groups. I therefore expect that leftist insurgencies are no more or less likely to control territory than other organizations. Civilian support can also affect rebels’ ability to mobilize participants into the insurgency. Rebel size may be related to the likelihood of territorial control because of the implications of larger rebellions for capabilities and legitimacy. Size can be a rough proxy for military strength and can also be an indicator of community integration.168 In their work on external government intervention into civil conflict, Szentkirályi and Burch argue that “when smaller factions rebel, their revolutions are more likely to be viewed by the international community as radical or extremist, motivated by some disenfranchised group within society, whose illegitimate campaign diverges from the majority’s interests and its judgment that the states rules justly.”169 Local communities may perceive small rebels 166 Lewis, How Rebellion Begins. 167 Ibid, pg. 19. 168 Stephen Gent, 2008. “Going in When it Counts: Military Intervention and the Outcome of Civil Conflicts,” International Studies Quarterly 52(4): 713–735. 169 Levente Szentkirályi and Michael Burch, 2017. “An Umbrella of Legitimacy: Rebel Faction Size and External Military Intervention,” International Political Science Review, pg. 8. 38 groups similarly and may view larger organizations as more ideologically legitimate and politically viable. Therefore, I expect that larger organizations will be more likely to control territory than smaller groups. This explanation offers an important inter-relationship with my argument: Thomas and Bond find that larger violent political organizations are more likely to include women than smaller factions.170 Legitimacy mechanisms underpinning rebel size and women’s participation may therefore work in tandem to encourage civilian support and increase the probability of territorial control. Beyond civilian support, rebels are at the mercy of external and environmental factors. One, as mentioned, is support from external governments. Balcells and Kalyvas conclude that when external states support insurgents it may earn them international attention and solidarity. This outside validation can increase rebels’ legitimacy among the local population.171 Fungible or material support may yield legitimacy for rebels or offer them resources to incentivize civilian supporters. As a result, I expect that groups receiving external support will be more likely to control territory than those without this assistance. External actors may further view women-inclusive insurgencies as more legitimate than their counterparts because their composition signals righteousness and integration of rebel grievances.172 Women’s participation could theoretically increase the likelihood of external support in these cases. Another key explanation for territorial control is whether natural resources are an aggravating factor of the conflict. Groups with access to natural resources can use them strategically to finance operations or to push the state towards compromise. For example, the Séléka in the Central African Republic levy parallel taxes on mining and smuggle diamonds to finance their activities. Rebels with access to valuable resources can also buy civilian support when they are unable to engender it through legitimacy-building mechanisms.173 This is particularly important for weak rebels because it enables them to incentivize recruits. Alternatively, rebels may also have a greater ability to wield coercive violence against non-combatants without consequence when external resources finance their activities.174 From this explanation, I expect insurgents operating in conflicts where natural resources are an aggravating factor will control more territory than other organizations for two reasons: territory may be even more desirable in resource-rich areas, and resourced rebels should have financial 170 Thomas and Bond, “Women’s Participation in Violent Political Organizations.” 171 Balcells and Kalyvas, “Revolutionary Rebels and the Marxist Paradox.” 172 Wood et. al. “Female Combatants and the Outcome of Civil Conflict.” 173 Wood, “Rebel Capability and Strategic Violence against Civilians.” 174 Jeremy Weinstein, 2007. Inside Rebellion: The Politics of Insurgent Violence. Cambridge University Press. 39 flexibility to procure weapons and buy off detractors. I do not anticipate that women’s participation in insurgent groups is related to natural resources. Favorable geographic conditions also help insurgents. Rough terrain offers insurgents places to hide and train. It may provide rebels an upper hand in areas unfamiliar to security forces.175 The LTTE and the FLN were durable in part because they dragged fighting out into the mountains where state cadre had little geographic knowledge or ability to intervene. Mountains enable insurgents to avoid detection and can provide cover against counterinsurgency operations. From this explanation, I expect that groups operating in rough terrain control the most territory because they can better avoid counterinsurgent operations. I do not expect that terrain is related to women’s participation in rebel organizations. Competition may impact rebel trajectory by making it difficult for groups to set themselves apart as a discernable alternative for civilians. Young and Dugan conclude, “Groups that experience heavy competition, using Darwinian terms, have a higher probability of being selected out.”176 Rebels face competition from co-ethnic groups, from ideological rivals, and from their own splinters.177 This can encourage defection or fractionalization, and competing groups stretch finite civilian support across multiple actors. Consequently, I expect that in competitive environments, insurgents will be less likely to control territory because of rivalries over land and a potential loss of civilian support. I similarly expect that competition weakens rebels because of the opportunities for defection and fighting over civilians and material resources. Groups seeking to set themselves apart to the public may recruit women strategically if they recognize the ideological and political benefits. If this is the case, the competitive environment could override the advantages of women’s participation. At the same time, competitive, women-inclusive groups may be more capable and seen as more politically viable than their rivals. This could allow them to monopolize competitive conflicts. A final explanation for variation in rebel outcomes is state strength. Regime ability to gather intelligence, anticipate attacks, and launch counter-campaigns are conditional on state capabilities. States with limited institutional framework, poor principal agent oversight, and meager abilities to maintain territory and gather reconnaissance data simply cannot conduct broad counterinsurgency operations.178 Weak states also lack legitimacy or institutional resources to succeed in ‘hearts and 175 Ibid; James D. Fearon and David Laitin, 2003. “Ethnicity, Insurgency, and Civil War,” American Political Science Review 97: 75-90. 176 Young and Dugan. “Survival of the Fittest.” 177 Evan Perkoski, 2017. “The Politics of Militant Fragmentation,” manuscript. 178 Lewis, How Rebellion Begins. 40 minds’ campaigns. Wartime destruction of judicial infrastructure, territorial losses, and other tangible indictors of strength encumber the state’s ability to succeed in these areas. Logistical obliteration prevents states from protecting territory or forcibly recovering territorial control from insurgents. Consequently, I expect rebels in low-capacity states to control the most territory. I do not expect that state capacity influences women’s participation in rebel organizations. Explanation Mechanism Observable implications Women’s participation Salience with civilian population through ideological and political legitimacy, resultant civilian support; increased rebel capabilities, resultant civilian support Front-line: women-inclusive groups are more likely to control territory than other groups, but substantive effects may not follow a progressive pattern by level of participation Auxiliary: women-inclusive groups are more likely to control territory than other groups, substantive effects following a progressive pattern by level of participation; strongest effect Leadership: women-inclusive groups are more likely to control territory than other groups, substantive effects following a progressive pattern by level of participation Religiosity Salience with civilian population, resultant civilian support Religious organizations are more likely to control territory than other groups Political ideology Salience with civilian population, resultant civilian support; quality of Marxist insurgencies Leftist organizations are more likely to control territory than other groups Ethnic organizations Salience with civilian population, resultant civilian support Ethnic organizations are more likely to control territory than other groups Rebel size Size signals viability; larger insurgencies can build more expansive governance structures and can send more fighters to the front-line to hold off counterinsurgents Larger insurgencies are more likely to control territory than smaller rebel organizations Natural resources Territorial desirability; financial ability to materially incentivize civilians; less reliance on civilian support for survival and success Rebels operating in conflicts where natural resources are an aggravating factor will control more territory than other organizations Mountainous terrain Rebel familiarity with rough terrain; ability to better avoid state reconnaissance; Groups operating in rough terrain are more likely to control territory 41 counterinsurgency is more difficult in rough terrain than groups in less mountainous areas Competition Rebels in competitive environments may seek out territorial control to differentiate themselves, but territory is finite; competing groups stretch finite civilian support across multiple actors, creating opportunities for defection In competitive environments, insurgents will be less likely to control territory than organizations in less competitive conflicts External support Fungible or material support may yield legitimacy for rebels or offer them resources to incentivize civilian supporters. Groups receiving external support will be more likely to control territory than those without this assistance. State capacity States with limited capacity cannot conduct effective counterinsurgency operations; weak states lack legitimacy or institutional resources to generate civilian support Rebels in low-capacity states will control the most territory Table 2.1: Summary of observable implications, explanations for territorial control Conflict outcomes Rebel successes during conflict are important largely because they contribute to insurgents’ ultimate objective of reconstituting the state. Conflict outcomes generally signal the – often temporary - end of armed violence, suggesting that the state can shift away from active counterinsurgency into managing post-violence instability. Policymakers and governments are rightly concerned with rebel outcomes because they can dictate the future of instability, can help predict armed group resurgence, and can be an important signal of state strength and territoriality. While rebels often aim for unbridled victory, rebel wins are rare and often internal conflicts drag on as protracted wars of proverbial whack- a-mole counterinsurgency. States and rebels may be becoming more pragmatic. The past thirty years represent a marked shift in how conflicts end, with warring actors increasingly agreeing to negotiated settlements under the watch the third-party interveners.179 Ceasefires and negotiated settlements are therefore relatively favorable outcomes for rebels, although they are largely ineffective methods of stabilization. Toft, for example, finds that “civil wars ended by negotiated settlement are more likely to recur than those ending in victory by one side or the other” and that, in fact, “rebel victories are more likely to secure peace than negotiated settlements.”180 But rebels reasonably want power, not necessarily peace, and consequently government concessions in all forms are preferable to defeat or defensive operations. 179 Monica Duffy Toft, 2010. “Ending Civil Wars: A Case for Rebel Victory?” International Security 34(4): 7-36. 180 Ibid, Abstract. 42 The literature identifies a myriad of factors to explain why insurgencies end in such varied ways. Chief among them is the relationship between civilian support for rebels and states’ consequent assessment of rebel strength and popularity.181 Civilian support may be key to the domestic, and international, legitimacy that makes states view rebels as a credible threat to state power. Civilians provide material, personnel, and intelligence resources that help insurgencies launch viable campaigns. As a result, Coggins suggests that rebel legitimacy increases rebels’ negotiating position relative to the state.182 Civilian support may also be critical to rebel victories. Toft suggests, Rebels need a high level of institutional capacity if they are to survive and win… Often, they begin the process of rebellion by criticizing the government’s shortcomings and highlighting (or at times inventing) popular grievances. In cases where incumbent governments overreact to these challenges, the result will be a sharpening of grievances and a shift in public perception that the rebels may have a just cause. Rebels who are successful beyond the goal of merely surviving will be able to use this added legitimacy and increasing popular support to amass more resources and develop their institutional capacity…. Thus, rebel legitimacy and capacity may be endogenous to rebel victory.183 Women’s rebel participation can play a key role in sharpening these grievances, both as victims of potential government overreaction but also as symbols of ideological legitimacy and through their contributions to rebel capabilities. As Wood and colleagues argue, women’s involvement can “signal the community’s sympathy and support for the rebellion as well as the group’s commitment to achieving its political goals despite the steep costs.”184 As front-line fighters and leaders, female participants are highly visible not only to their publics, but also the state and international actors. As auxiliaries, women are often acutely integrated into the populations that governments need to mobilize into informational loyalty to be successful counterinsurgents. Moreover, women’s participation in all roles likely increases rebel capabilities in battle and in the rear-guard. These factors may work together to further shape civilian perception of insurgency, cultivating legitimacy for rebels and strengthening insurgent resolve to fight on against the state. This theory yields a series of observable implications. First, I expect that women’s participation in all roles increases the likelihood that rebels will sign negotiated agreements. This is because states may be more willing to bargain with women-inclusive groups if they comparatively well-resourced and well-supported by the public. Further, women’s participation may influence third-party negotiators, 181 Heger and Jung, “Negotiating with Rebels;” Balcells and Kalyvas, “Revolutionary Rebels and the Marxist Paradox.” 182 Bridgette Coggins, 2015. “Rebel Diplomacy: Theorizing Violence Non-state Actors’ Strategic Use of Talk,” in Ana Arjona, Nelson Kasfir and Zachariah Mampilly, eds. Rebel Governance in Civil War: Cambridge University Press, pg. 98- 118. 183 Toft, “Ending Civil Wars,” pg. 28. 184 Wood et. al., “Female Combatants and the Outcome of Civil Conflict,” pg. 10. 43 generating credibility for the insurgent cause and leading external actors to pressure the government into settlements. I expect this effect to be stronger for female front-line fighters and leaders than for female auxiliaries because, although women in support roles scaffold insurgencies in crucial ways, they are less visible to states and third-party actors. Second, I expect that women’s participation in all roles increases the likelihood of rebel victory and decreases the likelihood of government victory. Ideological and political legitimacy, coupled with women’s prowess as unexpected conflict actors, should resource women-inclusive rebels to better withstand counterinsurgency than other groups. I expect these effects to be strongest for female front- line fighters and auxiliaries, as their participation may have the biggest substantive effects on capabilities, but I expect that women’s leadership still improves rebel chances in this area. The effects of women’s participation on conflict outcomes likely varies by level. For example, I expect that the likelihood of peace agreements and rebel victory is greatest when women participate at the highest levels and is lowest when women participate at the lowest levels. However, as I discuss further in Chapter 7, I assess women’s participation dichotomously in this case due to data constraints. There are several other, intersecting explanations for variation in conflict outcomes. Some of these explanations suggest competing mechanisms for legitimacy and civilian support while others highlight factors in the external environment. Legitimacy and external factors often play mutually enforcing roles, and I discuss them together in this section. Here, I discuss these explanations and generate observative implications from each. I also discuss potential inter-relationships between women’s participation in rebel organizations and these explanations. Table 2.2 below summarizes my expectations from my theory and from competing explanations. State capabilities likely affect conflict outcomes. Aronson and colleagues conclude that rebel military capability, relative to state capacity, is the strongest predictor of whether rebels will win decisively in civil conflict.185 Similarly, Clayton demonstrates that states are only willing to enter negotiations if rebels pose a serious military challenge.186 States generally have much greater capabilities and greater access to resources than insurgents, meaning they can often seek a military solution.187 Still, there is substantial variation between states. Strong states can launch more expansive counterinsurgency campaigns than weak states, meaning that they may be more inclined to fight rather 185 Jacob Aronson, Paul Huth, Mark Lichbach, and Kiyoung Chang, 2015. “How Rebels Win (and Why They Lose),” working paper. 186 Govinda Clayton, 2013. “Relative Rebel Strength and the Onset and Outcome of Civil War Mediation,” Journal of Peace Research 50(5): 609-622. 187 Ibid. 44 than bargain. In contrast, weak states should be more afraid of rebels, less likely to defeat them decisively, and subsequently more likely to settle conflicts quickly. Further, weaker states will have more difficultly incentivizing civilians through ideology or through signals of viability and capability. Therefore, I expect that rebels operating in weak states are most likely to encounter governmental willingness to negotiate.188 I further expect that rebels facing weaker states will be more likely to win outright. I do not expect that state capacity influences women’s participation in rebel organizations, though women-inclusive groups may be more likely to win in weak states given these conditions. State capacity may also affect conflict duration, which should itself affect conflict outcomes. Long insurgencies are less likely to have decisive endings, suggesting that comparatively strong or very weak groups fight relatively short wars.189 Governments may be able to easily defeat very weak groups, and may be more likely to bargain with strong insurgents when they cannot win quickly. Alternatively, long wars may exhaust government willingness to engage in counterinsurgency. Hammes argues, “The greatest strength of the insurgent is the fact that he doesn’t have to win. He simply has to stay in the fight until (the coalition) gives up and goes home. By simply not losing, [insurgencies] compel their opponent to choose- either continue to fight, perhaps indefinitely, or quit and go home.”190 Civilian support for insurgencies can be crucial in strengthening insurgencies quickly, but also for sustaining those that fight for long periods of time. However, given the salience of findings in the existing literature, I expect that longer conflicts are less likely to result in rebel victory or negotiated settlements. External state support may also explain why some insurgents achieve favorable outcomes and others do not, but the evidence on these effects is mixed. Balcells and Kalyvas conclude that when external states support insurgents it may earn them international attention and solidarity. This outside validation can increase rebels’ legitimacy among the local population.191 Walter also suggests that third parties can deter fractionalization and violent spoilers by facilitating peace negotiations and enhancing rebel legitimacy in the state’s view.192 These factors can make settlements more likely. Connable and Libicki further note that state-supported insurgencies often win outright because benefactors provide arms, training, intelligence, and funds.193 Conversely, Regan suggests that direct external support can 188 I do not directly test rebel capabilities in comparison to state strength because the existing data presents a significant mis-match with my rebel groups dataset that, given the scope of this project, I cannot immediately resolve. Further versions of this project will include a measure of rebel capabilities. 189 Ben Connable and Martin C Libicki, 2010. How Insurgencies End, RAND. 190 Thomas X., Hammes, 2006. The Sling and the Stone: On War in the 21st Century, Zenith: pg. 183. Hammes apparently did not know about female insurgents. 191 Balcells and Kalyvas, “Revolutionary Rebels and the Marxist Paradox.” 192 Barbara Walter, 2001. Committing to Peace: The Successful Settlement of Civil Wars, Princeton University Press. 193 Connable and Libicki, How Insurgencies End. 45 prolong civil conflicts.194 Consequently, I expect that rebels who receive external support may be less likely to win decisively, but may be more likely to achieve negotiated settlements. Women’s participation may affect external support for rebel organizations because female militancy may shape external governments’ perceptions of insurgent legitimacy.195 This should be particularly true when women participate in visible roles as combatants and leaders, as female auxiliaries will not be as evident to external actors as to local civilians. Consequently, external governments may be more likely to support rebels when they include women on the front-line and in leadership positions. Another explanation for variation in conflict outcomes is whether natural resources are an aggravating factor of armed violence. There is consensus that resources distract rebels from conflict- ending objectives. For example, Humphreys argues that the draw of natural resources weaken rebel structures, making rebels easier to defeat.196 Other studies suggest that these mechanisms prolong conflict and discourage rebel victories. Buhaug, Gates, and Lujala find that diamonds, other gems, and petroleum distract rebels from building military capabilities. A focus on extraction can disincentivize negotiation or settlement and extend conflict.197 I expect that insurgents in natural resource-aggravated conflicts will experience more unfavorable conflict outcomes because they are relatively weak militarily and may lack bargaining power. I do not expect that women’s participation in rebel groups affects this relationship. Whether rebels control territory may also affect the likelihood of favorable conflict outcomes. Asal and colleagues find that for ethno-political insurgents, territorial control is a key predictor of negotiated settlements because it offers rebels a strong bargaining chip. They conclude, “when a challenging organization establishes governing structures and controls movement in part of a state’s territory, however, it can easily inflict significant economic and political costs on the state while also possessing a valuable asset to exchange for concessions.”198 As discussed, territorial control can also signal civilian support for insurgents through both ideological legitimacy and through governance structure and capability signals that shape civilian perceptions of viability. The location of the territory may mitigate the effect of control. Cunningham and co-authors argue that when rebels control territory in peripheral areas as opposed to those near centralized state power, control offers less 194 Patrick M. Regan, 2002. ‘‘Third-party Interventions and the Duration of Intrastate Conflicts,’’ Journal of Conflict Resolution 46 (1): 55-73. 195 Wood et. al. Female Combatants and Conflict Outcomes. 196 Macartan Humphreys, 2005. “Natural Resources. Conflict, and Conflict Resolution: Uncovering the Mechanisms,” Journal of Conflict Resolution 49(4): 508-537. 197 Halvard Buhaug, Scott Gates, Päivi Lujala, 2009. “Geography, Rebel Capability, and the Duration of Civil Conflict,” Journal of Conflict Resolution 53(4): 544–569. 198 Asal et. al., “It Comes with the Territory,” pg. 3. 46 political capital because peripheral rebels may still be weaker than the state.199 They caution against conflating the ability to inflict costs on the government and evade counterinsurgency. Still, I expect that rebels controlling territory will be more likely to achieve peace negotiations and victories. I further expect that when rebels control territory and include women in any roles, the likelihood of favorable outcomes increases because of the interacting effect between these characteristics. The competitiveness of rebel environments may also influence conflict outcomes, for the reasons discussed above. Competition makes it more difficult to rebels to survive and to differentiate themselves.200 I expect that competition weakens rebels because of the opportunities for defection and fighting over civilians and material resources. If true, rebels with many rivals may have difficulty threatening the state. However, why competition may make successes difficult during conflict, rebels in such circumstances may fare better at conflict’s end. Asal and colleagues contend that states may want to negotiate with less popular groups or those facing rival organizations because they will “demand less and coerce less effectively.”201 Further, negotiating with some groups over others can divide civilian support among organizations and muddle movement agendas. Therefore, I expect that rebels in competitive environments will face an increased likelihood of negotiated settlements, but may be less likely to be victorious. As discussed, groups in particularly competitive environments may theoretically recruit women strategically. Thus, women-inclusive groups operating among many rivals may be better situated for victory. Finally, democratization may account for rebel outcomes in civil conflict. Democratic states may face higher audience costs that prevent brute-force counterinsurgency campaigns, making decisive government victory more difficult. These governments may also be concerned with international reputation costs and attempt to avoid violating the international laws of war.202 For example, democracies are less likely to use mass killing during counterinsurgencies than other states.203 Consequently, I expect that democracies will be less likely to decisively defeat rebel groups and will be more likely to sign ceasefires and peace negotiations. I further expect that rebels will be less likely to win in democracies because of these settlements and because democracies generally have strong 199 David E. Cunningham, Kristian Skrede Gleditsch, and Idean Salehyan, 2009. “It Takes Two: A Dyadic Analysis of Civil War Duration and Outcome,” Journal of Conflict Resolution 53(4): 570-597. 200 Young and Dugan. “Survival of the Fittest.” 201 Asal et. al., “It Comes with the Territory,” pg. 6. 202 Martha Finnemore, 1996. National Interests in International Society, Cornell University Press. 203 Benjamin Valentino, Paul Huth, and Dylan Balch-Lindsay, 2004. ““Draining the Sea:’ Mass Killing and Guerrilla Warfare,” International Organization 58(2): 375-407. 47 enough military capabilities to keep rebels at bay. I do not expect that women’s participation is related to a state’s degree of democratization. Explanation Mechanism Observable implications Women’s participation Salience with civilian population through ideological and political legitimacy, resultant civilian support; increased rebel capabilities, resultant civilian support; state and external actor recognition of support and capabilities Front-line: women-inclusive groups are more likely to achieve peace agreements and rebel victories than experience rebel defeat Auxiliary: women-inclusive groups are more likely to achieve peace agreements and rebel victories than experience rebel defeat; effect comparatively smallest for peace negotiations Leadership: women-inclusive groups are more likely to achieve peace agreements and rebel victories than experience rebel defeat; effect comparatively smallest for victory State capacity States with limited capacity cannot conduct effective counterinsurgency operations; weak states lack legitimacy or institutional resources to generate civilian support Rebels in low-capacity states will be more likely to achieve peace agreements and to win decisively Conflict duration Governments may be able to easily defeat very weak groups, and may be more likely to bargain with strong insurgents when they cannot win quickly Longer conflicts are less likely to result in rebel victory or negotiated settlements External support External support can increase rebels’ legitimacy among the local population; third parties can facilitate peace negotiations, but prolong civil conflicts Rebels who receive external support may be less likely to win decisively, but may be more likely to achieve negotiated settlements Natural resources Resources distract rebels from conflict-ending objectives, making them easier to defeat Insurgents in natural resource- aggravated conflicts will be more likely to experience defeat and less likely to reach peace agreements Territorial control Territorial control offers rebels a strong bargaining chip; can signal civilian support for insurgents Rebels controlling territory will be more likely to achieve peace negotiations and victories. Competition States may want to negotiate with groups facing rival organizations because they will demand less; competing groups stretch finite civilian support and resources across multiple actors Rebels in competitive environments will face an increased likelihood of negotiated settlements, but may be less likely to be victorious. 48 Democracy Democratic states face higher domestic and international audience costs that prevent brute-force counterinsurgency campaigns, making decisive government victory more difficult; democracies generally have strong enough military capabilities to keep rebels at bay Democracies will be less likely to decisively defeat rebel groups and will be more likely to sign ceasefires and peace negotiations; rebels will be less likely to win in democracies Table 2.2: Summary of observable implications, explanations for conflict outcomes Conclusion To summarize, I argue that women’s participation in all roles offer insurgencies a strategic advantage in their engagement with the domestic public. Specifically, I suggest that rebels believe women’s front- line, auxiliary, and leadership involvement makes civilians more receptive to their grievances and that civilian support is critical to successes during and at the termination of conflict. Moreover, I theorize that women-inclusive insurgencies should experience more conflict-related success, namely through the procurement of territorial control and an increased likelihood of negotiated settlements and decisive victory, because of the legitimizing and capability-building effects of female militancy. In the following chapters, I empirically interrogate these expectations through a multi-method research design. In Chapters 3 and 4, I examine the securitization of women militants and gender more broadly in rebel visuals to demonstrate that insurgents cross-nationally rely on and invoke women in their public outreach to shape civilian perceptions. I then turn to a quantitative assessment of how women’s participation affects tangible outcomes, emphasizing the myriad ways that gender affects conflict dynamics and highlighting the importance of disaggregating women’s participation by role. 49 Chapter 3: The Visual War In the qualitative part of this project, I theorize rebels’ securitization of gender in visual propaganda to legitimate their campaigns. I argue that rebels deliberately invoke women’s conflict participation as a visual device to influence public perceptions of insurgency. I suggest that rebels can securitize gendered tropes of victimhood and subordination to build legitimacy from sympathetic and uncommitted audiences. While a rich literature explores the possibilities, and effectiveness, of visual securitization, rebels and other non-state groups’ use of political images is tremendously understudied in political science. In this neglected space I establish framework for interrogating the visual securitization of rebel images. I propose a theory of visual securitization that explains why and how rebels can exploit women militants to fundamentally shape engagement with the public. Securitized images can direct legitimacy, attention, and resources to rebels’ struggle by deploying salient gender tropes in militarized context. Specifically, I introduce an interpretivist, iconological method that holds gender tropes as freestanding political icons. Iconology is a visual approach most commonly found in cultural sociology and art history that involves the “systematic interpretation of images as images by also taking their social embeddedness into account.”204 As Heck and Schlag explain, iconology seeks images’ symbolic forms, intentionality, and historical continuity. This enables us to connect the “iconic, material and discursive dimension of visuals and directs our attention to the processes of visual securitization.”205 My framework makes two primary interventions into the iconology and visual securitization literatures: I argue that tropes across political images can constitute icons, and I suggest that these images can ‘speak security’ autonomously without supplemental discourses. I follow Hairman and Lucaites in defining icons as “images appearing in print, electronic, or digital media that are widely recognized and remembered, are understood to be representations of historically significant events, activate strong emotional identification or response, and are reproduced across a range of media, genres, or topics.”206 Hansen distinguishes between discrete (e.g. the photograph of the hooded Iraqi prisoner at Abu Grahib) and generic icons (e.g. paintings of the Virgin Mary). In the latter, “certain elements are repeated over and over, from image to image, so that despite varying subjects, times, and locations, the basic scene becomes a familiar staple, a visual cliché.”207 I 204 Axel Heck and Gabi Schlag, 2013. “Securitizing Images: The Female Body and the War in Afghanistan,” European Journal of International Relations 19(4): 892-893. 205 Ibid, pg. 896. 206 Robert Hariman and John Louis Lucaites, 2007. No Caption Needed: Iconic Photographs, Public Culture, and Liberal Democracy, University of Chicago Press: pg. 27. 207 David D. Perlmutter, 2005. “Photojournalism and Foreign Affairs,” Orbis, 49:1: pg. 1. 50 adopt this approach, but I depart from the iconology literature to argue that the definitional focus on historically significant events is too narrow. As Goldberg notes, icons quickly acquire symbolic status and frames of reference that imbue them in local or international significance.208 Therefore, I conclude that icons can represent not only events or individuals, but highly familiar conventions and political ideas with national or international purchase. Specifically, I propose that we can apply iconology to the study of tropes. I suggest that tropes, in this case gender tropes, can be generic icons. I apply this iconological approach explicitly to visual propaganda, where the actors across conflicts deliberately invoke and securitize widely recognized gender tropes to cultivate legitimacy. My approach also provokes debate in the securitization literature over the capacity of images to be interpreted without supplementary discourse. I suggest that rebel posters and similar propaganda visuals can be autonomous, communicative acts.209 This is contentious: visual securitization scholarship generally agrees that visuals need text or external explanation to ‘speak security.’ This is because images are ambiguous and their meanings malleable.210 Mitchell argues, “[I]mages are not words. It is not clear that they actually ‘say’ anything. They may show something, but the verbal message or speech act has to be brought to them by the spectator, who projects a voice into the image, reads a story into it, or deciphers a verbal message.”211 I disagree. This assumption that visuals cannot convey their intended messages unaided persists because most research on visual securitization focuses on conflict-related photographs that are taken and circulated by photojournalists but widely interpreted by outsiders.212 This approach fails to account for political images directly tied to their means and purpose of production, like rebel propaganda. Propaganda images are uniquely fixed and stable- propagandists disseminate them and often expect them to be interpreted without contextualization. Posters, for example, occupy public spaces and must convey their messages quickly and visually. Most visual propaganda pairs text with images, providing limited guidance to observers. But propaganda often does feature solitary images, as I detail 208 Vicky Goldberg, 1991. The Power of Photography: How Photographs Changed Our Lives. New York. 209 See Williams, “Words, Images, Enemies” pg. 527 for a discussion of images as ‘communicative acts.’ 210 Kearns, “Gender, Visuality and Violence,” pg. 6. 211 W.J.T. Mitchell, 2005. What do Pictures Want? The Lives and Loves of Images, University of Chicago Press, pg. 140. 212 Williams, “Words, Images, Enemies”; Lene Hansen, 2015. “How Images Make World Politics: International Icons and the Case of Abu Ghraib,” Review of International Studies 41(2): 263–288; Jonathan D. Caverley and Yanna Krupnikov, 2017. “Aiming at Doves: Experimental Evidence of Military Images’ Political Effects,” Journal of Conflict Resolution 61(7): 1482–1509;.Kearns, “Gender, Visuality and Violence”; Heck and Schlag, “Securitizing Images”; Cynthia Weber, 2005. “Not Without my Sister(s): Imagining a Moral America in Kandahar,” International Feminist Journal of Politics 7(3): 358–376. Jill Steans, 2008, “Telling Stories About Women and Gender in the War on Terror,” Global Society 22(1): 159–176; V. Mackie, 2012, “The ‘Afghan Girls’: Media Representations and Frames of War,” Continuum 26(1): 115–131. 51 in this chapter and the next. These visuals are intended to be autonomously and immediately understandable. It is, therefore, a mistake to underestimate the communicative power of an icon itself. As a method, iconology aims to “reconstruct the symbolic content of images in their historical and social contexts.”213 Interpreting visuals is a descriptive act, but it is also a process of meaning- making that highlights the power of shared imagery. This method offers a systematic focus on visuality, one that helps us understand how images perform politics and security. Through this approach I demonstrate how rebel organizations use gender tropes as securitizing devices. I suggest that rebels invoke gender to shape the narrative of rebel/state relationships and to justify grievances and use of violence. An important assumption in much of the research on propaganda is that political visuals are an effective means of influencing the observer to benefit the image creator. As Pretorious points out in her study of anti-apartheid propaganda in South Africa, despite a great deal of attention to political propaganda, few studies attempt to make a causal link between the images and their effects on the audience.214 The effectiveness of visual propaganda is seemingly an accepted assumption among scholars. I do not evaluate whether the images in this study work as designed. Instead, I demonstrate that states, rebels, and other securitizing actors believe that publicizing women’s participation this way is advantageous. The evidence across conflicts suggests that underlying gender assumptions permeate across cultures and across struggles and manifest themselves in visual propaganda intended to legitimate the creator’s political violence. This includes the visualization of women as victims, and, specifically, the gendered illustration of female militants. Still, there is evidence that visual securitization is an effective method for legitimizing violence and violent actors, and that images are a highly effective method of conveying information and influencing observer opinion. For example, in an experimental study Caverley and Krupnikov find that visuals of U.S. soldiers increase respondents’ sense of threats to national security even if the military is depicted out of combat. They find that these feelings of insecurity change respondents’ attitudes about the military, resulting in a willingness to allocate more money to defense.215 The authors conclude that securitization of political images is a particularly powerful ‘agenda-setting’ tool. Aday demonstrates that elite cues to the media can shape the coverage of battle images, framing them as instances of sacrifice and heroism. This, in turn, can produce rallying effects in support of war.216 This 213 Heck and Schlag, “Securitizing Images,” pg. 899. 214 Jacqueline Pretorious, 2004. Propaganda Posters as Visual Rhetoric: An Exploration and Case Study. Ph.D diss, University of Pretoria. 215 Carverley and Krupnikov, “Aiming at Doves.” 216 Sean Aday, 2010. “Leading the Charge: Media, Elites, and the Use of Emotion in Stimulating Rally Effects in Wartime,” Journal of Communication 60(3): 440-465. 52 indicates that visuals can substantively shift opinions about conflict when their contents are framed as security issues. While scholars expertly assess gendered state images, war photography, and media distribution of conflict visuals, there is deficient theoretical expansion to rebel organizations. Few studies interrogate how insurgents leverage political images to legitimate their violence. Even less work asks how gender enables visual securitization in this setting. This is surprising, because rebels need to acquire support more intensely than other, more authoritative actors. To mobilize a population against the government, rebels need legitimacy and the material benefits that follow. Earning this support can be difficult, particularly if insurgents kill civilians or if their campaign exhausts the public’s willingness to tolerate insecurity. Visual securitization can be a uniquely effective method for rebels to mitigate these problems. Casas and Williams note, “A large collective action literature portrays information costs as playing a key role in determining the failure or success of mobilizing efforts: people need to know about the existence, costs, and benefits of a mobilization before deciding whether to support it or not.”217 Images are the most effective, quickest way to learn and interpret information.218 With images, rebels can convey their platforms, emphasize the need for struggle, and construct narratives that highlight the benefits of participation or support. First, visuals are widely accessible to observers of many literacy stages and attention spans.219 Writing on visual jihadi propaganda, Brachman and Boudali conclude, It is our belief that these images speak for themselves, quite literally. In most cases, one does not need to be able to read any of the text within the images to understand the broad meanings conveyed by the propagandists. The motifs may have a significant impact on people who are not literate in [the language] so long as they possess a cultural frame of reference that allows them to decode the components of the images.220 In some forms, like online images and posters, they are also comparatively low-commitment methods of transmitting information. Posters, for example, are often plastered on walls and lampposts where observers cannot avoid their messages. In her global history of posters’ development and social life, Guffey asserts that posters “exert a palpable physical presence, shaping spaces while reflecting and altering human behavior.” 221 Reading pamphlets or newspaper articles, seeking out group members, 217 Andreu Casas and Nora Webb Williams, 2017. “Images that Matter: Online Protests and the Mobilizing Role of Pictures,” manuscript. 218 Ann Marie Seward Barry, 1997. Visual Intelligence: Perception, Image, and Manipulation in Visual Communication, State University of New York Press. 219 Casas and Williams, “Images That Matter.” 220 Jarret Brachman and Lianne Kennedy Boudali, 2006. “Introduction,” in The Islamic Imagery Project: Visual Motifs in Jihadi Internet Propaganda, The Combating Terrorism Center, pg. 7. 221 Elizbeth E. Guffey. 2015. Posters: A Global History. Reaktion Books, pgs. 7-39. 53 or attending informational meetings are all political activities that observers may be unable or unwilling to take on. This is a hurdle for insurgent organizations keen to recruit new members and supporters. As Viterna contends, mobilization requires that civilians develop a ‘participant identity,’ defined as believing that “movement participation seems like a natural and even necessary thing” for “someone like you.”222 Images may simply reach more viewers than other forms of communication, but their accessibility may also encourage a wider set of individuals who can understand and interpret the messaging without deviating from their routines. Second, visuals allow rebels to tightly control which messages are circulated and which audiences they will reach. Images can direct observers to focus on specific aspects of an issue intended to shape their opinion.223 This is important because, as Bolt notes, to legitimate violence and lay claim to control of the future insurgents create a narrative of collective identity and memory.224 He suggests that people remember events as stories, and visuals offer a concrete way for rebel groups to memorialize history and continuity for political violence. Importantly, “what counts is not the truth or falsity of a metaphor but the perceptions and inferences that follow from it and the actions that are sanctioned by it.”225 Visual propaganda should influence how individuals view security risks, thus enabling rebels to define civilian interests. Third, visuals evoke visceral, emotional responses. Studies suggest that images have stronger effects on observer emotion and subsequent attitudes than written text.226 In their study of war and conflict framing, Powell and colleagues conclude that when viewed in isolation, images deliver stronger framing effects than text alone.227 When images and text are presented together, text shapes observer opinion but images influence behavioral intentions – the wiliness to act differently based on the treatment image. The authors find that emotions, particularly sympathy, play the strongest role in these effects, especially when images are presented alone. Anger and fear predict behavioral intentions in cases of image and text combination.228 Fahmy and colleagues similarly demonstrate that emotions- 222 Viterna, Women in War. 223 Charles Taber, 2003. ‘‘Information Processing and Public Opinion,” in Sears et. al. eds. Oxford Handbook of Political Psychology, Oxford: Oxford University Press, pg. 433-476. 224 Bolt, The Violent Image, pg. 81. 225 Ibid, pg. 121. 226 Aarti Iyer, Joanna Webster, Matthew J. Hornsey, and Eric J. Vanman, 2014. “Understanding the Power of the Picture: The Effect of Image Content on Emotional and Political Responses to Terrorism,” Journal of Applied Social Psychology 44(7): 511–521; Cornelia Brantner, Katharina Lobinger, and Irmgard Wetzstein, 2011. “Effects of Visual Framing on Emotional Responses and Evaluations of News Stories about the Gaza Conflict 2009,” Journalism & Mass Communication Quarterly 88(3): 523- 540. 227 Thomas E. Powell, Hajo G. Boomgaarden, Knut De Swert, and Claes H. de Vreese, 2015. “A Clearer Picture: The Contribution of Visuals and Text to Framing Effects,” Journal of Communication 65(6): 997–1017. 228 Ibid. 54 namely shock and sorrow- triggered by visual memories of September 11, 2001 predict study participants’ concern with terrorism.229 These studies suggest that securitized images can influence observer attitudes, the perceived legitimacy of political violence, and possibly their behavior in support of that violence. Therefore, I draw on this scholarship to argue that rebels view women’s participation as a publicity asset and securitize gender in their visual propaganda to legitimize their violence. Methodologically, I adapt Hansen’s ideal-typical ‘strategies of security depiction’ to identify and interpret political images.230 These ideal-types categorize common visual securitization strategies and distinguish between how and what tropes, ideas, or symbols rebels mobilize to construct threat and legitimize political violence in response. I employ three primary strategies: fear of the Other; victimization by the Other; and reverence231 (Table 3.1). They are not mutually exclusive: most often, rebel visuals evidence multiple strategies. I focus on how rebels securitize women and gender using these strategies to increase their legitimacy among supportive and uncommitted audiences. Here I discuss, non-exhaustively, the kinds of images expected from each strategy and provide examples. I draw these strategies out further in the empirical chapters using my case study evidence. The first strategy, fear of the Other, depicts the Other (state/opponent) as “demonic, barbaric, evil, and menacing.”232 The second, and similar, strategy is victimization by the Other. This depiction represents rebel members and/or civilians as “persecuted, violated, downtrodden, starving, and thus threatened.”233 These two strategies articulate the beginning and end of the same process: unwarranted state violence. Therefore, they overlap extensively in conflict propaganda and I consider them in tandem in this analysis. These approaches might manifest through the depiction of particularly brutal or extreme forms of violence, including sexual violence. They may portray women as particularly vulnerable, and egregious, victims of physical state violence, and it may also call attention to forced social or political changes caused by the conflict. For example, the People’s Defense Units (YPG) in Turkey and Syria recently posted an image on their official Twitter account featuring a dead, stripped, and sexually mutilated female combatant with superimposed text: “The body of captured YPJ fighter Barin Kobané was mutilated by Erdogan’s terrorists in Afrin.”234 In the reverence strategy 229 Shahira Fahmy, Sooyoung Cho, Wayne Wanta, and Yonghoi Song. 2006. “Visual Agenda-Setting After 9/11: Individuals’ Emotions, Image Recall, and Concern With Terrorism,” Visual Communication Quarterly 13(1): 4–15. 230 Hansen, “Theorizing the Image for Security Studies,” pg. 58. 231 Hansen refers to this strategy as “familiarization.” She also introduces a fourth strategy, “belittlement of the Other.” This strategy constitutes the Other as “insignificant, weak, small, cowardly, backward, or feminine, as someone ‘to be laughed at rather than hated or feared.” I have not identified any rebel visuals that overtly adopt this gendered approach. 232 Hansen, “Theorizing the Image for Security Studies,” pg. 58. 233 Ibid. 234 People’s Defense Units (YPG) in Rojava- Democratic Federation of Northern Syria (@DefenseUnits), Twitter, February 1, 2018. The YPG is the women’s unit of the YPJ. 55 “securitization is made in reference to something that [rebels and/or civilians] hold to be divine, sacred, or superior.”235 One example we may see in gendered rebel visuals is securitization of mothers. The mother as representative of and responsible for reproducing the nation is a consecrated symbol. This includes the personification of the nation or homeland as a mother, and of literal mothers taking part in anti- state activity. A relevant Lebanese Kataeb Party (a Christian paramilitary militia) poster features a woman holding a child and a growing flower. Text accompanies her: “The mothers raise their children to the sounds of glory.”236 We may also see women martyrs. As I draw out in Chapter 2, the imagery of women ‘married to’ and dying for struggle is visceral and, because women are misunderstood to be non-conflict actors, shapes the urgency of threat. Strategy of depiction Description237 Gender tropes Example Fear of or victimization by the Other Depicts the Other (state/opponent) as “demonic, barbaric, evil, and menacing;” represent rebel members and/or civilians as “persecuted, violated, downtrodden, starving, and thus threatened.” State violence, often sexual and often extreme, perpetrated against women and girls, women participating in militant and non- militant ways, female militant mothers, martyrs 2018 People's Defense Units (YPG) photographic image featuring a dead, stripped, and sexually mutilated female combatant with superimposed text: “The body of captured YPJ fighter Barin Kobané was mutilated by Erdogan’s terrorists in Afrin.” Reverence “Securitization is made in reference to something that rebels and/or civilians hold to be divine, sacred, or superior.” Mothers, including the personification of the nation or homeland as a mother; female martyrs and ‘brides’ of the nation 1975-1990 Lebanese Kataeb Party poster featuring a woman holding a child and a growing flower with the text: “The mothers raise their children to the sounds of glory.” Table 3.1: Summary of securitization strategies I use these strategies of depiction to demonstrate how rebels securitize gender as a legitimacy tool. I incorporate Hansen’s approach into an iconological framework to provide a comprehensive assessment of rebel visuals and gender in insurgent imagery. To explore how rebels deploy gender as a security device to influence public perceptions of violence, I conduct a detailed case study of IRA’s use of visual propaganda during the modern conflict 235 Hansen, “Theorizing the Image for Security Studies,” pg. 58. 236 Lebanese Kataeb Party, 1975-1990, Signs of Conflict: Political Posters of Lebanon’s Civil War (1975-1990). 237 Adapted from Hansen, “Theorizing the Image for Security Studies,” pg. 58. 56 in Northern Ireland. I use iconological framework to demonstrate that the IRA securitized women’s participation in republican nationalism to build legitimacy, bolster their platform as a ‘people’s revolution,’ and increase popular support. I argue that the organization exploited gendered notions of militarism, victimhood, and mothering in ways that did not always reflect the realities of women’s experiences. To make these claims, I draw on hundreds of posters, publications, and other propaganda produced during the Troubles and collected through archival research in the Linen Hall Library and the Eileen Hickey Irish Republican History Museum in Belfast, Northern Ireland. I also source republican posters from the International Institute of Social History archive, and CAIN. I collected 532 posters and similar visual propaganda produced by IRA organizations, their political wings, and supportive republican organizations in the United Kingdom and the Republic of Ireland. Of these visuals, nearly 22 percent feature women. Gendered depictions are much more common than other easily securitized themes we might expect in an ethnonationalist conflict. The IRA is a useful case for several reasons. First, the organization was deeply concerned with legitimacy and went to great lengths to build popular support.238 The organization had an uncomplicated, three-prong defense of its activities: all of Ireland has a “natural and inalienable right to national self-determination;” the IRA reflects and continues the historical republican rebellion against British imperialism; and because British occupation is based on physical force, it can only be defeated by physical force.239 Still, the IRA’s strategy of using brutality against civilians over decades to exhaust British interest in Northern Ireland meant that naturally sympathetic audiences experienced years of violence, counter-targeting by loyalist paramilitaries, government repression, and economic disruption. They were desperate to be considered freedom fighters in an anti-colonial war rather than the fringe terrorists that the British government derided. The group regularly stressed, for example, that they provided warnings to the British Army and the RUC before bombing civilian areas. They accused state forces of deliberately delaying evacuations to cause civilian casualties and turn public opinion against the IRA.240 Therefore, the IRA devoted extensive resources to developing and maintaining legitimacy throughout the Troubles, including operating a publicity department. Sinn Féin, the IRA political wing since 1949, operated separate publicity departments, and in the late 1970s a Sinn Féin goal became “developing a propaganda war against the state.”241 This resulted in 238 Joanne Wright, “PIRA Propaganda: The Construction of Legitimacy,” Conflict Quarterly 10(3): 24-41. 239 Ibid. 240 Provisional Irish Republican Army, 1973. Freedom Struggle, Irish Republican Publicity Bureau. 241 The BBC. 2005. “Who Is P. O’Neill?” 28 September. 57 widespread production of political images circulated throughout the United Kingdom, the Republic of Ireland, and abroad. Second, archivists and scholars have collected an extraordinary amount of material from the Northern Irish conflict. For example, since the first year of the Troubles, Linen Hall Library operates a ‘no questions asked’ policy allowing anyone to donate relevant material. The result is a trove of IRA- related archives. In addition to posters, Belfast’s archives hold leaflets, manifestos, magazines, calendars, embroidered handkerchiefs, buttons, badges, and staged photoshoots. This material provides abundant resources for understanding the purpose and presentation of republican political propaganda. Third, the IRA benefitted from many saliences with its target population that could have increased civilian support, including ethnic and national similarities and shared religious grievances. This offers a mechanism for ruling out other support-building explanations. Because the IRA produced visual propaganda so widely, I assess which legitimacy-building areas – religion, ethnicity, gender – the group securitized and in what ways. I find that these all arise in IRA visuals, but that women are invoked more frequently and in ways that contribute to or enhance these other areas. There are also limitations to using the IRA as a primary case study. The Irish question, while not a novel anti-colonial one, is in some ways unique. Western rebellions like the IRA are comparatively few and generally face much stronger governments than their counterparts.242 Indeed, during the Troubles the IRA wrestled a powerful, modern state while many of its insurgent allies in Africa, Latin America, and the Middle East fought weaker opponents. Further, while many insurgencies operate in competitive environments, IRA groups competed amongst each other or, in some cases, worked together towards the same objective. One final limitation is that the IRA case warrants caution when extrapolating it onto other groups. The IRA technically advances a desire to unite Catholics based on their religion and right to a Catholic, Irish state, and the IRA and Protestant paramilitaries traded religiously-targeted civilian killings. Still, it is hard to qualify the IRA as a ‘religious’ insurgency: in its propaganda, public speeches, and state-targeting violence the group looks much more like a typical liberation movement than a religious rebellion. For these reasons, I do not explore my theory within the IRA case alone. I demonstrate that the IRA is not unique in its exploitation of gender as a security device, and that the same underlying ideas – or literal visual reproductions of them – function as legitimacy- 242 Still, rebellions flourished in Spain (Basque Homeland and Liberty, (ETA)), Germany (Red Army Faction (RAF), Macedonia (National Liberation Army (UCK), and among low-intensity leftist rebels in Greece. 58 builders for rebels cross-nationally. To demonstrate the transportability of gender tropes in conflict propaganda among rebel organizations, in Chapter 5 I supplement my primary case study with an analysis of Palestinian political posters produced between 1960-2018. Because the Palestinian and IRA cases are somewhat similar in their basic aims and colonial histories, I also include a survey of gendered political visuals from Afghan, Lebanese, Nicaraguan, Bangladeshi, Namibian, Angolan, Zimbabwean, South African, and transnational jihadist militant groups. These insurgencies operate in highly varied conflicts, and I use them to demonstrate the similarities in securitization strategies across contexts. These analyses are cursory, in part because archival availability for material from these struggles is limited. Still, I elucidate similarities between cross-national rebel visuals and the IRA as well as discrepancies and areas for further theoretic interrogation. I argue that rebels across conflicts believe that women’s militancy shapes civilian perceptions and consequently they securitize women and gender in expressly gendered, and strikingly similar, ways. The Irish Republican Army’s visual campaign Between 1969-1998, the IRA fought a low-intensity war against British governance in Northern Ireland. Women participated as front-line fighters, auxiliaries, and political leaders during this period. Women laid bombs, smuggled weapons, served as lookouts, organized protests, fundraised, and sheltered members in their homes. Security forces arrested and interned them without trial, and many were convicted for IRA-related activities. Recent research profiles female IRA members, identifies patterns in women’s recruitment, highlights women’s diverse roles within the organization, and examines detention and security forces’ use of sexual abuse against female prisoners.243 However, scholars neglect the effect of female participation on group behavior. I argue that women’s participation in militancy and broader republican activism fundamentally shaped the IRA’s strategies and engagement with their domestic audience. I demonstrate that the organization securitized gender in their political visuals to generate legitimacy and increase support for their grievances. I suggest that the IRA believed that women’s involvement improved public opinion and resourcing of the group. Further, I argue that in the pursuit of salient security narratives, the organization exploited gendered 243 Alison, “Women as Agents of Political Violence: Gendering Security;” Mia Bloom, Paul Gill, and John Horgan, 2012. “Tiocfaidh ár Mná: Women in the Provisional Irish Republican Army,” Behavioral Sciences of Terrorism and Public Aggression 4(1): 60-76; Mary Corcoran, 2006. Out of Order: The Political Imprisonment of Women in Northern Ireland, 1972-1998, Willan; Paul Gill and John Horgan, 2013. “Who Were the Volunteers? The Shifting Sociological and Operational Profile of 1240 Provisional Irish Republican Army Members,” Terrorism and Political Violence 3: 435-456; Azrini Wahidin, 2016. Ex- combatants, Gender, and Peace in Northern Ireland: Women, Political Protest, and the Prison Experience, Palgrave Macmillan. 59 notions of militarism, victimhood, and mothering in ways that did not always reflect the realities of women’s experiences. In the remainder of this chapter, I first offer a brief history of republican insurgency in the region and discuss women’s participation in the IRA. I then turn to discuss republicans’ propaganda production during the Troubles and dig descriptively into my dataset of political visuals to identify thematic trends. Finally, I adopt an iconological approach to analyze posters featuring women. I argue that republican propaganda securitized gender and women’s participation to construct narratives of threat in the IRA’s fight against the British government. Specifically, I find that depictions of female militants and nationalist – those supporting a united Ireland, largely Catholics - civilians constitute a substantive part of the rebels’ propaganda campaign. I identify fear and victimization by the Other and reverence of the sacred as primary gendered securitization strategies in IRA and broader republican propaganda. Militant republicanism: a history The conflict in Northern Ireland was born from the centuries-old struggle to free Ireland from British rule. British governance in Ireland officially began in 1800 but dates back to the Crown of Ireland Act of 1542. The country experienced small and unsuccessful Irish rebellions until 1916, when the IRA launched its first insurrection intended to establish a 32-county, self-governing republic. The Easter Monday uprising, known as Easter Rising, lasted for six days after over 1,000 armed men and approximately 200 auxiliary women seized important buildings in Dublin and declared an Irish Republic. Almost 500 militants and civilians died in the clashes, and the state convicted and executed fifteen IRA leaders for their participation. Easter Rising invigorated Irish republicans – nationalists seeking to unite all of Ireland into an independent republic through military force – and prompted the Irish War of Independence in 1919. The 1921 Anglo-Irish Agreement partitioned the 26 southern counties into the Irish Free State. Many republicans saw the treaty as capitulation that betrayed a united Ireland. Their resistance triggered the 1922-1923 Irish Civil War, but they were defeated handily by the Britain-resourced Free State government. In 1937, the Free State received a new constitution and became the Republic of Ireland. The IRA continued to wage sporadic violence for the next four decades, though its military strength never rebounded. In the 1950s, the IRA re-armed and launched a Border Campaign to reunite the six Northern Irish counties with the Republic of Ireland. The Border Campaign lasted from 1956-1962 with disastrous results. Numerous IRA members were arrested and detained without trial, and the campaign 60 never generated much popular support in Northern Ireland. Violence re-erupted in the late 1960s, when the IRA waded into the burgeoning civil rights movement to push for Irish liberation in Catholic areas. Backlash to the push for social justice from unionists – largely Protestant, and conservative, Northern Irish residents loyal to Britain – prompted violence from the Royal Ulster Constabulary (RUC, the police) and British Army. As Moxone-Browne notes, “the more the Army sought to keep the peace, the more the initial goodwill of Catholics evaporated with the result that the IRA found itself cast in the role of defender of the Catholic ghettoes.”244 This period of unrest between 1968 and 1998 is known as the Troubles. In August 1969, a unionist parade through the Bogside area of Derry devolved into a riot between unionist and republican residents. The RUC and British troops deployed into the area, catalyzing simmering tensions, heightening republican anger towards British governance in Northern Ireland, and triggering the creation of two new IRA organizations. In 1971, an IRA sniper killed a British soldier and launched the low-intensity war that that killed approximately 1,935 civilians, 374 IRA members, 162 loyalists – unionists seeking to maintain British rule in Northern Ireland through paramilitary force – and 1,060 police and military forces.245 The Troubles ended in 1998 with the Good Friday Agreement, a peace accord between IRA groups, loyalists, the British government, and the Republic of Ireland. Northern Irish and Irish voters approved the agreement with 71 percent and 94 percent in favor, respectively. The result is a devolved system of government with power shared between Northern Ireland and the Republic of Ireland. Violence has largely ceased since 1998, though some IRA organizations opposed to the treaty have sprung up in the Good Friday Agreement’s wake. The modern IRA sees itself as heir to the militant republicanism of Easter Rising and the Irish War of Independence. This is a simple idea that is made vastly more complicated by the multiple republican factions that operated during the Troubles, each claiming to be the original IRA’s rightful successor (Figure 3.1).246 The Provisional Irish Republican Army (PIRA, also Provos) was the largest and the most dominant. The PIRA formed in 1969 after veteran republicans frustrated with the IRA’s primarily political response to the riots that year broke away and created a new organization. PIRA founders were also worried about an over-focus on political socialism among republicans and wanted 244 Edward Moxon-Browne, 1981. “The Water and the Fish: Public Opinion and the Provisional IRA in Northern Ireland,” Terrorism 5(1-2): pg. 46. 245 Malcom Sutton, “An Index of Deaths from the Conflict in Ireland,” Conflict Archive on the Internet (CAIN), University of Ulster. 246 In addition to those discussed here, other very small assorted organizations existed during the Troubles. Some were proxy groups. None are associated with many military events or recruited successfully. Subsequently, I do not include them in my analysis. 61 to prioritize military resistance. At the same time, IRA members displeased with Sinn Féin’s decision to end its policy of abstentionism in the Republic of Ireland splintered into a new faction called the Official Irish Republican Army (OIRA). The OIRA was explicitly Marxist. In 1970, Sinn Féin suffered a parallel split: Sinn Féin (also called Provisional Sinn Féin) became the PIRA’s political wing while Official Sinn Féin worked in tandem with the OIRA. The OIRA agreed to a ceasefire in 1972. Violence continued through 1973, when the organization began to push out some of its militarism-forward leadership. In 1974, a group of OIRA members opposed to the ceasefire formed the Irish National Liberation Army (INLA). The INLA continued the OIRA’s explicitly socialist agenda, and its creators introduced a political wing, the Irish Republican Socialist Party (IRSP), to support that goal. Figure 3.1: IRA factions operating during the Troubles The PIRA suffered discord and fractionalization as well. In 1986, the group reversed their policy of abstention by removing a ban on PIRA members from supporting elected republicans who took their seats in the Republic of Ireland. This was a controversial decision. In response, some PIRA members split to form the Continuity IRA (CIRA). A Sinn Féin faction, called Republican Sinn Féin, joined them. The CIRA, like the organizations before it, claimed to be the IRA’s legitimate executor. The PIRA experienced another split in 1997 when members opposing the Good Friday Agreement Irish Republican Amy (IRA) Provisional IRA (PIRA), 1969-1998 Sinn Fein Continuity IRA (CIRA), 1986-2017 Republican/Real Sinn Fein Real IRA (RIRA), 1997-p Official IRA (OIRA), 1969- 1973 Official Sinn Fein Irish National Liberation Army (INLA), 1974- p (1998 ceasefire) Irish Republican Socialist Party (IRSP) 62 vowed to continue military struggle as the Real IRA (RIRA). The group, while small, is the largest active militant republican organization in post-1998 Northern Ireland. The RIRA does not have a discernable political wing, though the 32 County Sovereignty Movement supports the RIRA’s goals. The leader of the 32 County Sovereignty Movement, Bernadette Sands McKevitt, is married to Michael McKevitt, the founder of the RIRA.247 Bernadette Sands McKevitt is certainly not the first woman to associate with the IRA. Women have contributed to militant republicanism since its inception. As Volunteers – the IRA’s term for front-line ‘active service’ participants –, auxiliaries, and political leaders, women bolstered IRA struggle in the early revolutionary period and during the Troubles. I suggest that their involvement shaped IRA engagement with civilians, namely through the visualization of female republicans in the rebels’ propaganda. To make this argument, I first evaluate women’s participation in IRA organizations and explore variation in their roles and opportunities. Women in the IRA In 1909, revolutionary Irish republican Countess Markievicz implored women “to realise ourselves as Irishwomen – not as Irish or merely as women, but as Irishwomen doubly enslaved and with a double battle to fight… Arm yourselves with weapons to fight for your nations’ cause.”248 In the decades that followed, women played critical roles within republican organizations seeking a free, united Irish state. Inghinidhe na hÉireann (Daughters of Ireland) educated children about nationalism from 1900-1914, promoting Irish language, song, and history, and publicly advocating home rule. In 1911, member Helena Moloney became one of the first women jailed for militancy when she threw a stone through a shop window displaying images of the United Kingdom’s monarchy.249 Female republicans founded Cumann na mBan (The Irishwomen’s Council), a radical women’s group that served as support for the Irish Volunteers in the pre-civil war years, in 1914. The organization evolved to become auxiliaries for the IRA during the Anglo-Irish Treaty talks (1920-1921) and the Irish Civil War (1922-1923). It was integrated at varying levels into IRA groups during the Troubles. Women’s participation in modern militant republicanism varied by organization. When the IRA split in 1969, the OIRA was the first organization to include women as combatants. The Officials disbanded Cumann na mBan: women were accepted as roundly-trained Volunteers, the IRA term for 247 Bernadette Sands McKevitt is hunger striker Bobby Sands’ sister. 248 Countess Markievicz, 1909. Women, Ideals, and Nation, transcript of a lecture delivered to the Students’ National Literary Society in Dublin, Ireland. 249 Margaret Ward, 1983. Unmanageable Revolutionaries: Women and Irish Nationalism, Pluto Press. 63 its front-line participants.250 There are no reliable estimates of women’s participation in the group, though photographs from OIRA events show numerous women in traditional IRA uniforms.251 When the INLA split from the Officials in 1974, many women who had never belonged to IRA groups joined because they viewed it as a political alternative. Still, Wahidin suggests that the INLA “never really attracted women Volunteers.”252 There is also evidence of some Cumann na mBan auxiliaries and political leaders in the CIRA. Josephine Hayden, the Republican Sinn Féin General Secretary, was incarcerated for five years after being arrested with illegal weapons in 1995. Women members are also generally present for CIRA press conferences.253 In contrast, the only known woman in the RIRA is Bernadette Sands McKevitt. She was, possibly, the third in command in the group’s early years. In 1998, the U.S. denied her a visa because of her involvement in “terrorist activities.”254 Women participated most extensively in the PIRA. At least 4.9 percent of Provos killed in active service were women.255 In 1988, Sinn Féin noted that women were “now represented at every level” of the group and that women had “forged an increasingly active role for themselves and determined that role in opposition to some of their male comrades.”256 The Provisionals maintained Cumann na mBan as an auxiliary organization whose members were “ipso facto members” of the group but remained largely peripheral.257 Female Volunteers were more fully integrated as Provos, though they are often attributed to Cumann na mBan in memorials. In 1987, the PIRA claimed that women composed about 25 percent of participants “including support personnel for the Active Service Units.”258 As Volunteers, numerous women took part in bombings and other violent activities. In the early 1970s, women carried out most explosive attacks.259 In 1973, Marian and Dolours Price were convicted for bombing London’s Old Bailey Courthouse and an army recruitment center that killed one and injured 200. Dr. Rose Dugdale, an economist, hijacked a helicopter in 1974 and used it to 250 Wahidin, Ex-Combatants, Gender, and Peace in Northern Ireland. 251 See, e.g. Bobbie Hanvey, 1974. Official IRA women members of Cumann na mBan. Shot taken on Easter Sunday in Scotch Street, Downpatrick. Boston College Libraries Special Collections. 252 Wahidin, Ex-Combatants, Gender, and Peace in Northern Ireland, pg. 86. 253 Miranda Alison, 2009. Women and Political Violence: Female Combatants in Ethno-National Conflict, Routledge. 254 James F. Clarity, 1998. “U.S. Denies Visa Request from Woman Linked to the Real I.R.A,” New York Times, August 22. 255 Gill and Horgan, “Who Were the Volunteers?” 256 Sinn Féin Women’s Department, 1988. Women in Struggle, pg. 22. Linen Hall Library. 257 Wahidin, Ex-Combatants, Gender, and Peace in Northern Ireland, pg. 87. 258 Scott S. Smith, 1987. “Interview with the I.R.A,” The Irish Edition, December, Original interview transcript. Linen Hall Library. 259 Daily Telegraph 1972; Belfast Telegraph, 1974 “Women Among IRA Suspects Rounded Up,” 12 April. Linen Hall Library. 64 drop bombs – which failed to detonate – on an RUC station.260 In an interview she tells reporters, “If reason would end the wrong England is doing in Ireland, it would have ended 800 years ago. If you recognize that wrong, and are determined to put it right, you have to face the terrible fact that only armed struggle will work.”261 Women’s participation in bombings waned a few years into PIRA operations, likely because in the mid-1970s the organization shifted into a cell-based structure that greatly increased efficiency and decreased the likelihood of informants leaking information to the police. While women offered a unique element of surprise and aroused the least suspicion, this tactical change made these advantages less necessary. Temporal estimates of women’s involvement support this theory: women accounted for 5.1 percent of Volunteers before 1976, but only 2.9 percent between 1976-1980.262 Gender density continued to vary by year: Bloom and colleagues estimate that women composed 6.3 percent of Volunteers, the highest level recorded, in the 1980s.263 A “Republican Roll Call of Honor” compiled by An Phoblacht/Republican News lists 335 front-line republicans killed between 1969-1995.264 At least fifteen, or 4.5 percent, of those listed are women. Using articles published in major Irish and British newspapers during the Troubles, I identify 79 unique incidents of PIRA violence allegedly involving female Volunteers. Within these there are 63 unique female perpetrators and 22 unnamed female suspects. Nearly 18 percent of these incidents involve more than one woman. Eleven women are accused of more than one violent crime, bringing the total number of accusations to 91. These include murders, attempted murders, bombings, attempted bombings, fire-bombings, bomb-makings and premature explosions, shootings, armed robberies, abductions, and a rocket attack where security forces place women at the scene or accuse them of coordination. One woman was arrested for helping engineer an anti-aircraft missile. Because most PIRA attackers are still unidentified, the incidence rate of female Volunteers is likely higher than estimates suggest. Women Provisionals also participated in auxiliary positions, though these roles are more difficult to assess. Quality of life labor like cooking, nursing, and housing is often rendered invisible in daily life, making it a hidden form of participation during conflict. Supportive militant activities like smuggling, intelligence gathering, and laying ‘honey traps’ to lure targets are often covert. Therefore, I expect that historical records undercount female auxiliaries. Still, available documentation suggests 260 Donal O’Higgins, 1980. “English IRA heiress freed up,” United Press International, 11 October. Linen Hall Library. 261 David Beresford. No date. “The prison graduate,” The Guardian. Linen Hall Library. 262 Bloom et. al. “Tiocfaidh ár Mná.” 263 Ibid. 264 An Phoblacht, 1995, “Republican Roll Call of Honor 1969-1995,” 13 April. Linen Hall Library. 65 that Provos women participated widely in these roles. Former Volunteers testify that women were highly effective spies and intelligence couriers.265 Women concealed and transported small arms.266 Numerous women were arrested on weapons charges for gun possession or transport. Women also continued the tradition of educating the public about nationalism. A 1972 recruiting leaflet from a PIRA auxiliary women’s group called Cumann na gCallina specifically targets teenage girls. It reads, Are you doing anything to help your country or your people? If so, is it enough? Have you ever wanted to join the Republican Movement? If your answer to the last question is yes then now you are getting your chance. gCallina is a widespread organization active in your area for young girls… who are interested in Irish Republicanism. The aims of gCallina are to help the Irish Republican Army (Provisional) in their fight for freedom and Justice and also to help educate the youth the Ireland in the National Culture which include[s] the Irish language… Anyone wishing to join this Organization should contact any member of the Provisional Irish Republican Movement. Act now your country needs you.267 One highly visible way that nationalist women contributed as supporters, but not members, of the IRA was through banging on waste bin lids, shouting, and whistling to warn residents when the RUC or British Army entered republican neighborhoods.268 Pickering suggests that security forces’ perpetual intrusions into republican homes broke down traditional gender barriers, politicizing women and mobilizing them towards this resistance.269 Women were also extensively involved in other republican and anti-imperialist organizations, many of which supported armed struggle or incarcerated IRA members. Female activists were the forefront of republican protest during the Troubles. They focused their efforts on four primary causes: ending internment; strip-searching and prison conditions; the application of political prisoner status to republican detainees; and the RUC’s use of plastic bullets.270 Between 1969-1998, women organized hundreds, if not thousands, of protests in Northern Ireland, the Republic of Ireland, and Great Britain. Some women achieved leadership roles within the PIRA’s affiliated political groups. Female commanders staffed Cumann na mBan. Women also held high-ranking positions in Sinn Féin: Máire 265 Bloom et. al. “Tiocfaidh ár Mná.” 266 Alison, “That’s Equality for You, Dear.” 267 Cumann na gCallina. 1972. Questionnaire from Cumann na gCallina to all Young Girls in the Turf Lodge Area. Linen Hall Library. 268 Tara Keenan-Thomson, 2010. Irish Women and Street Politics, 1956-1973: 'This Could Be Contagious,' Irish Academic Press. 269 Sharon Pickering, 2000. “Women, the Home, and Northern Ireland,” Women & Criminal Justice 11(3): 49-82; ____, 2001. “Engendering Resistance: Women and Policing in Northern Ireland,” Policing and Society 11: 337-358 270 See, for example, Sinn Féin POW Department, undated, Armagh Prison Conditions Creating H-Block Conditions; Sinn Féin, 1981. Plastic Bullets Kill! Linen Hall Library. 66 Drumm was the vice president of Sinn Féin and a Cumann na mBan commander when loyalists assassinated her in 1976. By the mid-1990s, Sinn Féin leadership included more women than any other political party in Northern Ireland.271 A particularly famous woman in republican leadership is Bernadette Devlin McAliskey. She served as a Member of the Parliament from 1969-1974 and played a central role during the Bogside protests. She was the youngest woman to ever sit in the House of Commons.272 Though not a member herself, she was a prominent supporter of the campaign for political status for PIRA members and other incarcerated republicans. McAliskey’s daughter, Róisín McAliskey, rose through Provos ranks and was arrested in 1996 and 2007 for allegedly participating in a mortar attack against a British Army compound in Germany.273 Women’s importance to the PIRA and other IRA organizations extended far beyond their active participation. The outbreak of violence in 1968, the brutal governmental crackdown on nationalist protestors, and a resurgence of community support propelled republicans into a widespread propaganda campaign. Women feature prominently in this political imagery. As Volunteers, auxiliaries, and victims they are symbols of republican resistance to British rule. In many cases, they are the ‘face’ of the IRA and its political bedfellows. These depictions reflect a history of republican groups deliberately promoting women in their visuals. Ward recalls the Irish Bulletin, a news sheet produced by the First Dáil – the first parliament of the revolutionary Irish Republic – from 1919-1921.274 Women compiled and circulated the Irish Bulletin, which was intended to promote Irish republicanism and counter British propaganda, but men wrote the stories. The last issue features an image of “The Dáil Girl,” a woman holding a revolver and a cudgel, underneath the headline “Makers of the Republic.” As Ward writes, “The women who had consistently provided the nationalists with their most invaluable weapon – propaganda – had been acknowledged.” In the remainder of this chapter, I explore the extent of this invaluable, gendered weapon in modern IRA propaganda. Visualizing the Troubles Although armed struggle continued in waves after the Anglo-Irish Agreement, republicans and nationalists produced virtually no political imagery between 1921-1968. Loftus attributes this to the 271 Ibid. 272 In 1972, Bernadette Devlin punched then-Home Secretary Reginald Maudling in the House of Commons after he defended British paratroopers who killed 14 unarmed protestors on Bloody Sunday, which she witnessed. When asked if she was sorry, she scoffed and said, “I'm just sorry I didn't get him by the throat.” She also refuted a reporter’s attempts to qualify her action as “emotional,” calling it a calm response to his speech. 273 The BBC. 2007. “McAliskey extradition bid refused,” 23 November. 274 Ward, Unmanageable Revolutionaries. 67 modernization and professionalization of publicity of the post-1960s period.275 The Troubles ushered in a new era of political propaganda for republican groups, one that contributed to the cultural canonization of historical symbols and ethno-national visuals on all sides. Visuals in the Northern Irish conflict are most often associated with political murals. An extensive body of research is dedicated to the murals, cataloguing their histories and identifying republican and loyalist political themes.276 Political murals are integral to loyalist history and date back to 1908, but the republican mural tradition did not begin until the 1980s.277 Goalwin suggests that this is because republicans, long economically and politically disenfranchised by the unionist government, historically had no access to the materials needed to create murals nor the financial ability to secure space.278 Mural painting reached its peak between 1981-1998: republican murals flourished after hunger striker Bobby Sands issued a call to ‘paint and poster’ the region to increase support for political prisoners. Internment and the 1981 hunger strikes “set the stage for a tremendous surge of republican mural painting, a pointed show of resistance to British and unionist dominance.”279 Despite their supremacy in the scholarly literature and popular thought (the murals are now popular tourist attractions), murals composed only part of the republicans’ propaganda operation. In fact, the republican foray into political visualization began over a decade before Troubles muralists found voice. Loftus traces the first republican political posters to an IRA-affiliated organization in 1968.280 In 1969, visiting French art students taught their nationalist counterparts to silk screen. These artists in turn agreed to produce posters for republican organizations, and until the 1970s the production of political images was a largely professional affair. In August 1971, the government introduced internment and the RUC arrested 342 men without charges. Fearing detention, many people stopped making posters. Most of those who continued joined IRA groups for protection.281 Republican poster-production during the Troubles was often well-organized and collaborative. Many 275 Belinda Loftus, 1981. Images in Conflict: Visual Imagery and the Troubles in Northern Ireland (1968-1981), Ph.D diss., Keele University. 276 Bill Rolston, 1991. Politics and Painting: Murals and Conflict in Northern Ireland, Associated University Presses; ____, 2009. “‘The Brothers on the Walls:’ International Solidarity and Irish Political Murals,” Journal of Black Studies, 39(3): 446-470; ___., 2010), “Trying to Reach the Future Through the Past:’ Murals and Memory in Northern Ireland,” Crime Media Culture, 6(3): 285-307; J. A. Sluka, 1996, “The Writing's on the Wall—Peace Process Images, Symbols and Murals in Northern Ireland,” Critique of Anthropology, 16(4): 381-394. 277 Gregory Goalwin, 2013. “The Art of War: Instability, Insecurity, and Ideological Imagery in Northern Ireland's Political Murals, 1979–1998,” International Journal of Politics, Culture, and Society 26(3): 189-215. 278 Ibid. 279 Ibid, 198. 280 Loftus, Images in Conflict. 281 Ibid. 68 posters went through official channels and received organizational approval.282 Sinn Féin and the PIRA produced most republican images. Wright contends that although the IRA was regularly on the propaganda defensive during the Troubles, “the reason why they [were] always able to recover, and to assume the offensive, is because of their careful and considered construction of legitimacy.”283 Sinn Féin publications and other written republican materials stressed the continuity of republican struggle and use of force to legitimize anti- state conflict. Wright argues that the IRA used legitimacy of their cause through propaganda to construct legitimacy of their violence.284 Loftus similarly suggests that Sinn Féin drew on these themes in their visual propaganda to legitimate PIRA activities. This was particularly true in the 1970s, when the IRA favored militarism and strove to differentiate itself from non-violent civil rights groups protesting the state. Loftus notes that a key component of this campaign was linking the IRA’s struggle to anti-colonial liberation movements worldwide: Sinn Féin produced several posters depicting IRA members as freedom fighters at a time when they had direct contact with Basque separatists, SWAPO, and other groups.285 This visual connection strengthened the IRA’s claim to be a legitimate ‘people’s rebellion.’ One IRA propagandist noted that the group’s intended message was, “This war is a young man’s war, and a young women’s war.”286 Danny Devenny, a leading republican visual artist who was wounded during active IRA service in 1973, describes his role as creating imagery to directly counter the British narrative of republican struggle. He contends, The British tried to criminalize republicans by taking away political status, brutalizing the prisoners, and claiming that we didn’t have any support, that the IRA was a ‘criminal conspiracy’… When the British were criminalizing the prisoners, they were claiming that it was not a colonial struggle, but one between two warring factions — the usual myth coming from a policy of ‘divide and conquer.’ We had to tell the world the truth, that this is an ongoing struggle against colonialism, a political struggle, and that the British were not going to make us criminals. The whole idea of the IRA is that they were a people’s army. They were proud of the people, they came from the people, the people supported and provided them with everything.287 282 Loftus, Images in Conflict. 283 Wright, “PIRA Propaganda,” pg. 24. 284 Ibid, pg. 28. 285 Loftus, Images in Conflict. 286 Ibid, pg. 665. 287 Jack Conway, 2010. “Unbowed and Unbroken: A Conversation with Irish Republican Visual Artist Danny Devenny,” Radical History Review 106: 164-166. 69 He concludes, “A lot of what I do is campaign literature. It’s almost like public relations, trying to find an angle from which to expose British injustice.”288 This centered on creating cohesive republican narratives of legitimate anti-colonial struggle, but the visuals also rallied nationalists around the cause and helped mobilized them into participation.289 This approach cemented the IRA’s struggle as integral to the broader, collective liberation struggle from within nationalist homes and neighborhoods. Republicans’ turn towards visuals resulted in widespread production of political posters circulated throughout the United Kingdom, the Republic of Ireland, and abroad. Posters were hung in windows and pasted onto walls and lampposts. They were also used in republican marches and rallies.290 Many of these visuals survived the conflict. Through archival research in Belfast, Northern Ireland, I collected 532 unique posters and similar materials (e.g. smaller prints and cards, calendar covers) produced by IRA groups, their political wings, and republican-sympathetic organizations in the United Kingdom and Republic of Ireland between 1968-1998. While republican groups produced many more hundreds, if not thousands, of leaflets and posters in this period, I rely on a strict coding rubric to include only those meeting the threshold for ‘republican propaganda.’ These data include all unique posters and similar materials produced by IRA groups, their political wings, other republican organizations, and other political groups who expressly support either republican militarism or republican militants. These data exclude civil rights posters and those produced by organizations not publicly supportive or related to republican activities or militants; posters produced by IRA groups in support of rebels in other conflicts, unless they directly reference the Northern Irish conflict; campaign posters for Northern Irish or Republican of Ireland parliamentary elections; material produced and circulated by sympathetic organizations in countries other than the United Kingdom and the Republic of Ireland; and posters that ask the observer to participate in events, demonstrations, rallies, marches, and commemoration ceremonies. For example, I exclude Northern Ireland Civil Rights Association (NICRA) posters. NICRA produced several posters in support of interned republican militants and suspected militants in the 1980s. However, they focused on non-violent solutions to the conflict and campaigned for loyalist internees. In other words, they took a neutral anti-internment approach focused on civil rights. The most difficult coding call to make was the exclusion of posters advertising protests, memorials and other events. Many of these posters feature images or themes found in other visual propaganda. Still, 288 Ibid, pg. 168. 289 Ibid. 290 The Linen Hall Library, 1999. Troubled Images. 70 they are fundamentally informative. Because information is their primary intention, they do not meet the threshold for ‘propaganda’ and I exclude them this analysis. I code the 532 images along a series of indicators to assess the content, producer, and year of production for each visual. The Northern Irish struggle is an ethnonationalist conflict where religious, national, and historical symbols are deeply important to militants and civilians on all sides. Republicans couple these symbols with anti-colonial – and gendered – themes in their propaganda. I identify 81 unique producers, but many are not independent of each other. For example, the Sinn Féin POW Department, the Sinn Féin Youth Department, Republican Publications, and Saoirse all produced posters. Technically, they are products of or related to the PIRA and Sinn Féin. Figure 3.1 in the Appendix shows the distribution of images by individual producer. For ease of analysis, here I collapse producers into eight categories: IRA organizations, their political wings, other republican organizations or unattributed republican creators, groups that are not themselves republican but are supportive of republican struggle, sympathetic anti-imperialist organizations, individuals, and others that do not meet these classifications (Figure 3.2). While supportive and broadly-based anti-imperialist groups did contribute several visuals, IRA groups, their political wings, and other explicitly republican organizations produced the lion’s share of conflict propaganda. It is likely that Sinn Féin groups also produced many of the unattributed republican images (categorized as ‘republican organizations’).291 I identify trends in these posters and assess which symbols, tropes, and narratives republican organizations used to generate legitimacy and sympathy among their constituencies (Table 3.2). Most posters depict state violence against IRA members or Catholic civilians. This includes physical violence, like the firing of rubber bullets into nationalist protests, and the violence inflicted through internment. These images also commonly feature fallen IRA members as martyrs for the cause, depict the IRA using physical force in the name of a united Republic of Ireland, and call attention to the 1980 and 1981 prison hunger strikes. These indicators are not mutually exclusive and, in most cases, posters reflect more than one. For example, hunger strike posters fall within the boundaries of ‘state or police violence’ and some are also martyr posters. 291 Alistair Gordon, Linen Hall Library librarian, Personal Communication, February 2018. 71 Figure 3.2: Distribution of republican propaganda by producer, 1968-1998 Imagery Number of posters Percentage of posters State or police violence (physical or carceral) 362 68% Republican martyrs 129 24% Republican use of force 119 22% Women 117 22% Hunger strikers 115 22% Internment and prison conditions 103 19% Continuity with the original IRA 57 11% Bloody Sunday 22 4% Catholicism or Christianity 19 4% Solidarity with other anti-colonial insurgencies 13 2% Table 3.2: Content trends in republican visual propaganda, 1968-1998 Republicans feature women heavily in their propaganda, in posters of every theme. Women are more often visualized than seemingly obvious symbols of ethnonationalist strife, including religious imagery, continuity with the original IRA, and highly sectarian events like the British murder of fourteen unarmed nationalist protestors on 1972’s Bloody Sunday. The Bloody Sunday events rallied nationalist anger at British forces and energized support for the IRA, leading to an influx of 72 new recruits.292 Yet, Bloody Sunday only features in 4.1 percent of republican political visuals. Similarly, the religious divide between Catholic nationalists and Protestant unions was among the most divisive issues during the Troubles. A surprisingly low number of republican posters, only 3.6 percent, invokes Christian or Catholic symbols. It is difficult to establish the direction of this relationship, but many of the IRA’s key propaganda issues enjoyed support from sympathetic audiences. Indeed, in 1978 only 55 percent of Northern Irish Catholics believed that the British government should crack down harder on the IRA.293 In 1979, Davis and Sinnot found that 42 percent of Catholics in the Republic of Ireland supported the IRA’s motives, and 21 percent supported their activities.294 Police treatment of Catholics was a key issue in 1980s Republican propaganda, and in 1985 57 percent of Catholics in Northern Ireland believed that the country’s police dispensed justice unfairly.295 Internment, another focus of Republican political imagery, was wildly unpopular among Catholics: an estimated 84 percent opposed reintroducing the practice in 1988.296 Around 27 percent of all republican posters and similar material produced during the Troubles were created in the 1970s (Figure 3.3). This production nearly doubled in the 1980s after Sands’ call to visual arms in the 1980s during the hunger strikes, and it dropped back down in the 1990s. Approximately 25 percent of posters appeared in the last decade of conflict. Images featuring women follow an almost identical trend in frequency, suggesting that the depiction of women did not spike or fade noticeably throughout the Troubles but was related to the overall poster production (Figure 3.4). However, the depiction of women does not rise and fall as sharply as the production of posters overall. This indicates that women received consistent attention in periods of lower propaganda production. This is interesting, because, as I discuss above, women’s participation in IRA groups did wax and wane over time. 292 Peter Pringle and Philip Jacobson. 2000. Those Are Real Bullets, Aren't They?: Bloody Sunday, Derry, 30 January 1972, Fourth Estate Ltd. 293 Moxon-Browne, The Water and the Fish. 294 E..E. Davis and Richard Sinnott, 1979. Attitudes in the Republic of Ireland Relevant to the Northern Ireland Problem: Vol 1. Descriptive Analysis and Some Comparisons with Attitudes in Northern Ireland and Great Britain. The Economic and Social Research Institute. 295 Belfast Telegraph, 1985. “Attitudes to Issue of Law and Order in Northern Ireland,” 6 February. 296 Ronald John Weitzer, 1990. Transforming Settler States: Communal Conflict and Internal Security in Northern Ireland and Zimbabwe, University of California Press. 73 Figure 3.3: Distribution of republican propaganda by decade, 1968-1998 Figure 3.4: Distribution of republican propaganda by theme and decade 1968-1998297 Comparatively, the IRA adopted and neglected other potentially salient themes throughout the Troubles. As Figure 3.4 illustrates, there was a sharp decline in propaganda memorializing Bloody 297 This graph excludes a handful of images whose actual decade of production is unidentifiable and who fall within two possible periods (e.g. 1980s-1990s). No material in these data is dated prior to 1970. Although sources indicate republican poster production began two years early, it is unclear if this material survived. 74 Sunday during republicans’ focus on prison conditions, internment, and hunger strikers in the 1980s. Religious imagery skyrocketed in the 1980s, often appealing to Catholic and Protestant sentiments about the right to life and correlations between the strikers and Jesus Christ. It plummeted in the 1990s. Propaganda exalting the IRA’s use of force declined steadily over time, likely coinciding with public willingness to tolerate long term insecurity or be taken in by its blatant advertisement. The volume of data, the proficiency of Sinn Féin and affiliates’ publishing operations, and the distribution of these trends suggest that republican propagandists and their allies made strategic choices about which images to include in their political visuals. I suggest, therefore, that the comparatively frequent depiction of women evidences strategic securitization on the part of republican organizations. I argue that the IRA and supportive republican organizations securitized gender to invigorate nationalist emotions and build legitimacy across ethnic lines. In the sections that follow, I analyze how the IRA and other republican organizations exploited gendered notions of militarism, victimhood, and mothering to shape the narrative of threat and to legitimize extraordinary violent action. I further suggest that they depict female militants and civilians in ways that selectively edited the realities of the conflict. Women in Republican propaganda IRA propaganda prominently featured women during all stages of the Troubles. IRA visuals depict women militants, auxiliaries, and nationalist civilians as active participants in insurgency and as victims of state violence. These depictions helped establish the organization as a ‘people’s revolution.’ It connected the group to liberation movements worldwide and offered a new, emancipatory vision of society. The IRA recognized the strategic benefit of women’s inclusion before they even started their own visual production. In the 1970s, NICRA printed one of its most popular anti-internment posters featuring photographs of incarcerated women. Sinn Féin cut off the reference to NICRA and hung it in their office windows to promote it as their own.298 Gendered republican visuals most often reflects ‘fear of’ and ‘victimization by the Other’ securitization strategies, through which the group depicts the state as “demonic, barbaric, evil, and menacing” or represents rebel members or nationalist civilians as “persecuted, violated, downtrodden, starving, and thus threatened.”299 I suggest that IRA and its supporters adopted these securitization strategies to depict violence against women, but also to portray women’s militancy as the unique result 298 Loftus, Images in Conflict. 299 Hansen, “Theorizing the Image for Security Studies,” pg. 59. 75 of state intrusion. I further find evidence of ‘reverence’ strategies, whereby “securitization is made in reference to something that rebels and/or civilians hold to be divine, sacred, or superior,” through the IRA’s use of mothering and martyrdom as key visual devices.300 Here, I discuss how republican images invoked gender and I suggest that these campaigns deliberately employed, embellished, or selectively focused on women to legitimate their activities as essential, and defensible. Fear of and victimization by the Other A primary approach to rebel legitimacy-building is to paint the state as monstrous and unfit to govern. This securitization strategy enables insurgents to shape threat through state intrusion, provoking outrage among sympathetic populations and generating sympathy from uncommitted or even oppositional audiences. Rebels portray themselves as victims of state violence and the female victim is perhaps the most salient figure in conflict propaganda. Myriad armies and rebel organizations invoke her to justify militarization.301 By capitalizing on gendered ideas about female victimhood, rebels can sell themselves as shelters for women targeted by state violence and broaden their appeal to local communities. I find that republicans securitized gender to emphasize fear of and victimization by the state in two ways. First, and most conventionally, republicans depicted women militants and civilians as victims of physical and sexual violence perpetrated by the state security apparatus. IRA propagandists sought to preserve nationalist support and rally unionist sympathy because the stereotype of female victimhood is engrained across religious and cultural lines. Representations of security forces targeting women are somewhat disproportionate to the realities of this period. Second, I broaden traditional interpretations of gender and victimhood to argue that the IRA’s depiction of female militants also evidences a ‘fear of and victimization by the Other’ strategy. I argue that rebels securitize gender to signal the envelopment of conflict and its disruption of ‘normal’ life. Rebels draw on the persistent belief that women are naturally non-violent and peripheral actors in conflict to frame their participation as part of a collective struggle so all-consuming that even women are involved. This imagery paints the conflict as necessary by emphasizing that women’s participation is anomalous. In the first vein, republican visuals highlight state-perpetrated violence against incarcerated republican women and nationalist civilians. A dominant theme in republican posters is the violent 300 Ibid. 301 Cynthia Enloe, 2000. Maneuvers: The International Politics of Militarizing Women’s Lives, University of California Press; Laura Sjoberg, 2013. Gendering Global Conflict: Toward a Feminist Theory of War, Columbia University Press. 76 strip-searching of women interned or serving sentences in Northern Irish and English prisons. Prison guards subjected IRA and suspected republican prisoners interned, remanded, and convicted to regular, invasive strip searches. Between 1982-1983, 24 women were strip searched an average of 772 times, meaning guards searched each woman an average of 8 times per month.302 Wahidin argues that “in the case of prison, the level of violence against women… was tolerated and normalised as part of the organisational behavior of prison officers, the RUC, and the British Army [sic].”303 She notes that women resisted strip search efforts, which resulted in security forces holding female prisoners down and beating them as they ripped off their clothes. Republican organizations coordinated many hundred protests and pickets aimed at highlighting state violence and ending these practices. They produced and distributed an enormous number of leaflets, handouts, magazines, and posters about internment, strip-searching, and the individual women who were housed under these conditions. Often these included photographs of incarcerated women: Sinn Féin introduced several posters featuring interned Provos women as prisoners of war and calling for their release. Many political visuals portray strip-searching through emotionally jarring illustrations. Figure 3.5 depicts one of the most extensively circulated posters during the Troubles. Produced by Sinn Féin in 1984 or 1985, the image of a nude woman shielding herself in a jail cell conveys immense vulnerability and abuse. A singular demand accompanies her: “Stop strip searches.” This image was reprinted in dozens of nationalist publications, on protest leaflets, and made into badges and buttons for protestors to wear. The Belfast Strip Search Committee, an organization that formed to advocate for the rights of republican prisoners, produced a similarly emotional poster in 1984 (Figure 3.6). The poster demands an end to strip-searching in HM Prison Armagh, the facility that incarcerated many republican women. The Belfast Strip Search Committee produced several other posters related to strip-searching, including one that reproduces Figure 3.5 and notes, “Strip-searching is naked torture.” This type of visual was very common in republican propaganda and nationalist publications in the 1980s and 1990s. Securitizing women this way not only invoked women as victims, it also called attention to sexual violence as emblematic of the colonial violence that Britain inflicted on Northern Ireland: former prisoners recall brutal, coordinated assaults as well as sexualized anti-Catholic remarks.304 As discussed in Chapter 2, in some conflicts propagandists fabricated or exaggerated sexual abuse to compel emotional responses. In Northern Ireland, this very real violence mobilized resistance 302 Bairbre de Brun, 1988. “Women and Imperialism in Ireland,” Women’s Studies International Forum 11(4): pg. 323-328. 303 Wahidin, Ex-Combatants, Gender and Peace in Northern Ireland, pg. 51. 304 Ibid. 77 across political lines.305 Newspapers widely reported the abuse and it became a lynchpin of republican activism – and imagery – against British rule. Importantly, male republican prisoners were also strip searched during this period, though to a lesser extent than women.306 Security forces subjected them to invasive internal examinations.307 While the IRA, its affiliates, and other political groups did focus heavily on prison conditions and the treatment of male detainees, they did not emphasize this sexual abuse in their activism or propaganda. There is no evidence of political visuals concerning this issue. I find no mention of it among thousands of leaflets, posters, magazines, picket calls, or letters related to incarceration. This is not to say that these searches did not outrage sympathetic organizations. Still, the gulf between strip-searching as a central issue in organizing around female prisoners and its comparative silence in campaigns for men is striking. It is, however, unsurprising. Rebels across conflicts highlight state-perpetrated sexual violence to rightly shed light on the abuse, but also to portray themselves as protectors of vulnerable women and to activate emotional responses from observers.308 This narrative extended beyond strip-searches and prisons. In the 1980s, Sinn Féin produced a series of conflict-related posters featuring nationalist and republican women. One poster depicts a woman seemingly afraid and awake in her bed while looming man enters her bedroom (Figure 3.7). She hides herself and draws the covers tightly around her body. There is no text: Sinn Féin expects this image to ‘speak security’ for itself. The overtones are clear: this man is obviously a sexual intruder and she is afraid of him. The Linen Hall Library, experts on IRA materials and the Troubles, suggest that the “most likely explanation is that the woman is a metaphor for Ireland and that the intruder is a metaphor for Britain.”309 States and rebels frequently invoke this imagery to justify political violence, drawing on the trope of women as national symbols and of sexual violation as allegory for colonial for imperial invasion. That the poster features no contextualization suggests that Sinn Féin found it powerful enough to invoke this visual alone and believed the transportability of this trope was so salient that it would be immediately legible to observers. Another conventional example of the ‘fear of and victimization by the Other’ strategy in republican visuals is the depiction of the RUC’s use of plastic (and to a lesser extent, rubber) bullets 305 Monica McWilliams, 1995. “Struggling for Peace and Justice: Reflections on Women’s Activism in Northern Ireland,” Journal of Women’s History 6(4)/7(1): 13-39. 306 Ed Moloney, 2003, A Secret History of the IRA, W. W. Norton & Company, 307 Ibid. 308 For example, the Liberation Tigers of Tamil Eelam actively recruited on the basis of protecting women from state sexual violence (Alison, “Cogs in the Wheel?”). 309 The Linen Hall Library, 1999. Troubled Images. 78 against nationalist protestors. Security forces fired many thousand plastic bullets in Northern Ireland. Between 1972-1981 the RUC killed at least seventeen people with rubber and plastic bullets, 82 percent of whom were male. Eight were children under the age of eighteen, including two girls- Julia Livingston and Carol-Ann Kelly. One female adult, Nora McCabe, was killed this way.310 McCabe, Livingston, and Kelly were the eighth, ninth, and tenth victims of RUC plastic bullets, respectively.311 Still, IRA and advocacy group visuals often, if not primarily, depict the RUC firing on female victims. For example, a 1981 Republican Movement poster depicts the RUC in action and profiles Livingston and Kelly (Figure 3.8). A 1983 poster from an unknown republican producer solely depicts Kelly by including a photograph of her face after the RUC shot and killed her. This image reads: “Loyalist Order, British Order, Civil Order, Plastic Death!” It suggests that the RUC’s campaign for “civil order” is a front for “loyalist order” and “British order” targeting nationalists. Another poster, produced by the United Campaign Against Plastic Bullets in 1986, shows an RUC officer firing directly at an unarmed woman with her hand up to block the blow. Sinn Féin organized anti-RUC protests on this issue and often used photos of McCabe in its advertising. In 1988, Sinn Féin released a report on security forces’ treatment of nationalist women that focused primarily on plastic bullets. It decried “the RUC indiscriminately firing plastic bullets at women and children during peaceful protest” and stressed the victimization of women whose husbands or sons were killed this way.312 While some of these visuals feature male child victims, few, if any, depict adult men despite the latter composing the majority of victims. Women militants The IRA and other republican groups securitized a widely translatable gender trope – women’s victimization - to argue that the British government threatened nationalist safety. They did this through visualizations of state-perpetrated sexual violence against female militants and other physical violence against female nationalist civilians. Republican political visuals also, somewhat paradoxically, used women’s participation in the IRA and militant activity to demonstrate that the security threat of the British government disrupts traditional gender roles and social life. The IRA called attention their female front-line fighters in ways intended to elicit sympathy and humanize the group. Armed women signaled the severity of threat because of the widespread perception that women are generally not 310 Data from the Eileen Hickey Irish Republican History Museum. 311 Sutton, “An Index of Deaths from the Conflict in Ireland.” 312 Sinn Féin Women’s Department. 1988. Nationalist Women and the RUC. 79 conflict actors. This imagery legitimized violence as collective necessity and illustrated the envelopment of struggle into all parts of daily life. It said, if the state were not so oppressive, women would not be forced out of their ‘natural,’ ‘normal’ roles and onto the front-lines. The need to specifically single women out in propaganda and qualify them as ‘female’ participants, as most IRA posters do, calls attention to the fact that women are not supposed to be there. Women represent the exigency of circumstances. In this way, militant women, while defiant and revolutionary, are also securitized as emblematic victims of state violence In these images I see two purposes intended to increase legitimacy and construct security: emphasize the emergency circumstances requiring women’s non-traditional participation and use women’s participation to shape narratives of emancipation as an insurgent platform. Both purposes securitize gender to highlight the envelopment of the conflict, and both suggest that without the state’s violent, colonial governance women would not need to be involved in conflict. In leftist and ethnonational groups, visuals of armed women can demonstrate an insurgency’s commitment to liberating entire communities from state oppression. Alison contends that the IRA provided republican women opportunities to act publicly and violently in ways unavailable in conservative, traditional life. 313 In this case, female participation was literally and symbolically an emancipation from the state’s ideology. Thus, women’s participation can imprint a sense urgency. Some of the most widely circulated political images during the Troubles feature women actively participating in the conflict. Many of these women are armed with rifles, and most visuals that include female militants feature women exclusively. They call direct attention to the ‘womanness’ of women participants. Posters and other propaganda often directly advertise Cumann na mBan. These visuals generally show women in traditional uniforms and holding weapons, using the women’s organization to sell continuity between the modern IRA and their predecessors. This birthright to physical force republicanism was key to the IRA’s platform. By specifically highlighting Cumann na mBan, these images link modern armed resistance to roles women played in Easter Rising and other insurrections. The uniforms depicted often closely resemble Cumann na mBan uniforms from the early 1900s. Figure 3.9 and 4.10 show two such images that emphasize the importance of these women to the IRA.314 They suggest that women are integral to the united Irish state and that female participation is needed to overcome British occupation. Another 1985 poster similarly features an armed Cumann na mBan member in uniform with the text, “Soldiers of Ireland.” 313 Alison, “Women as Agents of Political Security.” 314 Figure 4.9 is a glossy card, approximately 5x7 inches 80 Importantly, these posters misconstrue the realities of Cumann na mBan as a separate auxiliary wing of the IRA and suggest that Cumann na mBan women were regularly front-line fighters. In fact, the PIRA kept Cumann na mBann auxiliary roles and the work of female IRA Volunteers so separate that Moloney describes the group in the early 1970s as the “IRA equivalent of being stuck in the kitchen… they could be useful for gathering intelligence and carrying weapons, but they did little, if any, actual fighting.”315 Wahidin disputes this categorization, arguing instead that Cumann na mBan members were involved in more serious operations including-bomb making.316 Still, she notes that Cumann na mBan members were not generally front-line fighters. In fact, Cumann na mBan members were required to seek leadership approval before carrying out active service missions in emergency circumstances. In cases where this was not possible, the IRA disciplined women who took on front- line roles without permission.317 One former female Volunteer recounts joining the IRA instead of Cumann na mBan she found the latter restrictive: “You got to do more in the Army. In Cumann na mBan, you got to go on camps and you learnt how to use weapons and different things, but with the Army it was the real thing…. You got to do more. You learnt more and you got more respect. I never bothered with Cumann na mBan.”318 Women did become more integrated into the PIRA as the conflict raged on, but the auxiliary wing remained a largely female enterprise. In these and similar visuals, the IRA invoked the Cumann na mBan inaccurately to make greater points about historical continuity and female militancy that they believed would signal the necessity of struggle and attract supporters. Republican propaganda featuring armed IRA Volunteers often pairs visuals of women with a famous IRA idiom: “This is not a man’s war, but a people’s war and very, very much suffering has been borne by the women, be they the mothers, wives, political activists, or Volunteers and the men ought to remember that without the sacrifice of women there would be no struggle at all.” Figure 3.11 shows one iteration of this poster, produced by the IRA. The IRA, like many rebel organizations, was committed to being perceived as a legitimate people’s rebellion and not a band of fringe terrorists. This is evident in their demands for political prisoner status for incarcerated members, as well as in their non-visual propaganda.319 In the 1973 book PIRA propaganda book Freedom Struggle, the rebels 315 Moloney, A Secret History of the IRA, pg. 55. 316 Wahidin, Ex-Combatants, Gender, and Peace in Northern Ireland. 317 Ibid. 318 Ibid, page. 89. Emphasis in original. 319 In 1972, William Whitelaw granted republican prisoners Special Category Status, or prisoner of war status, to separate them in their rights from convicted criminals. In the 1976, the government rescinded Special Category Status and triggered republican hunger strikes. 81 write that the IRA was emerging as “The People’s Army.” They contend, “Any attempt to end British occupation and exploitation, no matter how heroic, stands little chance of success today unless it is firmly based on the common people.”320 The IRA clearly relied on this platform to encourage civilian resources and exhaust British determination, and they evidently discovered that women’s participation in republican militarism helped evidence this narrative. Republican women’s militancy did not always involve violent action. Much of their militant activity was covert or concealed, and therefore images in this period highlight women’s most visible and communal contributions. One of the most recognizable visuals of the Troubles is women banging waste bin lids in the streets to warn IRA members when the British Army conducted raids. Sinn Féin posters widely depicted this activity. It signaled that the republican cause was so imperative that women were willing to sacrifice for it in their private spaces. Figure 3.12 shows a 1990s poster depicting a woman holding a bin lid in front of a sun-burst, which was a common republican symbol for a new dawn, a new Ireland. This poster is from the same series as Figure 3.7, and Sinn Féin intends the viewer to interpret its security implications without text or contextualization. A 1990s poster (Figure 3.13) similarly features women banging bin lids. The poster also depicts an armed woman and the abovementioned quote about women in republican struggle.321 Bin lids are a highly gendered symbol of resistance to British occupation in Northern Ireland. They are everyday, domestic objects mobilized for resistance out of sheer urgency. Owicki suggests that this mode of participation enabled women to engage while “retaining at least the appearance of their traditional roles as protectors off the hearth. In many ways, bin lids then became an iconic symbol for the ‘proper’ roles and duties of Catholic women.”322 She argues that the IRA, conflicted by their desire to be viewed as a people’s army and their Catholic gender traditions relegating women to the home, found the use of bin lids a compelling way to widely integrate women and maintain public support. Female IRA Volunteers report a similar tension within the organization, recalling misogyny despite the public promotion of women militants.323 Owicki notes that the consensus of Catholic nationalism was that republican women “should not seek out conflict, but only act if the conflict comes to them.”324 This gendered line was a precarious one for the IRA, and they likely securitized 320 Provisional Irish Republican Army, Freedom Struggle, pg. 27. 321 This image appears as the backdrop of several other posters during this time-period. While there is no concrete evidence, republican folklore suggests that the armed woman pictured is Mairéad Farrell. 322 Eleanor Owicki, 2010. “Rattle Away at Your Bin: Women, Community, and Bin Lids in Northern Irish Drama,” Theatre Symposium 18: pg. 57. 323 Wahidin, Ex-Combatants, Gender, and Peace in Northern Ireland. 324 Owicki, 2010. “Rattle Away at Your Bin,” pg. 59. 82 women’s front-line and auxiliary participation in their propaganda to demonstrate that the conflict was so dire, it had indeed come to women. A smaller segment of IRA propaganda features women – armed and otherwise - integrated with armed men. One 1987 poster featuring an armed woman with other Provos in action was reproduced as the cover of the annual Republican Resistance Calendar in 1990. In 1982, the Republican Movement produced a poster titled “IRA in Action” that features armed, balaclava-clad Volunteers performing various militant activities. In the center of the poster is an unarmed woman in a traditional Cumann na mBan uniform. A 1981 Republican Publications poster titled “Óglaigh na hÉireann [Soldiers of Ireland]” depicts an armed woman with two male colleagues against a roll call backdrop that lists republicans killed in resistance activity. Like other roll calls this poster lists the dead by battalion. This emphasizes the IRA’s military structure as opposed to terrorist fragmentation, even though by the 1980s the group had shifted towards sleeper cells. This choice sells the group as a legitimate army into which women are integrated members, rather than a band of extremists. This is a different approach than most of the IRA’s visual propaganda, which securitizes women as so exceptional, it warrants unique and gendered attention. Still, the underlying message in these visuals are the same: the IRA is a people’s rebellion, and women’s participation is a measure of the intensity of state oppression and morality of republican political violence in response. Women’s participation integrates the insurgency into local communities: if women are involved, the conflict is everywhere. The IRA saw such propaganda benefit in armed female fighters that, beyond posters, they organized staged photo shoots to further the perception that rifle-wielding women were coming for state security forces. This is noteworthy because while IRA visuals often show women with rifles, it was very rare for republican women to carry arms in public. Many women were arrested for holding, smuggling, or couriering guns during the Troubles, but often they were transporting or concealing weapons for someone else. News records suggest that fewer than a dozen violent incidents attributable to IRA women were shootings, and in most of those cases women assisted or coordinated the attacks.325 Female members were also rarely armed during IRA parades or demonstrations, even when accompanied by armed male colleagues. They infrequently took part in street patrols, where visible armed fighters were most common. Still, the image of armed women offered Provos enough social capital that they promoted it widely. In 1975, the PIRA invited Brendan Murphy, an Irish News photographer, to photograph their training camp. PIRA members picked him up on the side of a deserted road. They threw a hood over 325 News records accessed in the Linen Hall Library. 83 his head, threw him in the back of their car, and drove him around in circles for about an hour before taking him to a facility outside of Belfast. Once there, PIRA fighters only directed him to photograph three women masked in balaclavas, despite there being many more male rebels there. Murphy recalls one of the women asking, ‘Where do you want us?’. The women “posed for about a half dozen pictures – aiming the rifles, holding the guns, rolling out a wire drum that wasn’t actually connected to anything.”326 Eventually, they were interrupted by a security force helicopter flying overhead, but Murphy took and published several of the photographs (Figure 3.14). In the early 1970s, Colman Doyle took a famous photograph of a female Volunteer aiming an ArmaLite AR-18 rifle. Doyle writes that the photo was taken during “active service” in West Belfast, but other news photographers at the time contended that Provos staged it. Regardless, it was so popular that the PIRA reproduced it when they published a children’s book, titled “A Republican ABCs.” The photo is drawn for the book’s first page, reading “A is for Armalite that sends them all running [sic]” (Figure 3.15). ArmaLites were a Provos weapon of choice, and the organization undoubtedly had many images to choose from. That the book opens with a feminine woman aiming the PIRA’s signature gun is compelling, and certainly not a coincidence. Reverence The other strategy identifiable in IRA propaganda is reverence. Securitization of revered ideas or visuals involves using the sacred or divine to shape narratives of threat. Often this involves depicting violations of the sacred or of linking secular, revered phenomenon to religion. For example, some IRA posters calling attention to dying hunger strikes liken them to Jesus Christ, dying for their people and their cause. I find that republicans securitized women to invoke reverence in two ways, as mothers and as martyrs. Interestingly, while the mothering imagery hinges directly on highly gendered understandings of nationalist womanhood, the martyrdom visuals often do not. The martyrdom images and mother images, however, do evidence an overlap between securitization strategies. They both depict women as victims of state violence and sacred symbols, either because the government threatens their natural roles as mothers or because their martyrdom signals the need for extreme resistance to oppression. 326 Brendan Murphy, 2003. Eyewitness: Four Decades of Life in Northern Ireland: O’Brien Press Ltd, p. 28. 84 Mothers are a cross-national gender trope in conflict propaganda and a key component of what Sjoberg calls the “beautiful soul narrative.”327 Sjoberg uses this term to encompass the idea that women are pacifists in need of protection and only participate in war to avenge men or because of defect. She notes that these stories also [E]mphasize women’s mothering, where women’s identities are crucially tied to ‘bearing and rearing children on the home front…’ They provide love and nurture, and at once serve as a support for the logistical and moral fighting for the war and as a symbol of the good and pure that requires the evil of fighting to save it. In this understanding of women’s relationships with war, there is no room for women fighting wars – they are at once fought over in war and protected from it.328 Securitizing women as mothers suggests that the state’s violence threatens women at their most vulnerable and culturally sacred moments. Here, the victimization and reverence strategies overlap. In ethnonationalist conflicts, this is particularly important symbolism because most societies perceive women to “reproduce nations,” biologically, culturally, and symbolically.329 Rich writes that the mother “exemplifies in one person religion, social conscience, and nationalism.”330 This can uniquely expose women to harm during identity-based conflicts and makes motherhood an emotionally charged, highly exploitable visual device in conflict propaganda. Republicans portrayed women militants and nationalist civilians as mothers and therefore uniquely egregious victims of state violence. For example, one United Front Against Internment poster from the mid-1970s demands the release of Mary Kennedy, a suspected PIRA member (Figure 3.16). The poster offers information about Kennedy along with her photograph, noting first that she is the mother of seven children, three of whom are fostered. It then turns to a discussion of her internment being in violation of the European Convention on Human Rights (ECHR). This ordering suggests that the creators believed Kennedy’s status as a mother would be more immediately compelling to observers than the human rights violations of detention without trial. The poster also specifically argues that her internment violates Article 8, the part of the ECHR enshrining the right to respect for one's “private and family life.” Importantly, the poster prioritizes this violation against other potential, and seemingly more persuasive, legal arguments against internment including Article 327 Laura Sjoberg, 2010. “Women Fighters and the ‘Beautiful Soul’ Narrative,” International Review of the Red Cross 92(877): pg. 53-68. 328 Ibid, pg. 56, quoting V. Spike Peterson and Anne S. Runyan, 1999. Global Gender Issues in the New Millennium, Westview Press, pg. 116. 329 Nira Yuval-Davis, 1997. Gender and Nation, SAGE Publications Ltd. 330 Adrienne Rich, 1995. Of Woman Born: Motherhood as Experience and Institution, W. W. Norton & Company, pg. 45. 85 5 (right to liberty and security) and Article 6 (right to a fair trial). It suggests that her status as a mother is the most concerning part of internment. Republican groups reproduced this imagery in the mid-1990s when Róisín McAliskey was arrested while pregnant on an extradition warrant for her suspected participation in an IRA bombing in Germany. Republican posters reading “Interned without charge: free Róisín and her baby” popped up in Northern Ireland, supported by her status as a cause célèbre in feverish news coverage. This kind of revered mother imagery is most often seen in other conflicts as the ‘armed mother,’ juxtaposing the militant with her capacity for and the expectation of nurturing her child. I discuss this depiction further in Chapter 5. Importantly, while these images convey the collectiveness of struggle, we do not see similar references to male rebels as fathers. This visualization humanizes insurgent struggle by never letting the observer forget that even as combatants, women are still women. Another common theme in republican visuals is mothers suffering because their husbands are interned or incarcerated for IRA activity. Sinn Féin produced several posters in the 1980s using this imagery. One such image (Figure 3.17) shows a mother and child awaiting someone’s release from prison. The observer is clearly to infer that they are waiting for their husband and father. An Cumann Cabhrach, the aid committee for the relief of republican prisoners, produced several posters in the 1970s using this visual. One shows a mother and child awaiting someone’s release from prison and reads, “Free the Political Hostages.” In another poster, an incarcerated husband and father asks the observer to contribute to the mother and child during internment. This imagery is very common, but there is no evidence of similar images depicting men as female internees’ dependents, even though many incarcerated republican women were married and had children. Mothers also arise in republican visuals through the gendering of “mother Ireland” and the feminization of national endurance. In Figure 3.16, the mother and child fill the outline of the Republic of Ireland. She is literally ‘mother Ireland.’ In another example, a 1978 republican poster referencing republican men incarcerated in HM Prison Maze (new buildings on the same property as the Long Kesh Detention Centre, where nationalists were held until 1976) H-Block reads, “Ireland, Mother Ireland, your sons shall not be forgotten.” In this way, mother Ireland and the female republican mother are similarly presented as victims of British imperialism whose protection is critical to the survival of Irish nationalism. 86 Martyrs The other set of posters generated through a ‘reverence’ strategy of securitization are images of front- line IRA Volunteers killed in action. Martyrdom posters are not unique to the IRA. In fact, they are most commonly associated with Palestinian factions, many of whom expressed solidarity and shared some resources with the IRA throughout the Troubles. Writing on visual expressions of martyrdom in Palestine, Abu Hashhash contends, “[T]he act of martyrdom… has gradually undergone a metamorphosis into a heroic act of resistance that breeds its own secular ethics, the martyr has become progressively consecrated.”331 He suggests that in Palestine, martyr posters represent the fallen as courageous heroes rather than victims. In Northern Ireland, martyrdom posters appear to fulfill both narratives. Martyrdom visuals were very common among republicans. They commemorated and memorialized Volunteers killed in action and stressed the continuity of fighting an independence war. But they also worked to provoke outrage and to evidence, as the IRA saw it, the British government’s unwarranted assassinations of freedom fighters. Many martyrdom posters feature early IRA Volunteers, particularly those who participated in Easter Rising and the Irish War of Independence. Others feature the 1981 hunger strikers: that year, 10 republican men died while demanding Special Category Status.332 During the Troubles, republicans also produced posters memorializing Volunteers killed by the RUC, British troops, loyalist paramilitaries, or premature bomb explosions. IRA martyrdom posters feature women militants. Their depiction is not as overtly gendered as mothering visuals: women and men martyrs are generally presented in similar ways. A 1990 Republican Movement poster features 36 Volunteers killed between 1970-1984 and includes both men and women. Still, most martyr visuals depict women and men separately and in this way call specific attention to women. For example, a 1990s image produced by unknown republicans is titled “Women’s Roll of Honour: The Ultimate Sacrifice” and lists women in killed in action (Figure 3.18). Although these images themselves do not invoke the same gender tropes as visuals of mothers or women with bin lids, the common choice to depict women separately suggests that there is something propagandists perceive to be unique about female martyrs and that they believe their audience will respond positively to this segregation. This may be because, as I discuss in Chapter 2, women dying for an insurgent cause is the ultimate symbol of exigency, state brutality, and the 331 Mahmoud Abu Hashhash, 2007. “On the Visual Representation of Martyrdom in Palestine,” Third Text 3-4: pg. 391- 392. 332 The first hunger strike to demand political prisoner status began in 1980 but ended without any fatalities. Three women in HM Armagh joined the 1980 strike. 87 envelopment of conflict. This can be a useful visual securitization strategy that both paints the state as monstrous and invokes sacred, gendered symbols. And this approach has broad appeal: female- perpetrated terrorist attacks generate eight times the amount of media attention as men do when they use the same tactics.333 Bloom suggests that this is one reason that rebels encourage female combatants, to focus local and international attention onto their cause.334 In addition to attention, women dying for insurgency may also direct support. The FARC found that when they sent women to the front- line, local outrage grew against state forces because firing on insurgent women was so unpopular.335 One illustrious example of this securitization strategy is the depiction of IRA Volunteer Mairéad Farrell. In 1988, British Special Air Service (SAS) soldiers shot and killed PIRA members Farrell, Seán Savage and Daniel McCann in Gibraltar. The group was planning to plant a bomb in the main town area. SAS tracked them to a gas station where, even though all three were unarmed, they shot and killed them on sight. Witnesses testified that the SAS shot Farrell and McCann while they were attempting to surrender. British forces did not find a bomb or detonator in the car, though Farrell’s keys led them to a stockpile of Semtex explosives. In 1995, the European Court of Human Rights (ECtHR) ruled that although the victims were engaging in terrorism, the British government had infringed on Farrell and her compatriots’ Article 2 rights because they were unarmed. Farrell had served over a decade in prison for IRA activities by the time she killed, and nationalists mourned her as a true patriot. Republicans and nationalists in Northern Ireland responded with outrage to what they viewed as an assassination. Belfast erupted with violence in the period following their deaths, and a loyalist paramilitary member attacked the communal ‘Gibraltar Three’ funeral and killed three mourners. While these events solidified the ‘Gibraltar Three’ in nationalist memory and generated extensive coverage from the IRA and from the media, all have payed unique attention to Farrell. Republican propaganda often invokes her image and wisdom (Figure 3.19). For example, Figure 3.18 features a quote from Farrell. Indeed, Bloom and colleagues contend that Farrell’s death inspired women to join the PIRA, and that she stands out because of the “strategic use of her image for propaganda purposes.”336 They note that Farrell, functioning under the belief that the Irish people had a legitimate right to end British occupation through physical force, became a PIRA spokesperson while 333 Mia Bloom, 2017. “Women and Terrorism,” Contentious Politics and Political Violence, Groups and Identities, Political Psychology, World Politics, Oxford Research Encyclopedia. 334 Ibid. 335 Herrera and Porch, “‘Like Going to a Fiesta.’” 336 Bloom et. al. “Tiocfaidh ár Mná.” 88 incarcerated before her death and was featured on several recruitment posters.337 While she is not depicted very differently than male Volunteers killed in action, but Farrell’s killing certainly received comparatively more attention from republican propagandists and memorialists than her two colleagues. For example, in 2011, a Republican Network for Unity speech named Farrell alongside preeminent republican figures, including James Connolly, Bobby Sands, and Countess Markiewicz.338 This is probably due in part to her long-term status within the IRA, but likely also results from the local purchase of female martyrs in catalyzing beliefs about women’s victimhood and the legitimacy of insurgent violence. Conclusion Many conflict components affect how civilians feel about insurgencies. In the Northern Irish conflict, the IRA navigated ethnonationalist divisions of religion, class, and culture to generate legitimacy and support for their cause. Women contributed enormously to the IRA in this respect. As front-line Volunteers, auxiliaries, and political leaders they assisted the organization’s survival throughout decades of rebellion. This case elucidates how female militancy, beyond being an important and understudied phenomenon, fundamentally shapes rebel organizations’ publicity strategies and engagement with their audiences. I conclude that the organization securitized female participation and victimization in particularly gendered ways to build legitimacy among sympathetic and skeptical communities, to strengthen their image as a ‘people’s revolution,’ and to increase support for their activities. Through emotionally compelling, historically rooted imagery, the IRA stimulated widely held beliefs of female vulnerability and victimhood to present militant and victimized women as products of imperialist state violence. Women provided the IRA with some of its most consistent propaganda opportunities during the Troubles. The IRA’s securitization of women sold a valuable narrative of group struggle and survival to legitimate the group’s cause. The IRA exploited this imagery, employing traditional gender beliefs about passivity and submission to deploy women as the face of anti-state ideology. By invoking Cumann na mBan, women in other liberation wars, and wide range of female militant activities, IRA visuals emphasized historical continuity and broad participation. These images served to enhance group legitimacy as a revolutionary agent and diminish accusations of fringe terrorism. IRA propaganda also legitimated republican violence by emphasizing state victimization of nationalist 337 Ibid. 338 Ibid. 89 women. Their publications focused disproportionately on women and girls killed by plastic bullets despite the comparatively low number of female victims. The IRA, Sinn Féin, and other political groups also made strip-searching a focal point in their activism. These images are emotionally mobilizing across political lines. They further humanized the high costs of republican struggle and directed attention to gendered state violence. The need to protect women from state brutality was an effective rhetorical, and visual, tool for nationalists. It also had broad appeal for Protestants and unsympathetic groups, many of whom came out in protest of strip-searching and other forms of female victimization. Like female militants in other organizations, IRA women contributed to tangible group successes. They expertly smuggled weapons, planted bombs, lured targets, and gathered intelligence because they were unexpected actors. This case demonstrates that women also provide a strategic advantage in realizing the political narratives that potentially shape conflict trajectory. The IRA visually securitized women as victims, militants, mothers, and martyrs in their political images to construct narratives of threat and legitimize political violence. We can see from these images that rebels clearly believe that women’s participation affects observer attitudes about rebel grievances. They rely on these highly gendered tropes of women’s passivity and natural position in society to present female militants as irregular and therefore extraordinary and legitimize the group’s activities among civilians. In the next chapter, I adopt the same iconological approach to explore how other rebel organizations visually securitize gender in their propaganda. I explore Palestinian political posters as well as gendered political visuals from Lebanon, Namibia, South Africa, Zimbabwe, Afghanistan, Mozambique, Angola, Bangladesh, Nicaragua, and from transnational jihadist groups. I suggest that insurgents across conflicts and across time draw on similar gendered narratives to legitimate their violence, and I demonstrate that rebels cross-nationally rely on the same visual securitization strategies to influence civilian support. 90 Figure 3.5: “Stop Strip Searches,” 1984-1985, Sinn Féin, Linen Hall Library 91 Figure 3.6: “Stop the Strip Searches in Armagh,” 1984, Belfast Strip Searching Committee, Linen Hall Library 92 Figure 3.7: Unnamed. 1980s, Sinn Féin, Linen Hall Library 93 Figure 3.8: “Death Bullets,” 1981, Republican Movement, Linen Hall Library 94 Figure 3.9: “Free Our Country,” 1990s, Cumann na mBan. Linen Hall Library 95 Figure 3.10: “The Soldiers of Cumann na mBan," 1970s, Cumann na mBan, Linen Hall Library 96 Figure 3.11: “A Woman’s Place is in the Struggle,” 1999. IRA. Linen Hall Library 97 Figure 3.12: Unnamed. 1980s, Sinn Féin, Linen Hall Library 98 Figure 3.13: “Women in Struggle,” 1990s, Unknown producer, Linen Hall Library 99 Figure 3.14: “Taking Aim,” 1975. Brendan Murphy. Eyewitness: Four Decades of Northern Life 100 Figure 3.15: “A is for Armalite…,” 1970s. Republican Movement, Linen Hall Library 101 Figure 3.16: “Release Mary Kennedy!,” 1974-1975, United Front Against Internment, Linen Hall Library 102 Figure 3.17: “Bring All Our Prisoners Home,” 1980s, Sinn Féin, Linen Hall Library 103 Figure 3.18: “Women’s Roll of Honour: The Ultimate Sacrifice,” 1990s, Producer Unknown, Eileen Hickey Republican Museum 104 Figure 3.19: “Mairéad Farrell: Irish Republican Army” 1988, Sinn Féin Women’s Department, Linen Hall Library 105 Chapter 4: Gender in Rebel Images: Cross-National Evidence The IRA is not unique in its securitization of women and gender in political visuals during conflict. Indeed, rebel organizations cross-nationally invoke the same underlying gender tropes, and, often, very similar images, to legitimize their campaigns. In this chapter, I extend the theoretical framework and empirical analysis of Chapter 3 to other groups to demonstrate that insurgents across conflicts see value in advertising women’s participation in particularly gendered ways. Moreover, I show that rebels securitize gender more broadly in their visuals across contexts, namely by invoking female civilians. I do this through a case study of Palestinian groups and an overview of visuals produced by rebels in Afghanistan, Lebanon, Namibia, South Africa, Zimbabwe, Bangladesh, Nicaragua, Mozambique, and Angola. The Palestinian case is a useful comparison because, like the IRA, it is an ethnonationalist conflict with deep-rooted religious overtones and generational struggle over nationhood and land. As a result, there are many potential characteristics that insurgents could securitize to appeal to civilians. However, there are some key differences between the cases. Unlike the IRA, Palestinian factions varied in the importance of their religiosity and the religious platforms of some groups evolved over time.339 While women’s visible participation in the IRA largely decreased by decade, women’s involvement in Palestinian attacks grew substantially after the First Intifada.340 In this chapter, I again adopt an iconological approach to identify securitization strategies and to demonstrate that Palestinian factions securitized women’s participation and gender more broadly to build legitimacy. They did so in ways that relied on the gender tropes of female victimhood and periphery to conflict and that exaggerated women’s front-line participation to signal the exigency of circumstances. Palestine is perhaps the only case where the archival collection of visual propaganda rivals Northern Ireland. The Palestinian Poster Project Archive, for example, hosts over 11,000 posters of many varieties, including propaganda. This enables me to engage more than a superficial investigation of visual materials. In the sections that follow, I first discuss women’s participation in Palestinian rebel groups and in Palestinian resistance more broadly. I then examine Palestinian poster production in the post- 1964 Palestinian-Israeli conflict and explore how these factions securitize gender and women’s participation in their political visuals.341 I find striking similarities between IRA and Palestinian 339 Devorah Margolin, 2016. “A Palestinian Woman's Place in Terrorism: Organized Perpetrators or Individual Actors?” Studies in Conflict & Terrorism 39(10): 912-934. 340 Ibid. 341 Yasser Arafat founded the PLO in 1964, marking the modern phase of the conflict. 106 securitization strategies, and suggest that depictions of women militants and Palestinian civilians compose a substantive part of Palestinian poster campaigns. I identify fear of and victimization by the Other and reverence of the sacred as the primary securitization strategies in Palestinian propaganda. Following the Palestinian case study, I demonstrate the transportability of gender tropes by analyzing how cross-national rebel visuals invoke these securitization strategies. Women in Palestinian resistance The major Palestinian factions, except for Hamas, are organized under the Palestinian Liberation Organization (PLO). These include Fatah, the Popular Front for the Liberation of Palestine (PFLP), and the Democratic Front for the Liberation of Palestine (DFLP), among others. Women historically play integral roles in Palestinian resistance across factions, but their participation varies substantially between and within groups. Women in the PLO, broadly, are most often found in auxiliary positions. According to the head of the Palestinian Women’s Union, “The participation of women in the PLO is very low. In 1964, the first Palestinian National Congress resolved to include women in all parts of our struggle. It said women should be equal in rights and duties. Since then, however, these resolutions haven't been put into effect!”342 Some PLO women train to handle small arms and explosives, but they rarely enter the front-lines on behalf of the organization. PLO women do, however, take on important leadership roles in the group. Wives and sisters of male PLO members are particularly likely to attain leadership roles, and women actively participate in the PLO’s executive and legislative institutions. There are high-ranking political women in every PLO faction.343 The General Union of Palestinian Women (GUPW), founded in 1965, is the most prominent women’s political organization in the PLO. In Fatah, women participate as logisticians, recruiters, and suicide bombers. Women participate in Fatah’s armed unit, the al-Aqsa Martyrs' Brigades, but primarily take on auxiliary roles.344 For example, Fatah women participated in clandestine intelligence activity against the Lebanese Army in 1969.345 During the First Intifada, women fundraised, helped coordinate mass demonstrations, and smuggled leaflets, often hiding them in babies’ diapers to avoid suspicion.346 Fatah was so conservative 342 Christine Dugas, 1981. “Women in the PLO: rifles, fatigues, but no veils,” Christian Science Monitor, 31 July. 343 Amal Kawar, 1996. Daughters of Palestine: Leading Women of the Palestinian National Movement, State University of New York Press. 344 S.V. Raghavan, V. Balasubramaniyan. 2014. “Evolving Role of Women in Terror Groups: Progression or Regression?” International Journal of Women’s Studies 15(2): 197-211. 345 Rosemary Sayigh, 1998. “Gender, Sexuality, and Class in National Narrations: Palestinian Camp Women Tell Their Lives,” Frontiers: A Journal of Women Studies 19(2): 166-185. 346 Cragin and Daly, Women as Terrorists. 107 when it came to women on the front-line that Leila Khaled, a Palestinian militant who famously hijacked multiple airplanes in 1969 and 1970, left the faction to join the PFLP. Less is known about women in the PFLP, but at least 20 women – including Khaled – trained at a camp in Lebanon and women actively participated as suicide bombers.347 The DFLP, arguably the most leftist PLO group, incorporated women on a platform of women’s liberation and viewed women’s emancipation as a key part of national liberation. Sayigh argues, “This gave DFLP women a stronger profile than in other Resistance groups: the party was characterized by ‘openness to women at all levels’… it was the first Resistance group to appoint women to leadership positions; and in the Occupied Palestinian Territories DFLP women members acquired sufficient power to create independent frameworks for women’s action.”348 Still, women primarily participated in support roles due to gendered restrictions keeping women largely off the front-line.349 A unique feature of women’s participation in Palestinian rebellion is that front-line women are more likely to commit what Margolin calls “personal initiative” attacks rather than carry out missions on behalf of organizations. Margolin concludes that between 1965-1986, women participated frequently with Palestinian factions. However, the rise of conservative religious organizations like Hamas during the First Intifada in the late 1980s largely pushed front-line women into auxiliary positions.350 As a result, women began carrying out attacks on their own volition rather than those sanctioned by an organization. This shift led to a marked increase in the number of female-perpetrated terrorist attacks in Palestine.351 Women and girls continue to carry out personal initiatives attacks, including knife attacks against Israeli soldiers. Still, even the religious organizations are occasionally open to female attackers acting under the group’s authority. In 2004, Reem al-Riyashi killed four Israelis in a suicide bombing jointly organized by Hamas and the al-Aqsa Martyrs' Brigades. In 2005, Hamas announced a women’s military wing. Palestinian women also participate in support activities outside of organized groups: they organize boycotts, form medical supply committees, lead protests, nurse the wounded and sick, maintain the educational system, and raise martyrs.352 The Palestinian resistance extends far beyond 347 Ibid. 348 Rosemary Sayigh, 2010. “Book Review,” Feminist Review 96, pg. e11; Frances Susan Hasso. 2005. Resistance, Repression, and Gender Politics in Occupied Palestine and Jordan. Syracuse University Press. 349 Frances Hasso, 2005. Resistance, Repression, and Gender Politics in Occupied Palestine and Jordan, Syracuse University Press. 350 Margolin, “A Palestinian Woman’s Place in Terrorism.” 351 Ibid. 352 Gonzalez-Perez, Women and Terrorism. 108 organized rebellion. Particularly during the First Intifada, women engaged in sustained democratic activism and organized resistance campaigns through mass civil society organizations.353 In technical terms, these women were civilians, but their activities directly assisted the struggle for Palestinian statehood. Like women auxiliaries within organized insurgent organizations, accounts of the Palestinian-Israeli conflict often render these women’s contributions invisible.354 Women in rural villages, refugee camps, and urban centers experience resistance in profoundly different ways, with the latter being a much less conservative environment.355 However, as perceived bearers of the nation in a long and brutal war, women are historically important symbols in Palestine. Rubenberg concludes, “Women have been an important sign, a marker, for the political goals of the Palestine National Movement,” shaped by “dispossession, dislocation, occupation, and the nationalist struggle to achieve political independence and to assert a national identity.”356 Consequently, the armed factions visually invoke Palestinian women as symbols of nationalism and struggle. Palestine has a rich tradition of political visualization, namely through the production of political posters, and women feature prominently in those produced by armed groups and larger nationalist groups alike. Visualizing the Palestinian conflict While postering has a long Palestinian history, scholars consider the birth of the modern Palestinian poster to be the years following 1967’s Six Day War.357 Walsh argues, “it is clear that even before the dust settled, the PLO began mobilizing the poster to project its ‘victorious’ version of the battle.” 358 This spawned rapid mobilization of the Palestinian poster, which “is a reflection of the PLO’s determination to control the narrative – to ‘re-produce’ the culture of Palestinian nationalism.”359 While many poster artists were untrained, armed groups began recruiting professional artists to create visual material. For example, Swiss artist Marc Rudin, also known as Jihad Mansour, regularly produced posters for the PFLP. The factions themselves established publicity units to glorify the resistance movement and highlight those who fought and died in the struggle. As Maasri concludes, The struggle for the liberation of a lost homeland that the various movements engaged in was coupled with relentless efforts in organized media activity to advocate the Palestinian cause. 353 Penny Johnson and Eileen Kuttab, 2001. “Where Have All the Women (and Men) Gone? Reflections on Gender and the Second Palestinian Intifada,” Feminist Review 69: 21-43. 354 Ibid. 355 Cheryl Rubenberg, 2001. Palestinian Women: Patriarchy and Resistance in the West Bank, Lynne Rienner Publishers. 356 Ibid, pg. 16. 357 Daniel Walsh, 2011. The Palestinian Poster Archives: Origins, Evolution, and Potential, Master’s thesis, Georgetown University. 358 Ibid, pg. 35. 359 Ibid. 109 The media office of the Palestinian Liberation Organization, based in Beirut, acted as the Unified Information Office and included a fine arts department that catered for the art-related activities of the organization, including the publishing of posters… Posters, among other means, played a key role in awakening or maintaining national identity and mobilizing the youth to actively partake in the resistance, as well as calling for international solidarity with the Palestinian people in their struggle for liberation and social justice.360 PLO’s Unified Information Office created their own posters, and they provided technical assistance to Fatah and other factions. Fatah, the DFLP, and the PFLP all actively engaged in poster production, and artists often created images for multiple groups.361 The 1993 Oslo Peace Accord enabled Palestinians to operate printing presses without Israeli oversight, leading to even more domestic production: previously, factions often commissioned printers overseas.362 Over time, the posters have evolved in style and artistry, which Walsh argues reflects “an awareness that more painterly, more diverse, and more appealing imagery could extend the audience.”363 Palestinian posters are designed to be posted in refugee camps, in Palestinian homes, and to be circulated abroad.364 One scholar argues that factions use posters to “romanticize” Palestinian resistance and “push their own agenda.”365 Palestinian political posters feature recurring nationalist symbols, and these images establish the continuity between early posters of the 1960s and those produced today. All posters contain “the core elements of Palestinian nationalist credo- ‘armed struggle,’ ‘national unity,’ and ‘an independent state.”366 This is represented through common iconography. Some examples include Jaffa oranges (symbolizing Arab-Jewish cooperation), barbed wire, chains, and shackles (symbolizing oppression and occupation), a clenched first (symbolizing resistance), olive trees (symbolizing Palestinian national rootedness), the Palestinian flag, and the keffiyeh, the black and white checkered scarf worn by Palestinian activists. The posters also often feature women as militants, civilians, and personified versions of the Palestinian nation. However, despite the frequent visualization of women, Rudin notes that there were no woman designing posters.367 A key difference between Palestinian political visuals and those produced by the IRA is that Palestinian factions enjoy greater support from their civilian population for their struggle and tactics 360 Maasri, Off the Wall, pg. 38. 361 Walsh, The Palestinian Poster Archives. 362 Alaa Daraghme, 2017. “A Brief History of Palestinian ‘Martyr Posters,’” VICE UK. 363 Walsh, The Palestinian Poster Archives, pg. 38. 364 Interview with Jihad Mansour, https://www.nadir.org/nadir/initiativ/rev_linke/sanat/jihad/jihad1.html 365 Dr. Walid Al Shurafa, quoted in Daraghme, “A Brief History of Palestinian ‘Martyr Posters,’” n.p. 366 Walsh, The Palestinian Poster Archives, pg. 41. 367 Interview with Jihad Mansour. 110 than the republican groups did. For example, in 2001 61 percent of Palestinians surveyed believed that “armed confrontations helped achieve Palestinian national rights in a way that negotiations could not.”368 Levels of support for armed attacks were also high: 92 percent of those surveyed supported hypothetical attacks against Israeli soldiers and settlers, while 58 percent supported hypothetical attacks against Israeli civilians.369 The PLO’s legitimacy has also grown in the eyes of Israelis over time: in 1987, only 21 percent of Israelis agreed that a Palestinian state should be established as part of a peace agreement. In 1996, those in favor of Palestinian statehood increased to 48 percent.370 Therefore, while republican political visuals may attempt to generate legitimacy among potentially sympathetic – and hostile – populations, Palestinian posters may be more focused on sustaining the legitimacy of and faith in the statehood objective. As I discuss further below, some scholars view posters as a method of endurance in the face of repeated, long-term trauma. I collect political visuals from the Palestine Poster Project Archives (PPPA). The PPPA includes 12,023 posters produced during the Palestinian-Israeli conflict, 4,781 of which belong to Palestinian factions and artists. The PPPA includes 2,308 images featuring women, produced by a variety of nationalist agencies, governments, insurgencies, and individuals. For continuity between cases, I limit my analysis to posters produced by organized Palestinian factions, including the PLO, Fatah, the PFLP, and the DFLP. I collect, non-systematically, posters featuring women produced by these groups and use them to explore how insurgencies cross-nationally invoke gender as a securitization strategy. I also use one Hamas-produced poster sourced from their publications, not the PPPA. In the sections that follow, I emphasize the similarities and differences between IRA political images and those produced by Palestinian groups and discuss how the gendered depictions of women transcend borders in political visuals. Fear of and victimization by the Other Palestinian posters emphasize Israeli violence and violence from paramilitaries perpetrated against Palestinian civilians. Often, this is represented through depictions of victimized women. As I discuss further blow, these visuals, and all visuals featuring women, depict women as mothers more commonly than in republican posters. In 1970, the PLO released a series of text-less posters that include suffering 368 Palestinian Center for Policy and Survey Research, 2001. Palestinian Public Opinion Poll No. 3, “Palestinians support the ceasefire, negotiations, and reconciliation between the two peoples but a majority opposes arrests and believe that armed confrontations have helped achieve national rights.” 369 Ibid. 370 Sarit Yalov, 2001, “A Palestinian State and Israeli Public Opinion,” in Arieh Stav, ed. Israel and a Palestinian State: Zero Sum Game? Zmora-Bitan Publishers and ACPR Publishers. 111 Palestinian women. In one example, an agonizing woman holds her suffering child (Figure 4.1). In the background, perceivably dead and suffering women lay on the ground or clutch their children. The poster’s rusty coloration suggests that they are bloodied. The image is visceral, clearly intended to convey the anguish of Palestinian mothers without contextualization. Palestinian posters also use similar depictions of women to reference state- and militia- perpetrated massacres and assaults on Palestinian refugee camps. In 1976, the Christian militia Lebanese Forces (LF) attacked the Tel al-Zaatar Palestinian refugee camp in Beirut, Lebanon. The assault followed several other Christian militia attacks on refugee camps. Approximately 2,000 Palestinians were killed during the Tel al-Zaatar siege; PLO and PFLP fighters killed approximately 200 LF militants.371 In 1982, the Kataeb Party, another Christian militia purportedly backed by the IDF, killed between 2,000-3,5000 Palestinian and Lebanese residents of the Sabra neighborhood and Shatila refugee camps in Beirut. The victims were primarily women, children, and the elderly.372 Palestinian camps were besieged again between 1985-1988 during the Camps War, when Lebanon’s Amal Movement, allied with Hafez al-Assad’s Syrian Defence Forces (SDF), and other groups took control of West Beirut and attacked refugee camps. The PLO, PFLP, and other Palestinian groups responded from within and outside of the camps, and the fighting continued with varying intensity for four years. Palestinian posters regularly invoke this carnage. For example, a 1976 PLO poster features a photograph of a bloodied, crying woman paired with smaller images of fleeing civilians. It reads, “Tel Azaatar [sic]- 3,000 Martyrs.”373 Another poster, produced by Fatah in 1985, includes a photograph of a Palestinian mother fleeing a burning refugee camp with her two children. She and her children are anguished as they walk through rubble. The text reads: “Palestinian Diaspora: For How Long?” A 1985 PLO poster depicts a bloodied woman and her child next to a blood-spattered wall (Figure 4.2). The poster reads: “They did not spare them, providence did.” This poster directly refutes the IDF and Kataeb Party’s claim that they purposefully spared survivors of the Sabra and Shatila massacres, instead suggesting that divine intervention saved women and children – the most vulnerable.374 Compared with the IRA’s and republican posters, Palestinian posters in the fear of and victimization by the Other vein feature more bloodied violence against women. There is a noticeable 371 United States Army Human Engineering Laboratory, 1979. Military Operations in Selected Lebanese Built-up Areas, 1975– 1978, Technical Memorandum 11–79. 372 Kathryn Kabib, 2017. “Dropping by Shatila, a Wounded Place,” The Washington Report on Middle East Affairs 36(4), June/July: 24-25. 373 Poster and translation provided by the PPPA. 374 PPPA curator’s note, “Providence Did.” 112 lack of sexual violence in Palestinian posters, especially when compared to those produced in other conflicts. In contrast to the republican visuals calling attention to violent strip searching and the use of rape as a visual metaphor for the invasion of Ireland by the United Kingdom, I find no evidence of sexually violent motifs in Palestinian posters. This may be explained, in part, by the comparatively low levels of sexual violence perpetrated during the Palestinian-Israeli conflict. Wood notes that although the Israelis perpetrated some sexual violence during the forced movement of Palestinians from some areas in 1948, presently neither the IDF nor Palestinian factions perpetrate sexual assaults despite civilian-targeted attacks on both sides.375 But, as I discuss in Chapter 2, depictions of sexual violence as justification for retaliatory political violence are not always predicated on fact. The absence of sexual victimization as a securitization strategy among Palestinian visuals is thus a curious phenomenon that requires future analysis. Women militants Like their Northern Irish allies, Palestinian factions feature armed female militants in their political imagery with a frequency and in situations disproportionate to women’s actual involvement on the front-line. While women are infrequently armed actors in Palestinian groups, armed women are a very common image in Palestinian posters. They are depicted both on their own and with armed men. For example, a 1978 Fatah poster includes a drawing of a woman holding a rifle in one hand and a dove and olive branch – symbols of peace – in the other. The poster reads, “Liberation, Victory, Return.” A 1968 PFLP poster, possibly one of the group’s first visuals, shows only a woman holding a rifle and announces the PFLP. In 1977, Fatah and the PLO United Information Office produced a poster that directly links women’s front-line involvement to the sieges of Palestinian refugee camps in Beirut. The poster depicts six women holding rifles and a grenade launcher (Figure 4.3). The text reads, “Tel Azaatar [sic]. The weapons of the revolutionaries and the masses created the epic of Palestinian steadfastness.” This poster suggests that the brutality of the Tel al-Zaatar massacre mobilized women into armed action. It indicates that the women’s participation is the result of a ruthless war where even the clearest conflict norms are broken. Fatah produced a similarly themed poster in 1986 that includes a photograph of two women dressed in fatigues and brandishing their rifles in the air (Figure 4.4). This image is among the most similar to posters featuring male militants, the women appear dangerous and 375 Elisabeth Jean Wood, 2010. “Sexual Violence during War: Toward an Understanding of Variation,” in Laura Sjoberg and Sandra Via, eds. Gender, War, and Militarism: Feminist Perspectives, ABC-CLIO, pgs. 124-137. 113 dedicated in their facial expression and attire. However, rather than drawing attention to militancy and the necessity of violent struggle, the poster reads, “In Defence of Peace.” This juxtaposition between militant, armed women and the call to peace highlights the boundaries of women’s access to violence and to war. Even when they most resemble their male counterparts, the narratives of women’s militancy converge on protecting peace and passivity. Palestinian posters also depict women alongside male militants. These visuals emphasize the envelopment of the conflict and the need for a people’s rebellion in response. The DFLP produced a poster in 1976 that depicts two men and two women collectively holding a rifle with the text, “All arms are prepared for training and carrying arms (armed struggle).”376 The insurgencies also feature unarmed women on their posters to underscore the communal nature of the conflict. For example, the PLO printed their first poster in 1965. It depicted five people holding up the text, “Palestine Liberation Organization.” Each person represents a different social class: there is a laborer, a businessman, a Bedouin, an armed militant, and an auxiliary woman. The poster read, “We are all sons of Palestine.” In 1967, the PLO reprinted the poster but changed the text from “We are all sons of Palestine” to “We are all for the resistance” (Figure 4.5). According to the artist, Ismail Shammout, the PLO updated the text to better convey the poster’s central meaning of a collective resistance that required all people’s - including women’s - contributions.377 A very common trope in Palestinian postering is the armed mother. The armed mother melds traditional expectations of femininity and nurturing with women’s battlefield presence to legitimate struggle in a way that perhaps no other imagery can. As Naaman succinctly summarizes, When women opt to fight alongside men, they challenge the dichotomy of woman as victim/man as defender. Women fighters are physically strong, are active (therefore agents), and, most important, are willing to kill (hence, they are violent). They challenge not only the images of women as victims of war but also the traditional patriarchal binary opposition that postulates women as physically and emotionally weak and incapable of determining and defending the course of their own lives.378 As a result, observers often attribute women’s violence to deviance from their ‘proper’ roles as women. Armed mothers, however, directly juxtapose a ‘sacred,’ private gendered experience – literally reproducing the nation – with the savagery of war. Armed mothers do not shy away from their femininity, they demonstrate that even the ‘ultimate’ woman, the mother, must participate on the 376 Translation provided by PPPA. 377 Translation provided by PPPA. 378 Dorit Naaman, 2007. “Brides of Palestine/Angels of Death: Media, Gender, and Performance in the Case of the Palestinian Female Suicide Bombers,” Signs 32(4): pg. 935. 114 front-line to protect her child and her nation. The armed mother is desperation. She is legitimacy embodied, she moralizes the struggle. She is sacrificial. One such poster, produced by the PFLP in 1986, depicts an armed woman wearing a keffiyeh and holding her baby, who has his hand out as to protect himself from an aggressor (Figure 4.6). In an interview, Rudin, the artist, notes that he intentionally chose to color the background green to symbolize the “fertility” of Palestine. At the same time, he contends that the keffiyeh is “not a representative symbol of the women, but only women under arms carry it.”379 This melding of a symbol highly associated with womanhood and a symbol primarily associated with public resistance and war perfectly encapsulates the visuality of the armed mother. She is, at once, a beacon of womanhood, protectionism, necessity, violence, and resort. Other posters echo these depictions of fertility and reproduction. A 1980 Fatah and Palestine Martyrs Works Society (SAMED) features a woman and her child with the text, “We produce, fight and continue our march with determination.”380 The woman wears a crown of weapons and tools. A 1978 GUPW poster depicts an armed woman in the foreground and reads, “The woman; Mother and fighter on the path to liberation; The General Union of Palestinian Women is a foundation of the revolution.”381 In 1980, the PFLP circulated a poster for international women’s day featuring an armed mother and her baby (Figure 4.7). The organization, and outside groups, have reproduced it on multiple posters celebrating Palestinian women’s resistance. In addition to holding her baby, the armed mother in the poster is also wearing a dress with two images. One image, directly above the baby, is a caged woman, who looks vulnerable and has a similar facial expression to the baby. In the other image is a woman wearing a keffiyeh, not behind bars. These visuals capture the dichotomy of the armed mother as a visual symbol: she is both protecting and in need of protection. She is forgiven, and even encouraged, for her shift into the public conflict space because she remains, foremost, a woman. Reverence While armed mothers exemplify victimization by the Other as a securitization strategy, they also evidence reverent narratives in Palestinian posters. Factions regularly depict unarmed mothers with similar reverence, as sacred symbols of nationalist resistance. For example, a 2009 Fatah poser features 379 Interview with Jihad Mansour, np. 380 Translation provided by the PPPA. 381 Translation provided by the PPPA. 115 a Palestinian woman protesting, yelling towards the sky and holding the Palestinian flag above her head. Superimposed over the photograph is the text, “The Palestinian woman protects our sacred flame” (Figure 4.8). Though she is not directly labeled a ‘mother,’ the sentiment of women protecting the ‘sacred flame’ is palpable: Palestinian artists use the “Palestinian flame” in other posters to symbolize resistance and nationalism. The wording of the text suggestions that “the” Palestinian woman – all Palestinian women – share this responsibility. Palestinian posters also personify the Palestinian nation as a woman- as mother Palestine. In 1984, Fatah published a remembrance poster commemorating the second anniversary of the Sabra and Shatila massacres. The text reads, “Despite the massacre we will be reborn anew” and the image depicts a woman rising above dead massacre victims. She is clearly intended to represent being “reborn anew.”382 Similarly, a 1983 PLO poster recalling the massacres in Beirut reads the “steadfastness of Beirut” and prominently features an unarmed Palestinian woman with an army positioned behind her. She is the steadfastness of Palestine under siege.383 A 1983 Fatah and PLO Unified Information Office poster commemorating the 18th anniversary of the PLO’s launch, titled “Mother Palestine,” depicts a woman holding the city in her hands (Figure 4.9). There are Jaffa oranges behind her, symbolizing peace in the Palestinian struggle. Another poster, produced by Fatah in 1980, declares the PLO the Palestinian people’s “sole legitimate representative.” The poster anthropomorphizes PLO as a woman, and the text is inscribed into her dress. Martyrs Another example of the reverence securitization strategy is the visualization of martyrs. Palestinian organizations have a long history of producing martyrdom posters, which have flourished since the Second Intifada and rival perhaps any other conflict in their frequency and consistency. Abu Hashhash concludes that martyrdom posters are the main form of visual representation of this act in public space, and that they innately “imply the idea of sacrifice” for religion or for Palestine (not mutually exclusive).384 As evident in Chapter 3 and in visuals across conflicts, martyrdom posters serve several functions. As public obituaries, they memorialize. As glorifications of political violence, they recruit. As signals of exigency, they evidence the state’s brutality and the necessity of sacrifice. As an on-going practice, they position the struggle within its historical continuity. Still, Palestinian 382 Translation provided by the PPPA. 383 Translation provided by the PPPA. 384 Abu Hashhash, “On the Visual Representation of Martyrdom in Palestine,” pg. 391. 116 martyrdom posters are unique. The length of the conflict, the total enmeshment of Palestinians living under Israeli occupation in high- and low-intensity generational violence, and the everyday normalization of martyrdom as a feature of resistance means that “rituals that idolise [sic] martyrs have become a necessity to help people cope with repeated loss.”385 Some martyrs even design their own posters, picking out their desired images before their deaths.386 These posters’ place in social consciousness and political activism is singularly Palestinian. Maasri identifies competitive martyrdom postering between parties to the Lebanese Civil War (1975-1990), of which the PLO and its factions were a central part. She notes that through these posters, “The number of fallen heroes becomes an indicator of the party’s share of participation on a front and proof of its commitment and sacrifice in the defence [sic] of an existential cause.”387 As discussed in Chapter 2, women’s sacrifice through death is a particularly strong signal of the direness and moral legitimacy of the struggle. This is because women dying for the cause helps evince political violence as necessary. Female martyrs are uniquely sacrificial, given the gendered expectations of their position as wives and mothers innately deserving protection. This view is so embedded in Palestine that some female martyrs are known colloquially and by their organizations as ‘brides’ of the homeland. For example, after Hanadi Jaradat killed herself and 21 others in a suicide attack in Haifa, Israeli in 2003, the Islamic Jihad Movement in Palestine (PIJ) claimed her as the ‘bride of Haifa.’ They also referred to her attack as a “wedding,” noting that the “wedding in Haifa will teach the Zionists an unforgettable lesson.”388 When Fatah’s Wafa Idris became the first female Palestinian suicide attacker in 2002, a professor and columnist compared her martyrdom to the Virgin Mary’s birth of Jesus: “From Mary’s womb issued a child who eliminated oppression, while the body of Wafa became shrapnel that eliminated despair and aroused hope.”389 Naaman argues that because women are “accepted in national liberation struggles as the symbol of the mother nation,” titling female martyrs ‘brides’ conveniently enables rebels and observers to fit them into resistance narratives without jeopardizing patriarchal nationalist projects.390 Still, this terminology is not applied equally across female martyrs and is often reserved for those that can be presented as particularly feminine.391 385 Ibid. 386 Daraghme, “A Brief History of Palestinian ‘Martyr Posters.’” 387 Maasri, Off the Wall, pg. 87. 388 Vered Levy-Barzilai, 2003. “Ticking Bomb,” Haaretz, 4 June., np. 389 Dr. Adel Sadeq quoted in James Bennet, 2002. “Arab Press Glorifies Bomber as Heroine,” The New York Times, February 11. 390 Naaman, “Brides of Palestine/Angels of Death.” 391 Ibid. 117 Like republican visualizations of IRA Volunteer Mairéad Farrell, Palestinian factions also elevate certain ‘martyr-heroes’ in their political posters.392 A prominent example is the myriad images of Fatah’s Dalal Al-Mughrabi. Mughrabi participated in the 1978 Coastal Road attack in Israel, hijacking a taxi and a civilian bus, taking 71 hostages, and engaging in a shootout with Israeli forces. Thirty-eight Israelis, including 13 children, were killed in the massacre. Mughrabi was one of eleven attackers, only two of whom survived the attack. Still, she is inarguably the most famous Coastal Road attacker and one of the most famous female Palestinian militants. She is commonly depicted in Fatah and PLO images. For example, a 1978 Fatah poster features a painting of Mughrabi with the text, “Operation ‘Martyr Kamal Adwan’ [the operation’s name] - 11 March 1978. Martyr of the Palestinian revolution. Dalal Al Mughrabi” (Figure 4.10). Another 1978 poster produced by Al-'Asifah – at the time Fatah’s military wing – includes an image of Mughrabi and calls her the “gift that spreads the voice of Al-'Asifah.” A PLO Unified Information Office poster from that same year features a photograph of her in camouflaged fatigues and calls her operation “the path to the homeland.”393 According to Peteet, prior to her death Mughrabi had a salacious reputation that turned into valorization following the Coastal Road attack. Though this information is speculative and should be treated as such, Peteet links this social shift to other Palestinian militant women who become “objects of veneration, but… [through their militancy they become] symbols that are above judgement by the normal standards of female conduct.”394 In 2004, Reem al-Riyashi killed four Israelis and herself in a joint Hamas and al-Aqsa Martyrs' Brigades suicide bombing at the Erez Crossing at the Gaza-Israel border. Riyashi was the wealthy mother of two young children and she left a pre-recorded video message in which she contended, “God has given me two children. I love them [with] a kind of love that only God knows, but my love to meet God is stronger still.”395 Hamas circulated a martyr poster of Riyashi holding her young son, both holding weapons and wearing Hamas headbands (Figure 4.11). The poster is feminine: greenery and pink flowers surround her. The text reads, “The Eraz Crossing Operation executed on January 14, 2004; Reem the Qaasam Martyr.”396 Her poster makes clear that Riyashi is an armed mother, making her martyrdom an ultimate sacrifice for Hamas’ cause. Indeed, Hamas argued that the poster demonstrated “both the depth of despair among Palestinian women and their desire to defeat the 392 Maasri, Off the Wall coins this term. 393 Translations provided by PPPA. 394 Julie Peteet, 1991. Gender in Crisis: Women and the Palestinian Resistance Movement, Columbia University Press, pg. 156. 395 Chris McGreal, 2004. “Human-bomb mother kills four Israelis at Gaza checkpoint,” The Guardian, 14 January. 396 The Al-Qassam Brigades, EQB is Hamas’ military wing. 118 occupation.”397 Israeli military sources suggested that Hamas forced Riyashi into a suicide attack to atone for infidelity in her marriage, a claim that her husband and family vehemently denied. Her husband responded by calling Riyashi an “honorable woman who sacrificed her life for the sake of Islam and Palestine.”398 Still, not all Palestinians received Riyashi’s martyrdom as evidence of the struggle’s necessity. Indeed, despite decades of PLO, PFLP, Fatah, and DFLP visuals depicting armed mothers and women ‘martyrs,’ many Palestinians criticized Hamas for employing a young mother as a suicide bomber and for circulating the photographs of Riyashi and her armed children.399 This response suggests that securitizing gender as a method of legitimacy-building may not always have its intended effects. Gender norms are entrenched, and even the most venerated female militants often remain controversial.d Sheikh Ahmad Yassin, the founder and then-spiritual leader of Hamas, concluded that Riyashi’s martyrdom and, consequently, the valorization of it in Hamas’ images, “opened the door for more women to die in the fight against Israel.400 And perhaps it did. The leader of Hamas’ women’s unit, founded the year after Riyashi’s death, noted that being a mother made Riyashi a particularly inspirational martyr: “The martyr Reem Al-Riyashi is like a crown on our heads and a pioneer of the resistance. Nobody can fathom the magnitude of her sacrifice.”401 Visualizing women cross-nationally These gendered depictions of women as victims of state or opposition violence, as part of a people’s rebellion, as armed mothers, as mothers of the nation, and as martyrs for the cause are not limited to the IRA, to Palestinian factions, or to anti-colonial conflicts more broadly. These securitization strategies, and, often, the same imagery, are reproduced across conflicts cross-nationally. In this section, I briefly explore how insurgents across conflicts use these gendered visuals with striking similarity to influence public perceptions of rebel grievances. Powerful examples of female victims of state violence come from the National Islamic Front of Afghanistan (NIFA) and from the South West Africa People's Organization (SWAPO). The former, which is also known as the Internal Islamic Front of Afghanistan, began as a mujahideen group fighting the Soviet Union in Afghanistan. In the 1980s, they produced Figure 4.12, a poster depicting an 397 Chris McGreal, 2004. “Palestinians shocked at use of suicide mother,” The Guardian, 26 January. 398 Khaled Abu Toameh, 2004. “Bomber’s Husband Denies She Was Coerced for Adultery,” Jerusalem Post, 18 January. 399 McGreal, “Palestinians shocked at use of suicide mother.” 400 Ibid. 401 Middle East Media Research Institute, 2005. “Commander of Hamas Military Wing Women’s Unit: ‘Our Members Yearn for Martyrdom,’” Special Dispatch No. 983. 119 Afghan woman and her baby looking despondent as a hammer-and-sickle bomb barrels down on top of them. There is no text, but the message is clear: the Soviet Union is a direct threat to women (mothers, specifically). SWAPO fought for, and won, Namibian independence from South Africa between 1962-1990. Women played a central role in SWAPO, on the front-lines, in auxiliary roles, and in leadership positions.402 In the 1970s and 1980s, SWAPO produced posters for their domestic audience and to distribute on international tours. Many of these posters reference the violence of the South African state by focusing on women victims. One, shown in Figure 4.13., depicts Namibian women sitting and holding their small children. It reads, “the wound of the daughter of my people wounds me too. Who will turn my head into a fountain and my eyes into a string of tears so that I may weep all day and all night for all of the dead of the daughters of my people.” Many rebel factions also feature armed women in their political visuals. For example, a 1960s Umkhonto we Sizwe (military wing of the African National Congress) poster shows two woman standing at attention and holding grenade launchers (Figure 4.14). It reads, “Women arise and act!” It also makes an explicit call for unity among the population, mirroring the collective, community- focused rhetoric of posters depicting women in other conflicts. Other images depict armed women with armed men. For example, an undated poster produced by the Publicity Department of the Zimbabwe African People’s Union (ZAPU) shows an armed man standing behind an armed woman with his hand on her shoulder. She touches his hand lovingly and the poster reads, “Victory to the People of Zimbabwe.” A FRELIMO poster issued in 1973 commemorates the Mozambican Revolution and shows and armed woman and armed man together (Figure 4.15). She also holds a book, while he holds a farming hoe. The poster reads, “To study, to produce, to fight.” Armed mothers are also very frequently depicted in rebel visuals cross-nationally. A 1985 Hizbollah poster shows a wife and child sitting with the body of their husband and father, a fighter who appears to have been just killed in battle (Figure 4.16). The poster, commemorating “Martyrs’ Day,” reads, “The blood of martyrs demeans the thrones of tyrants and creates the dawn of the nation.”403 The imagery is intense, and it illustrates the process of becoming an armed mother. The woman, looking on as her child cries over his father’s body, holds the rifle that clearly only moments earlier belonged to her husband. Her face stoic, she stares at the rifle as through she is recognizing that she, now, must take it up to protect her family and her people. Importantly, remarkably few 402 M. Bahati Kummba and Ona Alston Dosunmu, 2009. “Women in National Liberation Struggles in the Third World,” in Berch Berberoglu, ed. The National Question: Nationalism, Ethnic Conflict, and Self-determination in the 20th Century. Temple University Press. 403 Translation provided by Maasri, Off the Wall. 120 women fought on the front-line for Hizbollah. While there is evidence of one female suicide bomber in 1985, women’s participation in the group was very limited and one primarily raised funds, educated children, and occasionally smuggled weapons.404 As one scholar argues, motherhood was women’s primary contribution to Hizbollah.405 Another example of armed mothers in visual propaganda is a 1975 People’s Movement for the Liberation of Angola (MPLA) poster featuring the photograph of a female fighter holding her rifle in one hand her and child in the other (Figure 4.17). Her baby also touches the rifle. The MPLA fought the Portuguese army during the Angolan War of Independence (1961-1974) as an outwardly socialist, secular movement. The group included a women’s wing, the Organization of Angolan Women, and women participated as front-line soldiers and in auxiliary roles “in substantial numbers.”406 This imagery is strikingly similar to that of armed mothers in Palestinian posters, as it is to posters from a number of other conflicts. An illustrative undated poster comes from the Sandinista Children's Association during the Nicaraguan Revolution (1962-1990). It depicts a Sandinista mother, with her rifle slung over her back, holding her baby and surrounded by her two other smiling children. As is often typical of posters featuring women, she is flocked by feminine imagery including flowers and doves. The poster reads, “Thank you mother for defending our joy.” A final example is a 1971 Bangladeshi poster produced during the Bangladesh War of Independence (Figure 4.18). The poster includes a painting of a woman armed with a rifle that is topped with a bayonet. It reads, “Bangladesh Mothers and Daughters Fighting for Independence.” Through these visuals, women are condensed to their capacities as mothers and daughters even when actively fighting on the front-lines. In this way, the posters demonstrate the urgency of struggle and maintain the conservative gender expectations to which women in most societies are subject. Insurgents cross-nationally also represent women as martyrs in highly gendered ways. An illustrative 1986 NRF and SSNP poster commemorates suicide bomber Sana’ Mehaidli (Figure 4.19). It features Mehaidli and other female martyrs drawn onto a pink and purple background. Maasri writes, The poster features [Mehaidli] as a role model for women’s active participation on the war front… The soft strokes of the illustration and the overall feminine and romantic expression of the poster contrast with the violence of her death. The call for an active women’s role in resistance is progressive in a patriarchal society consumed by traditional gender roles, yet the 404 Kirsten E. Schulze, 1998. “Communal Violence, Civil War and Foreign Occupation: Women in Lebanon,” in Robert E. Miller and Rick Wilford, eds. Women, Ethnicity and Nationalism: the Politics of Transition, Routledge. 405 Maria Holt, 1999. “Lebanese Shi’i Women and Islamism: A Response to the War,” in Lamia Rustum Shehadeh, Women and War in Lebanon, University Press of Florida. 406 Henda Ducados, 2004. “Angolan women in the aftermath of conflict,” Accord 15, pg. 58. 121 visual representation in this poster does not match that challenge. It remains quite conservative and does not break from the typical stereotype of romanticized femininity. Perhaps it is intentionally designed in that manner to reflect the idea that women can be as active as men on the war front without totally compromising their ‘femininity.’407 Like martyr posters from the Irish and Palestinian conflicts, this image draws explicit attention to the ‘womanness’ of women who die for insurgent struggles. These visuals excel in gender-role management, securitizing women’s participation to shape narratives of threat by reminding the observer that women’s martyrdom is the consequence of state violence. Not all women-inclusive organizations publicize women’s participation. It is useful to briefly consider not only women’s representation, but how the absence of such visualization forms rebel organizational strategies. For example, despite women’s frequent participation in auxiliary roles and extremely occasional involvement on the front-line, official jihadi visuals almost never picture women. Lehane and colleagues identify women in only a handful of the 3,869 images published in five online jihadi propaganda magazines published between 2009-2015, only two of which present women as part of a jihadist group.408 These images, appearing in Al Qaeda’s Inspire, feature women wearing niqabs alongside information on the importance of the covering. The lack of female visualization is particularly striking in Daesh publications because they often feature direct appeals to or discussions of women. For example, the group’s now-defunct Dabiq included in each issue a column purportedly written by a woman intended directly for women readers. These pieces implored women to join jihadist struggle as supporters. Rumiyah, their current publication, also features sections aimed at women. For example, Rumiyah’s tenth issue features a full page labeled “sisters” that is backed with colorful pink and purple flowers and includes the text, “Be a supporter, not a demoralizer.”409 Lehane and colleagues argue that the distinct lack of women in jihadi images “is a deliberate attempt to reinforce traditional gender roles and strengthen existing gender hierarchies within terrorist organizations.”410 It portrays a clear message about women’s place in jihadi groups, and in society, to observers. Daesh has mastered the technological shift of visual imagery from walls to social media, circulating digital posters online. Still, despite its slick and highly visual propaganda enterprise, official 407 Maasri, Off the Wall, pg. 94. 408 Orla Lehane, David Mair, Saffron Lee & Jodie Parker, 2018. “Brides, black widows and baby-makers; or not: an analysis of the portrayal of women in English language jihadi magazine image content,” Critical Studies on Terrorism, OnlineFirst. 409 Rumiyah, Issue 10. 410 Lehane et. al. “Brides, black widows and baby-makers; or not,” pg. 1. 122 Daesh e-posters largely exclude women. The few exceptions depict women as wives as in need of protection from the non-Muslim world. One such image, published by Daesh’s propaganda department between 2014-2015, shows a fully veiled woman watching an armed male fighter walking away. The image reads, “Go and find us that Jannah.” Jannah, meaning paradise, is a very common theme in Daesh propaganda and teachings. For Daesh, Jannah protects Muslims while non-Muslims are punished in the afterlife. The poster implies that through his fighting, or martyrdom, the Daesh fighter will achieve paradise for him and for all Muslims, personified through his wife. Conclusion Across conflicts, rebels invoke gendered depictions of women as innately victimized, peripheral, and non-traditional conflict actors to legitimate their violence against the state. I argue that the continuity of this imagery across contexts shows that rebels believe that this visual securitization benefits their campaign because of nearly universal gender expectations about women, protectionism, periphery to conflict. Employing this imagery can help build ideological legitimacy among civilians by constructing the state as monstrous and by using women’s participation in rebellion to moralize rebel grievances. Consequently, I suggest that women’s participation can shape fundamental organizational strategies concerning rebel-civilian engagement. Thus, women’s militancy may be crucial to rebel options. This and the preceding chapter lay out one aspect of how women’s participation in rebel organizations affects conflict dynamics. Rebels clearly believe that strategic, gendered depictions of women’s militancy offer a pay-off in the form of legitimacy and, consequently, civilian support. In the remainder of this dissertation, I shift to explore how women’s actual participation affects conflict outcomes. Using a quantitative strategy and original data on women’s participation as front-line, auxiliaries, and leaders, I suggest women’s involvement does offer tangible advantages for rebel organizations. While I do not draw a direct or causal line between rebels’ visual securitization of women and the effects of women’s involvement in insurgent groups, I demonstrate two sides of a dynamic, and arguably multi-sided, coin: rebels believe that women’s participation helps increase legitimacy and thus they exploit it to encourage civilian support necessary for conflict successes; and women’s participation in rebel organizations does, in some cases, make success more likely. I show that women’s militancy and broader gender dynamics substantively affect conflict in deeply consequential ways. 123 Figure 4.1: No Title, 1970. PLO, Ismail Shammout (Palestinian), The Palestine Poster Project Archives (PPPA) 124 Figure 4.2: “Providence Did,” 1985. PLO Unified Information Office, Mohammed Hijji (Palestinian), The Palestine Poster Project Archives (PPPA) 125 Figure 4.3: “Weapons of the Revolutionaries and the Masses,” 1977. Fatah and PLO United Information Office, Artist unknown, The Palestine Poster Project Arcvhies (PPPA) 126 Figure 4.4: “In Defence of Peace,” 1986. Fatah, Artist Unknown, The Palestine Poster Project Archives (PPPA) 127 Figure 4.5: “We Are All For The Resistance” 1967. PLO, Ismail Shammout (Egyptian), The Palestine Poster Project Archives (PPPA) 128 Figure 4.6: “The Land,” 1986. PFLP. Marc Rudin (Swiss), The Palestine Poster Project Archives (PPPA) 129 Figure 4.7: “Yom al Mara'a al Alami,” 1980. PFLP, Marc Rudin (Swiss), The Palestine Poster Project Archives (PPPA) 130 Figure 4.8: “Sacred Flame,” 2009. Fatah, Khaled Hourani (Palestinian), The Palestine Poster Project Archives (PPPA) 131 Figure 4.9: “Mother Palestine,” 1983. Fatah and PLO Unified Information Office, Helmi Eltouni (Egyptian), The Palestine Poster Project Archives (PPPA) 132 Figure 4.10: “Dalal Al Mughrabi,” 1978. Fatah, Artist unknown, Palestinian Poster Project Archive (PPPA) 133 Figure 4.11: No Title, 2004, Hamas, Filasteen al-Muslimah 134 Figure 4.12: No Title, 1980s. Internal Islamic Front of Afghanistan 135 Figure 4.13: “The Wounds of my Daughter…,” 1970-1980s. SWAPO, African Activist Archive 136 Figure 4.14: “Join Umkhonto We Sizwe,” 1960s. African National Congress, South African History Archive 137 Figure 4.15: “Commemoration of the Mozambican Revolution,” 1973. FRELIMO, Agostinho Milhafre, International Institute of Social History 138 Figure 4.16: “The blood of martyrs demeans the thrones of tyrants and creates the dawn of the nation,” 1985. Hizbollah, Signs of Conflict 139 Figure 4.17: “MPLA,” undated. People's Movement for the Liberation of Angola, Photographs and Prints Division, Schomburg Center for Research in Black Culture, The New York Public Library. 140 Figure 4.18: “Bangladesh Mothers and Daughters Fighting for Independence,” 1971, Ministry of Information and Broadcasting during the War of Independence, (reprint Liberation War Memorial Museum, n.d.), International Institute of Social History 141 Figure 4.19: “In the first annual commemoration…”, 1986. Lebanese National Resistance Front and Syrian Social Nationalist Party, Signs of Conflict 142 Chapter 5: Assessing Women’s Participation in Rebel Organizations To explore the relationship between women’s participation in rebel organizations and conflict successes, I use new data on female militancy as the key variable of interest in a series of statistical analyses. Researching women’s contributions to political violence is challenging. One primary limitation is that historical accounts of conflict often erase women’s participation. This means that there is a dearth of accessible information about female militancy. My data improves upon existing efforts by expanding the scope of whose contributions are counted to include female auxiliaries, leaders, and founders using nuanced measurements. Existing systematic efforts There have been several recent efforts to systematically assess women’s participation in politically violent groups (Table 5.1). They are very useful but limited. Most of these efforts focus primarily on female combatants. Thomas and Bond offer binary indicators of women’s overall involvement and combat participation in 166 ‘violent political organizations’ (VPOs) operating in Africa between 1950-2011.411 They define VPOs as “any named sub-state organization of two persons or more using violence as a primary means of bargaining over a political good or issue.”412 They identify ‘participation’ as the presence of a single woman. This means that organizations with an occasional female cook or smuggler and those with an established female auxiliary unit are considered the same in the overall participation variable. Similarly, groups with occasional female suicide bombers and those with high ranking female military commanders are both coded positively as having female combatants. Because this coding is binary, it maintains a false equivalency between groups with greatly varied female participation. Other studies take this binary approach as well, likely because of the difficulty in collecting data on women in conflict. Wood and Thomas’ “Women in Armed Rebellion Dataset (WARD)” improves upon these efforts by offering categorical assessments of women’s participation in 211 conflicts active between 1979 and 2009.413 This indicator is constructed on a four- level scale as follows: 0) no evidence of female fighters 1) low (< 5%) 2) moderate (5 - 20%) 3) high (> 20%). This data is very useful for answering questions about female combatancy and draws much needed attention to women’s participation as front-line fighters. Still, this dataset neglects women in 411 Thomas and Bond, “Women’s Participation in Violent Political Organizations.” 412 Ibid, 488. 413 Reed M Wood and Jakana L Thomas, 2017. “Women on the Frontline: Rebel Group Ideology and Women’s Participation in Violent Rebellion,” Journal of Peace Research 54(1): 31–46. 143 non-combat roles. Consequently, WARD cannot speak to the diverse positions women engage during wartime. Henshaw recognizes the shortcomings of this approach and assesses women’s participation in front-line roles, support roles, and leadership roles across 72 insurgencies operating between 1990- 2008.414 However, like Thomas and Bond, she codes these variables as binary indicators. Thus, although her data offers the most substantive assessment of women’s complex roles, the coding masks much of the variation. Dataset N Scope Roles included Coding scheme Thomas and Bond (2015) 166 “Violent political organizations” in Africa, 1950-2011 Any participation, combat Binary Henshaw (2015) 72 Global insurgencies, 1990-2008 Auxiliary, combat, leadership Binary Wood and Thomas (2017) 211 Global insurgencies, 1979-2009 Combat Ordinal, 0-3 RWDP 146 Global insurgencies, 1960-2016 Auxiliary, combat, leadership Ordinal, 0-4 Table 5.1: Recent data collection efforts, women’s rebel participation Empirically, these datasets are also inappropriate for this project because of the source data they use to identify insurgencies. Henshaw and Wood and Thomas use the UCDP Non-State Actors dataset. Conflict is considered ‘active’ in these data during the first year in which there are at least 25 battle-related deaths, which means that actors do not appear in the data until this threshold has been met. These data consequently exclude rebellions that may generate civilian followings, carry out attacks, kill, and control territory, but fail to meet the deaths requirement. According to my data, insurgents in most cases do not carry out attacks that meet this threshold until a few years into operations. Subsequently, these datasets miss rebels’ emerging years. Excluding these years or entire insurgencies propagates the myths that lethality is the best measure of insurgent ‘success’ and that we need only study ‘successful’ organizations to understand rebellion. As I argue below, this measure misunderstands the effect of processes like gender on conflict and therefore these datasets are unsuitable for this project. Thomas and Bond also use the UCDP data to draw their sample but supplement it with data from the National Consortium for the Study of Terrorism and Responses to 414 Alexis Leanna Henshaw, 2016. “Where Women Rebel,” International Feminist Journal of Politics 18(1): 39–60. 144 Terrorism (START)’s Global Terrorism Database (GTD) and “academic and journalistic accounts of violent conflict.”415 Their data is not randomly sampled and focuses on a specific region. It is therefore not suitable for this project. Further, the binary focus on ‘war’ and ‘peace’ in existing quantitative datasets is theoretically flawed. Using a battle-deaths threshold of inclusion in a ‘war’ operationalizes intensity as the key measure of conflict. Under these circumstances, war “is conceived as an extraordinary interruption into the ordinary processes of society, economy, culture, and politics.”416 This is a strange conceptualization of conflict, particularly for small, internal wars fought between rebels and, often, state militaries scrambling to recruit new, non-traditional fighters into standing armies. As Barkawi notes, militarism, war, and the potential for war shape modern social life even when there is no active fighting. War is a pervasive presence, and the “long reach of the shadows of war” color social constituencies and re-structure social relationships in preparation for, perpetration of, and aftermath of violence.417 Conflicts and conflict-years excluded by the battle-deaths threshold can have profound effects on state and society interactions and greatly alter the local environment. For example, Boko Haram does not appear in the UCDP dataset until 2009, eight years after Mohammed Yusuf began recruiting and training supporters in a religious complex and school. Yusuf garnered local admiration and territorial control by railing against government corruption and linking local problems to Western interference in religion and governance. By 2009, when Boko Haram exacted violent revenge for Yusuf’s arrest and execution, the leader had already shaped the cultural landscape in Borno and built a strong Islamist sentiment on which Boko Haram members easily capitalized. These eight years are critical to understanding Boko Haram’s evolution into a viciously violent organization, but are erased from most quantitative assessments of the group’s behavior. In another example, Chile’s FPMR is excluded entirely from most datasets because they killed few people during their 15-year battle against the state. Still, the FPMR carried out over 100 attacks including high level assassinations, bombings, and kidnappings. Their activities inspired other leftist guerilla movements in Chile and helped shape the political landscape during the Pinochet regime. They integrated women into their front-lines. Battle-deaths thresholds erase them entirely. War’s widespread reach shapes gender relations. Feminist IR scholars write extensively about militarization’s effects on gender in pre-conflict, conflict, and post-conflict arenas. They call attention 415 Thomas and Bond, “Women’s Participation in Violent Political Organizations,” pg. 492. 416 Tarak Barkawi, 2016. “Decolonizing War,” European Journal of International Security 1(3): pg. 202. 417 Ibid, pg. 201. 145 to cultural shifts in masculinity and the militarization of men’s lives, demonstrate how women’s participation can alter social boundaries and integrate divisions between public and private spheres, and illustrate how reintegrating non-traditional actors like women and children after conflict can shape future insecurities or potential for violence.418 Enloe, for example, argues that gender is a driving force in the nationalist divides that often spur violence.419 Belligerents may invoke women to justify conflicts, either by stressing the need to protect them (e.g. the LTTE’s fight against the Sri Lankan government) or by comparing ‘their’ women’s social position as superior to the women represented by their opposition (e.g. US interventions in the Middle East). These wars often necessitate a change in women’s social roles even before they begin. They often end with increases in gender-based violence in militarized ‘post-conflict’ environments. In Boko Haram territories, women and girls’ public lives may be drastically restricted years before they are kidnapped, indoctrinated, and sent out for suicide attacks.420 In Colombia, Nicaragua, and other sites of leftist insurrection, women’s increasing public opportunities and men’s responses shape community interactions and the goals of leftist idealism far before war is waged. It is critical that we accept war for what it really is- not an extraordinary interruption of peacetime but an on-going process and possibility that structures society at all levels and at all times. Datasets measuring conflict by level of intensity are insufficient tools for such a project. Original dataset My primary challenge is the battle-deaths thresholds in most datasets. To the combat this limitation, I collect a new dataset of global insurgencies operating between 1960-2016. I identify these organizations by drawing together several data sources, including the UCDP Non-State Actor dataset, the Correlates of War Project, START’s Big, Allied, and Dangerous (BAAD) dataset and Global Terrorism Database (GTD). I utilize existing literature, US Department of State Human Rights Country reports, Amnesty International Annual Reports, truth and reconciliation commission reports, and other primary sources. My dataset includes over 400 insurgencies active between 1960-2016 and considers the year of formation as the group’s first year. Formation year is not a perfect solution, but it provides a more robust assessment of militant evolution than existing data. 418 Enloe, Bananas, Beaches and Bases; Herrera and Porch, “‘Like Going to A Fiesta;’” Stewart and Adelstein, “Civil War, Women, and Social Development.” 419 Enloe, Bananas, Beaches and Bases. 420 Many women voluntarily carry out suicide missions for Islamist groups (Matfess, Women and the War on Boko Haram). 146 I define insurgencies as non-state units; with a name for their group; with a leader or leadership structure separating leadership from cadre; who have executed or planned at least one credible act of violence; who desire to control territory; and who intend to change or alter the state’s authority. This definition excludes partially or autonomous states, including ‘frozen conflict zones;’ cases where militants fight only foreign governments within their country; coups and attempted coups that are not part of a larger armed rebellion; terrorists who operate hit and run or sporadic campaigns and do not engage with the civilian population; civilian defense forces; and state paramilitary and other military offshoot groups. Importantly, I include insurgencies who eventually use or plan violence even if they do not in their initial years. I include groups generally labeled as ‘terrorist organizations,’ like Chechen separatists, Daesh, or Boko Haram in this definition if they meet the above criteria. Most insurgencies use terrorist tactics intended to foster insecurity among civilians. I include insurgencies that became part of fronts or umbrella groups under one of two conditions. First, I include groups that operated independently and then merged to form a new organization. In these cases, I include the group only during the years it operated independently. Second, I include groups that operated independently and then joined a larger organization but kept their name and carried out some independent operations. In these cases, both the sub-group and its umbrella are included in the dataset. This can introduce bias because not all groups in the dataset are independent from one another. Given the overlapping, dynamic, and co-operational nature of insurgency, it is possible to reduce but not eliminate this bias entirely. As I discuss further below, I try to mitigate this dependence bias by clustering standard errors on the individual actor in my analyses. There is important variation between the sub-groups and umbrellas. For example, I include both the Guatemalan National Revolutionary Unity (URNG) and the Revolutionary Organization of Armed People (ORPA). ORPA joined the URNG three years into its lifespan but continued to operate as a unit. The URNG has a higher number of female combatants than the ORPA because other groups in the front contributed more female members. Similarly, while there is evidence of female leadership in the URNG, there is no data to suggest the same in ORPA. I collected data on years of operation, political ideology, and religiosity for all cases in this dataset. I then took a random sample of 50 Islamist groups, 50 leftist groups, and 50 groups with neither ideology for a full sample of 150 cases. I chose this stratified sampling method, instead of a truly random sampling method, because randomly pulling from the full universe of cases could easily produce a sample that under-represents or otherwise misrepresents women’s participation in insurgencies. This is a problem because my variable of interest is women’s participation. Extant 147 literature suggests that leftist groups are most likely to include female members, while Islamist groups should be least likely to have female recruits.421 By drawing equally and explicitly from these proxies and a third random group, I gathered a semi-random but theoretically appropriate sample for this project. The result is a geographically diverse sample with variation across time, longevity, and ideology. Of these 150 cases, I excluded four from the final sample. I removed two of these, the Shan State Army in Myanmar and the Maoist Community Party of Turkey, because there is so little reliable information about the groups that coding them along basic indicators, let along for female participation, was not possible. I excluded the third group, Revolutionary People’s Front (RPF) in India, because I discovered after drawing the sample that it is a wing of the People’s Liberation Army of Manipur (PLA). The PLA is also included in the sample, so I removed the RPF to prevent introducing bias by counting the same group twice under different names. Finally, I removed the Fatherland and Liberty Nationalist Front of Chile because of conflicting representations of the group as an insurgency. While I originally included and fully coded this group, scholars disagree on the extent to which the group is an insurgency or a paramilitary. After careful study, I am no longer convinced that I can responsibly qualify the group as an insurgency. I exclude paramilitaries from my dataset, so I err on the side of caution and remove the group from my dataset. My final sample for this project includes 146 groups from 55 countries (listed in the Appendix).422 Coding women’s participation As discussed in Chapter 1, I adapt Henshaw’s typology of women’s conflict participation: front-line, auxiliary, leadership, and founders. Table 5.2 summarizes these roles. This is not a perfect coding system. I plan to improve it over time with better data and more nuanced measures. For example, women’s leadership participation can vary widely. Consider leftist insurgencies in Colombia. Women co-founded and led the M-19 guerrilla movement, but women leaders in the National Liberation Army (ELN) usually did not rise above regional commands.423 In the Quinten Lame Armed Movement (MAQL), only one woman became a commander but many women were lieutenants and 421 Thomas and Bond, “Women’s Participation in Violent Political Organizations;” Wood and Thomas, “Women on the Frontline.” 422After fixing two ideological miscodes and dropping these four cases, the final sample includes 52 Islamist insurgencies, 48 leftist insurgencies, and 47 groups with neither ideology. The initial coding errors should not bias my results because these solutions increased the number of Islamist insurgencies in the sample. These insurgencies are least likely to have female members. 423 Luz María Londoño and Joana Fernanda Nieto V., 2006. Mujeres no contadas: Procesos de desmovilización y retorno a la vida civil de mujeres excombatientes en Colombia 1990-2003, La Carreta Editores. 148 squad leaders.424 My data accounts for variation in the level of women’s participation in leadership roles, but does not distinguish between these types of leadership. This is a function of necessity: collecting data on the level of women’s participation is difficult, particularly for auxiliary and leadership roles where previous data collection efforts are scarce or non-existent. Thus, while theoretically crude, this coding procedure is the most sophisticated, best-faith effort currently possible. Role Definition Front-line Combatants regularly engage in the front-line environment in support of the group; includes the use of or close proximity to violence Auxiliary Auxiliaries identify with the goals, ideology, or effort of the group and routinely offer general support contributing labor (nurses, cooks, spies, scouts, sex workers etc.), supplies, planning, logistics, or recruitment assistance; rarely use or are in close proximity to violence. Leadership Leaders exercise direct control over and provide oversight of other participants and/or exercise direct control over the strategy, policies, and/or ideology of the group. Founders Founders participate as a core member of the group’s inception Table 5.2: Role identification criteria I code these data from several sources including reports from international organizations and non-profits including the United Nations and Amnesty International; reports from terrorism monitors including the South Asia Terrorism Portal (SATP) and the Global Terrorism Database; secondary accounts in peer-reviewed journals, manuscripts, and books on insurgencies across disciplines including political science, international security, women and gender studies, feminist studies, sociology, and anthropology; government records; ceasefire agreements and demobilization records; insurgent manifestos; extant datasets of women’s participation; declassified Central Intelligence Agency assessments and US Department of State cables; and news reports available in English and Spanish through LexisNexis, the BBC, IRIN, and the New York Times. I code women’s participation on a five-level scale: none, occasional, low, moderate, and high (Table 5.3). These data include measures of female militancy in front-line roles, in auxiliary positions, and as leaders. It also includes a binary measure of whether women helped found an insurgency, for descriptive analysis, as well as binary measures of the categorical variables (Table 5.4). 424 Ibid. 149 Level of participation Coding rules Examples 0 None No evidence of women’s participation Armed Islamic Salvation Front; United Somali Congress/SNA front-line roles 1 Occasional Women participated occasionally - Composed less than 5 percent of an insurgent group and/or role within an insurgent group and/or - Qualitatively identified as, but not limited to, “occasional,” “few”, “infrequent,” and “rare” Ethiopian People's Revolutionary Party leadership roles; East Turkestan Islamic Movement front-line roles 2 Low Women participated at low levels - Composed between 5-9 percent of an insurgent group and/or role within an insurgent group and/or - Qualitatively identified as, but not limited to, “in low numbers,” “small group of women,” and “minor” People's Revolutionary Army (Argentina) front-line roles; Jabhat Fateh al-Sham support roles 3 Moderate Women participated at moderate levels - Composed between 10-19 percent of an insurgent group and/or role within an insurgent group and/or - Qualitatively identified as, but not limited to, “moderately,” “several”, “large minority,” and “often” Palipehutu - Forces for National Liberation support roles; Liberians United for Reconciliation and Democracy front-line roles 4 High Women participated at high levels - Composed at least 20 percent of an insurgent group and/or role within an insurgent group and/or - Qualitatively identified as, but not limited to, “at high levels,” “widespread,” “integral,” “substantial,” and “significant” Shining Path leadership roles, People’s Liberation Army (Nepal) front-line roles Table 5.3: Rebel Women Data Project coding rules In a handful of cases, I identify the existence of female members in certain roles but cannot assess the scope of their participation. For example, Brown recalls an event where three women contributed to Peru’s Revolutionary Left Movement (MIR) as couriers.425 The extent to which these women are representative of other auxiliaries, the regularity of this participation is unclear, and other sources are equally vague. Consequently, I conclude that the MIR benefited from female auxiliary 425 Jonathan C. Brown, 2017. Cuba’s Revolutionary World, Harvard University Press. 150 support but do not estimate women’s proportional participation. In another example, I identify 71 groups with female front-line fighters, and I estimate the proportion of women in 97 percent of them. Variable Variable type Women’s overall participation in an insurgent group (highest level recorded in any role) Ordinal, 0-4 Binary, 0-1 Women’s participation in front-line roles Ordinal, 0-4 Binary, 0-1 Women’s participation in auxiliary roles Ordinal, 0-4 Binary, 0-1 Women’s participation as leaders or commanders Ordinal, 0-4 Binary, 0-1 Women’s participation as founding members of an insurgency Binary, 0-1 Table 5.4: Rebel Women Data Project data I work to combat four primary limitations. First, the concept of ‘participating in an insurgency’ may not be equally transportable across contexts. For example, Willard suggests that even the idea of ‘joining an organization’ may be an unfamiliar, exported concept. She recalls interviewing Mayan women in Guatemala who participated in the civil war. When she asked them why they joined the insurgency, they told her that they did not join any group, they just did what other people did to support their community.426 Writing on the MILF, one scholar contends that while men engaging in auxiliary positions such as communications officers or medics may be seen as soldiers, women in these roles are often viewed as civilians supporting the insurgent effort.427 I mitigate potential discrepancies of this kind by coding women’s participation in a rebel organization by their activities and the case- specific context. Second, many women are forcibly recruited into conflict. Forced recruitment mechanisms are not simple. Groups like Boko Haram and the Daesh forcibly enslave women for sex and menial labor. At the same time, Daesh actively (and successfully) recruits many female volunteers to serve in supportive roles. Boko Haram rarely does. Some women, like many in the Lord’s Resistance Army (LRA) in Uganda, recruited for sex are also armed fighters. In some groups, like those that operated in Liberia, women are both forcibly and voluntarily recruited as combatants. In the FARC-EP, where 426 Emily Willard, personal communication, February 20, 2018. 427 Rosalie Arcala Hall and Julian Smith, 2016. “Women in Combat: Both Spoilers and Enablers of Peace,” The Diplomat, February 15. 151 women were both voluntarily and forcibly recruited, some coerced members eventually came to publicly support the organization and recruit others. A key part of my theory is visibility. Women’s visible support for an insurgency is essential to civilian perception of insurgencies as female-inclusive, legitimate, as well integrated into their communities, and as committed to a common struggle. Forcibly recruited women likely do not have this same effect, because “forced combatants… are almost randomly pulled out of their communities to join fighting forces.”428 In other areas, like tactical and strategic effectiveness, coercive recruitment status would have less of an effect on my outcomes of interests. Still, to avoid bias and the theoretical complexities of what civilians might know about female recruitment status, I evaluate women’s inclusion based on voluntary participation. By voluntary, I mean that there is no evidence of coerced recruitment. For example, many female Boko Haram suicide bombers are forcibly recruited, but Matfess suggests that some women do join the group voluntarily in both front-line and supportive positions.429 Therefore, in my data I consider Boko Haram to have occasional female front-line participation and low supportive female participation. Third, it is challenging to gather data on role-level participation in insurgencies. I extensively examined rebel histories and related documents to assess the position of female militants. In some cases, data is missing for one or more indicators. In a few cases, I code variables assessing women’s overall participation but am unable to determine the positions in which women participated. This is particularly true for assessing women’s involvement in auxiliary roles because this labor is often invisible. Therefore, I suspect that my assessments of women’s supportive participation are under- estimated cross-nationally. I do expect some biased undercounting by region in this area. Specifically, my research suggests that scholars have paid the most substantive attention to women’s participation across roles in Latin American insurgent organizations. This suggests that women’s auxiliary participation may be more difficult to assess in other regions than Latin America. While I did find this to be the case in my coding, I found that it was not an issue of finding evidence of participation but instead, I found fewer additional sources to cross-verify women’s auxiliary participation in other regions. Fourth, often women and girl actors are discussed indistinguishably in the literature on women in conflict. Theoretically, the distinction between voluntary women and girl recruits is not germane to this project, particularly because many women are recruited during their teenage years and rebel groups 428 Dara Kay Cohen, 2010. Explaining Sexual Violence in Civil War, Ph.D. diss., Stanford University, pg. 30. 429 Matfess, Women and the War on Boko Haram. 152 are often staffed by young fighters of all genders. Still, I follow the literature in establishing 15 as the age of adulthood during conflict.430 In this dataset, these are no cases where female recruits are primarily or exclusively girls. I discuss the descriptive findings in my quantitative data in Chapter 6. In brief, these data show that women actively participate in political violence. The RWDP includes data not used in my dissertation. I collect data on temporal variation of female involvement within insurgencies, coding the abovementioned variables again for women’s participation in a rebel group’s first five years. My future projects will assess how this variation affects support for insurgency and rebel successes. Ethical challenges Research on women and political violence presents a host of ethical challenges. These primarily arise from the tensions between conducting feminist security studies research and the necessity of quantification and generalization in this project. Westmarland highlights a number of feminist challenges raised in the pursuit of quantitative research, including ‘adding’ women to processes where we simply extend findings from research about men to women; prioritizing ‘knowledge’ produced through positivist, quantitative statistics over women’s individual voices, stories, and histories; assuming that quantitative data is ‘objective’ and therefore not prone to the ‘biases’ produced by qualitative work; and assuming that as researchers, we can rise above emotional investment to provide purely objective, scientific results.431 Scholars of women in war caution against one-dimensional representations of women’s wartime experiences as active participants or victims of violence.432 Women’s political participation is not homogenous: their statuses and activities are shaped by history, geography, their organizations, access, and cultural, religious, and ethnic differences. This is a particularly visceral contention for feminist quantitative researchers. How do we conduct ethical quantitative research that, by necessity, categorizes and generalizes the complex lived experiences of a historically misrepresented population? I share these concerns about quantitative research. I worry that quantification threatens feminist knowledge building by suggesting that generalizable understandings of women in war are useful but individualized, contextual, and unique studies are not. I fear that quantification conflates real women with varied experiences for the sake of generalizability. This can be dangerous because it 430 Protocols 1 and 2 to the Geneva Conventions establish fifteen as the age of consent for military service, and the UN- established Special Court for Sierra Leone has jurisdiction for those over the age of fifteen for war crimes. 431 Nicole Westmarland, 2001. “The Quantitative/Qualitative Debate and Feminist Research: A Subjective View of Objectivity,” Forum: Qualitative Social Research 2(1): 1-12. 432 See: Annick T. R. Wibben, e.d., 2016. Researching War: Feminist Methods, Ethics, and Politics, Routledge. 153 can offer inadequate or inaccurate assessments of women in politics, because it can re-articulate women as a monolithic “Other,” and because it may further entrench women’s experiences in historical darkness. Feminist quantitative research is challenging, because, as Harding notes, “Feminist analytical categories should be unstable – consistent and coherent theories in an unstable and incoherent world are obstacles to both our understandings and our social practices.”433 Still, this work is possible. Sjoberg writes, Research in Feminist Security Studies reformulates mainstream approaches to traditional security issues, foregrounds the roles of women and gender in conflict and conflict resolution, and reveals the blindness of security studies to issues that taking gender seriously shows as relevant to thinking about security. Together, these… show that gender analysis is necessary, conceptually, for understanding international security, important for analyzing causes and predicting outcomes, and essential to thinking about solutions and promoting positive change in the security realm.434 I strive to make ethical, feminist interventions in security studies that demonstrate the necessity of gender analysis for understanding conflict. This research is “uncomfortably lodged at the intersection of multiple fields of scholarship.”435 Through careful, nuanced coding I attempt to break down monolithic representation of women’s militancy. I supplement this data with qualitative assessments of visual propaganda and political messaging to further emphasize this diversity. I highlight women’s unique contributions in individual contexts, and I focus particular attention onto women’s supportive labor. In this sense, this research moves away from the over-focus on elites – highly intelligible female fighters – in similar studies to make visible the myriad of un- and under-appreciated ways that women sustain violent campaigns. There is much we can learn from generalizable measurements. Social scientists across disciplines have produced an exceptional body of work on women’s wartime participation in individual and comparative contexts. However, we simply know very little about the scope of women’s militant participation, historically and otherwise. We need to know how, when, and where women participate cross-nationally to identify trends or unique moments within individual cases. To do deeply contextual work with meaningful impact, we need to understand how that work is situated in the broader phenomenon of women and political violence. International organizations and state governments 433 Sandra Harding, 1986. “The Instability of the Analytical Categories of Feminist Theory,” 11(4): pg. 649. 434 Laura Sjoberg, 2011. “Looking Forward, Conceptualizing Feminist Security Studies,” Politics & Gender 7(4):. pg. 600. 435 Annick T. R. Wibben, 2016. “Introduction, in Wibben, ed. Researching War: Feminist Methods, Ethics, and Politics, pg. 2. 154 increasingly afford attention to the role of gender in counter-extremism programming, post-conflict peace negotiations, and wartime adjudication. Contextual application of these programs is paramount, but policymakers often develop them using a cross-national lens. Sjoberg concludes that feminist research should be a site of contestability whose intent “is to raise problems, not to solve them; to draw attention to a field of inquiry, rather than survey it fully; to provoke discussion, rather than serve as a systematic treatise.”436 I hope to offer a starting place for new and better research that provokes discussion, raises questions, and pushes forward a burgeoning field. Descriptive findings: where and how women rebel These data suggest that women’s insurgent participation is very common. I identify women militants in either combat, auxiliary, or leadership roles in at least 53 percent of cross-national rebel groups in this sample.437 An estimated 49 percent of organizations include women in combat positions while 46 percent incorporate women as auxiliaries. I further find that women participate as political leaders or commanders in 28 percent of groups. Women founded or co-founded 3 percent of insurgencies. These estimates should be considered the floor and not the ceiling of women’s militant involvement. This is particularly true for evaluations of women’s auxiliary participation because women’s supportive labor is so often invisible. For example, women participated widely in the Guevarista Revolutionary Army (ERG), a leftist group with a gender-equity platform operating in Colombia. Their ideology and similarity to other leftist organizations in the region suggest that women should participate that all levels. Still, from extant literature and news reports I only find evidence of female combatants and commanders. This does not mean that there are no female ERG auxiliaries. Instead, it is likely that there are many female ERG auxiliaries, but their participation is not recognized by scholars or journalists. Not only do women militants participate in all roles, but they are involved at all levels. Women constitute a widespread proportion of leadership in 8 percent of organizations in this sample. Rebels also incorporate women as occasional leaders in 10 percent of groups. Rebels are least likely to include women auxiliaries occasionally – in fact, I only identify one case of occasional female auxiliaries in this sample. Rebels are most likely to position women at high auxiliary levels (18 percent). When insurgents employ female combatants, they are by far most likely to do so at low, but still sustained, levels (21 436 Laura Sjoberg, 2011. “Looking Forward, Conceptualizing Feminist Security Studies,” Politics & Gender 7(4): pg. 602. 437 All percentages in this dissertation are rounded. 155 percent). Among the most interesting findings in these data is that women’s participation in rebel organizations does not always follow the patterns we would expect. For example, women are often shut out of prestigious positions in leftist groups that adopt emancipatory frameworks. Few women reached leadership positions in Argentina’s People's Revolutionary Army (ERP) even though women composed about 30 percent of the organization.438 In some cases, women are widely found in some groups but seemingly absent from similar organizations within the same struggle. This is the case in the Philippines: while women MILF militants participated at low combat levels and operated in designated auxiliary units, there is no available evidence of women’s involvement in the group’s predecessor, the Muslim Independence Movement (MIM).439 Figure 5.1 details the distribution of women’s participation by role in these data. The total percentages are slightly different than those presented above, because they take into consideration the relative proportion of women’s participation (ordinal measure) and not their mere presence (binary measure). Rebels are least likely to include women political leaders or commanders. The highest percentage of female-inclusive groups with female auxiliaries include them at high levels. This contrasts with the trends in female front-line fighters: groups are most likely to include women fighters at low levels. Figure 5.1: Women’s participation in rebel groups by level, 1960-2016 438 Paul H. Lewis, 2002. Guerrillas and Generals: The "Dirty War" in Argentina, Praeger. 439 Rosalie Arcala Hall and Joanna Pares Hoare, 2015. “Philippines,” in Women in Conflict and Peace, International IDEA. 156 Variation by group type and ideology There is significant variation in the distribution of women’s participation by group type. This divisions are largely ideological and identity-based. The likelihood of female involvement in ethnic groups is fairly even, with women taking part in 45 percent of ethnic organizations. Islamist groups are least likely to include women: 67 percent of Islamist organizations in this sample include no female members. In comparison, only 16 percent of leftist groups exclude female participants. These differences are stark when we break them down by role. Approximately 29 percent of Islamist groups position women in combat, compared with 81 percent of leftist groups. Still, what is interesting about this distribution is the differences in women’s participation across group-types by level. Table 5.6 details the variation in Islamist and leftist insurgencies. Islamist groups Front-line Support Leaders None 71% 75% 96% Occasional 15% 2% 4% Low 13% 13% 0% Moderate 0% 10% 0% High 0% 0% 0% Leftist groups Front-line Support Leaders None 20% 23% 40% Occasional 2% 0% 23% Low 28% 16% 9% Moderate 15% 16% 7% High 35% 44% 21% Table 5.6: Distribution of women’s participation by group-type Women take on very different positions, with varying levels of intensity, between group types. For example, Islamist groups are much more likely than leftists to occasionally employ women on the front-lines. This is likely because Islamist groups, like the ETIM and the Taleban employ women for occasional suicide attacks more frequently than other organizations. Using female combatants in these organizations sometimes symbolizes a shift in the groups’ strength or strategy. Daesh historically relegates women to supportive positions as wives, cooks, and recruiters, but as they increasingly lose territory in Iraq and Syria women are occasionally being sent on front-line missions. Pakistani authorities recently intercepted a woman who confessed she intended to carry out a Daesh suicide attack on Easter 2017. In July that year a woman disguised as a civilian fleeing Mosul with her child detonated a suicide bomb. Even Al Qaeda, which famously has very few female members, began 157 occasionally using female attackers in 2005.440 Still, female front-line involvement violates the religious and socially conservative tenets of most Islamist organizations. Consequently, few groups offer sustained roles for women as combatants. Conversely, only two leftist groups employ women as occasional fighters. Leftist groups are most likely to have female fighters in widespread roles - accounting for more than 20 percent of their fighting forces - than they are to have female combatants at any other level. This suggests that armed female participation is very integrated into most leftist organizations, with women making up core cadre constituencies. The distribution of roles for women across leftist organizations is also much smoother than in Islamist groups: women are likely to participate in leftist organizations as combatants, supporters, or in leadership positions. In Islamist groups, women are most commonly found at moderate levels in support roles. Boko Haram, Hizbul-Mujahideen, and other organizations use women as smugglers to carry contraband through checkpoints or on public transit.441 Women often provide intelligence or operate in domestic roles like cooking and cleaning for fighters. Variation over time The data also suggest that women’s substantive participation in rebel organizations may be declining over time. To be clear, the number of groups that include women are not waning, but the proportion of female militants within them may be. Groups with no women or occasional female participants have proliferated since the 1980s, while groups with organizations with low, moderate, and high levels of female involvement have declined. One possible explanation is that new insurgencies are increasingly Islamist, while leftist rebellions form less often. Few Islamist groups operated before 1980 and very few leftist organizations forming after 1985. Other scholars identify as similar trend: Hewitt and colleagues find that while most terrorist attacks between 1970-2001 were carried out in three Latin American countries, between 2001-2008 most attacks occurred during Islamist conflicts in the Middle East and South Asia.442 This discrepancy is particularly apparent when it comes to women’s political leadership and command in rebel organizations (Figure 5.2). Female leadership at all levels has dropped significantly over time and, while women did participate at high levels in the 1990s, women leaders were largely found in mature groups that formed in the 1960s and 1970s. 440 Mia Bloom, 2014. “Female suicide bombers are not a new phenomenon,” Washington Post, August 6. 441 Seran de Leede, 2014. Afghan Women and the Taliban: An Exploratory Assessment. ICCT Policy Brief. 442 J. Joseph Hewitt, Jonathan Wilkenfeld, and Ted Robert Gurr, 2012. Peace and Conflict 2012. Paradigm Publishers. 158 Figure 5.2: Group ideology and women’s leadership participation over time Islamism: 0= no, 1 = yes While certainly not a causal link, this correlation across all types of female participation suggests that as leftist insurgencies die out and Islamist insurgencies form more often, women will participate less in sustained militancy but may increasingly be occasional front-line attackers and lower-level auxiliaries. Regional variation These data include rebel groups from five regions: Western states (Europe or North America), Middle East, Asia, Africa, and Latin America. Women’s participation by role and level varies by country (Figure 5.3) and by region (Figure 5.4). States like India, Syria, and Ethiopia that have many operating rebel groups have high levels of female militancy in several of them. In other cases, like Brazil and Chile, women are found widely in the few active insurgencies operating between 1960-2016. In other countries, like Mali, women’s participation appears less frequent. 159 Figure 5.3: Highest recorded level of women’ combat participation by country Asia leads with the highest number of female-inclusive organizations, many of which operated in India and the Philippines. Europe hosts the smallest number – one - of groups with female members. This sample captures two European insurgencies during these years, and consequently women are found in 50 percent of these organizations. However, when looking at female-inclusive groups as a proportion of all groups, female militants are most likely to be found in Latin America. Nearly 87 percent of Latin American insurgencies include women in some official capacity. In contrast, an estimated 59 percent of Middle Eastern rebellion include women. These includes some Islamist organizations, leftist groups in 1980s Afghanistan, and several Syrian resistance groups that have formed since the outbreak of the 2011 civil war. Surprisingly, only 37 percent of African rebel insurgencies in these data include female members. Extant research on women in African rebel organizations suggest that this estimate is low and is likely the result of the groups sampled in the RWDP relative to the universe of possible cases.443 There is significant regional variation in the roles women play within insurgencies. Women are most likely to serve in front-line roles, support roles, and as leaders in high numbers in Latin American insurgencies. A useful comparison is the Middle East or Asia, where women rebel in high levels but occupy very few prestigious positions. Similarly, there are comparatively few occasional female 443 Thomas and Bond, “Women in Violent Political Organizations” identify women participants in 45 percent of African insurgencies and women combatants in 29 percent of African insurgencies. 160 participants in Latin American rebellions, while Middle Eastern, African, and Asian groups are much more likely to integrate women sporadically. Still, there are regional similarities. These findings suggest that cross-regionally, women auxiliaries are scaffolding rebellions from the rearguard. They also indicate that in all areas, rebels are hesitant to incorporate women leaders at sustained levels. Female political leadership and command occurs most often at occasional levels in all regions. Figure 5.4: Women’s participation by region, role, and level (y=number of groups). 161 Variation within conflicts: evidence from Syria Women’s participation often varies within conflicts, even among similar group types. The on-going Syrian conflict exemplifies this variation. These data include six insurgencies primarily targeting the Syrian government: the Free Syrian Army (FSA), the Islamic Front (IF), the People’s Protection Units (YPG/J), Jabhat al-Nusra, the Al-Tawhid Brigade, and the Nour al-Din al-Zenki Movement.444 Except for the leftist (and heavily female), Kurdish YPG/J, these groups are largely Islamist and have similar conflict aims.445 Some, like Jabhat al-Nusra and the Nour al-Din al-Zenki Movement, are more extreme than their counterparts I identify female participants in 50 percent of these groups, and in 40 percent of the Islamist organizations. Syrian women play a significant role in Free Syrian Army activities. They staff field hospitals, work as intelligence officers, smuggle weapons and money, run the group’s coordination offices, and take up arms on the front-lines.446 One group of women run a medical aid and food service out of a storage unit. At least two women formed their own women’s combat brigades, taking on leadership positions. One female FSA member founded an auxiliary unit – the Khawla Bint Al Azwar brigade - to give first aid. Shortly after, she began weapons training for unit members. In 2013, her unit had 30 members and 80 waiting to join.447 Another militant, Bahar, joined the FSA in 2013 as a front-line nurse until she armed herself and joined an existing battalion. Later that year she established her own brigade in East Ghouta, the Al Ghouta Free Women Battalion. In the group’s creation announcement video, Bahar and other veiled female militants wield weapons and disparage al-Assad.448 A third FSA women’s brigade, Al Mouminin Aisha (named after the Prophet Mohammed’s wife) reportedly fights in Aleppo. In contrast, Jabhat al-Nusra’s official position does not let women fight on the front-lines and largely restricts them to auxiliary activities that can be accomplished from inside the home. The group espouses a hardline, fundamentalist agenda and subjects women to exceedingly harsh restrictions. Still, few women work in intelligence gathering and evidence suggests that women do occasionally try to fight for al-Nusra. Hala, a self-identified female al-Nusra combatant, suggested that men in the group 444 The Al-Tawid Brigade was an independent organization, part of the Free Syrian Army, and part of the Islamic Front at varying times 445 The FSA was not an Islamist group in its inception, but over time Islamist factions have dominated the organization 446 Eva Sohlman, 2013. “Sisters in Arms Join the Fighting in Syria,” The New York Times, 10 April; Rania Abouzeid, 2013. “Portrait of a Lady: A Female Syrian Rebel Speaks to TIME,” Time Magazine, 27 February. 447 Sohlman, “Sisters in Arms Join the Fighting in Syria.” 448 Omar Al Muqdad, 2016. “Meet the woman who Founded the First All-Female Unit of the Free Syrian Army,” The National, 16 August. 162 regularly try to keep her away from the front-line.449 Commanders do not let her stay in al-Nusra bases with men, and denied her request to join her husband on a mission in Aleppo because she would have had to stay in the base room with 40 male fighters. Still, she says she is proud of being al-Nusra member and the group’s affiliation with al-Qaeda. She reports, “It bothers the West, which pleases me. Let them say Qaeda is terrorist. It honors me to be called a terrorist.”450 While women contribute at varying levels to the FSA and Jabhat al-Nusra, I do not identify female participants in IF or in Nour al-Din al-Zenki Movement. Though the case is not included in my RWDP, the variation across Islamist groups in Syria is even more apparent when we consider women’s participation in Daesh. The group has voraciously recruited foreign female members. In December 2015, the Soufan Group reported that up to 31,000 people from over 86 countries had traveled to join Daesh.451 Over 5,000 of these recruits are from Western states, and more than 600 of them are women.452 Female recruits identify ideological salience, the mandate of hijra, and a sense of community as their primary motivators. Most Daesh women take on traditional, auxiliary roles: the group actively recruited women on social media by stressing the need for women to carry out their own jihad of marrying and caring for soon-to-be martyrs, raising the next generation of the caliphate, recruiting, and educating others about Islam.453 The group strictly restricted women from combat. However, Daesh may be shifting its prohibition of women on the front-line. Winter and Margolin suggest that two Daesh announcements in their official publications evidence a shift in the group’s policy on women combatants, and they cite video evidence of female suicide bombers and armed fighters in 2017.454 Conversely, Cottee and Bloom caution against reading into this apparent change. They emphasize Daesh’s deeply engrained opposition to female combatants and imply that the Iraqi Army may have fabricated stories of female fighters to further demonize the terrorist group.455 Variation within conflicts: evidence from South Yemen In 1964, the Front for the Liberation of Occupied South Yemen (FLOSY) organized to contest British occupation and to attempt a government takeover. That year, the National Liberation Front (NLF) sprung up to support FLOSY’s effort. Things quickly soured, and the groups fought each other and 449 Rania Abouzeid, 2014. “Syrian Al-Qaeda women: Searching for Combat, Martyrdom on the front-lines,” Al Jazeera America, July 20. 450 Ibid. 451 Soufan Group, 2015. Foreign Fighters: An Updated Assessment of the Flow of Foreign Fighters into Syria and Iraq. 452 Ibid. 453 Loken and Zelenz, “Explaining Extremism.” 454 Margolin and Winter, “The Mujahidat Dilemma.” 455 Simon Cottee and Mia Bloom, 2017. “The Myth of the ISIS Female Suicide Bomber,” The Atlantic, 8 September. 163 British forces simultaneously as they vied for control. The UCDP Non-State Armed Actor dataset considers FLOSY and the NLF to be the same organization, but they are very different groups who incorporated women in extremely different ways. Few women participated in FLOSY, and those who did generally followed male relatives into the organization. Available evidence suggests that FLOSY women were exclusively auxiliaries who “contributed by distributing publications, carrying food and ammunitions, handling administrative work at FLOSY’s offices, and occasionally giving political talks.”456 In contrast, women were regularly involved in the NLF. When the NLF was born, the Arab Women’s Society – a local political organization – splintered and members supporting armed revolution joined the militant group. Arenfeldt and Al-Hassan Golley estimate that between 200-300 women were active in the NLF.457 These women performed auxiliary roles but also found their way to the front-line. In urban areas, women couriered arms, food, and money and secretly distributed NLF propaganda.458 In rural areas, women took up arms. The NLF included at least one female unit commander and at least one female martyr killed by British forces. Women played such a significant role that Britain announced large financial rewards for information leading to some female fighters’ arrests.459 Looking beyond ‘high-intensity’ conflicts My data suggests that women often make significant contributions to rebel organizations that are traditionally excluded from conflict datasets because they do not meet the battle deaths thresholds. For example, like Comandante Tamara many women fought on the front-line and worked in auxiliary roles for the FPMR. The government convicted Silvia Brzovic Pérez and Marcela Mardones, FPMR militants, of murdering Senator Jaime Guzmán in 1991.460 Multiple FPMR women died in shootouts with state forces. Women also called in bomb threats. In one case, an FPMR women called a Spanish news agency positioned across from the U.S Consulate and told them, “This is the Patriotic Front speaking. Two car bombs will explode in front of the U.S. Consulate. You have 10 minutes to abandon your offices. This is a revolutionary warning.”461 Other women served as FPMR spokespeople, and 456 Amel Nejib al-Ashtal, 2012. “A Long, Quiet, and Steady Struggle: The Women’s Movement in Yemen,” in Pernille Arenfeldt and Nawar Al-Hassan Golley, eds. Mapping Arab Women's Movements: A Century of Transformations from Within, American University in Cairo Press, pg. 209-210. 457 Ibid. 458 Ibid. 459 Ibid. 460 CE Noticias Financieras English, 2018. “Ex-fronting couple Ricardo Palma Salamanca is arrested in France;” CE Noticias Financieras English, “Case Guzmán: UDI qualifies as “unacceptable” statement by Escobar Poblete.” 461 United Press International, 1985. No Headline in Original 164 on one occasion an armed, mixed-gender group hijacked a train.462 Based on Chilean and international news reports, I conclude that women composed approximately 30 percent of participants in reported front-line attacks. The FPMR may also have benefitted from international help. The group had ties to the IRA, and in 1996 three Irish women broke four FPMR members out of a Chilean prison by hiring a helicopter to fly over the facility and drop down a basket.463 Conclusion These data suggest that women’s participation in rebel organizations is frequent and highly varied. Understanding how and where women contribute to political violence provides a more robust understanding of conflict dynamics. Moreover, it can help us understand conflict trajectory and explain consequential wartime events. Chapters 3 and 4 discuss how representations of women during conflict can shape ideological rebel legitimacy among civilians, and how rebels securitize women in highly gendered ways as a result. In the following chapters, I use the RWDP to explore how women’s actual participation affects rebel trajectory, specifically focusing on areas of rebel success. Chapter 6 assesses the relationship between female militancy and rebel territorial control, emphasizing that women’s involvement in all roles is related to an increased probability of control. Chapter 7 examines how women’s participation affects conflict outcomes. I find that that groups with female combatants, auxiliaries and leaders are much more likely to achieve victory than experience government defeat. 462 United Press International, 1990. “Retired air force general wounded by gunmen.” 463 New York Times, 1997. “Chilean Leftists Flee Jail in Copter,” 1 January. 165 Chapter 6: Rebel Women and Territorial Control Most insurgencies aspire for territorial control, but the barriers to entry are immense and few succeed.464 Why are some rebel organizations able to control territory while others fail? In this chapter, I interrogate the effects of women’s participation in rebel organizations on the probability of insurgent territorial control during conflict. I suggest that women’s participation can help build ideological legitimacy by signaling the ‘rightness’ and collectiveness of struggle and political legitimacy by contributing to perceptions of viability. To review, from my theoretical framework I generate a series of expectations concerning women’s militancy and territorial control. I suggest that in cases of front- line participation, women-inclusive groups are more likely to control territory than other groups, but substantive effects may not follow a progressive pattern by level of participation. For female auxiliaries and leaders, I expect that women-inclusive groups are more likely to control territory than other groups, with substantive effects following a progressive pattern by level of participation. In the sections that follow, I discuss my quantitative approach and its advantages and disadvantages before turning to empirical analysis and statistical results. I then offer a micro-case illustration of paired cases from the communist insurgency in Peru to explicate the legitimacy mechanisms potentially working in cases where women-inclusive and -exclusive groups vie for territorial control. Method To assess the relationship between women’s participation in rebel groups and rebel territorial control, I estimate a series of logistic regression analyses with standard errors clustered on the individual actor. As discussed in Chapter 3, rebel groups regularly overlap, which can introduce bias into the statistical models. Clustering them this way helps to mitigate the assumption of independence between observations. Here, I discuss my statistical approach for these analyses. Transforming women’s militancy variables In my data, there is one case of women’s occasional auxiliary participation (Table 6.1). This is Hizb-e Wahdat, where there is evidence of at least one woman who, in the group’s first year, established a martyr’s clinic and with her aides administered schools in Afghanistan and Pakistan.465 464 Asal et. al, “It Comes with the Territory.” 465 Hafizullah Emadi, 2002. Repression, Resistance, and Women in Afghanistan, Praeger. 166 None Occasional Low Moderate High N Combatants 74 10 30 11 18 143 Auxiliaries 76 1 18 19 25 139 Leaders 101 14 7 6 11 139 Table 6.1: Distribution of the level of women’s participation by role Because of this single case in its own category, the logistic regressions result in a perfect separation on this level of participation. This artificially inflates the coefficients and destabilizes the model. Given the nature of auxiliary participation, it is likely that few groups truly have ‘occasional’ female participants in support roles. Consequently, I bin together the occasional and low categories for the women’s auxiliary participation variable. Conceptualizing and measuring territorial control Scholars are enmeshed in the question of how to measure territorial control during conflict, explained succinctly by Tao and colleagues: “Territorial control is central to the understanding of violent armed conflicts, yet reliable and valid measures of this concept do not exist.”466 The spatial extent to which rebels control territory is highly varied. Some groups, like the PIRA, exert fragmented political control over individual neighborhoods and establish services – for example, Sinn Féin operated its own postal service in Catholic areas of Belfast – while others, like the LTTE, maintained nearly autonomous control of large swaths of their target countries. Still, nuanced data on territorial control is just now burgeoning. Stewart and Liou, for example, compile binary measures of territorial control that differentiate between domestic control and control across borders.467 The Minorities at Risk Organizational Behavior (MAROB) dataset, which includes etho-political minority groups, offers an ordinal measure of territorial control ranging from 0-2: territorial control not used as a strategy; organization controls movement but has not established governing structures; and organization maintains governing structures of infrastructure, respectively.468 Kalyvas adopts a five-level model of territorial control, ranging from complete incumbent control to complete insurgent control with 466 Ran Tao et. al. 2016. “A hybrid approach to modeling territorial control in violent armed conflicts,” Transactions in GIS 20, pg. 413. 467 Megan A. Stewart and Yu-Ming Liou, 2016. “Do Good Borders Make Good Rebels? Territorial Control and Civilian Casualties,” The Journal of Politics 79(1): 284–301. 468 See: Asal, et. al. “It Comes with the Territory.” 167 degrees of contestation in between, and he applies this model using case-study research.469 The most substantive, nuanced measures of territorial control are confined to a single region or country. For example, the Carter Center’s Syria mapping project provides dynamic measures of territorial control among factions in the civil war. Rubin measures rebel territorial “influence” on a four-level scale in the Philippines using Armed Forces of the Philippines year-end reports.470 Unfortunately, these ordinal data are not feasible for this project because of their scope To measure territorial control, I use the Non-State Actor Dataset’s (NSA) binary measure, operationalized as “This field indicates whether the rebel group controls territory.”471 I independently assess cases in my data not included in the NSA. While crude, the NSA measure is an established indicator in research on territorial control.472 The primary benefit of this measure is its cross-national scope, which is essential to these analyses. The primary limitation of this measure is that a binary estimation of territorial control conflates extensive governance with lower levels of control. Therefore, I make no claims concerning the degree of territorial control that rebels exert, only the probability that rebels will control territory in some capacity. No territorial control Territorial control No female militants (overall) 52 16 Female militants (overall) 42 36 No female combatants 56 18 Female combatants 37 34 No female auxiliaries 57 19 Female auxiliaries 33 32 No female leaders 70 31 Female leaders 21 19 Table 6.2: Cross-tabulation of women’s participation, by role, and rebel territorial control The distribution of control varies cross-contextually, but an estimated 36 percent of rebel organizations in my data controlled territory. Insurgents in some regions are much more likely to control territory than others: for example, 47 percent of Middle Eastern rebel groups control territory compared with only 17 percent of Asian organizations. Territorial control also varies slightly by group type. An estimated 12 percent of Islamist groups control territory, as do around 9 percent of leftist 469 Kalyvas, The Logic of Violence in Civil War. 470 Rubin, “Rebel Territorial Control and Civilian Collective Action in Civil War.” 471 David Cunningham et. al., Codebook for the Non-State Actor Data, pg. 6. 472 E.g. Megan A. Stewart, 2018. “Civil War as State-Making: Strategic Governance in Civil War,” International Organization 72(1): 205–226. 168 organizations. Somewhat surprisingly, only around 6 percent of ethnic groups achieve territorial control. This suggests that ethnic salience between organizations and a domestic population may not necessarily be as advantageous as expected. No territorial control Territorial control Female combatants (none) 56 18 Female combatants (occasional) 5 5 Female combatants (low) 16 14 Female combatants (moderate) 6 5 Female combatants (high) 8 10 Female auxiliaries (none) 57 19 Female auxiliaries (occasional/low) 13 6 Female auxiliaries (moderate) 10 9 Female auxiliaries (high) 8 17 Female leaders (none) 70 31 Female leaders (occasional) 9 5 Female leaders (low) 4 3 Female leaders (moderate) 4 2 Female leaders (high) 2 9 Table 6.3: Cross-tabulation of women’s participation, by role and level, and rebel territorial control Alternative explanations I assess the effects of women’s participation on rebel territorial control when accounting for the alternative explanations identified in Chapter 2. Table 6.4 details the summary statistics for all variables used in these analyses. I create a dummy variable to assess if an insurgency’s membership is primarily drawn from an ethnic group and/or if the insurgency’s platform explicitly advocates for an ethnic group. Indigenous groups are coded as ethnic groups. To explore the effect of natural resources I adapt Rustad and Binningsbø’s assessment of “natural resource issues,” coding the dummy variable as 1 if resources are a distribution, financing, or aggravating factor of the conflict. I use their criteria to independently assess cases not in their data. I also use Fearon and Laitin’s indicator of what percent of a rebel’s primary target country is mountainous as a proxy for measuring the effect of rough terrain on insurgency success. I create a four-level categorical variable to measure the competitiveness of rebel group’s environment where I code groups with no competing organizations as 0; groups with 1-2 competing organizations as 1; groups with 3-4 competing organizations as 2; and groups with 5 competing organizations or more as 3. I use the UCDP External Support Dataset and independent assessment to code a dummy variable 169 that measures if an external government supports an insurgency financially, through training, or with sanctuary. I use the natural log of the Composite Index of National Capability (CINC) from the Correlates of War Project’s National Materiel Capabilities (v4.0) dataset to measure state capabilities. I take the average CINC rating of the years an insurgency operated. Table 6.4: Chapter 6 summary statistics I also measure the effect of religiosity. Many religious insurgencies, like the Holy Spirit Movement in Uganda, are Christian, Buddhist, or other affiliations. However, there no groups in the data with an explicitly religious agenda that is not Islamist. I therefore code a dummy variable that indicates if the insurgency’s platform explicitly calls for militant violence in the name of Islam. Primarily Muslim groups without a stated Islamist agenda are coded as 0, not Islamist. In future iterations of this project, I will expand the dataset to include other non-Islamist organizations to assess the effect on type of religiosity. To measure rebel group size, I rely on the NSA, Stanford’s Mapping Militants project, and independent assessment to construct a continuous variable.473 Finally, I include two control variables. 473 My data include some rebel organizations not included in existing datasets as well as newer groups that, though they will likely appear in the UCDP and similar data projects’ updates, in 2016 were not yet included. While this does not substantively affect data collection for most variables, I am missing approximately 14 percent of observations for estimated rebel size. This is, theoretically, not inconsequential. However, as detailed below, the models perform significantly better when including this variable. The rebel size variable is also highly varied, with a very high standard 170 I take from World Bank data the natural log of the average population in an insurgency’s primary target country during operations. Population is a useful control because rebels’ ability to control territory may be related to the number of people that the state must manage. I use the Polity IV Project’s polity rating to assess how democratic a country is while fighting an insurgency. Much of the literature argues that democracies are inferior counterinsurgents because of high audience costs from war-avoidant populations and because insurgents can exploit voting patterns through highly visible attacks.474 Empirical results The sections that follow detail the results of these analyses. I assess the relationship between women’s participation as combatants, auxiliaries, and leaders on territorial control during conflict. I also estimate these regressions using the dummy variable version of these indicators as a robustness check. I find, across models, that women’s participation in all roles significantly and substantively increases the likelihood of rebel territorial control. Figure 6.1 details the results of Model 1 (combatants), Model 2 (auxiliaries), and Model 3 (leaders). These preliminary models include all identified alternative explanations for territorial control, with AICs of 160.87, 152.19, 158.51, respectively.475 In these models, women’s combat participation is a significant indicator of territorial control at occasional levels (p<0.05), low levels (p<0.05), moderate levels (p<0.05), and high levels (p<0.01). Women’s auxiliary participation is not a significant indicator of territorial control at occasional/low levels, but is significant at moderate (p<0.05) and high levels (p<0.01). Finally, women’s leadership participation is only a significant indicator of territorial control at high levels (p<0.01). The significance of other explanations vary across models. As evidenced in Model 1, leftist groups are slightly less likely to control territory than other organizations. In Model 2, rebels operating in more democratic states are less likely to control territory. Model 3 suggests that groups in competitive environments are slightly more likely to control territory. deviation. To mitigate bias that including this tenuous, but theoretically important, variable may cause my analyses, I supplement my estimations with an imputed version of this variable. These analyses are detailed in the Appendix. 474 See: Ethan Bueno de Mesquita and Eric Dickson, 2007. The Propaganda of the Deed: Terrorism, Counterterrorism, and Mobilization. American Journal of Political Science 51(2): 364-381; Robert A. Pape, 2005. Dying to Win: The Strategic Logic of Suicide Terrorism. Random House. 475 N= 118, 114, 114, respectively. 171 Figure 6.1: Model 1, Model 2, and Model 3: effects of women’s participation on territorial control (all explanations). However, these models are not the best estimators of territorial control. To identify the most explanatory models, I use manual variable deletion to identify the best-fitting and theoretically appropriate models based on AICs and generalized Fisher scores. I consequently exclude external support, whether the group is an ethnic group, and the percentage of mountains in a target country from the updated models (Models 4-9). Excluding these variables improves model fit with slight changes in the results discussed above. Table 6.5 details the results of Model 4 and Model 5, which assesses the effect of female combatants on territorial control using ordinal and binary variables, respectively. I again find that groups with female front-line fighters at all levels are significantly more likely to control territory: women’s combat participation is a significant indicator of territorial control as occasional levels (p<0.1), low levels (p<0.01), moderate levels (p<0.05), and high levels (p<0.01). The significance levels for occasional and low levels are slightly different in these models than the full model discussed 172 above. In the binary assessment, female-inclusive groups are highly likely to control territory (p<0.01). I also find in Model 4 that leftist organizations are slightly less likely to control territory. Model 4 indicates that rebel territorial may be less likely in less democratic countries. Table 6.5: Models 4 and 5, effects of female combatants on territorial control Table 6.6 details the results of Model 6 and Model 7, which assesses the effect of female auxiliaries on territorial control using ordinal and binary variables, respectively. These results again suggest that groups with female auxiliaries at moderate levels (p<0.05) and high levels (p<0.01) are most likely to control territory during conflict. These results hold when employing the binary estimator of female auxiliary participation (p<0.01). These models suggest that rebels are more likely to control territory when they operate in less democratic states. 173 Table 6.6: Models 6 and 7, effects of female auxiliaries on territorial control Table 6.7 details the results of Model 8 and Model 9, which assesses the effect of female leaders on territorial control using ordinal and binary variables, respectively. Women’s leadership participation is significantly related to territorial control only at the highest level (p<0.01) in Model 8. Model 9’s binary analysis also suggests that groups with women leaders are significantly more likely to control territory (p<0.05). The level of democracy remains a significant indicator of rebel territorial control, and Model 8 indicates that groups in competitive environments may control more territory than other organizations. Across models, these results suggest that rebel groups that include women in combat, auxiliary, and leadership roles are significantly more likely to control territory, though the significance of levels of participation varies by role. 174 Table 6.7: Models 8 and 9, effects of female leaders on rebel territorial control To analyze the substantive effects of women’s participation on territorial control, I estimate the predicted probability of this outcome based on women’s involvement as combatants, auxiliaries, and leaders. This approach tells us, with each increase in the level of women’s participation, how much does the probability that groups will control territory increase? This approach holds all other variables at their mean. Figures 6.2, 6.3, and 6.4 illustrate these predicted probabilities for Model 4 (combatants), Model 6 (auxiliaries), and Model 8 (leaders). The vertical black bar in each plot denotes the percent of groups that control territory in these data, around 36 percent. 175 Figure 6.2: Model 4 predicted probabilities Figure 6.3: Model 6 predicted probabilities 176 Figure 6.4: Model 8 predicted probabilities The predicted probabilities indicate that in all cases, there is a statistical separation moving from no female participants to a high level of female participants. In the case of female combatants, having a high level of female participants increases the probability of rebel territorial control from between ~10-25 percent to at least 45 percent. This effect is more significant for female auxiliaries, increasing the probability of territorial control to at least 60 percent when women participate at high levels. The same is true for female leaders. These results suggest that rebel organizations with widespread female participation are much more likely to control territory than their counterparts. It also highlights the importance of disaggregating women’s participation in rebel groups by role, a finding that can hopefully be sharpened in the future with more data. The substantive effects of women’s participation at other levels is less conclusive. As evidenced in these visualizations, these estimates include wide confidence intervals. This is a function of the small number of cases in this dataset, and this limitation makes it difficult to identify a patterned effect by level of participation. Still, for auxiliaries the point estimates and confidence intervals appear to be moving in the theorized direction, increasing the probability of territorial control as women’s participation increases. This is not true for women’s leadership: there is no substantive change moving from no female leaders to occasional, low, or moderate participation. While the confidence intervals make conclusions drawn from these data preliminary, these findings suggest that women’s leadership 177 may be less substantively important as an indicator of territorial control than women’s auxiliary and combat participation. This is the opposite of my theorized effect, where I expected that women’s leadership is a bold show of inclusion and may increase legitimacy and territorial control at all levels. However, when women participate as leaders at high levels, it is unlikely that they are not also participating in combat and/or auxiliary roles at high levels as well. For example, only two groups in these data include women leaders but exclude female combatants. Therefore, the leadership variable may be capturing some of what the other participation variables measure. Collapsing women’s participation into binary variables helps elucidate the importance of women’s participation across levels and mitigate the cloudiness of occasional, low, and moderate participation in the graphs above. This is especially important because most groups that include women do so at occasional, low, and moderate levels rather than high ones (see Table 6.1 above). Using the binary estimator of women’s participation holds all levels of participation equal and, while this does not help clarify the effects of women’s participation at individual levels, it does take women’s involvement across the board into consideration in analysis. Figures 6.5, 6.6, and 6.7 illustrate the predicted probabilities of territorial control using binary women’s participation variables as assessed in Model 5 (combatants), Model 7 (auxiliaries), and Model 9 (leaders). These visualizations tell us what the predicted probability of rebel territorial control is when we move from having no female participants to having female participants at any level of participation. The results are clear: women’s participation greatly improves the probability of territorial control, though slightly overlapping confidence intervals means that this relationship is less determinate for leadership roles. This suggests that women’s participation does matter, and likely not only at high levels. Figure 6.5: Model 5 predicted probabilities 178 Figure 6.6: Model 7 predicted probabilities Figure 6.7: Model 9 predicted probabilities One key concern is the direction of the relationship between women’s participation and territorial control. It is possible that groups with territorial control are simply more likely to recruit women, introducing an endogeneity problem. Territorial control can signal political legitimacy and capability to civilians (and, as I discuss further in Chapter 7, to the state).476 Consequently, civilians could become more supportive of rebels who control territory and this could yield higher rates of recruitment from noncombatants overall. Rebels who control territory may also be particularly interested in service provision, an area in which auxiliary women may be explicitly recruited after-the-fact.477 However, I 476 Kalyvas, The Logic of Violence in Civil War. 477 See, e.g., Stewart, “Civil War as State Building: Strategic Governance in Civil War,” for a discussion of territorial control and service provision among rebels. 179 suggest that women’s participation improves the probability of territorial control, and not other way around, for three reasons: timing, barriers to entry, and rationale. First, in the broader RWDP I construct an identical set of variables to those discussed in this project that measure women’s participation in an insurgency’s first five years. The average lifespan of an insurgency in these data is 14 years and the median is 10 years. I find that 90 percent of rebel groups that include women in some capacity do so within the first five years. Second, and relatedly, the barriers to entry for rebel territorial control are too high for most organizations.478 Asal and colleagues conclude that “an organization that controls territory has something of great political and economic value” and that control signals rebel capability.479 Nearly all rebels, including terrorist organizations with seemingly trans-territorial aims, aspire to territorial control, but in my data only around 36 percent succeed.480 There is not good cross-national data on the timing of rebel territorial, but when the timing of women’s participation and barriers to territorial conquest are taken together, it is unlikely that most female-inclusive rebel groups are able to control territory during their formative years before they introduce female militants. Third, women’s participation appears to improve the probability of territorial control regardless of what role women play within the organization. While territorial control may introduce a new set of incentives, like service provision, for rebels, these auxiliary programs (education, healthcare, and religious services, for example) can be introduced before a group controls territory and as a means of territorial takeover. Further, if these new incentives prompted women’s recruitment we would likely see a more substantive difference in the relationship between female involvement and territorial control across roles. For these three reasons, I expect that these models represent a relationship that moves directionally from female involvement to territorial control, and not vice-versa. Assessing the effects of women’s participation across group types The primary limitation of this study is that the RWDP is a work in progress, and consequently there is currently not a lot of data with which to analyze relationships within smaller data categories. Still, I preliminarily explore the relationship between women’s participation in rebel organizations and territorial control in sub-sampled group. This analysis is useful, even in a preliminary stage, because Islamist and leftist groups are least and most likely to include women, respectively. As I hypothesize 478 Asal et. al. “It Comes with the Territory.” 479 Ibid, pg. 4. 480 See: Jaume Castan Pinos and Steven M. Radil. 2018, “The Territorial Contours of Terrorism: A Conceptual Model of Territory for Non-State Violence,” Terrorism and Political Violence, OnlineFirst. 180 in Chapter 2, this may lead to inter-relationships between women’s participation and other explanations for territorial control. Table 6.8 details the results of these models, which use the binary estimators of women’s participation due to already small sample sizes. I include all alternative explanations for variation in territorial control because sub-sampling removes two covariates from the analysis and leaves comparatively few explanations. Models 10 11, and 12 analyze women’s participation as front-line fighters, auxiliaries, and leaders, respectively, in leftist organizations. Models 13 and 14 assess women’s participation in front-line and auxiliary roles, respectively, in Islamist groups. I exclude analysis of women’s leadership participation in Islamist groups because there is a perfect separation between territorial control and this binary variable. Therefore, I cannot reliably assess the relationship between women’s leadership participation and territorial control in Islamist groups using these data. Interestingly, these preliminary results suggest that at the cross-national level, women’s participation in rebel organizations has the most explanatory value in cases where rebels are neither leftist nor Islamist. Women’s participation is an insignificant indicator of territorial control across models. As hypothesized, this may be true because states launch such ferocious counterinsurgency campaigns against many leftist organizations that legitimacy and quality gains generated through female participation is negated by government force.481 The insignificance of women’s participation in Islamist groups is surprising given my theory, as I expected women’s involvement to be related to territorial control in these organizations. This result may be indicative of the complexity of gender norms and the rigidness of these expectations in conservative groups: the legitimizing effects of women’s participation may be less pronounced in Islamist groups. In the leftist samples, groups are slightly more likely to control territory when they are larger, when they are operating in conflicts where natural resources are an aggravating factor, and when they fight in countries with difficult terrain. In cases of Islamist insurgents, low state capacity is a slightly significant indicator of rebel territorial control. 481 The large coefficients are not the result of high correlation between these variables and the territorial control variable: they are minimally correlated. 181 Table 6.8: Models 10-14, effects of women’s participation on territorial control, group-type sub-samples Importance of role disaggregation Finally, to explore the importance of role disaggregation in the study of women’s rebel participation and territorial control, I re-estimate my primary models using a measure of women’s overall participation rather than a by-role variable. Model 15 uses a categorical estimate that assesses the 182 highest recorded level of women’s participation in any role within a rebel group. Model 16 uses a binary estimate, indicating whether or not women participated in a rebel organization at all, in any capacity (not shown). Figure 6.8 offers the predicted probabilities for women’s militancy in Model 15.482 Figure 6.8: Models 15 predicted probabilities These results indicate that while women’s overall participation when assessed as a binary is a significant indicator of territorial control (p<0.01), women’s overall participation in the categorical variable varies in significance by level of participation. In Model 15, women’s overall participation is a slightly significant indicator of territorial control at moderate levels (p<0.1) and a very significant indictor at high (p<0.01) levels. These findings reiterate the importance of role disaggregation, demonstrating that collapsing female participation into a singular experience does not accurately represent the effects of women’s combat, auxiliary, and leadership contributions during conflict. 482 N and AIC= 119, 153.724; 120, 156.974, respectively. 183 Case illustration: Leftist rebels in Peru Case evidence helps to illustrate the mechanisms underlying these statistical results. Paired comparison of two leftist insurgencies operating in roughly the same period in Peru, the Revolutionary Left Movement (MIR) and Shining Path, demonstrate how civilian support matters for territorial control and how women’s participation can play a crucial ideological and political legitimacy-building role towards that objective. Although women’s participation in leftist groups may not be related to territorial control at the cross-national level, as discussed in Chapter 2 contextual evidence suggests that it may play an important role in individual cases. The MIR largely excluded women from participation, failed to earn local civilian support, and collapsed after leader decapitation without ever successfully controlling territory. In contrast, Shining Path integrated women widely and at all levels from inception, enjoyed widespread civilian support, and controlled large areas of Peru. The group also explicitly used women’s participation, and gendered representations of it, to generate legitimacy. In 1962, Luis De la Puente Uceda founded the MIR out of an American Popular Revolutionary Alliance (APRA) splinter group. The MIR advocated for an armed peasant uprising against the bourgeoisie. The organization launched three fronts aimed at mobilizing peasants from within their communities, recruiting widely and integrating itself into agrarian villages. A legitimizing public relations campaign was central to this exercise: Brown and Fernández conclude that “De la Puente counted on peasant support for the MIR’s… program of guerilla struggle… [he] orchestrated the best- advertised guerilla struggle in the history of Latin America.”483 The MIR leveraged saliences that it thought would appeal to the civilian population, even when it meant mis-representing the organization’s ideology and platform. For example, to build support among indigenous populations the group “sometimes used Andean religious symbols for rhetorical effect in its communiqués [official statements]… but a deeper knowledge of the role of these mountain spirits in the peasant’s life on the land was nowhere to be seen.”484 However, despite their extensive legitimacy-building campaigns, the MIR never gained consequential support. There is little evidence that their outreach or organizational activities were effective.485 Despite operating fronts in three relatively diverse parts of the country, their failure to win over civilians in those areas meant that the group never gained even a semblance of territorial control. Government forces killed De la Puente and most MIR guerillas in 1965, after which the remaining members splintered into other organizations. 483 Michael Brown and Eduardo Fernández, 1993. War of Shadows: The Struggle for Utopia In The Peruvian Amazon, University of California Press, pg. 90. 484 Ibid, pg. 90. 485 Ibid. 184 The MIR sold itself as a people’s revolution requiring dedication and sacrifice. In a famous, widely-circulated 1964 speech, De la Puente ties the group to Cuban revolutionaries and other leftist organizations cross-nationally. He notes that revolution demands people from all social sectors and classes collaborating in armed struggle.486 Still, the MIR excluded women almost entirely and was reluctant to address gender issues. De la Puente explicitly relegates women to the periphery in his 1964 speech: despite the MIR’s ‘people’s revolution’ platform, he does not mention women except as the wives-of revolutionaries. Even when situating the MIR within leftist movements in Bolivia, Venezuela, and other Latin American countries with female militants, he speaks exclusively of young men and peasant men’s contributions to rebellion.487 This speech encapsulates the MIR’s approach to women in the movement- there is almost no evidence of women’s participation. The sole exception is documentation of three female couriers.488 In the late 1960s, after the MIR’s collapse, Abimael Guzmán and his wife Augusta la Torre founded the communist Shining Path as a student movement within Peruvian universities. Like the MIR, the group advocated a peasant rebellion focused on agrarian struggle. In 1977, the Shining Path launched an armed ‘people’s war’ against the Peruvian government. They carried out their first attack in 1980. The organization operated in urban centers and in rural villages and viewed community integration as a primary vehicle for merging these environments. La Serna concludes, Shining Path guerrillas counted on high school and university youths not only to provide the military backbone of the revolution but also, and more important, to serve as intermediaries between the cities and the countryside, where most students’ families were from. It was through these networks that the PCP-SL hoped to amass peasant support. The guerrillas achieved this support with varying degrees of success, but by late 1982 state authorities were already complaining about the ease with which the rebels were able to recruit and train peasants - ‘even elderly women’ - in the absence of a strong rural police force.”489 Mobilized by social ties, many indigenous communities supported Shining Path by providing shelter, funds, intelligence, and recruits. The organization eventually controlled large swaths of territory in both rural and suburban Peru, instituting educational systems and extra-legal justice mechanisms. But Shining Path was brutal, and over time killed thousands of civilians as punishment for non- compliance. Peasants in some villages rose up and joined with state counterinsurgents to beat back 486 Centro de Documentación de los Movimientos Armados, Discurso del Dr. Luis Felipe De La Puente Uceda en la Plaza San Martin De Lima - Peru, el 07 de Febrero de 1964. 487 Ibid. 488 Brown, Cuba’s Revolutionary World. 489 Miguel La Serna, 2012. The Corner of the Living: Ayacucho on the Eve of the Shining Path Insurgency, University of North Carolina Press, pg. 144. 185 the organization.490 The group remains technically active, though operating in a very low-intensity conflict. Why was Shining Path able to mobilize local communities and efficaciously control territory while the MIR failed in this endeavor? There are several contributing factors. Shining Path met little government resistance in rural areas in its initial years, enabling the group to develop without significant opposition. The organization also appealed directly to indigenous people concerned with lawlessness who felt that the government had abandoned them, carrying out punishments and building political legitimacy early on. Further, Shining Path successfully marshaled ideological support from civilians along racial and class lines in a way that the MIR was unable to materialize.491 And, most importantly for the purposes of this study, Shining Path widely integrated and publicized its employment of women militants at high levels in all roles. Women’s inclusion in the group from its inception was a central part of Shining Path’s platform, a lynchpin in their ideological recruitment, and a selling point in their mobilization of civilian communities. Guzman noted in 1970, “The success of our movement hinges on the active participation of women.”492 Women participated in Shining Path’s founding, and they composed comparatively large proportions of auxiliary, front-line, and leadership positions. By 1987, half of all Shining Path members charged with terrorist activity were women.493 Women composed an estimated 40 percent of political cadres and 50 percent of the organization’s National Central Committee, and women served as high-ranking commanders on the front-line. Female spies and other auxiliaries scaffolded the group.494 Shining Path explicitly encouraged women’s militancy and focused on women’s emancipation as a core tenet of its insurgent campaign. They stressed educating women, liberating them from capitalism and men’s tyranny in the home, and creating new political opportunities. The group also feminized its public face: Shining Path relied heavily on women for recruiting.495 Ortega notes that the organization used women’s integration in target territories to spread their ideology and suggests that women recruiters uniquely fostered comfort among civilians. Shining Path also emphasized women’s participation to demonstrate the exigency of struggle, “using arguments such as, ‘even those women are in the armed struggle, how can you, as a man, not be involved.’”496 490 Ibid. 491 Ibid. 492 Quoted in Gonzalez-Perez, Women and Terrorism, pg. 40. 493 Ibid. 494 Ibid. 495 Cragin and Daly, Women as Terrorists. 496 Luisa Maria Dietrich Oretga, 2011 “Gendered Patterns of Mobilization and Recruitment for Political Violence,” in Yvan Guichaoua, e.d. Understanding Collective Political Violence, Palgrave Macmillan, pg. 97. 186 The organization further adopted different gendered frameworks for recruiting in different communities. For example, when recruiting in the comparatively educated, liberal universities Shining Path stressed the need to destroy the state to keep it from professionally subjugating women. In contrast, when recruiting in more conservative, rural communities, they stressed community values and the necessary heroism of women warriors.497 The group also emphasized the need to protect women from state-perpetrated sexual violence and domestic violence in territorial areas with low regulatory oversight.498 This multi-pronged approach enabled Shining Path to deploy appropriate gender expectations into its target communities and use women’s participation to leverage ideological legitimacy into civilian support while establishing territorial control and recruiting new members. Women also provided Shining Path with political legitimacy by signaling group strength and viability. On the front-lines, women were highly effective attackers. They were frequent participants in major military operations, including often delivering the fatal shot in very visible, political assassination attempts.499 Off the front-lines, women contributed significantly to the auxiliary operations that keep insurgencies alive. Men and women divided supportive labor relatively equally in Shining Path, creating a broad-based and well-staffed logistical structure.500 Women extensively participated in political leadership, making policy and creating the codes and regulations that those in Shining Path territory abided. With women’s help, the organization built substantive governing structures and positioned itself as a viable alternative to the Peruvian state. This likely assisted Shining Path in taking over territories, as well as maintaining them in the aftermath. Conclusion In this chapter, I use my original data of women’s participation in front-line, auxiliary, and leadership roles to explore the relationships between female militancy and rebel territorial control during conflict. Using a series of logistic regressions, I find that women’s participation in all roles greatly improves the probability of territorial control. I suggest that this is the result of intersecting legitimacy and capability factors. Women’s participation shapes public perceptions of rebel grievances, signaling the appropriateness and exigency of struggle and legitimating rebels’ cause. Further, women’s participation 497 Ibid. 498 Gonzalez-Perez, Women and Terrorism. The group’s official code of behavior forbade sexual violence perpetrated against women, but also explicitly forbade sexual violence perpetrated against men. 499 David. J. Whittaker, ed. 2003. The Terrorism Reader, Routledge. 500 Juan Lázaro, 1990. “Women and Political Violence in Contemporary Peru,” Dialectical Anthropology 15(2/3): pg. 233- 247. 187 likely increases rebel capabilities on and off the front-lines, making rebels appear more politically viable and resourcing them to withstand counterinsurgent operations. I contend that territorial control is an important achievement for rebels, and that women-inclusive rebels fare better in this respect than their counterparts. In the following chapter, I assess the effects of women’s participation on conflict outcomes. Thus, I demonstrate the strategic advantages that women militants offer rebels during conflict and at its termination. 188 Chapter 7: Rebel Women and Conflict Outcomes Rebels achieve wins and shoulder losses throughout conflict, but why do some ultimately achieve favorable outcomes while others experience defeat or dissolution? In this chapter, I explore the effects of women’s participation in rebel organizations on the likelihood of achieving ‘favorable’ conflict outcomes: successful peace negotiations and decisive victories. I suggest that women’s participation can help build ideological legitimacy by signaling the ‘rightness’ and collectiveness of struggle and political legitimacy by contributing to perceptions of viability. To review, I expect that women’s participation in all roles increases the likelihood that rebels will sign negotiated agreements because states may be more willing to bargain with them. I expect this effect to be stronger for female front- line fighters and leaders than for female auxiliaries because of their comparative visibility to states and third-party intervenors. I also expect that women’s participation in all roles increases the likelihood of rebel victory and decreases the likelihood of government victory because of women’s legitimizing effects and capability-building. I expect these effects to be strongest for female front-line fighters and auxiliaries, as their participation may have the biggest substantive effects on capabilities. In the sections that follow, I discuss my quantitative approach and its advantages and disadvantages before turning to empirical analysis and statistical results. I then offer a short case illustration of the National Resistance Army’s (NRA) successful insurgency in 1980s Uganda to explicate the legitimacy mechanisms potentially working in cases where women-inclusive and - exclusive groups compete for favorable outcomes. Method To assess the relationship between women’s participation in rebel groups and conflict outcomes, I estimate a series of multinomial logistic regression analyses. This enables me to explore the likelihood of rebels achieving each possible outcome. To maintain the integrity of independence between outcomes, I drop three groups from these analyses. These are all cases of groups composing a larger alliance that operated both independently and on behalf of the umbrella organization (see Chapter 5). First, I exclude the Revolutionary Organization of Armed People (ORPA) and the Guerrilla Army of the Poor (EGP). This is because they are sub-groups of the larger Guatemalan National Revolutionary Unity (URNG). While each group varied in its composition of female participants and I cluster the individuals actors in Chapter 6 to mitigate bias, the URNG signed the 1996 Accord for a Firm and Lasting Peace as a singular unit. I also exclude the Islamic Renaissance Party for the same reason: it signed the 189 1997 General Agreement on the Establishment of Peace and National Accord in Tajikistan as part of the United Tajik Opposition, which is also included in the data. I use binary measures of women’s participation in front-line, auxiliary, and leadership roles in these analyses. This is because distributing a small number of cases over multiple possible conflict outcomes results in small bins across categories. Consequently, I explore the effect of women’s participation at any level, across roles, on outcomes. Using these variables makes excluding organizations that reached decisive victories as part of a sampled larger group even more appropriate. For example, while women’s participation within ORPA and EGP varied, in these analysis that variation is compressed into a dichotomous measure. It is therefore hard to justify the bias that including them would introduce. Conceptualizing and measuring conflict outcomes When does conflict end? This is an elusive question and is not one I attempt to definitively answer in this project. Protracted periods instability mark many ‘post-conflict’ areas. As Lake convincingly argues, armed violence fluctuates across these “not-wars” but the shadow of conflict itself militarizes and burdens local populations. She highlights the “conditions of uncertainty that perpetuate” these “everyday war[s].”501 Rarely does internal conflict operate at a fever pitch for its tenure. Violence ebbs in intensity, rebels and counterinsurgents shift strategies and targets, insurgents switch between violent non-violent expressions of grievance. Still, rebel groups aim to reconstitute or defeat the state and as a result, there are identifiable outcomes signaling conflict termination, even for a time. In this project, I adapt the widely used UCDP Conflict Termination Dataset (CTD) measurements and data for conflict endings.502 There are benefits and disadvantages to these data, the latter of which stem from the UCDP’s 25 battle-deaths threshold discussed in Chapter 2. The CTD identifies six possible outcomes. First, peace agreements are “concerned with resolving or regulating the incompatibility,” and “signed and/or accepted by all of the main parties in the conflict.”503 Second, ceasefire agreements are “concerned with ending the use of force by the warring sides but they can also offer amnesty for participation in the conflict.”504 In cases where a ceasefire agreement is immediately followed by a peace agreement, the peace agreement is considered 501 Milli Lake, 2018. Everyday War: Cyclical Insecurity and the Fragmentation of Violence in Low Intensity Armed Conflict, working paper, pg. 2. 502 Joakim Kreutz, 2010. “How and When Armed Conflicts End: Introducing the UCDP Conflict Termination Dataset,” Journal of Peace Research 47(2): 243-250. 503 Joakim Kreutz, 2016. UCDP Conflict Termination Dataset Codebook v.2-2015, pg. 2. 504 Ibid, pg. 3. 190 the main outcome. The third outcome is a decisive victory for the government, which involves eliminating rebels or rebels publicly laying down arms. The fourth outcome is a decisive victory for rebel organizations, which includes eliminating the government or otherwise taking power as the primary governing structure. The CTD considers “low activity” as the fifth outcome, which is where conflict continues but fatalities fall below 25 in a given year. The sixth outcome is “actor ceases to exist,” which occurs when “conflict activity continues but at least one of the parties ceases to exist or become another conflict actor.”505 Using the CTD data and coding rubric as my guide, I adapt this scheme to better match my data collection. My data includes 27 insurgencies excluded from the CTD. In total, 81 percent of my cases are in the CTD. This is because, like the Morazanist Front for the Liberation of Honduras (FMLH), some groups never met the battle deaths threshold; or because, like Uganda’s Popular Resistance Army (PRA), the CTD considers them to be the same group as their successor or another close organization. There is a clear mis-match between the CTD threshold for inclusion, and denotation of the “low activity” outcome, and my threshold for inclusion. I mitigate this in several ways. First, I include the CTD measure for peace agreements, government victory, and rebel victory. I update the data based on independent research and code of my cases that did not originate in the CTD. Of the 27 cases unique to my data, 9 experienced one of these decisive outcomes. Second, because I do not use the battle-deaths threshold to categorize insurgent activity, I do not view ‘low activity’ as a conflict outcome. Consequently, I collapse the CTD’s “low activity” and “actor ceases to exist” categories together into a single outcome, “actor ceases to exist.” I independently assess cases in the CTD considered “low activity” to establish if they experienced government defeat, dissolution, or another outcome. I re-categorize them accordingly, and most remained in the “ceased to exist” category. Third, I descriptively include but analytically exclude cases where the CTD outcome is ceasefire. This is the most difficult outcome to match with my data because it is not clear what qualifies as a ceasefire in the CTD coding scheme and several cases in my data that are on-going as of 2016 signed ceasefires that broke down. One illustrative example is the PLO. This is particularly difficult to square between datasets because the CTD data is time series data, offering each group multiple outcomes by year. As a result, I exclude both groups that signed ceasefires but did not sign peace agreements and groups that were still operating in December 2016 from the outcomes analysis. With 505 Ibid, pg. 4. 191 these adjustments, I do not expect the inclusion-threshold differences between these datasets to substantively affect the results. Table 7.1 details the data distribution across possible outcomes. Conflict outcome Peace negotiation 22 Ceasefire 17 Government victory 22 Rebel victory 7 Actor ceases to exist 42 Actor continues to exist 33 Table 7.1: Distribution of conflict outcomes (count) across rebel organizations In my data, approximately 32 percent of insurgent groups achieve a negotiated settlement with the government, agree to a ceasefire, or win decisively in battle. Another 15 experience a decisive government defeat, while almost 30 percent cease to exist by fizzling out, fractionalizing without a surviving parent organization, or merging with other rebels to become a new group. An estimated 23 percent of insurgencies in these data were still fighting at year-end 2016 and have not ever achieved a discrete outcome that was not a ceasefire. Importantly, conflict outcomes do not always serve as a tenable cessation of violence. Still, these outcomes are a good proxy for the state’s ability to wage counterinsurgent war and on rebels’ ability to withstand or bargain their way out of these campaigns. In this chapter, I explore the effect of women’s participation on favorable conflict outcomes – negotiated agreements and victories – relative to unfavorable or neutral outcomes – government victory or actor ceases to exist. Table 7.2 shows the distribution of these outcomes across rebel organizations by their inclusion of women participants. The most striking descriptive finding is that every rebel group that achieved a decisive victory over the government included women as auxiliaries. Consequently, every rebel group in this category included women as participants overall. However, the number of rebel organizations in this category is the smallest of any possible outcome: 7. Groups with female combatants and auxiliaries signed more peace agreements than other groups, but this is not the case for female leaders. Governments defeated more insurgents with female combatants than those without, but many more groups with female leaders and combatants won wars than those without women participants. 192 Peace agreement Government victory Rebel victory Actor ceases to exist No female militants (overall) 9 7 0 26 Female militants (overall) 13 14 7 19 No female combatants 9 8 2 28 Female combatants 13 13 5 14 No female auxiliaries 8 9 0 28 Female auxiliaries 13 11 6 14 No female leaders 13 13 2 35 Female leaders 8 6 5 7 Table 7.2: Distribution of conflict outcomes (count) by women’s participation across rebel organizations Alternative explanations I assess the effects of women’s participation on conflict outcomes when accounting for the alternative explanations identified in Chapter 2. Table 7.3 details the summary statistics for all variables used in these analyses. I use several of the same measures identified in Chapter 6. This includes the Rustad and Binningsbø measure of natural resource aggregation, because rebels who can self-finance or seize control of lucrative natural resources may be better positioned to negotiate or stronger in their capabilities; the CINC proxy for state capacity because weak states may be more inclined to offer peace negotiations or ceasefires while being less able to defeat insurgents; the UCDP External Support Dataset measure of external government support, because external assistance may either increase rebel capabilities or diminish home government trust in rebels to uphold their end of a bargain; the Polity2 democratization score, because states in democracies face higher audience costs in their counterinsurgent campaigns that may shape their perception of outcome options; and a categorical competition variable, because insurgents operating in highly competitive environments may be more willing to settle and less likely to win. In addition, for this chapter I code a dummy variable to indicate if the insurgent group seeks independence, autonomy, or otherwise self-governance. Governments may be especially keen to negotiate or declare ceasefires with separatists who view their objects as indivisible. I code a continuous variable which counts the number of a years a rebel group operated. Strong insurgents should fight relatively short wars because of government capitulation in negotiations, but long wars can also exhaust government willingness to fight on. Finally, I also include territorial control as an 193 explanation for variation in conflict outcomes, as rebels with territory have a valuable strategic bargaining chip. I again use the NSA binary measure of this variable. Table 7.3: Chapter 7 summary statistics Empirical results The sections that follow detail the results of these analyses. I assess the relationship between women’s participation as combatants, auxiliaries, and leaders on conflict outcomes. I drop all cases where rebels continue to fight as of 2016. In these logits, I hold government victory as the reference category. Consequently, I analyze the likelihood of three possible conflict outcomes (negotiated settlement, rebel victory, and actor ceases to exist) compared to the likelihood of decisive government victory. I find that rebel organizations with female combatants, auxiliaries, and leaders are more likely to win decisively than experience defeat. I also find that the relationship between women’s combat participation and territorial control is a useful indicator of peace negotiations. Table 7.4 details the results of Model 1 (combatants), Model 2 (auxiliaries) and Model 3 (leaders). In Model 1, women’s front-line participation is a significant indicator rebel victory. Model 2 suggests that rebel groups with female auxiliaries are significantly more likely to decisively defeat governments than be defeated by them during internal conflict. This makes sense, given the distribution of groups in this category. They are also less likely to merge with other groups or cease to exist. Model 4 reports the similar results for female leaders: rebels with female leaders appear significantly more likely to experience victory than defeat and are less likely to cease to exist than to be handily defeated. 194 Across models, several other rebel and state characteristics contribute to explaining variation in conflict outcomes. Weaker states are much more likely to sign peace negotiations and lose to rebels than win outright. Rebels appear more likely to cease to exist in strong states, suggesting that protracted conflicts may work in the government’s favor in some cases. Natural resources emerge as a clear indicator of favorable rebel outcomes: when rebels fight in resource-aggravated conflicts, they are significantly more likely to sign negotiated peace deals and to win. This finding goes against my original hypothesis. Conversely, external support may work against insurgent groups: in Model 1 and Model 3, rebel that receive external support are less likely to merge or dissolve than to be defeated. In Model 2, rebels that receive external support are less likely to win than be defeated. Finally, these results indicate that less democratic countries are more likely to defeat rebels decisively than they are to experience defeat themselves. This is likely the result of comparatively low audience costs that enable brutal counterinsurgencies. 195 Table 7.4: Model 1, Model 2, and Model 3: effects of women’s participation on conflict outcomes 196 To further explore the effects of women’s participation on variation in rebel conflict outcomes, I generate the odds ratios for the effect of each female militancy predictor on each possible outcome. In this case, odds ratios help clarify the likelihood of rebels achieving one outcome, compared to being defeated by the government. Table 7.5 details the odds ratios for Models 1, 2, and 3. Female combatants (Model 1) Female auxiliaries (Model 2) Female leaders (Model 3) Peace agreement 1.30 0.77 3.10 Rebel victory 15.10* 36416546.66*** 42.5** Actor ceases to exist 0.07*** 0.18** 0.42 Table 7.5: Odds ratios: effect of women’s participation on conflict outcomes, * p<0.1; ** p<0.05; ***p<0.01 An odds ratio greater than 1 indicates a positive relationship between the predictor and the outcome, while an odds ratio smaller than 1 indicates a negative relationship between the predictor and the outcome. Thus, when rebel groups include female combatants they are 30 percent more likely to sign a peace agreement than be defeated by the government. However, groups with female auxiliaries may be 23 percent less likely to sign peace agreements than be defeated by the government. These findings, along with the high odds ratio for female leaders in this category, suggest that women’s visibility in militant groups may play an important role in the state’s willingness to negotiate. Nevertheless, none of these peace agreement findings are statistically significant. This is a surprising finding, as I expected that increased ideological legitimacy among civilians and increased political legitimacy through rebel capabilities would result in state willingness to negotiate with insurgencies. Given the ratios but statistical insignificance, I anticipate that a larger sample may yield more conclusive results. Rebel groups with women participants in any role have drastically increased odds of victory relative to their odds of being decisively defeated. Including female auxiliaries increases rebel odds of victory by an improbably ridiculous, and highly statistically significant, number. This is likely the result of the distribution of women’s auxiliary participation across categories and the small number of rebel victories: all 7 rebels that achieved victory included female auxiliaries. Groups with female leaders in this sample are much more likely to win than to be defeated by the government, as are groups with female combatants. These results are all statistically significant and they suggest that women’s participation does improve the likelihood of rebel victory in conflict, but we simply need more cases of rebel victory to assess how broadly this effect applies. 197 Territorial control is a notably insignificant indicator of all rebel outcomes except for peace negotiations in Model 1. Even then, it is only slightly significant at p>0.1. This is surprising, particularly considering the show of tangible strength and, often, civilian support that such control represents. Because of the significant and substantive relationship between women’s participation in all roles and the probability of territorial control, I explore the interaction of this relationship as an explanation for outcome variation. Table 7.6 shows the result of Model 5 (female combatants), Model 6 (auxiliaries), and Model 7 (leadership), which include the interaction terms. Model 4 produces the only significant interaction term: rebel groups with female combatants and territorial control are more likely to win than be defeated by the government There is no significant interaction effect in Model 6 and Model 7, suggesting that women’s auxiliary participation, leadership, and rebel territorial control are better estimators of rebel outcomes independently. In sum, Models 1-7 suggest confidently that rebels with women combatants, auxiliaries, and leaders are more likely to win outright than be defeated; that rebels with women combatants and territorial control are more likely to win outright than be defeated; and that rebels with women combatants are more likely to be decisively defeated than to merge with other groups or otherwise dissolve over time. 198 Table 7.6: Model 4, Model 5, and Model 6: effects of women’s participation and territorial control on conflict outcomes 199 I also generate the odds ratios for the effect of each female militancy variable and territorial control interaction on each possible outcome. Table 7.7 shows the ratios for Model 4, Model 5, and Model 6. Female combatants x Territorial control (Model 4) Female auxiliaries x Territorial control (Model 5) Female leaders x Territorial control (Model 6) Peace agreement 6.29 0.43 125.24 Rebel victory 10278319.41*** 0.30 49.65 Actor ceases to exist 0.94 1.58 2.47 Table 7.7: Odds ratios: interaction effect between women’s participation and territorial control on conflict outcomes The odds ratios further suggest, though most of the results are not statistically significant, that the relationship between female militancy and territorial control may offer an additionally robust explanation for conflict outcomes. This is expected, given the substantive effects of women’s rebel participation on the probability that insurgents will control territory and the theoretical importance of territory to favorable conflict outcomes. Still, we simply need more data to draw reliable conclusions about these relationships. These results remain unchanged when using the binary indicator of whether women participated in any role (see Appendix for results). As hypothesized, women’s participation in rebel organizations improves the likelihood of victory relative to government defeat. However, women’s participation is not a significant indicator of negotiated agreements, relative to government defeat, in any of these analyses. This finding is surprising, as I expected that increases in rebel legitimacy and capabilities prompt states to bargain rather than face strong insurgencies. Still, the odds ratios do confirm my original expectations that women’s front-line and leadership participation may increase likelihood of negotiated settlements while female auxiliaries may not. With a larger dataset and better matched conflict outcomes data, I anticipate that I can draw more responsible conclusions about these relationships. Case illustration: Victorious rebels in Uganda Case evidence from the National Resistance Army (NRA) in Uganda illustrates some of the mechanisms underlying these results. Rebel victory is rare, but women’s front-line, and leadership 200 auxiliary participation appears to be strongly related to this achievement. The NRA’s strategies, and their inclusion of women across roles, highlight the importance of civilian support in many rebel victories and elucidate the critical benefits women can offer off the front-line. In the 1970s, Yoweri Kaguta Museveni recruited and trained fighters to overthrow then- President Idi Amin. His group, the Front for National Salvation (FRONASA), joined Tanzania’s military to successfully take control of the regime. Still, Museveni responded to Uganda’s 1980 election of Milton Obote, who Amin had originally overthrown, with anger and charges of corruption. He re- recruited some FRONASA soldiers, now part of the Uganda’s official security force, the Uganda National Liberation Army (UNLA), launched the NRA, and began the five-year Ugandan Bush War against the government. The war ended in 1985 when the NRA claimed victory and Museveni became president. The NRA’s political party, the National Resistance Movement (NRM) remains Uganda’s ruling political party. Why was the NRA successful in achieving a decisive rebel victory? The organization benefitted from strong central command and leadership, from highly fractionalized government, and from a coup carried about by a rival faction in the war’s closing days. Still, scholars conclude that a central aspect of NRA success was its mobilization of local publics. The group recognized the importance of civilian support from the beginning, and believed that organizing local political networks was crucial.506 Kasfir contends that this commitment resulted from “ideological conviction, relative military strength, dependence on civilian material assistance, and need for accommodation with civilian preferences in its operational area.”507 The NRA surpassed most others in their facilitation of rebel, and civilian, governance. The groups’ “appeal for popular support included requests for civilian material assistance as well as justification of its claim to legitimacy” and they established democratic participation in areas under rebel control. Beyond these shows of political viability, the NRA operated a ‘Publicity and Propaganda Sub-Committee’ responsible for publicizing insurgent struggle and for calling attention to government human rights abuses. Museveni spent time during the NRA’s early years traveling between countries seeking external support.508 UNLA-perpetrated civilian targeting bolstered the NRA’s claims to political and ideological legitimacy, particularly when compared with the group’s attempt to democratize villages in the Luwero Triangle. 506 Edward Katumba-Wamala, 2000. “The National Resistance Army (NRA) as a Guerilla Force,” Small Wars & Insurgencies 11(3): pg. 163. 507 Nelson Kasfir, 2005. “Guerrillas and Civilian Participation: the National Resistance Army in Uganda,1981-86,” Journal of Modern African Studies 43(2): pg. 271. 508 Ibid. 201 Widespread civilian support enabled the NRA to grow quickly and efficiently. With strong military capabilities, the group overran rival bases and absorbed entire units defecting from government forces. 509 Women’s participation in the NRA contributed immensely to these capabilities and to the ideological enfranchisement of the public. Indeed, Mugambe contends that the “greater majority of women were engaged in mobilization work.”510 At the end of the war, there were approximately 800 female NRA fighters, armed with guns and known to lay landmines.511 Women combatants were highly visible, but the group still maintained a semblance of gender ‘role management:’ the NRA often excluded women from particularly “risky” missions.512 Women also served as commanders and trainers. Schubert suggests that when the war reached peak intensity, “gendered spaces were temporarily dissolved” and women participated in all front-line battles.513 However, women primarily took on auxiliary roles aimed at generating civilian support and building military capabilities. One primary NRA need was food from civilian supporters. Women were particularly important in this respect: they got involved in recruitment, in finding and collecting food and in organizing civilians to grow it for them.514 Female auxiliaries also worked as spies, engaging in reconnaissance. They “provided key intelligence information essential for strategic planning,” and they worked as health care attendants.515 In this latter role, women not only worked with injured soldiers but integrated themselves into Luwero Triangle communities. They helped women in childbirth and they practiced traditional herbal medicine to prevent malaria.516 NRA women also worked within the local population to protect vulnerable people. For example, “women participated a lot in the moving children, elderly relatives and any essential items needed for survival, to ‘safer’ areas. Women carried those who were seriously sick and injured to the camps, where they could get medical treatment.”517 The NRA further called attention to women’s vulnerability as evidence of the legitimacy of their grievances. The UNLA committed widespread rape, and the rebels emphasized this violence in their domestic and international justifications for 509 Katumba-Wamala, “The National Resistance Army (NRA) as a Guerilla Force.” 510 Beatrice Mugambe, 2000. Women’s Role in Armed Conflict and Their Marginalisation in the Governance of Post-Conflict Society: The Case of the ‘Luwero Triangle’, Organization For Social Science Research in Eastern And Southern Africa: pg. 10. 511 Ibid. The NRA composed an estimated 9,000 fighters at war’s end, meaning that female fighters made up around 8.8% of combatants. 512 Mugambe, Women’s Role in Armed Conflict and Their Marginalisation in the Governance of Post-Conflict Society. 513 Frank Schubert, 2006. “‘Guerrillas Don’t Die Easily’’’: Everyday Life in Wartime and the Guerrilla Myth in the National Resistance Army in Uganda, 1981–1986,” IRSH. 514 Ibid; Mugambe, Women’s Role in Armed Conflict and Their Marginalisation in the Governance of Post-Conflict Society. 515 Rosalind E. Boyd, 1989. “Empowerment of Women in Uganda: Real or Symbolic,” Review of African Political Economy, 45/46: pg. 107. 516 Mugambe, Women’s Role in Armed Conflict and Their Marginalisation in the Governance of Post-Conflict Society. 517 Ibid, pg. 12. 202 violence.518 More than one source references rape, fear of rape, or anger over rape as a motivation for women joining the NRA.519 Women’s participation undoubtedly shaped NRA engagement with the civilian population, both in terms of ideological legitimation and the distribution of services and other evidence of viable governance. Widespread civilian support made the NRA a ferocious insurgency, and scholars generally credit the groups’ civilian loyalty with their ultimate ascension to power.520 Conclusion In this chapter, I use my original data of women’s participation in front-line, auxiliary, and leadership roles to explore the relationships between female militancy and conflict outcomes. Using a series of multinomial logistic regressions, I find that women’s participation in all roles greatly improves the likelihood of rebel victory relative to government defeat. I contend that this is because women’s participation increases rebel legitimacy among civilians and strengthens rebel capabilities. I further find that women’s participation in front-line and leadership roles increase the odds of negotiated settlements, though these results are not statistically significant. Finally, I find that the relationship between women’s front-line participation and territorial control offers significant explanatory power for rebel victory. In sum, I suggest that rebels believe women’s participation shapes public perceptions of their grievances in ways that ideologically and politically legitimize insurgent struggle, and I further demonstrate that women’s participation does substantively increase the likelihood of rebel successes during and after conflict. 518 Katumba-Wamala, “The National Resistance Army (NRA) as a Guerilla Force.” 519 Mugambe, Women’s Role in Armed Conflict and Their Marginalisation in the Governance of Post-Conflict Society. 520 See, e.g. Kasfir, “Guerrillas and Civilian Participation.” 203 Chapter 8: Conclusion This project is motivated by two key questions. First, how does the focus in extant scholarship on female combatants obscure the gendered dynamics of women’s militancy? Second, how does women’s participation in front-line, auxiliary, and leadership roles affect conflict trajectory? I argue that women’s participation in rebel organizations offers insurgencies a strategic advantage. I suggest that insurgents believe that women’s militancy can make civilians more receptive to their grievances, and I emphasize importance of civilian support to success during conflict and at conflict termination. Specifically, I find that rebel groups cross-nationally securitize women militants and female civilians in their visual propaganda to cultivate ideological and political legitimacy from the public. Rebels capitalize on the assumption that women are traditionally non-violent actors to signaling the ‘righteousness’ of struggle and integrate the insurgency into the private sphere. I demonstrate that insurgents invoke the same gender tropes across time and across conflicts in pursuit of this objective. Further, I find that women’s participation in rebel organization is related to conflict success, and that female-inclusive groups are more likely to control territory and achieve victory. My results also indicate that women’s front-line and leadership participation can increase the odds of negotiated agreements, though these results are not statistically significant. The data used in this project are the result of archival research and quantitative data collection. At the Linen Hall Library and the Eileen Hickey Irish Republican History Museum and through the online archives of the International Institute of Social History and CAIN I accessed approximately 4,000 visuals from the Northern Irish conflict. I constructed an original dataset of republican propaganda produced during the Troubles, identifying 532 images that fit this description. I also collected political images from the PPPA and Afghan, Namibian, Lebanese, Angolan, South African, Mozambican, Nicaraguan, Bangladeshi, and jihadist propaganda from archives including but not limited to the African Activist Archive, the South African History Archive, International Institute of Social History, and Signs of Conflict. I also employ an original dataset of women’s participation in 146 cross-national insurgencies operating between 1960-2016 that assesses women’s front-line, auxiliary, and leadership participation. Through these data collection efforts, I argue that rebels of varying group-types and tactics use their visual outreach to shape narratives of threat and foster civilian support for their campaigns and that women’s participation is related to rebel successes. These findings offer a host of theoretical and policy implications. First, this research highlights the need to disaggregate women’s participation by role and emphasizes the need for scholarly research on women auxiliaries. While women combatants make critical contributions to rebellion, female 204 auxiliaries also substantively affect wartime dynamics. Moreover, I suggest that civilians may respond differently to women militants in different roles, and consequently research on legitimacy and support must take these complex gender processes into consideration. Braithwaite and Ruiz’s newly published study finds that groups with women combatants are more likely to win civil conflicts than those that do not include female front-line fighters, but that this is only true when women are voluntarily recruited.521 Groups relying on forced recruitment are more likely to be defeated by the government. The authors do not theorize deeply as to why this might be, but this finding lends even more credence to the argument that legitimacy and visibility, instead of sheer capability, play crucial roles in explaining how women’s participation affects conflict dynamics. Second, my findings indicate that research on women’s political violence must look beyond their tactical dexterity and interrogate how gendered expectations structure observer perceptions of rebellion. More attention should be paid to the strategic effects of female militancy on civilian support for rebel grievances. Third, I identify a surprising lack of research into rebel visuals and the use of propaganda in internal conflicts. As insurgent outreach shifts onto social media, rebel images now circulate quickly and widely. Examining the role of posters and other rebel visuals helps identify the mechanisms of this propaganda. Exploring securitization in these images can help elucidate the future of social media recruitment and proselytizing. From a policy perspective, these findings suggest that governments and peace-building practitioners should take women seriously as militants. Civil war, rebellion, and terrorism compose the bulk of politically violent threats to the human security. As a result, policy-makers and security studies scholars have produced a breadth of research on insurgent behavior and recruitment. This research emphasizes the importance of civilian support in sustaining rebellion and contributing to rebel success, and an insurgency’s membership – who is bidding for civilian support – should be of paramount importance in understanding what makes rebels attractive. Further, my results suggest that DDR, peace-building initiatives, and other post-conflict programming should adopt a wide lens for assessing women’s participation in conflict. Paying specific policy attention to women’s contributions off the front-line will likely improve the robustness of post-conflict prosperity by helping participants integrate back into their daily lives. Finally, this project indicates that women’s participation in rebel organizations may substantively improve the possibility of durable peace. Toft suggests that rebel 521 Alex Braithwaite and Luna B. Ruiz, 2018. “Female Combatants, Forced Recruitment, and Civil Conflict Outcomes,” Research & Politics, OnlineFirst. 205 victory in civil war results in increased democratization and a decline in conflict resurgence.522 If women-inclusive rebels are more likely to be victors than experience government defeat, than women’s participation in all roles may contribute to long-term stability. 522 Toft, “Ending Civil Wars.” 206 Appendix Chapter 4 Distribution of republican images by producer 207 Chapter 5 Insurgencies in the Rebel Women Data Project Afghanistan Hezb-e Wahdat-e Islami-yi Afghanistan Afghanistan Hizb-i Demokratik-i Khalq-i Afghanistan Afghanistan Mahaz-i Milli-yi Islami-yi Afghanistan Afghanistan Taleban Algeria Armed Islamic Salvation Front Algeria Salafist Group for Preaching and Combat Algeria, Mali Movement for Oneness and Jihad in West Africa Argentina Montoneros Argentina People's Revolutionary Army Bangladesh Gonobahini Bangladesh Jagrata Muslim Janata Bangladesh Bangladesh, Pakistan Purba Banglar Communist Party Bangladesh, India Bru National Liberation Front Bolivia Tupac Katari Guerrilla Army Brazil Araguaia Guerrillas Burundi Palipehutu- Forces for National Liberation Cambodia Kampuchean United Front for National Salvation Cameroon, Nigeria Bakassi Movement for Self-Determination Central African Republic Convention of Patriots for Justice and Peace Chile Manuel Rodriguez Patriotic Front Chile Revolutionary Left Movement China, Afghanistan East Turkestan Islamic Movement Colombia Revolutionary Armed Forces of Colombia Colombia Guevarista Revolutionary Army Colombia National Liberation Army Colombia Quintin Lame Armed Movement Djibouti Front for the Restoration of Unity and Democracy DR Congo (Zaire) Alliance of Patriots for a Free and Sovereign Congo DR Congo (Zaire) Congolese National Liberation Front DR Congo (Zaire) Congolese Rally for Democracy-ML Egypt al-Gama'a al-Islamiyya Egypt Wilayat Sinai Egypt, Israel Ansar Bayt al-Maqdis El Salvador People's Revolutionary Army Eritrea Eritrean Islamic Jihad/Reform Movement Ethiopia Afar Liberation Front Ethiopia Eritrean Liberation Front Ethiopia Ethiopian People's Revolutionary Democratic Front Ethiopia Ethiopian People's Revolutionary Party Ethiopia Western Somali Liberation Front Georgia Forest Brothers Guatemala Guatemalan National Revolutionary Unity Guatemala Guerrilla Army of the Poor Guatemala Revolutionary Organization of Armed People Honduras Chinchoneros Popular Liberation Movement Honduras Lorenzo Zelaya Revolutionary Front Honduras Morazanist Front for the Liberation of Honduras India Al Umar Mujahideen India Bodo Liberation Tiger Force 208 India Communist Party of India-Marxist-Leninist India Dima Halam Daogah - Black Widows India Hizbul-Mujahideen; Kashmiris India Kangleipak Communist Party India Kuki National Front India Maoist Communist Centre India Mizo National Front India National Socialist Council of Nagaland- IM India People's Liberation Army of Manipur India People's Liberation Front of India India People's War Group India, Pakistan Lashkar-e-Taiba Indonesia Laskar Jihad Iran Jundullah Iran Kurdistan Party of Free Life Iraq 1920 Revolution Brigades Iraq Islamic Supreme Council of Iraq Iraq Jamaat Ansar al-Sunnah Iraq KDP/Provisional Command Iraq Naqshbandi Army Iraq Reformation and Jihad Front Iraq al-Qaeda, Iraq Iraq, Jordan Jama'at al-Tawhid wal-Jihad Israel Hamas Israel Palestinian Liberation Organization Israel Popular Front for the Liberation of Palestine Ivory Coast Ivorian Popular Movement of the Great West Laos Patriotic Neutralists Lebanon Amal Movement Lebanon Lebanese National Movement Lebanon National Resistance Front/ Jammoul Lebanon, Israel, Syria Hizbollah Liberia, Sierra Leone Liberians United for Reconciliation and Democracy Liberia, Sierra Leone, Guinea United Liberation Movement for Democracy in Liberia – AK Libya Ansar al-Sharia Libya Libya Dawn Macedonia National Liberation Army Macedonia Malaysia Clandestine Communist Organization Mali 23 May 2006 Democratic Alliance for Change Mali Al-Mourabitoun Mali Ansar al-Dine Mali Arab Islamic Front of Azawad Mali Popular Liberation Front of Azawad Mali Popular Movement of Azawad Mexico Zapatista Army of National Liberation Mexico Popular Revolutionary Army Mozambique Mozambique Liberation Front Myanmar Beik Mon Army Myanmar Democratic Karen Buddhist Army Brigade 5 Myanmar Rohingya Solidarity Organization Myanmar, Thailand Mong Tai Army Nepal Communist Party of Nepal (Maoist) 209 Niger Union of the Armed Resistance Forces Nigeria Ansaru Nigeria, Chad, Cameroon, Mali Boko Haram Pakistan Lashkar-e-Islam Pakistan Balochistan Liberation Army Pakistan Ansar ul-Islam Pakistan Taleban Movement of Pakistan Peru Revolutionary Left Movement Peru Shining Path Philippines Abu Sayyaf Philippines Bangsamoro Islamic Freedom Fighters Philippines Mindanao/Muslim Independence Movement Philippines Moro Islamic Liberation Front Somalia Digil Salvation Army Somalia Hizbul Islam Somalia Ras Kamboni Brigades Somalia United Somali Congress/Somali National Alliance Somalia, Ethiopia Al Ittihad Al Islamiya South Africa Azanian People's Organization South Yemen Front for the Liberation of Occupied South Yemen South Yemen National Liberation Front - Yemen Sudan Sudan People's Liberation Army - Nasir Sudan Sudan Liberation Army- Minni Minnwai Syria Jabhat al Nusra Front Syria Free Syrian Army Syria Nour al-Din al-Zenki Movement Syria Islamic Front Syria Liwa al-Tawhid Brigade Syria People's Protection Units of the PYD Tajikistan Democratic Party of Tajikistan Tajikistan Islamic Renaissance Party of Tajikistan Tajikistan United Tajik Opposition Thailand Gerakan Mujahideen Islam Pattani Turkey Revolutionary People's Liberation Party/Front Turkey, Syria, Iraq Kurdistan Workers Party Uganda Front for National Salvation Uganda National Resistance Army Uganda Popular Resistance Army Uganda, DR Congo Alliance of Democratic Forces Uganda, Sudan Uganda National Rescue Front II Uruguay Tupamaros Uzbekistan, Tajikistan, Pakistan Islamic Movement of Uzbekistan Venezuela Bandera Roja Zimbabwe (Rhodesia) Zimbabwe African National Union Zimbabwe (Rhodesia) Zimbabwe African People's Union 210 Chapter 6 Missing rebel size data In Chapter 6, my primary analyses assess the effects of female militancy and other explanations on territorial control and includes the rebel size variable. Here, I re-assess these models using an imputed version of the rebel size variable that imputes the variable’s mean into the missing observations. One important limitation of this imputation is that missingness in this variable is likely not randomly distributed: missing data is probably most likely for small rebel organizations. Therefore, this assessment is intended not as conclusive material on its own, but as further evidence that the missing observations do not substantively change the primary model results. I additionally re-estimate my models excluding the rebel size variable. I find little change in my independent variables of interest – female militancy – across these models and only slight changes in the level of significance in other explanatory variables when rebel size is excluded.523 Model 1, Model 2, and Model 3 explore the relationship between women’s participation in combat (Model 1), auxiliary (Model 2) and leadership (Model 3) roles and territorial control using an imputed version of the rebel size variable (Figure 6.8). The only change I identify between the models here and those above that use the original rebel size variable is that the level of democracy loses its slight significance and as an explanation for territorial control, as compared with its non-imputed counterpart. Model 4 (combat), Model 5 (auxiliary), and Model 6 (leadership) also use the imputed version of the rebel size variables but consider all explanations for territorial control, not only those included in the best fitting models. Here, I find no changes in the significance or signs of any explanations for territorial control when using the original vs. imputed version of this variable (not shown). 523 In these models, I also use an imputed version (held at the mean) of the CINC state capacity measure, but there are only four missing cases of this variable. 211 212 Chapter 7 Effect of female militants on conflict outcomes