UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Trust in civil wars : the implications of conflict character and threat on political and social trust Yaylaci, Sule 2017

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2018_february_yaylaci_sule.pdf [ 3.76MB ]
Metadata
JSON: 24-1.0362112.json
JSON-LD: 24-1.0362112-ld.json
RDF/XML (Pretty): 24-1.0362112-rdf.xml
RDF/JSON: 24-1.0362112-rdf.json
Turtle: 24-1.0362112-turtle.txt
N-Triples: 24-1.0362112-rdf-ntriples.txt
Original Record: 24-1.0362112-source.json
Full Text
24-1.0362112-fulltext.txt
Citation
24-1.0362112.ris

Full Text

TRUST IN CIVIL WARS: THE IMPLICATIONS OF CONFLICT CHARACTER AND THREAT ON POLITICAL AND SOCIAL TRUST by  Sule Yaylaci  B.Sc., Summa Cum Laude, The Middle East Technical University, 2006 M.Sc., The London School of Economics and Political Science, 2007  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Political Science)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  December 2017  © Sule Yaylaci, 2017 ii  Abstract  My research investigates the repercussions of protracted civil wars on bystanders’ political and social trust. The literature is fraught with inconsistent findings on how violence impacts trust. I argue that civil wars have distinct effects on trust primarily because wartime trust formations vary by the character of the conflict (ethnic vs. ideological) and the macro historical dynamics of the country, which together shape collective threat framing. Ethnic wars should induce a higher political trust for politically represented ethnic group via the state discourse’s emphasis on collective threat, even in the presence of personal threat. In ideological wars, a similar discourse on collective threat forwarded by the state is less likely, and in the absence of a higher national threat framing, personal insecurities extending from the war should diminish people’s trust in governing political institutions. Regarding social trust, ethnic violence renders in- and out-group distinctions visible and decreases out-group trust. Alternatively, ideological violence diminishes general trust (trust in unknown others).  I deploy mixed-methods, combining case studies and cross-national quantitative data analysis. The two cases are the territorial Kurdish insurgency in Turkey (1984-) and the Maoist insurgency in Peru (1980-1992). I spent six months in each country and conducted archival work, comparative historical analysis, and numerous interviews and focus groups in 2013–2014. To see whether the theoretical predictions and empirical findings from Turkey and Peru can travel beyond their boundaries, I analyzed a pooled time-series cross sectional dataset (1981-2015), using multi-level models. As well as being one of the first qualitative studies of trust in conflict settings, my work is also original in distinguishing between the effects of different types of civil wars on trust, iii  disentangling the impact of collective and personal threat, and showing that the effects vary in the society along ethnic and political lines. My empirical findings also shed light on the generation of collective threat framing using a macro historical lens, and suggest that state-building conditions both the nature of the insurgency, national and ethnic identities, and how the conflict will be framed by the state via the official discourse.     iv  Lay Summary  Civil war, which may involve ethnic or civil conflict, is the dominant form of violence in today’s world. Violence challenges government authority, induces insecurity, and bears strong implications for individuals’ trust in government (political trust) and trust towards one another (social trust). Despite the importance of political and social trust for better functioning of institutions and societies, the effect of civil wars on trust has not been studied much. The existing studies usually focus on victims while this study focuses on bystanders—nonvictimized people—who often constitute the majority of any country undergoing conflict. This dissertation, examining the Kurdish conflict in Turkey and the Maoist conflict in Peru, finds that ethnic territorial wars and ideological wars have distinct effects on bystanders. While territorial wars increase political and ethnic out-group trust, ideological wars decrease political and generalized social trust. These findings shed light onto post-conflict challenges of re-establishing order and peace.  v  Preface  I identified and designed this research program in consultation with my supervisory committee. The research conducted this dissertation was approved by the UBC Behavioural Ethics Board, Certificate Number H12-00452, Certificate Number H12-03711 (and the renewal Certificate Number H1-03711-A001).   vi  Table of Contents  Abstract .......................................................................................................................................... ii	Lay Summary ............................................................................................................................... iv	Preface .............................................................................................................................................v	Table of Contents ......................................................................................................................... vi	List of Tables .............................................................................................................................. xiii	List of Figures ............................................................................................................................. xiv	List of Symbols ............................................................................................................................. xv	List of Abbreviations ................................................................................................................. xvi	Acknowledgements ................................................................................................................... xvii	Dedication ................................................................................................................................... xix	Chapter 1: Introduction ................................................................................................................1	1.1	 Focus, Distinctive Features and Contributions of the Dissertation .................................... 2	1.2	 Relevance and Significance of the Research ..................................................................... 5	1.3	 The Literature: Contradictory Findings and Gaps ............................................................. 6	1.4	 Research Design ................................................................................................................. 8	1.5	 Preview of the Argument ................................................................................................... 9	1.6	 Structure of the Dissertation ............................................................................................ 12	Chapter 2: Theory of Trust in Wartime ....................................................................................13	2.1	 Armed Intrastate Conflict ................................................................................................ 17	2.1.1	 Typology of Civil Wars ............................................................................................ 20	2.2	 Political Trust: Definition and Registers .......................................................................... 23	vii  2.3	 Impact of Political Violence on Political Trust ................................................................ 25	2.3.1	 Theory of Political Trust in Wartime ........................................................................ 29	2.3.1.1	 Conflict Character, Threat and Micro-Level Repercussions on Trust ............... 31	2.3.1.2	 Generation of Discourses and Collective Threat Framing ................................. 35	2.3.1.2.1	 Clash of Discourses ..................................................................................... 39	2.3.1.3	 On Personal Threat ............................................................................................ 40	2.3.1.3.1	 Determinants of Personal Threat ................................................................ 41	2.3.1.4	 Implications of Threat on Political Legitimacy ................................................. 43	2.3.1.5	 Identities, Group Influence, and Political Opinions ........................................... 44	2.3.1.5.1	 Identities in the Face of Armed Conflict ..................................................... 46	2.4	 Armed Conflict and Social Trust ..................................................................................... 47	2.4.1	 Theory of Social Trust in Wartime ........................................................................... 50	2.4.1.1	 The Role of Context ........................................................................................... 52	2.5	 The Link Between Political and Social Trust .................................................................. 54	2.6	 Observable Implications of Political and Social Trust .................................................... 55	2.7	 Conclusion ....................................................................................................................... 56	Chapter 3: Research Design ........................................................................................................57	3.1	 Case Selection: Why Turkey and Peru? ........................................................................... 60	3.2	 Focus of Analysis ............................................................................................................. 63	3.3	 Fieldwork and Data Collection ........................................................................................ 64	3.3.1	 Selecting Subnational Sites ....................................................................................... 65	3.3.2	 Data Collection ......................................................................................................... 69	3.3.2.1	 Sampling and Data ............................................................................................. 69	viii  3.3.2.2	 Conducting the Interviews and Focus Groups ................................................... 71	Chapter 4: Causal Antecedents of the Theory: Comparative Historical Dynamics behind the Onset of the Insurgencies in Turkey and Peru ...................................................................76	4.1	 Foundations of Turkey: Modernization, Nationalism And One Nation .......................... 78	4.1.1	 Kurds in a Turkish Nation-State ............................................................................... 81	4.1.2	 Suppression of Kurdish Identity and Initial Kurdish Uprisings ................................ 84	4.2	 Foundations of Peru: Colonial heritage and Nation-Building Efforts ............................. 87	4.2.1	 Brief History of Indigenous People in the Social and Political Structure ................. 94	4.3	 Conclusion ..................................................................................................................... 100	Chapter 5: Causal Mechanisms: Insurgency Dynamics, Collective Threat Configurations and Discourses ............................................................................................................................102	5.1	 Introduction .................................................................................................................... 102	5.2	 PKK and Sendero: The Characters of, the Threats Posed by, and the Military Responses to the Insurgencies .................................................................................................................. 102	5.3	 Discourses ...................................................................................................................... 106	5.4	 Conflict in Turkey .......................................................................................................... 107	5.4.1	 Pillars of the Official Discourse .............................................................................. 107	5.4.1.1	 Nation, State, Army, and Territory .................................................................. 108	5.4.1.2	 Martydom and Religious Rituals ..................................................................... 112	5.4.1.3	 Denial of the Kurdish Issue in the Official Discourse ..................................... 113	5.4.2	 The Official Discourse Embellished with Hegemonic Tools ................................. 114	5.4.3	 Media Representations ............................................................................................ 115	5.4.4	 PKK discourse ........................................................................................................ 118	ix  5.5	 The Conflict in Peru: Discourses ................................................................................... 122	5.5.1	 The Underlying Pillars of the Peruvian official discourse ...................................... 123	5.5.1.1	 The Country of Two Republics: The Inferior Position of the Indians ............. 123	5.5.1.2	 Caudillismo and the Military Culture in the Formation of Discourse ............. 124	5.5.2	 The Official Narrative ............................................................................................. 127	5.5.3	 Media Representations ............................................................................................ 127	5.5.4	 Sendero Discourse .................................................................................................. 130	5.6	 Conclusion ..................................................................................................................... 133	Chapter 6: Prelude to Empirical Chapters: Data Collection and Analysis ..........................135	6.1	 Recruitment of Participants for Interviews and Focus Groups ...................................... 135	6.2	 Conducting the Interviews and Focus Groups ............................................................... 136	6.3	 Coding ............................................................................................................................ 138	6.4	 Analysis and Inferences ................................................................................................. 140	6.5	 On Retrospective Accounts ............................................................................................ 142	Chapter 7: Hypotheses Testing: Effects of Civil Wars on Political Trust ............................145	7.1	 Configurations of Personal Threat in the Conflicts in Turkey and Peru ........................ 150	7.2	 The Kurdish Conflict in Turkey ..................................................................................... 154	7.2.1	 The Collective Threat Framing ............................................................................... 154	7.2.2	 Personal Threat for Mainstream Turkish Bystander ............................................... 159	7.2.3	 The Other Side of the Story: Ethnic Kurds, Personal and Ontological Security, Spread of Counterhegemonic Discourse, and Changing Popular Perception of the State .. 167	7.2.3.1	 Heterogeneous Baseline Priors of Kurds ......................................................... 167	x  7.2.3.2	 Personal Security Threat Posed by the State and Its Consequences on Political Trust of Kurds ................................................................................................................. 169	7.2.3.3	 The Foundations of the Counterhegemonic Discourse .................................... 174	7.2.4	 Implications of Threat on Legitimacy ..................................................................... 180	7.3	 The Maoist Conflict in Peru ........................................................................................... 184	7.3.1	 Collective Threat Framing or Rather Lack Thereof ............................................... 184	7.3.2	 Personal Threat ....................................................................................................... 190	7.3.3	 Physical and Ontological Security, and the Ethnic Bond of the Indigenous Peoples................................................................................................................................. 199	7.3.3.1	 Meaning (lessness) of Death ............................................................................ 205	7.4	 Conclusion ..................................................................................................................... 207	Chapter 8: Hypotheses-Testing: Effects of Civil Wars on Social Trust ...............................209	8.1	 Social Trust in Turkey in the face of the Kurdish Insurgency ....................................... 214	8.1.1	 A Sketch of Pre-War and Wartime Dynamics ........................................................ 214	8.1.2	 Declining Out-Group Trust in the form of Discrimination ..................................... 218	8.1.2.1	 Personal Experiences in Wartime as an Aggravating Factor Diminishing Out-Group Trust ..................................................................................................................... 224	8.1.3	 Contextual Effects ................................................................................................... 224	8.1.4	 Displacement, Politicization, and Social Trust ....................................................... 230	8.1.4.1	 In-Group Trust of Kurds .................................................................................. 231	8.2	 Social Trust in Peru in the face of the Maoist Insurgency ............................................. 233	8.2.1	 A Sketch of Pre-War Dynamics .............................................................................. 234	8.2.2	 Declining Generalized Interpersonal Trust ............................................................. 235	xi  8.2.2.1	 Displacement, Context-Dependency, and Declining Out-Group Trust ........... 239	8.2.3	 In-Group Trust of Indigenous Groups .................................................................... 244	8.3	 Conclusion ..................................................................................................................... 247	Chapter 9: Quantitative Hypothesis-Testing: Gauging the Generalizability of the Theory250	9.1	 Data and Method ............................................................................................................ 251	9.1.1	 Dependent Variables ............................................................................................... 252	9.1.2	 Independent Variables ............................................................................................ 253	9.1.2.1	 Country-level Control Variables ...................................................................... 255	9.1.2.2	 Individual-level Control Variables .................................................................. 258	9.2	 Model and Estimation Strategy ...................................................................................... 260	9.3	 Findings ......................................................................................................................... 261	9.4	 Controls and Robustness ................................................................................................ 267	9.5	 Discussion ...................................................................................................................... 267	9.6	 Conclusion ..................................................................................................................... 270	Chapter 10: Conclusion .............................................................................................................271	10.1	 Review of Findings ...................................................................................................... 272	10.2	 Contributions ............................................................................................................... 276	10.2.1	 Implications for Democracy ................................................................................. 280	10.3	 Limitations of the Study............................................................................................... 281	10.3.1	 Future Research .................................................................................................... 284	References ...................................................................................................................................286	Appendices ..................................................................................................................................307	Appendix A Extra Material for Chapter 1 .............................................................................. 307	xii  Appendix B Extra Material for Chapter 3 ............................................................................... 309	Appendix C Extra Material for Chapter 5 ............................................................................... 314	Appendix D Extra Material for Chapter 6 .............................................................................. 315	D.1	 Interview Guide for Turkey ...................................................................................... 318	D.2	 Interview Guide for Peru ........................................................................................... 322	Appendix E Extra Material for Chapter 9 ............................................................................... 326	 xiii  List of Tables  Table 1. Empirical Strategies for Different Parts of the Theory ................................................... 59	Table 2. Hypotheses for Political Trust, Corresponding Observable Implications, and Empirical Evidence Employed for Testing .................................................................................................. 147	Table 3. Total Number of Deaths in the Civil Wars of Turkey and Peru ................................... 151	Table 4. Variations in Subnational Sites as Leverage for Testing the Effects of Personal vs. Collective-level Threat on Trust ................................................................................................. 155	Table 5. Variations in Subnational Sites in Peru as Leverage for Testing the Effects of Personal vs. Collective-Level Threat on Trust .......................................................................................... 190	Table 6. Summary List of Hypotheses for Social Trust, Corresponding Observable Implications, and Empirical Evidence for Testing ........................................................................................... 211	Table 7. Distribution of Country-Year Observations on the DV and IVs .................................. 254	Table 8. Summary Statistics of the Variables ............................................................................. 259	Table 9. Civil Violence and Trust_Country-Year Cluster Models ............................................. 262	Table 10. Three-Level Random Intercept Models ...................................................................... 265	 xiv  List of Figures  Figure 1. Summary Diagram of the Theory_Ethnic Territorial Wars and Trust .......................... 15	Figure 2. Summary Diagram of the Theory_Ideological Wars and Trust .................................... 16	Figure 3. Conflict and Political Trust: Crude Mechanisms .......................................................... 29	Figure 4. Selected Subnational Sites in Turkey ............................................................................ 67	Figure 5. Selected Subnational Sites in Peru ................................................................................ 68	Figure 6. Distribution of Terror and Presidential Popularity in Peru (1985-1993) .................... 197	Figure 7. Attitudes of Kurds and Turks Towards Each Other .................................................... 223	Figure 8. Summary Diagram of the Expected Average Effects in the Theory ........................... 251	Figure 9. Predicted Probabilities of Political Trust for Different Values of Ethnic and Ideological Violence ...................................................................................................................................... 266	Figure 10. Predicted Probabilities of Interpersonal Trust for Different Values of Ethnic and Ideological Violence ................................................................................................................... 267	 xv  List of Symbols  Interview Citation: #: Sign preceding the number of in-depth individual interview E#: Sign preceding the number of expert interview R: Respondent (usually in an interview) P: Participant of a focus group  xvi  List of Abbreviations  CPO: Causal-Process Observations  EPR: Ethnic Power Relations EVS: European Values Survey GDP: Gross Domestic Product IEP: Institute of Peruvian Studies (Instituto de Estudios Peruanos)  JITEM: Gendarmerie Intelligence and Counter-Terrorism Organization (Jandarma İstihbarat ve   Terörle Mücadele Teşkilatı) MHP: Nationalist Movement Party (Milliyetçi Hareket Partisi) MRTA: Túpac Amaru Revolutionary Movement (Movimiento Revolucionario Túpac Amaru) PKK: Kurdistan Workers’ Party (Partiya Karkêren Kurdistanê) PCP: Communist Party of Peru (Partido Comunista del Perú) PUCP: Pontificial Catholic University of Peru (Pontificia Universidad Católica del Perú) Sendero: Shining Path—Sendero Luminoso del Partido Comunista del Perú TRC: Truth and Reconciliation Commission in Peru (Comisión de la Verdad y Reconciliación). WVS: World Values Survey WWI: The World War I     xvii  Acknowledgements This dissertation is a product of a long research process which would not have been possible without the incredible amount of support and help I received from the faculty members at UBC, colleagues, friends and their networks in Turkey and Peru.  First, I would like to thank my supervisor who has been instrumental in every step of this research for encouraging me to think big and never putting any restraint on the research I wanted to conduct. I owe much gratitude to him for teaching me everything I needed to know related to academic life, from how to handle quantitative data to how to navigate scholarly circles. In every struggle I went through—academic, emotional, health-related or financial—he has been extremely supportive, understanding and helpful. I could not have asked for a better supervisor, and what I produced would not have been possible without him.  My committee member and mentor Max Cameron not only taught me the big questions of comparative politics but also inspired me to conduct research on Peru. From writing the grant proposal for fieldwork funding, to introducing me to Peru for the first time and providing me with contacts there, he was a crucial figure in the building of my dissertation. Thank you, Max, for your belief in me and your endless support. I owe thanks to Alan Jacobs on many fronts. I started off lucky in my academic journey, having the chance to learn qualitative methods from him. As a committee member, he offered much help with big methodological issues. I would like to acknowledge the input he provided in all the design-related aspects of my work. Also, Fred Cutler, Paul Quirk, Antje Ellermann and Brian Job, thanks for all the opportunities you have provided me with. My doctorate has proved much more rewarding thanks to you.  I conducted twelve months of fieldwork in Peru and Turkey, which make up the backbone of my dissertation. The fieldwork was possible through the financial support from the Ibn Battuta Travel Award, the Bottom Billion Fieldwork Funding from the Liu Institute, International Development Research Center’s Doctoral Award, and Fieldwork Funding from the UBC Department of Political Science in the University of British Columbia.    In Peru, the Institute of Peruvian Studies (IEP) and the Pontifical Catholic University of Peru (PUCP) were instrumental to my project’s success. Sandro Macassi at the Institute for Public Opinion in PUCP, and David Sulmont in particular provided me with invaluable sources and key contacts all across Peru. Macassi also helped me with recruitment and organization of focus groups in Lima. Patricia Ames and Moisés Palomino Medina at the IEP provided immense support in the completion of my fieldwork. I am also very grateful to Stefany Morales Rojas, Andrea Vásquez, and Lizett Vásquez for not only helping with my research in Peru, but also embracing me as a family member and offering me a home in Peru. Nelly Vásquez made the hardest part of my fieldwork in Ayacucho so much easier, and became my Peruvian mom. My research assistant Jimena Vargas helped in every stage of my fieldwork in Lima, transcribing long hours of conversations, finding interview participants, and doing content analysis of newspapers. Eddie Vásquez carried out further content analysis after I left, helping me complete my research. My fieldwork would not have been possible without this army of support.  xviii  In Turkey, my deepest gratitude goes to my claimed sister Tuğçe Söğüt, my claimed twin Melis Oğuz, and my family who were not only my main sources of moral support throughout my dissertation but also did much of the heavy lifting during my fieldwork, helping me recruit participants and arranging the logistics. Gökhan Güler, Çiçek Şahin and Zeki Sarıgil generously put me in touch with experts all across Turkey, facilitating my data collection. My mom Oya Yaylacı travelled to the conflict zones with me and provided moral and nutritional support.    I am very thankful to the UBC-Yale Fox Fellowship for giving me the opportunity to spend a year at Yale University, take advantage of the amazing resources, meet brilliant fellows from all over the world, and focus and write my dissertation without distractions. Thank you Jason Lyall for providing me with an office space at the Political Violence FieldLab, and Prakhar Sharma for attending my every need and being a wonderful friend.   The writing stage of my dissertation was almost equally as long and arduous as the data collection stage. I owe many thanks to a number of brilliant women for reading through my drafts, giving me substantive and editorial feedback. First and foremost, Robin Studniberg. As a professional editor and a roomie-turned-into-a-best-friend, I cannot thank you enough for all your help with my professional career. You are one of the most generous, smart and wonderful people I know. Thanks for being so awesome! Pam Carlquist, my new mom, offered her hand every time I needed one. With her spiritual power, huge heart, and spectacular talent in the English language, she has provided so much input to my work and my happiness, and I am eternally grateful. Afsoun Afsahi, Edana Beauvais, and Lily Ivanova, not only inspired me with their beautiful minds and souls but also spoiled me with their friendship and generosity. They read many versions of chapters, spent hours giving me most useful feedback and correcting my errors. I am so fortunate to be surrounded by brilliant and benevolent women like you.  My friends who have been by my side through many years of this journey, I would not be able to produce this dissertation without your positive energy, love and support. Pınar Gürleyen and Serbülent Turan have always opened the door of their house and their heart whenever I needed. Esin Gözükara, my biggest fan, you are my rock! From the very first day we met ten years ago, you have never stopped filling me with confidence. Laura Nimmon, you are my idol and soul sister! Thanks for your beautiful presence in my life. Andrea Nuesser, I am grateful to the methods camps for bringing us closer. You not only expanded my culinary culture but taught me a lot about how to cherish life. Rumana Monzur, you are the reason that I believe nothing is impossible! Hilal Özçetin, I still feel your love and soul in my life. Your absence is so palpable, and I miss you a lot. Gizem Takıl, Türkay Sipahi, and Düriye Canbaz, Anastasia Shesterinina, Pascale Massot, Yasemin Ipek Can, my life-long friends, who never ceased to extend their hand and heart to me. From afar you always listened to my struggles and counselled me. I am most lucky to have you! And all my scholar friends— Katrina Chapelas, Carla Winston, and many others that I could not list here— thank you for enhancing my intellectual journey.  Finally, my soulmate, my better half, and my life partner Will. You are the best gift life has offered to me. Your unconditional love, heartfelt care, and patience made this journey all the more exciting and pleasant. Thanks for all your help with the many mathematical inquiries I had and bringing sunshine to my life everyday. I dedicate this dissertation to you. xix  Dedication         To Will, my family, and all the women who inspired me 1  Chapter 1: Introduction “War has an immediate effect upon the attitude of mind of everyone who is brought into connection with it. The end to be attained supersedes the motives of ordinary life.” Abraham Lowell.1  On November 19th, 2011, yet another clash between soldiers and Kurdish PKK insurgents dominated the headlines of Turkey’s mainstream papers: “26 martyrs and 22 injured soldiers in an attack by about 200 terrorists” screamed one.2 As always, state funerals were organized for the deceased soldiers, and attendees (most of them strangers to the victims) chanted “martyrs do not die; motherland does not divide.” Young Turkish men with flags flooded Kurdish neighbourhoods in some provinces, marching through the streets in an attempt to intimidate the residents.   In Peru, in contrast, where the war was between Maoist insurgents and the Peruvian state, neither were soldier fatalities heeded as much nor the official funerals generated a similar kind of rallying effect on general public. These are snapshots of the story that motivates my research questions. Why do people rally around a state that has failed to provide security in some contexts? Are the rallying reactions exclusive to secessionist wars or would we observe similar dynamics in revolutionary guerilla wars? More broadly, how do civil wars affect the attitudes of individuals, especially those who are not victimized (whom I call bystanders)? 3 These fundamental questions are often                                                 1 Lowell (1923, 222). 2 Retrieved on March 21, 2017 from http://www.sabah.com.tr/gundem/2011/10/19/hakkaride-hain-saldiri-21-sehit 3 Civil wars are defined as militarized disputes between two or more parties within a bounded sovereign territory. 2  overlooked by scholars of civil wars and political behaviour, despite their indisputable relevance to the respective literatures, as well as to the policy implications for reconciliation processes.    1.1 Focus, Distinctive Features and Contributions of the Dissertation My dissertation seeks to examine the impact of protracted civil wars, or intrastate armed conflicts, on trust. On four grounds my dissertation is distinctive: 1) My focus is on bystanders (non-victimized citizens) as the main group of interest instead of victims, based on the idea that bystanders are the main audience for the theatre of conflict, and it is their trust that dictates the popular trust levels. 2) I distinguish between political and social trust, as trust in institutions and individuals are qualitatively different, notwithstanding their possibly mutual effect on each other. 3) Similarly, I distinguish between two types of civil wars, ethnic territorial and ideological/ revolutionary types civil wars. 4) This is a comparative study employing a mixed-method design. The strategy combines intensive case-study analysis with statistical work where large-N quantitative study tests for the generalizability of the argument.  Intrastate political violence, as with all forms of violence, certainly speaks to more than the immediate victims;4 it targets the whole population, instilling seeds of insecurity, challenging the authority of government, and shaking the grounds of trust. In fact, victims have very little intrinsic value for the purposes of political violence (Crenshaw 2002). Nonetheless, research on violence tends to focus on the victims (e.g. Becchetti, Conzo, and Romeo 2014; Blattman 2009; Colletta and Cullen 2000). Bystanders, who surely have a large impact on the political culture of the country, have not been systematically studied in conflict settings. My research extends the focus of such explorations to a broader level by including both direct targets of violence and                                                 4According to White their deaths and injuries are important for the symbolic expression of danger to random and ordinary citizens (2002). 3  bystanders, focusing primarily on the latter. I define victims as people who suffered physical pain personally, who lost their first degree family members to war, who had to relocate due to an imminent threat to life of a family member, and who faced serious financial challenges due to physical damages to property during an armed conflict. And I define bystanders as all the non-victimized citizens residing in a particular country at the time of a violent act who either witness the act in person, or become cognizant of it through media or social environment. 5 For the most part, bystanders constitute the majority of any democracy’s population. Secondly, trust as a dependent variable is dissected into two separate, though interlinked, forms: political trust, referring to trust in political actors, institutions of the state as well as state apparatus, and social trust, referring to trust for other individuals in the society. Based on the premise that different types of conflict may affect forms of trust differently, this dissertation examines how conflict character impinges on political and social trust, and why the repercussions on trust are different. Thirdly, research shows that ethnic and nonethnic ideological civil wars have different causes (Sambanis 2001). I argue that when there is causal heterogeneity, we should expect consequential heterogeneity. The war’s attributes – including the motives for conflict and social cleavages – may impact citizens’ attitudes. Building on that and deducing from extant distinctions between ethnic and ideological types of conflict in the literature, I propose that the consequences of civil wars are, in large part, contingent on the character of conflict. I theorize how conflict character may be instrumental in defining the nature of threat by specifying the                                                 5 This definition draws on the approach of Staub (2003) and Vetlesen (2000). Vetlesen (2000) also distinguishes between internal i.e. within defined boundaries of a country and external bystanders i.e. outside the borders of a country where the violent act occurred. Internal bystanders can be individuals, organizations, institutions or the state itself while external bystanders, in addition to these, include countries. I only focus on internal bystanders. 4  necessary social and political processes, and how conflict character may have different consequences on political trust.  This distinction between ethnic territorial and ideological wars does not refute the possibility that ideological wars have an ethnic dimension or vice versa. Indeed, most civil wars have crosscutting ethnic and class dimensions. What may appear as an ethno-nationalist conflict may, at its roots, be a manifestation of social and economic grievances. Take the example of the Maoist insurgency in Peru: The insurgency featured itself as a Maoist insurgency, and in most datasets that involve a typology the civil war in Peru is coded as a non-ethnic, ideological, revolutionary war. However, the insurgent discourse was tied to and intended to speak to the grievances of the indigenous groups. The leadership mobilized and recruited predominantly from the indigenous people, and it waged its war, for the most part, in the indigenous-populated areas of Peru.6 Hence, the adjectives “ethnic” and “ideological” assigned in the names to two separate kinds of conflicts does not mean that they do not have any other dimension. The names are intended to reflect the main motivation of the insurgency and the master cleavages in the conflict. My analysis considers the identity dimension and explains how it played out in both types of conflict. Finally, this study combines the strengths of qualitative and quantitative methods to develop and test a new theory on trust formation in civil war. To the best of my knowledge, this is the first study combining in-depth qualitative research on conflict and trust and the largest scale of quantitative data available to test the generalizability of the argument.                                                 6 In the same vein, the conflict between The National Union for the Total Independence of Angola (UNITA) and the government of Angola (dominated by the Popular Movement for the Liberation of Angola (MPLA)) was at first sight an ideological conflict. The UNITA branded itself as an anti-communist movement, and did not officially represent any ethnic group. Yet, UNITA was comprised primarily of Ovimbundu group members, while the government was comprised primarily of Mbundu group members (Collelo 1991). 5  1.2 Relevance and Significance of the Research Civil wars, defined as militarized disputes between two or more parties within a bounded sovereign territory, have afflicted states since the first formation of a nation-state. Virtually all nations have experienced a form of civil war (see Skocpol 1979; Tilly 1985). Even the most peaceful and established countries in the modern world have gone through episodes of internal conflict (e.g. the American Civil War [1861-1865], the Boshin War in Japan [1868-1869], and the Sonderburg War in Switzerland [1847]). In today’s world, civil wars are so common that they have supplanted interstate wars as the dominant form of political violence (Human Security Report Project 2013). As of 2017, there are 17 ongoing civil wars, including the protracted Kurdish-Turkish conflict and rather new Syrian Civil War, consuming democratic and nondemocratic regimes alike.7  Wars shape politics and society. Civil wars puncture the political and social order, realign the society along new lines of identity, and carry broad social and political implications for the aftermath. Anthony Smith contends that “‘societies’ — that is to say, given ethnic or other communities — owe much of their form and solidarity to the exigencies of war” (1981, 377). Not only the regimes, institutions, formal and informal structures, but also micro-level values, beliefs and attitudes are subject to change in the face of such transforming experiences with violence. The question is how these beliefs and attitudes change in the face of violence?  It is important to understand the repercussions of political violence on attitudes and values because they loom large for legitimate survival, especially of a democratic regime, given                                                 7 Some notable examples of civil wars in democracies are: PKK: Partiya Karkêren Kurdistanê (Kurdistan Workers’ Party) in Turkey; IRA: Irish Republican Army in the UK; LTTT: Liberation Tigers of Tamil Eelam in Sri Lanka; PCP-SL: Sendero Luminoso Partido Comunista del Perú (Communist Party of Peru) also known as Shining Path in Peru; ULFA: United Liberation Front of Asom (Asamese Separatist Movement) in India.                                                                          6  the importance of civic culture as a precondition for stable and functioning democratic institutions (e.g. Almond and Verba 1963; Diamond 1993; Inglehart and Welzel 2003; Putnam, Leonardi, and Nanetti 1993). The social and political consequences of a civil war then could have significant implications for long-term stability and societal peace in a country even after the war has long ended (Dyrstad 2012).  1.3 The Literature: Contradictory Findings and Gaps Although the causes of civil wars receive much scholarly attention in both international relations and comparative politics literature, study of the consequences of civil wars is a relatively incipient though burgeoning field. The economic damages of civil wars are easier to quantify, and hence we know more about such effects.8 Political and social implications are, however, harder to gauge. Wars permeate all layers of life—politics, family, individual and collective psychology, memory, attitudes, behaviours, mental health— as well as culture and institutions. Despite these challenges, there is a growing literature on social and political consequences of conflict, albeit fraught with contradictory findings and predominantly focused on victims.  Many empirical findings show that wartime trauma is found to be associated with lower generalized trust in others (Kunovich and Hodson 1999; Rohner, Thoenig, and Zilibotti 2013), and that war-torn countries suffer from distrust (e.g. Colletta and Cullen 2000; Collier 2003; Widner 2004). Other studies, in contrast show that violence of higher intensity, counter intuitively perhaps, leads to greater levels of generalized trust, explained by post-traumatic growth theory—a theory positing that positive changes may follow trauma (e.g. Blattman 2009;                                                 8 Economic effects are often associated with growth regressions by hampering savings, investment, and income (e.g. Alesina et al. 1996, Barro 1991). For other economic effects, see (Arunatilake, Jayasuriya, and Kelegama 2001, Collier 1999, Grobar and Gnanaselvam 1993). Some suggest that negative socioeconomic legacies spill over to the post-conflict years (Blattman and Annan 2010, Bundervoet, Verwimp, and Akresh 2009); while others contend that wars do not leave lasting impacts on household socioeconomic status measures (Bellows and Miguel 2009). 7  Gilligan, Pasquale, and Samii 2014). Another study suggests that the effects may be contingent on the type of the war. Distinguishing between different identity wars, the authors argue that civil wars reduce trust though religious wars more so than ethnic wars (Traunmüller, Born, and Freitag 2015). The same study also suggests that generalized trust is affected more than particularized trust (trust in family) except for trust within the neighbourhood (Traunmüller, Born, and Freitag 2015). Yet other works fail to find a general pattern between wartime trauma and social trust (Shewfelt 2009). With respect to implications of civil wars on political trust, findings are also mixed. A number of studies find a negative effect. Hutchison and Johnson (2011), in their aggregate cross-national study on 16 African countries, find internal political violence to be deleterious to political trust (measured as generalized trust in government). Similarly, attacks by Tuaregs against the Malian state had detrimental effects on people’s trust in the national president (Gates and Justesen 2013). More evidence of a negative effect comes from the Nepalese Civil War in the immediate aftermath of a ceasefire, where exposure to violence has been found to reduce trust in the national government (De Juan and Pierskalla 2016). Other scholars argue that the effect of violence on political trust is contingent on the ideology of the governing party, and that voters evaluate hard- and soft-line governments differently (Arce 2003; Chowanietz 2010).  Most of these studies focus on single cases or short-run impacts, and many contradictory findings exist, pointing to the importance of contextual effects. Cross-national studies are helpful to see a broader pattern, and they show divergent effects of war on the society (Dyrstad 2012) but fail to capture causal mechanisms. We can infer from these studies that popular reactions to civil war violence could be heterogeneous. My dissertation draws on this inference, seeking to explain the reasons behind heterogeneous effects, and thus represents an effort to give a more 8  nuanced interpretation of the impacts of conflict on wartime transformations of trust in the society.   1.4 Research Design My design is mixed-method where I combine rich qualitative case studies and individual-level data with cross-national time-series quantitative data analysis. Guided by my theory that argues ethnic territorial and ideological revolutionary wars have different consequences on trust, I selected the Kurdish insurgency in Turkey as a case of ethnic war (fought between Partiya Karkêren Kurdistanê [Kurdistan Workers’ Party—the PKK hereafter] and the Turkish state since 1984), and the Maoist insurgency in Peru as a case of ideological war (fought between the Shining Path (Sendero Luminoso del Partido Communista del Perú — Sendero hereafter) and the Peruvian state (1980-2000). Turkey and Peru, despite their diverse historical origins—Turkey being a post-imperial country descending from the Ottoman Empire and Peru being a postcolonial country— they still host enough equivalence that renders their contemporary legacies similar, which helped the choice of cases. Both countries suffered from organized intrastate political violence for over two decades and yet democracy survived in both countries. Both Turkey and Peru experienced elite-driven democratic revolutions and multiple military interventions, and both countries share institutional features, yet differ in many others. Turkey and Peru are both multi-ethnic, low trust societies, especially in terms of social trust (they are two of the lowest ranking countries according to World Values Survey data (See Figure A 1 and Figure A 2 in Appendix A). For the purposes of this dissertation, the most preeminent feature of these two countries is that they both underwent long-running violent insurgencies, which were culminations of the preexisting social issues and horizontal inequalities across groups. In Turkey, violence is mainly ethnic territorial in nature; in Peru, the conflict was mainly ideological and 9  revolutionary, notwithstanding the fact that ethnic and ideological cleavages existed in both. I spent six months in each country (in 2013-2014) doing semi-ethnographic fieldwork. I conducted gathered in-depth qualitative data through numerous interviews and focus groups from cities varying in their distance to battlefields from both cases. I focus on the period between 1980 and 2000. 1980 was the year Shining Path commenced its armed operations, and the PKK was already active and preparing its base for an organized guerilla war.  To test the major hypotheses of the theory quantitatively and to see whether the theoretical predictions and empirical findings from Turkey and Peru can travel beyond their boundaries, I analyzed a pooled time-series cross sectional dataset (1981-2015), using multi-level models. 1.5 Preview of the Argument I argue that protracted civil wars change both the political and social trust of individuals, and conflict character (i.e. whether the conflict is ethnic territorial or ideological revolutionary in character) is a primary factor affecting the way conflict impinges on trust. Political trust is likely to increase on average as a result of ethnic territorial wars while ideological wars lead to decrease wartime political trust. This difference is because of the collective threat framing of the state discourse in two types of conflicts. Ethnic territorial conflicts usually threaten the identity of the nation as the territorial integrity is at stake, which is a defining feature of a nation-state, while ideological conflicts, often targeting the government, only pose a kindred collective threat when they emerge in a country that identifies itself with a particular governmental regime. In the case of ethnic territorial conflicts, the state produces a discourse of existential threat (threat against the national unity, territorial integrity and nationhood) and capitalizes on agitation to galvanize people to support itself.  Through this discourse, the insurgents who are perceived to 10  be posing the collective threat (and personal threat) become enemies, and the state, its institutions and actors are glorified in their eyes. In ideological wars, a similar kind of collective threat framing is less likely, and personal security threats usually guide the changes in people’s trust. For social trust, I distinguish between generalized interpersonal trust (trust in unknown others), and category/identity-based trust. I argue that ethnic violence should induce category-based social trust, wherein in and out-group distinctions become clear and out-group trust dwindles; while ideological violence should diminish general trust because identity markers are not as clear.  The effect of political violence on trust is not homogenous across the society. The aforementioned general arguments, especially those of political trust, apply to the politically represented groups. Politically underrepresented groups, who are also more likely to be victimized, undergo different experiences. Identity is quite essential to how the war affects people’s attitudes, and ethnic territorial wars and ideological wars are quite different in terms of salience of primordial identities.  In the case of Turkey’s Kurdish conflict, I find that political trust increased after the onset of the conflict for bystanders who identify themselves with the nation because of the collective threat framing. In the face of an existential threat, the state incorporated the notion of the non-negotiability of borders into public discourse. Nationalist tools were at the disposal of the state, being well ingrained in the nation-building of the state. The Turkish state propagated the idea that it was doing its best to counter the insurgency, and the trust in the army and by association the state and the government skyrocketed. For the ethnic Kurds who developed sympathy for the PKK over the course of the conflict, trust in the Turkish political institutions and actors decline 11  precipitously. Ethnic bonds among co-ethnics enabled cultivation of collective identity and spread of counterhegemonic discourse of the insurgency and generated heightened salience of ethnic identity. In Peru, due to a lack of collective threat framing, the conflict did not have much effect on the bystanders as long as it did not pose a tangible personal threat despite the more widespread civilian targeting during the war. For those who felt personal threat, political trust declined because there was not a strong collective threat framing to offset the negative effect of personal threat on political trust. The relative failure of the Peruvian state in nation-building and the ideological character of the war explain the lower magnitude and effect of collective threat. Welfare concerns were serious at the same time in Peru due to the heavy economic crisis the country was undergoing, and the economic evaluations proved to be a stronger factor in remote bystanders’ trust calculations. Even though the state’s relative weakness vis-à-vis Sendero was glaring at times, which may have diminished the public’s perception of economic performance, economic threat was more real and immediate. Victimized indigenous people lost their trust in the state as both their physical and ontological security was endangered, and they did not feel any promise or intention of protection by the state. However, unlike Kurds, they did not have a strong ethnic bond or collective consciousness. The sufferings of Kurds in the Eastern provinces due to state violence affected the attitudes of their unharmed co-ethnics. Kurds also had a new alternative power to turn to who claimed to represent their interest: the PKK, while in Peru Shining Path never branded itself as a representative of indigenous groups.  This dissertation displays the nuances in political and societal consequences of civil wars. The conflict character, guerrilla warfare, and the type of threat, and identities are important in explaining the variations. I suggest my theory and findings are generalizable particularly to 12  protracted ethnic-territorial wars and ideological-governmental wars that emerge in countries where one group dominates over the other(s).  This dissertation also suggests that the structure of civil conflicts is itself endogenous to features of state formation. Grievances that led to the onset of insurgencies extend from nation-building (See Chapter 4 for inductive comparative historical case studies) also shapes the discursive framework employed by the state to respond to the conflict. Ideology and strategies of nation-building mould societal forces, political groups, and shape the political landscape of the country for many years to come. Both the character of the conflict, the framing of the threat by the state, and the consequences of the war are embedded in the dynamics of nation-building. 1.6 Structure of the Dissertation The dissertation is organized as follows. Chapter 2 provides the reader with clarification of the essential concepts and the logic of the theory. Chapter 3 describes the research design used to examine the effects of civil war on trust. Chapter 4 presents nation-building in Turkey and Peru and dynamics of state-society relations as a causal antecedent to the insurgencies. The goal of the chapter is to clarify a potential causal confusion engendered by the fact that the pattern of any given civil wars is not “as-if random.” Conflict character as the independent variable is embedded in the nation-building and history of state-society relations. Chapter 5 discusses the insurgency dynamics, configurations of threat and discourses. Chapter 6 describes the data collection and analysis. Chapter 7 and 8 are hypothesis-testing chapters for political and social trust respectively. Both chapters use in-depth interviews and qualitative evidence to describe the implications of war on individual’s trust registers. Chapter 9 tests major hypotheses of the theory using pooled time series cross sectional data. Chapter 10 concludes by summarizing the dissertation’s contributions and providing a guideline for future research.  13  Chapter 2: Theory of Trust in Wartime This chapter theorizes the links between civil war and trust. My theory distinguishes between the kinds of civil wars and types of trust, suggesting that ethnic territorial wars and ideological revolutionary type of civil wars have distinct effects on political and social trust.9 Ethnic territorial wars on average increase political trust, particularly for the politically represented ethnic group, while ideological wars decrease political trust. For social trust, I look at generalized interpersonal trust (trust in unknown others), and category-based/identity-based trust and expect that effects of wars on the two types of social trust will differ. My theory suggests that ethnic violence should induce category-based social trust, wherein in and out-group distinctions become clear and out-group trust dwindles; while ideological violence should diminish general trust because identity markers are not as clear (see Figure 1 and 2).  I presume that the difference we observe in the trust implications of civil wars stems from collective threat framing and that we should disentangle the effects of collective and personal threat, as shown in Figure 1 and 2). Collective identity threat refers to a situation where the motivations of the conflict challenge the identity of the state/nationhood, as in the case of ethic territorial wars. Ethnic territorial wars attack the territorial integrity of a nation and existence of the politically powerful ethnic collective. States under attack, in response, frame the threat as a collective threat to nation via discourse, and rally the politically represented ethnic group around its symbols and leaders. Through this discourse, the insurgents who are perceived to be posing the collective threat (and personal threat) become enemies, and the state, its institutions and                                                 9 Inasmuch as I consider the possibility that social and political trust may be empirically correlated (see section 2.5), my view is that political and social trust are theoretically distinct. Being so, institutional trust may be fundamental for legitimacy while interpersonal trust is necessary for a better functioning society.  14  actors are glorified in their eyes, which should increase political trust. Even when faced with a personal threat, because of the overarching collective threat people feel to the nation, their allegiance to the state should not swerve, and collective threat framing of the conflict should moderate the negative effect of the personal security threats on political trust. The meaning attributed to the nation via state discourse should supersede the value of human life. Politically weak ethnic groups should not subscribe to the discourse emphasizing collective threat for the nation and thus be more sympathetic to the counterhegemonic discourse of the insurgency, which in turn should lead to plummeting political trust.  To the extent that ethnic territorial conflict is geographically confined and the perpetrators easy to identify, it should not affect generalized interpersonal trust. However, both through the polarizing effect of state discourse emphasizing ethnic divisions, and through the personal security threats groups feel from influx of internally displaced populations, category-based out-group trust should diminish.  Ideological wars attack the government and the regime. The institutions and the actors are often targeted, crippling the state. Because ideological conflicts often do not feature a strong collective identity threat (as the threat is to the government, not to a defining feature of the nation), collective threat framing will be weak and personal security threat should determine trust. Threat to individual securities cast shadow on the capacity of the state and the government to protect and it decreases trust. Because the perpetrators are hard to distinguish, generalized interpersonal trust is undermined. Identity and context-contingencies, which are subject to change in the course of the war, further condition the individual responses to these threats. The theory explains the mechanisms connecting armed conflict to each type of trust and elaborates on the nuanced effects of conflict on politically underrepresented groups.  15  Forms of               Mechanisms                                    Forms of Trust Civil Violence                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          Figure 1. Summary Diagram of the Theory_Ethnic Territorial Wars and Trust Note: The main mechanism between ethnic territorial wars and increasing political trust is the collective identity threat framing, whereas the main mechanism between ethnic territorial wars and decreasing out-group trust is the personal security threat. Yet, it is essential to emphasize that personal security threat is never independent of the overarching collective threat. They operate together to shape the perceptions of out-group members. The black dotted line extending from ethnic territorial wars means that the impact of these wars on personal securities will be less as compared to their impact on collective level threat. Out-group trust on the bottom right indicates that the category of social trust ethnic territorial wars have a significant effect on is identity-based out-group trust.    Collective Identity Threat (Framing)  Personal Security Threat Out-Group  Trust Ethnic Territorial Political Trust 16  Forms of               Mechanisms                                    Forms of Trust Civil Violence                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                Figure 2. Summary Diagram of the Theory_Ideological Wars and Trust Note: The main mechanism between nonethnic ideological wars and decreasing political trust and decreasing generalized interpersonal trust is the personal security threat. The theory presumes that because ideological wars feature less of a collective identity threat, framing of such threat will be weak. The thick line extending to generalized interpersonal trust indicates the most significant and strong effect of ideological wars will be on generalized trust over category_based trust. Out-group trust could also be affected if group distinctions are feasible.      Ideological (Revolutionary) Collective Identity Threat (Framing)  Personal Security Threat Generalized Interpersonal Trust Political Trust Out-Group  Trust 17  This chapter starts with defining armed state conflict, clarifying the distinctions between ethnic and ideological wars, and delineating the scope conditions. Then, it lays out the theory of political trust in wartime, followed by the theory of social trust in wartime. The former theory pivots on the mechanism of threat, and after explaining the main parts of the theory; the section discusses implications of armed conflict for political legitimacy. The theory of social trust in wartime distinguishes between interpersonal and category-based trust and is centered on identities and context-contingency. In the last section, reciprocity of political and social trust is discussed.   2.1 Armed Intrastate Conflict The scope of this work is protracted (more than 5 years) armed intrastate conflict/civil wars that emerged in context where one group dominates over the other(s). I define armed intrastate conflict as the intentional use of sustained violence, including armed force, between a government and a politically and militarily organized non-state entity for political goals within the boundaries of a recognized sovereign state. My definition of civil war follows Stathis Kalyvas who suggests that “when domestic political conflict takes the form of military confrontation or armed combat we speak of civil war” (2007, 416). Defined this broadly, civil wars are interchangeable with armed intrastate conflict, yet civil war definition varies substantially and is usually subject to many restrictions. Most definitions of civil war impose a fatality threshold to distinguish it from street crime or riots in terms of scale of violence.10 I do                                                 10 An oft-used definition dates to the Correlates of War Project, which requires a minimum of 1,000 battle deaths in a single year for an internal conflict to be considered as a civil war. (Singer and Small 1972). Later, scholars relaxed this requirement because many cases that have the nature of a civil war were excluded due to arbitrarily high threshold of casualties, and it caused problems in coding the duration of the war, as even though the war is on, the number of battle-deaths does not always meet the threshold. To overcome such issues, some used the 1000 battle-deaths threshold for the whole duration of the war instead of one-year (Gleditsch et al. 2002); some combined both 18  not employ any cumulative death threshold in my definition,11 as my goal is in part to theorize the effect of political violence graded by the intensity of violence.12    I refer to non-state entities fighting with a government as insurgents. Insurgents are non-state, armed factions that coercively challenge state authority, seek to redefine the political structure, and fight for sovereignty. Insurgents are often called rebels or depending on the time when they were active and the conjuncture of political order then, they may take politically charged names such as “revolutionaries,” “terrorists,” “freedom fighters” (Metelits 2010; Turk 2004). Semantics are part of the legitimacy game for political actors, and in my dissertation I use insurgents and rebels interchangeably while eschewing the use of politically charged terms. I use the terms civil war, intrastate war, armed intrastate conflict interchangeably. An event that depicts an armed intrastate conflict may also be denoted as a case of internal conflict, rebellion, or domestic terrorism, so some conceptual clarification is in order. Armed intrastate conflict is subsumed under a broader category of internal conflicts, which encompasses any form of conflict ranging from hostility or dispute to violent fight over divisible or indivisible resources                                                                                                                                                        criteria by imposing at least 1000 war-related deaths in both at least a single year of the war and overall (Sambanis 2004), and others made it a bit stricter by imposing another threshold of 100 deaths every year on average (Fearon and Laitin 2003). See Sambanis (2004) for further examination of conceptual and empirical complexities of defining civil war. 11 The fact that arms are employed and used against the fighting parties for political reasons within a territory in a sustained fashion suffices to classify a case as intrastate political violence. Death every year is not a necessary requirement. Most cases of prolonged armed conflicts generate battle deaths, and number of deaths not being in a systematic fashion does not change the fact that the conflict is ongoing. Finally, because the violence exerted needs to be sustained, anomic or one-time events are outside of the scope of this project.  12 Other attempts at categorizing cases of violence by intensity also exist: Communal violence in Africa, for instance, is divided into four types by intensity and duration of the violence and the goal. Ethnic violence: “an event of short duration [...] in which two identifiable communal groups are antagonists in violence to secure some short-term goal”; irredentism, “an event in which an identifiable communal group seeks to change its political allegiance from the government of the territorial unit in which it resides to a political system, in which the authorities share the communal identification of the irredentist group concerned”; rebellion, “an event in which an identifiable communal group seeks by violence to gain increased autonomy from the national political authorities”; and civil war, “an event in which an identifiable communal group [...] seeks by violence to form a new political system based on boundaries of ethnic community” (Morrison, Mitchell, and Paden 1989, Fearon and Laitin 1996). 19  based on ethnic, religious, class, ideology or other material and immaterial factors within the boundaries of a recognized state territory.13 The nonviolent conflict (a dispute between parties which does not involve coercion or threat) may culminate into a violent one. My work focuses on the phase where the conflict is militarized (i.e. the phase of armed conflict).14  Insurgent groups are often proscribed as terrorist.15 Conceptually, terrorism is a form of political violence with certain characteristics, most of which overlap with the definition of insurgency such as the use of purposive violence for political goals.16 The analytic distinction between terrorism and civil war then lies in distinguishing the strategy from the broader definition. I conceive of terrorism as one warfare strategy or tactic among others.17 Some cases of civil war may involve acts of terrorism, while others may not (Kalyvas 2004).18 Here I define insurgent terrorism as acts of violence perpetrated by nonstate actors for political goals, inflicted on a vulnerable group(s) to create a sense of threat, and to intimidate a wider target audience (it                                                 13 It is important to distinguish violence from conflict as studying them may require separate theoretical frameworks (Brubaker and Laitin 1998). 14 Others scholars categorize the events according to the goal, intensity or duration of the violence. One example identifies four types of communal violence in Africa varying in these parameters: ethnic violence, irredentism, rebellion and civil war (Morrison, Mitchell, and Paden 1989). 15 Indeed, many insurgent groups cited in the civil war literature such as the Kurdistan Workers’ Party (PKK) in Turkey, Shining Path (PCP-SL) in Peru, Irish Republican Army (IRA) in the UK, the National Liberation Army (ELN) or the Revolutionary Armed Forces of Colombia (FARC_EP) in Colombia, and Tamil Tigers (LTTE) in Sri Lanka are proscribed as terrorist groups by the EU and the US as well as the respective governments they are fighting with. Often it is in the interest of the state to denote a rebel group as terrorist, and sometimes international system dictates these norms. For instance, protracted secessionist conflicts, which were called ethnic conflicts in the 1990s, were renamed as terrorism following 9/11 (Murer 2012). 16 See Boyle (2012) for a thorough discussion of the study of terrorism as separate from the other kinds of political violence. 17 See Fortna (2015) for effectiveness of terrorism as a military tactic. 18 For instance, one study argues secessionist groups are more likely to employ terrorism (Pape 2005), although empirical evidence fails to support this argument (Fortna 2015). Similarly, religious conflicts, the geography of the terrain and gross domestic product per capita are argued to be linked to the likelihood of terrorism (Laitin and Shapiro 2008, Stanton 2013, Pape 2005). Jessica Stanton (2013) also shows that rebel groups challenging democratic governments are more likely to use terrorism as the sensitivity to civilian losses is expected to be higher in democracies, which increases the chances of concessions. 20  may be a national government, a rival group, a political party, civilians). The extent to which terrorist tactics are employed is relevant to personal threat, which will be explained in section 2.3.1.3.    2.1.1 Typology of Civil Wars Disaggregating civil wars, though critical, is problematic as there are many layers to the character of a conflict.19 All attempts at classification must choose select criteria to achieve workable categories, and usually the research question drives criteria selection.20 I formulate my distinctions along the lines of master-cleavage: ethnic vs. ideological wars—a common way to disaggregate civil wars. Robinson (2001), for instance, identifies two cleavages: class and ethnicity – the two major organizing principles of most conflicts – and posits that ethnic conflict is qualitatively distinct from class conflict. Along the same lines, Nicholas Sambanis (2001) produced a seminal work distinguishing identity wars (ethnic and religious) from nonidentity wars (revolutionary and others). He finds that ethnic and nonethnic wars have different causes; ethnic heterogeneity and lack of democracy, which other scholars found to be insignificant for the onset of civil war, increase the likelihood of identity wars while it does not play a significant                                                 19 See Kalyvas (2007) and Newman (2014) for a thorough discussion of civil war types and difficulties facing attempts to categorize civil wars. 20 James Fearon (2004) identifies five types of civil wars where three of them are short, stemming from anti-colonial struggles, coups, revolutions, and collapse of regimes, notably the collapse of the Soviet Union and Yugoslavia, and two of them are long, emerging from peripheral insurgencies—“civil wars involving rural guerilla bands typically operating near the state’s borders”— involving land or natural resource conflict or conflicts where the rebel group is financed through contraband (opium, diamonds or coca) (Fearon 2004, 277).  Others classify civil wars by the kind of warfare. One classical distinction is between conventional and irregular wars. In conventional wars, parties that are roughly symmetric in military power fight across defined frontlines. By contrast, irregular wars – also known as guerilla wars – are characterized by power asymmetries between the parties and nonstate combatants, and may involve active involvement of civilians (e.g. popular guerilla warfare). In most cases of civil wars these two are mixed; indeed it is quite rare to encounter employment of conventional warfare alone in civil wars – rare examples are Spanish and American Civil wars – even though it is the dominant form of warfare in interstate wars (Kalyvas 2005). 20 Nearly half of civil wars since World War II were indeed fought as irregular wars (Balcells and Kalyvas 2007). 21  role in the onset of nonethnic wars (also see Besançon 2005; Cederman, Gleditsch, and Buhaug 2013; Fearon and Laitin 2003; Licklider 1995; Reynal-Querol 2002).   I define ethnic wars, following Brubaker and Laitin (1998), as armed intrastate conflicts where the insurgents brand themselves with an ethnic identity, and where the putative ethnic difference is fundamental to the conflict. I define ideological wars as armed intrastate conflicts over class or ideology. Cleavage, however, is only one major aspect of a conflict that determines its character. What may appear as an ethno-nationalist conflict may, at its roots, be a manifestation of social and economic grievances.21 It is thus worthwhile to underscore that ethnic wars may have an ideological dimension, and non-ethnic wars may have an ethnic character. Take the example of Sendero in Peru: Sendero featured itself as a Maoist insurgency, and in most datasets that involve a typology the civil war in Peru is coded as a non-ethnic, ideological, revolutionary war. However, Sendero’s discourse was tied to the grievances of the indigenous groups even though the motive of the insurgency could not be defined around ethnic claims. The leadership mobilized and recruited from the indigenous people, and it waged its war, for the most part, in the indigenous-populated areas of Peru even though at no point it was possible to distinguish the two parties to the war by ethnic affiliation.22 That’s why we should heed Brubaker and Laitin                                                 21 Worse yet, a simplistic categorization of a conflict may have strong policy implications and lead to wrong decisions as in the case of the Bosnian conflict where the label of “ethnic” civil war discouraged intervention (Newman 2014). 22 In the same vein, the conflict between The National Union for the Total Independence of Angola (UNITA) and the government of Angola (dominated by the Popular Movement for the Liberation of Angola –MPLA) was at first sight an ideological conflict. The UNITA branded itself as an anti-communist movement, and did not officially represent any ethnic group. Yet, UNITA was comprised primarily of Ovimbundu group members, while the government was comprised primarily of Mbundu group members (Collelo 1991). Conversely, the PKK in Turkey is an ethnic insurgency although they embrace Marxist principles. The PKK movement grew out of political as well as economic and social grievances, so class, ideology and ethnicity are inextricable components of the movement. Similarly, some quintessential examples of ethnic wars – such as Burundi’s Hutu-Tutsi conflict, or former-22  (1998, 427) who advise to pay careful attention “to the forms and dynamics of ethnicization, to the many and subtle ways in which violence—and conditions, processes, activities, and narratives linked to violence—can take on ethnic hues” (see Murer 2012 for what is "ethnic" in ethnic conflicts). In my empirical analysis, I discuss the crosscutting dimensions essential to the warfare as they continue to be at play throughout conflict to eventually shape the outcomes of the war. The motive of the insurgency also taps into the conflict character. Insurgencies may seek territorial control (various levels of self-determination) or a change in the government or political system at large. The motive is so integral to the character of the war that civil wars can be disaggregated along the lines of demand as governmental and territorial (Gleditsch et al. 2002).23 For definitions, I rely on the Uppsala/PRIO Armed Conflicts project, where Governmental conflicts concern “the type of political system, the replacement of the central government, or the change of its composition while Territorial conflicts involve demands for secession or autonomy” (Gleditsch et al. 2002, 619). It is commonplace to hear that ethnic conflicts break out over territorial demands while class conflicts often seek governmental demands i.e. are more concerned with an overhaul of the system governing the country. Indeed, most separatist insurgencies are initiated by distinct ethnic minority groups (Bartusevičius 2016; Horowitz 1985). The temptation to think ethnic conflicts are tantamount to territorial conflicts, and ideological conflicts are the same as government                                                                                                                                                        Yugoslavia’s conflict –  had economic motivations (Bartusevičius 2016). 23 Uppsala/PRIO Armed Conflicts dataset shows that since the WWII virtually two-thirds of all intrastate conflicts since 1946 have been challenges to the central government, the remaining being classified as territorial disputes (Gleditsch et al. 2002).  Findings attest to the importance of this categorization: Buhaug (2006), for instance, shows that country size and ethnic fractionalization increase the likelihood of territorial conflicts, but not the likelihood of governmental conflicts. 23  conflicts is natural. Nevertheless, cases contrary to this logic abounds. Indeed, nearly half of all ethnic conflicts since World War II sought government control (Buhaug 2006, Table 1).24 Ethnic wars indeed diversify much more in terms of territorial vs. governmental aims (see Bartusevičius 2016; Buhaug, Cederman, and Gleditsch 2014; Wimmer, Cederman, and Min 2009).25 Here, I am concerned with ethnic territorial wars.   Why conflicts take on the character they do is an important question for the reasons behind the character of conflict continue to exert their impact to wartime and post-war dynamics. I call these reasons “causal antecedents of conflict.” My theory elaborates on the question of how civil wars lead to changes in trust, so the type of civil war or the conflict character is my independent variable. Nevertheless, the reasons behind the emergence of conflict i.e. the grievances that pave the way for conflict and the macro historical dynamics are as important factors in conflict character as the leadership of the insurgency, which eventually determine the type of war to be fought. The causal antecedents of conflict character are discussed in Chapter 4. Here, I focus on the link between civil war after its inception and trust. 2.2 Political Trust: Definition and Registers Political trust entails a belief that political institutions and actors therein will keep their promises and protect the interest of the citizenry even in the absence of scrutiny (Levi, Sacks, and Tyler 2009). Political trust has strong theoretical relevance to legitimacy concerns (Easton 1975), and is integral to the day-to-day functioning of democratic governments (Uslaner 2002), such as securing citizen compliance with tax-paying (Fjeldstad 2004; Scholz and Lubell 1998) and other                                                 24 The recent ethnic wars in Rwanda and Burundi, for example, were not territorial, but rather about government control. 25 Non-ethnic wars overwhelmingly seek governmental control; only a few non-ethnic territorial wars exist (e.g. conflicts in Nigeria and Pakistan in 2004). 24  governmental demands and regulations (Levi 1997).26 Trust is also instrumental in entrenching regime norms and maintaining political stability (Hutchison 2011).27 From a culturalist standpoint, political trust is politically exogenous; it extends from general trust orientations (principally interpersonal trust), which are embedded in cultural norms and acquired through political socialization (cf. Almond and Verba 1963; Inglehart 1997). From an institutionalist perspective, however, political trust is endogenous to institutional performance (see e.g. Hetherington 1998). I do not consider these two perspectives as mutually exclusive, and employ both insights in my theory. I presume that citizens’ satisfaction with the political actors and institutions serve as a yardstick in evaluation of their performance evaluations, notwithstanding the role of learned trust and cultural norms that condition the judgments behind these evaluations.  Duties of the state involve fulfillment of basic services (e.g. security, education) and provision of necessities (welfare, infrastructure) and administrative functions. Provision of security is indeed an ultimate reason for conception of state. In the early social contract theories, such as of Hobbes, Locke, Rousseau, the right to rule was granted to state in exchange for security from war (Milliken and Krause 2002).                                                                                                              26 See Hetherington (1998) and Hetherington (2005) for further details on political relevance of trust. 27 Early empirical works did not find the effect of declining trust levels in the late 1960s consequential on policy outputs, public opinion (Citrin 1974, Citrin and Green 1986), or political behaviour (Miller 1980, Pierce and Converse 1989). More recent literature, in contrast, provides conclusive evidence that changing trust in government has significant repercussions on domestic policy making (Chanley, Rudolph, and Rahn 2000, Rudolph and Evans 2005, Hetherington 2005), foreign policy and national defense preferences (Hetherington and Husser 2012), public opinion (Hetherington 1998) and vote choice (Hetherington 1999, Peterson and Wrighton 1998, Southwell and Everest 1998). The immense volume of these studies well establishes the political relevance of political trust (see Hetherington 1998, Hetherington 2005 for a detailed explanation on the political relevance of trust), and justifies treating political trust as a dependent variable. 25  2.3 Impact of Political Violence on Political Trust Political violence poses threat to security, identities, the authorities, and the institutions. In times of armed conflict, performance of the state with respect to provision of state services (the basics being security, order and general welfare) may be compromised. Long-term conflicts may also impinge on the cultural norms by changing citizen-state relationship. A state whose authority is challenged by an organized armed group and who has lost monopoly of coercion has to justify its rule and maintain its legitimacy in the eyes of its citizens. Even if the group that rises against the state authority is very small in numbers, the disorder it generates will have repercussions on political attitudes and behaviour.28   Most of the empirical work on political violence and political trust is situated in the terrorism literature.29 The dominant argument suggests an increase in political trust as a result of terrorism (e.g. Chanley 2002; Hutchison and Johnson 2011; Putnam 2002; Schubert, Stewart, and Curran 2002). These findings are closely related to the “rally-round-the-flag” effect–a sudden and a dramatic rise in public support in response to military and diplomatic crisis—first                                                 28 In the realm of political behaviour, most studies find positive effects of exposure to violence on political engagement, which are again explained by post-traumatic growth theory (Powell et al. 2003, Tedeschi and Calhoun 2004). Bellows and Miguel (2009), for example, show that war victims (by displacement and family deaths) are more likely to cast a ballot, be politically mobilized, and partake in local collective action (attending community meetings etc.) in Sierra Leone. Along the same lines, Wood (2003) observes that government violence in El Salvador prompted its victims to support and even join opposition forces out of moral outrage. Analogously, Blattman (2009) finds that abduction of children in Northern Uganda leads to substantial increases in voting and community leadership, largely due to elevated levels of violence witnessed. Palestinians surviving aerial attacks are also found be more political active (Punamäki, Qouta, and El Sarraj 1997). The terrorism literature harbours insights on the changes in voting behaviour induced by conflict (Berrebi and Klor 2006, Merolla and Zechmeister 2009, Dunning 2011, Kıbrıs 2011, Getmansky and Zeitzoff 2014). Also see (Holmes and De Piñeres 2002, Bali 2007, Siqueira and Sandler 2007, Berrebi and Klor 2008, Gould and Klor 2009, Montalvo 2011, Hersh 2013, Weintraub, Vargas, and Flores 2015). 29 There is also a vast related literature on terrorism and voting behaviour (Holmes and De Piñeres 2002, Berrebi and Klor 2006, Bali 2007, Siqueira and Sandler 2007, Berrebi and Klor 2008, Gould and Klor 2009, Merolla and Zechmeister 2009, Montalvo 2011, Dunning 2011, Kıbrıs 2011, Hersh 2013, Getmansky and Zeitzoff 2014, Weintraub, Vargas, and Flores 2015). 26  formulated by John Mueller (1973).30 Even though rallying literature largely focuses on presidential approval, it has long been argued that political trust is, in large part, a function of presidential approval (Citrin 1974; Citrin and Green 1986).31  Against the average positive effect of terrorism comes a myriad of other studies proposing conditional effects of terrorism on political trust or presidential approval. For instance, Edwards and Swenson (1997) argue that only certain groups, such as those who are already prone to support the president, will rally. Baum (2002) asserts that the most ambivalent groups, those with a touch of disapproval for the incumbent government and with moderate political awareness, are most likely to rally behind a president. Others suggest that the effect differs across types of political institutions (Dinesen and Jæger 2013). There are also some studies from outside of the US which shows a decrease in political trust (e.g. Gates and Justesen 2013; Montalvo 2011) or imply decline in trust (e.g. Gassebner, Jong-A-Pin, and Mierau 2008; Montalvo 2011) as a result of terrorism. However, most of the terrorism literature looks at the impact of a singular attack rather than the effect of continuous attacks, which is what civil wars characteristically feature (see Stanton 2013). All these works cited so far focus on international terrorism, yet civil wars feature domestic terrorism. Studies focusing on domestic attacks often demonstrate negative effects. Hutchison and Johnson (2011), in their aggregate cross-national study on 16 African countries,                                                 30 Although Mueller (1973) does not list terrorism among the events that can trigger a rally, his description of crises—“specific, dramatic and sharply focused” and pertinent to national security (Mueller 1973, p.209)—applies to terrorism (Chowanietz 2010). Mueller argues that sudden military interventions, major military developments in ongoing wars, major diplomatic developments, dramatic technological developments, meetings between the US president and leaders of other major powers, and the inauguration of each presidential term can trigger a rally. 31 Recent empirical studies also confirm that they go hand-in-hand: When public trust decreases, approval ratings also drop (Brody 1991, Hetherington 1998) and vice versa (Hetherington 1998), notwithstanding the occasional exceptions (Hetherington 2005). Chatagnier (2012) also finds that the rallying effect augments political trust. 27  find internal political violence to be deleterious to political trust (measured as generalized trust in government). Similarly, attacks by Tuaregs against the Malian state had detrimental effects on people’s trust in the national president (Gates and Justesen 2013). More evidence of negative effect comes from the Nepalese Civil War in the immediate aftermath of a ceasefire, where exposure to violence has been found to reduce trust in the national government (De Juan and Pierskalla 2016). Others contend that the effect of terrorism on political trust is conditional on party or context. Arce (2003), for instance, finds that the subversive acts of Shining Path in Peru decreased support for the left-wing government, and did not necessarily undermine the support for the right-wing government. Chowanietz (2010), on the other hand, contends that the party effect is context-dependent: in the UK, France, and the US, right-wing parties benefited more from rallies, while in Spain it was the left-wing party that enjoyed the rally effect more.32 The divergence in effects of international and domestic terrorism is usually attributed to the source of threat they pose, where the former features external threat and the latter an internal one, and my theory is inspired by this distinction.33 External threat stems from a source that is beyond the control of the government, rendering the government and the state vulnerable in the eyes of the citizens. Also, threat coming from an outside source magnifies the perceptions of in-group homogeneity (Rothgerber 1997; also see Hetherington and Rudolph 2008), which fosters a sense of collectivity in the public. In the face of an external collective threat, people search for a charismatic leader as a coping mechanism (Merolla and Zechmeister 2009; also see Berinsky                                                 32 The sample in this study includes a mix of domestic and international terrorism. 33 Starting with the Samuel Stouffer’s work in 1955, it is known that perception of threat is a significant spur behind attitudes and beliefs. Studies consistently show the direct effect of threat perception on decreasing political tolerance (Arwine and Mayer 2014, Feldman and Stenner 1997, Sullivan, Piereson, and Marcus 1982), changing voting choices (Merolla and Zechmeister 2009), and inducing authoritarianism, and altering policy preferences (Gadarian 2010, Gordon and Arian 2001, Hetherington and Suhay 2011, Huddy et al. 2005, Kam and Kinder 2007). 28  2009). They are more inclined to turn to the government as the legitimate actor that can respond to threat (Willer 2004), per social identity theory (Tajfel and Turner 1986), and tend to collectively lend in more political trust (Hetherington and Rudolph 2008; Mueller 1973). Oppositional elite voices are also usually muted as the information pertaining to external threat is limited, and elites tend to eschew appearing unpatriotic (Brody 1991).  Internal threat induced by domestic attacks originates from an insurgent group inside the territory, most likely stemming from political or economic grievances. Since the threat originates from within the territory, spurred by factors that the state, in theory, is responsible for, citizens may perceive domestic attacks as being more easily preventable. Put differently, when the source of the threat is homegrown, people may blame the government for not ensuring internal security. Furthermore, the competing elites are usually more vocal and critical (Chowanietz 2010), unlike in the case of international terrorism, where the silence of the opposition reinforces government’s handling of the threat. Even though the elites may denounce the terrorist act, the culpability of the government often underlies their interpretation of the attack. Moreover, a unified collective support from the public is usually missing in domestic crises, since these are usually a surfacing of an underlying societal issue and are more likely to exacerbate the internal discord (Mueller 1973).  I argue in the case of civil wars that the dynamics of external threat may be present. Whether the internal threat is secessionist or ideological in nature is important. Hutchison & Gibler (2007), for example, show that territorial issues bear greater significance domestically than other types of international issues. Ethnic territorial conflicts may distort the perception of internal threat. Even though the threat is homegrown, it may cause an illusion of an external threat; the internal group’s desire to secede may be perceived as an out-group posing an external 29  threat. Inspired by these insights, I construct my theory below. 2.3.1 Theory of Political Trust in Wartime My theory of political trust in wartime is based on the following propositions. I begin with the assumption that threat is the main mechanism through which causal impulses for trust are affected.34 Threat and fear are psychological phenomena that are tightly linked to basic needs of security. Loss of security is a causal impulse that will impinge on people’s evaluation of government’s performance as well as feelings towards/judgment about other individuals (see Figure 3). I conceptualize the threat conflicts may pose on two levels: collective and individual/personal level, following Huddy et al. (2005).    Figure 3. Conflict and Political Trust: Crude Mechanisms I argue that it is the collective threat framing that determines how civil wars will impinge on political trust. Collective-level threat is the threat the state or the regime is facing and it may be internal or external. Based on my inductive insight from my case studies (see Chapter 4), I propose that wars posing threat to the identity of the state/nation are more likely to induce a                                                 34 Starting with the Samuel Stouffer’s work in 1955, it is known that perception of threat is a significant spur behind attitudes and beliefs. Studies consistently show the direct effect of threat perception on decreasing political tolerance (Arwine and Mayer 2014, Feldman and Stenner 1997, Sullivan, Piereson, and Marcus 1982), changing voting choices (Merolla and Zechmeister 2009), and inducing authoritarianism, and altering policy preferences (Gadarian 2010, Gordon and Arian 2001, Hetherington and Suhay 2011, Huddy et al. 2005, Kam and Kinder 2007). Armed	Conflict	 Threat		• Performance	Evalua8ons	of	Authori8es	and	Ins8tu8ons	• Regime	and	Poli8cal	System	Evalua8ons	Poli8cal	Trust	30  strong collective threat framing by the state, assuming that the insurgents are considered militarily strong.35  In contexts where collective threat is pervasive and the nation’s identity is at stake, collective threat should drive the trust reactions, which tend to rally people around the flag. I should emphasize that collective threat framing is different from the act of denouncing the insurgents as terrorism. State elites often benefit from framing the threat as external and demonizing the insurgents as “enemies” in the eyes of the public in order to fortify their authority and legitimacy (Vasquez 1993). However, discursive framing of the threat they pose and the extent to which the conflict is depicted as a national matter underpins the concept of collective threat framing here. Following Hetherington and Suhay (2011), I propose personal physical threat as an important component of perceived threat (also see Pyszczynski, Solomon, and Greenberg 2003). Personal threat is the threat an individual perceives as threat to her and her family’s security, bodily rights, and property. One psychological process underlying personal physical threat is mortality salience such as heightened concerns about death (see Greenberg, Pyszczynski, and Solomon 1986 for mortality salience), extending from fear (see Mueller 2006; Mythen and Walklate 2006 for culture of fear). When individuals feel personal threat, regardless of the perpetrator, their baseline evaluation of government performance in providing them with security will be negative, which should result in decrease in trust. Collective threat framing has, however, stronger effects on trust calculations, and such framing should boost political trust in                                                 35 The magnitude and severity of the threat that insurgents pose to the state also depends on the power balance between the warring parties. The state may downplay the importance of the insurgency. This may especially be the case if the insurgents are considered weak allowing the state to be dismissive of their actions or if there is power symmetry between the state’s and the insurgency’s military capacity— not to appear weak to the public. 31  absence of personal threat, and should moderate the negative effect of personal threat if the feared perpetrator is the insurgents (see section 2.3.1.3 for details). 2.3.1.1 Conflict Character, Threat and Micro-Level Repercussions on Trust I suggest that ethnic territorial conflicts usually threaten the identity of the nation because secessionist claims challenge the territorial integrity that defines a nation-state while ideological conflicts only pose a kindred threat when they emerge in countries that identify with a particular ideology or regime type, as in the cases of the Soviet Union, Cuba, or the United States. My theory also suggests that the structure of civil conflicts is itself endogenous to features of state formation. I draw on my inductive insights from the conflicts in Turkey and Peru to suggest that grievances that led to the onset of insurgencies extend from nation-building (See Chapter 4 for inductive case studies). The dynamics inherent to nation-building also shape the discursive framework employed by the state to respond to the conflict. Ideology and strategies of nation-building mould societal forces, political groups, and the political landscape of the country for many years to come (See Chapter 4).   In the case of secessionist territorial conflicts, the state produces a discourse of threat against national unity and capitalizes on agitation to galvanize people to support itself in the face of an existential threat to its territorial integrity and nationhood (see 2.3.1.2 on Discourse). It should marshal all nationalist and ideological tools at its disposal to construct a discourse around national unity, portraying the insurgency as a collective threat, and rally the general public around status quo political institutions. For those identifying with the nation, this collective threat framing should have a strong rallying effect around the state throughout the conflict as long as the emphasis on collective threat continues. The underlying psychological mechanism is likely that the ethnic group represented by the state (politically powerful ethnic group) tends to 32  perceive the internal threat posed by another co-national ethnic group (often an ethnic minority) as an external threat, and turn to the actors representing itself and rely on them. Even when faced with a personal threat, because of the overarching collective threat they feel to the nation, their allegiance to the state does not swerve. The insurgents who are perceived to be posing a personal threat become enemies, and it glorifies the state in their eyes even further. If that is the case, personal threat should even solidify their attachments to the state and lead to higher political trust. The meaning attributed to the nation via state discourse overwhelms the value of human life.  The underrepresented/politically weak ethnic group will undergo a different process.36 As long as they subscribe to their ethnic identity, collective threat framing generated by the state should not resonate with them as they do not consider themselves as part of the collectivity of the politically represented ethnic group, thus rallying around national institutions should be less appealing. In the absence of personal threat, the conflict should not directly lead to any change in their political trust. Often, the ethnic groups claimed to be represented by the insurgency are influenced by the counterhegemonic discourse of the insurgency. This discourse and heightened salience of their ethnic identity could turn them against the state. When under high personal threat, their distrust in the state and political institutions will be aggravated. Remote co-ethnics would be indirectly affected by the sufferings of their people in these territories through ethnic bonds and insurgency discourse (see section 2.3.1.2 for discourses and section 2.3.1.5 for identities).                                                 36 Salience of social identities is subject to change. See the section on “Identities” for a discussion on identity transformations during wartime. 33   In the case of ideological conflicts, for a state to marshal ideological tools to evoke patriotic sentiments to rally people, it needs to strongly identify with a particular regime.37 Nations identifying themselves with a particular regime (democracy, communism, fascism etc.) would be better able to weather attacks to the regime by rallying the public around the nationalist ideological tools. Nations without a clear regime or ideological identity, or that are experimenting with a new regime do not have such national ideological tools at their disposal and hence are more vulnerable to ideological threats and attacks to the regime, and they are less motivated to frame the insurgency as a collective threat as it may weaken their appearance. In such nations without a strong regime identity, I contend that collective threat framing of ideological wars will be weak.  In the absence of strong collective threat framing, personal threat should determine the trust calculations. When under personal threat (especially low intensity), the baseline evaluation of government performance in providing security should be negative, as suggested above, which will result in decrease in trust. Because there is no strong overarching collective threat, there is no force to offset the negative effect of personal threat. Prolonged extreme conditions of insecurity hurt the perception that state is doing one of its most fundamental jobs— protecting its citizens. When the collective threat framing is not strong, individuals tend                                                 37 It is thus worth to underscore that ethnic wars may have an ideological dimension, and non-ethnic wars may have an ethnic character. What may appear as an ethno-nationalist conflict may, at its roots, be a manifestation of social and economic grievances Take the example of Sendero in Peru: Sendero featured itself as a Maoist insurgency, and in most datasets that involve a typology the civil war in Peru is coded as a non-ethnic, ideological, revolutionary war. However, Sendero’s discourse was tied to and intended to speak to the grievances of the indigenous groups. The leadership mobilized and recruited predominantly from the indigenous people, and it waged its war, for the most part, in the indigenous-populated areas of Peru. In the same vein, the conflict between The National Union for the Total Independence of Angola (UNITA) and the government of Angola (dominated by the Popular Movement for the Liberation of Angola (MPLA)) was at first sight an ideological conflict. The UNITA branded itself as an anti-communist movement, and did not officially represent any ethnic group. Yet, UNITA was comprised primarily of Ovimbundu group members, while the government was comprised primarily of Mbundu group members (Collelo 1991). 34  to perceive their precarious security conditions more personally and lose their trust in state’s ability to protect.  Nevertheless, the fact that personal threat is ubiquitous induces a sense of collective threat, even when the state does not frame the threat of insurgency as a collective threat. The fear that the insecurity may persist long and that the idea that state of fear may be the norm pushes people towards the state and previous status quo. Works in psychology find that pervasive threat induces conservative shifts among all people, enhancing authoritarian and patriotic tendencies (e.g. Jost et al. 2003; Nail et al. 2009). Increasing authoritarianism signifies higher submission to authorities (Altemeyer 1981). Thus I hypothesize that: As the level and scope of personal threat intensifies, individuals will gradually begin to perceive a more ubiquitous threat generating a sense of collective threat, and feeling of collective threat should moderate the negative effect of personal threat on political trust.38 Cognitive dissonance is what lies behind the seemingly paradoxical response to extreme insecurities, for instance, by supporting the state at the expense of personal or collective interests. The System Justification Theory, drawing on the dissonance theory (Jost and Banaji 1994), posits that people facing threat are motivated to cognitively justify the prevailing system because of the need for order and stability; furthermore, ideologies underlying this motive act as tools to weather existential anxiety (Jost and Hunyady 2003). Members of disadvantaged groups may entertain similar justification processes for the status quo even though it is in contradiction with their personal interests (Jost and Hunyady 2003). Some of the                                                 38 Conflict may also affect attitudes and orientations of bystanders through interpersonal communication, which is shown to be as effective as mass media in supply of information and public opinion formation (Yang and Stone 2003). Speaking to the victims, especially in contexts where identity polarization is blatant and groups are defined along identity lines, could alter the cognitive processing of conflict and thus attitudes for bystanders. 35  underlying reasons behind such expressions of conservatism are: “epistemic (need for order and closure, cognitive complexity, etc.), existential (fear, terror management, anger, self-esteem, etc.) and ideological (system justification, group dominance, social dominance, etc.)” (Echabe and Guede 2006, 260). My theory distinguishes between political trust and legitimacy. I conceptualize legitimacy as a higher order concept than trust, and don’t expect lower political trust observed under conditions of high personal insecurity to necessarily damage state’s legitimacy. Legitimacy is most likely to be affected when political trust is damaged more permanently, and when recognition of state’s authority to govern is at stake rather than short-term or transient failures to provide security. Decreasing political trust is not tantamount to loss of legitimacy. Physical insecurity does not immediately lead to a loss of legitimacy. Ontological insecurity, however, entails a loss of legitimacy as explained in section 2.3.1.4. Below, I further expound the logic of the theory and causal mechanisms by unfolding each connecting part of the theory. I start with the mechanisms of discourse and collective threat framing. Next, I discuss how individuals’ personal security risk perceptions are formed, distinguishing between physical and ontological security. Following that, implications of threat on political legitimacy are spelled out. Before presenting the theory of social trust in wartime, I offer a discussion on the role of identities and identity transformation, which affects how individuals process information about political institutions, actors and their performances, and how they perceive individual, group and national/collective threat.   2.3.1.2 Generation of Discourses and Collective Threat Framing I define discourse here as a “social and political construction that establishes a system of relations between different objects and practices, while providing (subject) positions with which 36  social agents can identify” (Howarth and Stavrakakis 2000, 3-4). Official discourses of a party involved in a conflict usually aim to delegitimize the other party (or parties) while justifying their own actions, however violent those actions may be. These discourses serve as a frame of reference for the public and play major roles in threat framing and opinion formation, and they may condition or aggravate attitudinal responses. Discursive frames pertaining to the dynamics of war emanate from the historical narratives used to explain the interaction of the parties involved in the conflict, as well as from considerations of national identity, context-specific parameters, and, most importantly, conflict character. Armed conflict assumes its character in response to the historical grievances at play and in accordance with the political orientations of the insurgency leadership. Though each case of armed conflict is unique, we can observe parallels in the discursive formations in conflicts of a similar character. For example, ethnic territorial conflicts tend to evoke a nationalist discourse by the state, while nonethnic ideological conflicts are more likely to spur a discourse rooted more around the identity of the state and the regime. Yet, the qualitative content of such discourses will vary because of differences in context, historical antecedents, and the particular dynamics of the conflicts.  The assumption is that for an insurgency to be threatening, political elites need to attribute national political salience to the insurgency. Media coverage and elite rhetoric are critical in attributing national significance to an event and shaping the public’s national importance judgments (Behr and Iyengar 1985; Erbring, Goldenberg, and Miller 1980; MacKuen and Coombs 1981). National importance judgments have political consequences, and it is reasonable to expect them to influence political trust (Hetherington and Rudolph 2008). As an issue becomes nationally significant, the public will be more attentive to the performance of the 37  government in dealing with the issue and will adjust their trust levels accordingly. The larger the number of people naming an issue as nationally important, the more performance evaluations will influence political trust (Hetherington and Rudolph 2008).  Once an issue gains national importance, elites generate a discourse to frame the issue. The official discourse of a state facing a separatist insurgency often frames the issue as an existential threat to the nation and use emotional appeals to spur a sense of unity around a national collective goal of keeping the nation together. Often the nation-states have a plethora of available tools and national symbols geared towards exploiting emotional ties to the nation. Many states invest in parades, celebrations of national days and deploy national anthems, historical myths, flags and military uniforms to fashion national identity as part of nation-building (Schatz, Staub, and Lavine 1999; Sears 1993). By producing a discourse pivoted around an existential threat to the nation and utilizing such national symbols, states may incite nationalist mobilization.  In ethnic territorial conflicts, the effect of secession is more restricted to a bounded territory. In non-ethnic ideological conflicts, the threat is to the regime or the state. Effectively, it relates to every citizen of the country where the insurgency occurs, as the change in regime will impact them all personally. However, unless a certain ideology is embedded in nationalist symbols, states that are under attack from an ideological insurgency do not have commensurate ideological tools to induce a similar nationalist support at a collective level.  The extent to which elite-generated discourses are communicated to and find resonance within the public is also crucial to consider. In wartime, it is of primary interest for the state to disseminate its official discourse on the conflict (describing its nature, its threats and its reasons) and the media are key in these communications (see Gamson and Modigliani 1989). Research 38  consistently shows the media’s role in agenda setting, as well as priming, and framing, referring to the ability of news coverage to highlight the importance of political issues and political judgments (McCombs and Shaw 1972; Soroka 2002; Iyengar and Simon 1993; Kinder and Iyengar 1987; Kahneman and Tversky 1984). The way the mass media communicate information on an ongoing conflict is consequential on public opinion about the war (Gadarian 2010).39 That’s why elites often use the mass media to convey their opinion or discourse.   To what extent the media will regenerate or perpetuate this discourse (by enhancing the visibility of the official discourse), or challenge it hinges upon the media’s capacity, relationship with the state, and press freedoms. In most democratizing countries (the “grey-zone” countries—so-called semi-democracies), the freedom of the media is limited. In transitioning regimes particularly, an independent media is usually an aspiration rather than a norm. The mass media networks are often coopted into the state’s domination mechanism.  Nevertheless, the media is not just a tool at the elite’s disposal; it has its own interests. For instance, the media benefits from sensationalist coverage, which tends to overblow events and embellish stories with evocative imageries and exploitation of emotions. The emotional cues in news media usually accentuate threat. Especially in contexts where investigative research is hampered such sensationalist coverage may be appealing to make up for the paucity of in-depth information on the conflict. Media representations are vital for bystanders’ attitude formations. In civil war contexts where the media freedoms are curtailed and coverage of the news is strictly censored, public                                                 39 Though the research on media effects on opinion and policy preferences mostly focuses on foreign policy coverage, and the outcome variable is usually public support for war, the pillars of the theories in this literature are applicable to public opinion on domestic policy issues.  39  access to both a true account of the conflict and balanced information and assessment of it is highly restricted. The same scenario is valid for cases where the media lacks the capacity to undertake independent investigative reporting. 2.3.1.2.1 Clash of Discourses Against the hegemonic state discourse, insurgent discourse may unearth silent memories from the past and arouse them. Past grievances are usually the instigator of insurgent movements and spreading these historical grievances is an important goal of an insurgency in order to legitimize its mission. While challenging the hegemonic state discourse at the macro level underlies this goal, bringing the collective memories to light takes place at the meso level (community and group-level), new discourses emerge. New memories created during an armed conflict along with the history of repression are also quickly integrated into the narrative and the conflict updates the collective memory. Despite high levels of collaboration between official historians and state officials, alternative sources of historiographic knowledge have consistently challenged the hegemonic discourses on memory (Bevernage 2010). Critical historians, civic human rights initiatives, and of course, community elders with “unofficial” stories make it possible to resist the totalizing tendencies of official history (Bickford 2007; Evans 2002).40 While the state discourse is likely to find resonance among those who are represented by the central political institutions, insurgent discourse is likely to spread among the politically weak/underrepresented groups, to marshall support for the rebellion, and decrease the sympathizers’ political trust. Social representations of the past in the insurgent discourse                                                 40 I should note that the local memories are not necessarily in line with the cleavages of the conflict because local cleavages are not always parallel to the national cleavages. As Wood argues: “What appears at the national level to be the key issue—for example, class relations, constitutions, or ethnic secession—may not be salient at the local level, which may be dominated by conflicts” (Wood 2008, 547).  40  influence feelings, desires and actions of the disadvantaged groups (Bar-Tal and Antebi 1992; Liu et al. 1999), affecting the construction of who they are. Inasmuch as social representations play a big role in construction of social identities, social identity affects social representations (Breakwell 2001; Moloney and Walker 2007). The literature on collective memory and collective identity echoes the same mutually constitutive relationship between the two (Olick and Robbins 1998) (also see section 2.3.1.5).  2.3.1.3 On Personal Threat How do people calculate the threat they are facing? What goes into the perception of risk? Individual interpretations may vary even though the type and magnitude of violence they are exposed to or they construe is the same. The literature on civilian targeting has very useful insights on this point. Scholars suggest that violence may have contrasting effects on civilian support depending on the nature of the violence (indiscriminate vs. selective), the intentionality attributed to it, the precision with which violence is applied, and the identity of the perpetrator (Downes 2007; Kalyvas 2006; Kocher, Pepinsky, and Kalyvas 2011; Lyall, Blair, and Imai 2013).   Personal physical insecurity is the focus in the aforementioned literature, but ontological insecurity is as important to perceived threat as physical insecurity. Physical security entails protection of the body against threats that may harm—personal threat implies an expected breach of physical security. Ontological security entails protection of identity and personhood from threats. I borrow the term ontological security from Anthony Giddens, and international relations scholars who build on his work. Ontological security denotes the need to experience stability in the cognitive environment and feel like a continuous and whole person in time, which is essential 41  for individual agency (Mitzen 2006). When an individual is surrounded by deep uncertainty, she feels threat to the security of her being and self.41 I suggest that loss of physical vs. ontological security is associated with different consequences for political trust. Individuals who do not feel physical security at certain times during the war do not necessarily lose their trust; the process is neither automatic nor simple. People factor in both the availability of alternative authorities on which to confer legitimacy as well as their ontological security (state’s performance/intention in providing them with identity security and the social representations for past relations with the state as well as state’s efforts and intentions in providing physical security). Hence, a person who never felt threat to their physical security may lose their political trust because of ontological insecurity while a person who was under complete physical insecurity may lend higher political support because of resonating with the collective threat framing of the state. Ontological insecurity has consequences that are more far-reaching. Individuals who lose their ontological security due to conflict should not confer any legitimacy to the state, or authorities, or even the regime (See section 2.3.1.4). 2.3.1.3.1 Determinants of Personal Threat The repertoire of violence deployed by insurgent groups and the state influences perceived personal threat. The methods insurgents or the state use, which may range from primitive arms such as machetes to car bombs or suicide bombing, may compound the perception of risk. An                                                 41 Jennifer Mitzen (2006, 345) articulates that: Ontological insecurity refers to the deep, incapacitating state of not knowing which dangers to confront and which to ignore, i.e. how to get by in the world. When there is ontological insecurity, the individual’s energy is consumed meeting immediate needs… Ontological security, in contrast, is the condition that obtains when an individual has confident expectations, even if probabilistic, about the means– ends relationships that govern her social life. Armed with ontological security, the individual will know how to act and therefore how to be herself. 42  insurgency (such as Sendero) that engages in urban warfare with car bombs poses a different menu of threat to the general public than an insurgency that fights with an organized professional army in rural areas such as the PKK. Patterns of violence, oscillating between selective and indiscriminate, are just as important as the methods. The repertoire of violence, organizational capacity, and relative strength of the insurgency vis-à-vis the state, all together, condition or aggravate the insecurities individuals feel. Of course, insurgents are not the only party that may pose individual-level threat. State violence is very common in civil wars. The state may resort to violence to protect its own security at the expense of individual securities. Buzan (1983, 31) contends that the contradiction between individual and state security is unavoidable and is “rooted in the nature of political collectives.” A belligerent state may violate the rights of its very own citizens to protect its self-interest and pose security risks. The counter-insurgency tactics may aggravate the grievances of the same group whose identity and other rights have already been violated. State violence, exercised to protect its sovereignty, may be the biggest threat to its legitimacy. This is observed in many intra-state conflicts. Protracted clashes between Hutus and Tutsi in Rwanda and Burundi, between Tamils and Sinhalese in Sri Lanka, between Ladinos and indigenous groups in Guatemala, between Turks and Kurds in Turkey, all contain forms of state violence. Kalyvas (2007) argues that indiscriminate violence will induce feeling of insecurity and distrust much more than selective violence, and I think this is true for violence perpetrated by the state and the insurgents for social trust. If the state is an agent of indiscriminate violence against civilians, I expect it to have heavy repercussions on political legitimacy, which I discuss below.  43  2.3.1.4 Implications of Threat on Political Legitimacy Political legitimacy is relevant to my theory as I suggest that threat has implications not only for political trust but also legitimacy. Political legitimacy is accorded a vital role in the maintenance of regimes or states as well as understanding the change in them (Easton 1957; Weber 1978), notwithstanding the critiques (Przeworski 1986). Its importance, in large part, stems from its link with obedience and hence the potential to reduce violence. Once the authority justifies itself and garners the support of the citizens, citizens will comply with the laws and regulations and the authority, and motivation for rebellion will be obviated. Political trust is often employed to operationalize legitimacy. Yet, I suggest it is a concept of a higher order, decline of which carries stronger implications for the survival of the regime than political trust.  I define political legitimacy here as the conviction of the governed that the ruler has rightful authority to govern and that they are bound to obey (Alagappa 1995; Gilley 2006). Justification of authority is ascribed an integral role in the conception of legitimacy by many theorists (e.g. Beetham 1991; Buchanan 2002; Krehoff 2008).42 Components of justification are sometimes described as moral and performance legitimacy (Williams 2010), which tap into “diffuse” and “specific” political support, respectively (Easton 1965). While diffuse political support represents a conviction about moral conformism between the functioning of the government and individuals’ principles about what is right, specific political support refers to performance evaluations. Moral legitimacy is inherently linked to performance legitimacy. Bad performance of governments in providing basic services                                                 42 Other aspects of power that needs to be legitimated for the public to have the conviction that the authority in a modern state has political rightfulness are legality of acquisition and exercise of power (constitutional legality), and of consent to power (Beetham 1991, Gilley 2006).  44  will have repercussions on the moral legitimacy of the state. The question is how a government error would reflect on the legitimacy of the state? Easton posits that “dissatisfaction with the authorities may be of sufficient duration and of sufficient intensity to breed a more generalized feeling than the authorities…. [So it is possible that] loss of specific support for political authorities - the incumbents of roles - has thereby become converted into a decline in support for one part of the regime: the authority roles” (1975, p.449).  I argue that during wartime, when an individual or group feels ontological insecurity, the moral legitimacy of the government and the state should be damaged. Ontological insecurity operates differently from physical insecurity. It is very likely that some citizens feel personal threat from the state security forces, and those cases are when ontological security is shaken. Perception of ontological security is heavily conditioned by the discourse of the insurgency, collective memory, and group identities. Even in the absence of conceivable threat to self, people may feel ontologically insecure if their collective identities are under threat or if they feel the state does not have sincere intentions to protect them. Ontological insecurity may emanate from group-level threat depending on the state’s actions and intentions towards the group. Individuals who feel ontological insecurity not only should hold negative evaluations about the performance of the state in security provision but also question the authority of the state authority.   2.3.1.5 Identities, Group Influence, and Political Opinions  Identity crosscuts all processes linked to conflict and mediates the effects of conflict on trust on every level. Identity is “a crucial social-cognitive mediator that enables people to comprehend and act in their social worlds as self-conscious and motivated agents” (Simon 2008, 41). It is important to discuss identity because, as a great number of works demonstrate, subjective group membership is an important variable in shaping political attitudes and behaviour (Campbell et al. 45  1960; Conover 1988; Leighley and Vedlitz 1999; Miller et al. 1981).  The relevance of identity to this dissertation, in large part, stems from the centrality of identity in individual attitude formation as elaborated in reference group theory, and its sibling, social identity theory. Seminal works in the early American public opinion literature offered group reference points as a source people draw on to make meaningful political decisions (see Converse 1964). Individuals construct groups by placing themselves in one (in-group), and setting apart those they dislike in another (out-group). Using affection towards groups as reference cues, individuals can comprehend the political world, make political judgments on complex issues and form their attitudes (Brady and Sniderman 1985). Social identity theory echoes the insights of the reference group theory and attributes crucial role to social groups. Social identity, according to the founding father of the theory Henri Tajfel (1981, 285) is predicated upon “knowledge of membership of a social group (or groups) together with the value and emotional significance attached to that membership.” Again, ordinary individuals use their membership in a group to make sense of the political world.  Works in social psychology have already shown that socially shared representations of history sway peoples, with objectively akin interests, to adopt qualitatively distinct attitudes with respect to issues of international politics and internal diversity (Liu and Hilton 2005). Liu and László (2007, 86) contend that “the same political situation could engender quite a different probability space of responses from different peoples, depending on their representations of the historical experiences that have shaped them as a people.” Building on these accounts, I suggest that there would be multiple different interpretations of the same event since memories of victims and bystanders, as products of their subjective interpretations, are in constant interaction with social identities and memory frameworks.  46   In this dissertation, identity transformation due to conflict (especially an ethnic one) serves as one of the mechanisms through which civil war exerts influence on trust. Identities change the lens through which one sees the political institutions; alter the framework through which one processes the past; and affect the collective memories one subscribes to. As an individual undergoes identity change, she positions herself differently vis-à-vis the society and the state and perceives the political system, actors and other ethnic groups differently. This change in perspective ultimately affects political and social trust.  2.3.1.5.1 Identities in the Face of Armed Conflict Armed conflict shapes identities by restructuring the extant ethnic or ideological or social groupings. New groups may be reconstructed because of the conflict. Construction of identities is embedded in a growing number of studies as an outcome of war. Smith (1981) spearheaded the systematic work on how wars affect the formation and cohesion of ethnic communities, building on Simmel’s (1964) path-breaking work on conflict. Brubaker (2004) also underscores that high levels of groupness may be more a result of conflict (especially violent conflict) than its underlying cause. Cederman, Gleditsch, and Bauhaug (2013) also consider the possibility of identity crystallizations as a result of conflict rather than the reverse sequence in their analysis of conflict processes. Balcells’s (2012) work on the role of victimization in generation of new political identities or redefinition of existing ones also speaks to the same pattern. In my theory of political trust in wartime, identity transformation is an important mechanism as conflict alters trust attitudes by changing identities as well. In the course of a conflict, identities crystallize and are realigned around the main cleavages of the war, and individuals may change the way they perceive the war events if they adopt new identities. For instance, I encountered many ethnic Kurds in Turkey who came to identify themselves as 47  Kurdish only as a result of the Kurdish insurgency, and who altered their perception of the state and of their co-ethnics and ethnic Turks. I argue that the interplay between social representations, collective memory, and social identities reflects on trust attitudes by reinforcing reconstruction of social groups. Individuals, by crystallizing their social identities in the face of the new dynamics and discourses imposed by the conflict, may reconstruct their attitudes.   2.4 Armed Conflict and Social Trust Trust requires a belief that others will not deliberately harm you, which contrasts with the wartime experiences where purposeful harm inflicted on others is part and parcel of life. Generalized social trust is an important building block for large-scale, complex, interconnected social networks and institutions, and therefore is a key disposition for generating social capital. It also underwrites other essential dispositions for democratic culture such as tolerance (see Warren 1999). I also look at category-based trust, also called identity or group-based trust (Freitag and Bauer 2013; Kramer 1999; Stolle 2002). This conception of trust extends from the social identity theory (Tajfel 1974; Tajfel and Turner 1979), and identification lies at the center of it. Category-based trust entails trusting a person with whom one has an identity link even though there is no personal relationship. The logic is that people are more prone to trust those with whom they share a group identity or social category than others (Freitag and Bauer 2013; Kramer 1999).43 Most of the literature on violence and social trust looks at generalized interpersonal trust. The conventional approach suggests that mutual distrust is pervasive in war-torn countries (Colletta and Cullen 2000; Collier 2003; Widner 2004). Empirical evidence from Croatia and                                                 43 Shared identity is defined broadly; it encompasses common fate, mores, ethnicity, regional proximity, behavioural similarities, or cultural traditions. The assumption is that sharing a social categorization magnifies the commonality among members of the category (Stolle 2002).  48  Uganda for instance shows that wartime trauma is associated with lower generalized trust in others (Kunovich and Hodson 1999; Rohner, Thoenig, and Zilibotti 2013). Similarly, utilizing trust games after the 2007 political outbreaks in Nairobi, Kenya, Becchetti, Conzo & Romeo (2014), find that experience of direct or indirect physical violence and/or forced relocation significantly reduces trustworthiness learning. Studies on social capital echo the negative effects of war on trust and social relations (Cassar, Grosjean, and Whitt 2011). Some empirical social psychology studies corroborate insights of these studies and found that external threat of terrorism leads to lower social tolerance (Feldman and Stenner 1997; Huddy et al. 2005). A related strand of studies find personal exposure to terrorism leads to higher prejudice (Besser and Neria 2009), and higher exclusionist political attitudes as a result of personal exposure to terrorism (Canetti-Nisim et al. 2009).44 Other studies indicate contrasting effects. A number of findings show that violence of higher intensity, counterintuitively perhaps, leads to greater levels of generalized trust, explained by post-traumatic growth theory (e.g. Blattman 2009). Bellows and Miguel (2009), for instance, show that households that had direct experience with wartime violence in Sierra Leone, are more likely to participate in community meetings and engage with local political groups. In the same vein, evidence from Burundi indicates that exposure to violence increases altruistic behaviour (Voors et al. 2012). Gilligan, Pasquale, and Samii’s (2014), analogously, find that members of                                                 44 There is also a vast literature on the psychological effects of war, ranging from trauma and accompanying stress disorders to other demographic changes such as fertility (Brunborg and Urdal 2005). Studies also consistently show the adverse psychological effects of war on children (Barenbaum, Ruchkin, and Schwab‐Stone 2004, Blattman and Annan 2010, Dyregrov, Gjestad, and Raundalen 2002). Another related strand of research finds negative effects of political violence on tolerance (Hutchison 2014, Merolla and Zechmeister 2009). A more recent study distinguishes between ideologies and shows that detrimental effects of continuous terrorism on tolerance are stronger on those who affiliate with the Right (Peffley, Hutchison, and Shamir 2015). 49  communities who were more exposed to civil war are more inclined to contribute to public goods and are more trusting. The positive effects of conflict on trust may not be exclusive to victims. Coser (1956) posits that conflict demarcates the society across identity lines, and group identities come to the fore. Conflict may increase the interpersonal trust relations by solidifying the internal cohesion of a group as a whole (Coser 1956), or divide the society as in-group and out-group members. The demonizing responses of elites to attacks or threats may further contribute to such division. Whoever is associated with the enemy through their ideological, ethnic, religious, or other group ties are excluded from the in-group. So, it is very possible that while generalized trust declines in the face of civil war, in-group trust soars.  A recent cross-national study posits conditional effects of war on the type of conflict. The authors find that religious wars on average reduce social trust more so than ethnic wars (Traunmüller, Born, and Freitag 2015). The same study also suggests that generalized trust is affected more than particularized trust (trust in family) except for trust within the neighbourhood, and wars fought along ethnic or religious cleavages is most harmful on this type of trust. Yet other findings, e.g. from Aceh, Indonesia, however, fail to find a general pattern between wartime trauma and social trust—the effects of violence appears to be more context dependent (Shewfelt 2009). Evidently, the literature has mixed findings. As possible factors driving such heterogeneous findings, scholars propose type of the civil war, context, and personal experiences. Along with these factors, my theory also suggests polarization along group lines and identity transformation as additional conspicuous mechanisms explaining the change in social trust. Elite discourse on the conflict could polarize the society along group lines, 50  generating an illusion of in-group and out-group distinctions even in a well-blended society, particularly in tandem with the polarization of political identities and dissolution of old social networks, which often occurs during civil war (Wood 2008). New social group categories may emerge out of these discourses. Similarly, forced displacement in wartime can change the social fabric of the society, and force individuals to reformulate their social groupings and trust relations as explained in the previous section. It is through these changes in social groupings and identities that social trust is reshaped during wartime. My theory explains how interplay of threat and identities dictate the direction of change by distinguishing between generalized trust and category-based trust, and ethnic and ideological war.  2.4.1 Theory of Social Trust in Wartime My theory presumes that ethnic territorial wars and ideological wars have distinct effects on social trust. Identities operate along with threat to determine generalized and category-based trust. Because of the differences in cleavages, ethnic and ideological wars play out differently.  Ethnic warfare heightens the salience of ethnic identity. As the war unfolds, ethnic group definitions are crystallized.  Inasmuch as identity markers are clear—language, phenotypes, cultural codes are easy markers—in-group and out-group distinctions emerge. Category-based trust becomes relevant. Indeed, identity conflicts in general do not bode well for particularized trust as cleavages across the lines of ethnicity or religion polarize the society and destroy the social fabric even in local contexts (Kalyvas 2006). According to social identity theory, conflict with a rival out-group leads to stronger identification with one’s in-group, which translate into mistrust towards the out-group. 51  From the perspective of a powerful ethnic group, the group associated with the insurgents is not trustworthy as it is considered to be with the enemy. Based on these blanket judgments, trust levels towards the out-group will diminish. In reaction, a politically weak ethnic group facing blanket judgments from the other group based on their ethnic identity should react in a similar way. Identity-based exclusions or discriminations emanates from lower trust towards out-groups. These attitudes should further alienate the out-groups. As a result, when group boundaries are clear, we should observe that groups curtail their out-group trust. Personal negative experiences would magnify the negative judgments towards out-groups. Even though personal relationship is not a condition for identity-based trust, personal positive experiences with a member of a social category may help reinforce positive attitude towards other members, and negative experiences may help reinforce negative attitudes towards other members. Hence, distrust for out-groups should be stronger for people who have personal negative experiences with the members of that out-group.  Ideological affiliation, in contrast, is not primordial in character and is not as sticky an identity as ethnicity. Group polarization along the lines of ideology can still occur but markers may not be as clear. Also, changing camps is an option— while undressing ethnicity is much more difficult. Because defining the groups is challenging, distinctions such as in-group and out-groups do not as easily apply to ideological conflicts. Because of the fluidity of ideological identity (i.e. it is easier to switch), group boundaries are blurry and cultivating in-group trust is harder. By the same logic, recognition of the out-group is problematic. Someone out of sympathy with insurgents may worry that he or she is the next victim. Those in sympathy may worry that the person next door is accomplice of state violence. Thus, when distinguishing groups is hard, generalized trust plummets. In contexts where the insurgency is active, I expect distrust to be 52  more pervasive.  By the same token, I expect effect of ideological wars on category-based trust to be weak. 2.4.1.1 The Role of Context Social context is also a strong factor for processing information and opinion formation (Kinder and Kiewiet 1981; Kinder and Rhodebeck 1982; Sears et al. 1978; Tolbert and Hero 1996). I hypothesize that people living in volatile contexts or contexts that undergo major compositional changes due to the war (such as influx of internally displaced population) should have stronger negative feelings towards out-groups.  Context is defined as a bounded locality one is established in; it could be a village, town or a city. Local/social context is a subcategory, and it inheres in more proximate surroundings such as neighbourhoods or districts within a context. Social context is also integral to identity formations/transformations because it determines whether in-group/out-group categorization is instrumental or meaningful (Simon and Klandermans 2001). In the context of civil war, context also determines exposure to violence, likelihood of victimization, local forms of resistance (Balcells and Justino 2014). The seminal study of Gould (1995) on insurgent identities is an example. He contends that collective action capacity of working-class is predicated on residential location and neighbourhood network structures. Norms, culture, political identity and consequently social, political and economic opportunity structures (which altogether composes the character of a locality) influence that extent and possibility of activism, mobilization and civil society organizations in a context. For instance, being situated in a historically politically charged city may emphasize political identities, and strengthen group identities around the relevant salient social identity.  53   Social context may carry different memory frameworks about the conflict and the political history and state-society relationships. Most people acquire their “memory” about recent as well as distant past events through the second-hand dissemination of knowledge. The intergenerational transmission of memory through the stories told by elders is among the chief mechanisms (Harris 2006). Such transmissions are very constructive of identity in the early political socialization stage. Memory pertaining to one generation is transferred to the following by recounting the past and providing a “representation” of the past through the lens of their social identities, and that generation in turn transfers these representations to the subsequent one by bringing in their own interpretation.45 These memories formulate the local narratives, which may then infiltrate into the beliefs, threat perceptions and risks extending from the conflict.  Locals may be disturbed or threatened by influx of a new group because they pose threat to the social and economic order. In civil wars, the forcibly displaced migrants are highly likely to be labeled as “terrorists,” which aggravates the perception of threat. External threat from a known out-group (as observed in 9/11 attacks) or existential threat to the self, on identity basis, could be a factor that may elevate salience of a group identity. This may in turn intensify group attachments (Pyszczynski, Solomon, and Greenberg 2003) because group-level existential concerns affect group member’s emotions and behaviour for individuals derive both meaning and self-worth from their group memberships (Tajfel and Turner 1986). Putnam (2000) suggests that displacement and migration may lead to new conflicts between long-time residents and newcomers. Such tensions, feeling of threat because of the unknown, and association of the                                                 45 This idea also forms the basis of “social representations theory” originally generated by Moscovici (1984). According to this theory, knowledge of the past is always social and subjectively shaped and developed by interaction (Duveen 2000). Hence, accounts of the past are percolated through cultural frameworks and social identities. 54  migrants with the insurgency should lower generalized trust within society, and category-based trust if the migrants hold a common group identity.  2.5 The Link Between Political and Social Trust  The relationship between political and social trust is still subject to debate.46 Some accounts indicate that established political trust (in a democratic regime) is a strong predictor of a country’s interpersonal trust level (e.g. Jamal and Nooruddin 2010; Muller and Seligson 1994), and yet others show that the direction of effect runs more from trust in individuals to trust in government (e.g. Brehm and Rahn 1997; Keele 2007; see Kay and Johnston 2007 for a review and discussion). The possibility that they are mutually reinforcing remains strong as well (Hetherington and Rudolph 2008).47  My theory assumes that insecurities leading to changes in political trust impinge on social trust relations. Threat locals may feel from the forcibly displaced migrants is highly conditioned by the collective threat locals feel from the insurgency. The causal impulses for political and social relations are interconnected, and in the empirical chapters (Chapter 7 and 8) I take these interrelations into account.                                                  46 Three models exist: social-psychological model, social-cultural model, and institutional performance model (Newton and Norris 2000). Social-psychological model considers trust as a personality trait and posits that all types of trust are correlated. Socio-cultural model views trust as a product of socialization and predicts a positive correlation between the individual social capital and his/her trust levels. Institutional performance model downplays the role of personality and socialization, and assumes that political and social trust emanate from different sources so they are not necessarily correlated. Institutional performance drives trust in government and interpersonal relationships determine social trust. 47 Newton &Norris (2000), examining 17 trilateral democracies, contend that there is no strong association between social trust (interpersonal trust or horizontal trust) and institutional trust (also see Newton 2001).  55  2.6 Observable Implications of Political and Social Trust Declining political trust implies that institutions are not fulfilling their promises with respect to service and infrastructure provision. In the context of civil war, if political trust is declining, we expect to observe: 1) Citizens’ dissatisfaction with the security forces’ (police, gendarmerie, army, etc…) ability to protect; 2) Citizens’ increasing skepticism of the government’s counterterrorism policies; 3) Declining faith in the state’s policies to protect the interest of the citizens or the nation; 4) Citizens’ decreasing willingness to engage with the political representatives; 5) Citizens’ decreasing desire to pay taxes. Plummeting political trust may lead to casting doubt on the authority of state to govern. If that is the case, legitimacy is at stake. Observable implications of declining legitimacy, as I view it, are: 1) Increasing expression of skepticism on state’s rightfulness in governing; 2) Decreasing use of state’s system for justice; 3) Generation of alternative systems of security and leadership. Some of the observable implications for declining generalized trust are:  1) Developing a fear that unknown others will deliberately harm you instead of being neutral or positive about their intentions; 2) Starting to avoid interactions with unknown others even though that was not a practice before; 3) Taking extra measures for self-protection (locking doors, accompanying kids to school); 56  4) Develop an anxiety that your family’s safety is at risk when they are away from home. Some observable implications for declining category-based trust are: 1) Developing negative blanket judgments for members of a defined out-group and use of discriminatory language (regarding their character, trustworthiness, intentions, etc.); 2) Developing a fear that the out-group members are posing a risk to one’s interests and practicing in-group favouritism; 3) Increasing reluctance to engage with them in business/transaction settings; 4) Increasing reluctance to involve in family unions (e.g. not allowing the siblings to marry an out-group member). 2.7 Conclusion This chapter laid out the theory of political and social trust in civil wars. The major predictions of the theory are that ethnic and ideological wars have contrasting effects on political and social trust. While ethnic violence boosts political trust, ideological violence dampens it. Furthermore, whereas ideological violence decreases generalized interpersonal trust, ethnic violence does not affect it yet category-based out-group trust is undermined in ethnic violence. The theory suggests that these contrasting effects largely stem from the differences in collective identity threat posed by ethnic and ideological wars and state’s discursive responses. Empirical chapters that test the hypotheses of the theory (Chapter 7 and 8) provide a summary of the hypotheses and the corresponding observable implications (see Table 2 and Table 6).   57  Chapter 3: Research Design Establishing causal connections between violence and trust is challenging. Conflicts vary a lot in their scale, character, and level of impact, and hence they vary in the threat they pose. In addition, other factors that are exogenous to an ongoing conflict do not cease to influence political trust, so mapping the landscape of conflict onto observed changes in political trust can be a considerable challenge. As Ragin (2014, 26) suggests, “it is the combinatorial, and often complexly combinatorial, nature of social causation that makes the problem of identifying order-in-complexity demanding.” To begin to tackle this question, I employed mixed methodology combining the rigor of rich qualitative case studies and quantitative cross-national studies. I embarked upon my research with a theory distinguishing between ethnic territorial and ideological revolutionary wars. I designed my qualitative research to assess and update the specified mechanisms in my theory. I combine multiple empirical strategies to generate and test the different pieces of my theory. Primarily, I conducted case studies of the Kurdish insurgency in Turkey (1984–) and the Maoist insurgency in Peru (1980–1992), representing ethnic territorial and ideological conflicts, respectively. The similarities between these cases demonstrate the “match on these variables,” and could work as “controlling for” the role of these factors on political trust, plus help me partially sidestep the “few cases many variables” problem often observed in comparative case studies (Collier 1993). Even though the similarities helped me test some parts of my theory, because of significant differences, my case studies cannot be described as comparative method with most similar cases design. Finally, I tested the main hypotheses derived from my theory using quantitative data to see whether the theoretical arguments and the findings of the 58  qualitative research can travel beyond the boundaries of Turkey and Peru. Table 1 below summarizes the empirical strategies employed for theory building and theory testing parts.  This chapter starts with a description and justification of the case selection. I then move on to delineate the fieldwork and give details about the process of case studies and micro-level data collection including my strategies for sampling, recruitment, and participant observation for the purposes of reliability. Next, I describe the coding process and how I arrived at my inferences. In the Analysis section, I discuss the method of process tracing, and how I harnessed it to make inferences to help me identify the causal pathways between the conflict and trust, and to test and update my theory.      59  Table 1. Empirical Strategies for Different Parts of the Theory  Parts of Theory and Questions Questions and Empirical Strategy Causal Antecedents of Conflict Character  (How does nation-building/state-formation shape the character of conflict? Chapter 4)  Why was the conflict in Peru ideological while the one in Turkey ethnic in character?  Comparative Macro Historical Institutional Analysis of Turkey and Peru  (Inductive inferences from the political histories of the two countries to identify the causal antecedents of conflict character)  Generation of Collective Threat Framing Through State Discourses (Chapter 5) Is the collective threat framing different in Turkey from the one in Peru?     State Discourse as a Causal Mechanism to Influence Micro-Level Trust Formation  Case Studies of Turkey and Peru (Archival Work, Content Analysis of Newspapers, Secondary Literature) and Process Tracing to establish collective threat framing and discursive content in Turkey and Peru  Micro-level evidence (narratives from interviews and focus groups with people belonging to the politically represented/powerful ethnic group) that refer to discourses  Impact of Personal Threat as a Causal Mechanism How does personal security threat affect political and social trust?   Interviews and Focus Group with Individuals who vary in terms of their exposure to violence  Context-Contingent Effects Do the predicted effects of civil war on political and social trust vary by context? For instance, does the predicted negative effect of civil war on social trust magnify in contexts that had a major change in social composition because of internal displacement?   Subnational Cases—variation in terms of political mobilization, percentage of displaced people, proximity to clashes and political order was used as leverage Testing Micro-level Observable Implications  Do the hypothesized individual-level relationships between civil war and trust hold?  Narratives from Interviews and Focus groups Note: This table draws on Ana Arjona’s presentation of her framework in Rebelocracy (2016). 60  3.1 Case Selection: Why Turkey and Peru? The Kurdish insurgency in Turkey (1984–) and the Maoist insurgency in Peru (1980–2000) are two cases representative, respectively, of an ethnic territorial and ideological form of conflict). Turkey and Peru have seldom been studied in a comparative fashion in political science. This may not come as a surprise given their regional and historical differences, yet a closer look at their contemporary history and social structure would reveal that they are more similar than different. In fact, despite their diverse historical origins—Turkey being a post-imperial country descending from the Ottoman Empire and Peru being a postcolonial country—, they can host enough equivalence to render their contemporary legacies similar. The similarities provide me with a leverage to identify the distinctive features, which may explain the discursive differences and collective threat framing. I should underscore that the method employed here is not most similar designs, yet the similarities are essential to emphasize as they are useful in making comparisons. The differences allowed me to build parts of my theory, and within-country variations helped me test micro-level implications of my theory.  Turkey and Peru are multi ethnic countries, featuring one group dominating over the other. The subordinate groups enjoyed significantly different status vis-à-vis the empires they were ruled by: the Kurds in Ottoman Empire were granted autonomy and functioned as an independent principality while the indigenous people in Peru were practically slaves during the Spanish empire rule. The contemporary history of Kurds in Turkey and indigenous groups in Peru nevertheless show striking parallels. Policies of assimilation (mestizaje in Peru reflecting the entrenched racism, and Turkification in Turkey reflecting the overarching nationalism), endemic subjugation and repression of subordinate ethnic groups define the cornerstones in the histories of both countries. Though Turkey did not have as stratified and as long racial 61  taxonomies as Peru did, Turkification of Kurds is kindred to the idea of mestizaje – the idea of gradual evolution of indigenous people by rejecting and discarding their culture and language and taking on the dominant culture as proposed by mestizo elites in Peru (which never became a state policy due to disagreements upon the meaning of the notion). The tensions between Kurds vs. Turks in Turkey and Indians vs. mestizos in Peru have centered on association of the “weak” ethnic group with backwardness. The Kurds were peripheral, traditional and religious while Turkey is on route to modernity and securalism (Yeğen 2007). Indians in Peru, also traditional, were perceived to be barbaric while Peru advanced toward a modern and civilized country (Méndez 1996). 48 The statist and centralist economic development policies in Turkey rendered the regional development levels very uneven; while the West was advancing with industrial development and investments, the East (which has been predominantly occupied by Kurds) was lagging behind. The Peruvian state, in its origins (it became independent in 1826), was oligarchic; it was grounded upon a union of the wealthy and powerful elites across the country (Burt 2007). The state was practically serving the goal of maintaining and advancing economic interests of these elites, and the state-society relations were structured around this goal, and were hence pervasively clientelist (Stokes 1995). Economic development was also fairly imbalanced: While capitalism (mining being the principal industry) was swelling mostly on the coast and some other adjacent regions, other regions made no considerable economic progress and perpetuated the old feudal relations.                                                  48 In Peru, the historical injustices were on a larger scale simply because indigenous people comprised virtually half of the population while in Turkey Kurds have always been a minority (estimated to be around %15-%20 of the population) in their contemporary histories. 62  Hence, inequality between groups in both countries soared, buttressing the traditional ethnic and geographical divisions, which perpetuated the intersectionality of geography, class and race/ethnicity in Turkey and Peru. Access to basic state services such as education and health, and basic infrastructures (roads, water, electricity) has also been weak for the subordinate ethnic groups until recently. The Kurds in Turkey and the indigenous people in Peru were both treated as secondary citizens. Indeed, in Peru “Indians” who were not literate in Spanish were disenfranchised until 1979. On the other side, Kurds who constitutionally always had equal citizenship with Turks, had various restrictions placed on their identity, language and culture.  Against this backdrop, the seeds for both guerilla groups were planted during the surge of socialist movements in the late 60s and 1970s. The PKK and Sendero were kindred in their ideological orientation; both were offshoots of large communist parties. Their respective leaders, Abdullah Öcalan for the PKK and Abimael Gúzman for Sendero, were both prominent figures in the party, and were activists. They were both incarcerated by the state and developed their ideas for initiating an insurgency in prison. They both commenced the armed struggle in the early 1980s, while their respective countries were transitioning from a military rule to a civilian one. The PKK embraced a more ethnic doctrine and waged a secessionist war whereas Sendero was Maoist and adopted a class-based doctrine. That being said, intersectionality of class and ethnicity were visible in both cases. For example, “ethnic-regional” identity was the binding force for the middle and lower ranks of Sendero (Degregori 1990).  Besides these parallels in the guerilla groups and political histories, the countries display other similarities. Turkey and Peru experienced a democratic revolution from above and underwent multiple military interventions. Experiments with democracy have been interrupted by military coups in Turkey and caudillos in Peru.  63  The significant differences between Turkey and Peru can, however, obstruct the possibility of robust causal inferences. Peru is a postcolonial country while Turkey is postimperial. Strong elitism (creolismo), authoritarianism, and corporatism have featured in postcolonial Peruvian politics and nation-building (Palmer 1980), while in Turkey the keystones of the political culture have been a strong state tradition, nationalism, and secularism (see Chapter 4 for an extensive comparison between the political contexts of the two countries). These political differences are important to note, since they relate to the societal consequences of the two conflicts. I take these differences into account in my empirical discussion as well. 3.2 Focus of Analysis I gathered intensive case knowledge through extensive fieldwork. When comparing the two cases to determine how conflict affects trust, I have taken into account the historical processes, institutions, and social and political configurations of both the countries and the conflicts. I focus on the period between 1980 and 2000. 1980 was the year Sendero commenced its armed operations; even though the PKK did not start its attacks against the state until 1984, for comparison purposes I start my analysis in the same year in Turkey, since by that time it was already active and preparing a base for organized guerilla war. The year 2000 saw the political conjuncture change significantly in both countries. In Peru, Alberto Fujimori resigned in 2000. Also, even though Sendero was effectively dissolved when Abimael Gúzman was captured in 1992, many researchers extend their analysis till 2000, as it was not clear for a number of years afterward whether Sendero might resume. Also, the era between 1992 and 2000 coincides with the rise of Fujimori and saw the concoction of a memory of victory and glory, extending from the times when Sendero posed the highest risk to Peru between 1989 and 1992. In Turkey, Abdullah Öcalan, the founding leader of the PKK, was captured in 1999, and the Justice and 64  Development Party came to power in 2002. The power balance within domestic politics and vis-à-vis the PKK changed a lot following the imprisonment of Öcalan. Armed struggle paused between 1999 and 2004 and resumed more fiercely afterwards. Even though beliefs, attitudes, and behaviours of the society continue to change, the first stretch of the PKK’s guerilla war (1984–1999) established the major elements of the overarching discourse that has carried over to today.  3.3 Fieldwork and Data Collection I conducted six months of fieldwork in Peru between September 2013 and March 2014 and six months of fieldwork in Turkey between March 2014 and September 2014, with the support of International Development Research Agency Doctoral Research Award. There are five parts to the qualitative data I gathered: historiographic material, archival analysis of newspapers, focus groups, expert interviews, and in-depth interviews with ordinary people.  I mined secondary literature—including articles, books, diaries, reports, and master’s and doctoral theses—on insurgency, state, media representations of the conflict, and political trust in both contexts. I also went into newspaper archives in Lima and Ankara to see the discursive framework that the state and the media employed during the conflicts to describe the insurgencies. In doing that, I analyzed select issues of Milliyet and Hürriyet between 1984 and 2000 in Turkey (thanks to the availability of previous archival analysis on the issue (see Somer 2005) and select issues of La Republica between 1980 and 1992 (see Table A 1 in Appendix B for the months and years analyzed), and compared the visual representations of the conflict and conducted content analysis. Media representations of the conflict guided the discursive content in both contexts, and I used my archival analysis to construct the overarching and competing 65  discourses, the official and alternative narratives, over the course of the conflicts. I watched all the relevant movies and documentaries I was able to access.49 Using my case study materials, I generated the causal priors of my theory where I spell out how state formation practices and nation-building strategies may condition the nature of contemporary insurgencies as well as the responses to these insurgencies (See Chapter 4).  Micro-level data collection took place in stages. My first exploratory field trip in Turkey was in the summer of 2012. I conducted twelve interviews with experts who fell into one of three categories of people, which I defined as: 1) national and local politicians; 2) informed observers such as academics, journalists, and local notables; and 3) NGO representatives. I asked them questions about political culture and their perception of the conflict and its consequences. My exploratory field trip to Peru was in November of 2012. I met academics and researchers, established my first contacts, and identified feasible research networks. Subsequently, I conducted my second fieldwork between September 2013 and September 2014, spending 6 months in each country. Details of data collection are presented below and in Chapter 6. 3.3.1 Selecting Subnational Sites I picked my subnational cases in light of the requirements of my theory. First, to test the micro-level effects of interaction of collective threat and personal threat on trust as specified in my theory, variation in terms of personal threat was necessary. Second, to test the contextual effects as suggested in my theory, the cases need to vary with respect to exposure to violence, ethnic                                                 49 Some of the most important sources are the PKK’s own magazine, Serxwebun, Sendero’s media outlet El Diario, Hasan Cemal and Yalçın Küçük’s publications on the Kurdish issue, memoirs of soldiers, investigation reports of the commissions, and the report of Truth and Reconciliation Commission in Peru. The documentaries and movies I watched in Turkey were Bakur (2015), Kanun uykusu(2014), Atesten Tarih(2012), Nefes:Vatan Sağolsun (2009), and 5 nolu Cezaevi (2009). Major documentaries and movies I watched in Peru were Matar para vivir (2013), Milk of Sorrow (2009), Lucanamarca (2008), State of Fear (2006), Paloma del papel (2003), La vida es una sola (1993), and La Boca del Lobo (1988). 66  composition, political mobilization, and concentration of internally displaced people. Selected as guided by my theory, the subnational units helped me make within-case comparisons (Snyder 2001) as well as understand the role of social context and other contextual variables. I chose seven cities in each country as sites for gathering micro-level data: Ankara, İstanbul, Mersin, Diyarbakır, Mardin, Şanlıurfa, and Gaziantep in Turkey, and Lima, Ayacucho, Arequipa, Cajamarca, Cusco, Tarapoto, and Iquitos in Peru. These cities vary with respect to exposure to violence, socioeconomic development, ethnic composition, and the concentration of internally displaced people (see Figure 4 and Figure 5). In Peru, exposure to violence changed over the course of war in some localities (for example Sendero only reached Lima towards the end of the 1980s), the temporal variation also helped me test parts of the theory with respect to implications of personal threat in the absence of collective threat. Especially Lima being the heart of Peru, which varied temporally both in terms of presence of insurgents and internally displaced people was a perfect case. These variations and change in people’s trust formulations in the face of these changes provided me with leverage for hypothesis-testing. I spent close to a month in each city, except for Lima, where I spent close two and a half months, given its centrality to Peruvian politics and society.         67   Figure 4. Selected Subnational Sites in Turkey Note: The provinces chosen for fieldwork are marked by a black dot, where the size of the black dot indicates the percentage of Kurdish population (Ankara 8.50%, Mersin, 10.1%, İstanbul 13.8%, Gaziantep 15.4%, Mardin 75.8%, Diyarbakır 91.3%).50 The numbers below the province names indicate the ranking of the city in terms of socioeconomic development where İstanbul has the highest score in total of 81 provinces. The clash zones, represented primarily Şırnak and Hakkari are marked by red. The map is built with Tableau based on generated Longitude and Latitude.                                                 50 Official percentages of Kurdish population by province is not available in Turkey. The percentages here are estimated by the author using a survey data from 2006. Specifics of and details about the calculation are available in Appendix B. 68   Figure 5. Selected Subnational Sites in Peru Note: Marked cities were the chosen sites of fieldwork. Those marked with dark red were the main theatres of operation for Sendero. The marks are used to show the variation in proximity to violent zones.  The map is built with Tableau based on generated Longitude and Latitude.      69  3.3.2 Data Collection For micro-level original data collection in each site, I started with a focus group to explore the conflict-related themes that are of relevance to the locals and tailor my questions to the context. Focus groups are very instrumental in canvassing a broad range of experiences in a population as interactive group conversations bring a comprehensive array of opinions, experiences, attitudes, and context-specific incidences to light. Finally, focus groups helped me recruit interviewees. Some participants had a unique source of information or would tell a narrative that is disconfirming my theory. In such cases, in order to probe, 51 I would ask for a follow-up interview. Sometimes, after one focus group, I would start in-depth interviews, and then organize another focus group. Sometimes, if I need more collective information, I would conduct more focus groups before interviewing people on one-on-one basis. 3.3.2.1 Sampling and Data I used purposive sampling techniques when selecting my interviewees. “Purposive sampling” is defined as a type of sampling in which “particular settings, persons, or events are deliberately selected for the important information they can provide that cannot be gotten as well from other choices” (Maxwell 1997, 87). My selection criteria were determined by considerations of ethnic groups in the countries, level of victimization, amount of exposure to violence, age range, and gender, as I find these identifiers relevant to my theoretical constructs. I imposed approximate quotas for each category proportional to the size of the locality. My purpose was not only to gather systematic data to test my theory but also to diversify the inventory of subgroups in order                                                 51 Probing may be needed when a participant utters words about a potentially significant and or sensitive issue in an ambiguous fashion. If the researcher catches a cue in what a participant says or how he or she says it, the researcher may want to know more about what is referred to or the participant’s underlying thoughts. Hence, researchers may use the probing technique and gently ask further questions. 70  to access different viewpoints, make my sample more representative, and update my theory (LeCompte and Goetz 1982). My field observations thus guided me to make adjustments in my quotas. For instance, I found significant in-group variation within ethnic groups with respect to their accounts of conflict, even though these interviewees were are from the same context, so I added education level as a criteria and also interacted it personal experiences, ethnicity, and socioeconomic status to create new categories as I went on collecting more data. Lastly, I used theoretical saturation to determine the exact sample size, along with my own quotas. The saturation point occurred when the responses became repetitive and the marginal benefit of each additional interview started to decay fast. At that point, I ceased data collection.  I conducted phenomenological focus groups, which seek to understand the experiences of respondents. For compositions of focus groups, I used the geographical and ethnic divisions in my target populations as the main criteria and I also segregated the groups by gender and age to make sure participants are comfortable to speak about their experiences. I ran adult focus groups (25 and above) by separating men and women, and ran youth focus groups (18-25) in each city. The average size of the groups was 8 people. Recruitment strategies and the details pertinent to the conduct of the interviews and focus groups are discussed in Chapter 6. In total, I completed 36 expert interviews, 66 in-depth interviews, and 19 focus groups with ordinary people in Peru. In Turkey, I conducted 46 expert interviews, 92 in-depth interviews, and 15 focus groups with ordinary people. I also conducted a survey with every focus group participant to collect information on their political predispositions, which helped me draw a deeper understanding of the origins of and change in their political attitudes.  I conducted 12 expert interviews in Lima, 10 in Ayacucho, 6 in Cusco, and 5 in Cajamarca. I held 22 in-depth interviews with ordinary people in Lima, 12 in Ayacucho, 14 in 71  Cusco, 5 in Cajamarca, 8 in Arequipa, and 5 in Iquitos. The distribution of focus groups was as follows: 6 groups in Lima, 3 in Cusco, 3 in Cajamarca, 3 in Iquitos, 2 in Tarapoto, and 2 in Ayacucho. In those cities, I conducted minimum of one focus group with adult men only and one with adult women only. In those cities where I held more than two focus groups, I conducted one with youth (18-25), and, in some cases, one mixed-gender adult group. In the capital city of Lima, I conducted two with adult men, two with adult women, and two with youth. Except for the youth groups and the focus groups in Iquitos, each group was made up of a mix of different degrees of bystanders and victims. In Turkey, I conducted 15 focus groups: 3 in İstanbul, 4 in Ankara, 3 in Diyarbakır, 2 in Gaziantep, and 3 in Şanlıurfa. Focus groups were composed of 6 to 10 people varying in age and gender. Also, I completed 46 expert interviews and 92 in-depth interviews with ordinary people. The distribution of the expert interviews was as follows: 8 in İstanbul, 8 in Ankara, 10 in Diyarbakır, 8 in Şanlıurfa, 8 in Mardin, and 4 in Mersin. The distribution of the in-depth interviews with ordinary people was as follows: 12 in İstanbul, 15 in Ankara, 20 in Diyarbakır, 13 in Şanlıurfa, 12 in Mersin, 10 in Gaziantep, and 10 in Mardin. The length of my interviews varied between 45 minutes and 4 hours, averaging about 90 minutes. Focus groups were on average 2 hours long. 85 percent of all my conversations were audio-recorded; in the other fifteen percent of the subjects did not consent to being recorded (21 people in Turkey, and 13 people in Peru). Details of recruitment are presented in Chapter 6. 3.3.2.2 Conducting the Interviews and Focus Groups  All interviews and focus groups began with participants consenting to the protocol (usually oral consent was granted rather than written). At the beginning of every interview, I introduced myself as a Ph.D. candidate from Canada and underscored the independence of my research and 72  the purpose of it. I also reiterated that, as per the consent form, they had the right to choose what goes in the recording, end the interview at any moment, and reserve to right to decline answering any of my questions. Sometimes establishing rapport took longer than usual, and without rapport the answers tended to be brief and uninformative. For instance, with the Kurdish respondents, it was a bit hard to establish rapport sometimes because they would usually guess I am not Kurdish and would hesitate to open up. As Leech says, rapport is much more than putting the respondent at ease; “it means convincing people that you are listening, that you understand and are interested in what they are talking about, and that they should continue” (Leech 2002, 665). Making sure that no judgment or threat could be interpreted to be in my phrasing or body language when it came to sensitive questions, I sometimes consciously extended the initial part of the interview, where I ask nonthreatening or less sensitive questions. In order to make the interviewees feel comfortable, I would sometimes start with stories or a very general question, what Spradley (1979) calls “grand tour questions” such as “where they grew up and how the local life was back then” (as cited in Leech 2002, 667). I also changed the way I pronounced “PKK” depending on the subject in order to avoid causing any offense. The Kurdish pronunciation reads the letter “k” as “kay” while the Turkish pronunciation reads it as “ka” (as in “karate”), and, in today’s politics, the pronunciation of “PKK” pronunciation can be a simple cue to signal one’s side or camp.  I resorted to “soaking and poking” (heavy immersion in the details of a case) to get an insider perspective, familiarize myself with the world of the respondent, and ensure the content validity of my interviews, which, since they were based on open-ended questions, were sometimes lengthy and hard to replicate. In order to overcome this reliability issue, I increased 73  the size of my sample, fell back on my prompts quite often, and made sure to ask the same questions in every interview, even if in a different order. The participants had a range of experiences: some were ex-guerillas, many were victimized by war, some were discriminated against based on their ethnicity, some were forcibly displaced (and these experiences were not necessarily mutually exclusive), and some were not affected (these people were the main focus of my inquiry). The storylines and insights offered by the participants differed greatly, and in order to integrate all these diverse narratives, I carefully tailored my questions to each subject. To alleviate any discomfort around sensitive issues, I would break up my inquiry into segments that I scattered throughout the course of the interview, thus lessening the intensity of the questioning. At times, interviewees had extraordinary experiences or unique insights to share, and I would form new questions on the spot to follow-up on those cues. At other times, they would use a certain vernacular, and, when I was not sure what a particular word or phrase meant, I would follow up, ensuring equivalence of meaning between researcher and subject (LeCompte and Goetz 1982). The semi-structured format of the interview enabled me to revisit a question later if the respondent diverged from a cue that I had wanted to follow-up on.  I paid due diligence to the security concerns some respondents had regarding the expression of certain opinions. In Peru, where the conflict is long over, the participants did not indicate experiencing any tangible fear in expressing their views. In Turkey, however, because the conflict is still ongoing, extracting honest and unfiltered responses was more of a challenge. Misrepresentation of personal opinions out of concerns for one’s security is especially a risk in an ongoing civil war setting (Wood 2003). Nevertheless, 2014 was an exceptionally peaceful year to conduct a conflict research in Turkey, as the peace process was under way (2013–2015), 74  and, thanks to the armistice, the political situation was very calm. It was five years after the Kurdish Initiative, which extended rights and freedoms to Kurds starting in 2009, and the Kurdish participants were more comfortable than ever before to talk about “taboo” subjects such as their identities or sympathy for the PKK. They were also extremely eager to talk about the past and tell their stories, experiences, and sufferings. Both in Peru and in Turkey, many of my subjects mentioned at the end of the interview that “the interview was cathartic.”52 Still, obtaining genuine answers was sometimes an issue. Expression of support for the PKK and its justification seemed especially hard for participants to utter due to the ongoing nature of the conflict. When I sensed discomfort, I would either change the topic to ease the respondents’ way back into talking about the PKK or reformulate the question. Furthermore, as Wood (2003, 35) cautions: [T]he telling of personal and community stories in an ethnographic setting is also shaped by [….] [the respondent’s] present political loyalties, beliefs concerning the likely consequences of participation in the interview and of expressing particular views, and present political objectives.   Elected officials in Turkey were especially eager to give me the official state narrative without offering any additional useful input. In such cases, it required more prodding than average to get a personal response, as the subjects typically circumvented the question with politically correct rhetoric. In Peru, the public was more used to the practice of talking about the past, especially in conflict-ridden zones thanks to the Truth and Reconciliation Commission’s (CVR) work. The Commission sought to unearth the truth by drawing on almost seventeen                                                 52 During the course of my field research, I took every measure necessary to ensure the privacy, confidentiality, and security of my respondents as per my ethical obligation. I kept the audio recordings in a secure, password-protected bag, with the field notes carried with me at all times. Here I present quotes from the conversations without attribution or personal identifiers, again to respect the privacy of the participants and protect their anonymity. 75  thousand testimonies and construct a new historical narrative—an alternative to the official accounts (Milton 2007). In Peru, I had the advantage of both immense amount of data already gathered and the public’s comfort with talking about the past. My fieldwork in both countries was very immersive and rich; it was semi-ethnographic in nature. In Peru, I lived with local families except for my first two months in Lima, but even then I was hosted by many locals and carried on my research outside of the formal settings of focus group or structured interviews. I had numerous informal conversations with locals on politics throughout my time in the country. They shared their experiences and told me stories from their past pertaining to the times of conflict. In Turkey, I had more advantages being Turkish myself. I had the opportunity to be present during casual political conversations at dinner tables or on Sunday brunches. Even after my planned fieldwork was over, I kept in touch with some of my key contacts in both countries and continued to ask clarifying questions when needed. See Chapter 6 for further details. 76  Chapter 4: Causal Antecedents of the Theory: Comparative Historical Dynamics behind the Onset of the Insurgencies in Turkey and Peru In this chapter, I present my case study materials that led me to inductively generate a theory on nation-building and conflict character as a causal antecedent. Insurgencies do not emerge in a vacuum. The historical antecedents of political violence (prewar dynamics) are important to understand not only to make better sense of the onset of the conflict but also the consequences. Much of the conflict character, dynamics and consequences are buried in the social and political context that gave rise to the reasons explaining the onset of violence.  Inductively, I theorize that nation and nation-building strategies condition not only the nature of the insurgencies (conflict character) but also the responses of the states to the conflict (official discourse) and hence trust calculations of the individuals (see Chapter 2 for discourse as a causal mechanism defining collective threat). In other words, here, I argue that threat framing and conflict character are endogenous to the historical antecedents of the conflict (state-society relations, historical injustices and grievances extending from nation-building).  Turkey, a post-imperial country, was founded upon the principle of a nation with one supra identity (Turkishness). The success in the War of Independence fought against the allied powers granted the country a proud start, cultivating a strictly nationalist identity of the new Turkish Republic, which then seeped into political, public and private sphere, from language policies to dress codes, from military conscription to city names. From the early days of the Republic, this understanding of new modern Turkish nation with non-negotiable territorial boundaries had been ingrained in the discursive sphere, which is what the state later used to construct the official discourse against the insurgency, capitalizing on the tools it had been 77  planting (see Chapter 5). Importantly, the same policies and discourses featuring strong state and the Turkish identity also led to the emergence of ethnic insurgency, prompted by decades-long suppression of Kurdish identity, and also shaped the political and social attitudes of the people in the face of the war, as explained below. Peru, a post-colonial country (a post-viceroyalty specifically), underwent tumultuous state-formation. Strong elitism (creolismo), authoritarianism, and corporatism have featured in postcolonial Peruvian politics and nation-building (Palmer 1980), and national integration proved to be an arduous task. Centralist policies and entrenched racism, weakness of Peruvian national identity, lack of strong state both impeded a strong state and an overarching discourse about Peruvian state’s identity, and also led to the emergence of a Maoist insurgency. The insurgency did not pose collective identity threat to Peruvian national identity because it was not ethnic territorial and the Peruvian state did not have an ideological identity. The Peruvian government, under attack by Sendero, never successfully framed the insurgency as a collective identity threat, first because it did not have ideological tools to marshall, and second because of the issues embedded in nation-building (inferior status of the indigenous people and their being the main target initially interfered with the perception of threat (See Chapter 5).  Peru had the necessary conditions for an ethnic territorial insurgency to emerge in place, with a relatively large (indeed much larger than the Kurdish population in Turkey) marginalized ethnic group, however an ethnic territorial insurgency never broke out. The answer to this puzzle also lies in the colonial past and the post-colonial nation-building. Kurds enjoyed political and economic autonomy and maintained a privileged status during the Ottoman rule, and developed an exclusive identity and agency. The post-imperial Turkish state tried to marginalize it but never fully achieved. Indigenous peoples in Peru were enslaved by the Spaniards, which casted 78  the path for their inferior status in the Republic of Peru after independence. Even though they may revolt the elite, they neither had a cohesive identity, collective capital, or a coherent and consistent ethnic narrative to claim a more inclusive status, sometimes because of their very own elites’ conflicting positions.   Starting with the case of Turkey, the chapter lays out the political histories of the countries, nation-building strategies, state-society relationship with a particular focus on Kurds and indigenous groups respectively. Despite the similarity in their contemporary marginalization, the agency that Kurds were able to flourish as opposed to the relatively weak indigenous identity in Peru, explain much of the differences in the dynamics and outcomes of the insurgencies. The last section of this chapter explains how the historical grievances led to the rise of the insurgencies.  This chapter does not claim to do a deep comparative historical analysis of the two countries. Rather, it aspires to portray the remarkable parallels and variances in nation-building, center-periphery relations, political genealogy of Kurds and indigenous groups, the past and present social structure in both countries, and the onset of the insurgencies in order to lay the foundations of discourses, threat framing and identity frameworks—the causal mechanisms of my theory.   4.1 Foundations of Turkey: Modernization, Nationalism And One Nation The Turkish Republic was formed as a nation-state from the remnants of the Ottoman Empire in the wake of the losses in Balkan Wars (1913), World War I (1914-1918) and the gains of the War of Independence (1919-1922).53 “Nation” (millet) had a different connotation in the                                                 53 See (Lewis 1961) for an extensive coverage of emergence of modern Turkey. 79  Ottoman Empire where religion was the binder of communities. The Muslim population was, for instance, composed of different ethnic groups such as Turks, Kurds, Arabs, Circassians, Albanians and so on (Timur 1998). The founders, despite this multi-ethnic legacy of the Empire, believed that only with a homogenous post-Ottoman society that was united around a common affiliation to nation, the ideal of nation-state would be achieved. Religion could not have played that role as modernization and secularism were the overarching goals of the national movement. Turkish nationalism was proposed as a new binding force. The discourse of Turkish nationalism had indeed been expanding since the late 19th century Ottoman era. It evolved from an originally cultural and a linguistic movement to a strategy for political integration, supplanting Ottomanism54 (Yeğen 2007). The logic is of classic nationalism: citizens under the nation-state would be defined with historical, cultural and genealogical bonds and would speak the same language. Hence came the definition of Turkish citizenship that is above and beyond the ethnic and religious characteristic, intended to encompass everyone living under the rule of the Republic of Turkey, and supposed to serve as a primary identity.55 Turkism by “displacing both religion and rival nationalist formulations as the official ideology of the republic, [became] a vital tool in the process of Turkish state-building” (Hanioğlu 2006).56 Even though the identity project was presented as civic nationalism in its                                                 54 Ottomanism suggested that an “Ottoman” identity could unify all subjects of the Ottoman state differing in religion and ethnicity while Turkish nationalism posited that “Turkishness” is the only plausible ground on which a political unity can be built, which builds on the idea that Turks were the unsur-i asli (chief ethnic group) (Yeğen 2007).  55 See (Yeğen 2007) for further discussion of Kurdish question in Turkish nationalism. 56 The political cadres were replete with the supporters of Kemalist principles. The intraelite conflict between the more liberal and conservative elements among the revolutionary officials during the War of Independence ended with the victory of liberals by the mid-1920s. The unity of elites, composed of military, bureaucrats, intellectuals, party elites, was remarkable; they were deeply inspired by the six principles, and pervasively undertaking the mission of transforming and modernizing Turkish society (Ozbudun 1993). 80  character and aspirations, its implementation bore many references to ethnic nationalism.57 Along with Turkish nationalism modernization was embraced as an ultimate goal of the Republic. Indeed, the creation of a new identity and a homogenous nation was perceived as a precursor to the grand modernization project as uniformity was sine qua non for modernization reforms to be successful (Ergin 2008). 58 Many reforms were then undertaken as part of the modernization project; inter alia were the abolition of the Caliphate in 1924, the introduction of law of unification of education (Tevhid-i Tetrisad Kanunu), the adoption of Swiss Civil Law, and other reforms in wide ranging components of social life from clothing to measurement. These reforms were also intended for homogenization of the society towards a modern and secular one.  The founders had the attitude that Anatolia is made up of one single ethnicity (Mardin 1973), and turning a blind eye to the differences in society was necessary for building a nation (Toprak 2012); however, the reality was yet to assert itself. The founders were facing a grand challenge of imbuing the very heterogeneous society composed of non-Muslim minorities (as defined in Lausanne Treaty signed in 1923), 59 and different ethnicities – the majority of the                                                 57 Civic identity is derived from attachment to a common territory, citizenship, belief in the same political principles or ideology, respect for political institutions and enjoyment of equal political rights, and the will to be political part of the nation (Shulman 2002). Ethnic nationalism, in contrast, stems from people’s inherited ethnic characteristics including their language, religion, customs, and traditions (Ignatieff 1993).  58 The Republic of Turkey was founded with the aspiration of “reaching the highest level of contemporary civilizations,” which was embodied in the idea of “modernization.” Indeed, steps towards modernization had already been taken in the last era of the Ottoman Empire with the westernization movements. Yet, due to the reaction from the Islamic communities who were reluctant to accept changes in their educational and administrative system, modern institutions were added while preserving some of the traditional institutions such as medrese— school heavily oriented to religious studies. The juxtaposition of old and new was not embraced by Atatürk; he believed in a radical change affecting “all aspects of Turkish society, and sweep[ing] away most if not all, of its traditional beliefs and institutions (Okyar 1984).” Furthermore, Atatürk believed that theocratic Ottoman state with its different religions, culture and languages is not a suitable ground for moving the modernization steps forward. So, a newly established homogenous state to be united by the bonds of religion, language, and culture was not only auspicious but also indispensable for the goal of modernization.  59 When World War I was coming to an end, the Treaty of Sevres, signed in August 1920, provided for “local autonomy for the predominantly Kurdish areas” (Article 62), and in Article 64 even looked forward to the 81  ethnic minorities was Kurdish—with this new identity and modernized/westernized60 life style and system. These top-down reforms were hard to digest particularly for the ethnically heterogeneous Muslim population. There was clearly an underestimation of the entrenched cultural paradigms as well as the role of ethnicity in one’s identity.  4.1.1 Kurds in a Turkish Nation-State Kurds have lived in the geographic region stretching from upper Mesopotamia (referring roughly to contemporary south-Eastern Turkey, northern Iraq, north-eastern Syria) and to northern Iran. Tribalism defines their main organizing characteristic (Gunter 2004). Tribes (aşiret) are socio-political and economic units constructed by kinship bonds or common ancestry (see Jongerden 2007 for details). The aghas (feudal landlords or tribal chieftains) and sheikhs (religious leaders— the leader of the Sufi order) played major integrating roles in and across the tribes. Kurdish tribes, traditionally, were however subsumed under emirates /principalities, and prior to their absorption into the Ottoman Empire in the 16th century, they operated somewhat independently (Gavan 1958).  Some of the Kurdish territories were incorporated into the empire in the years 1514 to 1517. As per the administrative structure of the Ottoman empire, the Kurdish territories were divided into three provinces (eyalet): Diyarbakır, Raqqa and Mosul, which were further divided into districts. Old ruling families of Kurdistan (usually aghas) were assigned important                                                                                                                                                        possibility that “the Kurdish peoples” may be granted independence from Turkey. Turkey’s quick revival under Ataturk-ironically enough, with considerable Kurdish help, because the Turks promoted the theme of Islamic unity-altered the entire situation. The subsequent and definitive Treaty of Lausanne in July 1923 recognized the modern Republic of Turkey without any special provisions for Turkish Kurds. Minorities in the Treaty of Lausanne was defined exclusively in religious terms. Kurds, being Muslim, were never granted a status of minority. 60 Modernization went hand in hand with Westernization; they were perceived corollary of each other by the founders of the Republic espousing the Kemalist ideology. 82  administrative and political positions, and selection into such offices was hereditary, so only members of the ruling family could be elected. These local Kurdish rulers maintained their autonomy under the Ottoman rule thanks to the highly decentralized political and administrative structure of the Empire, which allowed communities (millet) to govern themselves.61  Kurdish principalities enjoyed greater autonomy than their non-Kurdish counterparts thanks to their role as a buffer between the Iranian and Ottoman Empire (Yeğen 1996).62 These peripheral Kurdish governments were, for the most part, exempt from tax collection, or military service in the Ottoman army unlike many other peripheries (van Bruinessen 1978).  When the Ottoman empire collapsed at the end of the World War I (WWI), Kurds were promised a scheme of local autonomy with a prospect of independence in the predominantly Kurdish areas by the Treaty of Sevres (1920). On the other side, the National Resistance Movement organized by the Ottoman Turks was promoting the idea of a new sovereign state for Ottoman Muslim populations (Lewis 1961). Even though Turkish national identity was strongly embraced in this movement, Kurds opted for fighting along with the Turks for independence in the post-WWI wars. The appeal of the movement for Kurds came from the facts that Islam was employed strongly in the rhetoric of the movement and that ethnic communities were promised local autonomy. Yet, once the war was won and the Republic was established (in 1923), the Kurds realized that self-rule of ethnic communities at the local level was off the agenda, and that                                                 61 The millet system acknowledged the pluralism within the peripheries, embraced the multi-ethnic multi- religious nature of them, accommodated this diversity, and enabled political and traditional differences. That being said, there were variations in the autonomy granted to peripheries; the terms of subjugation were usually determined at the time of annexation. For further information on administrative structure of the Ottoman empire, see (Barkey 2008, Karpat 2000). 62 Extended autonomy was not special and exclusive to the Kurdish principalities. There were many other privileged provinces which were not subject to the same rules of taxation or laws that other provinces were subject to such as: Eflak, Bogdan, Erdel (Transylvania), Crimea, Algeria, Tunisia, and etc. (Ortaylı 1979). 83  they had been simply absorbed in the new Turkish nation.  Finally, the centralization policy, which had commenced in the final era of the Ottoman Empire, dismantled the economic integration among Kurdish tribes. Once united by emirates under an empire, Kurdish tribes were now spread across three newly defined nation-states: Turkey, Iraq, and Syria. Each of these new entities sought to create an integrated national economy within its own territory and establish new economic centers. In the new structure of Turkish Republic, the new centers were İstanbul and Izmir, and the traditional centers of the Kurdish economy, Aleppo, Damascus and Baghdad, were now outside of the territory (Yeğen 2007). To resist these imposed nationalization processes, Kurds resorted to smuggling to perpetuate their economic integration.  In summary: The politics of Islam, the autonomous political structures of tradition, and the resistance of the ‘periphery’ to an integrated national economy were all the components of the constitution of Kurdishness. The constitution and exclusion of Kurdish identity was intrinsically related to the project of transforming an a-national, de-central and disintegrated political, administrative, and economic space into a national, central and integrated one. Indeed, the exclusion of Kurdish identity was an outcome of that project (Yeğen 1996, 226).  The idea of modern, secular, and Turkish nation governed by a central authority and a national economy was hence problematic on many levels.63 The salience of Islam in their culture and identity,64 the tribal structure they construct themselves in, and their perennial traditions of                                                 63 Kurdish principalities enjoyed greater autonomy than their non-Kurdish counterparts thanks to their role as a buffer between the Iranian and Ottoman Empire (Yeğen 1996). These peripheral Kurdish governments were, for the most part, exempt from tax collection, or military service in the Ottoman army unlike many other peripheries (van Bruinessen 1978).  64 Sheiks have been highly venerated in Kurdish circles. They have had absolute power over laymen. Sheik families in Kurdish politics were also prominent figures in modern Kurdish nationalism, which had been on the rise in the 84  governance and economy were all at odds with the national project of the new Republic.  4.1.2 Suppression of Kurdish Identity and Initial Kurdish Uprisings Two major developments in the early Republic sparked the onset of the Kurdish struggle. First, the constitution drafted in 1924 defined all the citizens as “Turkish” and did not mention other ethnicities. It stipulates that: “The people of Turkey, regardless of religion or ethnicity, is regarded as Turk in respect of citizenship.”65 In the same year, the institution of Caliphate was abolished. The caliphate was important for binding the multi-ethnic Muslim groups. It was especially for Muslim ethnic groups (particularly Kurds) in the periphery for it allowed space for local autonomy for the periphery through strong roles Sheiks had in the system (Yavuz 2001; Yeğen 2007). Kurds’ allegiance to the new Republic, in a sense, hinged on the institution of Caliphate and promises for local autonomy while for Ataturk and his loyalists, the Caliphate was the biggest impediment before a secular and modern regime, and local autonomy for ethnic groups was a threat to the national unity. Suppression of other identities than “Turkish,” and abrupt secularization resulted in a backlash from the Kurdish population. Kurdish uprisings commenced with the famous Sheikh Said in 1925.66 There are still debates regarding the nature of the rebellion; some suggest that it was religious in origin, some argue that it was a revolt to reclaim Kurdish identity, and yet others                                                                                                                                                        20th century. The sheiks embodied both religious and political power, pursued the goal of Kurdish autonomy, and preserved the tribal structure when clinging on the uniting power of Islam. They played a balancing role between Islam and Kurdish nationalism and autonomy (Tucker 2013). In tarikats (religious sect) and tekiyyes (dervish lodges), sheiks not only taught religious principles but also cultivated nationalist ideas (Olson 2013, Yeğen 2007).  65 The first clause of the Article 88 of the 1924 Constitution states: “Türkiye ahalisine din ve ırk farkı olmaksızın vatandaşlık itibariyle Türk itlak olunur.” 66 This was the first organized insurrection against the newly formed Turkish Republic. However, Kurdish revolts date back to Ottoman era. Sheikh Ubeydullah’s uprisal in 1870 is known as the first nationalist revolt. There were also other small-scale rebellions at the level of tribes. 85  contend that both were at play (van Bruinessen 1978). Whatever the true incendiary cause was, it was a reaction to the project of establishing a modern nation-state by imposing new identities and new life styles. The first attempt to revolt against the state resulted in heavy crackdown by the military forces and execution of the leader.  The Turkish state, determined to obviate any further similar insurgencies, took some strict measures, the reverberations of which persists today. The East Reform Plan, Şark Islahat Planı, being the most potent of all, was the first official step to assimilate Kurds, and “Turkify” (Türkleştirme) the Kurdish region via means of education, transportation, restructuration of cities, and new administrative regulations (Yeğen 2009). The plan, practically speaking, formed the framework of the state policies towards the Kurdish issue, the gist of which was denial and persecution of the Kurdish identity.  Prior to full operationalization of the Reform Plan, many follow-up Kurdish uprisings transpired in the following decade, varying in scope. The second major uprising occurred in Dersim, in response to the notorious Resettlement Law of 1934, passed to induce cultural homogeneity. It ended with a big blow from the state forces to strengthen the state authority and break the extant feudal ties.67 In 1935, a law (No. 2884) was passed to dismantle the tribal structure of the area, to “civilize” the inhabitants, and to change the name of the region from Dersim to a Turkish one: Tunceli.  Restrictions on the use of Kurdish were part of the strategy towards the obliteration of Kurdish identity, which was considered to be at the root of these insurrections. Soon after the adoption of the Latin alphabet in 1928, the government heralded a campaign to dissuade people                                                 67 The extent of the state killings was perceived as a “massacre” and culminated in genocide controversy.  86  to use any other language but Turkish with a slogan of “Citizen, Speak Turkish” (Vatandaş, Türkçe Konuş). In 1944, Law No. 7267 stipulated that “village names that are not Turkish and give rise to confusion are to be changed in the shortest possible time by the Interior Ministry after receiving the opinion of the Provincial Permanent Committee” (Jongerden 2009, 10). As a consequence of this policy, between 1940 and 2000, the names of more than 12,000 villages mostly in the Eastern region. Approximately one third of all villages were changed to a Turkish name (see Tunçel 2000 for the full list of renamed villages).  With the liberation of the political space in the 1960s and increasing expressions of Kurdish identity (see Chapter 5), further measures were taken towards the goal of suppressing the Kurdish identity. In 1961, Law No. 298 forbid the use “any other language or script than Turkish in propaganda disseminated in radio or television as well as in other election propaganda” (Article 58). With reference to this law, many prominent Kurdish intellectuals critical of state policies were arrested and resettled. Analogously, the Turkish Workers’ Party (Türkiye İşçi Partisi), which was voicing concerns over the ban of Kurdish language, was banned on pretext of “encouraging separatist activities” in 1971 (Gunter 1990,17). In 1983, a new and more comprehensive legal prohibition was introduced by the military regime on the use of Kurdish language with Law No. 2932. The prohibition came into effect a month before they turned the power back to a civilian government, supplementing the new constitution (1982) which further engraved the nationalist elements of Turkish identity and state.68 This new law forbade the use of any language but Turkish “as a mother tongue,” and banned publishing in any language other than Turkish, and it stayed in effect till 1991.                                                  68  For the Turkish original of the 1982 Constitution and a description of all the amendments to date see (Retrieved on May 5, 2017): http://www.anayasa.gen.tr/1982constitution.htm  87  In summary, the draconian rule of the state to ensure cultural and national homogeneity, fostering a culture of fear and relative conformism in the Eastern part prevailed through the 1990s.69 Stipulations of displacement and forced resettlement to rearrange the demographic structure of Turkey and “dilute” the concentration of Kurds, and the de facto and de jure language bans were the most blatant strategies to the assimilation end, which were interpreted as “an attack on the social space where-in Kurdishness is constituted” (Yeğen 2007, 226).70 The atmosphere of fear further induced de-politicization of Kurds. The generation of Kurds that has witnessed the cruel suppression of dissent was conditioned to “behave.” The Turkish state resorting to its military power established itself as the only authority that makes the rules of the game. Many Kurds truly embraced the imposed national identity of the state and abided by the rules. They also raised their kids with the same ideology and stayed away from the politicized groups. The few decades following the brutal suppression of Dersim uprising were quiet in terms of identity rebellions. Until the burgeoning of leftist movements for equality, rights and freedom in the late 1960s, there was no noteworthy legal or illegal organized Kurdish movement. 4.2 Foundations of Peru: Colonial heritage and Nation-Building Efforts “There is a history of conflict, both racial and geographical, which can be   explained  by one or more of the following: urban-rural, coastal- interior, center-periphery, and White-Mestizo-Indian” (#E32, Male, Lima, February 7, 2014).  Peru proclaimed its independence from the Spanish Empire, transitioning from colonial viceroyalty to a postcolonial republic in 1821. From the conquest of the Inca Empire by the Spaniards in 1532 onwards, Peru occupied a pre-eminent place in the Spanish Empire. All                                                 69 See Celik (2004) for detailed explanations on small-scale Kurdish mobilizations after the Seikh Said Rebellion.  70 For further details of the Plan, see (Bayrak 2009). 88  Spanish colonies in South America were governed by the Viceroyalty of Peru until the eighteenth century.71 Being a Viceroyalty, Peru had significant number of proud royalists to the colonial regime, and it was indeed the last country to proclaim independence in South America. The liberation from Spanish rule was brought by exogenous forces, and the resistance and reluctance of the royalists as well as the fragmentation in the population stalled the establishment of an independent country initially and bred instability throughout the early years of post-independence Peru (Contreras and Cueto 1999).  Nation-building in Peru and the search for Peruvian national identity were arduous given the strictly aristocratic state elites within an ethnically fragmented polity.72 The incompatibility between presence of the ideals of enlightenment, liberalism and modernization, the founding principles of the Republic, and the colonial legacy with a society based on caste system and an oligarchic state posed difficulties for state formation. The chasm between the ruling mestizo (mixed race) oligarchy in the center and subordinated indigenous publics (called “Indian” in a derogatory way by the creoles) in the peripheries was a strong defining feature of the colonial chapter of Peru, which spilled over to the post-independence era. Indigenous people who constituted the heavy majority in Peru were subjugated by the criollos (creole ruling aristocratic elites) throughout the colonial period. Although the nobility descending from Inca times were assigned aristocratic status, the indigenous local elite in charge of tax collection, they were few in number (Thorp and Paredes 2010).                                                  71 Peru, as a viceroyalty like Bolivia or Mexico, was one of the areas where power was centralized in contrast to “peripheries” such as Chile or Venezuela. 72 See (Galindo 1987) for the early years of the Republic. 89  Indigenous aristocrats occasionally challenged the Spanish oligarchic rule with small-scale uprisings, but these were to no avail until late 18th century.73 The counter-hegemonic rebellion of Túpac Amaru in 1781 against the Bourbon reforms, which proved to be the largest and most influential movement, had immense effects on the criollo elite, but not in the direction the insurgents were hoping for. The massive scale of Túpac Amaru’s revolt (also known as the Great Rebellion), though defeated and brutally repressed, left a deep mark on creole memory and perception of the Indians. Having lost their trust in the indigenous peoples, the colonial state gradually wiped out indigenous aristocracy lest an incident of a similar kind repeat itself. This insecurity of the ruling elites vis-à-vis Indians and their fear of being dominated by them would determine their attitude for many years to come. Méndez (1996, 220) posits that “the disdain towards and unfavourable image of the Indian grew together with the fears of an ‘outburst’ and the resulting need for the subordination of these populations.” One essential benchmark revealing this attitude was the Peru-Bolivia Confederation between 1836 and 1839, originated from the historical unity and commercial ties between two countries in the pre-Hispanic era. The creation was led by Marshall Santa Cruz, who was indigenous in ethnicity. Though celebrated by elites in southern Peru, the creation was not welcome by all Peruvian creoles, especially the commercial elites in Lima and the coast at large (Larson 2004). Governance under the leadership and persona of Santa Cruz symbolized subordination for the creoles and was disconcerting for the aristocratic creoles of Lima. The discourse against the Confederation embodied elements of the overarching disdain towards the Indian. In the words of Méndez (1996, 206), “[t]he most outstanding feature of the anti-                                                73 O’Phelan (1985) argues that these rebellions emanate from renewed dialects of freedom and renewed control, and identifies three waves : 1726-37, 1751-56 and from 1777 on which culminated in the Túpac Amaru II Revolt. 90  Santacrucista political discourse was precisely the definition of Peruvian national identity on the basis of an exclusion of, and contempt towards, the Indian, symbolically represented by Santa Cruz.”  Following the downfall of short-lived Confederation,74 consolidation of Peru became a priority for the state. The composition of the oligarchy also changed during these early years of the Republic. The old aristocracy (colonial nobility), decimated in the internal wars, was replaced by military and a small but wealthy powerful civilian commercial elite who emerged through guano trade and agriculture. Integration of the Indians into the modern economy and “civilizing” the “Indian” sierra were the major goals of this new plutocracy (Rodriguez Montoya 1921). The first “civilista” initiative was undertaken under Manuel Pardo’s rule (1872-6). Railways to close the gap between the center and peripheries, introduction of modern political economy, disruption of feudal labour relations, educating the Indians, and European immigration were among the highlights of the de-Indianization strategy of the Peruvian oligarchy (Davies 1974; Larson 2004). The modernizing state embracing liberal values envisioned a unified Hispanized nation, and mestizaje,75 in a sense, was employed as a nation-building tool, which engendered these reforms to accelerate the assimilation of indigenous cultures; thus recognition of different ethnic and racial identities has been problematic in Peru (De la Cadena 2000).  The interest and desire of the centralist elite and the regional elites were not always compatible with respect to the Indian issue. Some of the integration policies directed towards Indians were frowned upon by the entrenched landed elites (hacendados) and actively resisted                                                 74 See (Méndez 1996)for a thorough discussion on the discourse of Confederation. 75 Mestizaje is the idea of gradual evolution of Indians by rejecting and discarding their culture and language and taking on the dominant culture (De la Cadena 2000). 91  (see the next section for details on the land tenure system). Indian literacy and education, for example, were perceived as threats to the social and economic system of the sierra and in contrast to the interest of the hacendados, who benefited from being bilingual as overlords over Quecha-speaking indigenas. Most of them did not let their indigenous peasants get schooling. These types of local resistance and lack of a coherent policy to diagnose and tackle the “Indian problem” hampered and retarded nation-building in Peru. The War of the Pacific (1879-1883) that was fought among Peru, Chile and Bolivia over the resource-rich Atacama Desert (Guano and nitrate were hot commodities in the Peruvian political economy at the time76)—would aggravate the Indian problem for the Peruvian state. Not only did Peru lose devastatingly to Chile but also factional divide within the Peruvian elites resulted in a civil war and then numerous indigenous uprisings ensued right after. Though indigenous peoples fought in the war against the Chilean army by forming resistance armies in the highlands, indigenous peoples were portrayed as the reason behind the failure retrospectively.77  Landowners’ fear of peasant empowerment and invasion of haciendas by indigenous fighters was critical in this portrayal. The following excerpt succinctly states the prevailing view of the time:  The principal cause of the great defeat is that the majority of Peru is composed of a wretched and degraded race that we once attempted to dignify and ennoble. The Indian lacks a patriotic sense; he is a born enemy of the white and of the man of the coast. It makes no difference to him whether he is a Chilean or a Turk (Larson 2004, 196).                                                 76 See (Rodriguez Montoya 1921)for comprehensive study of guano’s role in Peruvian economic history. 77 See (Larson 2004, 178-187) for a detailed narration of the involvement of peasant-organized guerilla bands, and the question of whether indigenous people were “patriots or barbarians?” 92   Losing the war attested to the failure of Limañean (from Lima) oligarchy in nation-building. The rhetoric of indigenous as “traitors” also further tainted the image of indigenous peoples in the eyes of the creoles (Thorp and Paredes 2010), epitomizing the “Hispanista” discourse. Though granted equal citizenship in the Constitution of 1823, indigenous people were stripped of franchise after the War of the Pacific,78 and remained disenfranchised until 1979 by the Spanish literacy requirement. In summary, the indigenous peoples were effectively excluded from the “national projects” of the nineteenth century (Nugent 1994). In response to the racist “Hispanista” discourse, there arose an “indigenismo” movement. Indigenismo is a post-colonial discursive movement against the historical subjugation of indigenous peoples, their exclusion from the “imagined communities” during the nation-building, and the racism of the Peruvian oligarchy. In Peru, it was originally spearheaded by Manuel Gonzáles Prada in the late 19th century. Incorporation of the Indian masses, recognition of their rights and promotion of their interests were the main thrust of the movement. In the early 20th century, projects of indigenismo started to flourish, especially during Augusto Leguía’s rule.79 Though much action was taken in an attempt to recognize Peru’s indigenous communities, especially in the years of President Leguía,80 the undertones of the indigenista view again was                                                 78 Electoral law of 1895 restricted the vote to the literate population. 79 The indigenista discourse was not univocal and had internal divisions. Early twentieth-century debates over the role of the “Indian” in Peruvian society, and over the merits and desirability of biological and cultural mestizaje (mixing), divided Peruvian indigenistas into two major camps, those who equated mestizaje with degeneration (Luis E. Valcarcel’s camp), and those who embraced and advocated it as a positive social force (Jose Uriel García) (Devine 1999). 80 In the 1920 constitution, indigenous communities were officially recognized. In 1921, Leguía founded the Patronato de la Raza Indigena  (Guardianship of the Indian Race) to protect indigenous campesinos by investigating their complaints, enforcing their rights, and aiding their education (Heilman 2010).  93  colored by perception of inferiority of the indigenous race and the need to whiten them.  Between 1895 and 1919, the so-called “Aristocratic Republic” under the Civilista party once again ruled the country. Though they managed to reinstitute the national political order, exclusion of the masses and especially the indigenous peasants expanded the rift between the center and the peripheries. The center exclusively represented the commercial interests of the coastal elites and some mine owners (Miller 1982). The regional elites, especially those from the sierra (hacendados or merchants) were more marginalized yet connected to the central government through clientelistic ties (Nugent 1994). The state became increasingly centralized around Lima with acceleration of capitalist development and under the influence of centralizing military and bureaucracy and grew more distant to the sierra. The failure to incorporate the mass population in the economic system, to represent the common interest and to respond to popular demands further alienated the indigenous people from affiliating themselves with the state. In summary, the centralized government in Lima could not manage to unify the fragmented population but rather aggravated the existing cleavages. The concentration of the polity and economy on Lima and the coast went hand in hand with geographic, ethnic and economic marginalization of the sierra as well as political exclusion. The cultural and ethnic divisions intersecting with social classes persisted to the benefit of the dominant classes and to the detriment of a collective Peruvian identity (Cotler 1978).  Few initiatives emerged in the political arena that sought transformation of the country. The American Popular Revolutionary Alliance Party (APRA), under the leadership of Haya de la Torre, and the Communist party, under the leadership of Mariátegui in the early 20th century were founded. Political mobilization sharply increased in 1930s as economic crisis hit the country. Military forces took over the government in 1968 with a strictly populist and reformist 94  agenda to eliminate the “external dependence and internal domination,” causes of “national disunity” and of the hostility between the people and the armed forces, launching the so-called “Peruvian experiment” (Cotler 1983, 19).81 Yet, none of the attempts brought a sustainable solution to the entrenched inequalities and poverty of the indigenous groups. 4.2.1 Brief History of Indigenous People in the Social and Political Structure The natives of the geopolitical space where the Viceroyalty of Peru was established82 had been inhabiting the coastal and rainforest regions that were previously home to pre-Colombian civilizations (Dueñas 2010). Though heterogeneous in their ethnicities and cultures, Spanish colonizers classified them all under the single “Indian” category, which they perceived to be inferior to Spaniards. The “superiority” of Spaniards most likely originated from their ruler status over the colonized subjects and their Catholicism rendered them closer to God. A parable from the sixteenth century comparing Europe and America as two sisters conveys this understanding succinctly: “The first [Europe] is beautiful and gracious and rapidly receives a visit from Christ who weds her soul. The second has to compensate for her ugliness and country manners by offerings of mountains of gold and silver to tempt possession” (Thorp and Paredes 2010, 91).  Being “intrinsically inferior,” the native people were relegated to a subordinate position by the Spaniard colonizers in economic and social life. The colonial economic structure was agrarian and built on a system of economic exploitation whereby landowners and mineowners                                                 81 Retrospectively, this era of military government is evaluated under two phases: The Velasco regime during which reforms in land, health care, income distribution, and education took place, and Bermudez regime who ousted Velasco and dismantled many of Velasco’s reforms. Agrarian reform to end the feudal relationship between the landowners and peasants (hacendados and campesinos) and state’s increasing role in the economy are the only two that endured in the second phase (McClintock and Lowenthal 1983). 82 Broadly speaking Andeans and the Amazonians were the two major indigenous groups. The Andeans, semantically refer to the natives people of Andean highlands, were the major indigenous group at the time of the arrival of the Spanish conquistadors in the region, composed of native Quechua- and Aymara-speaking peoples. 95  were criollos and peasant labour were the Indios (Indian), who composed the highest and lowest social class respectively (Stavenhagen 1992). “Creole” denoted people of European (principally Spanish) ancestry who were born in Peru (Thorp and Paredes 2010). “Mestizos” were the middle category in the colonial caste system, referring to a mixed race— one parent of indigenous origin and one of Spanish or creole origin.  The system of economic exploitation induced and perpetuated a hierarchical social structure (a tripartite society), hindering possibilities of social mobility (Thorp and Paredes 2010). That being said, not all indigenous peoples were equally subjugated. Spaniards recognized the distinction of Inca nobility, along with their skills for political organization, and promoted them to a higher status in the class system and deployed them for the purposes of exploitation until the Túpac Amaru Rebellion in 1781 (Méndez 1996). The indigenous aristocracy (caciques) was in charge of collecting taxes, delivering the mita (labour force provision for the mines) and enjoyed rights to land and privileges to use of labour force (O'Phelan 1985).83  Mestizos, initially small in numbers, grew considerably as the expanding colonial economy brought about urbanization and migration. With the centralization in administration brought by Bourbon reforms, they started to assume roles as small town mayors, tax collectors and postmasters. More occupational space opened up for them in the hierarchy following the defeat of Túpac Amaru revolt and the gradual destruction of indigenous elites (Thorp and Paredes 2010). The indelible memory of the massive Túpac Amaru revolt, as well as the inherited                                                 83 See (Thorp and Paredes 2010) for further information on the process of colonization of the indigenous population. 96  contempt for the Indians from the colonial state, shaped the creole vision of the Indian in the post-independence era and reflected on the nation-building strategies, as discussed in the preceding section. Méndez (1996, 221) asserts that “[t]he ideas of the Enlightenment, with its zeal for classification, hierarchies and control, probably helped to mould the creoles’ new perception of the Indians, paving the way for that theoretical rationalization of fears which were the product of an unquestionably decisive historical experience.” Initially, emancipationist Creole elites wanted to construct a unified and rather homogeneous nation by eliminating the colonial caste system and bringing in equality under law. To the purpose of decolonization, the term “Indian” was disbanded in the official state discourse and supplanted by “indigene” and the concept of “indigenous race” was coined as part of decolonization and state formation efforts in the early nineteenth-century Peru.84 However, the dependence of the economy on Indian tributes in the first half of the nineteenth century reproduced the colonial domination of indigenous peasantry in a new formulation (Thurner 1997). The state made a pact with the indigenous community: in exchange for the head tax, the state was to protect Indian land by restraining its privatization and assign an indigenous authority to mediate their relationship with the state. This semi-autonomy conferred upon the indigenous communities was not serving to the ultimate goal of nation-building yet it was enjoyed by the local Indian republics as this new pact brought along rights over land that they did not have in the colonial era (Thurner 1997). Yet, in terms of the goals of decolonization and equal citizenship, this pact was detrimental as it concocted a subaltern indigenous citizenship, carrying over the “dual republic” of the colonial era.                                                 84 In nineteenth-century everyday parlance, the term Indian was employed to address to common rural folk by non-Indians, and not urban Andean elites (Thurner 2003). 97   The life of this pact was very short in the end. The spike in guano sales in mid-nineteenth century resulted in replacement of guano with the Indian tributes as the primary source of revenue for the state, which practically ended the pact on the part of the state. The Peruvian state ceased to be the protector of the Indian land, and allowed its privatization, which undermined legitimate republican rule in the eyes of the indigenous communities. The state’s levying Indian tribute once again in 1879 to cover war expenses was then considered illegitimate and paved the way for a new round of indigenous uprisings (e.g. Atusparia uprising), which were again brutally repressed (Thurner 1997). As a result, illiterate indigenous people were eventually stripped from enfranchisement until 1979.    With the fast growth of the coastal economy (light manufacturing, steel production and fish-meal processing) and promotion of Lima as the political and economic center, the sierra and more so the selva (the Amazons) was consigned to “malign state neglect” (Heilman 2010, 9). While the coast developed with market capitalism and free-trade liberalism and became modernized, the sierra lagged behind, maintained traditional feudal economic structures, and perpetuated agrarian class relations.   The institution of “gamonalismo,” referring to domination by local power holders (petty hacendados (landowner), gamonales (provincial authorities) and rural caudillos),85 perpetuated the colonial system of exploitation, and fostered clientelism. The exploitative relationship is succinctly characterized in the following quote:                                                 85 The term “gamonalismo” encompasses a broad range of local power relations in the sierra to “define the parameters of local power, based on monopolistic forms of wealth [landholding, mercantile monopoly, access to servile labour, military bands, etc.] and the privatization of provincial power [control over local offices, judicial processes, etc.]” (Larson 2004, 164). The local power holders may be mestizo or cholo (people of Indian ancestory who is acculturated in mestizo lifestyle) (Larson 2004). See (Manrique 1991) for further information on “gamonales.” 98  [T]he hacendado requires more of the Indian than just a percentage of his crop. [...] The most widespread form of service is the pongaje whereby the Indians and their families are required to work free in the hacienda household as maids, butlers, chauffeurs, cooks and general handymen. The Indian receives no payment for these services (Davies 1974, 12).   The now double-layered subordination of indigenous communities, both by the central and the local power structures, aggravated their plight. Against this advance of subjugation, everything from expansion of commercial haciendas and state neglect to complaints about abusive local authorities, indigenous revolts proliferated throughout all the regions of the sierra (Cusco, Puno, Arequipa, Ayacucho etc.). The revolts culminated in the massive peasant movement of Rumi Maqui (Stone Hand) in 1915, which sought to restore the Inca Empire (Heilman 2010).  The indigenista movement described in the previous section was inspired by the Rumi Maqui rebellion (De la Cadena 2000; Heilman 2010). Indigenismo was replete with significant dissidences as to how to solve the “Indian problem,” and thus inconsistent policy attempts followed. Many assimilationist efforts failed. In the words of Devine: [I]ndigenous culture has not been erased through education, changing modes of production, geographic relocation, or superficial alterations of physical appearance, as had been suggested by the wide variety of ‘assimilationist’ indigenistas who have sought to transform ‘Indians’ into what they considered to be ‘acceptable’ citizens, a process through which White and mestizo society would ultimately absorb and do away with Peru’s Indianness (1999, 71).  Pervasive racism extending from the creole nationalism and certain strands of indigenismo perhaps was the only persistent element of Peruvian state discourse.   Indigenistas were not a uniform group; they perceived and interpreted the reality and status of Indians in distinct ways and embraced different solutions for their national integration. On one end of the spectrum lies the position that views Indians as inherently good and 99  unadulterated and seeks to preserve this virtue by keeping them separate. Luis E. Valcarcel spearheaded this school of thought. He believed mestizaje is tantamount to degeneration of Indians, cholification.86 The opposite end of the spectrum hosts the idea that advocates for mestizaje and gradual elimination of the Indian race through acculturation and biological blending (Devine 1999). José Uriel García was a major proponent of this approach. Marxist indigenistas, closer to this end of the spectrum, perceived Indians as primarily peasants and they too saw their redemption in the de-emphasis of Indianness as a distinct culture and promotion of class consciousness. What unites indigenistas is the understanding that “Indian identity” and “national identity” are inherently incompatible.  This concocted barrier and the fundamental schism among the indigenistas caused Peruvian state’s policies with respect to the “Indian problem” to be inconsistent. In the 1940s, Valcarcel was the Minister of Education and had adopted policies to shield the indigenous culture by emphasizing their distinctness, and to maintain two distinct nations within Peru. Velasco’s military government, however, overturned Valcarcel’s approach, and pushed to bury ethnic identities and instead adopt class identities. The confusion on the part of the Indians is perfectly summed up by Devine in the following quote: After having been taught by a paternalistic State for a quarter of a century that they were Indians, and that their only salvation was in maintaining their ‘pure’ Indianness, Peru’s rural inhabitants under Velasco’s regime were now told that the State had made a grave mistake in promoting that false identity. No, they were not really Indians (or if                                                 86 Cholo is a defamatory term to denote cultural hybrids (mestizos) who have roots in the sierra and behave like a white person; it is loaded with judgment. Cholification refers to the process of Indian “degeneration.” I should underscore that racial taxonomies in Peru are vague and a bit arbitrary. One mixed race person (with identical phenotype), depending on the cultural and educational attributions (not the biological), could be Indian, mestizo or cholo (contingent on the lens of the reader) or white (Devine 1999). The availability of upward mobility through education goes against the racialized geographies, rendering these categories all the more complex. 100  they were, they should not be); they had been peasants all along, and just had not realized it (Devine 1999, 70).  Presuming a class-based system by refraining from race and culture related taxonomies, the Velasco regime attempted to solve the “Indian problem” by land reform. The military government also implemented educational policies, which paradoxically gave recognition to the indigenous languages. However, these efforts were not enough to contain the revolutionary movement of Sendero, as will be explained in the next chapter. 4.3 Conclusion The postcolonial Peruvian nation-making not only failed to incorporate indigenous groups but aggravated their situation due to inconsistent policies. As the criollo elites consolidated their power, the institution of gamonalismo became entrenched, and the plight of the indigenous peoples continued (Stavenhagen 1992). It is these ingrained structures in Peruvian national building that gave rise to the insurgency and made it ideological rather than ethnic. Unlike Kurds in Turkey, indigenous peoples also lacked collective capital. Despite some major revolts, agency in initiating change has been missing in Peru.  I believe the underlying reason for the lack of collective agency on the part of the indigenous peoples is the subjugation and subservience they had been subjected to after the Spanish conquest. While Kurds enjoyed political and economic autonomy and maintained a privileged status during the Ottoman rule, indigenous peoples in Peru were enslaved by the Spaniards, which casted the path for their inferior status in the Republic of Peru after independence. The Turkish state perhaps managed to successfully integrate a great number of Kurds and assimilate some others, yet Kurds maintained their collective identity that they had preserved and further cultivated thanks to the privileged status they enjoyed during the Ottoman 101  Empire.  When it was time to revolt against the decades-long suppression of Kurdish existence in the Turkish Republic, the Kurdish nationalist movement was able to rise around an ethnic identity, which culminated into the biggest Kurdish uprising in the history—the PKK’s war against the Turkish state. As we shall see in the following Chapter, the policies of repression and national identity project engendered the counterhegemonic discourse of the PKK, and helped construct the collective memory narratives of many Kurdish activists. Sendero, on the othe