Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Armed groups, child soldiers and the pursuit of legitimacy Plowright, William 2018

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

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata


24-ubc_2018_september_plowright_william.pdf [ 2.33MB ]
JSON: 24-1.0368943.json
JSON-LD: 24-1.0368943-ld.json
RDF/XML (Pretty): 24-1.0368943-rdf.xml
RDF/JSON: 24-1.0368943-rdf.json
Turtle: 24-1.0368943-turtle.txt
N-Triples: 24-1.0368943-rdf-ntriples.txt
Original Record: 24-1.0368943-source.json
Full Text

Full Text

  Armed Groups, Child Soldiers and the Pursuit of Legitimacy   by William Plowright B.A., The University of British Columbia, 2010 M.Sc., London School of Economics, 2011     A DISSERTATION 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) July, 2018 © William Plowright, 2018    ii  The following individuals certify that they have read, and recommend to the Faculty of Graduate and Postdoctoral Studies for acceptance, the dissertation entitled:  Armed Groups, Child Soldiers and the Pursuit of Legitimacy  submitted by William Plowright  in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Political Science  Examining Committee: Dick Price, Political Science Supervisor  Arjun Chowdhury, Political Science Supervisory Committee Member  Katharina Coleman, Political Science Supervisory Committee Member Philippe Le Billon, Geography University Examiner Brian Job, Political Science University Examiner     iii Abstract  In armed conflicts around the world, armed groups recruit child soldiers into their organisations to fill a wide range of roles. In response, transnational advocacy networks have been mobilised to name-and-shame such armed groups and seek an end to this humanitarian concern through asserting the norm against the use of child soldiers. Some armed groups respond to this advocacy, and demobilise their child soldiers, while others ignore international pressure. There is a puzzle here: why do some armed groups demobilize child soldiers, while others do not? What makes armed groups more or less responsive to advocacy? I argue that some armed groups are engaging with the child soldiers norm in order to gain legitimacy from international audiences. Evidence shows that the armed groups that do this are those who have a large domestic support base but are losing their armed struggle. Armed groups do not engage with the norm if they are winning their struggle, or if they are losing and do not have a large domestic support base.  In order to demonstrate this theory, I present evidence from qualitative interview-based research conducted in Syria and Myanmar in 2012 and 2013. I show that in these two dramatically different conflicts, armed groups follow the same forms of behaviour. In both cases, it is only when groups are losing but have support, that they will dialogue with transnational advocacy networks in order to engage with the child soldiers norm and acquire international legitimacy.    iv Lay Summary  In conflicts around the world, armed groups are recruiting children to wage war. This is a matter of great concern, and many advocacy groups have tried to pressure armed groups to cease using children as soldiers. Some armed groups respond and demobilize their child soldiers, while others ignore the pressure and continue. Why is this? I argue that some armed groups are responding to pressure and demobilizing their child soldiers because they wish to appear legitimate to international audiences. Put simply; they care what the outside world thinks and are prepared to prove it. I conducted interviews with 12 armed groups in Syria and Myanmar. In both contexts, armed groups respond to international pressure when they meet two conditions: that they are losing their fight, and that they have a large domestic support base. Groups that are winning, and groups with no local support, will continue to ignore international pressure.     v Preface  This dissertation is original, unpublished, independent work by the author, William Plowright. The fieldwork reported in Chapters 3-4 were covered by UBC Ethics Certificate numbers H12-01366 and H14-02240.    vi Table of Contents Abstract ........................................................................................................................................................ iii Lay Summary ................................................................................................................................................ iv Preface .......................................................................................................................................................... v Table of Contents ......................................................................................................................................... vi List of Tables .............................................................................................................................................. viii List of Figures ............................................................................................................................................... ix Acknowledgements ....................................................................................................................................... x Dedication .................................................................................................................................................... xi Chapter #1: Introduction ............................................................................................................................... 1 1.1 - The Puzzle ......................................................................................................................................... 1 1.2 The Theory & Hypotheses ................................................................................................................... 3 1.3 The Methodology .............................................................................................................................. 14 1.4 The Ethics .......................................................................................................................................... 21 1.5 The Road Ahead ................................................................................................................................ 26 Chapter #2: Armed Groups and Child Soldiers ............................................................................................ 27 2.1 Introduction ...................................................................................................................................... 27 2.2 Traversing the Definitional Terrain ................................................................................................... 28 2.3 The Common Wisdom ...................................................................................................................... 38 2.4 Conclusion ......................................................................................................................................... 57 Chapter #3: Myanmar ................................................................................................................................. 58 3.1 Introduction ...................................................................................................................................... 58 3.2 Cases ................................................................................................................................................. 60 3.3 The Myanmar Conflict ....................................................................................................................... 63 3.4 Karen National Liberation Army ....................................................................................................... 72 3.5 Karenni Army .................................................................................................................................... 78 3.6 Mon National Liberation Army ......................................................................................................... 82 3.7 United Wa State Army ...................................................................................................................... 85 3.8 All Burma Students Democratic Front .............................................................................................. 88 3.9 Pa-O National Liberation Army ......................................................................................................... 90 3.10 Summary ......................................................................................................................................... 93 Chapter #4: Syria ......................................................................................................................................... 97  vii 4.1 Introduction ...................................................................................................................................... 97 4.2 Cases ............................................................................................................................................... 100 4.3 The Syrian Conflict .......................................................................................................................... 103 4.4 The Free Syrian Army ...................................................................................................................... 113 4.5 Syrian Islamic Liberation Front........................................................................................................ 120 4.6 Shabeeha ......................................................................................................................................... 122 4.7 Hezbollah ........................................................................................................................................ 125 4.8 Jabhat al-Nusra ............................................................................................................................... 129 4.9 The Islamic State of Iraq and al-Sham ............................................................................................. 132 4.10 Summary ....................................................................................................................................... 137 Chapter #5: Conclusion ............................................................................................................................. 141 5.1 Concluding Thoughts ...................................................................................................................... 141 5.2 Contributions of Research .............................................................................................................. 142 5.3 – Future Avenues of Research ........................................................................................................ 153 5.4 Final Thoughts ................................................................................................................................. 158 Bibliography .............................................................................................................................................. 160     viii List of Tables Table 1: Predictions of Norm Acceptance by Armed Groups.......................................................................7 Table 2: Predictions for Cases Under Analysis..............................................................................................8 Table 3: Observable Implications of Potential Competing Explanations....................................................13 Table 4: Summary of Interviews…..............................................................................................................19 Table 5: Cases from Myanmar Conflict…….................................................................................................62 Table 6: Conflict Trends and Popular Support in the Myanmar Conflict….................................................95 Table 7: Cases from Syria Conflict…….......................................................................................................102 Table 8 – Conflict Trends and Popular Support in the Syrian Conflict. ....................................................140 Table 9: Syria and Myanmar Compared……...........................................................................................…144 Table 10: Alternative Explanations………...................................................................................................146 Table 11: Hypotheses of this Research……...............................................................................................147 Table 12: Primary Audience of Legitimacy-Seeking Behaviour…..............................................................149 Table 13: Preliminary Conclusions Regarding Normative Independence…..............................................151       ix List of Figures Figure 1: Locations of Field Research in Eastern Myanmar……...................................................................59 Figure 2: Area of Operations of the Armed Groups in the Myanmar Conflict….........................................69 Figure 3: Conflict Map of Myanmar Conflict.…...........................................................................................71 Figure 4: Location of Research: Aleppo Governate, Syria...........................................................................98 Figure 5: Status of the Syrian Conflict at of the time of Field Research (July 2013)…..............................107 Figure 6: Conflict Map of Syrian Conflict….…............................................................................................111     x Acknowledgements Firstly, I would like to thank my supervisor Dr. Richard Price, and committee members Arjun Chowdhury an Katia Coleman for their guidance and support. I would also like to thank my parents Hilary and Robert for trying to convince me not travel outside my comfort zone, and in doing so, unknowingly motivated me to pursue the life I did. I am very lucky to have them. In conducting the research, I would have been lost without the guidance and leadership of Muhannad and Ahmed in Syria, as well as Maung and U in Myanmar. Further, this research could not have been done without the financial support of the following; the Department of Political Science at UBC; the Social Sciences and Humanities Research Council; the Australia Awards-Endeavour Scholarships and Fellowships; the Liu Institute for Global Affairs (UBC); the Rotary International Peace Fellowship; the International Development Research Council; the JH Stewart Reid Memorial Fellowship Trust; the Public Scholars Initiative; and the Consular Corps of British Columbia.    xi Dedication  Dedicated to Muhannad and Saint-Fort.  At different times, in different ways, they acted as my guides and mentors in Syria and the Central African Republic respectfully. They shared their knowledge and experience with me, and in the years or months that followed, both died in tragic incidents. They shared a passionate belief that their people shouldn’t suffer the nightmare of armed conflict. Struggling against overwhelming odds, they sought an end to the pain and instability that plagued their families, their communities, and their countries. The loss of each of them is a heartbreak felt by their friends and colleagues around the world. Chapter #1: Introduction  Armed Non-State Groups, Legitimacy & Child Soldiers   “You should be afraid of us. Everyone in the West should be afraid of us… But tell your friends and family in Canada that we do not hate them, and that we do not use child soldiers. This is very important because it is against Islam, and we want the world to know it.”  X. (Name withheld), Islamic State of Iraq and al-Sham (ISIS1) commander. Aleppo, Syria. 20132  “It is very sad, children are supposed to be in school. Instead, you see [them] dying in the front lines.”  Colonel New Dah Mya, Karen National Liberation Army – Kayin State, Myanmar, 20123   1.1 - The Puzzle Late on a summer night in Aleppo, as the regime’s bombs fell around us, an exhausted leader of the Free Syrian Army explains the armed group’s recruitment strategy, showing the mechanisms by which children are prevented from joining. In Myanmar, a general of an ethnic rebellion proudly displays children who have been demobilized from his armed group, and who are now studying, mere miles away from where government forces are conducting a campaign of ethnic cleansing against their people. These events may at first appear unrelated and unconnected. They are not. Both represent trends in conflicts around the world, in which armed groups are changing their behaviour in order to gain international legitimacy by engaging with the norm prohibiting the use of child soldiers.                                                           1 The Islamic State of Iraq and al-Sham (ISIS) is also known by the names: Islamic State of Iraq and Syria, the Islamic State of Iraq and the Levant (ISIL), the Islamic State (IS), or the Arabic name ad-Dawlah al-Islāmīyah fil 'Irāq wa ash-Shām (DAESH). Hereafter, the name ISIS will be used as it was the predominant term used at time of writing. 2 X (name withheld).  Islamic State of Iraq and al-Sham. Interview with author, 05-07-2013, Aleppo, Syria. 3 Colonel New Dah Mya, Karen National Liberation Army.  Interview with author, 29-06-2012, Kayin State, Burma.  2 The use of children and youths as fighters is a problem that pervades conflicts the world over.4 However, there is no clear tally of the number of child soldiers in the world, due to the inherent difficulty in counting or estimating. The number 300,000 has been used since at least the early 2000s.5 Another oft-quoted statistic estimates that there are child soldiers in 40% of the world’s state armed forces and non-state armed groups. Other estimates have noted at in addition to the 300,000 actively fighting child soldiers, there are as many as half a million child soldiers not currently engaged in armed conflict. 6 The existence of so many children in armed groups around the world prompts a morally imperative question: what conditions lead armed groups to cease this practice, and what role do international actors play in this process? This practical question prompts a broader theoretical one: what is the relationship of non-state armed groups (ANSG) to the contemporary international order? What role does that order play in the decision to demobilize child soldiers? In order to answer this question, we must analyse cases where armed groups have actually ceased using child soldiers and have demobilized them, and also cases where they have not. Organizations such as Geneva Call make it their priority to approach ANSGs on issues like the child soldiers norm in order to compel them to engage with it. The result is that some armed groups do construct dialogue with members of transnational advocacy networks like this, and they end up demobilizing their children. Herein lies a puzzle; though children are used as soldiers in conflicts around the world, some armed groups release them from their ranks even as the armed conflict continues. They do this even though it puts the armed group at a strategic disadvantage in the conflicts they are fighting, since they are reducing their capacity to fight while continuing to do so. Armed groups use children as active fighters when they cannot recruit other soldiers, and they then may come to rely on children who are easy to abduct, indoctrinate, and replace.7 Recent research has also begun to suggest that child soldiers actually increase an armed group’s ability to fight (Haer and Böhmelt 2016). Yet, as noted, some groups demobilize them during their campaign of violence. There is an unexplained issue here, which provokes the key research question of this project: why are armed non-state groups who are known to violate international humanitarian law releasing some of their fighters when they are in the middle of fighting a                                                           4 In 2013, the prominent advocacy organization Child Soldiers Internationals released a press release titled “Children still used in most contemporary armed conflicts by state armed forces and non-state armed groups.” (2013) 5 This number is widely repeated as being a valid estimation without any real interrogation of the origins of this number or the reasons for its use. It is – at best – a wild guess. Studies which are guilty of this include Brett & Sprecht (2004, xiii); Wessells (2006, 30); Rosen (2012, 1); Schauer & Elbert (2010). The origins of the number are likely the Machel Report, the first real report on the issue of child soldiers (United Nations General Assembly 1999). One inquiry into the origins of the number describe it as a “zombie statistic” that isn’t based on any hard evidence (Kessler 2016). 6 For example, this is reported by the Council on Foreign Relations (2005). The problems with this estimation, its lack of any evidential base, and its continued use are discussed by Drumbl (2012, 26-27), as well as Nilsson (2013). There are inherent problems with estimations of this kind, as discussed by Ames (2009, 10-11). 7 A range of literature has discussed this issue, including, Singer (2006), Honwana (2006), Rosen (2012), Drumbl (2012), and Beber & Blattman (2013) amongst many others.  3 civil war? Why do they willingly sacrifice vital strategic resources, often for no clear or immediate gain? Why do some adopt this behaviour while others do not? I argue that these armed groups are adopting a form of costly signalling to both domestic and international audiences in order to show their commitment to the norm of the prohibition against the use of child soldiers, and therefore displaying a desire to be seen as internationally legitimate.8 They are showing a concern for longer time horizons in favour of an undetermined future gain.  However, not all armed groups engage in this behaviour. The puzzle here comes in explaining why some ANSGs make decisions that seem to run contrary to their strategic needs in waging insurgency. This behaviour shows that there is more occurring than a short-term cost-benefit analysis, and this necessitates digging deeper than hypotheses deduced from immediate material self-interest. Below, I will develop my argument for why some armed groups engage in this behaviour and why some do not.  1.2 The Theory & Hypotheses This research uncovers why some armed groups may become more willing to comply with IHL. The core argument is that groups who demobilize child soldiers do so in order to gain legitimacy not in order to seek a specific material good, but instead the gains from being seen as a legitimate actor. In general, armed groups will seek legitimacy from four broad audiences: (1) the international community (including IOs and transnational advocacy networks); (2) populations providing both material and non-material support to the ANSG; (3) the population in the ANSG’s zone of operations that do not yet support the ANSG in question; and (4) states, including enemies, allies and state sponsors. Although the focus here will be placed predominantly on the first of these categories, in the ensuing analysis, these audiences will be discussed since the legitimacy-seeking behaviour of each individual armed group depends on which of these audiences they are seeking legitimacy from.  Armed groups vary dramatically in the manner that they govern people in their areas of operation, from pseudo-state structures, to mere predation. Mampilly has argued that these strategies are a product of both the initial preferences of the rebel leadership, and their interactions with a wide variety of local and international actors(Mampilly 2011). It is well established that armed groups seek material gain from international actors to this end.9 However, recent research has also demonstrated that armed groups seek international legitimacy (Jo 2015). Some armed groups are endeavouring to be seen as legitimate political actors by international organisations and transnational advocacy networks, however not all of them are. I argue that this form of legitimacy-seeking is compelled by two motivating factors; the first being material strength, i.e. whether or not they are waging a successful insurgency (hereafter referred to as winning or losing).10 The second is the level of support in the population over which they are asserting to be the legitimate authority (hereafter referred to as with or without                                                           8 Hereafter referred to simply as the child soldiers norm. 9 For a full discussion, see Salehyan (2009) 10 Given that armed rebellion is considered an ends in of itself to many groups, ‘winning’ will also include whether the ANSG is successfully battling their opponent to a prolonged stalemate.  4 support. Contained within this are two implicit statements that will allow me to generate hypotheses. Firstly, when groups are winning, we should not expect them to change strategies. Secondly, a pre-requisite of international legitimacy is some form of local support. International legitimacy is not typically granted to armed groups who have no local support. When taken together, these dyanmics suggest that we should not expect groups that are winning to alter strategies, and we should not expect groups with no local legitimacy or support to seek it internationally. Overall, therefore, it will be shown that groups that are losing with support are much more likely to engage in this legitimacy-seeking, since they are engaged in a losing struggle and need to increase their support abroad, while also cementing the gains they have made at home. A priori, it might be assumed that groups that are losing would be the most likely to keep their child soldiers, and that groups that are winning will demobilize, since they are more likely to be able to bear the costs. I argue the opposite is occurring; that groups that are winning are likely to be indifferent to the norm, while groups that are losing are more likely to engage with it. However, it is not all losing groups who will engage with the norm. Only armed groups that have local support are likely to seek international legitimacy, since some form of local legitimacy is necessary, and domestic support is used to convince international actors that the ANSG is a legitimate representative of the people it claims to represent. Further, since a group that is losing with support is unlikely to be able to get more from its local population, it makes appeals internationally. The core argument that I use to generate hypotheses, therefore, is the following armed groups will engage with a salient norm in order to access  international legitimacy, based on the interaction of two variables of  winning or losing and with or without support. From this theory, I draw four key hypotheses. Firstly, I hypothesize that armed groups that are losing with support are the group most likely to engage in this form of legitimacy-seeking and norm acceptance. As noted above, local support is a pre-requisite for international legitimacy. The international community simply doesn’t grant legitimacy to armed groups that have no local support (and therefore some form of local legitimacy). Therefore, from the perspective of the gatekeepers to legitimacy, an armed group must have this support. This can be viewed as a ‘pull’ factor for groups with support. However, there are also ‘push’ factors. Groups have control over an area and population, however they are at risk of losing it since they are losing the overall struggle to other forces. The leaders therefore engage in strategic decision-making in order to seek new support outside their area of control, since they need further assistance just in order to hold on to what they already have. This is, in essence, a strategic ‘push’ factor compelling the ANSG and its leadership to seek new forms of support in their ongoing struggle and engage in legitimacy-seeking behaviour through engagement with a norm.  When armed groups engage in legitimacy-seeking, however, they must convince their organisation to accept the norm, they cannot merely dictate that everyone must believe it. It is not necessary for them to convince every individual to accept the norm, or to socialize every individual. Rather, they must institutionalize the norm into the organization, while convincing the group that the new norm is broadly in line with their pre-existing beliefs, in the form of the group’s central identity-narrative. Put simply, their pitch must credibly connect to the group’s central beliefs, even though it is not necessary for the leaders to convince every individual.  5 Secondly, I hypothesize that groups that are winning with support will not engage with the norm. As R. Wood notes, strong armed groups are able to tap into pre-existing resource bases and support for all strategic decision-making (2010, 601). They simply don’t need to gamble and change strategies in the same way as groups who are losing and facing the reality that without a change of strategy they are potentially doomed. There is little incentive, therefore, for a group that is winning with support to engage with the norm, which tacitly includes an admission of deviant behaviour in the past. In order to demobilize child soldiers, one must first admit that one has them. For a group winning with support, there is no clear reason to publicly admit deviant behaviour and risk the associated costs. It makes much more strategic sense to simply continue with strategies that are working. In addition to the argument that winners typically don’t alter strategies, there are also reasons to believe that even if these armed groups were to engage with a norm, they would have a more difficult time trying to do this. This is due to the role of bandwagoners. Because of their strength, these groups are the most likely to attract bandwagoners, who will not be as committed to the group’s cause as true believers.  These bandwagoners are the least likely to have their conduct guided by the group’s beliefs – or norms in general – and are more likely to engage in norm violations. They may care about the norm for other reasons, but we should expect them to be less susceptible to changes in the group’s core identity-narrative, since they are not as closely tied to it. Bandwagoners don’t merely align with the strongest power, they will also align with a group that they identify with in some way. However, the mere fact that they are bandwagoners suggests a limited investment in the organization. We should also expect this group to engage in cheap talk, condemning the use of child soldiers, while taking little or no action to institutionalize the norm. This is due to the fact that the group wishes for international legitimacy even if it has no interest in paying any costs for it, and if only to pay lip service to the prevailing trends with regards to norms. In terms of groups that are winning without support, I hypothesize that these groups will not engage with the norm and are unlikely to engage in cheap talk either. An armed group that has no support in its area of operations but is having success must have some kind of external supporter to which it is beholden. As Jo has noted, armed groups with a single state sponsor are susceptible to the pressures that their state patron asserts (2015, 29-30). We should only expect groups that are winning with support to care about the norm to the extent that their state patron does.  When states care about norms, they typically don’t support armed groups that don’t. Since the states who sponsor armed groups are more likely to be indifferent to these types of norms, we should expect the armed groups to be so as well.11 It is therefore likely to be indifferent to international norms and pressure, and subject only to the pressures of that external supporter. If that supporter is indifferent to the norm, the armed group will be as well.                                                           11 Since the provision of support to armed groups in another country is itself a violation of norms related to sovereignty and peaceful politics, it follows that they will be less concerned with other norms as well. Saleyhan, et al. have argued that state sponsorship of insurgents is more likely when the state can maintain plausible deniability to the violations committed by armed groups (2011, 713). This further demonstrates a lack of concern over the conduct of armed groups receiving sponsorship, since the state’s concern is to its reputation and not the norms themselves.  6 It is theoretically possible that we could see an armed group who cares about the norm for its own reasons or may have some local variation of the norm already institutionalized. However, evern for such a group, we should not expect them to engage internationally with transnational advocacy networks who are pushing the norm. The first reason is that the transnational advocacy networks will not assign legitimacy to a group with no local support. As noted above, local support is a pre-requisite for international legitimacy since local support implies local legitimacy. A second reason is due to the strategic interests of the armed group. As they are winning, and they have a state backer, it doesn’t make sense for them to alter their strategies and jeopardize both their status as a winning and the support from the state backer.  Altering a winning strategy doesn’t make sense strategically, however in engaging with actors other than their backer there is a risk that they could jeopardize the support that they already have. As noted earlier, engaging with the norm and demobilizing child soldiers involves a public admission of deviant behaviour. If groups that are winning without support do this, they run a very real risk of shaming their state patron who was supporting their behaviour during the period of norm violation, just implicating them in the deviant behaviour itself. This does not sound like a wise or strategic course of action for a croup that is completely reliant on an external backer to due to a lack of local support in their area of operations. Therefore, due to gatekeeping by transnational advocacy networks, and due to high level of risk, we should not expect groups that are winning without support to engage with the norm in any serious way. Finally, when a group is losing without support, we should not expect the armed group to engage with the norm, for the simple reason that a group that is losing its armed struggle and has no support is unlikely to survive for long. As Bueno de Mesquita has shown, when a rebel group is losing and is close to the end, we should expect them to hold out on current strategies – especially irregular ones – and “gamble for resurrection (2013, 325). This means that they should double-down on a strategy like the use of child soldiers, and fight to the bitter end in hopes of a change. At this stage, they are essentially fighting for their lives, and we should expect them not to adopt any course of action that risks limiting their fighting capacity and further reduce their ability to hold of the potentially inevitable end. In the case of child soldiers, we should expect to see them retain the fighters they have while potentially even trying to recruit more child soldiers in order to expand their fighting capacity. Further, they are unlikely to have the time horizons long enough to be concerned about the distant future and are likely to use any and all means available to them, meaning that they are likely the group that would be the most active at recruiting child soldiers. Their time horizons will be much more immediate, and will be limited to continuing the current fight, and holding on by the skin of their teeth, while potentially looking for ways to increase rather than limit their ability to wage conflict. As such, groups that are losing without support should not engage with the norm in any real way, and it is unlikely that the group will be around for much longer. These four hypotheses can be plotted into the following table;    7   Trends in the Conflict    Winning   Losing Level of popular support held Support  Reject norm / cheap talk   Adopt costly signalling; Demobilize child soldiers   No Support   Reject norm / no cheap talk  Continue recruiting / Cease Operations entirely  Table 1: Predictions of Norm Acceptance by Armed Groups In order to conduct this analysis, I have selected 12 individual armed groups to serve as cases to test the above hypotheses. I will discuss the case selection in detail below, however it is important to note here that these cases were selected to provide variation in potential explanatory variables. Further, they were selected from the contexts Syria and Myanmar in order to test the argument in two radically different contexts, with large variation in culture, duration of conflict, and in conflict dynamics. Using these hypotheses, I generate a set of predictions for each case under analysis here. Table #2 lists each of the 11 primary cases under analysis here in both Syria and Myanmar. Each ANSG is listed according to whether it is winning or losing, and whether it is with or without support. A prediction is determined according to the hypotheses above. Finally, the observed outcome for each case is also included. As can be seen in the below table, the predictions bear fruit in all the observed cases.        8 Case # ANSG Winning or losing? Support or no support? Prediction Outcome observed Prediction Confirmed?  SYRIA      1 FSA Losing Support Engage with norm Engage with norm Yes 2 SILF Losing No Support Continue recruiting / cease operations entirely Continue recruiting / cease operations entirely Yes 3 Hezbollah Winning  No Support Reject norm / no cheap talk Reject norm / no cheap talk Yes 4 ISIS Winning Support Reject norm / cheap talk Reject norm / cheap talk Yes 5 Jabhat al-Nusra Winning Support Reject norm / cheap talk Reject norm / cheap talk Yes 6 Shabeeha Winning No Support Reject norm / no cheap talk Reject norm / no cheap talk Yes  MYANMAR      7 ABSDF Losing No Support Continue recruiting / cease operations entirely Continue recruiting / cease operations entirely Yes 8 KA Losing Support Engage with norm Engage with norm Yes 9 KNLA Losing Support Engage with norm Engage with norm Yes 10 MNLA Losing Support Engage with norm Engage with norm Yes 11 PNLA Losing Support Engage with norm Engage with norm Yes 12 UWSA Winning No support Reject norm / no cheap talk Reject norm / no cheap talk Yes  Table 2: Predictions for Cases Under Analysis One last aspect of the theory is left to discuss here, regarding groups that are losing with support and choose to accept the norm. As noted briefly above, a group that does embrace the norm will be limited in the extent that it is able to. I argue that the manner in which the armed group will accept the norm (or not) will be conditioned by internal characteristics of the armed group. Not all ANSGs can adopt all norms, some will be incongruent with their beliefs. The identity-narrative of the group (its internal culture and beliefs) will condition whether and how they understand a norm. Armed  9 groups vary in their normative independence, or the ability of an armed group to incorporate new norms into its own identity-narrative. Once leaders have taken the strategic decision to engage with a norm, the identity-narrative and subsequent normative independence of an armed group therefore act as conditioning factors, determining the extent to which the norm will be credible to those in the ANSG.  1.2.1 Key Assumptions Before proceeding to potential alternative explanations, one key assumption and two scope conditions of this analysis must be clearly stated. The assumption is that the primary goal of every armed group is survival, even if this means abandoning the armed struggle for a political one. Waltz noted that states are motivated primarily by a concern for survival, due to their existing in a self-help system, thus limiting their ability to cooperate (1979, 104-107). The same is true for armed groups, though their focus on survival may be more acute since they face a more immediate existential threat. Whereas the aim of the group as a whole is survival, different actors may have very different conceptions of what that survival means. For the leader, survival is likely to mean survival of the ANSG as a whole with the leadership remaining at the top. For the sub-level commander or regional strongman, the interest of survival will be tied to the survival of the sub-group under his or her command, again with themselves retaining the leading role. This can lead them to embrace strategies to strengthen the ANSG as a whole, or to weaken it should they decide to factionalize.  For the fighters at the bottom, however, survival will be more tied to the ‘band-of-brothers mentality’ that emerges at the level of the fighting unit.12 Though individuals may risk or even embrace their own death, they will be concerned with the survival of their small fighting unit as well as its connections to the movement as a whole (Qirko 2013, 140-142). Whereas they may care about the causes for which the movement purports to be fighting, they will care more about local problems and local issues (Kalyvas 2003). As the literature on military psychology demonstrates, soldiers fight predominantly for their comrades and not for the movement as a whole.13 This does not mean, of course, that the fighters may not decide to defect at some stage. However, this very act of defection is a rejection of the identity or cause of one group in exchange for that of another.   Given these three different notions of survival at the three different levels, how then are we to conceptualize survival of the ANSG as a whole? An armed group’s narrative is the centrally unifying discursive device holding the ANSG together. We can conceptualize the survival of the ANSG as the survival of that narrative as well as the relationship of each level to it; meaning the survival of the leadership at the top, the survival of the regional-strong men as leaders of their subgroups, and the survival of the band-of-brothers at the bottom. The identity-narrative may change, as may the individuals in the organisation, but the collective survives. Krasner noted that the USA was an example                                                           12  See Grossman (1995, 149-155); Siebold (2007); Kirke (2009); and Qirko (2013).  13 Joanna Bourke, in her analysis of Western soldiers in WWI, WWII and Vietnam, that creating a “group mind” was necessary to increase the power of the individual and compel them to choose ‘fight’ over ‘flight.’ This group mind that Bourke discusses is of vital importance to the small group level (the ‘band of brothers’) as opposed to the army as a whole (2000, 98).   10 of a state that remained a coherent entity across times, even though the individuals within it, and the society it represented, changed greatly over that time (Krasner 1984). The same is true of the armed group; though the individuals, strategies, and even beliefs of the organisation may change, the group has a whole remains as long as there is an identity-narrative to tie them together. The first scope condition to be here noted is that the ANGs under consideration are engaged in active militarized struggles with no immediate end in sight. Whereas a group may demobilize child soldiers in order to encourage a peace process, or they may even be involved in a peace process, the restriction here is that this peace does not follow immediately and that the ANSG is not about to be disarmed either voluntarily or by force. Because they must prepare for the peace process to fail, or because they may receive no share of the peace dividend, they operate on an assumption that they will have to continue fighting, which factors into their strategic decision-making due to their primary goal of survival.  Additionally, should they be involved in a peace process, they will maintain their status as an armed non-state group and will not be simply be disarmed or incorporated into state military structures. Of course, over time, this can change, as can an armed group’s legitimacy. As Mampilly has observed, legitimacy from local populations is endogenous to the course of war (2011, 53) The same is true for international legitimacy, which can be both won or lost over time. The second scope condition is connected to the first. It states that the armed group is not the target of a major international coalition whose raison d'être is the destruction of the armed group in question. For example, during the period the Taliban was the target of a major international intervention from, we should not expect it to making appeals to legitimacy of the international system that was targeting it (even if it was losing with support over that period). The same is true of ISIS following the commencement of a major intervention in 2014 - we should not expect it to follow the same dynamics for other armed groups while it is being targeted by the very system to which it might make pleas of legitimacy. Therefore, it could be said that since coalition forces captured the ISIS strongholds of Raqqa and Mosul in 2017, it may be losing with support, but the theory noted here no long applies to them following the start of the coalition in 2014.   1.2.2 Alternative Explanations  In order to demonstrate the theory described above, it will be important to engage with a number of potential arguments for why armed groups may engage with the norm for reasons other than legitimacy. This subsection will lay out the most credible alternative explanations, to be later revisited. These alternative explanations will also be engaged with further in following chapters. The initial alternative explanations which I will need to rule out include: (1) cheap talk; (2) fluctuation in resources; (3) payoffs; (4) the parochial specificities of various ethno-nationalist conflicts; and (5) fear of punitive action by international actors. The first and simplest alternative explanation is that the ANSG may be simply engaged in cheap talk. For armed groups that demobilise their child soldiers, it is clear that cheap talk is not a sufficient explanation, since they are incurring a cost in order to follow the norm. Groups that do not demobilize,  11 however, may also engage in cheap talk even though they do not follow the norm. Even this cheap talk should not be quickly dismissed. It could be argued that since they have nothing to lose by simply condemning child soldiers, this argument might posit, perhaps the ANSG is simply paying lip service to the supporting the norm. It may, perhaps, make them look morally superior to their supporters. Alternatively, pretending to support the norm may trick members of transnational advocacy networks, which may in turn, increase the ANSGs support abroad or even potentially divert the attention of peacekeeping or counterinsurgent forces. However, even in such cases the talk may be cheap, but it still serves a purpose related to international legitimacy. This argument risks being tautological, however; if the armed groups didn’t see value in the opinions of others, then why engage in the talk even if it is cheap? Even if we grant that armed groups may engage in cheap talk to receive material benefits even though they don’t care about the opinion of others, this does not provide a full explanation of why they believe those benefits are available. Further, it fails to explain why ANSGs would continue seeking these benefits even when they are not forthcoming. The explanation of cheap talk does not account for why the norm is salient to begin with, and why international actors care about the conduct of local actors when they have no material interest at stake.  A second competing explanation is a fluctuation in available resources. One way this could occur is through the presence of lootable resources. If a purely materialist rationale is to hold, we should expect groups in regions with lootable resources to be more resistant to normative change since they be motivated solely by material gain and not by belief in a cause.  Armed groups which do not have large amounts of resources may recruit substandard and cheaper soldiers – i.e. child soldiers – in the manner described by Weinstein’s industrial explanations of rebellion (2007). Therefore, an armed group that suddenly seizes control of a large resource pool may simply demobilize its child soldiers in order to hire adults.  This does not appear to be the case, given the access and exploitation of such resources in both Syria and Myanmar, but also variations in whether or not groups adopt the norm. Therefore, one observable implication which allows us to reject lootable resources as an adequate explanation in all cases is the presence of an ANSG that is not reliant on lootable resources but is demobilizing. This would mean that even though they don’t have alternative means of hiring new recruits, they are still releasing the ones they have. As the cases from Myanmar will demonstrate, there is no direct correlation between presence of lootable resources and demobilizing, since there are groups that both demobilize with and without resources, and there are groups that do not demobilize, both with and without lootable commodities. The third alternative explanation involves a direct payoff of the ANSG by an external actor, essentially a material bribe or reward to encourage them to engage with the norm. This approach has been discussed by both Hamberg (2013) and Beber & Blattman (2013) who both advocate that armed groups can be paid off. This argument can be investigated through seeking to uncover whether a patron is paying off the ANSG. If this argument were to hold, we should expect to see evidence of such payments, since NGOs, international organizations and even states are unlikely to provide large amounts of resources for such a campaign without publicizing their work. Further, should an armed group accept such payment and that payment is not continuous, we should expect to see the armed group return to the deviant behaviour relatively quickly. As will be shown empirically, this account is not sufficient in the  12 cases under consideration, due simply to the lack of such pay-offs, pointing further to the longer legitimacy benefits as an explanation rather than short-term material gains.  A fourth potential explanation is that culture may play a role, and the specifics of ethno-nationalist conflict may explain why groups recruit or demobilize child soldiers. This could follow two separate logics, the first being that certain cultures may have established traditions of bringing children into military settings. The second is that ethno-nationalist conflict produces forms of communal mobilization in which all those able to fight are brought into the struggle, which then may end due to local reasons distinct from the actual material progress of the conflict itself.14  As such, engagement with the norm may be due to some cultural reason, explaining why they mobilize children in certain situations, or why they demobilize in certain situations. For example, it could be the case that the community considers it justified to use children for a short period when there is an existential threat. Or it may be the case that tradition forbids children from picking up weapons, and therefore they follow the norm regardless of international legitimacy. An observable implication allowing me to reject this would be to find two armed groups, both with the same ethnic identity, both fighting the opponent(s), however one follows the norm, but one does not. As will be explored in the empirical chapters, this was observed in both Syria and Myanmar. If there is diversity of outcomes across diverse cases, we should see a process that transcends the specificities of culture, even if there are some cases where there are cultural reasons for bringing children into armed groups. The fifth alternative explanation is a fear of punitive action by international actors. This could be an international intervention force, which is pressuring its ANSG allies to follow norms. Alternatively, the International Criminal Court may be involved. In March of 2012 Thomas Lubanga was convicted by the ICC for recruitment of child soldiers. Therefore, it may be the case that armed groups do not care about legitimacy, however are demobilizing out of fear. This explanation is based on coercion, rather than the voluntary motivation associated with the incentive of legitimacy. It is widely accepted that a key motivation for the existence of the ICC is to deter future violations (Kersten 2016, 23). If we accept the explanation that intervention forces are pressuring their ANSG allies, then we should expect to see armed groups demobilize when they have such allies placing pressure. If the ICC version of this explanation holds, we should expect the ICC to be active in the region, and for armed groups to display concern over the ICC’s actions related to child soldiers.  There is a broad problem with this explanation, namely that there is little evidence that the ICC has a deterrent effect on gross violators of IHL.15  As such, were this explanation to hold, it would run contrary not only to the argument                                                           14 There are many examples of cultural traditions which can be viewed through a lens of militarization. One of the most relevant here, are the examples of coming of age rituals in Sierra Leone and Liberia, which are referred to as secret societies. These involve ritualized training processes in which male youths are taking from their families and taught the cultural traditions of violence. This has been noted as a form of militarization of children by Richards (2004, 81-83); Ellis (2001); Ellis: (2004); Denov (2010, 57-60); and Hoffman (2011, 83-91). 15 This has been discussed by Cronin-Furman (2013) and Kersten (2016, 23-26). Jo & Simmons have argued that the ICC can have a deterrent effect if an armed group sees prosecutions as possible, they may undertake costs in order to comply with norms and therefore reduce the possibility of prosecution (2016). However, the  13 presented here, but to a broader field of research on the relationship between justice and conflict. However, for the purposes of this study, three observable implications can be noted to reject this explanation; the absence of international intervention forces placing pressure on the armed groups; the absence of the ICC in the region; and a lack of concern displayed by ANSGs to the ICC. At the time of research, all of these were absent from the areas understudy, and therefore do not hold as alternative explanations. The following table lays out all five of the alternative explanations to be rejected, and the related observable implications allowing the analysis to reject them. The empirical chapters will return to these explanations, providing evidence for why they can be rejected.  Explanation or Mechanism Observable Implication 1 Observable Implication 2 Observable Implication 3 1 Cheap Talk Did ANSG voice support for norm while not demobilizing all CS? Was there no deviation from norm? Did members of ANSG care if statements deviate from identity-narrative? 2 Resource Boom  Did ANSG gain access to resources immediately prior to demobilization? Did a large recruitment take place which replaced child soldiers with other recruits?  3 Payoffs  Did an international actor offer the ANSG resources to demobilize?  If no straightforward offer was made, was it the perception of the ANSG that such resources would become available?  4 Ethno-nationalist insurgency Do all members of the culture/ethnicity in the area of concern follow the same logic with respect to the norm?   5 International coercion Are international forces or the ICC placing certain pressures on groups that use child soldiers?  Are the ANSGs allied with COIN/PKO forces the only ones demobilizing? Is there credible fear of prosecution by the ICC, and does this fear pervade all groups equally, regardless of which ‘side’ of the conflict they are on?  Table 3 – Observable Implications of Potential Competing Explanations.                                                           problem with this is that it is simply saying that armed groups who want to be perceived as following norms are more likely to follow them.   14 1.3 The Methodology I will demonstrate the manner in which ANSGs seek to acquire international legitimacy through engagement with the child soldiers norm and use this as leverage to acquire legitimacy from other audiences. In order to explain why armed groups will sacrifice material resources that can contribute to their ongoing struggles, I will select a single norm and analyze the processes by which ANSGs both adopt the norm and attempt to demonstrate compliance to others. I will do so in twelve individual case studies; six armed groups each from Syria and Myanmar. The child soldiers norm is an ideal test for this research to demonstrate legitimacy-seeking since following the norm is a material trade-off; in order to follow the norm, you have to demobilise fighters and redirect time and resources in order to bring about institutional mechanisms that ensure no new children will be recruited. Overall, this makes the norm a hard case to demonstrate legitimacy-seeking through the use of norms, since cheap talk will not fool others, and costs must be incurred to follow the norm.  The overall approach of this study is to utilize qualitative research to analyse norm acceptance, while highlighting conditions of the armed group and its environment that led it to the outcome of interest. In order to highlight the strengths of this form of research, it is useful to compare it to a research design that adopts a starkly different approach. A recent study has sought to address the role of legitimacy-seeking by armed groups. Hyeran Jo has sought to investigate legitimacy-seeking by armed groups and its relationship to whether they comply with international humanitarian law (2015). Using statistical analysis of armed groups from 1989 to 2009, Jo analyses a universe of 250 cases of armed groups in order to understand if armed groups care about legitimacy (2015, 87). Although interesting, these results leave as many questions as answers.  Methodologically speaking, the use of a correlational logic tells us little about when or why armed groups act in the ways described.  In order to conduct a statistical analysis such as this, Jo uses a binary measure of violation or no violation which can obfuscate the dynamics of the child soldiers norm.16 The problem here is that child soldiers are present in almost every conflict and in almost every armed group. A single child preparing food for a group of soldiers does not constitute the same violation as training a thousand children to engage in suicide attacks. Both are violations of the norm in question, but they clearly represent vastly different positions on the norm itself.  It is into this gap that in-depth qualitative research in the field can dive in order to elucidate such processes, and probe deeper into legitimacy-seeking.  Jo’s research focusses on armed groups who are making appeals to constituencies who care about the armed group’s behaviour (2015, 19). Further, it only looks at why leaders might decide to seek legitimacy by following norms but does not ask whether or not the followers will do so once the leader makes that decision. My research goes a step farther, looking at cases where armed groups are trying to obtain external legitimacy by following a norm, and the conditions that make it more likely that they will. By using process training in multiple cases, in multiple contexts, it is possible to demonstrate the similar processes that take place in radically different contexts, while also making predictions about which groups engage in this behaviour and which do not.                                                           16 Jo does allow for variation however, in that some groups may comply persistently while some may comply at times and defect at times (2015, 90-92).   15 In the ensuing section, I will discuss the case selection I have adopted and the methodology with which I approached the research. My research will demonstrate these arguments by conducting qualitative research on active ANSGs using child soldiers in order to explain the adoption of the norm and the demobilization of those child soldiers.  My case studies will therefore be the individual ANSGs, and I will adopt a diverse cases approach in which I select different sets of ANSGs in disparate conflict systems.   1.3.1 Approach  My approach employs qualitative research to observe the adoption of a norm by the armed group and the attempt to follow it by both leaders and fighters in order to draw some preliminary conclusions regarding the causes. As Marini and Singer have argued, initial ideas of causal relationships are usually triggered by empirical cues and inductive reasoning. After looking at the use of empirical cues and inductive reasoning in the formation of causal hypotheses, we consider the process by which cumulation of a body of evidence leads to the confirmation of causal hypotheses (Marini and Singer 1988, 347-363). The methodology that I use is comparative analysis of multiple case studies, uncovering the preconditions that lead to the outcome of interest. A different approach - process tracing – uses diagnostic evidence to uncover sequenced steps in a causal chain. As discussed by Jacobs, process tracing lends itself to the study of ideational theories regarding elite actors, as it can be expansive in temporal range and level of analysis (2011, 3). As Mahoney observed, rather than highlighting necessary conditions (i.e. we cannot have outcome Y without precondition X), observing the process can help to show contributing conditions (2015, 203).  The approach here does not trace a full step-by-step process tracing of a single case, but rather, follows the broader process in twelve cases in order to highlight key points of decision-making (i.e. the adoption of the norm). At the same time, the approach I adopt uses key insights from the process tracing approach in order to use diagnostic evidence to trace ideational changes, and the resulting behaviour, in a complex process. It further highlights the conditions that led to the decision, and finally, whether these conditions allow the decision to be successful. This allows me to use qualitative research to gather data on the reasons for decisions, the effects of those decisions, and then compare across multiple cases to uncover shared antecedent conditions. In order to study the armed group, it will be essential to take statements by fighters or leaders and triangulate them with observed patterns of behaviour. In this way, it is possible to compare what someone says they believe, or what they say their identity is, in order to see if words match deeds. As such, although the identity-narrative and the beliefs of an individual or a group are not directly measurable or observable, they can be inferred by comparing actions and statements. Further, the actions or statements of individuals can be compared to those of the group as a whole, in order to determine if the vies of the fighter on the ground match the leader. In cases where it is not possible to access leaders, official statements can be used as substitutes. By watching the process unfold within multiple case studies, I will be able to uncover the conditions leading to the outcome in question or the conditions leading to a different outcome, while also addressing the role of sequencing and timing.  This  16 allows us to demonstrate the conditions which will increase the likelihood that armed groups will engage in norm acceptance and legitimacy-seeking, rather than asserting that all armed groups will act the same way should certain conditions be present.  There are further practical and theoretical reasons for using this approach. A practical reason for conducting qualitative research is the immense difficulty of safely gathering reliable quantitative data in on-going conflict zones (Norman 2009) . A more theoretical reason is the nature of the puzzle under examination. Dessler has noted that statistical studies focussing on explanations of war have led to an understanding of antecedent conditions, but not of the causal process leading to the outcome (1991). This research is seeking to explain the process leading to a certain outcome; namely the ANSGs that demobilize child soldiers and seek to embrace IHL after having been known to commit violations. I use a diverse cases approach, looking at variance in the outcome of interest (i.e. norm acceptance), and then look to the antecedent conditions that lead to the diversity in outcomes. This further allows analysis to identify the conditions making the outcome more likely, while also identifying the process involved.   1.3.2 Case Selection The twelve cases for the purposes of this study are individual ANSGs, which have been selected from multiple conflict systems, to provide a diverse range of cases within each context while also selecting from radically different contexts. It is possible to begin with deductive reasoning of how we should expect ANSGs to behave, based on their immediate material interests. However, given that armed groups act in ways that such reasoning does not predict (i.e. they demobilize while still fighting), it is necessary to look deeper. As such, I have refined my hypotheses using inductive research in several supporting cases, before using two conflicts – Syria and Myanmar – to pursue in-depth analysis.  I conducted research in several contexts, including the primary case studies Myanmar and Syria, as well as exploratory and inductive research with ANSGs in Afghanistan, Colombia, Indonesia, Papua New Guinea, the Philippines, and Thailand.17 Supporting interviews were conducted in Canada and Australia. Although not formally explored here as in-depth case studies, this research was highly informative in helping to refine the argument and identify commonalities across regions. Where possible, in the countries in which I conducted research, I did so with as many ANSGs as possible. By selecting on a range of these independent variables, this allows me to adopt a diverse cases approach. This allows me to hold potential alternative explanations constant. 18 At the same time, by ensuring that there are cases with varied explanatory characteristics, I avoid the pitfall of selecting cases                                                           17 The research was made possible due to support from the Social Sciences and Humanities Research Council; the University of British Columbia; the Liu Institute for Global Issues; the Public Scholars; the Rotary Peace Fellowship; and The Australia Awards Endeavour Fellowship; and the Michael Smith Foreign Study Award; and the Public Scholars Institute; and the Stewart Reid Fellowship.  18 This particular approach of “diverse cases” was discussed and developed by Collier, Mahoney & Seawright (2004). It is particularly suited to this study, as it seeks to explain the salience of the given norm in a variety of contexts, including with multiple actors each of whom represent different outcomes in the variable of interest, namely, the acceptance or rejection of the norm of child soldiers.  17 based on the outcome of interest.19 This in turn allows variation in the independent variable as well as the dependent, allowing me analyse the process from explanation to outcome. Following from this, I am able to discuss the process in multiple case studies within each conflict system. Doing so will allow me to use the features of each particular context and how these unique features show the working out of the process of interest, in what Skocpol and Somers discussed as a parallel demonstration of theory (1980). There are several key reasons why the primary case studies Myanmar and Syria were chosen. The first, and most straightforward, is the substantive importance of these two conflicts. Myanmar has regularly been cited as being the country with the most child soldiers of any country in the world (Drumbl 2012, 5). Although estimates in Syria are difficult, the conflict has received the largest amount attention of any conflict in recent years, including on the issue of child soldiers. These two countries also provide the most efficient source of data, in that there are multiple ANSGs in each conflict, including both groups that use child soldiers and groups that do not. This allows variation in outcomes. The second reason is methodological. As will be discussed in detail in later chapters, Syria and Myanmar represent radically different contexts in terms of the form, function and duration of the armed conflicts in each. By using both, I will be able to analyse a similar process in radically different environments. Collier, Brady and Seawright noted that a core goal of any process tracing is to use as many kinds of evidence as possible to reject alternative explanations while establishing the hypothesized relationship (2004). In my approach, I draw from this argument, while not restricting myself to process tracing a single case study. The differences both between countries and within them will allow me to accomplish this in each conflict, demonstrating the generalizability of the theory beyond these two contexts. Although the differences between the two conflicts will be discussed in the empirical chapters and conclusion, a few are worth noting here. The two conflicts therefore represent divergent ends of the conflict spectrum in terms of intensity, as well as in terms of acceptance or rejection of the norm. The Syrian Conflict is a high-intensity conflict, involving some groups that accept and some that reject the child soldiers norm as well as the international institutions that support or uphold them.  The Myanmar conflict, meanwhile, is a low-intensity conflict, which receives little attention internationally. Further, whereas the Syrian conflict is dominated by ANSGS with tens of thousands of fighters, the Myanmar conflict is dominated by much smaller groups. Finally, whereas the Myanmar conflict had been waged for sixty years, the Syrian conflict had only lasted three. The two conflicts represent extremes with regard to many characteristics, therefore increasing the analytical leverage.  1.3.3 Data Tools and Sources I conducted my research primarily through interviews with armed groups, while also seeking to triangulate all data with further interviews and secondary sources. These interviews included individuals representing all levels of armed groups, from the fighters at the bottom, to mid-level commanders, and both military and political leaders at the top. In interviews, the interviewee was given the following                                                           19 This follows from the recommendation of King, Keohane and Verba to avoid selecting on the dependant variable, and instead, to select on explanatory variables to the greatest extent possible (1994, 141).   18 choices for the purposes of citation (1) to be quoted by name; (2) to use a pseudonym of their choice; or (3) to be referenced anonymously by their position (e.g. fighter or commander). Further, it was made clear that in addition to academic research, the results would be used to advise organizations with whom I worked.20 Through qualitative data-gathering, combined with triangulation of data, it is possible to provide as reliable information as can be hoped for. By using multiple sources that have competing interests it is possible to wash out bias or self-interest of an individual actor who may have incentive to misrepresent. I conducted interviews with a multiple ANSGs (both rivals and allies), as well as representatives of state militaries, members of domestic civil society, and those in transnational advocacy networks.  Additionally, I used press releases (by ANSGs, NGOs, IOs and governments); news reports; opinion polls; as well as reports and assessments and other forms of documentary evidence.  I therefore use as wide a range of evidence as possible in order to draw a clear causal chain. Table 4, below, summarizes the location of all interviews conducted for the purposes of this study:                                                          20 Most notably the Child Soldiers Initiative, and Médecins Sans Frontières. Although I never interviewed anyone on their behalf, it was made clear to those interviewed that the purpose of the research was both to produce a PhD dissertation, but also to advise various organizations.  19 Table 4 – Summary of Interviews   Conflict Complex Notes Location of Interviews Armed Groups contacted and interviewed Number of Interviews 1 Afghanistan-Pakistan Pashtun Insurgency Ranges between low and high-intensity conflict in predominantly rural areas. Includes informal interviews. Common use of terrorist strategies (IEDs, VBIEDs, etc.) Afghanistan (Helmand Province) 1. Taliban 8 2 Myanmar Conflict Multiple low-intensity conflicts in various regions of the country, predominantly rural & guerrilla in nature. Strong connection to narcotics & weapons trade Myanmar (Kayah, Kayin, Mon, Shan & Yangon states) Thailand 2. All Burma Students’ Democratic Front (ABSDF) 3. Karen National Liberation Army (KNLA) 4. Karenni Army (KA) 5. Mon National Liberation Army (MNLA) 6. Pa’Oh National Liberation Army (PNLA) 7. Shan State Army-South (SSA-S) 8. United Wah State Army (UWSA) 26 3 Syria Conflicts High intensity urban conflict. Use of prohibited weaponry (chemical weapons, cluster munitions, etc.) Presence of international counterinsurgency force. Common use of terrorist strategies. Presence of high numbers of foreign fighters. Syria Turkey 9. Free Syrian Army (FSA) 10. Islamic State of Iraq and al-Sham (ISIS) 11. Jabhat al-Nusra (JaN) 12. Liwa al-Tawhid 13. Syrian Islamic Liberation Front (SILF) 75  20   Conflict Complex Notes Location of Interviews Armed Groups contacted and interviewed Number of Interviews 4 Moro Insurgency in the Southern Philippines Low-intensity conflict. Predominantly rural. Strong connection to traditional retaliatory violence between family/kinship groups (known as “rido”) Philippines (Mindanao) 14. Moro Islamic Liberation Front (MILF) 15. Moro National Liberation Front (MNLF) 19 5 Papuan Independence Conflict & Tribal Violence Low-intensity conflict, largely surprised. Sporadic use of terrorist strategies. Indonesia Papua New Guinea Australia 16. Organisasi Papua Merdeka (Free Papua Movement) 12   Additional Interviews:    Supporting Interviews Interviews (formal and informal discussions) conducted with individuals in transnational advocacy networks & academics Australia Canada Netherlands UK N/A 25   Explored Case Studies Interviews conducted with individuals in transnational advocacy networks, local government/military, and former members of armed groups Columbia South Sudan Thailand N/A 31    TOTAL NUMBER: 16 196  21  1.4 The Ethics Research on a topic such as this must strike a balance between the desire to access vulnerable youth, and the dangers this presents to both to the interviewee and the interviewer. Clearly, when engaging with children in contexts such as war-torn Syria or Myanmar, there is a huge risk of creating risk for the interviewed individual. The challenges grow exponentially when conducting research in the presence of armed groups, as it is often the case that the data that researchers desire is not the kind of information that the armed group wishes to have disseminated. Elsewhere, I have argued for the necessity of engaged research with armed groups on the subject of child soldiers (Plowright 2013a). Further, I have argued that such work with ANSGs can assist in encouraging even the most violent of armed actors to follow the norms of conflict, both directly and indirectly (Plowright 2013b). Finally, I have discussed the ethical approaches that are necessary in engaging in research in active conflict zones, with members of armed groups, especially when the qualitative research involves discussions on children or even interviews with them (2015). This section will provide a brief summary of these key points and the need for research of this kind to be policy-relevant, theoretically innovative, and ethically conducted. A wide range of literature has been produced on ethical research on children and youth, as well as on the practical challenges this presents.21 Further, there is a large literature on children and conflict. However, there is little discussion merging the two; i.e. research on youth in the context of war. The fieldwork conducted for this project included interview-based qualitative research relying on triangulation of data; however, the same problems exist in gathering data for quantitative research in conflict zones. Such research brings a range of ethical considerations and logistical difficulties that are not commonly acknowledged in the scholarly literature. I will here describe three key dangers of this kind of research: the lack of training, the lure of adventurism, and the production of non-useful research. It will then turn to a brief discussion of several strategies on how this research sought to avoid these, while also ensuring the protection of the individual being interviewed.   1.4.1 The Lack of Training In many areas of professional endeavour, ranging from the military, to humanitarian organizations, to journalism – individuals should expect to undergo a large amount of training before ever working in a conflict zone. In addition to training in their vocations, they also must be aware of their own safety and security and how it impacts others. Whereas academic researchers may receive training in methodology, it is exceptionally rare for individuals to have any kind of formal training in how to work in a conflict zone ethically or safely. Literature on the safe and ethical research in conflict areas is rare. One exception is the volume Surviving Field Research edited by Sririam, King, Mertus, Martin-Ortega and Herman  (2009). The title,                                                           21 For a literature review and practical guidance, see for example Alderson (2004), Bastien & Holmarsdottir (2015), and Pihl (2015).  22  however is misleading; though the book provides sage advice on ensuring ethical research and on the importance of establishing trust, the only chapter in the volume which addresses security and safety of researchers directly is on operational security. Although an important topic, operational security – which focuses on the overall climate of research and the risks involved – is very different from personal security, which focuses on the behaviour of the individual and how to respond to threats. Whereas the former helps individuals conduct risk assessments of unstable climates and the likelihood of conflict escalating, it does little to assist them in determining how to behave should they be threatened, kidnapped, or worse. Researchers in the field run a broad range of risks which stop well short of ethical or methodological concerns regarding their research. A well-funded researcher is a source of goods to steal, while also being a commodity themselves that can be kidnapped and ransomed or held for propaganda purposes. Further, researchers may find themselves facing risk of sexual assault. At the same time, more incidental dangers include landmines, unexploded ordnance, crossfire, terrorist attacks, and the dangers of simply being in the wrong place at the wrong time. Of equal importance is the risk of burn-out and long-term trauma. Though many universities offer high-standard counselling support, few if any require researchers returning from hostile environments to undergo a mental health check-up, in order to ensure that the researcher is not suffering from the side-effects of experiencing vicarious trauma or full blown post-traumatic stress disorder.  Training and preparation, therefore, is essential, not only to ensure safe and ethical conduct in the field, but also to ensure the well-being of the researcher upon their return home.  1.4.2 Adventurism Whereas many may admit that they are unprepared to work in a conflict zone, people are also likely to be blinded by their own desire for excitement, and thus be guilty of the second pitfall, that of adventurism. It is undeniable that travel is a rewarding and enriching experience, and some are attracted to conflicts as a place of increased excitement. Winston Churchill famously noted that, “Nothing in life is so exhilarating as to be shot at without result” (1898, 107). Researchers may also be attracted to this. They may also find themselves involved in a ‘race to the bottom’ with other researchers to gain access to more intense conflict zones, in a competition over who has accessed more exciting areas. Research may be motivated more by a sense of adventurism than a desire to produce insightful or innovative conclusions. In their study of research in conflict zones Helbardt, Hellmann-Rajanayagam and Korff have warned against the “sinister glamour” of conflict. They further asked, “Is there a rationale and a justification for the researcher to enter the war zone and indulge in the analysis of gory details without any apparent academic usefulness?” (Helbardt, Hellmann-Rajanayagam and Korff 2010). Researchers risk therefore becoming war tourists, advancing their own academic careers through a voyeurism of suffering.  23  1.4.3 Non-useful research This connects to the third pitfall, the question of exploitation and non-useful research. As Nan noted in her study of the Georgian-Abkhaz conflict, scholarly study is able to contribute directly and immediately to the goals of conflict resolution (2010). This kind of research can risk being exploitative, however, if it benefits solely the researcher, and does not provide anything to the individuals who are actually suffering. As such, research conducted with youth in conflict zones must be useful to the issue being researched. Do the results help those suffering, or those working to end that suffering, or both? Or, conversely, does the research have the effect of worsening the situation and benefiting those guilty of the suffering under study? This is a thorny issue; how are we to conceptualize ‘useful’ research? And how much time do we allow in order for this usefulness to appear; is it ethically tolerable if the research does not benefit those in the conflict in question, but may be applicable to future conflict settings? If we were to assess research on a rubric of immediate ‘usefulness’, there is a risk in constructing a hierarchy of disciplines, in which results are ranked not necessarily according to how useful they are in the long-term. This risk comes in advocating an analysis in which immediate practical results are deemed more desirable; a public health paper may outrank a social science one, which in turn could leave any critical studies approach at the bottom of our hypothetical hierarchy of usefulness.  This is something we should seek to avoid. Wallace famously pointed to a core problem of the social sciences was a tendency to favour theoretical research over applied research (1996). Weiss described useful applied research on humanitarianism as follows: “When done properly, applied research should not only help the helpers but also improve the lives of those who are dependent temporarily on outside assistance” (T. G. Weiss 2013). The goal here is not to observe that purely theoretical research is not useful. However, those suffering in conflict are less likely to be concerned with theoretical insight than with research which will assist them (or others in their position).  Though settling the debate over what constitutes ‘useful’ applied research is well beyond the scope of this project, it is important to remember that often the research primarily benefits the career of the researcher themselves. Therefore, the question should always be asked how the research will benefit those whose suffering it is analyzing.  1.4.4 Strategies Adopted in this Study In order to avoid these three pitfalls, the research conducted here followed several key strategies. These do not correspond directly to the three pitfalls noted above, but instead, are collectively intended to address them. It is important to note that the approach used here is not a panacea, or one that should be reproduced in all cases. Each also comes with its own secondary problems which must be explored. These three strategies are: (1) training in both ethical research and personal safety and security; (2) to work using the support of NGOs and humanitarian organizations; and (3) the use of professional psycho-social support. Firstly, in preparation for training, I took part in a number of workshops and training sessions in order to develop my ability to conduct ethical research. This included fulfilling all of the behavioural 24  ethics requirements of the university at which I was based.22 Further, I underwent training with: Child Soldiers Initiative; Working with Child Soldiers; and Child Wise Australia (specifically on working with abused children). I also took part in workshops with Save the Children and WarChild, and a range of local NGOs, on the subjects of work and interaction with child soldiers.23 In order to ensure my knowledge of security, I underwent security training with the Humanitarian Training Initiative, the Canadian Disaster and Humanitarian Response Training Program, and Médecins Sans Frontières. I also completed the Basic and Advanced Courses on security provided to United Nations staff.24 Safety and security is similar to ethical research in that there is no one approach or solution which can guarantee success. These workshops and trainings did not provide a certification which made me able to perform this research ethically or safely. Rather, they provided a variety of perspectives, and ensured I was constantly asking myself questions about how I was conducting my affairs. A core recommendation for others engaged in this kind of research (especially at the graduate level) is that they should not simply seek to do a single training on security or ethics, but rather, should constantly be engaged in training, workshops and discussions on these subjects. This will help ensure they are seeking multiple perspectives, are questioning their assumptions, and aren’t falling into bad habits. Doing so addresses both concerns around being unprepared for working in conflict areas, with the concern about adventurism, since it will help people gain better understanding of what risks are appropriate and which are to be avoided. The second strategy is that researchers can seek to advise or support NGOs. With regards to ethical research, they are often a great repository of local knowledge. For example, they may provide advice on what forms of questions are culturally appropriate and which are not. On the other hand, NGOs may actually seek the advice and perspective of the researcher, who is able to advise them on strategies used by other organizations in other regions to tackle similar problems.  Different relationships with NGOs will be appropriate in different contexts and will also vary with the subject under study.  This, naturally, comes with its own problems. It is possible that staff-members of NGOs may have their own political agenda, or a personal or organizational self-interest that could conflict with the interests of the research. Concerns about the NGO can be eased by stressing the standards towards which the NGO itself adheres. For example, the Sphere Standards seek to hold NGOs accountable to beneficiaries and donors and ensure that work is done efficiently and ethically.25 A researcher can work                                                           22 Specifically, the requirements for human behavioural research from the Office of Research Ethics at the University of British Columbia. 23 Including at the conferences Children and Youth Affected by Armed Conflict: Where do we go from here? (October, 2013), and Bogota, Colombia Challenges for the Protection of Children in a Changing Context: The Case of Colombia (November, 2013). 24 Basic Security in the Field II (BSITF) & Advanced Security in the Field (ASITF) (June 2014), through the United Nations Department for Safety and Security (UNDSS). 25 The IFRC’s Code of Conduct and the Sphere Standards seek to assure that assistance is delivered effectively, with the consent and approval of local actors. Put simply, it is a set of commitments from the NGO to allow it to be held accountable, buy setting measurable standards for the quality of the care provided, as well as the 25  through organizations that are known to adhere to such systems of accountability. They will then be able ensure that the research is disseminated in such a fashion to be useful to those seeking to end the suffering in question. Alternatively, they may be able to both give and seek advice about ethical and appropriate conduct, while also seeking to advise others whose work is benefiting the suffering.  I adopted the approach of working alongside several NGOs of varying kinds, though all had in common that they worked predominantly in areas affected by armed conflict.26 Typically my involvement with the NGOs was an informal partnership in which I sought local knowledge from them and gave practical advice (in the form of advising “monitoring and evaluation” programs). In one case – MSF – I worked directly with them and ceased all formal interviews. One risk is that those interviewed will be confused about who the researcher is representing; a university, or an NGO, or some combination? I adopted the approach of explaining that I was primarily a PhD researcher. This is especially the case when I was introduced to a specific individual by someone from an NGO. I explained the results of my research would be used to inform the NGOs I was affiliated with at the time, while also making it clear that I was not a staff-member of that NGO nor a representative of them. Operating with the support of such NGOs helped to ensure my work was conducted ethically and safely, while also presenting me with multiple fora in which I could advise others on the practical benefits of the research conducted here. The third and simplest strategy is that researchers should at all times receive psychosocial support from trained professionals with regard to their work in conflict zones. Although all individuals are different in the reactions to suffering and trauma, the risks of permanent mental or emotional damage are high when working in conflict areas. In a myriad of ways, war is potentially traumatic, and individuals can become traumatized simply by hearing stories of another person’s suffering (termed “vicarious trauma”). This is commonly acknowledged in a range of professions, from law enforcement, to psychology, and even the clergy.27 Researchers in the social sciences must learn from this and ensure that individuals receive the psychosocial support they need to heal from potentially traumatic events. A psychosocial debrief with an expert should be a mandatory part of any research in areas affected by armed conflict. In order to ensure my own mental health, I sought support through the relevant counselling services of the University of British Columbia, the University of Melbourne, as well as the Psychosocial Support Unit of Médecins Sans Frontières. Further, while in conflict areas, I used the self-assessment materials developed by the Headington Institute in order to manage my own stress levels and observe my own behaviour for signs of vicarious trauma or post-traumatic stress.28 A core recommendation,                                                           involvement of local beneficiaries in the process. The are available online at IFRC (2014) and Sphere Project (2011) respectively. 26 Including, at different times, Médecins Sans Frontières, the Child Soldiers Initiative, Save the Children, UNICEF, and WarChild. Although I was under contract with MSF, I did not engage in any interviews while under contract. With the other organizations, I provided advice on an informal basis and assisted in leading monitoring and evaluation programs. 27 As discussed in Second-Hand Shock: Surviving and Overcoming Vicarious Trauma (Izzo and Miller 2010). 28 The Headington Institute develops self-care techniques and resources for humanitarian workers who may not have direct access to psychosocial support (Headington Institute 2016).  26  therefore, is that any research working in conflict areas should seek professional psychosocial support before and after their stay in a conflict area, and should also develop personal strategies for monitoring themselves while they are there.  Carpenter questioned the extent to which academic research should be involved in addressing the human rights abuses it analyzes. As she noted, “I recognized myself as a participant in the process of social construction about which I was writing.” (C. Carpenter 2012, 377). In the research presented here, I do not seek to avoid this problem, but instead, to embrace it in order to ensure that research is conducted ethically and safely. Whereas the approaches advocated here do not represent a perfect strategy that will be useful in all areas, in this research they collectively work together to ensure I was professionally prepared, conducted myself in an ethical manner, had appropriate advice on the ground, and the support I needed before and after field work. Further, by working with NGOs known to maintain high standards of accountability, I was able to gain appropriate advice, while also ensuring the theoretical knowledge I gained could contribute to practical work done on the ground.  1.5 The Road Ahead  In order to demonstrate the argument summarized above, this study will proceed as follows. The second chapter will flesh out the argument by setting necessary definitional parameters of key terms, including the armed non-state group (ANSG), the conflict that they engage in, and the child soldiers that they recruit. Further, I will define the concept of legitimacy as it pertains to armed groups, while discussing the definition of the norm that is used to gain legitimacy, as well as the international or transnational actors that act as a gatekeeper to this legitimacy. The chapter will also provide a review of the existing literature on armed non-state groups engaged in contemporary warfare, and the manner in which this research seeks to add to it. It will also engage with the concept of identity-narrative, in order to understand how even armed groups who are willing to accept norms will be limited in the extent that they can do so. Following this, I will present two empirical chapters, a first Myanmar and a second on Syria. Discussion of each context will present a wide array of cases in radically different contexts, however demonstrating similar processes. These cases will demonstrate the manner in which different ANSGs involved in varying conflicts, fighting of different resources and causes, in widely disparate locations, are still subject to trends in international order regarding legitimacy. In building the argument, I will discuss whether each ANSG is either winning or losing, and whether they are with or without support. In the concluding chapter, I will draw upon the conclusions generated here in order to provide directions for future research, including the socialization of both armed groups and the individuals that constitute them.   27  Chapter #2: Armed Groups and Child Soldiers  Assessing the Study of Legitimacy, Norms in Contemporary Armed Conflict   “The children are always asking, is it the government, or the Shabeeha, or Hezbollah, or Iran that is coming to kill them?” Free Syrian Army city councillor (name withheld) – Aleppo Region, Syria.29  “Our enemies [the Democratic Karen Buddhist Army & the Burmese state] still use child soldiers. They don’t follow regulations. They will use anyone they can. They have no education or knowledge of international standards. But we do. We continue to struggle here and abroad for recognition, identity, freedom and democracy.”  David Thackabaw, Vice President of Karen National Union, Kayin State, Myanmar.30   2.1 Introduction  When discussing armed groups and legitimacy, we must first determine what we are talking about when we talk about armed groups. In conflicts around the world, armed non-state groups adopt a wide variety of forms and functions. They engage in terrorism, guerilla strategies, open war, control of markets, and in some cases governance. They do so in the name of democracy, freedom, ethno-nationalism, and even Marxism. They are accused of trading in weapons, lootable resources, and narcotics while engaging in human rights violations, and even ethnic-cleansing. Amongst their numbers are leaders, cadres, porters, fighters, drivers, and everything in between. This wide variety of actions, forms, ideologies, and behaviour begs simple questions: given this variety, what is an armed group? Which members are inside, and which are outside the armed group? And who exactly are the ones that define the ideology and identities? All of these combine to form one question which is the overall question investigated by this research: when there is some form of degenerate or widely condemned behaviour, who is it that we have to convince to stop, and how is it done successfully? What are the conditions of the armed group that lead it to follow this course of action?                                                           29 City Councillor (name withheld). Interview with author, 04-07-2013, Marea, Syria. 30 David Thackabaw, Vice President of Karen National Union.  Interview with author, 29-02-2012, Mae Sot, Thailand. 28   In order to demonstrate the theory argued in the opening chapter, it is necessary to have clear definitional boundaries to establish clear conceptual parameters. When we discuss legitimacy of armed groups, the adoption of norms, and the use of child soldiers, a number of questions arise over the meaning of these terms. What exactly is an armed group? Does it include entities labelled as terrorist or criminal organizations? Further, what exactly is legitimacy, and who are the gatekeepers to this legitimacy, deeming the armed group to be either legitimate or not? Finally, there is wide-scale disagreement over what a child soldier is, including disagreements over which age to use as a threshold, and even which positions (fighter, porter, and even sex slave) should be included and which not. This chapter will establish a clear definition of key terms, while also reviewing the literature on armed conflict, legitimacy and norms, in order to identify gaps in previous research on the subject, and to highlight where this research can contribute.   2.2 Traversing the Definitional Terrain In this subsection, I will define key terms, most notably the issues regarding what is (and what is not) an armed group, a child soldier, and an armed conflict. There are often challenges associated with defining key concepts in the social sciences, and few areas can be as challenging as terms associated with armed conflict. This is due not only to the ambiguous nature of the terms in question, but also because of the lack of consensus and the highly charged nature of such disagreements. When defining who is and is not a terrorist, for example, there is more than a discursive war of words at play, there is a parallel physical war over words. Rather than seeking to ignore the challenges associated with these terms, I will highlight them, in order to show the manner in which debate, and discussion of these concepts is part of the process of norm engagement and legitimacy-seeking.  2.2.1 The Pejorative Taxonomy of ANSGs Defining an armed group is a challenging enterprise. When constructing such definitions, we are attempting to mould our definition to reality and, conversely, seeking to ensure that the ontology we begin with is consistent both with our definitions and our understandings of the units we are observing. This problem is clear with the concept of an ‘armed group’; the primary unit of analysis in this study. What definitional approach ought to be adopted? Should it include terrorists? How are we to tell the difference between a criminal organization and an armed group? What size constitutes a ‘group’; are four youths a group? Should we avoid the pitfalls of using pejorative language such as ‘rebels’ or ’terrorists’, and if so, how? What of armed groups who receive state support; are they arms of the state, or armed groups in their own right?  An armed non-state group (shortened to ANSG or ‘armed group’) will be here defined as any organized group which engages in military operations to pursue political, religious, ideological or economic objectives, and is not under the formal operational control of a state or another group. This 29  definition seeks to avoid key pitfalls, the first of which to be discussed here is to utilize pejorative terminology which risks assigning or stripping a group of its legitimacy. The second is to define a single ANSG based on a single strategy in its portfolio of violence, while ignoring both that ANSGs are dynamic in strategy and organization, and they simultaneously use multiple strategies and tactics in their operations. As noted earlier, the goal of defining armed actors is a highly contentious undertaking.31  In scholarly literature, ANSGs are referred to using a broad range of nomenclature, including (though not limited to): rebels, insurgents, guerrillas, warlords, paramilitaries, militia, and militants. In the colonial and Cold War eras, much of the literature favoured the terms insurrectionists, subversives, or even narco-terrorists.32 At the extreme end in terms of deprecatory terminology are the labels terrorists, criminals, pirates, and gangs.33  In using such pejorative labels, a common approach is to conflate terms with competing definitions; for example the interchangeable use of terms such as militant, extremist, radical and terrorist.34 Findley & Young note that the terms insurgency, rebellion, guerrilla war and terrorism are usually used interchangeably (2007). It has been observed many times that the popular maxim “one man’s terrorist is another man’s freedom fighter” is a cliché.35 Although there is a kernel of truth in this cliché, definitions must be able to rise above simple dualities of terminology in favour of conceptualizations based on more value-neutral characteristics. As will be argued below, this is overlooked in much of the literature on conflict, which sees organized violence as socially deviant behaviour. The researcher, therefore, risks injecting their own values into a study through the terms they use. It has been noted that parties to conflict seek to label each other with negative terms to delegitimize each other in a discursive conflict of propaganda.36 However, this practice is not limited to belligerents at war with one another. The literature on conflict in the social sciences is littered with examples of pejorative naming. Some use terms such as insurgents, movements and challengers as                                                           31 As observed by Hoffman (2006, 1-42), Schroeder (2005), and Bhattia (2005). 32 The terms subversives and insurrectionists are especially common in the classic literature on counterinsurgency, including Kitson (1971) and Thompson (1966), with insurrectionist being favoured even by the so-called insurrectionists themselves when it came to the writing of the Vietnamese General Nguyên Giáp Võ (1962). The term ‘narco-terrorist’ originated through use by South American politicians condemning Marxist groups, however later spread to American politicians, and is often used in academic discourse (Martin 2010). Interestingly, the US Department of Justice’s Drug Enforcement Agency distinguishes between drug-related violence and narco-terrorism (2002). 33 The term ‘pirates’ was largely out of favour in the analysis of maritime conflict, except for in a historical context that usually focussed on the age of exploration. The term saw a resurgence in reference to armed groups from Somalia who became popularly labelled as pirates in the media. Interestingly, though the term ‘pirate’ has become commonplace in both the academic and popular discourse on Somalia, it is rarely used in the cases of other armed groups who engage in armed conflict in a maritime or littoral context. Bueger notes a narrative in which ANSGs in Somalia refer to themselves as a coast guard seeking taxation from foreigners which they will take by force if need be, while actively rejecting the label ‘pirate’ (2013, 17-24). 34 Hamilton-Hart notes this problem in the literature on Islam and terrorism in Southeast Asia (2005). She does so specifically referring to Abuza (2003).  35 As noted by Bhattia (2005, 7) and Ganor (2002, 284). 36 This has been noted by Fiebig-von Hase & Lehmkuhl (1997) and Nomani (2004). 30  synonyms.37 Others includes regular references to ANSGs to as being “greedy”38, as “spoilers”39, or even being referred to as “thugs”40 and “bandits”41. Paul Richards connected such language to what he termed a “new barbarism” in the literature on conflict, which encompassed a “re-primitivization” of the developing world where violence is degenerate and backwards (Richards 1996).42 Tuastad analysed the history of such terminology with regard to Middle Eastern conflicts, noting that these terms are a representation of a “neo-orientalism” of terminology that isn’t used in reference to conflicts involving the West (2003). These observations are hardly recent, however they remain compelling because the pejorative terminology they describe is still in wide use. These terms are rarely used in conjunction with Western armed actors; neither state nor non-state, in neither the past nor present. Neither the Royalists from the English Civil War nor the Confederate Army of the American Civil War are referred to as ‘insurgents’, or as ‘greedy thugs’, though perusing any book on conflict in Africa or the Middle East will reveal a generous peppering of such terms. Whereas the pejorative nature of terms like ‘terrorist’ is readily apparent to most observers, what is not as commonly acknowledged is that the same problems occur with labels that are often treated as being value-neutral, including insurgent, guerrilla and rebel. The term insurgent has a wide range of meanings, including both positive and negative connotations.43 Most standard definitions (from both the social sciences, as well as the literature on counterinsurgency) contain a number of key conceptual ingredients. These usually include that it is some kind of a struggle between a non-ruling group and the ruling authorities, in which the former consciously uses political resources and violence to sustain, destroy or reformulate the basis of legitimacy of the ruling authority.44  Whereas these stipulations may at first seem to be tenable, major conceptual problems can be noted. Mockaitis has argued that there are few terms related to conflict that are more controversial than insurgency, and its counterpart counterinsurgency (2010, 142). As both are defined relative to the other, the terms contain little meaning other than one’s status as an authority and the other as a non-                                                          37 See for, example, Bob (2005, 8). 38 The descriptions of rebel groups as being motivated by greed emerged most prominently as a result of Collier & Hoeffler (2000), though has become common terminology in study of civil war focussing on political economy approaches and quantitative methodologies that Claire Metelits lamented as being the “approach dominated by World Bank officials.” (2010, 164.) 39 Notable examples include Stedman (2002); Greenhill & Major (2006); Zahar (2010); as well as Hoddie & Hartzell (2010). 40 Recent examples include Azam (2006); Mueller (2007, 1); and Staniland (2010). 41 Examples include Angstrom (2001); Arquilla (2011); Collier, Hoeffler & Sambanis (2005, 3). 42 Although this was first pointed out by Richards (1996), the term has been picked up and used more recently in Tuastad (2003), while also being picked up by Samiei as a form of Orientalism (2010). The term is also used actively by Logan (2011, 718) to describe the Mexican Drug War as “new barbarism and depravity” quite apart from the manner meant by Richards. 43 So difficult is the endeavour to establish clear definitional parameters, that the very first sentence of the Encyclopaedia of Insurgency and Counterinsurgency is the simple yet honest sentence, “The term ‘insurgency’ is difficult to define.” (S. C. Tucker 2013). 44 Discussions of insurgency which include these characteristics of insurgent activity include O’Neill (1990); Mackinlay (2002) and (2009); as well as Metelits (2010). It has also been noted by the official counterinsurgency manuals of both the Canadian (2008, 1) and American militaries (2006, 1). 31  state entity. Mackinlay acknowledges that ‘insurgency’ is not a value-free concept, noting that to Mao it meant “genesis, freedom, and a fresh start to a new political era.” In contrast to this, in the West the term was defined negatively as a danger to commerce, security and stability (2009, 9). The positive and negative connotations are reflected not only in the literature on conflict, but also in the realm of policy. In 2004, American Secretary of Defense Donald Rumsfeld told reporters they shouldn’t use the term insurgents to describe the armed groups in Iraq, because the term gave leaders of the armed Sunni opposition a legitimacy they didn’t deserve.45 Instead, he favoured use of the somewhat ungainly name “enemies of the legitimate Iraqi government”, noting that calling them insurgents assumed they had a “legitimate gripe.”46 It wasn’t until 2013 that this advice was followed in Afghanistan, when military leaders in Afghanistan requested a change from the label insurgents to “enemies of Afghanistan” (Brook 2014). The Mexican government has argued against use of the term to apply to ANSGs in Mexico, preferring instead that the term “drug traffickers” be used (Negroponte 2011).    In addition to these being political charged terms, they also assume that the armed groups in question are seeking to overthrow the state. This assumption is greatly misplaced. In large conflict systems like those of the Great Lakes Region (centred in the DRC) or the Golden Triangle (centred in Myanmar), there are many armed non-state actors who do not delude themselves into thinking they will ever seize the central state, or even a part of it. Many have much more immediate and local aims, be it to represent a certain ethnicity, or to control certain markets, or even to satisfy some religious doctrine. Conceptually speaking, the term ‘insurgent’ suffers problems as the only agreed upon characteristic is that insurgents are people fighting the government, which does not describe many of the ANSGs that dominate conflict in places like the DRC and Somalia, where conflict is dominated by conflict between non-state actors. All that is left of the term is that insurgents are non-state armed groups; a nearly insignificant amount of conceptual leverage gained in a term that comes with so many problems. Other commonly used terms fare little better in terms of being conceptually clear or value-neutral. Much of the literature in social science prefers the term “rebels”.47 The term has two separate connotative sets of meaning. In popular culture in the West, the term is likely to conjure images of popular films, such as James Dean in the counterculture film Rebel Without a Cause (1955), or the iconic Che Guevara in as Guerrillero Heroico. In popular literature, we have romanticized images of a rebel as made famous through Camus’s The Rebel, in which a rebel is seen as a self-liberating slave (1967).  However, the image of the ‘rebel’ is not always a positive one. In the context of an Islamic country or conflict, the term ‘rebel’ is conceptually associated with the Islamic concept of taghut (توغاط)48, which means “to rebel” or “to overstep boundaries”. The term therefore takes on a completely new meaning: one of someone who rebels against Allah and aligns himself with Satan, thus                                                           45 Donald Rumsfeld, as cited in Perry (2010, 10). 46 Donald Rumsfeld, as cited in Associated Press (2005) 47 The terminology of ‘rebel(s) and rebellion’ is the preferred nomenclature, for example, in Gurr (1970); Keen (2005); Weinstein (2007); Kalyvas & Kocher (2007); Salehyan (2009); Beardsley & McQuinn (2009); Hegghammer (2010); Akcinaroglu (Rebel Interdependencies and Civil War Outcomes 2012); Driscoll (2012); Fjelde & Nilsson (2012); Checkel (2013); Hazen (2013); and Parkinson (2013). 48 Also transliterated as taghout taaghoot and ṭāġūt 32  demanding allegiance at the expense of God (Mir 2008, 55-56). Many of those who align themselves with the Salafist strain of militant jihadism – including most affiliates of al-Qaeda and ISIS – follow the writings of Sayyid Qutb, the Egyptian often noted as being the key theorist of contemporary militant Sunni Islam. For Qutb, and his ideological descendants, the rebels are those who challenge Islam, meaning the rebel is not the armed jihadist but instead is anyone that stands against him.49 This understanding of what the term ‘rebel’ means is reflected in communiqués and statements made by al-Qaeda leader Ayman al-Zawahiri.50 In this view, the rebels are Assad and Obama, not ISIS or the FSA. Definitions of the term ‘rebels’ also state implicitly (and often explicitly) that the ‘rebels’ are rebelling against the state, which is simply not true of many ANSGs.  As with insurgent, therefore, the term rebel comes to mean little other than that it represents an armed non-state actor, though it is laden with denotative shortcomings.   In the endeavour to avoid any form of pejorative taxonomy, a more neutral approach can be found in the literature of practitioners. Value-neutral language is commonly sought in international human rights law, as well as in the language used by development workers or humanitarians who deploy into areas where they expect to encounter violent actors.  The approach taken by most transnational non-governmental organizations (NGOs) is to embrace neutrality over partiality. Indeed, neutrality and impartiality between belligerents is a core philosophical tenet of humanitarian action in general.51 This is clearly represented in the definitions used and will be duplicated here. In beginning to finalize the working definition, it is worthwhile to begin our attempts to define the armed groups under study with definitions emerging from this more neutral approach. A common definitional approach within international humanitarian law (IHL) and the literature on humanitarianism is to use the terminology of armed non-state actors, the most common definition of which is: “any armed group, distinct from and not operating under the control of, the state or states in which it carries out military operations, and which has political, religious, and/or military objectives.”52 The standard United Nations definition of armed groups is “Groups that: have the potential to employ arms in the use of force to achieve political, ideological or economic objectives; are not within the formal military structures of States, State-alliances or intergovernmental organizations; and  are not under the control of the State(s) in which they operate.”53 The organization Geneva Call, who works exclusively with ANSGs, defines them as “any organised group with a basic structure of command operating outside state control that uses force to achieve its political or allegedly political                                                           49 Qutb’s book Milestones is considered to be the foundation of modern political Islamism, especially of its violent Sunni varieties (Calvert 2013, 1-8). Milestones calls the enemies of Islam rebels, and terms them “Jahiliyyah”, or those ignorant of divine guidance (Qutb 1978, 5).  50 Ayman al-Zawahiri, as cited in Mansfield (2006, 174). 51 As discussed in Allen & Schomerus (2008, 46) and by Barnett & Weiss (2008, 3). Further it is enshrined in the ICRC’s Fundamental Principals; humanity, impartiality, neutrality, independence, voluntary service, unity and universality (2014). 52 As in Bellal, et al.  (2011, 48). 53 As used in UNICEF’s Guidelines to Humanitarian Negotiations With Armed Groups (McHugh and Bessler 2006, 1) which is drawn from the Office for the Coordination of Humanitarian Affairs Glossary of Key Humanitarian Terms (2003, 7). 33  objectives.”54 The approach taken by such organizations is one which stresses the political position of the organization themselves, namely with respect to both neutrality and impartiality. This neutral approach to defining the non-state actors we wish to discuss broadly has several advantages. The first is that it avoids pejorative terminology. It also does not seek to establish definitional boundaries where they do not exist in reality, and therefore it avoids the pitfall of conceptually asserting an absolute distinction between military groups or those with criminal intentions. Though it may seem a small conceptual point, it is necessary to drop the term ‘actor’ from the term ANSG. This is because, as the following sub-section will argue, we must come to clear understandings of ANSGs not as single and cohesive unitary actors, but instead as fluid and dynamic groups of individuals – and to an extent, groups of groups – which constantly change both in form and function.  Secondly, I will also drop from the definition here the requirement that the armed group is not under the influence of a state, since this stipulation would actually remove a large number of ANSGs that exist in the world. Several examples include ANSGs that ally with the state in Myanmar; paramilitary groups in Colombia; as well as the Shabeeha and Hezbollah in Syria. Whereas, naturally, we do not wish to include violent groups that are under the complete control of a single state or government, all this really means to say is that we do not wish to include the armed forces of the state itself (be they military or paramilitary, formal or informal). At the risk of employing a double negative, any ANSG under the complete control of a state is, naturally, not non-state in any real sense.  However, it is important to acknowledge that many (if not most) ANSGs are under the influence of a state actor, be it the state in whose territory they are fighting, an external patron, or even a local ally. Most ANSGs conduct military operations in at least one country where they are at least to some extent under the control of one or more states; either they receive support from one, weapons from another, or at the very least are allied and fighting a common enemy. As such, I will tighten this assumption to say that the group is not under the operational control of a state or any other armed organization.55  Drawing this section to a close, an overall key point is worth summarizing. Definitions reflect worldviews, as does terminology. This act of naming and defining is itself part of a discursive battlefield, representing both a struggle for legitimacy an attempt to take it away. This occurs in all aspects of a conflict, from the way belligerents represent themselves, to the ways they discuss others. However, it                                                           54 As in Geneva Call (2007) and DCAF & Geneva Call (2011). 55 This would allow us to keep groups like the Shabeeha in Syria and Colombian paramilitary groups in our definition, while dropping private military companies (PMCs) who are guns-for-hire contracted by states. Examples of PMCs include Academi (formerly known as Blackwater) or Saracens International, who when deployed are under the operational and contractual control of the state they are hired by. As such, we wish to include groups who may receive arms from or be under limited control of a state. In some cases, they may even follow the orders of states when conducting interoperable deployments, in the same sense that a Canadian contingent to NATO intervention force in Afghanistan may follow orders of the America generalship in command of the mission at the time, even though Canada and the US are separate states. Naturally, we wouldn’t consider the Canadian Armed Forces in such situations to be under the control of the US, and neither should we consider ANSGs to be so when in similar positions. 34  also includes other actors, from how they are portrayed in the media, to how they are defined in international law, and how they are discussed in academic research. If we view ANSGs solely as greedy rebels, insurgents, spoilers or, we risk missing their broader struggles for legitimacy, and how they seek to be seen by others.  At the same time, however, we should not allow analysis to risk parochialism, and analyse cases of conflict separately from one another. Put simply; we need terms to describe what we observe in the world. The application of labels and definitions must be conducted in a way which is able to both allow for local variation while assessing commonalities across case studies.  The dividing line between insurgency and terrorism is at best fuzzy, and at worst, is impossible to establish. Rather than ignoring this, the approach adopted here is to embrace the lack of a sharp conceptual distinction between different terms, and instead to address the heterogeneity of the strategies of armed non-state actors while stressing the commonalities they share, including how they organize, as well as their relationship to the international system.   2.2.2 The Child Soldiers Norm In order to analyze the role of young combatants in insurgent groups, it is necessary to clarify the definitional parameters within which my research will operate.  Discussion of a child soldiers norm necessitates a definition of a child soldier and  a norm. Beginning with the former, the Optional Protocol to the Convention on the Rights of the Child provides the most commonly accepted definition. It defines a child soldier as: “any person under 18 years of age who is part of any kind of regular or irregular armed force or armed group in any capacity, including but not limited to cooks, porters, messengers, and anyone accompanying such groups, other than family members. The definition includes girls recruited for sexual purposes and forced marriage” (United Nations General Assembly 2000). The guide to the Optional Protocol notes that this definition includes children who have not necessarily carried weapons and is intentionally broad in order to include as many children affected by conflict as possible (United Nations Children's Fund 2003, 14).  This standard definition is highly contentious as it includes within its conceptual parameters almost any child involved in conflict in any number of conceptually vague roles. A second problem, as noted by James B. Pugel, is that the dividing point between childhood and adulthood. He argues that this shows that the term is laden with Western bias and ignores the manner in which childhood is a social construct which varies across cultures (Pugel 2010, 160). Additionally, the definition of the word ‘soldier’ is unclear. Though the definition includes porters and cooks, they are not traditionally considered to be soldiers. Further, those taken as sexual slaves are rarely considered to be members of an armed force or armed group in any real sense, however a child abducted for such purposes would indeed be considered to be a child soldier. The assertion that the so-called ‘comfort women’ taken by Japanese soldiers in the Second World War were soldiers is unlikely to find much support. Why, then, should it be assumed that girls are soldiers if they were taken as sex slaves by Boko Haram or ISIS?   35  The approach that I adopt is to not ignore the conceptual haze and lack of definitional consensus regarding what is and what is not a child soldier. Instead, I seek to take advantage of this lack of agreement, since this disagreement is not simply an academic one. It also persists within ANSGs, who have wildly differing opinions on the subject. By embracing multiple definitions, and lack of complete consensus, we are better able to see how armed groups are attempting to follow the norm, even when An elucidating example comes from several interviews conducted in Aleppo, Syria in 2013 with commanders from various armed groups. During an interview with members of the Free Syrian Army (FSA), one commander proudly boasted that his battalion lacked child soldiers, while also proudly showing off a 14-year-old who had volunteered to carry ammunition to snipers on the frontline.56 This represented a disagreement over both what is and what is not a soldier; in this commander’s view, a soldier fires a gun, he does not tote ammunition. Another perspective was presented by a commander from ISIS, who noted that recruiters sought only those over the age of 25, as they could not be sure if those claiming to be 18 were really mature enough to be considered adults.57 In another interview, with a commander from the Syrian Islamic Liberation Front (SILF), it was asserted that girls could not be child soldiers because even though women could fight, only a man could be a soldier.58 According to this view therefore, a girl could kill enemy soldiers and still not be a soldier herself, while a man could be over 18, but not mature enough to be an adult. These groups may not have always followed these rules, but what is important to note here is their varied understandings of what a child soldier is. A fourth example, from West Papua, came in discussions with leaders of the Free Papua Movement (typically known by its Indonesian name and acronym Organisasi Papua Merdeka, or OPM), who noted that children trained to fight could not be considered child soldiers as the conflict was low-intensity, and it was unlikely that the children would be sent into battle.59 Whereas the previous examples discussed the specific tasks that the fighter would be assigned, or the age at which one becomes an adult, this view accepts the norm but questions when it actually applies. Taken together, these views show a variety of interpretations of the norm and adaptations to local context.60  There is therefore a lack of consensus on the content of the norm itself. In terms of norms, I will adopt the standard definition of a norm as a standard of appropriate behaviour for actors with a given identity.61 The child soldiers’ norm, therefore, is prohibition against the recruitment of child soldiers. But                                                           56 Commander X (name withheld). Free Syrian Army. Interview with author. Interview with author, 05-07-2013, Aleppo, Syria. 57 Commander Y (name withheld). Islamic State of Iraq and al-Sham. Interview with author, 05-07-2013, Aleppo, Syria. 58 Commander Z (name withheld). Syrian Islamic Liberation Front. Interview with author. 05-07-2013, Aleppo, Syria. 59 Leader A (name withheld). Organisasi Papua Merdeka & Papua Traditional Council. Interview with author. 15/03/2014. Jayapura, Indonesia. 60 This can be seen as an example of norm localization, as theorized by Acharya (2004), to be discussed further below. 61  This is taken from Finnemore and Sikkink (1998, 891).  36  whose definition of a child soldier is most useful; the commonly accepted definition among transnational advocacy networks, or the view according to those who actually recruit and deploy them? The manner by which armed groups seek to match their definitions to those of the outside world is a largely dialogic exercise. They are matching their paradigmatic worldviews to those of the international community at large, and in the process, are ironing out the kinds of definitional agreements noted in the previous paragraph and are creating a flattened consensus.  However, even as they come to acknowledge the issue of child soldiers, they may not accept the definition of the norm that is utilized in the international community. They may localize, re-interpret, or even misinterpret it (Acharya 2004). Since the internationally accepted version of the norm is the one that used to measure norm adherence, and therefore access to legitimacy, this is the understanding of the norm that will be used here. However, at the same time that we must use the UNICEF standard as a broad definition, we must also acknowledge a second and much narrower definition that is much more prevalent from the perspective of the ANSGs. This narrow definition is that a child soldier is a person who is under 18 and is an active fighter in an ongoing conflict. This narrow definition represents a conceptual rally-point; few would disagree that a child with a gun is a child soldier.  A further lack of consensus regarding the definition of a child soldier relates to how children are recruited. There is one major myth that exists in the literature on child soldiers as well as in popular conceptions of them, and this myth is that child soldiers are abducted or kidnapped.62 In predominant views, child soldiers are not recruited but instead are forced at gunpoint to fight. Whereas there are many examples of this happening, the truth is that most child soldiers are volunteers who are pressured by community beliefs or socio-economic pressures, not by the barrel of a gun. As will be discussed below, there are many documented cases of children lying about their age in order to get into armed groups. Although international humanitarian law forbids the use of children as soldiers, as they are too young too consent, they still exercise their own agency concerning their involvement in conflict. This study does not limit itself to children who have been kidnapped or abducted, but instead, looks at the broad definition of the norm which includes all children within armed groups, recruited by a variety of means. Though I will use the broad definition as the final point of acceptance of the norm, it must also be acknowledged that individuals and organizations may not adopt the norm wholesale, or may adopt only a part of it; the narrow definition. Finnemore and Sikkink have theorized a ‘norm life cycle’ from emergence, to acceptance and finally to internalization (1998). In this approach, there is a consensus on the norm’s meaning and contents, however this is not always the case in practice. By bringing this in to the process of norm acceptance, the process of following the child soldiers norm involves four steps, namely: (1) moving from contestation to recognition of the narrow definition; (2) to localization and acceptance; (3) recognition of the broad definition and finally (4) acceptance and institutionalization of the more expansive definition of what a child soldier is.                                                           62 Discussed at length by Drumbl (2012, 11-17). 37  The dividing principle for my approach, therefore, between the narrow and broad definitions is the act of carrying and using weapons. The narrow definition includes this stipulation whereas the broad definition treats all child soldiers as equals. The reality is that not all child soldiers will be treated or viewed equally by the ANSGs. Certain child soldiers will be more difficult to demobilize than others, while there will naturally be many individuals who would be considered to be child soldiers by the broad but not by the narrow definition. At the same time, however, though the broad definition is problematic, it will be used for this research because it is essential to engage with the norm on child soldiers according to the language and definition parameters in which it asserted at the international level.   2.2.3 Armed Conflict Before wading more deeply into the literatures on child soldiers and insurgency in general, it is useful to provide a brief discussion of the definition of conflict itself. I define intrastate conflict as: any violent confrontation in which there is: (1) violent opposition between two or more organized groups (2) at least one of whom is a non-state group, (3) where there is effective resistance by at least two sides opposing each other. Though this term is less contentious than those noted earlier, it is still important to clarify that this study does not seek to delimit the term ‘conflict’ into different categories based on the scale of conflict. Rather than simply trying to analyse the largest conflicts, or the lowest intensity conflicts, or ethno-religious conflict, or any other sub-category, the argument here applies to all armed non-state groups involved in all forms of conflict. The standard definition of civil war as used in much of the literature on intrastate conflict is that provided by Small and Singer. They included in their definition of civil war into any conflict “that involves (a) military action internal to the metropole, (b) the active participation of the national government, and (c) effective resistance by both sides.”63  In my definition, I have dropped the requirements that the conflict be internal to a single state, that one of the actors involved in the confrontation is a state, and that there be 1,000 battle-deaths per year.  Each of these are restrictive parameters which eliminate huge numbers of cases for study, in return for only a small amount of leverage. Firstly, the stipulation that the conflict be internal to a single state is dropped for several reasons. There are very few conflicts which take place entirely within a single state. Armed groups move freely across borders in underdeveloped countries, as do weapons supplies, refugees. Were we to accept Small and Singer’s definition, we would have to exclude every single conflict analysed in this study, as none match the criteria that they take place in a single metropole. There would be little remaining for study; even the conflicts of Papua New Guinea (PNG), likely the lowest intensity and most ‘local’ conflict in the world, would not be a true ‘civil war’ as it spreads across the border with Indonesia and does not always involve the state. Secondly, I drop the requirement that one of the actors involved is the national government. From Somalia, to Afghanistan, to the DRC and Columbia, there are many violent dyadic confrontations which do not involve the state                                                           63 This definition was first described in Small & Singer (1972), and later expanded on in Small & Singer (1982, 210). 38  directly. What we are trying to analyse here is not the behaviour of an ANSG under restrictive circumstances (i.e. when tries to seize the state), but rather, the behaviour of all ANSGs in general. Removing this restriction allows analysis of conflicts with intrastate origins which have been internationalized (due to an intervention force, or due simply to spreading over borders), as well as for armed conflict between non-state groups.  Finally, I also drop the requirement of 1,000 battle-deaths, as I am looking to discuss both high intensity conflicts such as that in Syria, as well as low intensity conflicts in PNG and Indonesia.  A small caveat must be added here, and that is to note that my research will be viewing conflicts as systems rather than as dyadic relationships in the way Small and Singer do. As Salehyan noted in his study of transnational rebellion, “it is important to analyse regional conflict clusters involving multiple state and non-state actors rather than treat civil wars in various countries as separate events or make arbitrary distinctions between civil and international conflict.” (2009, 15).  For the purpose of my analysis, my case studies will be the individual ANSGs. However, it is important to acknowledge that they operate in a conflict system, in which a there are a multiplicity of violent confrontations in a single region or sub-region, with multiple violent dyadic confrontations between multiple governments and/or armed non-state groups. There are few – if any – violent dyads that exist in isolation. Where there is smoke, there is fire, and not merely a single flame. In the twentieth-first century, where there is war, there are a multiplicity of armed groups, and not merely a single group of ‘rebels.’ Taken in total, this definition allows us to view the similarities and trends across conflicts around the world, viewing the trends that are occurring globally as contemporary armed conflict evolves and changes. Moving forward, it will be possible to show that very large armed groups are subject to the same trends as small ones, in conflict systems both large and small, in regions vastly disparate from one another.  2.3 The Common Wisdom  This section will provide the background discussions on the argument presented here, while also connecting it back to broader discussions in the literature of contemporary war. After discussing he dynamics of conflict involving child soldiers, discussion will turn to the political organization of insurgency, and the manner in which we must consider both material and non-material motivations for fighting, at both the group and individual levels. Discussion will then turn to the role that identity-narratives play in the non-material motivations for war, especially with regards to legitimacy-seeking. This will be specifically discussed with regard to theoretical discussion on norms, and finally the concept of normative independence.  2.3.1 Child Soldiers in an Era of Intra-State War The child soldiers’ norm provides an ideal test for demonstrating legitimacy-seeking, and it is well poised to address the normative pressure by international actors on ANSGs for several reasons. 39  Most importantly, this is because of the material trade-off that comes with following the norm: in order to follow it, you have to remove certain fighters, and then institute programs to make sure they don’t come back. This makes it a crucial case for any theory seeking to explain the engagement by ANSGs with international humanitarian law (IHL). I have chosen to focus on the prohibition of child soldiers as the norm I research due to its ability to make contributions on three levels. The first reason for using the child soldiers’ norm is the substantive importance of the topic.  The use of children and youths in conflict is a problem in intrastate conflicts around the world.64 Secondly, as the pressure on non-state groups to follow the child soldiers’ norm is a relatively new phenomenon, it presents an opportunity for innovative research. Although there is a great literature on norms in conflict, child soldiers represent a different dynamic than landmines or chemical weapons. With the child soldiers’ norm, the prohibited weapons system is not an inanimate object, but instead is a human being with their own agency, meaning following the norm is more difficult since child soldiers as individuals have concerns and interests that are not necessarily the same as the ANSG.65 The third reason is methodological, namely the generalizability of my study to all armed groups using child soldiers, as well as beyond to other norms or aspects of IHL. This leads to the fourth reason, that conclusions generated in this study will be able to enhance the theorization of the relationship between non-state actors and norms in international politics. As noted above, the child soldiers norm is relatively new on the transnational advocacy agenda. Though it appeared in the mid-1990s with the Graça Machel Report (United Nations Children's Fund 1996), active dialogue on the subject between ANSGs and transnational advocacy networks have only begun in earnest over the last decade. These transnational advocacy networks are composed not only non-governmental organizations, but also of the different organs of international organizations such as the UN, AU and EU that engage with these social movements.66 These networks are therefore both transnational and international in nature, while containing components that are both non-governmental and intergovernmental.  Prominent among such organizations, Geneva Call have begun to work side-by-side with armed groups to assist in the demobilization of child soldiers, while also providing monitoring mechanisms.  With regard to the literature on child soldiers, there are two dominant approaches in the contemporary literature, the first of which sees economic development as a means to avert conflict, and thus prevent children from ever being recruited.67 This first approach subsumes the children in question                                                           64 In 2013, the prominent advocacy organization Child Soldiers Internationals released a press release titled “Children still used in most contemporary armed conflicts by state armed forces and non-state armed groups” (2013). 65 Child soldiers as a “weapons system” was discussed by Dallaire (2010).  66 This understanding of transnational advocacy networks draws heavily from Keck and Sikkink (1998, 8-12). 67 In much of the literature, economic development is asserted as being a solution to the problems faced by children in conflict. In such literature, the issue of child soldiers is usually subsumed as being only one symptom of underdevelopment and conflict among many. Prominent examples include Collier, et al. (2003); Collier (2007); and Anderson & Carter (2009). Singer touches on this directly in relation to child soldier issues (2006, 135-138). Mack also advocates this as the most effective way of ending the use of children in 40  into a search for the panacea to conflict as a whole, rather than treating the child soldiers’ norm as an issue with unique considerations worthy of special attention. The second approach is palliative; it treats the effects of conflict once it has ended by focussing on post-conflict reconciliation.68 While one approach addresses the issue prior to armed violence commencing, the second focuses on them once conflict has ended.69 In order that the research here can begin to address these gaps, it is important to discuss the common wisdom related to how and why ANSGs organize, and why they fight. There are therefore gaps in the literature on the study of child soldiers. There is little on how child soldiers are organized and used within an ANSG. Further, there is little research on how the use of child soldiers alters the strategic orientation of the organization as a whole, in terms of interests, tactics and strategic aims. The literature on child soldiers within the armed group, and how they come to be demobilized, is largely limited to biographical accounts of the child soldiers themselves.70 Though they provide emotional and compelling description of the experience of the child combatant, they provide little in understanding of the broader process. Most discussion of demobilization beyond these accounts comes from an advocacy perspective; seeking to advocate for demobilization or assessing the viability of specific demobilization processes. These discuss the how and not the why of when armed groups will demobilize child soldiers, leaving a gap in the analysis.71  A second critical gap in this scholarship neglects the normative influence of international actors on armed groups.  The research conducted here seeks to contribute to filling both these gaps, by discussing how armed groups demobilize children to seek legitimacy, and to highlight the role of international actors in encouraging this behaviour.                                                            conflict (2010).  Beber and Blattman explicitly state this as a solution to the problems of the use of children as combatants (2013, 101). 68 There is a great deal of literature that focuses on the reintegration of children post-violence, focussing on psycho-social support and reconciliation back into their communities. More importantly for the argument being developed here, are examples from the literature that advocate DDR/SSR as a means to solve the problem of child soldiers in conflicts around the world. Examples include Vittachi (1993, 90-103); Brett & Specht (2004, 129-136); Dallaire (2010, 152-184); Denov (2010). Wessels, meanwhile, links reintegration back to development, noting that if child soldiers are not reintegrated the society will struggle in post-conflict peacebuilding (2006, 154). 69 The only literature that deals with the time in between pre-conflict development and post-conflict reintegration comes in studies of child soldiers focussing on law. However, even these only provide codified legal approaches on how to render the use of children illegal, while providing little assistance in addressing those who have already been recruited. Notable examples from the literature on child soldiers and international law include Cohn & Goodwin-Gill (1994, 55-92); Kuper (2000); Happold (2005); Chikuhwa (2010) and Drumbl (2012, 134-167). Achvarina and Reich provide a rare example of an argument developed to address child soldiers once they are recruited. However, as their argument specifically focuses on child soldiers in relation to preventing refugees from being recruited, it stops short of addressing the problems presented by children who are already in the ranks of an ANSG (Achvarina and Reich 2006). 70 For examples, see Keitetsi (2003); Toïngar (2006); Beah (2007); Bilkuie (2008); Jal (2009); and Oloya (2013), as well as the collection of first-hand accounts provided by Eichstaedt  (2009). Secondary accounts focussing on the issues related to a single child soldier include Sheperd (2008) and Williamson (2012).  71 A notable example contrary to this trend is Hamberg (2013). 41  2.3.2 The Political Organization of Insurgency When discussing how armed groups behave, two debates remain prominent. The first is in comparing political and economic rationales for the motivations behind violence. The second discusses the political and economic opportunities external to the armed group, that shape its campaign of violence. I will here discuss both, before proceeding to analyse how the two can be merged when discussing the normative paradigm of ANSGs, and how it connects to the concerns for legitimacy. This subsection will touch on each, discussing how these debates have progressed or stagnated, with specific reference to legitimacy and the child soldiers norm.  Greed or Grievance? Much of the literature on child soldier recruitment has been connected to the economic rationale for warfighting.72 In this literature, insurgency is a product of ‘greed’ factors in which individuals only fight when in their own economic self-interest, compared to the ‘grievance’ factors associated with fighting for causes or beliefs. There has been a great deal of research that has moved past this debate, and toward a general consensus is that both ‘greed’ and ‘grievance’ matter when explaining violence (MacGinty and Williams 2016, 40). Whereas the literature on armed conflict has moved past the ‘greed or grievance’ debate, the study of child soldiers has not. Child soldiers are seen as a degenerate product of ‘greed’ factors, in that armed groups recruit children for short-term material aims and the children or youths join for the same reasons.73 This ignores that child soldiers are agents, and that many choose to fight, often due to political motivations and not economic ones. Purely economic rationales for warfighting have been expressed in a number of ways. Whereas Gurr (1968) conceptualized this as occurring due to relative deprivation, Collier and Hoeffler (2004) saw violence more as a product of desire for economic gain. They sought to test potential explanations on the causes of war to see if political or economic explanations held stronger predictive value. The greed or grievance approach was flawed for two reasons which are worthy of discussion here, although others may note additional challenges to this simplistic dichotomy. The first is that the ‘greed’ explanations of civil war led to a pathologized view of developing world conflict in which ‘their wars’ are all about resources, compared to the glorious struggles of the Western past.74  The ‘greed vs. grievance’ debate not only led to an inaccurate view of why individuals actually fight, it also led to a series of counter-productive policies in which it was assumed  that if actors could be bought off, the war would end  (Ratsimbaharison 2011). This argument is used prominently in connection to child soldiers, in which it is argued that if we only paid the armed groups, they would release the children, ignoring that this has been tried, and it has failed. 75                                                           72 For prominent examples, see Collier (2000) and Collier & Sambanis (2005). 73 Examples that reject the Greed argument include Arnson & Zartman (2005), Bensted (2011); and Keen (2005). 74 This has been noted by many, including and Zeleza (2008, 15-19), Dunn (2010, 50-53); and Wai (2012, 163-166). 75 One example of this argument was earlier noted; Hamberg (2013), who argues that engagements with the SPLA in South Sudan were a success after the armed group was paid off. However, when armed groups are paid 42  Secondly, the greed vs. grievance debate is problematic as it assumes that political rationales for conflict actually can be distinguished from economic ones. This is not the case. To explain violence as solely the factor of either of these explanations – political or economic – is to obscure half of the causal story. These political and economic factors are not competing explanations, but instead are twin logics that are both usually at play when explaining organized violence.76 There is no armed group in the world that is not deeply reliant on resource mobilization, and as such, all have economic rationale for warfighting. This does not mean that they are fighting only to gain resources, and there is no armed group in the world that does not use a socio-political narrative to motivate its soldiers to fight. All ANSGs do this regardless of how ‘greedy’ outside actors perceive them to be. Even gangs in American cities - the rationale for whose existence is purported to be purely economic – use such narratives to urge their recruits to follow orders and adopt risky behaviour (Densley 2012). Both economic calculations regarding resources and the  grievances which are a constitutive factor of identity-narratives are constantly at play in the decision-making of leaders. The literature on counterinsurgency is useful at demonstrating how insurgents balance these concerns, and why we must understand the political and economic rationales of warfighting. It is commonly acknowledged in the counterinsurgency literature that armed groups need both material and normative resources to wage their struggle against both counter-insurgents and co-insurgents.77 The noted war correspondent and counterinsurgency theorist Bernard Fall once noted that a government that is losing to an insurgency is not being out-fought, it is being out-governed (1964, 36). Kilcullen has argued that any analysis of insurgency must simultaneously assess the role of legitimacy, the identity of the fighters, and the need for resources (2010, 1-12).  By failing to take into account the political rationales of why armed groups use child soldiers, the study of their demobilization has become stagnated in the ‘greed’ debate and has failed to take advantage of advances in the broader study of armed conflict.   Political Opportunities and Resource Endowments ANSGs and their leaders must engage in a strategic calculation between immediate aims, and more long-term objectives, while also engaging in legitimacy-seeking at both a local and international level. In seeking these goals, they seek a range of governance strategies depending on their own preferences and their interactions with local and international actors (Mampilly 2011). Decisions and strategies will vary wildly based on these interactions, as well as internal interactions between potential rivals, or between leaders, followers, and those in-between. As such, we must not only look at the ANSG’s own micro-incentives or broad strategic aims. It is also necessary to discuss the operational                                                           to engage in any kind of behaviour, their incentive structures shift. The SPLA went back to recruiting child soldiers soon after the Hamberg article was published (Human Rights Watch 2015). 76 In one example, Cramer uses the heuristic model of DNA, noting that both the economic and political explanations are the two sides of the double-helix that holds the DNA together and defines the armed group’s constitution (2006, 201). 77 Prominent theorists in the development of this strain of counterinsurgency theory include Kitson (1971), Galula (1964); Trinquier (1964); Nagl (2005); and Kilcullen (2009). 43  context in which the ANSG operates, both in terms of material resources and the political terrain of beliefs and norms. It is possible to think about this position in two ways. One focuses on economic rationales while the second emphasizes socio-political opportunities.  In terms of the latter, work in this area has been done by Jeremy M. Weinstein, who discussed the “industrial organization of rebellion” based on the economic and social endowments available to an insurgency at its outset. He argued that groups with low endowments would attract opportunist fighters, who would have low commitment and pursue short-term gain. On the other hand, groups with high levels of endowments would attract activist rebels with long-term commitments (2007, 7-16). These early endowments essentially determine the future strategies of the armed group by pre-defining who is likely to be in it; ideologues or mercenaries. There is a shortcoming here, however. Saying a group has low “social endowments” is merely saying that the group has weak cohesion. Further, concluding that groups with weak cohesion must pay their fighters more doesn’t tell us anything about how these groups organize, or how they change over time. It just tells just tells us that mercenaries like to get paid whereas ideologues are more likely to support the cause. The problem with Weinstein’s approach is that resource endowments are not static over time, and organizations evolve and adapt in the face of ongoing conflict. Further, their social endowments fluctuate, as do the identity-narratives that armed groups use to hold their groups together. In Weinstein’s world of rebellion, fighters make solely economic decisions about how and when to fight at the outset of joining, and he does not note the ability of either the fighter or the armed group to shape or change their identity-narrative or their normative worldview. According to Weinstein’s typology, groups who use child soldiers would all fall in the ‘opportunist’ category; since the lack of endowments forces them to bring in less-than-ideal recruits. They can’t even afford mercenaries, so they use children. Given this, we should not expect the group to ever voluntarily demobilize its recruits unless the leaders are being directly and immediately bought off (or unless their endowments change dramatically). Beber and Blattman, in discussing the logic of child soldiering, have also made this prediction (2013). However, this prediction regarding child soldiers does not match the empirics, as will be discussed in later chapters; many armed groups are demobilizing even though they are receiving no tangible material goods. This prediction also fails to account for the simple fact that many child soldiers volunteer to fight for ideological reasons.78 It is a fallacy to assume that child soldiers are abducted, brainwashed or coerced to join. The opposite is true; many are volunteers, and in general, child soldiers display the same variety in motivations as adults (Schmidt 2007). In cases such as these, Weinstein’s analysis does provide some understanding how the endowments available to an organization will affect its organizational strategies but fails to explain behaviour that deviates from this.  As such, it is necessary to marry the economic agenda of ANSGs with an understanding of how it determines and is also influenced by their ontological view of reality. Bringing in Kalyvas’ analysis of the                                                           78 McBride discusses this specific problem of the existence of volunteer child soldiers, and the debate about whether or not a child holds the agency and responsibility to volunteer themselves for combat (2013, 6-7). 44  ontology of violence helps to show how master cleavages can be constructed to create alliances between actors within a conflict system.79 Kalyvas argues that though a conflict may contain a master cleavage – be it one of ethnicity, nationality, or ideology – most people at the local level join the struggle for individual and local reasons. As such, the narrative holds together groups even if not all fighting for the cause are primarily motivated by it (2003). However, this does not account for individuals actually believing in the cause. I therefore extend this logic, accounting both for local micro—incentives as well as for broader macro-cleavages, in order to see both the economic and political connections between them.  This, in turn, allows us to see how armed groups use these narratives in order to include legitimacy as a strategic good from international actors, and assert it to domestic and international audiences. These narratives not only create divisions and cleavages, as Kalyvas discusses, they can also be used in order to strengthen alliances and increase the perception of the ANSG as legitimate to actors on the other side of the cleavage. Analysing how the narrative is constructed by the ANSG, therefore, involves more than behaviour that can be deduced simply based on resource endowments or even micro-incentives in the short term. Though immediate interests are certainly a factor, there is a need to look beyond them, to look at both long-term interest and how both identity and perceptions of appropriate behaviour change in the process. Tarrow has noted that the openness of a political system to certain kinds of mobilization will determine whether or not it is possible for a social movement to emerge (1998, 16). Adding in the political opportunity structure with respect to ANSGs contributes two key points to Weinstein’s industrial organization of rebellion.  The first of these is an acknowledgement of the level of threat posed to the insurgency by other actors, both the state and other insurgent groups. This, in turn, allows a second key point; the acknowledgement that the strategies of the insurgent group are dependant not only on resources available, but on the strategic context in which they find themselves vis-à-vis other belligerents. As strategic actors, the decisions undertaken depend fundamentally on the actions taken by their interlocutors. The reason for one ANSG demobilizing its children is fundamentally tied to the actions of other ANSGs, and whether or not they are demobilizing their child fighters as well. If a group is losing with support, and it sees another armed group in a similar position that demobilises, this can encourage that group to do the same. We cannot simply predict behaviour as a function of resources and threat but must also assess the system in which the ANSG operates, and the strategic advantage that can be gained over other actors through the process of becoming perceived as a legitimate actor. These two aspects are vital to mapping out the strategic landscape on which insurgent groups operate. Understanding the role of the political opportunity structure helps to provide understanding of when and how armed groups may be able to accept norms, rather simply viewing norm acceptance as a function of the leader’s decision. As noted earlier, not all armed groups can accept all norms. In order to explore this further, and the role                                                           79 This was conceptualized in Kalyvas (2003), though was later developed in Kalyvas (2006). 45  that legitimacy plays in this, discussion will turn to the concept of legitimacy, before analyzing what internal characteristics of the armed group will condition whether a norm is accepted, and if so, how.  2.3.3 Legitimacy and the Child Soldiers Norm In discussing the relevance of the concept of legitimacy to non-state armed groups, it is essential for this project to have a clearly defined concept of legitimacy. This section will establish such a definition, asserting that it must be understood as an intersubjective notion which results from multiple actors and not a single authority.  It will then go on to argue that legitimacy is sought by certain armed groups, and that they view salient norms – such as the child soldiers norm – as a means to access that legitimacy. Legitimacy will therefore be here defined as the intersubjective perception of a rule, organization or institution which actors believe are desirable, proper or appropriate because of its favourable outcomes, and/or procedural fairness. Several aspects of this definition deserve a brief exploration. This definition relies on Suchman’s definition, which emphasized the characteristics of desirability or appropriateness, within a set of socially accepted norms (Suchman 1995). As Hurd noted, legitimacy is both an internal and an intersubjective concept (2008, 8). Research on ANSGs must focus on the intersubjective qualities of legitimacy, as for an ANSG, legitimacy is not granted by a single actor, nor does legitimacy follow automatically from the predominance of a single organization or institution in a given political arena. Instead, legitimacy is bestowed by multiple actors whose collective intentionality create that legitimacy. As Coleman has noted, in addition to multiple potential audiences, both actions and actors can be deemed legitimate (2007, 20). This study explores 1) when and how legitimate action can affect an actor’s legitimacy, and 2) how political actors adopt certain behaviour in order to increase their legitimacy, even though there is no single actor able to bestow it. The ANSG’s legitimacy is a product of the multiple audiences, including but not limited to: states, other ANSGs, international organizations, and NGOs. It is difficult to empirically verify the existence of both causal and constitutive effects of legitimacy, as it is an intersubjective quality and is therefore only indirectly available for study.  Additionally, there is little research on how armed non-state actors can be influenced by international actors using legitimacy as an incentive.80 Though the literature on counterinsurgency provides some guidance in overcoming this, its shortcoming is that it treats the normative worldview of armed groups as static or responding solely to exogenous change, without allowance for non-state actors to re-define their own outlooks. Instead, the paradigms are treated as “doctrine” or “ideology.”81 As such, we must embrace an understanding of legitimacy which not only allows for change, but also accepts that it is not                                                           80 One of the few examples of this is Hoffman and Ulrich (2011), although they look mainly at the practical implications in engaging armed groups. A second example is Bongard and Somer, who compare Geneva Call’s approach to other strategies that can be adopted to compel armed groups to follow international norms. However, they summarize the approach and experience of Geneva Call as an organization, and do not assess their effectiveness overall, or place their actions in broader theoretical perspective (2011). 81 As in United States Army/Marine Corps (2007) and Kilcullen (2006). 46  granted by a single actor, as there are at times competing sources of legitimacy, and the members of any group which bestows legitimacy may have dissenting opinions. Kilcullen has argued that both counterinsurgents and insurgents see the capture of legitimacy as a sine qua non of success (2010, 1-12).  This struggle by armed groups to be seen as a legitimate authority is evident in both the recruitment and demobilization of child soldiers. The insurgent actors who demobilize are doing so in order to increase others’ perception of their legitimacy of transnational advocacy networks, and by extension both civilian populations and states. This dissertation argues that they are therefore making immediate and costly material sacrifices in exchange for future undetermined gains that they believe they will receive once they are perceived as being legitimate. In this case by acting like a governing authority and protecting children, the ANSG hopes to be treated as such in other areas which will improve its position politically in ways they cannot accomplish while continuing to fight using their child recruits. This begs the question; if the armed groups are seeking legitimacy, what does legitimacy mean to them? In the struggles of insurgency and counterinsurgency, belligerents seek to gain control of both normative and material resources to gain advantage over their opponents. This study will question the extent to which immediate needs may be traded against the potential future gains that may come from being seen as a legitimate actor.  However, it will also be argued that armed groups make such decisions within the broader framework of an identity-narrative which conditions their behaviour. These are the first steps that can lead to internationalization of norms of appropriate conduct, and therefore socialization into a larger community.82 However, this will not be true in all cases, and it is important to note that ANSGs will be engaging with the norm from a number of perspectives. Some may have held the norm beforehand, though not acted on it, while others may have tried and failed to demobilize other child soldiers. Still others may have a locally defined image of the norm, others may reject it outright, while others still may have no clear stance on it.  In terms of how norms spread, and what characteristics of norms make them successful, a large body of research has been conducted. For example, there is analysis on how the landmines norm spread among states, and the role of non-state actors in this process.83 Price has noted in the case of the landmines norm that the salience of a norm can be related to how it is grafted onto pre-existing ones (1998). He has further theorized the process by which the acceptance of this norm emerges as a form of customary law (2004). Sundstrom has further noted that the success of a norm is related to how it fits into pre-existing societal views (2005). O’Dwyer has argued that the success of the landmines norm internationally was due to unique conditions such as geopolitical conditions, the effects of globalization, the attitude of other actors in the international system, how campaigners financed the issue, how they linked it to pre-existing norms (2006, 77). Additionally, Carpenter has discussed how a norm can become salient due to its position in international NGO networks (2011). This literature gives us a view of how a norm moves through groups based not only based on internal characteristics of a political entity – be it                                                           82 The process of socialization was theorized and expanded upon most notably by Johnston (2001), as well as Checkel (2001) and Checkel (2005). 83 Prominent examples include Price (1998) and Rutherford (2000). 47  state or non-state – but also on external characteristics of the political environment in which said entity operates, as well as its connections to other actors.  For the purposes of my research, I take the salience of the child soldiers norm as a given; I do not seek to explain why or how it rose to prominence within transnational advocacy networks. When these networks engage with armed groups, the armed group is not free to discuss whatever norm it chooses, it will discuss the ones of interest to the NGO. For example, when Geneva Call speaks to armed groups, they have the option of discussing the child soldiers norm, the landmines norm, or the norm prohibiting violence against civilians. The armed group does not choose whatever norm it wishes, it selects from a menu the same way that a customer does in a restaurant. The transnational advocacy networks have already limited that menu to several options.  The addition that this research seeks to add to this understanding is to look at ANSGs as the norm receiver – given that a norm has become salient in transnational advocacy networks – and ask how and why would ANSGs follow that norm in order to be considered a legitimate actor? What internal and external characteristics lead it to this form of behaviour? Their means of accessing this legitimacy is through using a norm. As Hurd noted, there is no reason that the study of international norms should not be considered side-by-side with the study of strategic behaviour; the two are not opposites, but complementary (2008, 310). Through acting ‘as if’ the norm matters, it may well be the case that further down the line the norm is internalized into the belief structure.84 Though there is previous work on why violent non-state actors will follow a norm, it is largely based solely on a short-term cost benefit analysis that fails to explain why ANSGs would demobilize even though there are no immediate gains.85 As Bongard and Somer noted, there is little work done on ANSGs to understand the effectiveness of monitoring mechanisms such as Geneva Call’s ‘Deeds of Commitments’, or on their effectiveness in encouraging ANSGs to follow the landmine or child soldiers norm (2011). This study of ANSGs and the child soldiers’ norm can therefore provide valuable theoretical and policy-relevant conclusions because insurgents using children as recruits have a clear conflict of interest between following IHL and waging war. The pursuit of the one comes into conflict with the other. Leaders are therefore making two key strategic decisions when embracing a norm (1) that the legitimacy will bring gains in the future, and that (2) the members of the group will find the norm to be credible. In order to bring together how armed groups engage with norms to access legitimacy, the final step in this discussion will assess why some norms are? be incorporated by an ANSG and why are not? In order to address this, it is necessary to understand the pre-existing beliefs of those                                                           84 Checkel has discussed the manner in which acting ‘as if’ an idea matters and will adopt a norm due to strategic calculation, however this may lead to later internalization (2005, 808-810]). Further, Finnemore and Sikkink have discussed how actors may follow a norm simply because it is popular, and they are concerned with their reputation, however at the desire for conformity may lead to later internalization of the norm in question (1998, 896-901). 85 Two specific examples that make the argument that rebel groups will only respond to material incentives in order to follow the child soldiers norm are Beber & Blattman (2013) and Hamberg (2013). 48  within the organization, in order to assess how a decision to engage with a norm by a leader will be seen as credible to those in the armed group.  2.3.4 Norms and Beliefs Given that some ANSG leaders may choose to accept a norm, how is this norm brought into the armed group’s pre-existing beliefs and identity? Must all of the members of an armed group ‘believe’ the norm? If not, what relation does it have to pre-existing beliefs? After reviewing the common wisdom on norms in the study of international relations, this section will turn to the discussion of norms and armed groups. Finally, it will seek to answer the above question about previous beliefs by looking at the role of an ANSG’s identity-narrative, in order to theorize how a leader’s strategic decisions must match the common beliefs held by the group.  Norms in International Relations The fact that some ANSGs are committing to various aspects of IHL – and that it is very costly for them to do so – suggests the importance of legitimacy, even to armed groups who are frequently depicted as the most violent and illegitimate of any group of actors. As noted above, one way to analyze this is to follow their acceptance of a single norm, and the matter that they demonstrate acceptance of this norm. It is important to clarify, however, what a norm means, to the organisation as a whole and the individual minds within it. Farrell has noted that norms not only proscribe and prescribe behaviour, they also define us and our communities (2005, 1). Far from being simply an idea or belief that exists in a single mind, norms are collective understandings of behaviour held by a community as a whole as well as in the practice of that knowledge.86  It is beyond the aims of this project to solve various unsettled debates on the role of norms, and as such, I will remain agnostic on several ongoing debates. Gilardi noted several unsettled debates with regard to norms, including how they travel across levels – city, state, region, state, for example – as well as the relationship between ‘good’ and ‘bad’ norms. She also notes the difficulty of empirically measuring norm diffusion (2013). It is into this third debate that the research here falls, since an essential aim of this project is to demonstrate norm diffusion of norms, in this case from transnational advocacy networks to ANSGs. In the last several decades, a large amount of literature has emphasized the role of norms in international relations (Risse and Sikkink 1999). This has spread through both into the study of norms and conflict, as well as the relationship between non-state actors and the diffusion of norms.  Within the literature on norms and conflict, a key approach has been to analyze where norms come from, how they diffuse, why they do so, and regarding which forms of behaviour. The literature on norms places a great                                                           86 In this sense, the emphasis on both understanding and practice is drawn from Wendt’s discussion of social knowledge and social structures existing, “not in actors' heads nor in material capabilities, but in practice.” (Wendt 1995, 73-74). 49  emphasis on international and transnational actors making norms salient on the international level. Keck and Sikkink have discussed how transnational advocacy networks can bring a single substantive issue to the fore (1998). Finnemore and Sikkink sought to answer the questions of where norms come from and how they change by conceptualizing the life cycle of a norm from emergence, to “norm cascade” to internalization (1998). Norms can rise to prominence through persuasion, however they can also be a factor of practical necessity and technological change.  Zacher has discussed the territorial integrity norm that exists between states, noting both exogenous changes and ideational factors as contributing to the norm’s acceptance (2001).87 Taken together, this research questions both how and why a norm comes to prominence, and how characteristics of the norm, the actors and the context can affect whether a norm is accepted. The manner in which norms become adopted by the insurgent group and brought in to their beliefs is of fundamental importance both for the goals of conflict resolution, and for scholarly research on norms in general. The literature of international relations is dominated by approaches that see norms developed by individuals or groups who then seek to convince states at the international level of the prescience of the given norm.88 The movement of norms has also been discussed in the manner in which non-state actors including NGOs, terrorist organizations, and civil society actors are able to use symbolic politics and moral suasion to redefine the interests and actions of states (Reus-Smit, 2005).  Though valuable in showing how norms acquire salience in international relations, approaches that focus on the international movement of norms often allow solely a limited agency on the part of the local actor as they remain a recipient of norms. Additional literature has additionally sought to analyse how norms fit to local normative paradigms (Sundstrom 2005). Further, Acharya has discussed how norms are localized and moulded to fit local worldviews (2004). These approaches provide a valuable contribution to the study of norms, as they allow agency on the part of the norm recipient, rather than seeing them as solely a receiver or rejecter of an externally defined norm. These allow not only for the fact that norms may not be transferred (as well as establishing the local cultural and material reasons for why this may be so), they also allow for the norm to be changed as it moves from an international actor to a local one. Rather than seeing norms as either adopted or not based on a moral entrepreneur’s attempts to convince, the exact opposite approach will be adopted. The approach of my research is not to adopt a view standing outside the ANSG looking in with the goal of convincing the ANSG. Instead, my approach is motivated by a desire to stand analytically within the armed group and look out, in order to understand how the armed group comes to see the adoption of a given norm as a desirable option based on its own paradigm, its strategic calculations and its long-term political aims.                                                            87 For further discussion on the spread of norms, see the previously discussed example of the diffusion of the norm related to the prohibition of anti-personnel landmines. 88 This has been discussed by Price (1998); Finnemore & Sikkink (1998); Keck & Sikkink (1998); Risse, Ropp & Sikkink (1999); Carpenter (2007); and Sikkink (2011). 50  Norms and Armed Groups As noted above, a key question in the discussions of norms and armed groups relate to who  must internalize the norm, who makes the decision to do so, and the role of the leader in this. Recent research has begun to acknowledge the role of norms and legitimacy in armed conflict, as they relate to armed groups. Adler-Nissen, has discussed the role that fear of stigma plays in guiding the actions of political actors in international relations (2014). Further studies have directly looked at norms and armed groups. Hamberg has discussed how armed groups can be encouraged to follow IHL using direct payments by international actors (2013).  Jo has discussed the extent to which rebels can become compliant with IHL in order to seek legitimacy (2015). It is useful to look at both studies closely. Hamberg discussed the case of the Sudan People’s Liberation Army (SPLA) demobilizing child soldiers, arguing that they only did so because the US was offering positive incentives in the form of financial gain to the SPLA leadership. He argues that is proof that of the role that positive incentives play in encouraging actors to demobilize, citing the SPLA as the only successful example of demobilization of child soldiers even as civil war continues (2013, 172).   Unfortunately, Hamberg’s claim that SPLA ceased using child soldiers is not accurate, and therefore he draws incorrect conclusions. Many other groups have demobilized child soldiers during conflict, as the following empirical chapters will discuss. In terms of the role of material incentives, if anything, the provision of them only encourages the further recruitment of child soldiers, in exchange for being rewarded when they are demobilized.89 Hamberg cites 2001 as a successful year in which the SPLA demobilized its children (2013, 149). This is an odd claim, given that the SPLA was known to have consistently used child soldiers up until the time of writing in 2017.90  Approaches such as Hamberg’s focus solely on leaders, however there is more at play than a leader’s incentives. What if there is opposition from below? It may be the case that a leader is willing to sign an agreement denouncing the use of child soldiers. It is something else entirely for the group to actually demobilize, to embrace the norm into collective identity, and for it to be supported by both the mid-level commanders, and the rank-and-file soldiery. Further, children will often seek to join an armed group even if the armed group is trying to avoid recruiting child soldiers. A leader can’t simply snap her fingers and follow a norm; she must go to great efforts to ensure the norm is institutionalized into the organization at all levels, and that individuals are socialized to support it.  The second study – that of Hyeran Jo – helps to lay a foundation for this study. Jo uses a quantitative approach to test for evidence of legitimacy-seeking behaviour in armed groups. Jo uses the following three variables: (1) the presence of a political wing, since this demonstrates longer-term aims; (2) the aim of the armed group to secede, since this means they will act in a more state-like manner; and                                                           89 Although no study has been completed on payments for demobilising child soldiers, programs of ‘money for weapons’ in disarmament campaigns have been shown to create circular weapons trade, and have led. to a net increase in weapons imported into an area, as entrepreneurs seek to make money from such programs. This has been discussed by Rogers (2009, 244). 90 As reported by Human Rights Watch in late 2015 (2015). In addition to the promise the SPLA made in 2001 to follow the child soldiers norm (that Hamberg cites as a sign of success), they also made similar statements as recently as 2012 (VOA 2012). 51  (3) whether they have external sponsors who care about human rights (2015, 29-30). These are then studied with relation to three aspects of international law, namely: (a) the killing of civilians; (2) the use of child soldiers; and (3) granting or denying of access to ICRC delegates. She generally finds results in favour of her hypotheses regarding legitimacy seeking by armed groups. For example, with regard to the child soldiers norm she notes that whereas 20% of armed groups without a political wing use child soldiers, 15% of those with a political wing do so (2015, 158). She further finds that of 165 non-child soldiering groups, 21 % are secessionist, while of 27 child soldiering groups, 26% are secessionist (2015, 159). The third proxy – support by governments that care about human rights – was inconclusive due to lack of cases, most likely because states that support human rights (i.e. democratic states) rarely support armed groups (2015, 159-160).  The proxies used by Jo to try to provide a measure for norm compliance are problematic however. It is likely that secessionist groups do not engage in abuses since they usually only fight in areas populated by their own supporters, making violence for violence’s sake problematic regardless of international concerns. The third proxy (support by regimes that care about human rights) is also of limited use, because democratic regimes are less likely to support rebel groups, and when they do, they are only likely to support rebel groups with ideational affinity for democratic values (San-Akca 2016, 67-105). This means such rebel groups are less likely to violate norms in the first place. Also, it is an unhelpful tautology to say that armed groups who care about the opinion of international supporters are more likely to care about the opinion of international supporters. As noted earlier, a central aspect of this argument is that ANSGs will make material sacrifices in order to follow a given norm. The issue of child soldiers is demonstrative of this due to the fact that in order to follow the norm, strategic actors must make sacrifices of a resource vital to continued insurgency; namely, the recruits who do the actual fighting. In order to understand how and why ANSGs will adopt or reject a given norm, it is necessary to adopt a view which sees these processes as endogenous to the course of war. Elisabeth Jean Wood adopted such an approach to understand why individuals would undertake the risky action of rebellion, seeing explanations as intrinsic to the process itself (2003). Mampilly has also observed that governance structures and practices are constantly transformed over the course of the war by a  variety of conflict produced dynamics endogenous to the fighting (2011, 15). Norms related to individual recourse to violence are other-regarding in that they have meaning only in reference to a wider community. It is therefore necessary to adopt an approach which is able to understand how ANSGs define and reconstruct their own normative worldviews, and how this is endogenous both to conflict itself, and the competition for legitimacy that comes with it. It is important to note, however, that certain norms have more salience than others in given contexts (be it domestically or internationally), while some will require a greater cost to follow than others.   Child soldiers are a microcosm of normative change, and ensuing research and analysis in this area sheds light on intra-state conflict in a way that requires a fundamental re-thinking of the interests of insurgent groups who are often seen as indifferent to normative concern.91 We must move past these                                                           91 Allen & Vlassenroot discuss this problem, with specific reference to prevailing pejorative myths related to the Lord’s Resistance Army (LRA) of Uganda (2010). 52  partial understandings, and instead theorize the manner in which armed groups both acknowledge and compete for the existence of normative resources. By adopting this view, it is possible to construct rationalist accounts of the use of norms which are not entirely dependent on materialist explanations of behaviour which would assert only material gain as an explanation for norm compliance. The approach adopted here is therefore to treat ANSGs as political actors interested in the same normative goods as governments, but who are also able to manipulate political narratives to a certain extent. At the same time, we must see that there are long-term effects of these decisions that may not have been intentional, or that may be occurring as a process of strategic social construction. Even though this view is often rejected by the dominant trends in the literature on conflict, such an assumption does not require a great leap of faith. All that it requires is that we view armed groups as behaving in ways similar to all other political actors when it comes to the complex interaction between resources, legitimacy, norms and identity. The politics of insurgency and the use of norms by ANSGs is a hard case for demonstrating the existence of legitimacy and its effect on the behaviour of various actors. ANSGs by their very nature reject the legitimacy of the sovereign state they are fighting and are not held to international commitments in quite the same way that states are. Most NGOs and inter-governmental organizations refuse to engage with ANSGs directly. Most of those that do engage with ANSGs do so only with consent of the government in whose territory the ANSG operates. Further, those organizations that do (for example the humanitarian organizations MSF and the ICRC) rarely publish results of their interactions. If legitimacy can thus be demonstrated to be a key factor in the decision-making of ANSGs, it challenges the commonly perceived notions of insurgent groups as operating more narrowly according to a cost-benefit decision-making process in terms of  how they conduct themselves in conflict.  Of additional importance is that there appears to be a kind of lock-in effect once insurgents ‘sign up’ to a specific norm and agree to monitoring mechanisms. Though insurgents are able to change the norms they assert (for example, by getting rid of child soldiers), doing the opposite (reversing on the norm) could incur large costs. This may include losing the support of a state-sponsor or diaspora communities abroad.92 As of yet, there is no example of an insurgent group that has demobilized its child fighters and agreed to verifiable monitoring mechanisms (in the form of Geneva Call’s Deeds of Commitment), only to change its mind and remobilize them again. Archarya has discussed the stickiness of norms, and the manner in which actors will not want to act contrary to a given norm once it is adopted if they have expectations that other actors will follow them (2005, 102-103). In relation to ANSGs and IHL, this demonstrates the manner in which decisions on the subject of child soldiers can lead to a ‘tying down’ of the ANSG using a web of norms which the ANSG itself brought into its own worldview. Therefore, to answer the question that opened this subsection, this localization process involves not only decisions by leaders, but also the pre-existing identity-narrative of the group.                                                            92 Jo has discussed the manner in which democratic states or states that care about human rights can impose costs on rebel groups (2015, 105). 53  Narratives and Normative Independence In armed conflict, leaders of armed groups do not simply issue orders; the orders  must be accepted and followed. Orders issued much match, in some way, the interests and desires of followers. Mampilly has argued that governance in rebel groups is not merely a top-down process, as is often assumed, but is an interactive process (2011, 15). The same is true with norms; when leaders decide to engage with a norm, they must convince their followers to follow their decision. They cannot simply order them to believe something. When they try to convince their followers, they are unlikely to make arguments asserting they wish to follow a norm because it will gain them access international legitimacy, or because it will yield undetermined future gains. Rather, they will use the language of beliefs, asserting to the group that it is the right thing to do, and often, that it is a belief already held by that group. In order to rally troops to their cause, and to keep them together, armed groups are no different from any other form of political organisation, in that they hold a set of beliefs shared broadly amongst the members of the community. It may not be the case that all people hold all of the same beliefs. Many, in fact, may hold contrary beliefs. When the leaders of an ANSG seek to incorporate a new norm into the belief system of the armed group, it must be compatible with their pre-existing beliefs. Fighters in an Islamist armed group, for example, would find a norm about gender equality in the military to be a hard norm to accept, while they may favour a norm to prohibit the use of landmines. The opposite may be true for an armed group in the Burmese highlands, who rely on landmines, and have a less rigidly patriarchal culture. The worldviews of each are starkly different, which conditions the extent to which the norm can be accepted by the group. But what exactly is this intangible concept of ‘worldview’, and how does it relate to beliefs, identity and ultimately the strategic behaviour of ASNGs? A worldview is typically defined as the orientation of an individual or society encompassing the entirety of the individual or society's knowledge and point of view.93 This is a symbiotic relationship here between both the beliefs and institutions that support them; the one is meaningless without each other. Indeed, beliefs and norms are a central part of most definitions of social institutions, be they formal or informal.94 It is not the goal here to wade too deeply in the conceptual debate on the difference between the terms worldview, ideology, belief system or Weltanschauung. Rather, the purpose here is to understand the social nature of these worldviews, and the manner in which it they bring meaning both to the individual action in the social context, as well as to the social context itself. A paradigm is typically understood as a set of assumptions, concepts, values, and practices that constitutes a way of viewing reality for the community that shares them. Culture is often defined as a set of historically transmitted                                                           93 Sometimes used in its original German – Weltanschauung – the term worldview derives from Kant. It has become central tenant of German schools of philosophy. Definitions of it are discussed at length in Naugle (2002). 94 Examples who point to the centrality of the roles of beliefs in institutions include Greif & Laitin (2004, 635). Even those who do not use the word ‘belief’ define institutions based on socially shared rules intended to bring agreed upon results (2004, 727). This is most true of the approaches of sociological and historical institutionalism, though not necessarily of purely rational approaches (Hall and Taylor 1996). 54  pattern of meanings embodied in symbols, a system of inherited conceptions expressed in symbolic forms by means of which men communicate, perpetuate, and develop their knowledge about and attitudes toward life. (Geertz 1973, 89). Sire defined a worldview as “a commitment, a fundamental orientation of the heart, that can be expressed as a story or in a set of presuppositions (assumptions which may be true, partially true or entirely false) which we hold (consciously or subconsciously, consistently or inconsistently) about the basic constitution of reality, and that provides the foundations on which we live and move and have our being.”95 These concepts are useful as framing devices, however a cultural unit is too broad; for example two armed groups may share the same culture or ideology but be mortal enemies and have different stances on a norm. The concept of worldview, meanwhile, is too rigid and fixed to religion, in which the experience is an individual’s not a group’s. We must move a step farther in our understanding of socio-political ideations in this case, as we must come to grips with the idea that the outlook of the ANSG is constantly in flux and being re-defined. Further, we must identify more specifically what holds a group together (beyond simply saying ‘culture’) in order that we can understand what sets armed groups apart from each other, even when they share culture. Bringing in the concept of narratives helps us to accomplish that.  Although the term identity narrative is used in the study of armed groups and in political theory in general, there is no clear definition of the term.96 Definitions of narratives can be drawn from literary theory. At its most simple form, a narrative is defined as a representation of an event or series of events. Whereas at first this may seem a simplistic definition, there is much contained within it. A representation of events inevitably involves a story with meanings, as well as a discourse and language that assert meaning to the story itself beyond a simple list of happenings (Abbott 2008, 13-15). Nelson stresses the act of constructing narrative as an act of self-definition; the narrative constructors are selective in what they depict and are interpretive of those meanings they carry, thus allowing self-definition through moral agency (2001, 1-13). Davis argues there are three schools in literary theory regarding narrative; the first being narrative as a relaying of events, the second being a manner of speaking about those events (a discourse) and the third focussing on the relationship between narrator and audience (J. E. Davis 2002, 10). It is the consumption of the interpretation that is vitally important; narratives are created not only for the in-group, but the out-group as well. I define the identity-narrative as a set of changing beliefs, concepts, values, and norm-based practices that constitutes a way of viewing reality for the community that shares them and is asserted against counterparts from other communities.97 At its center, the narrative is both a story and a rallying                                                           95 This definition is found in Sire (2015, 142), although similar definitions are found in Naugle (2002).  96 Two prominent examples include Christia (Christia 2012) and Whitebrooke (Whitebrook 2014) 97 In addition to the sources and terms noted above, this definition draws from a ride range of definitions of aspects of belief and identity in the study of politics and, more broadly, social psychology and the study of both identity and narrative (to the extent that it influences identity). As such, this definition draws from several sources, drawing from other concepts while also seeking to clearly define identity-narrative as a term distinct from other concept. These include Sanin and Wood’s definition of ideology, as systematic set of ideas that includes the identification of a referent group (a class, ethnic or oth r social group), an enunciation of the grievances or challenges that the group confronts, the identification of objectives on 55  flag, serving to unite the group while also providing the individual means to alter and change the whole. There are few if any individual norms or beliefs in the ANSG’s paradigm that are completely rigid and unchangeable. Whereas a group’s overall narrative (through which individual norms are interpreted) may be resistant to change as a whole, the individual norms that constitute it are in constant flux. For example, though the Taliban’s worldview of militant Salafist Islam is unlikely to change, the norms that comprise it change on a regular basis, be they with respect to children in conflict, the appropriateness of certain weapons, and even the role of women in society (Gray 2012).  Further, the concept of identity-narrative helps us to move past the concept since it allows for disagreement within the group, and opposition outside it. Concerning the former, although they share an identity narrative, individuals may disagree about a specific norm, while adhering to the group’s identity-narrative as a whole. Few Christians would assert they hold every belief contained within the Old or New Testaments, such as norms related to slavery, rape and gender inequality. However, Christians would still hold to the dominant identity-narrative associated with Christianity and Christ, while disagreeing about a number of specific beliefs. Further, identity-narrative allows for opposition outside the group. Benedict Anderson referred to nations as “imagined communities”, since its members believe they share commonalities across distances making themselves members of the same community (1983). Whereas nations may imagine themselves as having a enemies in those ‘outside’ the community, for an armed group, there is a very real enemy that is trying to kill them. Further, an identity-narrative may be fixed to a nation, but it need not be; two armed groups may be mortal enemies while still belonging to the same nation. Identity-narrative has a distinct value-added in this sense as it allows for agency on the part of the individual. When engaging with a social narrative, an individual is not simply a passive recipient of that narrative. Rather, they play an active role in defining and redefining that identity. 98 The concept of narrative adds to the critical discussions of worldview undertaken earlier, namely that the individual plays a role in the storytelling, and that the narrative is not solely in the hands of the elite. Although the leader may make certain decisions (for example, signing a Deed of Commitment with an organization like Geneva Call, who uses moral pressure to convince ANSGs to follow IHL), the response of those at the bottom to it is equally important in determining whether or not it will be followed. Many see the role of identity in conflict as superfluous, simply a flag to rally the troops when it is material interests that                                                           behalf of that group (political change or defence against its threat, and a (perhaps vaguely defined) program of action (Sanin and Wood 2014, 215). My definition moves past this, however, in stressing that uniqueness and specificity of an identity narrative that separates it from others. Two groups may fight and share an ideology or culture, but they cannot share the same identity-narrative. In this sense, my definition draws from the emphasis by Adbdelal, et al., that asserts that identity includes both content and contestation against others (2009, 18). 98 For discussion on how identity and narrative are treated, see Mackenzie, who provides a critical view of the literature on the subject from a perspective from the philosophical literature on personal identity (2008). Vignoles, Schwartz & Luyckz undertake a similar task from an integrative approach looking at trends across the social sciences (2011). 56  dominate. However, identity is a powerful motivating force which compels individuals towards both life and death.99 The reasons for which these narratives influence individuals – and how individuals influence narrative – are vitally important to understand how groups are held together, and how even the individuals at the bottom rung of the social ladder are able to influence the group as a whole. Canadian literary scholar Thomas King noted that the contents of the story are not as important as how the story is told and retold, saying “I get stopped every time I cross [the] border, but stories go where they want” (2012, 12). The quote demonstrates a key point; that leaders and institutions do not control narratives, rather, they move between people irrespective of elite control.  King also notes that the approach of the story-teller to the recipient of the narrative is to say; “"Take [it]. It's yours. Do with it what you will... But don't say in the years to come that you would have lived your life differently if only you had heard this story. You've heard it now." (2003, 29). It is the same with armed groups. The individuals at the bottom are involved in three key roles: as inheritors of the story, as weavers of the narrative, and as wearers of that identity. They both receive an identity passed on by elites, while also defining and redefining an identity for which they are the primary agents of creation.  The narratives of armed groups are not spread solely by the leaders at the top; they are told and retold by those at the bottom, taking the story and changing it by retelling it. Understanding this principle is vital in coming to an understanding of how an armed group both receives a norm, accepts it collectively, and in the process (re)defines it as a part of that group’s identity-narrative. Acharya theorized a process of norm localization, in which a local actor or group would reinterpret or redefine a norm, fitting it to local paradigms or ways of understanding (2004). Although leaders are identity entrepreneurs, followers must buy what the leaders are selling. There is a bi-directionality not only within the group, but also between norms and identity. Once a norm is brought into an armed groups identity-narrative, this can then have a recursive effect on the group identity itself. Although identity-narrative conditions norm acceptance, it is possible that norm acceptance can then condition the identity-narrative at a later time. My argument, therefore, goes one step farther; not only the decisions of elites, but the actions of those in the group contribute to this process. To help bridge this gap, I theorize the concept of normative independence, in order to theorize why some armed groups may be more or less able to adopt a certain norm.  This research project intends to uncover more clearly the exact process of how and why ANSGs decide to trade an immediate material asset – i.e. their child soldiers – in order to seek recognition as a legitimate actor. As discussed in detail above, ANSGs will alter their behaviour, and the norms they adopt, based on their strategic need to do so. However, it is important to note that groups vary both in their ability and in their need to make such changes. I conceptualize their ability to make these changes as their normative independence, defined as the combination both of the material capability and the strategic desirability of an ANSG to change, replace, remove or strengthen the norms asserted within its narrative-based identity. Important to note as ensuing research will show, however, is that the ability of                                                           99 The role of identity as a motivating force in armed conflict over material interest was discussed prominently by Azar (1985), Gat (2006), Cramer (Cramer 2006) and Van Creveld (2008). 57  armed groups to change norms does not mean that they will seek to. A group that is winning with support may have the ability, but not the desire, to alter its behaviour with regard to a specific norm. Further, a group in this category is much more likely to have a large number of bandwagoners than true believes. Such bandwagoners are more likely to be less committed to the armed group, and to commit violations of norms (both those held by the ANSG and others).  An armed group that is losing with support is more likely to have true believers, since most of the bandwagoners are likely to have jumped ship. By making a credible pitch to these fighters that remain, leaders may be in a stronger position to convince the group to engage with the norm. Should they make a pitch that the belief is in line with the identity-narrative, the true believers still in the organization are more likely not only to tacitly accept the new norm, but also to integrate it and follow it. As such, leaders of groups that are losing with support are more likely to have a higher normative independence success than leaders of a group that is winning with support.   2.4 Conclusion  Engaging in such a study as this requires a rethinking of the dominant approaches of armed groups in scholarly analysis. ANSGs are seen as greedy, and their fighters as mere automatons who must either accept or reject the resources offered by the armed group as payment. The approach adopted here stands far apart from this, seeing both the fighters and the armed group as a whole not solely as a collective of homo economicus, nor remaining passive recipients of external events of identity construction and narrative weaving. Rather, I see both the fighter and the group as agents in this process, which compels them to not only engage with norms external to their identity, but also, to engage with actors local and domestic in order to seek higher levels of legitimacy. Now that this study has clear definitions of the concepts under analysis, and has assessed the common wisdom on relevant concepts, it is possible to move forward with the review of the empirics of this research study. In doing so, it will be necessary to refer back both to the gaps established earlier, as well as the rival explanations which I have ruled out as insufficient. In doing so, it will be possible to lay out the circumstances in which armed non-state groups will actively engage with the child soldiers’ norm in order to increase their international legitimacy.   58  Chapter #3: Myanmar  Low Intensity Conflict and the Child Soldier’s Norm   “They accuse us of using child soldiers. There were child soldiers in the [Karenni] army, but we did not conscript them… We are requesting the UN to send international verifiers to come so we can show the world we do not use [child soldiers] anymore. No one has come yet, but we keep asking.”  Khu Ooh Reh, General Secretary KNPP, Karenni Army – Kayah State, Burma, 2012.100   “You cannot trust a tiger who says he is a vegetarian.”   Colonel New Dah Mya. Karen National Liberation Army, upon being asked if he believes other armed groups will uphold pledges to demobilize child soldiers – Kayin State, Myanmar, 2012.101   3.1 Introduction  Few conflicts in the world are as different from Syria as the one in Myanmar.   Almost since independence, the Republic of the Union of Myanmar (also known as Burma102) has been plagued by near constant ethnic strife and conflict. Whereas the Syrian Conflict represents a relatively young conflict that has rapidly escalated into a high-intensity urban war, the Myanmar Conflict has remained predominantly rural in nature, and has not come close to the level of death and displacement that has been visited upon Syria. Myanmar’s plight has been referred to as “one of the most under-reported tragedies of the second half of the twentieth century (B. Rogers 2015, xxi). Although it is exceptionally difficult to estimate death tolls, between 200,000 and a million have been killed as a result of the fighting since it began in 1948.103 Although an exact number is difficult to establish, it has been noted                                                           100 Khu Ooh Reh.  Interview with author, 06-07-2012, Kayah State, Burma. 101 Colonel New Dah Mya, Karen National Liberation Army. Interview with author 19-06-2012. 102 As will be discussed, the terminological dispute regarding the name of the country is one which has not been settled as of the time of writing in 2016. Whereas proponents of the regime use the term Myanmar, most opposition groups prefer the colonial name Burma. I take no side on the naming issue, but following convention, will use official state terminology. 103 The higher number of one million was estimated by Senior General Saw Maung (Steinberg 2010, 12). 59  many times that in the 2000s and early 2010s, Myanmar had more child soldiers than any other country in the world.104  Figure 1 – Locations of Field Research in Eastern Myanmar. 105                                                           104 This claim was first made by Human Rights Watch (2002), and has been repeated as recently as Chen’s study on child soldiers along the China-Myanmar border (2014). 105 Original map is freely available under Wikipedia Commons, although edits to original have been made by the author. Original available at Wikipedia (2015). 60  Myanmar is host to dozens of armed groups, who shift allegiances regularly, and it is difficult to conceptualize the conflict as being between two “sides”, or as being an attempt by rebels to overthrow the government. The armed groups, government, and regional powers form a constellation of alliances, confrontations, and both overt and tacit ceasefires. Further, over the last number of years, there have been increased attempts by individual ANSGs and coalitions to seek to engage with transnational advocacy networks in search of legitimacy. In this regard, the issue of greatest prominence has been the child soldiers norm.106 After providing a background on the history of the conflict and its transnational characteristics, this chapter will analyse seven ANSGs. It will show that groups that are losing with support are engaging with the child soldiers norm, while others are not. For the groups that fall into this category, the role of the group’s identity-narrative will be discussed, since it is only once the decision to follow the norm has been adopted that the leaders must ensure that the norm fits pre-existing beliefs of the group.  3.2 Cases Myanmar presents a compelling context to conduct this form of research for several reasons. The first of these is that there has been relatively little attention placed on the Myanmar Conflict since it began. Although the Myanmar Conflict has caused as displaced almost a million people, they have not fled to Western Countries. Rather, they have mostly been displaced internally or just across the borders with Thailand. The Thai response has restricted the flow of refugees from moving beyond the areas immediately adjacent to the border, meaning that the kinds of refugee flows associated with Syria are prevented from happening before they have begun. As such, both the violence and the displacement are far from regions of strategic global interest. This means that the gains of following a norm are even lower than in Syria, while the costs of getting attention are even higher. Armed groups need not only demonstrate their compliance with the norm, they must first attract the attention of international actors whose gaze may be directed elsewhere. This is much easier said than done. The second reason is the number of child soldiers present in the country. For several decades, both the Tatmadaw Kyee – the Burmese Army, hereafter referred to as the Tatmadaw – and the opposition groups have been strongly reliant on the use of child soldiers. Although exact numbers are difficult to conjure, it is likely still the case that the Tatmadaw is one of the largest abusers of child soldiers in the world.107 In addition to the human tragedy that this represents, it also creates a context in which the state is a worse violator of the norm than the armed groups under study. Whereas in Syria, the regime uses paramilitaries who have child soldiers within their ranks, the regime itself has largely avoided the widespread conscription of children, preferring instead to use proxies to do its dirty work. The Tatmadaw is at the opposite end of the spectrum from this. A common estimate is that there are                                                           106 For example. Geneva Call has engaged with more armed groups on this topic than on any other (Geneva Call 2017). 107 Continued violation of the norm by the Tatmadaw was detailed in a 2015 report produced by Child Soldiers Initiative (2015b). 61  70,000 child soldiers in Myanmar.108 As such, Myanmar is a compelling context, due to the sheer number of child soldiers, and the reliance of all actors on their use. This substantially raises the costs for any armed actor considering institutionalizing the child soldiers’ norm. Thirdly, the large number of armed groups in Myanmar presents a diverse range of cases. Although frequently depicted as an ethnic conflict, there are few ethnic groups who are neatly represented by a single ‘side’ to the conflict, and instead, most ethnic groups have multiple competing armed groups claiming to represent them, while many others represent a diverse range of ethnic groups. As such, it can be demonstrated that pre-conflict culture is not a determining factor in whether or not one group or another chooses to adopt the norm. The cases under study in Myanmar, as well as their key features, are detailed below in Table 5. The information provided represents the size of the armed groups and their trends at the time of field research, as well as their size at the time of writing. This assists in indicating the armed groups as growing or shrinking over time, which in turn assists in establish each armed group as either winning or losing, and with or without support.                                                             108 The number was first estimated by Human Rights Watch (2002a, 3), but has been repeated by Chen (2014, 4),  62  ANSG Estimated Size 2013 Estimated Size 2016 Area of Operation Predominant affiliation Use of Child soldiers Winning or losing Support or no support ABSDF 1000109 450110 Kayin State KNLA / UNFC Violations, constant. Losing No Support KA 800 - 1,500111 600112 Kayah State UNFC Limited violations; declining. Losing Support KNLA 6,000 -7,000113 5,00015 Kayin State UNFC Limited violations; declining. Losing Support MNLA 1000 80015 Mon State UNFC Limited violations; declining. Losing Support PNLA 1000 40015 Shan State & Kayah State UNFC Limited violations; declining. Losing Support UWSA 15-25,000114 20,000 - 30,000115 Shan State Regime Widespread violations. Winning No support  Table 5 – Primary Cases from Myanmar Conflict                                                          109 Although the exact number is unclear, and no solid estimates exist for this time period, the ABSDF was known to have at least 1000 fighters into the 2000s, and this has been declining over time. 110 AS estimated by Myanmar Peace Monitor (2015). 111 As estimated in Tripathi (2014, 142) 112 As noted by Gravers (2014a, 170-171), and by Myanmar Peace Monitor (2015).. 113 As reported by Burma Centre for Ethnic Studies (2012, 1). 114 As estimated by Tin Maung Maung Than (2011, 34). Duell has put the estimate at 16-20,000 (2011, 82). 115 The estimate of 20,000-25,000 was reported by Jane’s Intelligence (A. Davis 2015), however there have been reports which put the number as high as 30,000, including Pagnucco and Peters (2015) as well as Bo Ze Kai (2016). 63  3.3 The Myanmar Conflict 3.3.1 Background   The naming of the Myanmar Conflict itself is a contentious issue, since there isn’t even consensus on the name of the country itself. Although it has been referred to as Burma, this name is not accepted by all as it has both colonial origins and ethnic bias. The demonym Burman comes from the dominant ethnicity in Burma, the Bamar (also historically named the Burmans). The authoritarian regime of Saw Maung decided in 1989, however, that since the name Burma was a name given by colonial rulers, the historical name of Myanmar (or Myanma in Burmese) better represented the country in its present form. 116 Many opposition groups, however, rejected this name change, since they stood in opposition to all of the authoritarian regime’s policies, while rejecting the name Myanmar as it was a Burmese word, while Burma was not.117 This created the somewhat odd situation in which ethnic minorities would call themselves Burmese but not Burman, while the regime (who are mostly Burmans) preferred to be called Myanmarese. 118  Although the conflict should not be viewed through a lens which sees it solely as a primordial ethnic conflict between warring tribes, the ethnic characteristics of the conflict cannot be ignored. Coming to an understanding of the size of each ethnicity in Myanmar is a difficult enterprise, and no census data can be trusted since the last serious attempt to conduct a census was done by the British in 1931 (Kramer 2010, 52-53). Whereas in Syria the regime in power is a minority controlling a majority, in Myanmar it is a majority group controlling many minorities. However, the regimes in both countries have had a clear vested interest in obfuscating the result of censuses in order to legitimize their stranglehold on power.  The regime has certainly fudged numbers (over-estimating the dominance of the Bamar while reducing that of everyone else), it also completely ignores a number of ethnicities that it considers to be foreign from Myanmar. Included in this are ethnicities such as the Rohinygya, who are entirely left out of the census.  Other ethnicities are subsumed within broader groups, for example the Wa are considered to be Shan even though they have a distinct language and culture.119 The Kokang, meanwhile, are considered to be ethnically Shan whereas in reality are ethnically Chinese.120 Further, the government doesn’t bother to include estimates for many parts of the conflict regions, areas almost completely dominated by non-Bamar ethnicities (Spoorenberg 2015, 2). The Myanmar regime divides the population into eight ethnic “races”, which are then divided into “ethnic groups”. Although it provides                                                           116 As reported by the Embassy of the Republic of the Union of Myanmar, Brussels (2013). 117 Tucker has noted the bias of the term Myanmar as a Burmese word whereas Burma is “ethnically neutral (2001, 229).  118 Although I do not intend to take sides on this terminological dispute, I will use Myanmar as the official name of the state, and Myanmar Conflict to refer to the conflicts, since this is the historical term for the ongoing dispute. Further, the names of locations are contentious. A full list of the changed names is listed in Crocket (2015, xv-xvi). I will also use the official names. 119 An in depth discussion of the regime’s racial categorization system is provide by Gravers (2007). 120 As discussed by Myint Myint Ku (2016). 64  data on the “races”, it does not for the ethnic groups.121 These classifications are widely challenged. The following are the official sizes of each of the largest ethnic communities in Myanmar122: • Bamar 68%; • Shan (including the Wa) 9%; • Kayin (Karen) 7%; • Rakhine 4%; • Mon 2%; • Kayah (including the Karenni) 1.75%; and • Kachin 1.5%.  Although easily contestable, these numbers give a rough understanding as to the proportional size of each group relative to the others. It is important to note that whereas the exact numbers of the Bamar are in question, their dominance is not, nor is the fact that most individual minority groups have quite small populations compared to those that control the regime. Regardless of which numbers are used, consensus is that the Bamar majority is by far the largest. This presents a stark contrast from Syria, in which a relatively small ethnic minority controls the state. Myanmar’s east – the location of fieldwork – straddles the borders with Thailand (see Figure 1 above). An area of cultural heterogeneity, its largest ethnicities include the Bamar, the Karen, Kayin and Shan, while the main religions are animism, Buddhism, Christianity and Islam. The religious divisions do not overlap with ethnic ones, creating further divisions in an already divided region (Tin Maung Maung Than 2005, 68-69). Even ethnoreligious labels defy the variance within them; amongst the Karen are smaller sub-groups including the Karennic speaking Karenni peoples (meaning ‘Red Karen’ in Burmese). This is further divided into more local groups, among them the S’Gaw, the Pwo, and the PaO showing the extent to which such broad pan-ethnic groupings hide the divisions both between and within the diverse groupings of ethnic, linguistic and religious difference. (South 2008, 13). Although in some ways grouped together as Karen, they have fiercely independent identities.   3.3.2 Overview of the Myanmar Conflict  The conflicts in Myanmar have been smouldering for decades and is often considered to be the longest lasting civil war in the world. While ruling colonial Burma, the British maintained a system of ethnic hegemony by allowing different ethnicities to control their regions, while also keeping different groups in charge of different parts of government (business, military, etc.). When the country was made independent in 1948, the traditional territories of the Bamar people were connected to areas of ethnic                                                           121 This categorization follows from a tradition begun by British colonial officials. The full details of the list are available at Embassy of the Republic of the Union of Myanmar, Brussels (2015). 122 As reported in Dittmer (2010). 65  minorities who had never previously been a part of any Burmese polity. These included the Karen, Karenni, Arakan, Kachin and Chin ethnicities, all of whom still have armed groups purporting to fight on their behalf to this day.123  Whereas during World War Two, independence leaders including Aung San and Ne Win sided with the Japanese forces, many of the minority groups sided with the allies, and therefore felt that they had fought through both occupations by foreigners and betrayal by compatriots. As independence neared, therefore, they felt the colonial administrators would recognize their loyalty by granting them independence on their own terms, rather than simply lumping them all together in a state to be controlled by the Bamar.124 When independence was declared, a number of rebel groups immediately began to take up arms against the government. Amongst the earliest were the Karen and the Kachin, soon followed by a communist insurgency, which eventually evolved into the United Wa State Army (Thant Myint-U 2006, 261-262).   A brief experimentation with democracy came to an end in 1962 when the democratically elected regime of Win Maung was overthrown by the military, led notably by Ne Win. This was followed by decades of isolation, in which regime known by the Orwellian name the State Law and Order Restoration Council (or SLORC) sought to rid the country of foreign influence, while operating a martial monopoly over the economy.125 The result was an economic crash, alongside the development of a system of extreme corruption and authoritarianism. Dissent of any kind was not allowed, as the regime followed the “Burmese Path to Socialism”, a rather confusing political doctrine which advocated for cultural traditionalism and isolation with the occasional token reference to socialist thought.126  Ne Win’s rule came to an end in 1981, when control of the regime passed to several intermediaries before power was seized by general Than Shwe in 1992. All showed a passion for authoritarian rule and decision-making using traditional astrology, a twin strategy which further assured both international isolation as well as economic backwardness. This isolationist policy increased the desire of opposition groups to make international connections, even though the regime was unwilling to. Whereas the United Nations was led by a Burman – U Thant – from 1961 to 1971, the regime’s official policy was to stay out of international affairs, even going to the extent of barring the Burmese Secretary General from entering the country.  When U Thant’s family tried to bury the UN General Secretary upon his death in 1974, the regime sent soldiers to disrupt the funeral and steal the body, resulting in the deaths of dozens of mourners.127 Throughout this time, the regime relied on American allies for international support, in a trade-off which saw Burma shun China in exchange for American silence on human rights abuses committed within Burmese borders.128                                                           123 A more complete account of the British occupation is provided in Thant Myint-U (2001, 219-244). 124 Discussed in detail by Callahan (2003, 45-86). 125 This period of isolationism in Burmese policies is discussed by Egreteau and Jagan (2013, 71-139) as well as Taylor (2014, 255-308). 126 As discussed by Pye (1985, 5-6). 127 The complete story of the intervention at the funeral is discussed by U Thant’s grandson, Thant Myint-U (2006, 311-315). 128 As discussed by Thant Myint-U (2006, 301-302) and by Steinberg (2010, 114-119). 66   A final key date to note here is 1988, a year which saw the regime announce that increased political and economic liberalization was to come to the country. Popular movements of students and monks took to the street to celebrate the changes, and to protest policies of the regime. The military responded by quickly opening fire on protestors in cities across the country, killing as many as 3,000 people in the process (Ferrara 2003, 305). In response, many opposition groups who had sought to embrace peace, returned to violent confrontation of the regime. This had the effect of pushing a number of opposition groups to increase their campaigns of violence, and for others to break ceasefires that had been established with the regime’s forces. Further, entirely new multi-ethnic armed groups were created (including the All Burma Students Democratic Front), which began new campaigns of violence against the regime. 129 Although the nominal rulers of Myanmar have changed several times since independence, the regime’s authoritarian strategies have remained fairly consistent. This includes the policy of isolation, as well as the means by which the regime has sought to confront and challenge the armed groups who hold much greater popular support in some areas than the regime itself.130 Control of the country is in the hands of the Tatmadaw, and its key allies, with little dissent or disagreement tolerated within the military or the general population (Callahan 2007). Members of civil society and opposition parties have been the target of horrific human rights abuses throughout the time since the junta seized control. Throughout the almost 70 years of conflict in Myanmar, the strategy of the Tatmadaw has been quite consistent. Its official counterinsurgency strategy is the “Four Cuts” approach, which aims to remove the support for armed groups by cutting the opposition off supplies of food, funding, recruits and intelligence. In practice, this is carried out in brutal fashion, targeting not only the armed groups themselves, but also their support in the general population. It has involved both scorched earth tactics as well as mass executions.131 Essentially, this is a divide-and-rule strategy, in which the regime seeks to encourage factionalization and splintering of the opposition groups (Ardeth Maung Thawnghmung 2013, 107-109). The government encourages ANSGs to defect toward its side; in return they are permitted to keep their arms in exchange for becoming a “Border Guard Force”, and a licence to attack and raid other groups (Chengyang 2011, 110-111). The regime has done this in order that it can retain control as a kind of armed hegemon over a wide range of ANSGs of varying strength and independence from the state itself. As of 2016, the conflict is continuing, although it is declining. Trends indicate a liberalization of the Burmese economy and political system, as well as an increasing number of ceasefires being signed by armed groups. Presently, there are as many as 365,000 internally displaced persons in Myanmar, as                                                           129 More details on the size and dynamics of the Burmese insurgencies at the time of the 1988 uprising, and the effects it had on them, are detailed in Smith (1999). 130 A detailed discussion of the regime’s strategies and policies of isolation are provided by, Rogers (2010) and Fink (2009). 131 The “Four Cuts” Strategy is the official strategy of the Tatmadaw and continues to this day (B. Rogers 2010, 51-52). It officially began as early as the late 1950s and early 1960s and has been carried out in consistent practice since. Although the strategy has been regularly denied by both the regime and members of the Tatmadaw, its existence as a strategy is well documented, as noted by Smith (1999, 258-259). 67  well as 1.4 million stateless persons who reside in Myanmar but are rejected by the state. A further 450,000 people have fled the country, mostly to the western regions of Thailand.132 The conflict has largely targeted civilians. Between 1996 and 2011, more than 3,700 villages in Eastern Myanmar were destroyed by the Tatmadaw, displacing many.133   The conflict and politics of Myanmar exhibit two key trends in contemporary intra-state conflict. The first is that the splintering and fragmentation of armed groups has become the norm in conflict zones. In Myanmar the conflict has become an alphabet soup of armed groups claiming to represent a range of ethnic, religious, economic and even linguistic grievances. The role of mid-level power brokers in encouraging fragmentation of armed groups, leading to a multiplicity of rebellions in a single context, has been theorized recently by scholars including Christia (2012) and Fjelde & Nilsson (2012), amongst others. Although they provide succinct analysis of the breakdown of ANSGs, as well as temporary alliance-making in the context of an ongoing conflict, they miss a new trend with ANSGs in Myanmar; to create new coalitions to seek greater international attention. In Syria, alliances are mere temporary constellations which are usually subsumed into a larger armed group (such as the Free Syrian Army). In Myanmar, however, the alliances between non-state groups have been much more formal, with alliances being given formal charters and even having chairpersons elected from representatives of the various armed groups. Some in the past, such as the National Democratic Front, represented attempts to create a single front against the Burmese state. However, it was plagued by lack of coordination and varying levels of commitment by the constituent armed groups (Topich and Leitich 2013, 89).  A more recent attempt is the United Nationalities Federal council (UNFC), an alliance of opposition groups from all Burmese regions currently engaged in armed struggle against the regime. The purported aim of the UNFC is to provide a united front for both negotiations and armed conflict, asserting that there is no peace with the regime until all ethnicities have negotiated peace (United Nationalities Federal Council 2011). As such, they present a challenge to the regime’s strategy of divide-and-rule, while also representing a single body more able to engage with international actors, on behalf of armed groups like the Chin National Front, who likely do not have more than 200 fighters.134 Since its foundation, the UNFC has entered into discussions with the regime as well as a number of representatives of foreign governments and organizations. As such, it serves as a vehicle for relatively weak groups to assert themselves both domestically and internationally. The UNFC will be discussed further with specific references to the case studies who are members, however in order to demonstrate the multi-ethnic nature of the UNFC, and its representation of many ethnicities who are rebelling against the regime, it is worth providing a list of the armed groups who are members135: 1. Arakan National Council (ANC) 2. Chin National Front (CNF)                                                           132 The numbers on displacement are provided by the United Nations High Commissioner for Refugees (2015). 133 As reported by the Thai Burma Border Consortium (2012, 2).  134 The estimate for the size of the China National Army is provided by Myanmar Peace Monitor (2015). 135 Information on membership is provided by Myanmar Peace Monitor (2016). 68  3. Karen National Liberation Army 4. Karenni Army  5. Kachin Independence Army 6. Lahu Democratic Union (LDU) 7. Mon National Liberation Army 8. Pa-Oh National Liberation Army 9. Palaung State Liberation Front (PSLF) 10. Shan State Army - North  11. Wa National Army (WNO)  Although the frontlines in Syria are constantly shifting, the territories of armed groups in Myanmar has been relatively consistent over time. What has changed, however, is the regime’s ability to exert control over these disparate regions, and the emergence of regime proxy forces. In this context, armed groups who are relatively small are able to conduct insurgencies lasting decades and are able to challenge the legitimacy of the regime both at home and abroad.  Establishing levels of support for the various Burmese opposition groups is a challenging enterprise. International attention to the Syrian Conflict has led to a number of opinion polls being conducted in various parts of the country, as well as among displaced peoples. As such, support levels are easier to create. With Myanmar, however, no such polls exist. This is especially true for opposition areas which are far from the border areas with Thailand where most such polls can be conducted safely. As such, levels of support for each armed group will be inferred based on the predatory tactics of the armed group itself. In the case that an armed group operates warlordist structures over civilian populations, extracting rent and conscripting locals, the group will be seen to lack support.136 The United Wa State Army is an example of an opposition group which utilizes these strategies. However, in cases where armed groups operate a more inclusive form of governance, and do not extract rent, conscript, or where they use semi-democratic leadership selection mechanisms, we can say they have more support. Below is a diagram which shows the rough operational area of each of the armed groups under consideration, as well as the dominant armed groups in the region at the time of research. Analysis will then provide a background on each individual armed group in chronological order as they were formed, as well as establishing the armed group as winning or losing and with or without support.                                                            136 The understanding of warlordism used here is drawn from Reno’s depiction of warlords as those that capture and control regional markets, without interference from the state, for the purposes of personal or organizational gain (1998, 15). At the same time, these warlord structures are state-like, in the sense that they corrode the legitimacy and effectiveness of the state, while also leading to the fragmentation of all other political and economic arrangements (Marten 2006, 48). 69   Figure 2 – Area of Operations of the Armed Groups in the Myanmar Conflict137                                                            137 Created by author based on information in Stratfor (2015) and by Burma News International (2015, 88). 70  3.3.2 Transnational Conflict in Myanmar  Myanmar stands in stark contrast to Syria in terms of the transnational characteristics it displays. Whereas the Syrian conflict is one which has spread across multiple countries (and to some extent, multiple continents), the Myanmar conflict has remained relatively contained within the borders of the country.  Firstly, Myanmar lacks the direct military involvement of foreign states seen in Syria. Although there have been some concerns about Thai interventions into the conflict in the past, this did not occur. Secondly, arms have been supplied to some armed groups, but not in the large quantities. Weapons provisions in Myanmar have been mostly limited to China provided small amounts of arms to Kokang rebels.138 The Karen National Liberation Army have also received financial and material support from India, as have the Kachin Independence Army and the All Burma Students Democratic Front (Routray 2011, 130). Although it is difficult to estimate the amount of such support with any certainty, there can be little doubt that the support provided is far smaller than that provided to other conflicts such as Syria.  Secondly, there has been intense involvement of international organizations and non-governmental organizations in Myanmar, and this has continued to grow dramatically in recent years. As the Myanmar regime begins to end its policy of isolation in the 2010s, humanitarian and development organizations are clambering over each other to gain access to the minority regions, with some referring to it as a “gold rush” of funding opportunities for NGOs and the UN (Schroeder and Alan Saw U 2014, 201). In this context, the stakes of legitimacy increase even further for armed groups. The material gains of having development aid are clear enough. However, being deemed legitimate by international organizations not only increases their profile internationally, but cements it domestically, thus further allowing them to assert their organizational aims even though they are facing strategic defeat. The political and economic are the twin strands that make up the double-helix of an armed group’s metaphorical DNA, as are the normative and material resources that it seeks. The below diagram (Figure 3) displays the conflict, mapping out the key actors that are involved with this study, as well as the relationships of actors supporting or combatting the ANSGs. It emphasizes the relationships as they existed at the time that fieldwork was conducted in eastern Myanmar in 2012.                                                           138 Although, this has been denied repeatedly by the Kokang rebels themselves (Radio Free Asia 2015). 71   72  3.3.4 Status of Children and Youths in the Myanmar Conflict  Although the Myanmar Conflict is not raging with the same intensity as that in Syria, the status of children as soldiers is likely much worse in Myanmar. It has been noted since at least 2002 that Myanmar has more child soldiers than any other country in the world, and that the Tatmadaw has more child soldiers than any other military or armed group in the world.139 Child Soldiers International has noted that as of 2015, the Tatmadaw continues to forcibly recruit children, which the Tatmadaw itself strongly denies (Child Soldiers International 2015, 11). The number 70,000 is frequently used in estimations of the number of child soldiers in Myanmar, however it is important to note that this number has been used since a Human Rights Watch report in 2002, although it is still used in estimates as of 2016.140 Whereas the number most certainly is not accurate, it is still indicative of the extent of the problem.   3.4 Karen National Liberation Army 3.4.1 KNLA – Losing with Support  With 4-5 million people, the Karen are the second largest minority in the country and are based largely in the east of the country (Gravers 2014a, 155). While the Myanmar Conflict is the longest running the world, it is specifically the KNLA that are the oldest continually operating armed group (Steinberg 2010, 12). Formed in 1947, the Karen National Union (KNU) and its armed wing the Karen National Liberation Army (KNLA) were one of the first armed groups to commence insurrection against the independent Burmese state. As with many ethnic labels, the term Karen does not denote a single ethnicity, but instead, refers to a broad ethnolinguistic group of more than 20 subgroups. After the Shan, the Karen are the largest minority in Myanmar.  In addition to a wide cultural diversity, the Karen are also religiously diverse, with 15-20% Christian and 5-10% animist, with the remainder being Buddhists (Ardeth Maung Thawnghmung 2008, 3-4). The brutal ongoing conflict has left as many as 90,000 Karen displaced within Kayin State, and at least 130,000 have crossed the border to Thailand (Schroeder and Alan Saw U 2014, 199). During the colonial era, British officials considered the Karen to be a “martial race”, and therefore, gave them preferential treatment with regard to positions in the military (Ardeth Maung Thawnghmung 2013, 31). During the Second World War, the Karen sided with the Allies, operating a guerrilla insurgency against the Japanese occupiers. The KNLA leaders cite this as a key reason why the Karen should be granted an independent state (Karen National Union 1985, 9). The KNLA’s purported reasons for fighting are best summarized in a statement from their own manifesto; “The bitter experiences of the Karen throughout our history in Burma… taught us one lesson. They taught us that as                                                           139 This was first reported by Human Rights Watch (2002). 140 The initial report in which this number was used was produced in the report “My Gun Was as Tall as Me.” By Human Rights Watch (2002a). The number has been cited as recently as 2016 by Chen (2014, 4). 73  a nation, unless we control a state of our own, we will never experience a life of peace and decency, free from persecution and oppression.” (Karen National Union 2006, 8).  Early in its campaign of violence, the Karen managed to seize a large portion of the Karen speaking regions, however this quickly began to be eroded by the regime’s response. From the 1940s until the 1990s, the Karen saw themselves involved in a slow retreat, with their territory increasingly pinched between the advancing regime armies and the Thai border. This, however, is their strong point, since the Thai government has granted them tacit control over border areas in and around Mae Sot, meaning they will never be outright defeated (Ardeth Maung Thawnghmung 2013, 57-58). A major setback occurred for the KNLA in 1995, when a large portion of the armed group’s forces defected in order to create a government-allied proxy known as the Democratic Karen Buddhist Army (DBKA). Many of those who left were Buddhists, angry at what they felt was a Christian dominance of the KNLA. In strategic terms, the defection was exceptionally costly for the KNLA, as the DBKA not only broke away with a large amount of recruits and resources, but also revealed to the government vital intelligence which enabled them to seize much of the remaining KNLA territory (Topich and Leitich 2013, 113). In addition to the DBKA, the KNLA have fought with other armed groups for control of cross-border trade, including their nominal allies the Mon National Liberation Army (R. H. Taylor 2009, 389). Although the split and the border wars created new enemies for the KNLA, they used these threats to solidify their control over the Christian Karen, since the leaders were no longer trying to keep the Buddhists in the fold. With the departure of the Buddhist support, the Christian leaders were able to cement support within the Christian population. The KNLA have earned financial resources through controlling lucrative cross-border timber trade to Thailand (Topich and Leitich 2013, 116). The KNU have also received financial and material support from India, as have the Kachin Independence Army and the All Burma Students Democratic Front (Routray 2011, 130). Foreign support, however, has largely dried up, as have the KNLA’s access to lucrative cross border trade. Thailand has abandoned its border policy of using Karen and Shan as a buffer, leaving the KNLA without a foreign sponsor (Legêne and Ytzen 2014, 115). This occurred at the same time that Myanmar experienced an overall thaw in its cold relations with neighbours and great powers alike (Clymer 2015, 298-306). In this context, the gains from supporting armed opposition to the Burmese regime had declined.  Due to its setbacks in the 1990s, the KNLA has been experiencing a dramatic decline in strength, wealth and capability. Some have pointed to the KNU being at a “nadir” of its 70 year struggle (Tin Maung Maung Than 2011, 33). It has had to cling to border areas with Thailand and must base much of its operations from the refugee camps on the Thai side of the border due to increasing pressure from the Tatmadaw (R. H. Taylor 2009, 335). It holds a fraction of the territory it used to and has little or no effect on government policy (Topich and Leitich 2013, 113). It is in a tenuous position, as it becomes further isolated from political processes in Myanmar, while also becoming distant from the population which it represents. Although a ceasefire was signed with the regime in 2012, there have been frequent violations, and neither the fighting nor the abuses against civilians slowed or abated. More than 30,000 were displaced in 2013 alone (B. Rogers 2015, 57). Tenuous divisions within the KNLA’s ranks threaten 74  to split it still further (Gravers 2014b, 190). It did, however, join the UNFC and is a leading member of the ethnic alliance (Tripathi 2014, 141). Overall, therefore, the KNLA is certainly on the decline, and is fighting a losing battle it will not be able to continue forever. Since its foundation, the KNLA has had a large amount of support not only from the Karen-speaking peoples (Ardeth Maung Thawnghmung 2008, 9). However, it also has wide support from other minorities within Myanmar. It has long been allied with pro-democracy groups from a number of different ethnicities, and therefore has been seen as legitimate by many.141 Its support from the population has grown in that it asserts a strict ‘no drugs’ policy, seeking to keep its people involved in more legitimate trade (Ardeth Maung Thawnghmung 2008, 27). The structure of the KNU/KNLA is – compared to most ANSGs – a relatively democratic and inclusive. The leadership of its political wing are democratically elected to a four-year congress, with cabinet members being elected from different areas (Karen National Union 2016). Although it is certainly not a perfect democracy, it is also a far cry from other more abusive and authoritarian armed groups in the country. This is as much by necessity as ideology; since the Karen are an ethnically and religiously diverse group, the KNLA has had to adopt inclusive policies in order to avoid the kind of fragmentations that occurred with splinter groups like the DKBA. As such, the KNLA will be described as losing with support  3.4.2 Child Soldiers and the Karen National Liberation Army It is clear that that the KNLA has engaged in the recruitment and deployment of child soldiers in the past. Although the extent of the use of child soldiers is not clear, it is likely that they were widespread within the ranks of the KNLA well into the early 2000s. In an infamous incident, a 13-year-old child soldier bragged to a reporter that he willingly joined in the hope that he could shoot some Burmese soldiers.142  This is not something that the KNLA has denied; far from it. They have openly admitted that in the past child soldiers were recruited (Karen National Union 2009a). In spite of this, the KNLA began to take widespread action in order to comply with the child soldiers norm. Previously, the KNLA has sought to conclude action plans on the issue of child soldiers with the UN, however it was blocked by the regime, since the UN is unable to violate sovereignty (South 2014, 261). As the child soldiers’ norm became internationally salient in the early 2000s, the KNLA was one of the first Burmese ANSGs to respond. They openly called for international observers to come and verify they were not abusing children, while also asking for assistance with verification.143 These calls have been mostly ignored. It has become such a habit for the KNLA to request international verifiers in Myanmar on the child soldiers’ norm, that they now issue a press release on 15 September each year – the International Day of Peace.144                                                           141 As noted by both Ardeth Maung Thawnghmung (2008, 38) and Duell (2011, 45). 142 As cited in Gravers (2007, 7). 143 These requests were made in KNU statements including in 2007 (Karen National Union 2007) and 2009 (Karen National Union 2009b). 144 Examples include a KNU press release (2010) 75  The current policy with regards to child soldiers is clear enough; “The policy of the [KNLA] is that we do not recruit child soldiers… We co-operate with the United Nations and NGOs on this matter” (Karen National Union n.d.). It is important to note, however, that the KNLA has not eradicated child soldiers from its ranks and attained complete compliance with the norm. Whereas some ANSGs deny use of child soldiers outright, the KNLA acknowledges that it did so in the past and that is trying to stop.145 In an interview with the second-in-command of the KNLA’s political wing, it was noted that the decision was made to embrace the issue of child soldiers in order to get the attention of international actors of their struggle against the Tatmadaw.146 When I asked the leader of the KNLA if she was concerned about the ICC investigating the KNLA for use of child soldiers in the past, she responded simply, “That is not a problem for us.” She went on to explain that the ICC was never a concern, since no investigation would ever be opened in Myanmar, and the KNLA had already taken steps to remove child soldiers from their ranks.147 The decision was not taken out of fear. Rather, the KNLA engaged with the norm of their own volition in order to bring attention to their plight. This represents a key step in the process of demonstrating norm compliance: an admission of violation, and the expressed willingness to engage in steps to comply.  A 2013 report by Child Soldiers International notes that the KNLA has endeavoured to prevent child soldiers’ from joining its ranks, but has still engaged in limited violations as prevention mechanisms are not universally applied (2013, 28).Whereas this may seem to be condemnatory of the KNLA, it is further important to note that the UN and other international observers are blocked not by the armed groups, but by the regime.148 In order for verification mechanisms to be more effective, therefore, the regime would need to allow neutral observers to comment on KNLA activities. Either way, the signing of the Deed of Commitment in 2013 by the KNLA demonstrates a willingness to engage with the norm, and also a strong desire to demonstrate compliance with international observers. It is important to note that it is highly unlikely that any armed group will attain perfect compliance. As long as there are large numbers of children and youths trying to join armed groups, some will get in through lying, or be let in by junior officer’s whose conception of the norm may differ from the leaders. As noted in the previous chapter, every year the United Nations Officer of the Secretary General releases to the General Assembly a report on the status of children and armed conflict. The KNLA was listed as a violator of the child soldiers’ norm each year from 2012 to 2015.149  However, the same document also notes that the reports are unverifiable. Although the KNLA has been listed on the UN Watchlist, it is important to note that there is no real mechanism for them to be removed from this list, given that Myanmar blocks any UN verifiers from entering KNLA territory. Leaders of the ANSG expressed exasperation at this in interviews, since although they were fighting to demonstrate compliance with the norm, there were no international observers coming to establish the veracity of                                                           145 This was noted in interviews, including; Colonel New Dah Mya, Karen National Liberation Army. Interview with author 19-06-2012. It has also been noted in KNU press releases (Karen National Union 2007). 146 David Thackabaw (Vice President of KNU). Interview with author, 28-06-2016. Mae Sot, Thailand. 147 Naw Zipporah Sein (General Secretary of KNU). Interview with author, 28-06-2016. Mae Sot, Thailand. 148 This was reported to the author in interviews, and also is also noted in Lata (2012). 149 As reported in United Nations General Assembly (2012, 14), United Nations General Assembly (2014), and United Nations General Assembly (2015, 23) 76  their claims. It seems the only reason the KNLA are included in the UN lists is because they were included previously, regardless of whether they violate the norm or not. In interviews with the KNLA, senior commanders (including the Vice President of the KNU David Thackabaw) celebrated signing the Deed of Commitment that was being discussed with Geneva Call.150 The KNLA signed a Deed of Commitment with Geneva Call in 2013, committing themselves to the protection of children and the adoption of mechanisms to protect the recruitment of child soldiers (Geneva Call 2013). The KNLA has continued to work with Geneva Call, including participating in child protection workshops for high-level commanders (Geneva Call 2015a) as well as more junior officers (Geneva Call 2015b). Since Geneva Call uses the broad definition of the norm, so do the KNLA. During discussions with KNLA officers, they were quick to point out that though the KNLA had been guilty of violations in the past, its enemy the Democratic Buddhist Karen Army (DBKA) was a worse violator, and it had not sought to engage with any international organizations or actors with respect to the child soldiers’ norm.151 KNLA officers also repeatedly stressed that the Tatmadaw was also a gross violator of the norm, and that the KNLA was therefore more legitimate.152 One noted the struggle against the Tatmadaw “terrorists” was long and hard, but the KNLA would continue and “the struggle will go on as long as the desire for freedom keeps burning.”153 The KNLA were  It was also noted by leaders that the pursuit of positive relationships with international actors was of vital importance to the KNLA, even if it resulted in an increased hostility with the government.154 This comment, therefore, represents a willingness to risk increasing hostilities in order to pursue legitimacy, even though – materially speaking – the organisation would be better off it pursued a peaceful relationship with the government in the manner that other armed groups had done.155  The practical difficulties of preventing child soldiers from joining was also discussed at length by members of the KNLA. One high-ranking officer noted that although they endeavoured to keep child soldiers out of the organization, it was impossible so long as people had no documents, and there were huge numbers of refugees.156 Although some may see this as an excuse for norm violation, it is also an acknowledgment that norm violation occurs, and that constant efforts must be made to demonstrate compliance. As noted earlier, the first step in switching from norm violation to norm compliance is acknowledgement that violations occurred. In order to measure norm acceptance, we cannot simply set our target at ‘Zero violations’, since this is an unrealistic goal for ANSGs like those in Myanmar. Even                                                           150 David Thackabaw, Vice President of the KNU.  Interview with author, 28-06-2012, Mae Sot, Thailand. 151 This accusation was made in interviews with Colonel New Dah Mya, Karen National Liberation Army. Interview with author 19-06-2012. The widespread use of child soldiers by the DBKA has been noted by both Child Soldiers International (2013, 25) and by South (2014, 261). 152 This accusation was made in interviews with several KNLA fighters, as well as David Thackabaw, Vice President of the KNU.  Interview with author, 28-06-2012, Mae Sot, Thailand.  153 Colonel New Dah Mya, Karen National Liberation Army. Interview with author 19-06-2012 154 David Thackabaw, Vice President of the KNU.  Interview with author, 28-06-2012, Mae Sot, Thailand. 155 For example, as noted earlier, the Democratic Karen Buddhist Army had defected from the opposition to the government and was therefore able to seek lucrative gains from participation in cross border trade, as well as mining and logging. 156 Colonel New Dah Mya, Karen National Liberation Army. Interview with author 19-06-2012. 77  though we see some limited violations, it is clear that the KNLA has been engaging international actors in search of legitimacy and attempted to bring about institutional changes within the organisation in order to bring about compliance. The primary audience to which it directs its legitimacy-seeking behaviour, therefore, are the international community as embodied by the transnational advocacy networks it engages with. For any ANSG, once the leadership has decided to engage with a norm, it is not guaranteed that the followers will do the same. Why, then, was the adoption of the norm a success for the KNLA, and how does this relate to the group’s identity-narrative? Since it was the largest opposition group in Myanmar for much of its existence, the KNLA wanted to pitch itself as a legitimate representative of all people in Myanmar suffering under the regime. As such, it began to project an image of being the protector of all abused Burmese peoples. As one leader noted, “We welcome people from all areas to our movement, including children who will be protected”.157 In asserting itself as the protector of children (especially child soldiers who flee the regime), it is therefore adapting its identity-narrative to its strategic needs. Beliefs about being protectors are being matched to strategies of norm acceptance. As noted above, this decision overall was done to seek international attention. But what if receiving international support resulted in a deterioration of the relationship with the government still further? Might that preclude peace or a ceasefire? The second-in-command of the KNLA responded that this was not a concern of his since they were the legitimate representative of the Karen people and they wanted the world to recognize this. He put it simply; “This is not important. If a ceasefire breaks down, it breaks down.” 158 It is interesting to note that leaders prefer a favourable international opinion of the KNLA over a ceasefire with the government. In order to be successful, the change in policy (from recruitment to demobilization) must match the pre-existing identity-narrative of the organization. This leads to the counterintuitive result that people hold a belief that condemns their previous actions. When asked why the child soldiers norm should be followed now when it was not followed in the past, one field-level KNLA responded with confusion. He noted that it was important for the soldiers to do the right thing, since that is what they believed in, and that is what they were fighting for. He compared the KNLA to other groups, noting, “We do what we believe is right… [The others] are thinking only of survival, and they will do what they think they have to do to survive.”159 When pressed to explain why child soldiers were used in the past, but were not now, he responded that the soldiers did not always know how to conduct war properly. He noted that although they were always trying to do the right thing, at times they did not know how. He cited a workshop with Geneva Call on child protection, noting “It was very helpful, makes commanders understand more about humanity and how to treat enemies as well as women and children.”160 The change in policy, therefore, is one in which the fighters are encouraged to see themselves as the protectors of people they abused in the past.                                                           157 David Thackabaw (Vice President of KNU). Interview with author, 28-06-2012. Mae Sot, Thailand. 158 David Thackabaw (Vice President of KNU). Interview with author, 28-06-2012. Mae Sot, Thailand. 159 New Dah Mah (Colonel KNLA). Interview with author, 29-06-2012. Mae Sot, Thailand. 160 New Dah Mah (Colonel KNLA). Interview with author, 29-06-2012. Mae Sot, Thailand. 78  Following the meetings with Geneva Call, the KNLA leadership wrote a new Code of Conduct regarding children, which was passed down to lower ranks, who were encouraged to read it and reflect, and ask questions about things they didn’t understand. These questions were then filtered to continuing sessions with Geneva Call representatives.161 The lack of continued violations demonstrates the success of this approach. As the leader of the KNLA noted, embracing the norm of child soldiers was part of a wider process of reconciliation between the KNLA and civilians in general. She noted this as being important, so that all people could see that they were the legitimate leaders of the Karen and the Burmese opposition.162 Overall, the leaders made a strategic decision to follow the norm and engaged with Geneva Call in order to demonstrate publicly that they were doing so. They did this of their own volition, since there were no actors in the region compelling them to do so, and there was no concern of an outside intervention of ICC investigation. In order to convince the people within the organization that the norm ought to be followed, the norm was attached to pre-existing beliefs and identities as being the leaders of the Burmese opposition, and the protectors of all people. The result was that even people who violated the norm in the past came to engage with it and began to see the protection of children as a core belief and a vital part of their identity.   3.5 Karenni Army 3.5.1 KA – Losing with Support Although the smallest state in Myanmar, Kayah State (formerly known as Karenni state) has been home to some of the most intense violence and displacement. The Karenni people – also known as the Kayah or the Red Karen – number approximately 150,000 (Gravers 2014a, 155).  As such, they represent the smallest ethnic grouping under study here. Prior to the Karenni areas being brought into the British colonial of Burma, the area had a long history of independence (Kubo 2014). In the Panglong Agreement of 1947, signed by independence leader Aung San and ethnic minority leaders, the Karenni were promised full autonomy in an independent state (B. Rogers 2015, 3).  The Karenni became involved in the same anti-regime protests that swept the minority areas of Myanmar in the 1940s and 1950s, however it was not until 1957 that activists founded the Karenni National Progressive Party (KNPP) and its armed wing the Karenni Army (KA). In addition to speaking multiple languages, the KA also includes a mix of Buddhists and Christians. Further, the group welcomes large numbers of fighters from other                                                           161 As reported by David Thackabaw and New Dah Mah. 162 Naw Zipporah Sein (General Secretary of KNU). Interview with author, 20-06-2012. 79  ethnicities into their own ranks.163 Therefore, although it is nominally an ethnic defence group, in reality – like most of the other armed groups – it represents a multiethnic population. Just as with the groups noted earlier, the Karenni were victim to an immense campaign of displacement and human rights abuses. Although numbers are mostly unreliable, and it is difficult to estimate the number of displaced within Myanmar, as of 2007, more than 90,000 were displaced were thought to be in hiding from the Tatmadaw, while a further 22,000 had managed to cross the border to Thailand (Dudley 2007). The status of the Karenni is so dire that Kayah State has been referred to as “one vast concentration camp” (B. Rogers 2015, 58). The KA signed a ceasefire with the regime in 1995, however the Tatmadaw violated it within three months.164 It likely unravelled over a disagreement regarding logging concessions (Callahan 2007, 37). It later became clear that the regime was using the ceasefire as a means of amassing troops in the region, in hopes that they could deliver a decisive blow to the KA forces and route them while under nominal ceasefire (B. Rogers 2015, 60-61). Another ceasefire was signed in 2012, however fighting has continued, with negotiations for another ceasefire continuing into 2015 (Ye Mon and Dinmore 2015). As with other armed groups noted above, the KA has faced a constant problem of splinter groups and defectors. In 1978, a group of left-leaning members of the KA split off to from the KA to form the Karenni National People’s Liberation Front (KNPLF) (Gravers 2014a, 170). The KA are seen to hold legitimacy among the displaced, as it operates a policy of locating civilians being targeted by the regime and escorting them to relative safety in camps along the border with Thailand, and in some cases in Thailand itself.165 Further, as the Karenni Army is the only armed group seeking to defend the rights of the Karenni peoples, it holds large support from the minority population in the shrinking areas it controls. In the face of constant pressure from the immensely larger Tatmadaw, the KA have been shrinking, almost having in size during the 2010s from over a thousand to as few as 600 fighting soldiers. In order to make up for this declining strength, the KA has joined the UNFC. Overall, the KA are losing with support.  3.5.2 Child Soldiers and the Karenni Army The Karenni Army engaged with the child soldiers norm extensively, though did not do so until long after it had begun to lose the conflict and slowly ceded territory to its opponents. Just as with the KNLA, the KA is listed in the Secretary General’s Annual Reports from 2012 to 2015.166 A 2013 report by Child Soldiers International notes that the KA has endeavoured to prevent child soldiers’ from joining its                                                           163 The Diversity of the Karenni is discussed by Gravers (2014a, 155). The multi-ethnic makeup nature of Karenni fighters is discussed by former KA fighter Pascal Khoo Thwe in his biography, himself a member of the non-Karen Padaung tribe (2002). 164 Although not completely clear who first violated the ceasefire, Zaw OO and Win Min (2007, 12)., as well as Gravers (2014a, 170) note that it was the regime forces. 165 This has been discussed by Rogers (2015, 60) and Pascal Khoo Thwe (2002, 198-246) 166 As reported in United Nations General Assembly (2012, 14), United Nations General Assembly (2014), and United Nations General Assembly (2015, 23). 80  ranks but has still engaged in limited violations (Child Soldiers International 2013, 28).  Whereas the UN representatives were unable to go into the country, Human Rights Watch does not require the approval of a state in order to conduct verification on its activities. In 2002, Human Rights Watch documented cases whereby the Karenni Army was known to have recruited and deployed child soldiers (Human Rights Watch 2002a, 136-137). By 2007, however, a following report cleared the KA of continuing violations, noting that they had taken adequate steps to remove child soldiers from their ranks and to try institutionalize prevention mechanisms (Human Rights Watch 2007, 10-11). Representatives of the KA criticized the UN and their continued accusations of violations, given that no UN staff had come to meet with the KA, and citing the Human Rights Watch report (Karenni National Progressive Party 2016). Even though this specific example is from as far back as 2007, the same back-and-forth between the KNLA and UN has continued, with no possible means for the KNLA to demonstrate and prove compliance to the UN. During interviews, the leaders of the KA expressed a deep desire to engage with all international actors and organizations. As with the KNLA, many were exasperated that the international community would accuse the KA of using child soldiers, but then refuse to send verifiers to observe the KA’s efforts to demobilize any children within their ranks.167 Leaders and commanders did, however, acknowledge openly that there were child soldiers in the ranks of the KA in the past, but that the organization was now doing its best to ensure that not more found their way in.168 As with the KNLA, this demonstrates the Karenni Army’s difficulty in implementing norms at all levels. In the past, the KA has sought to conclude action plans on the issue of child soldiers with the UN, however it was blocked by the regime, since the UN is unable to violate sovereignty (South 2014, 261). The KA/KNPP have long tried to get attention from the outside world, and the leadership has frequently expressed exasperation at its inability to obtain international attention.169 At the same time, just as with the KNLA’s struggle, the KA continues to make repeated calls for UN verifiers, and to attempt to demonstrate compliance with any international observers who will listen. The Karenni Army also began to release annual requests for the UN to reconsider its stance.170 In 2012, the Karenni Army signed a Deed of Commitment with Geneva Call on the issue of child soldiers (Geneva Call 2012c). In interviews, leaders stressed this was an important moment in the history of the KA, since it showed a willingness to work with international actors.171 Along with the MNLA, the KA were the first armed groups to sign Geneva Call’s Deed of Commitment with regards to children and armed conflict since the organisation commenced activity on the issue (Geneva Call 2012). They therefore exhibit the same                                                           167 Aung Sun Myint, Secretary Two, Karen National Progressive Party.  Interview with author, 06-07-2012, Mae Hong Son, Thailand. 168 Khu Ooh Reh, Secretary One, Karen National Progressive Party. Interview with author, 06-06-2012. 169 This was noted in interviews Khu Ooh Reh.  Interview with author, 06-07-2012, Kayah State, Burma. It was also noted by Dudley (2010, 96-97). 170 This occurred in 2008 (Karenni National Progressive Party 2008), and in 2009 (Karenni National Progressive Party 2009). 171 Aung Sun Myint, Secretary Two, Karen National Progressive Party.  Interview with author, 06-07-2012, Mae Hong Son, Thailand. 81  interactions with the norm as other groups that are losing with support. As with the KNLA, the their primary audience is the international community as embodied by transnational advocacy networks. When asked why the Karenni Army had taken the decision to sign an accord with Geneva Call, one leader responded that it was “an important part of being a legitimate leader.”172 The General Secretary reported that the decision was taken in order to increase international exposure to the Karenni Army.173 Interestingly, it was reported that the decision was taken in conjunction with the leaders of the Mon National Liberation Army (MNLA, to be discussed below), and both organizations decided to sign the deeds of commitment at the same time. They became the first two organizations to sign Geneva Call’s Deed of Commitment concerning children and armed conflict (Geneva Call 2017). At the time of research, the Karenni Army was developing a new Code of Conduct regarding children in conjunction with Geneva Call, which it was hoping to begin distributing to its fighters.174 Other ANSGs in Myanmar followed, including the KNLA and the Pa-O National Liberation included in this study, as well as the Chin National Front and dozens of other ANSGs around the world (Geneva Call 2017). The Karenni Army – a tiny armed group from the highlands of Myanmar – were therefore at the front of a norm cascade, as discussed by Finnemore & Sikkink (1998). Once some actors start to embrace the norm and it is seen as legitimate, others begin to follow and engage with the norm as well. Just as with the KNLA, once the strategic decision to follow the norm is undertaken by the leaders of the Karenni Army, it was then necessary to ensure that it was followed. A key mechanism for spreading the norm within the movement was training, in which fighters and officers would take part in workshops on the frontline with a representative of Geneva Call. As one leader reported, “They enjoy it very much. They would do it 10 times a year if we let them.”175 The leaders were proud of being the first armed group to sign and cited this as a reason that the lower ranks followed the norm. In this case, the norm itself not only appealed to a broader identity-narrative of being the legitimate representatives of the Karenni people, it also appealed to their sense of being involved in an important struggle and having the attention of the outside world.176 Once the strategic decision was undertaken, the rank-and-file were quick to respond, and increased interaction with Geneva Call. Following this process, there were few to no reported violations of the norm.                                                            172 Aung Sun Myint (Secretary Two, Karenni Army). Interview with author, 06-07-2012. 173 Khu Ooh Reh (General Secretary, Karenni Army). Interview with author, 06-07-2012. 174 Khu Ooh Reh (General Secretary, Karenni Army). Interview with author, 06-07-2012. 175 Khu Ooh Reh (General Secretary, Karenni Army). Interview with author, 06-07-2012. 176 Aung Sun Myint (Secretary Two, Karenni Army). Interview with author, 06-07-2012. 82  3.6 Mon National Liberation Army 3.6.1 MNLA – Losing with Support The Mon people have been described as the oldest civilization in Southeast Asia and were independent until conquered in the 1750s (Topich and Leitich 2013, 5). At present, there are approximately 2-3 million Mon in the country (Gravers 2014a, 155). As a distinct ethno-linguistic group, the Mon have sought independence from the Burmese regime since the country itself found independence from Britain and were therefore one of the first groups to rebel against the new state (South 2003, 86-98). The Mon National Liberation Army (MNLA), and its political wing New Mon State Party (NMSP) were founded in 1958, by a group of Mon rebels dissatisfied with the other Mon armed factions who were willing to negotiate with the regime.177 Although openly rebelling against the regime, the MNLA has managed to avoid intense confrontations, due to its distance from major trade routes or resource-rich parts of the country. Still, the Mon have suffered greatly at the hands of the regime, even if they have been spared the same intensity of conflict that has hit the Karen and Chin. During the 1990s, the regime conscripted as many as 150,000 to complete a railway to the Andaman Sea, while committing abuses against them and depriving them of their livelihoods (Lang 2002). In 1989, the MNLA signed a ceasefire with the government, however sporadic fighting continued due to MNLA splinter groups and factions who were not completely controlled by the leadership. Purportedly, this was because of continued ceasefire violations and human rights abuses committed against Mon civilians by the Tatmadaw (South 2004, 164). Another ceasefire was signed in 1995, however this saw a cessation neither in the fighting nor the regime’s strategies of forced displacement and sexual violence (B. Rogers 2015, 67). During the 1990s, the MNLA experienced a number of large defections, including in 1996 when the Mergui Army broke away from the MNLA and defected to the regime (M. Smith 1999, 448). The MNLA has been able to partner with a number of INGOs working in the Mon ‘liberated zones’, providing both material resources, as well as normative resources both domestically and abroad (2004, 165). Further, it has expanded its influence within Myanmar and internationally through joining the UNFC in 2011 (Tripathi 2014, 143). The Mon insurgents have found financial resources through the lucrative trade in timber across the border to Thailand (Topich and Leitich 2013, 116). However, due to the position of Mon State on the Malay Peninsula, the Mon occupy a position similar to the Karen groups. The regime has increasingly seized control of border crossings, cutting the Mon off from its historical source of income. It has not been able to support itself with a foreign patron, nor has it seized control of a lucrative source of income                                                           177 South discusses this period, and the manner in which Nai Shwe Kyin formed the NMSP due to dissatisfaction with the Mon People’s Front coming to ceasefire with the regime (2004, 161). Tucker provides similar discussion, noting specifically the role that a government amnesty played in encouraging some Mon to defect to the government (2001, 168). 83  such as the narcotics trade. As such, their territory is shrinking, and the group has been pushed back by others.178 The MNLA is generally seen favourably by the Mon people, and an active Mon civil society has developed able to challenge the MNLA and help to alter its policies (South 2007, 168-169). Similar to the KNLA, the MNLA selects its leaders from a vote within a 27 member Central Committee, who advocate for democratic policies even though they boycott the regime’s elections.179 Although not perfectly democratic, the voting structure determines the decisions of the organization as a whole and offers a chance for Mon civilians to voice concerns to leaders. Whereas the MNLA is generally well supported by the population, it is experiencing challenges to its legitimacy by Mon parties who are willing to cooperate with the regime (Hongsar 2011). This is having the effect of further isolating the MNLA, pushing it into such a position that it although it is able to create transnational linkages with NGOs and international organizations, as well as joining the UNFC ethnic alliance, it is facing an existential threat even as the conflict declines in intensity. As such, the MNLA therefore is losing with support.  3.6.2 Child Soldiers and the Mon National Liberation Army The MNLA are similar to the Karenni Army, in that they slowly began to lose the conflict, and that long after this, they began to engage with transnational advocacy networks and the child soldiers norm. The MNLA, like all other armed actors in Myanmar, has certainly engaged in the use of child soldiers in the past, though the extent to which they did so is not completely clear. The UN Reports on Children and Armed Conflict between 2012 and 2015 do not list the MNLA as a violator of the norm. As the MNLA wasn’t involved in the same intensity of conflict as the KA or KNLA, it is likely that the MNLA did not engage in as much widespread recruitment for the simple reason that it did not require as many fighters. Violations certainly occurred, however. In 2002, Human Rights Watch noted that there were confirmed reports of child soldiers operating within the MNLA, however it was unable to confirm numbers, or the extent of the problem (2002, 144-146).  A follow-on report by Human Rights Watch five years later confirmed that the MNLA had instituted recruitment measures to prevent the widespread recruitment of children. It did note, however, that in some cases children may be allowed by soldiers at the ground level to join smaller units to perform basic non-combat roles (Human Rights Watch 2007, 112-113). This represents a similar problem noted earlier; that even if the armed group institutionalizes the norm in order to demonstrate compliance, it may not do so perfectly, and violations may occur at a much lower level. In interviews, commanders with the MNLA agreed that although the organization had struggled with the issue of child soldiers in the past, it was doing a better job now and was attempting to ensure                                                           178 Reports continue into 2017 of the shrinking territory of the Mon, and the continued lost battles. Examples include Nang Mya Nadi  (2017) and Weng (2017). 179 Further details are provided by the Myanmar Peace Monitor (2015). 84  that no new children were recruited.180 One commander described a process whereby recruits would be constantly checked to ensure that they were of the appropriate age, and even though the process was thorough, they would occasionally find recruits under the age of 18 trying to join and would refuse them entry.181 As part of this process, the MNLA had even begun instructing officers on child protection issues, in essence then participating as key local agents of norm diffusion.182  The MNLA has made clear attempts to engage with international actors with regards to the child soldiers’ norm. The MNLA signed a Deed of Commitment regarding children in conflict with Geneva Call in 2012 (Geneva Call 2012d). As with the other armed groups that signed the Deed of Commitment, the MNLA have adopted the broad definition of the norm, as this is the definition asserted by Geneva Call. During an interview, the head of the MNLA expressed deep pride in the agreement, noting it was a welcome change to be able to engage with international organizations.183 They have also continued to join in child protection training for leaders, high-ranking officers, and junior officers (Mon News Agency 2013). The leader of the MNLA – Nai Hong Sa – also took part in a delegation to Japan along with the leader of the PNLA. He reported the same as the PNLA leader; they would never receive assistance from the Japanese, since the Myanmar regime would block it. However, it was important to go to show that the MNLA were real leaders of the opposition.184 The leaders were quick to brag about the Code of Conduct they had written regarding child soldiers and encouraged me to visit their headquarters so I could see the original rather than their tattered copies.185 As noted above, the MNLA and the KA were the first armed groups to sign Geneva Call’s Deed of Commitment. Therefore, as a group that is losing with support, the MNLA also match the predictions at the beginning of this study. As with other groups in this category, its primary audience are the transnational advocacy networks. The role of identity-narrative of the Mon is similar to other armed groups, even though the contents of that identity-narrative differ.  One local fighter explained that it was important to protect children “Because we have been fighting for 60 years, and we were here for thousands of years before that.”186 One of his comrades went on to explain that the Mon civilization had a long history of protecting the weak, and they believed this was the most important reason to fight the government.187 One fighter bragged about how he was the special representative tasked with the responsibility of helping to free child soldiers that defect from the government side.188 His colleagues clearly supported him, noting it was an important job. One commented that “The government has many child                                                           180 Nau Hong Sa Poung Khine, Lt. Col. Nay Nyan Htun, Nai Alam Mon, Mon National Liberation Army.  Interview with author, 19-07-2012, Mon State, Myanmar. 181 Nai Salon Htaw. Mon National Liberation Army.  Interview with author, 20-07-2012, Mon State, Myanmar. 182 Nai Hong Sa, General Secretary, New Mon State Party.  Interview with author, 07-07-2012, Chiang Mai, Thailand. 183 Major (name withheld). Mon National Liberation Army.  Interview with author, 21-07-2012, Mon State, Myanmar. 184 Nai Hong Sa (Mon National Liberation Army). Interview with author, 07-07-2012. Chiang Mai, Thailand. 185 Nai Hong Sa (Mon National Liberation Army). Interview with author, 07-07-2012. Chiang Mai, Thailand. 186 Sergeant Moustache (Name changed). Interview with author, 19-07-2012. Mon State, Myanmar. 187 Fighter X (name withheld). Interview with author. 19-07-2012. Mon State, Myanmar. 188 Major X (name withheld). Interview with author. 21-07-2012. Mon State, Myanmar 85  soldiers, but we believe its wrong (sic). We are different.”189  Another expressed the role of his identity, noting; “We are Mon. We do not run away, and we do not have a child soldier (sic).”190 In all of the interviews conducted, none of the fighters or officers thought that the decision to demobilize all child soldiers was a strange or puzzling one. For them, it was an obvious decision since it was the right thing to do. In one group interview, six fighters thought it strange that I would travel so far to ask such obvious questions, since removing child soldiers was obviously the right thing to do. After asking why I had come, the question was jokingly put to me, “Did you come to learn how to do this, so you can do it in your country?”191 For these fighters and officers, as with the other armed groups discussed here, there was no contradiction between changing their policy with regards to child soldiers, since their beliefs precluded harming children, and their identity as an organization is as defenders of the Mon people. To paraphrase those quoted above; the current Mon identity doesn’t allow for running away, or for keeping child soldiers once the decision to demobilize has been taken, even though they admit having child soldiers in the past.  3.7 United Wa State Army 3.7.1 UWSA – Winning Without Support Nestled in the northeast corner of Shan state, the region known as Wa State192 is the territory of an armed group known as the United Wa State Army (UWSA). In its current incarnation, the UWSA is quite young, however it has roots that go back to the very beginning of conflict in Myanmar. Soon after the country achieved independence, a communist insurgency began in the North of the country. The Communist Party of Burma perpetuated in one form or another for forty years, until it collapsed in 1988. From its remnants appeared three armed groups, the largest of which became the United Wa State Army (Duell 2011, 82). Since its appearance, it has grown dramatically, to become the largest ethnic nationalist force in the country.193 The UWSA quickly allied with the regime and managed to avoid any direct confrontation with the Tatmadaw in exchange for directly waging war on other non-state groups in the region, especially the anti-government Shan State Army – South (SSA-S) and the Shan State Army – North (SSA-N). The UWSA is one of the largest private armies in the world, with more soldiers than the Taliban and a territory the size of Belgium (Thant Myint-U 2011, 107). Although the UWSA is nominally a member of the UNFC ethnic alliance, it does not attend meetings or take part in negotiations or decision-making (Gravers 2014a, 153). Indeed, it has been engaged in violent confrontations with other UNFC members, while it remains at peace with the regime. This tacit ceasefire with the government has allowed the UWSA to pursue the drugs trade, and to improve relationships with China, while also waging an                                                           189 Sergeant Moustache (Name changed). Interview with author, 19-07-2012. Mon State, Myanmar. 190 Nai Salon Htaw (MNLA). Interview with author, 19-07-2012. Mon State, Myanmar. 191 Nau Hong Sa Poung Khine. Interview with author, 19-07-2012. Mon State, Myanmar. 192 Officially, it is actually a sub-state region, not an actual state. 193 As noted by Kramer (2007, 1). 86  aggressive campaign against other armed groups (Thant Myint-U 2011, 107-108). Rather than operating any kind of counter-narcotics, the regime is directly complicit in UWSA’s trade in opium.194  The leaders of the UWSA are largely Chinese speaking and enjoy a close relationship with Chinese officials. In the past, the UWSA made largely unanswered appeals to both Western governments and the United Nations (Thant Myint-U 2011, 212). In failure to get international attention, they instead looked east. It is well known that China supports the UWSA with weapons and financial resources.195 These relationships come both from ethnic connections, as well as due to the UWSA’s origins as a communist force (Kramer 2007, 5). The regime in Yangon has gone so far as to accuse China of facilitating Chinese foreign fighters and mercenaries to fight with the UWSA (Reuters 2015). Whether or not the accusations about foreign fighters are true, the UWSA stands out as one of the only ANSGs in Myanmar who actively receive a large amount of support from a foreign state, in terms of weapons and materiel.196  As such, they are a rarity in Myanmar, since they are one of the only opposition groups who is clearly winning and able to remain functionally independent from the government, with whom they maintain a tacit alliance. They do not engage in alliances with other ANSGs and are able to push the government back when it attempts to make advances, making it one of the most powerful armed non-state groups in the region. The UWSA is well known to be deeply involved in the production, trafficking and distribution of narcotics.197 It has been referred to as the “world’s largest drug-trafficking armed militia” (Black and Fields 2006). The United Nations reports that at the time of research, more than 57,800 hectares in Myanmar was devoted to the production of opium, most of this in the control of the UWSA (United Nations Office on Drugs and Crime 2014). Although the UWSA has claimed to have eradicated opium production from the region, it is clear that it is still deeply involved in the production, smuggling and trade in narcotics.198  The method of rule of the UWSA is a centralized and hierarchical one inherited from its communist predecessors, and it leaves little room for local initiative or involvement in governance for civilians in general, and especially not for non-Wa ethnic groups within the UWSA’s territory.199 It is also known for its high levels of predation on the civilian population, displacing as many as 125,000 people to force them into narcotics production (Callahan 2007, 29). The forced resettlement is known to have created a great deal of resentment against the UWSA, even within the ethnically Wa populations that it purports to represent (Kramer 2007, 42). It has been reported that the UWSA “taxes” the civilian population, especially farmers who produce opium. However, this tax is reported to compose of the                                                           194 This has been discussed by Rogers (2010, 210) and by Ditlevsen (2014). 195 As noted by Kramer (2007, 4) and Barany (2015). 196 Chinese support for the UWSA has been discussed by Schaerf (2013) and Routray (2014, 70-71). 197 The relationship of the UWSA to narcotics, especially the production of opium and amphetamines, has been discussed in detail by Kramer (2007). 198 This has been discussed by Gravers (2014a, 150). Although there has been a decline in production of opium by the UWSA, it is still clear that’s involvement in trafficking continues. 199 As discussed by Callahan (2007, 28-29), Kramer (2007, xvi), as well as the Joint Kokang-Wa Humanitarian Needs Assessment Team (2003, 8). 87  entirety of crops produced by the farmers themselves.200 Most of the ethnic armed groups in Myanmar hold territory to which displaced people flee, fearing persecution from the government. The exact opposite trend is observable in Wa territories. All of the IDP camps in the area are outside Wa territory, and they contain large numbers of Wa who fled the UWSA for government held areas.201 Compared to other armed groups, it is clear that the USWA relies heavily on external patrons and on the narcotics trade. Quite apart from the other armed groups previously discuss, the UWSA does not build a relationship with the people it claims to represent, and therefore, does not hold much local support. Although it would not be accurate to state that the UWSA is a state proxy, its strength and its alliance with regime in power gives it a unique protected status among Burmese ANSGs. This allows it to continue its armed campaign of violence against other ANSGs, especially the SSA-S. Its support from China and the massive amounts of income it has received from the production and trafficking of opium further strengthen this position. This is in spite of the fact that the armed group has no support outside the Wa population, and to due to its predatory policies, little support from the Wa themselves. The UWSA, therefore, is winning without support.  3.7.2 Child Soldiers and the United Wa State Army Operating far outside the control of the regime, and in firm control of various illicit trade networks, there is little doubt that the UWSA continues to use child soldiers. The use of child soldiers by the UWSA has been well documented in recent decades, and it is clear that the UWSA purposefully recruits large numbers of children, and that many underage individuals are used as fighters, as well as in the production and trafficking of narcotics.202 The UWSA are known to conscript one soldier from each family, and should there be no man of fighting age, a child will be taken instead. A Japanese journalist witnessed the UWSA taking a 7-year-old girl to fight when no older male was available (2002, 131). A more recent study concluded that the UWSA has within its ranks as many as 2,000 child soldiers, making it the greatest violator of the norm other than the Tatmadaw (Chen 2014, 22). UWSA is listed in the Secretary General’s Annual Reports in 2012 through to 2015.203  Although Geneva Call attempted construct dialogue the UWSA on the subject of child soldiers, these were rebuffed by the UWSA and no engagement or negotiation towards signing a Deed of Commitment was conducted. Quite the opposite; apart from its allies in the narcotics trade and its Chinese sponsor, the UWSA has refrained from contact with not only the Myanmar regime, but also other armed groups (Boh Ze Kai 2016). Representatives of the UNFC expressed concern about the                                                           200 These practices were witnessed and reported first-hand by Hideyuki Takano, during a prolonged stay reporting from Wa territory (2002, 135). 201 The location of IDP camps outside of UWSA territory has been mapped by UN-OCHA (2017). Although a recent report, the camps have remained fairly consistent since the time of research. One first-hand report of civilians fleeing the UWSA is Nan Lwin Hnin Pwint (2016). 202 This was reported on in the past by Human Rights Watch (Human Rights Watch 2002). A second report in 2007 confirmed this (Human Rights Watch 2007). 203 AS reported in United Nations General Assembly (2012, 14), United Nations General Assembly (2014, 23), and United Nations General Assembly (2015, 23). 88  UWSA’s refusal to take part in talks with them.204 These signals are anything but esoteric; the UWSA has made it clear that it is indifferent to the norm and is not going to engage with international actors on the subject, or on other aspects of IHL. As such, the UWSA, as a group both winning and without support, it has not sought to engage with the norm, and refrains even from cheap talk on the subject, ignoring it altogether. Its primary audience is its state sponsors and not domestic audiences or transnational advocacy networks.  3.8 All Burma Students Democratic Front 3.8.1 ABSDF – Losing without support  In 1988, a wave of public protests against the regime were met with a swift and brutal response. Democratic protestors selected the auspicious date of August 8th (08-08-1988) to begin a campaign for democracy across the country, although when the regime deployed the Tatmadaw and police to the streets, as many as 3,000 protesters were slaughtered (Steinberg 2010, 79). For many of the pro-democracy students and activists, flight was the only option. Since people could not leave the country without an exit visa, many chose to flee east, to the areas held by opposition groups.  It was in this context that the All Burma Students’ Democratic Front (ABSDF) was formed in 1988, by a group of students hiding near the Thai border.  Their aim was to be “the first students’ army in the world to fight against a dictatorship.205 As it was originally composed mostly of students fleeing the regime’s crackdown, it is not driven by the desire to defend an ethnic group, but instead, to bring democracy to Myanmar (M. Smith 1999, 17). As such, it is the only Burmese armed group that is not founded on a narrative of ethnic resistance.  From the beginning, the ABSDF had a strong relationship with the KNLA, who provided them with training and weapons. However, since the ABSDF were predominantly Burman in ethnicity, they were provided only with limited weapons that were mostly badly functioning or out-of-date.206 Although in the first ten years of its existence the ABSDF also cooperated with the Kachin Independence Army (KIA), in the mid-1990s the KIA ordered the ABSDF out of its territory, further increasing the ABSDF’s near total reliance on the KNLA.207 The ABSDF has always been a small and weak armed group, and part of this was by the design of the early leaders. Although many students fled to the east in the early 1990s to join the ABSDF, the leadership took the counterintuitive decision to refuse entry to new fighters who came to join them. They did this in due to the strange rationale that they did not want to become a powerful force in the outlying areas, and instead wanted their supporters to remain in the major urban areas of the Burman regions (M. Smith 1999, 409-410).                                                            204 Nai Hong Sa, General Secretary of United Nationalities Federal Council. Interview with author. 07-07-2012. Chiang Mai, Thailand. 205 ABSDF activist Soe Myint, as quoted in Rogers (2015, 32). 206 This has been discussed by Kyaw Yin Hlaing (2004, 409) 207 The removal of the ABSDF from KIA territory is discussed by Human Rights Watch (2002a, 150). 89  As a highly disorganized and inexperienced armed group, the ABSDF were met with military failure almost from the outset. In the 1990s, fewer than half of the group’s fighters were armed (2003, 146). Military confrontations between the ABSDF and the Tatmadaw were disastrous for them, and it soon become torn apart by corruption and internal accusations of government infiltrators. The movement split in two, with many joining a rival organization, the Democratic Party for a New Society (R. H. Taylor 2009, 428-429). At its height it may have had as many as 10-20,000 members, however it has declined considerable in strength over time (Ford 2013, 93). At present, this has reduced dramatically to approximately 400 members (Myanmar Peace Monitor 2015).  The ABSDF has essentially been shrinking and losing territory since it was founded. It is puzzling that it still exists at all. Although it had many members at the beginning, most were Burman, and therefore were not supported by locals in the minority areas where the armed group was actually operating. It was not included in negotiations by other armed groups, and never received more than tacit acknowledgement from others in the armed struggle.208 As such, it never achieved strength, and has been on the verge of collapse for nearly a decade. Far from engaging in armed struggle, the ABSDF is largely reduced to existing as a means of protecting the students who founded it. Since they are unable to return home or to Thailand for fear of persecution, by remaining totally reliant on the KNLA for safety, while voicing involvement in the armed struggle, these last few leaders are able to seek limited protection and are unable to engage in armed violence. As such, the ABSDF will be described as losing without support.  3.8.2 Child Soldiers and the All Burma Students Democratic Front The ABSDF has been limited in the extent of its violations. As noted earlier, due to lack of resources, the ABSDF was unable to accept many recruits, and has historically turned away many thousands of volunteers even though it has a relatively small fighting force. Human Rights Watch reported that some child soldiers who had deserted the Tatmadaw had ended up joining the ABSDF (2002a, 45). The Tatmadaw is the largest violator of the child soldiers’ norm in Myanmar, and likely in the entire world. Since it conscripts many child soldiers and deploys them into the minority regions of the country to engage with armed groups, it has had a constant problem of desertion.209 Many individuals desert from the Tatmadaw in hopes that they might join with an armed group. However, most armed groups are reluctant to accept ethnically Burman fighters into their ranks, as in some cases they are fearful of spies, while in others they have open hostility to Burman people. Since the ABSDF is a predominantly Burma organisation, it is the general practice of many armed groups to send Burman deserters to the ABSDF. Whereas many are turned away, many are also accepted into the ranks as soldiers. This includes children. Therefore, whereas the ABSDF may not engage in the active recruitment                                                           208 The decline of the ABSDF from the 1980s to present was discussed in greater detail by Kyaw Yin Hlaing (2004, 408-412). 209 Desertion of children from the Tatmadaw has been discussed in detail by Human Rights Education Institute of Burma (2008, 45-46). 90  or conscription of child soldiers, it certainly has a number of child soldiers’ in their ranks. From the perspective of the ABSDF, this be seen as an act of mercy, since to refuse the children entry would likely result in their death at the hands of either the Tatmadaw or an opposition group.210  In interviews with ABSDF commanders, the ABSDF was presented as a civil society organization which did not take part in the armed struggle – a clear falsehood.211 In further interviews with members of the organisation, little was acknowledged on the subject of the ABSDF and its armed struggle, and the only admissions regarding child soldiers was that the ABSDF was a peaceful organization that helped child soldiers escape from the Tatmadaw.212 Although the ABSDF was approached by Geneva Call to discuss signing a Deed of Commitment, the ABSDF terminated the process quickly (Geneva Call 2013a). One commander, known by the nom de guerre “Snake” noted that Geneva Call had come, but that the ABSDF was not particularly interested in what they had to say. He commented that although Geneva Call had some interesting tips on how to avoid landmines, when it came to child soldiers “we would rather address it ourselves.”213 It was clear they were engaging local actors to desperately drum up support of whatever kind they could, including other armed groups.214 The ABSDF has not engaged in any other attempts to demonstrate compliance with the norm, even though its patron and protector, the KNLA, has done so. As an organization not only losing without support but also on the verge of total collapse, it is unlikely that the ABSDF will continue long into the future as an armed group.  As such, it is not engaging in legitimacy-seeking behaviour in the same way as other armed groups, since its leaders are more concerned with the immediate needs of survival.  It is therefore scrambling to get immediately local support in whatever form it can, and its legitimacy-seeking behaviour (to the extent that it exists) is directed at local audiences who do not support them.  3.9 Pa-O National Liberation Army 3.9.1 PNLA – Losing with Support The Pa-O are a Karen-speaking ethnic group, who do not consider themselves to be ethnic confrères of their fellow Karen speakers.215 They number less than 600,000, and primarily inhabit the south-west region of Shan state (Gravers 2014a, 157). The Pa-O began rebelling against the regime as early as 1949, joining the initial wave of rebellion against the Burman dominance (Ardeth Maung Thawnghmung 2008, 7). Although they have been waging the armed struggle for almost seventy years,                                                           210 As reported by Commander Snake, All Burma Students Democratic Front.  Interview with author, 20-06-2012, Mae Sot, Thailand. 211 Commander A (name withheld), All Burma Students Democratic Front.  Interview with author, 20-06-2012, Mae Sot, Thailand. 212 Commander B & C (names withheld), All Burma Students Democratic Front.  Interview with author, 20-06-2012, Mae Sot, Thailand. 213 Commander Snake, All Burma Students Democratic Front.  Interview with author, 20-06-2012, Mae Sot, Thailand. 214 As reported by sources both in the KNLA and the ABSDF 215 The term Pa-O is also transliterated as PaO, Pa-Oh, Pa'O and Paoh. 91  the Pa-O armed groups never attained the level of organization and strength shown by other armed groups such as the KNLA or KIA. The Pa-O movement was plagued by internal dissention and strife, leading to an alphabet soup of Pa-O organizations constantly defecting and re-aligning themselves with respect both to each other and to the regime.216 One of the largest groups representing the Pa-O minority – Pa-O National Organization and its armed wing the Pa-O National Army (PNO/PNA) signed a ceasefire with the regime in 1991 and were granted exclusive control over an area that became called the Pa-O Self-Administered Zone (Kramer 2009, 4-9). Upon signing a ceasefire with the government, the PNO/PNA was permitted to retain their arms, territory, and illicit trade, provided they did not engage in direct hostilities with government forces (Steinberg 2010, 111).  Pa-O who cooperated with the government were able to invest in a lucrative jade mine, bringing considerable wealth (Steinberg 2010, 112). As such, whereas the opposition has been historically weak and under-resourced, those Pa-O leaders who defected to the government were welcomed with immediate financial gain. However, many Pa-O were dissatisfied with this arrangement, since the Self-Administered Zone was a rump of traditional Pa-O territory, and the regime was still committing abuses against Pa-O in other areas, as well as against other minorities. Unity amongst the Pa-O was finally achieved in 2009, when Khun Okker led a committee to bring together all independent Pa-O organizations under a single banner (Burma News International 2015, 199). This group became known as the Pa-O National Liberation Army (PNLA), supported by a political wing known as the PNLA (Pa-O National Liberation Organisation). As a relatively young movement, and with only 400 fighters, there is little scholarly analysis available on the PNLA.217 Since it was founded as a coming-together of formerly competing organizations, it not only has a broad base of support, but also a formal constitution designed to spread power between a large committee of twelve committee members who are elected, and in turn, elect a Chairman (Myanmar Peace Monitor 2015). The leaders of the PNLA are aware of their tenuous position, noting that they have ceased trying to engage the government in battles, and instead only engage in “guerrilla war”; a euphemism for running and hiding.218 It represents, however, a fascinating case study for the purposes here, since it is a very small movement that risks being overwhelmed at any point by the Tatmadaw. Further, it has such a low profile that even regional news media rarely mention it. As such, as a group that is losing with support, it still is working to attract international legitimacy and attention even when it doesn’t see any real gain from either.                                                            216 Some of the now defunct Pa-O organizations include the Pa-O National Army/Pa-O National Organization, the Pa-O People's Liberation Organization, Shan State Nationalities Liberation Organization, and the Shan Nationalities People's Liberation Organization (Myanmar Peace Monitor 2015). 217 This estimate on the size of the PNLA is provided by Myanmar Peace Monitor (2015). 218 Khun Myint Tun, Pa-O National Liberation Army.  Interview with author, 11-07-2012, Mae Sot, Thailand. 92  3.9.2 Child Soldiers and the Pa-O National Liberation Army The PNLA has, in its relatively short history, been intimately connected to the child soldiers’ norm. As noted above, it was losing the conflict almost from the outset, however by 2014 it was engaging with the norm. Of all the organizations under analysis here, the PNLA is the only one that has engaged with the child soldier’s norm since its foundation. The head of the PNLA expressed that signing the agreement was vital for the organization’s continued existence.219 Soon after it was formed in 2009, it began to engage with international actors. The PNLA signed a Deed of Commitment with Geneva Call regarding children in armed conflict in 2014 (Geneva Call 2014). The PNLA therefore deftly avoided the issue of child soldiers since it has acted in a norm compliant manner from the outset. Its predecessor organizations – most notably including the Pa-O National Army and the Pa-O People's Liberation Organization – almost certainly used child soldiers in the past.220  Because it is technically a new organization, the PNLA is able to create a break with the past. Even when its predecessor organizations were norm violators, this does not necessarily tarnish the PNLA’s legitimacy at home or abroad. The PNLA can disregard past violations by other organizations (even if they included many of the members of the PNLA) and can focus instead on PNLA compliance with the norm. Although it is important not to exaggerate the break with the past that came with the foundation of the PNLA, it is also important to acknowledge it as a separate entity from its predecessors. In addition to its being a coming together of multiple armed groups, it was created with a new leadership structure and covered a larger and more diverse population than had its predecessors.   The PNLA’s rapid engagement with the norm, with both domestic and international actors, is evidence not only of the increasing importance of international legitimacy for armed groups, but also their increased desire of many to pursue it. The PNLA had also been petitioning the Japanese government, not for material support, but rather, for vocal support of the PNLA in their desire to be seen as a legitimate representative of their people. Although the PNLA were not looking for full recognition, they hoped that the Japanese government would acknowledge them as “right” in their struggle.221 The relationship with Japan may not have been directly related to the child soldiers norm but was part of an overall strategy to appear as the legitimate leader of the Pa-O people. As a group losing with support, therefore, it has engaged in legitimacy-seeking behaviour towards the international community in the form of transnational advocacy networks with regards to child soldiers, and even to other states.                                                           219 Colonel Khun Okker, Pa-O National Liberation Army. Interview with author, 07-07-2012, Chiang Mai, Thailand. 220 As most of these organizations operated prior to the child soldier’s norm becoming internationally salient, there is little research or analysis on the extent of their violations. Human Rights Watch has noted that the Pa-O National Army was likely a violator (2002a, 157-158). The same report notes that all armed groups likely used child soldiers to one extent or another. Ascertaining the extent of past violations pre-2009 is well beyond the scope of this project. For the purposes here, it is enough to note that it is very likely that violations occurred, and an attempt has been made to prevent any further violations. 221 Colonel Khun Okker, President Pa’O National Liberation Organization.  Interview with author, 07-07-2012, Chiang Mai, Thailand. 93  For those at the head of the organization, the reasons behind this are clear. The leader of the PNLA noted that the decision to sign the Deed of Commitment was made in order to be seen as a “the real government of the Pa-O who has a right to lead.”222 It was for the same reason that he led a delegation to Japan upon an invitation from the Japanese Ministry of Foreign Affairs. Even though he knew he would get no material assistance, since all aid from Japan must be approved by the Burmese government, the leader Khun Okker still felt it important to go to meet with foreign governments in order to show their organization as the legitimate representative of the Pa-O people.223  They also continued their meetings with Geneva Call. As with other organizations, the PNLA was developing a Code of Conduct with Geneva Call, based on that written by the KNLA.224 Perhaps more than other organizations, the PNLA were very quick to note the importance of their beliefs and identity. Even within Myanmar, very few have heard of the PNLA, and they receive almost no international exposure. One fighter emphasized the importance of protecting children, since it was the essence of their cultural beliefs.225 Another, when asked why he thought the PNLA should protect children, referenced the two main groups of Pa-O; “Because we are Pa-O! Red or white, doesn’t matter!”226 Another officer bragged that things had changed so much in the PNLA that he “knew nothing about child soldiers” since it had been so long since the PNLA had any.227 Of all organizations, it may have been easier for the PNLA to convince its members to follow the norm. Due to its small size, it is likely that every single fighter in the organization has a direct personal relationship with the leader. However, it is clear that the group’s identity-narrative of being a small group fighting overwhelming odds has a vital role in the process. The PNLA collectively seized an opportunity to punch above its weight, and fulfil the role of an important international actor, in order to be seen as the legitimate representative of their people.  3.10 Summary The trends in Myanmar indicate clear trends. Armed groups are engaging at various levels with international actors and are doing so in search of legitimacy. In cases such as the KNLA, MNLA, KA and PNLA, armed groups are seeking to cooperate with international actors and seek legitimacy for their struggle. Although each has a unique identity-narrative, expressing different meanings and beliefs in different ways, in each case where the norm was successfully implemented, the fighters in the organization expressed the norm as a vital part of their beliefs, even though their organization was known to violate the norm in the past. In the case of the ABSDF – an organisation that is losing without support and is on the verge of collapse – little engagement is occurring with the norm. Armed groups who are winning without support also have no engagement with the norm                                                           222 Colonel Hkun Okker (Pa-O National Liberation Army). Interview with author, 07-07-2012. Chiang Mai, Thailand. 223 Colonel Hkun Okker (Pa-O National Liberation Army). Interview with author, 07-07-2012. Chiang Mai, Thailand. 224 Colonel Hkun Okker (Pa-O National Liberation Army). Interview with author, 07-07-2012. Chiang Mai, Thailand. 225 Maung (name changed) (PNLA). Interview with the author, 11-07-2017.  Shan State, Myanmar. 226 Name withheld (PNLA). Interview with the author, 11-07-2017. Shan State Myanmar. 227 Tin Shwu Oo (PNLA). Interview with the author. 11-07-2017. Shan State, Myanmar. 94  One corollary trend noted in Myanmar is that the armed groups who engage with the norm admit both previous violations, as well as troubles implementing the norm. Whereas it may seem like an excuse to say that children are trying to join armed groups, it is still true. Children in Myanmar, for a variety of reasons, still try to join armed groups even when the group in question has a clear policy prohibiting child soldiers. Far from being a young conflict, the confrontations in Myanmar are almost 70 years old, and have been smouldering at a low intensity of violence for decades. Although there are ceasefires regularly signed, these mean little in the context of Myanmar, where ceasefires are repeatedly signed and violated. Although fighting is certainly on the decline, the Tatmadaw is still targeting specific armed groups, most notably the Kachin Independence Army (B. Rogers 2015, 230). Since the KIA is a member of the UNFC, the other constituent members of the UNFC are refusing to settle a long-term peace with the regime until a nation-wide ceasefire is in place. Further, armed groups are stilling fighting each other, most notably those in Shan State (Petrie and South 2014, 224-225). Therefore, although there are encouraging signs that the country is moving towards peace, the conflict is not over. Further, in Myanmar there are interesting trends that corroborate the conclusions of this study involving armed groups not under direct study. Although contact was not made with the Kachin Independence Army (KIA), it is interesting to note that it displays trends similar to the other armed groups. It has successfully managed to continue its campaign against the government, due to its strong organization and support from the Kachin people where it operates (Chen 2014, 25-26). It has also not sought to engage with the child soldiers’ norm, refusing to engage with Geneva Call while also continuing to openly recruit child soldiers as young as 13 (Child Soldiers International 2015a). A second example, the Chin National Army (CAN) is an armed group representing the Chin ethnicity. Although the only popular representative of the Chin people, the group has been losing ground rapidly to the Tatmadaw forces, thus suggesting that it is losing with support. Interestingly, the CAN has been a passionate supporter of interactions with Geneva Call, signing three Deeds of Commitment including one related to child soldiers.228  It is interesting to note that once Geneva Call became engaged in Myanmar and worked with some armed groups, other groups began to see the benefits of such interaction and approached Geneva Call requesting increased ties and interactions (Geneva Call 2012a). This suggests evidence of norm cascade amongst armed groups, in which armed groups who do engage in the process of accepting the norm gain legitimacy, thus compelling others to follow suit.  For the purposes of this study, it is useful to chart the armed groups under study into a binary table of winning or losing and with or without support. Although they were not full cases under study, the KIA and CNF have also been included in the below Table 6:                                                             228 Human Rights Watch reported that the CAN were losing ground fast as early as 2009 (2009). Geneva Call reports that the CAN has signed Deeds of Commitment with regards to landmines, child protection and gender issues (2016). 95     Trends in the Conflict     Winning   Losing  Level of popular support held     Popular support    Reject norm (cheap talk):   Kachin Independence Army (KIA)    Follow norm:  Karen National Liberation Army (KNLA)  Karenni Army  Mon National Liberation Army  Pa’O National Liberation Army  Chin National Front    Little or no popular support    Reject norm / no cheap talk:   United Wa State Army (UWSA)   Cease Operations:   All Burma Students Democratic Front (ABSDF)   Table 6 – Conflict Trends and Popular Support in the Myanmar Conflict   The cases under study in Myanmar demonstrate that armed groups interact with international actors in search of legitimacy. They do so even when it runs contrary to their immediate gains. In some cases, as with the PNLA, the pursuit of legitimacy runs contrary to long-term aims as well, since the group could abandon the concerns of legitimacy and follow other Pa-O who defected to the government sides. Both the leaders and fighters would be able to gain greatly from lucrative mining and lumber concessions in the region.  Far from being intransigent norm rejecters, armed groups in Myanmar have actually been passionate enthusiasts of the child soldier’s norm. In some cases, as was noted with the Karenni Army, the armed group itself has been the one trying to attract international actors’ attention to the norm 96  rather than the other way around. Even though the Karenni Army reported to being ignored by international actors with regards to the child soldiers’ norm for an extended time, the organization was still attempting to get international attention on the issue from the United Nations in addition to Geneva Call. This speaks not only to the repeated by attempts by armed groups to engage with legitimacy, but also the inconsistent application of ‘naming-and-shaming’ policies by the UN and other advocacy groups, who condemn armed groups for violations without offering a means for armed groups to demonstrate a desire to comply with norms and IHL.   97  Chapter #4: Syria  Child Soldiers on the Frontlines of Transnational Conflict   “[According to] the heart of Islam, you shouldn’t ever make children a soldier.”  Commander of Jabhat al-Nusra (name withheld) – Aleppo, Syria.229   “I don’t want to kill anyone anymore.” Former child soldier with Jabhat al-Nusra (name withheld) – Aleppo Region, Syria.230   4.1 Introduction In July and August of 2013, the Aleppo region of northern Syria was the focal point of a brutal contestation between the Syrian regime, a myriad of armed non-state groups (ANSGs), and multiple international intervention forces. During the day, the city echoed with the thunder of mortar and artillery, while at night the regime’s tanks pounded the frontline as the Soviet-built MiGs targeted military and civilians targets alike with Russian weapons.231 The people of Aleppo had suffered from a shocking range of violations of the international laws of war. Included in this were the targeting of civilians with chemical weapons, cluster munitions, thermobaric weapons, white phosphorous, barrel bombs, anti-personnel landmines, and even genocide. 232 The Syrian conflict is a rarity in the twenty-first century, as it had a clearly demarcated frontline with competing forces on either side of it. Whereas                                                           229 Commander of Jabhat al-Nusra (name withheld). Interview with author, 04-07-2013, Marea, Syria.  Interview with author, 05-07-2013, Aleppo, Syria. 230 Former child soldier associated with Jabhat al-Nusra (name withheld). Interview with author, 04-07-2013, Marea, Syria. 231 As witnessed by author at the time of fieldwork, July-August 2013. 232 The use of chemical weapons in Syria up to 2015, is detailed in Barnard and Sengupta (2015). The remnants of cluster munitions – presumably of Russian origin, due to the Cyrillic script on the casings – were personally witnessed by the author in Aleppo, July 2013. The use of thermobaric weapons has been reported by Cumming-Bruce (2013). Use of white phosphorous in Aleppo, and nearby towns (including Marea) was reported to the author by City Councillors in Marea [Ahmed (name changed), Marea City Council. Interview with author, Marea, Syria. 04-07-2013]. The use of barrel bombs has been a common theme in the conflict, as noted by Colum Lynch (2015). The use of anti-personnel landmines was reported and condemned relatively early in the conflict by the International Campaign to Ban Landmines (2012). Accusations of genocide were levied against ISIS for targeting ethnic and religious minorities in both Syria and Iraq (Rothschild 2016). 98  most contemporary conflicts are diffuse and rural in nature, the conflict in Syria represents a nascent trend in high intensity urban warfare.233 Aleppo city was cut in half by a frontline and even a No-man’s-land. For the civilian population of the area, the cold brutality of mechanized warfare was visited upon the city with astonishingly destructive results on an hourly basis.   Figure 4 – Location of Research: Aleppo Governate, Syria. 234  At the same time, soldiers on the ground engaged in pitched battles of urban warfare, fighting block to block. They fought in the same style as the urban guerrilla more prevalent in the Cold War, and as famously conceptualized in the writings of Marxist revolutionary theorists of guerrilla warfare.235 They used a range of strategies including civilian intimidation, terrorism, and hit-and-run tactics, while                                                           233 This emergent trend of contemporary urban insurgency has been discussed in detail by Kilcullen (2013). 234 Original map is freely available under Wikipedia Commons, although edits to original have been made by the author. Original available at Wikipedia (2011). 235 As discussed famously by Marighella (1969) and Taber (1970). 99  avoiding direct confrontation with the regime forces, their Iranian allies, and the air advantage that both held. In addition, Salafist groups were waging jihad against the Alawite Shia, with uneducated local recruits crying takbir alongside international fighters – including Canadians – as al-Qaeda and ISIS joined forces with local Islamist organizations that appeared in the wake of the regime’s retreat.236 In the “liberated” areas of what the opposition called “Free Syria”, entrepreneurs were appearing to conduit materiel, essential goods and even luxuries. Local strongmen were able to adopt a dual role as both the head of battalions of ready-to-die youths, and as the CEO of a new armed economic enterprise that advocated both an economic and militarized rationale for the armed struggle (Dark 2013). During a pitched battle between an Iranian intervention force and community-organized paramilitary groups, I spoke with a group of officers of the Syrian Islamic Liberation Front struggling to make sense of the fact that their mortal enemy Israel was conducting airstrikes against the regime that was bombing our location, while we debated the logic of why Israel and al-Qaeda were fighting on the same side.237  Taken together, all of these trends and events have combined to make Syria a context of both intense suffering, and high interest to policy-makers and scholars.  For the purposes of this study, the Syrian Conflict presents a compelling set of cases for analysis. Of all contemporary conflicts, it is this war that undoubtedly receives the greatest amount of attention in the popular media at the time of writing. As of April 2016, the conflict has led to at least 300,000 deaths, and more than 4.1 million people have been forced to flee the country.238 The conflict has had an overwhelming impact on the children and youths of the region. In addition to inestimable numbers of child soldiers, children are increasingly being used in ever more violent roles. A number of cases have emerged of children being used as executioners, or even as suicide bombers.239 In addition to the violence, ANSGs are engaged in dialogic struggles to gain access to legitimacy as a strategic commodity, in order to assert it internationally in their ongoing struggle. This chapter begins by presenting an overview of the conflict in Syria. Included in this will be a short history of the conflict as well as its underlying causes, and an overview of all the key domestic, transnational and international actors currently engaged in the conflict. Analysis will then proceed in establishing each relevant ANSG as either winning or losing, and with or without support. Discussion will then provide an overview of the status of children and child soldiers in the conflict, as well as analysis of key trends regarding which armed non-state groups are known to have used child soldiers,                                                           236 Abu Asif (name changed), Islamic State of Iraq and al-Sham. Interview with author, 11-07-2013. Aleppo, Syria. 237 Hattab (name changed) Syrian Islamic Liberation Front. Interview with author. 06-07-2013. Aleppo, Syria. 238 The number of at least 250,000 people killed by January 01, 2016, was reported by the Syrian Observatory for Human Rights, an independent organization that produces neutral estimates of deaths (2016). United Nations High Commission for Refugees reports the numbers of 4.1 million refugees as of June 2015 (2015). 239 In a number of videos released online, children have been seen executing adult prisoners, as reported in Withnall (2015), Carissimo (2015) and Dearden (2015). In more notorious cases, children have even been made to behead prisoners. The use of children as suicide bombers was reported by Human Rights Watch (2014a, 20). 100  and in what roles. Analysis will then turn to the specific ANSGs to be used as case studies, with data from fieldwork being connected back to the overall argument presented here. Overall, it will be shown that all ANSGs compete over legitimacy with the four audiences discussed in Chapter One. It is ANSGs who are in the losing with support category who are the most likely to engage with international actors in order to demonstrate compliance with the norm against the use of child soldiers. In these cases, I will discuss the role of identity-narrative, since after a decision to adopt the norm is made, the norm must be reflected in the pre-existing identity-narrative of those within the armed group. In the case of ANSGs who are winning with support (for example, the Islamic State of Iraq and al-Sham240 at the time of research) it is possible to demonstrate the use of “cheap talk” regarding the norm, in the face of widespread norm violation. In the case of winning without support (e.g. the Shabeeha), it will be shown that violations occur, and cheap talk is rare as there is almost no engagement or interaction with the norm, not even to reject it. Finally, in the case of ANSGs that are losing without support it will be shown that widespread violations of the norm take place, and though cheap talk may occur, it is unlikely, as the ANSG will rapidly disband.  4.2 Cases Syria presents an interesting context for this research for several reasons. The first is the substantive importance of the crisis in Syria. In addition to the hundreds of thousands that have been killed, more than 11 million have been displaced (United Nations High Commissioner for Refugees 2015). The scale and the intensity of the conflict command attention, as do the resounding impacts it is having not only in the immediate region, but across continents, capturing the attention of populations near and far. The scale of the conflict, and the attention it receives, are matched by the intensity of the violence being used. In images broadcast around the world, intensely shocking acts of violence have made the conflict notorious, even though most of the forms of violence are neither new nor rare. Whereas Vietnam may have been famous as being the conflict that was brought into people’s living rooms via televisions, the Syrian Conflict is one of the first that is being live broadcast onto people’s computers and phones. A second reason that Syria presents a valuable source of research is the scale to which child soldiers are being used. The abuse of children in the conflict is an incalculable humanitarian tragedy, and no clear estimate of the number of child soldiers is available. It is clear, however, that at least thousands and perhaps tens of thousands of children have been drawn into the conflict as child soldiers. Whereas some ANSGs use small numbers of child soldiers in an incidental fashion, others do recruit openly, advertising their success through their future generations of fighters. ISIS uses children openly, referring                                                           240 As with other armed groups and locations whose names have been changed, the strategy adopted here will be to use the terminology used by the individuals at the time of research. Although at the time of research the term ISIS was in use by the organisation, in June 2014 the armed group announced its English name would be Islamic State, in order to reflect its aspirations to build a global Caliphate, and not merely one limited to Iraq and Syria (Al Jazeera 2014). However, the approach adopted here will be to continue the use of the term ISIS. 101  to them as “cubs of the Caliphate”.241 Far from limiting itself to abusing child soldiers in ways seen around the world (as fighters, porters, etc.), ISIS has shown a cruel though adept skill at innovation, using child soldiers in ever more shocking ways. ISIS has also been known to repeatedly use children as executioners, shooting or beheading adult captives. A third reason that Syria presents an interesting context for research is that it is a conflict which features a wide array of ANSGs, who in turn vary in their acceptance of the norm. There is no clear estimate for the exact number of ANSGs involved in the conflict. There are many, and they range in their ideologies, tactics, patrons and strategies. Estimates have placed the number of armed groups in Syria in the hundreds, and some as many as thousands.242 Given the intensity of the conflict, the ANSGs in it become hard cases for the overall theory, since they engaged in an existential struggle. If the theory is wrong, they should be recruiting and using child soldiers especially when they are losing. To put the argument of this project simply; if the ANSGs don’t care about legitimacy, they should be using as many child soldiers as they can, in as many roles as they can. This would especially be true of ANSGs that are losing.  Included below in Table 7 is a summary of the key features of the ANSGs in the Syrian Conflict to be here discussed. The information provid