First dilemma: how best to contextualize individual narratives?
When drawing on personal narratives, the obvious risk is to fall into an extreme form of methodological individualism that partly distorts social phenomena be-3 I was helped by a research assistant in that process. We split the interviews among each other.
cause it fails to relate individual experiences with the wider structure. The analyst in fact has to be particularly careful not to exaggerate the instrumental rationality of individuals too strongly, or there is a real danger of remaining confined in a deterministic approach that bounds human choices and decisions by motives and previously existing causes, regardless of the structural aspects.4 To avoid such trap, I engaged in a process of document reconstitution to get a good grasp of the local context (by ‘local’ I mean the immediate environment of the people under study). I was particularly interested in developing a coherent version of the recent history of the western region by generating a detailed timeline of violent events since 2002, in order to get a sense of the general atmosphere in which res-pondents had evolved.5 After all, in many cases, among the driving rationales for explaining military engagement were the circumstances. I therefore extensively tapped into journalistic sources, particularly Ivoirian newspapers archives (Fraternité Matin, 24 Heures, Notre Voie, L’inter, Le Front, Soir Info) and those of national and international press agencies (Agence Ivoirienne de Presse, Agence France Presse, BBC, PANA Press, Reuters). Luckily, Côte d’Ivoire is home to a plethora of media and hosts at least a dozen daily newspapers with wide circulation throughout the country and easy internet access. Far from denying the partiality of certain sources and the politicized nature of some documents (Ivoirian newspapers are well-known for their political engagement and aggressive tone), my goal was to extract the most ‘factual’ information; hence, I treated the various articles as valuable primary documents that ac-counted for a particularly violent period. I completed this documentation with UN and INGO situation reports on western Côte d’Ivoire, impartial forces up-dates, and secondary sources (International Crisis Group analysis, IRINNews, and UNOCHA Bulletins). This reconstruction work had the merit of clarifying the different conflict phases and of introducing a certain temporality to the analysis of the conflict in the west of the country. It was a necessary step to understand, a posteriori, local processes of mobilization.
4 Long (2001) in fact warns us that many ‘micro’ studies fall short because of a tendency to adopt a voluntaristic view of decision-making, by highlighting too much the transactional nature of actors’ strategies while not examining enough how these individual choices were shaped by larger frames of meaning and action and by the distribution of power and resources. One acts in a certain way not only because of individual characteristics. The ability of people to build up room to manoeuvre only takes on full meaning when it is related to structural aspects and specific historical patterns. These actually partly explain how such room to manoeuvre is framed (Abbink, 2005; Carney, 1999; Giddens, 1979).
Man and Guiglo were both extremely ‘militarized’ and ‘humanitarianized’ in 2006 and 2007 – the period when I was doing fieldwork – which raised several dilemmas. How could I best approach such a messy field without unnecessarily putting myself and others at risk? How could I avoid being taken for a humanitarian practitioner while using the premises they used to conduct my first interviews? In Man, at the time of doing fieldwork, the local administration was completely managed by rebel officers. No recruits had yet been officially demobilized and none had received financial compensation. The mere prospect of releasing low-ranking recruits was not even debatable with the highest in command five years after the start of the conflict. In Guiglo, 981 militia elements had gone through a demobilization process and had received some kind of financial compensation. The local administration was also fully military and pro-government militias had an extremely bad reputation. They were particularly prone to hostile demonstrations against the UN and the French impartial forces, a characteristic I could fully observe in early 2006, when I was not yet doing fieldwork for this study but was nonetheless in the area for consulting activities.6 It was in such settings that I started my research at the end of 2006. Needless to say, timing was quite crucial for the success of my data collection and I could probably not have approached the youths the way I did if I had conducted the interviews earlier (at least in Guiglo). If I was regularly monitoring the changes taking place in terms of local security, my fieldwork occurred during a relative period of calm and I was never caught in the wrong place at the wrong time. Because of the sensitivity of the topics I wanted to talk about and the risk that my actions could be misinterpreted by the military hierarchy if I went into too much detail, I adopted a rather low profile and opted, for this particular project, not to seek formal approval of the local authorities when carrying out the research. When I met the military Préfet in Guiglo and the Com’Zone in Man, I introduced myself as a researcher from the University of Amsterdam interested in studying the impact of war on local youth. I purposely kept the definition broad to avoid 6 On 18 January 2006, following a peaceful demonstration in front of the UNOCI base to protest against a controversial communiqué (the International Working Group on Côte d’Ivoire had just announced that the mandate of the Ivoirian National Assembly due to expire on 16 December 16 2005 would not be extended), the event degenerated into a violent confrontation between the UN Bangladeshi peace-keepers and an angry mob. It resulted in the deaths of five protesters and the wounding of thirty-nine others (Human Rights Watch, 2006). In response to the shootings, a militia leader called everyone on the radio ‘to come into town to avenge the death of those struck down by the assassins’ bullets.’ Shortly after, militia leaders once again used the radio to incite violence against all UN and human-itarian organizations. Several offices were burned, twenty cars tagged humanitarian were extensively damaged, office equipment was looted (computers, electric generators), as well as food and medical supplies stocks (the WFP warehouse was completely emptied).
giving the impression that I was only focusing on militarized youth. As an extra precaution, I temporarily downplayed my French identity, overemphasizing my Dutch background instead. This discourse was generally well received and in Guiglo, the Préfet was particularly keen on sharing his thoughts on the subject. As I did not feel comfortable undertaking such research without the quasi cer-tainty that an umbrella organization would help me reach a safer place elsewhere should the local situation abruptly deteriorate, I approached two international development agencies and asked them to include me in their evacuation plan. One of them was the German NGO GTZ-IS. I had heard they had just begun a pilot project in Man and Guiglo for reinserting young recruits and I used them as gatekeeper to get access to 216 respondents.
Third dilemma: doing research tapping into humanitarian practice
There is an inherent tension in doing research while tapping into practice and in combining an approach that critically examines actions of development by using external interventions as means of accessing the bulk of respondents. There seems to be some kind of irreconcilable contradiction between the two and a real danger of bias for the analysis. I would argue that such a tension can yield productive results if the researcher pays sufficient attention to a range of things. To begin with, planned interventions offer a relatively easily accessible tank of potential respondents. While some would argue that it biases the selection of cases, not selecting respondents who are engaged in projects would probably also lead to bias as humanitarianism has become so much part of the local environ-ment that it would be a mistake to systematically discard anyone who participates in an NGO-driven intervention. Now indeed, critics point more to the question of balance. The problem is not so much framed in terms of whether or not project participants should be excluded from a selection of cases, but to keep a fair proportion of those who are benefiting from some kind of assistance and those who are not, to overcome the risk of encountering the same type of profile among the respondents.7 I would argue that such precaution is perhaps not necessary in certain contexts where external interventions have very limited impact on people’s lives. Also, it tends to conceive ‘project participants’ or ‘target popula-tions’ as a homogeneous group, with ‘post-project’ lives drastically different (and better) from the ‘pre-project’ ones. In reality, there is a great variety of patterns and different individual responses to similar structural and circumstantial condi-tions. In western Côte d’Ivoire, the reinsertion projects under study had very limited impact on project participants’ lives. I nuance that point more in Chapter 7 It could be hypothesized that those who benefit from programme interventions have a better social capital than those who do not, so that the socially marginalized are left out.
9, but the general impression was that the project fulfilled more a function of networking (with project participants adding ‘fellow participants’ and ‘project staff’ to their social network and range of opportunities) than a function of sup-port per se, although this indeed varied depending on individuals.
The fact that my initial encounters took place in centres run by NGOs for reinsertion purposes was nonetheless an issue I needed to mitigate. During my first visit to the centres, NGO staff – with approval of their hierarchy – organized an informal meeting with the youths present that day, where I was given the opportunity to introduce myself and to explain why I was there. For the sake of clarity, I explained that I was interested in hearing the life stories of young people who had spent some time in the militias/rebellion in order to compile such testimonies in a book. I emphasized that I was not interested in names, but in understanding from their points of view what drove them in and out armed groups and why they acted the way they did. I also emphasized that nobody had to meet with me if they did not want to. Interviews were not compulsory. I stated several times that I was not part of the project staff, that I could not materially help (so as not to raise undue expectations), and that any information shared with me would be kept confidential. While it is unlikely that everyone present at the meetings understood clearly what I intended to do, several youths volunteered to talk to me on my next visits. In terms of order, as I came to realize afterwards, the first persons I met in Guiglo were close to the militia leaders, some were even related to them. When I interviewed adolescent recruits in Man (they were hosted in a separate centre), the first person I met was the ‘President’ of the youths, followed by main members of his ‘bureau’. In order to help basic project man-agement, the creation of a certain hierarchy amongst the children had been encouraged by the local NGO running the centre to facilitate collective inter-actions with project staff. A ‘President’, a ‘Treasurer’, a ‘Secretary’, a ‘Chief of Hygiene’ had therefore been named by their peers and were mediating collective demands. During my next visits, more and more adolescents registered their names to schedule an interview with me, probably reassured by what early interviewees had reported to them and by my frequent visits to the centre. If one message had been clearly passed on, it was that I was not a threat.
In Man, Guiglo and Blolequin, most interviews were done inside, in a quiet room, and it was usually not a problem to conduct them in French (commu-nication was difficult with ten respondents in Man and interviews had to be cut short as I did not have enough knowledge of their respective local language to be able to carry on in-depth interviews without the help of a translator). Some interviews were filmed, some were taped and I took notes of the rest. With the youngest respondents, nearly all exchanges were filmed, unless they did not want this to be done.8 I was more cautious with older recruits, as I feared – perhaps too much – that there was a greater risk my intentions would be misinterpreted. I, for instance, intentionally avoided asking from the start if I could tape interviews. It was only if I felt a conducive climate in the one-on-one exchange that I would ask the respondent if he or she would not mind if our exchanges were recorded. Although I had my camcorder with me nearly all the time, I was determined to only use it if I was completely sure that my intentions would not be misunder-stood.
Fourth dilemma: how to empirically define youth for the scope of this study?
Since this work focuses on militarized youths, there is the need to define what is understood to be a ‘youth’ in this particular study. To be young is not a matter of biological age, and many scholars would agree with such a statement. Chauveau defines youth in terms of relational position. To him (and I share his view), being young is socially and culturally constructed, in relation to other generations, and in relation to access to relevant assets and resources that confer a certain social status (Chauveau, 2005a). If the notion of youth is a heuristic concept, there is also no universal definition of childhood, and the concept remains locally de-fined. Conceptualizing youth and childhood this way, in terms of local categories of perception, is in sharp contrast with the ‘target group’ categories built to meet the needs of external interventions. Those categories are in fact often constructed in ad hoc ways and according to age benchmarks defined by international standards, which are quite far from the lived realities of the people they attempt to define. In the normative approach promoted by planned interventions, a child is considered a child until the age of 18 (the age after which he/she is no longer eligible to receive child benefits), and a youth usually ceases to be young at 25 (the age limit is sometimes extended to 30). Defining childhood and youth along these lines inevitably suffers from a lack of solid grounds at the local level. Rosen has been among the fervent critics of such ‘politics of age’, a term he himself coined. When reflecting on the definition of childhood (Rosen, 2007), he argued that one effect of the ‘straight-18’ focus widely promoted in international law has been to shape the concept of childhood in a very strict way (bounded by numerical age) at the expense of more interesting reflections, discounting the more varied and complex local understandings of children and childhood and using age categories as instruments to advance specific agendas. Rosen espe-cially pointed out that the mainstream discourse has tended to stifle the fact that 8 It never happened. I forgot my camera twice during the period I did interviews, and each time, child-ren expressed some disappointment.
older teenagers are likely to differ from younger children in many ways and that there is a tendency to infantilize 16 and 17-year-old recruits in contexts where adolescence and military life are not necessarily seen as antinomic. He has also stressed that by focusing too much on the older teenagers (the recruitment age debate focused on the ages 15 to 18),9 the youngest ones have been forgotten, despite the fact that many child recruits are in their early teens when they are recruited (which this case study also illustrates).
But if youth and childhood are conceived as relational positions locally defined, does this thus mean that there are no limits to calling someone ‘young’ or a ‘child’? Abbink (2005) strongly argues in favour of such a benchmark, argu-ing that having no strict definition for youth does not mean that middle-aged people should be categorized as young. Even if some of the ‘middle-aged’ share common characteristics with younger persons (not yet having secured stable work, not yet having been in a position to raise a family, etc.), many of them miss the transition to adulthood because of poverty and deprivation (ibid) – I would add bad luck. As years pass by, their lives take a tragic turn and they eventually lose their youth. Their future no longer lies before them.
The line between ‘young’ and ‘child’ is a more blurred one and needs to be framed along cognitive development terms. Recent research has shown that children in different cultures are likely to engage in complex moral reasoning at a much younger age than expected (Boyden, 2007; Rosen, 2005; Rosen, 2007). In developing countries in particular, where most people are used to fending for themselves from an early age, context and experience have proved to play a 9 In the late 1990s, there was considerable debate on which minimum age for recruitment to set in inter-national law and a range of actors actively lobbied for abolishing the then marker of 15 as the mini-mum tolerated age for recruitment, and for raising this age limit to 18 in official documents (Harvey, 2000). INGOs were particularly active in pushing this view, along with the International Committee of the Red Cross, the Swedish government and the Quaker UN Office in Geneva (Brett, 2005; Rosen, 2007). Those in favour of the change were mainly arguing that the 1989 UN Convention of the Rights of the Child (UNCRC) had failed to completely prohibit child recruitment and were particularly keen on underlining a contradiction in the Convention since on the one hand, a child was defined in the text as anyone under 18 in need of special protection, and on the other hand, recruitment of 16 and 17-year-olds was tolerated. Proponents were therefore keen to have such an ‘anomaly’ corrected in international law by raising the benchmark to 18. On the opposite side, people feared that too much focus on age would distract international attention from more fundamental issues such as forced recruitment. Ryle (1999) notably argued that what eventually mattered was the way conscription took place, not chronological age. Whether recruits are 16, 18 or 21 is of lesser importance, as long as these people willingly enter the force. After several stalls in the negotiation process, an optional protocol to the Convention of the Rights of the Child on involvement of children in armed conflict was adopted in 2000 (United Nations, 2000). It called on states ‘to take all feasible measures to ensure that members of their armed forces who have not yet attained the age of 18 years do not take a direct part in hostilities’ (Article 3). With respect to state armies, it forbade compulsory recruitment before the age of 18 while tolerating voluntary enlistment of 16 and 17-year olds, provided states maintain minimum safeguards to protect the minors in their ranks. With respect to non-state armed groups, it uncondi-tionally forbade recruitment under the age of 18 and criminalized the practice (Article 4).
major role in the development of human cognition and in influencing modes of action. Age and maturity are no longer conceived as necessarily going hand in hand, and adulthood is no longer overestimated by assuming grown-ups’ immu-nity to being influenced.
For the scope of this study, I gave up trying not to bracket ‘youth’ into two figures, as I kept meeting 14-year-old mothers who were already fending for themselves for several years and 42-year-old ‘Présidents des Jeunes’ in the areas I toured (which appeared somewhat odd since they generally did not resemble adolescents or twenty-year-olds or people in their early thirties). If I tried in the beginning to look for some kind of remnants of an intricate age system in the societies I visited (a system in which generational issues are mitigated by the assigning of a social role to age groups, the maintenance of clear boundaries between them, and the existence of strict codes of behaviour), it became clear that such a system was long gone in western Côte d’Ivoire, and that the civil war had challenged it even more (despite the fact that some underlying ideas con-tinued to persist in terms of reciprocity and mutual obligations). So when is one young in this book and who was included in the sample of militarized civilians? I pragmatically opted for men and women between 14 and 35 years old.
Beyond methodological challenges, studying militarized youths is an ethical minefield. The ‘do no harm’ imperative is trickier to reach in conflict zones due to exacerbated political polarization, the presence of armed groups, and the general unpredictability of events. I tend to agree with scholars who point out that research cannot be ethically conducted everywhere and should not even be attempted in certain settings. One difficulty is to find satisfactory ways to address consent. If tackling the issue is often used as an example of good practice leading to an ‘ethically correct’ research, how best to do so is a major point of debate in culturally diverse settings. Another difficulty is to mitigate the potential stress respondents might experience as they recall disturbing events. But doing research on armed groups in war-affected areas is also disturbing because the researcher is fully confronted with certain practices and ideas about violence, whatever his/her own views are on deontological ethics. To what extent can one suspend judg-ment? How to minimize emotional shakes? Can one keep a fresh mind after hearing about the same atrocities for the umpteenth time? Researchers are not immune to feelings, and some discourses can be hard to swallow, even for the most open-minded. Wood (2006) mentions these emotional dynamics because she is persuaded inadequate attention to them may lead field researchers to make errors in judgment.
The ‘do no harm’ imperative
How to define ‘harmful’? Drawing on his experience as a member of a Canadian Research Ethics Board, Haggerty (2004) puts it this way: ‘In the eventuality that a research project poses a greater risk than what a person might encounter in his/her daily life, these risks must be managed by the researcher or the research cannot be conducted.’ Since they are related to a broader context than the research project alone, ethical concerns are put in some kind of perspective and the impact of social science research projects on people’s lives is not over-estimated. Haggerty in fact specifically warns about the current tendency to overrate the potential harm of research, mentioning a certain form of ‘ethics creep’ that invades social science research in the name of ethics.
Did I put research participants at a greater risk than otherwise in their daily lives? I doubt it. It would be largely overestimating the impact of this research. Notwithstanding a few exceptions, I genuinely believe the bulk of respondents were smart enough to decide for themselves whether they wanted to speak with me and what information they wanted to share. They also usually had a much more developed sense of risk than I did. I was indeed a bit nervous at check-points when I had on me transcripts that could prove the implication of a youth with an armed group ‘from the other side’. But how likely was it that my bag would be searched? And if it was, how likely was it that respondents would be traced back or blacklisted on the basis of first names and villages of origin? Would a pile of paper raise a soldier’s attention? There was surely a greater chance of my CFA francs being ripped off. If particular biographical details make my respondents recognizable – and I know an obstinate reader would be able to trace them back should he or she be patient enough to crosscheck all interview fragments in this book – it was highly unlikely in the field that a person would take time to do that with the aim of confronting someone.
But since a method I used was to ask young people to describe the distressing events they experienced during the war (thereby obliging them to recall disturb-ing memories as they were presenting themselves as either victim or perpetrator of violence), I had to minimize the psychological harm that might derive from such a recollection. It is always tricky to put yourself in the place of someone else and to evaluate the psychological damage certain questions can trigger. Different individuals have different reactions to the same experiences, and events that appear to me (and others) to be terrible do not necessarily affect the people who experienced them the same way. While doing fieldwork, I finally opted for inquiring about harrowing events without pushing my respondents if they did not want to elaborate much. If they wanted to talk and to describe what happened to them, I took the stand that they were willingly doing so and that they were conscious of the consequences; it was not unethical of me to listen. Some res-pondents mentioned having experienced nightmares for a while after having fought in the front line, others felt that they had to take some distance before returning to their former environment, to make a ‘fresh’ start. With regard to the ‘do no harm’ imperative, Haggerty (ibid) makes a nice parallel with social scientists and journalists, basically arguing that while both conduct interviews, videotape people and undertake some forms of participant observation, jour-nalists are much less bounded by ethical protocols and it is expected that a story be told, unless the interviewee clearly mentioned that certain information has to be off record. The assumption is also that respondents get quoted, unless special reservations are made. In academia, the initial assumption is the opposite: research participants remain anonymous unless they provide explicit permission to be identified, and the content of information (the nature of what to write) is heavily weighed, especially when some information appears sensitive and subject to an interpretation that might endanger the respondent or distort an ongoing process. When I mentioned that I would not use real names in the book, many respondents told me that they would not have any objections if their real name was used. For the sake of precaution, I eventually adopted the scholarly attitude and opted for standard anonymity.
Table of contents :
1 INTRODUCING THE STUDY
Defining militarized youths
Structure of the book
3 SOME THEORETICAL CONSIDERATIONS
The paradox of post-conflict interventions
Some theoretical reflections on war and mobilization processes
4 A CONDUCIVE HISTORICAL AND POLITICAL TERRAIN TO THE MILITARIZATION OF CIVILIANS
From cosmopolitan politics to a politics of ethnic polarization
From the Coup to the war
2000-2010: a decade of FPI rule
The place of humanitarianism in Côte d’Ivoire
5 THE IMMEDIATE CONTEXT
The general atmosphere: Chronology of violent events in the west (2002-2007)
Multi-ethnic agrarian societies
The fieldwork locations and the western humanitarian context
6 ARMED FACTIONS OPERATING IN THE WEST
7 MILITARIZED CIVILIANS: DIVERSITY OF TRAJECTORIES, DIVERSITY OF MOTIVES FOR ENLISTMENT
Who ‘took up’ arms in the west?
The specificities of the youngest recruits
What drove young civilians to military life?
8 BLURRED SPACES:MILITARIZED YOUTHS AND THEIR RELATIONSHIPS WITH THEIR IMMEDIATE ENVIRONMENT
Social relationships within the armed groups
The military-civilian nexus
The humanitarian-military nexus
9 RETURN TO CIVILIAN LIFE FOR MILITARIZED POPULATIONS: TWO STANDARD HUMANITARIAN INSTRUMENTS UNDER THE LENS
The global approach to reinsertion and reintegration
DDR in Côte d’Ivoire
Reinsertion under the lens: how were cash allowances spent?
Assessment of a pilot initiative fostering economic reinsertion
The specificities of the youngest recruits