What Happened to Children Who Returned from the Lord’s Resistance Army in Uganda?

T IM ALLEN Firoz Lalji Centre for Africa and Department of International Development, London School of Economics, London, WC2A 2AE, UK t.allen@lse.ac.uk JACKL INE AT INGO Firoz Lalji Centre for Africa, PO Box 1395, Gulu, Uganda DOROTHY AT IM Firoz Lalji Centre for Africa, PO Box 1395, Gulu, Uganda JAMES OC ITT I Firoz Lalji Centre for Africa, PO Box 1395, Gulu, Uganda CHARLOTTE BROWN Firoz Lalji Centre for Africa and Department of International Development, London School of Economics, London, WC2A 2AE, UK COSTANZA TORRE Firoz Lalji Centre for Africa and Department of International Development, London School of Economics, London, WC2A 2AE, UK CR IST IN A . FERGUS Firoz Lalji Centre for Africa and Department of International Development, London School of Economics, London, WC2A 2AE, UK MEL ISSA PARKER Department of Global Health and Development, London School of Hygiene and Tropical Medicine, London, WC1H 9SH, UK


Introduction
During the 20 years since the publication of Graca Machel's influential report to the UN General Assembly (Machel 1996), the plight of children in contemporary war has persistently been highlighted as a matter of grave concern in international meetings. Machel's report led to the establishing of a UN 'Special Representative on the Impact of Armed Conflict on Children', as well as requests for regular updates on the situation of children affected by armed conflict to be presented to the UN General Assembly and Human Rights Council. A plethora of studies have followed, with countless briefings from human rights groups and aid agencies, as well as books and articles. The latter have described in detail the suffering of children, how they have been co-opted into fighting forces and how some have subsequently been reintegrated (e.g. Boyden and de Berry 2004;Honwana 2005;Singer 2006;Vigh 2006;Wessels 2006;Coultrer 2009;Dupuy and Peters 2010;Drumbl 2012;Rosen 2012). There have also been new international agreements on best-practice responses, most notably the Cape Town Principles and Best Practices: On the Prevention of Recruitment of Children into the Armed Forces and Demobilization and Social Reintegration of Child Soldiers in Africa (UNICEF 1997), the Inter-agency Guiding Principles on Separated and Unaccompanied Children (UNICEF 2004) and the Paris Principles and Guidelines on Children Associated with Armed Forces or Armed Groups (UNICEF 2007), as well as legal developments associated with international criminal courts and tribunals. However, despite all the concern, and despite claims about the devastating traumatic consequences of violent experiences, there have been surprisingly few studies on the long-term effects on children who have been recruited into fighting groups (exceptions include Boothby (2006)  This article focuses on the experiences of children recruited by the Lord's Resistance Army (LRA) who were mostly returned to their relatives between 1997 and 2006. They all passed through a reception centre that was run by Gulu Support the Children Organisation (GUSCO). Fieldwork was carried out between 2012 and 2016, with follow-up interviews and discussions in northern Uganda continuing into 2019. Previous research in the region by T.A., M.P. and J.A. on recently returned children in 2004-06 is also drawn upon (Allen 2005;Allen and Schomerus 2006), as well as insights from D.B. and J.O., who worked at reception centres during that time.
The findings raise doubts about the implementation of best-practice guidelines in circumstances of disarmament, demobilization and reintegration of young recruits to armed forces. Building on the Cape Town Principles, the Inter-Agency Principles state the following: Primary responsibility for ensuring children's survival and well-being lies with parents, family and community. The national and local authorities are responsible for ensuring that children's rights are respected. Efforts must be made in an emergency to protect family unity and avoid child-family separation. The principle of family unity-or integrity of the family-states that all children have a right to a family, and families have a right to care for their children. Unaccompanied and separated children must be provided with services aimed at reuniting them with their parents or primary legal or customary caregivers as quickly as possible (UNICEF 2004: 16).
In northern Uganda, that meant placing children with potentially hostile relatives in acutely deprived conditions, thereby exposing them to what might be termed 'social torture' (Dolan 2009). Even at the time, the policy was challenged by some (including by T.A. in meetings occurring in the region in 2004-05). An alternative might have been to keep returned children in Gulu and other larger towns. Some aid agencies, notably those supporting the so-called 'night commuters' (i.e. children who walked into the main towns at dusk and slept there to avoid abduction), were lobbying for such a response. However, humanitarian organizations providing food aid had been involved in setting up the camp system back in the 1990s, partly to avoid the population becoming concentrated near the big towns (Allen 2015;Branch 2011). Along similar lines, concerns were expressed in coordination meetings about expanding facilities for children and young adults in Gulu having the effect of attracting ever larger influxes, including families living in the camps wanting opportunities for their children that could not be met. As a consequence, education and skills training in protected urban spaces was an option only offered to a minority-mostly those who had remained with the LRA for a long period, who had obtained a command position or whose relatives refused to accept back their child. Paradoxically, it was often the most vulnerable who were taken to internal-displacement camps and left there without any adequate follow-up. More generally, evidence presented below raises issues about legacies of humanitarian assistance-a topic that has, until recently, not been the focus of sufficient attention (Fassin 2011;Good et al. 2014;Allen 2015;Branch 2011;Abramowitz and Panter-Brick 2015;Allen et al. 2018).

The Scale of LRA Child Recruitment and the Reception-centre Process
By the time Machel wrote her report in 1996, the LRA conflict in Uganda had been going on for a decade (Allen 1991;Allen and Vlassenroot 2010). Joseph Children who Returned From the LRA in Uganda 3 Kony's LRA began fighting the government of Yoweri Museveni in the late 1980s, abducting thousands of children and young adults, and perpetrating atrocities on Kony's own Acholi people and their neighbours. The Ugandan army-which was officially named the Uganda People's Defence Force (UPDF) in 1995responded with anti-insurgency operations that were associated with human rights abuses. As the LRA attacks continued, internal-displacement camps supported by humanitarian aid became part of the government's strategies and ultimately a way of controlling the Acholi population (Otunnu 2005;Allen 2006b;Finnstro¨m 2008;Dolan 2009;Branch 2011).
Those found outside the camps were in danger of being accused of supporting the rebels by the UPDF, but the camps were not well protected and the LRA found it relatively easy to continue to abduct or solicit new recruits. By 2003, when Jan Egeland, the UN Under-Secretary General for Humanitarian Affairs visited, around 1.5 million people (about 80 per cent of the population) were living in hundreds of camps, often in appalling conditions. Following incursions by the UPDF into southern Sudan (now South Sudan), where the LRA's main bases were located, the LRA were making repeated attacks, forcibly recruiting ever more children. Egeland described the situation as a 'moral outrage' and one of the worst humanitarian crises in the world.
In the following year, further shocking atrocities were perpetrated by the LRA and the situation was referred to the International Criminal Court, with warrants being issued for senior LRA commanders in 2005. Then, in 2006, the LRA were drawn into peace talks. When these talks were deemed to have failed in 2008, fighting resumed, but not in Uganda (Allen and Vlassenroot 2010). The LRA remains active at the time of writing in the borderlands of South Sudan, the Democratic Republic of Congo and the Central African Republic. Meanwhile, security has been maintained in northern Uganda for over a decade. An amnesty was agreed for those who had been recruited by the LRA, with the exception of a few senior commanders, and some of the agencies supporting these processes view the region as a place where demobilization and reintegration have been relatively successful (World Bank 2013;Sharif 2018: 7).
During the fighting in Uganda, considerable efforts were made to keep track of the numbers of those being recruited by the LRA. Estimates vary, but an effort to compile all available data suggested that, by 2006, over 24,000 children and over 28,000 adults had been recruited. It was also speculated that the total number might possibly be as high as 75,000 (Pham et al. 2007: 22). Whatever the precise number, these estimates recognize that probably more than half of those joining the LRA were actually adults. The LRA war was never a campaign waged predominantly by 'innocent' children, as many commentators suggested at the time (e.g. Government of United Kingdom 2004). Nevertheless, it is certainly the case that thousands were children when they were recruited and a large number were still children when they returned.
Initially, the return of children from the LRA was managed by the UPDF. The names of those who returned were announced on the radio and sometimes they were paraded in towns. Relatives were asked to collect them, but they often failed to come forward. A group of parents thus came up with the idea of reception centres to help facilitate their return in a caring and appropriate way. GUSCO in Gulu town was the first of these centres to be established. It was founded in 1994 and secured funding from the Danish International Development Agency (DANIDA) in 1997, with Save the Children subsequently playing an important role in its development. Located nearby in Gulu, the World Vision reception centre was established a year after GUSCO. The other 10 reception centres were established between 1997 and 2004. Four of the larger ones were based in towns where the surrounding populations are also Acholi. Six others operated in neighbouring districts, where Teso and Langi populations were also affected by the LRA conflict. With international aid-funding, centres became walled compounds and were run by salaried staff. They received back what were referred to as 'formerly abducted people' from the UPDF, including both children and young adults. An unknown number of the latter were persuaded to join the UPDF and an entire UPDF battalion (the 105 th ) was made up of former LRA recruits (Allen and Schomerus 2006). All reception centres tried to provide a safe and secure space, enabling those who had spent time with the LRA to be debriefed and to receive healthcare, food, shelter and psycho-social assistance. In accordance with international agreements, it was anticipated that they would play a major role in facilitating a process whereby those returning from the LRA could be reconnected with their families and returned to their previous lives in internal-displacement camps.
Reception-centre staff were encouraged to see the returning children and young adults as innocent victims, who had been abducted and forced to commit violent and murderous acts against their will (Allen 2005; Allen and Schomerus 2006; Akello et al. 2006Akello et al. , 2009. The term 'ex-combatants' was generally considered inappropriate, even when a person was known to have been an LRA fighter. This was linked to a promoted perception that most of those recruited by the LRA were children, or were like children. While that may have been so in many cases, the persistent emphasis on abduction, and the corresponding use of the term 'captivity' for periods spent with the LRA, conveyed a lack of agency. In the Acholi language, the term mostly used is mako, which might best translate as 'taken'. The term mako tek tek means those taken forcefully, including for sexual purposes (i.e. rape). Meanwhile, the word odonyo is applied to those who willingly joined the LRA. In this article, the term 'abducted' is retained for recruited children, but it needs to be recognized that surviving with the LRA necessitated choices and a considerable degree of engagement with LRA activities. In-depth interviews with those who returned make it clear that passivity was rarely an option. Reception-centre staff were aware of this. Nevertheless, donors had provided funds in part because children and young adults returning from the LRA were all perceived as innocent and traumatized, and reception-centre staff were encouraged to do everything possible to alleviate any self-blame that those returning might feel. The message repeatedly conveyed was that they were not in any way culpable for the awful deeds they had committed or witnessed. An atmosphere of 'forgiveness' (timo kica) was promoted, with the intention of establishing codes of behaviour enabling them to slot back into life with their families.
To this end, at GUSCO, support was provided to those wanting to go through local healing rites. There was much emphasis in 2004 and 2005 on a particular reconciliation ritual called mato oput, which was associated in certain locations with social rehabilitation following a murder. However, the rites actually offered to returning children were more like those to welcome someone who had returned home after travelling. 'Group therapy' and sometimes individual 'counselling' were also offered with a social worker. At the World Vision centre, local healing rituals were discouraged in favour of prayers, but otherwise approaches were similar. Those passing through the centres were also encouraged to externalize their experiences. They were expected to talk about violent acts and recognize them as 'bad' (marac), as a way of demonstrating that they were recovering. Even at the time, there was some scepticism that such a strategy would work. In observations and interviews at GUSCO in 2005, it was striking that interviews with children away from staff revealed that they did not always agree with the outlooks being promoted. A few openly stated that they found killing ber (good/nice) and observations (by T.A., M.P. and J.A.) indicated that they were well aware of what sorts of stories were liked by the staff and what needed to be said for permission to leave the reception centre to be granted.
Within the reception-centre paradigm, leaving usually meant being reunited with families as soon as possible. Those who had command ranks in the LRA, or the status of senior wives, were sometimes settled in urban locations when it became apparent that living among relatives might be difficult. However, the preferred policy for the majority was take returned LRA recruits back 'home'; and this typically involved transporting them to the very camps from which they had been abducted. In 2004-05, this meant taking them back in the middle of the day and then remaining with them in the internal displacement camp (IDP) camp for a brief period, before rushing back to the security of the town before dark. The situation these children were returned to was far from ideal. Conditions in the displacement camps were appalling. Crude mortality rates were high and outbreaks of cholera and other infections common. This was due to the poor or non-existent water and sanitation facilities in the camps (MSF 2004;WHO 2005). With so many huts crowded together, fire was a constant hazard too.
Until 2002, when the UPDF began operating across the border in Sudan, the number of people returning from the LRA through the UPDF to reception centres was low. From reception-centre data, it was less than 10 per year until 1997 and then rose to around 1,000 annually. After 2002, the numbers increased dramatically to about 6,000 then dropped to a few hundred in 2005, and subsequently to fewer than 50 per year. The total number of those who returned through reception centres is hard to assess precisely. The quality of records varied between centres, there were instances of double counting and some records have been either destroyed or lost. The disappearance of 11,000 World Vision reception-centre records since 2006 is the most striking example of the latter. Efforts to find these World Vision records in 2012 proved fruitless and it was concluded that they had been destroyed.
Assessment of the available data back in 2005 and 2006 (when the World Vision records were still available) estimated that reception centres received back around 25,000 people, including many young adults (Allen and Schomerus 2006;Pham et al. 2007). Twenty-four per cent were female and 76 per cent were male, with the average age of those returning being 15 years for females and 20 years for males. At arrival at a reception centre, almost 66 per cent were registered as being under the age of 18 years old. Just over 34 per cent were over the age of 18 years, but many of them would have been recruited as children. The records on each individual reporting at the centres noted the amount of time spent with the LRA. These data indicates that 20 per cent of people being reintegrated through reception centres had been with the LRA for over a year, around 50 per cent for less than three months and around 15 per cent for less than a week. On average, females were with the LRA for more than twice the time spent by males.

Assessing the Reintegration of Children
There was considerable interest in the region at the time that most children and young adults were returning through reception centres-largely triggered by the visit to the area by the UN Under-Secretary General for Humanitarian Affairs in 2003, the referral of the situation to the International Criminal Court in 2004 and the enormously successful publicity campaign launched by Invisible Children the same year (Invisible Children 2004). The proliferation of aid agencies led to an increased number of reports aimed primarily at raising humanitarian funding. A good example was World Vision's 2004 publication, Pawns of Politics: Children, Conflict and Peace in Northern Uganda, which made inaccurate claims about the LRA and the purported prevalence of HIV/AIDS (World Vision 2004;Allen 2006a). Other aid-agency publications were more evidenced-based and thoughtful, but mostly still promoted normative assessments, linked to agendas espoused by non-governmental organizations and civil-society activism. There was, in particular, much enthusiasm for promoting customary rituals as a means of reintegration (e.g. Lomo and Hovil 2004) and there was a great deal of emphasis on amnesty, less in the formal sense and more as a general 'forgiving' of those returning from the LRA who had perpetrated violence. In retrospect, there was a remarkable and misleading expectation that families would welcome back their loved ones and that formerly abducted children would be easily assimilated.
Meanwhile, additional funding and logistical support became available for more rigorous investigations and it was possible to build on earlier research about the Acholi people and the LRA, including detailed studies from the late 1990s and early 2000s. Those studies included the then-unpublished, but electronically available, doctoral work of Finnstro¨m and Dolan (Finnstro¨m 2008;Dolan 2009). Also, in 2004 and 2005, Save the Children, UNICEF and United States Agency for International Development commissioned independent assessments of the experiences of those returning from the LRA (Allen 2005; Allen and Schomerus 2006). The fact that fieldwork occurred before security was established inevitably affected the way in which data was collected. It was necessary to have a military escort to reach some IDP camps and visits outside the big towns tended to be short, although effort was made to spend longer periods in some camps and even Children who Returned From the LRA in Uganda 7 remain overnight, which could be potentially dangerous (on two occasions, for example, T.A. was present at IDP camps during LRA attacks in 2004 and. In general, it was necessary to return to safe locations by nightfall.
In addition to ethnographic work based on observations at reception centres, in urban locations and in the camps, the research adopted a systematic approach to finding and interviewing young people who had returned from the LRA, passed through the formal reception procedures and been reunited with their families. To this end, in 2005, a 1:50 sample was generated from reception-centre records. Reception-centre staff expected that it would be possible to find them but, after five months of considerable effort, a total of only 415 were traced out of 886. That was partly a reflection of the short periods that it was possible to spend in camps for security reasons, but many of those who had been returned appeared to have disappeared and left no trace. Of those found, 238 were interviewed in person. In other instances, information was provided by a relative or neighbour.
The work highlighted some important issues. In those circumstances in which the child or young adult was living with their close relatives, there could be relief and delight that he or she had survived. Often this appeared to be a genuine emotional reaction, but it was also partly linked to the fact that they were sent home with money, a mattress and various household items. Within a few weeks or months (when the money had run out), problems tended to surface. Indeed, those interviewed felt that there were both subtle and explicit ways in which it became apparent that family members feared and rejected them. Neighbours were additionally prone to abusing or shunning them, despite attempts to make local councils monitor such things. There were complaints about name-calling, including references to where a returned child had spent time and received support (several said they were sometimes addressed aggressively as 'You! GUSCO child!' and were told to do extra labouring work). More generally, concerns were expressed about the less overt accusations made by neighbours. Almost all reported examples of being stigmatized, making reference to, for example, kwero (rejection) and cimu tok (pointing at the back of your head). Some felt oppressed by what they knew their neighbours thought about them and many of those who had killed or witnessed terrible events with the LRA continued to be disturbed by their memories. This was sometimes manifested in nightmares, high consumption of alcohol and aggressive behaviour associated with life in the bush (i.e. with the LRA). In the Acholi language, the LRA were sometimes referred to as olum (bush/ forest) and the English expression 'bush mentalities' was commonly heard.
There were also concerns raised about cen, a kind of malevolent spiritual force or vengeful ghost. Cen possess or emanates from those that have witnessed or perpetrated violence, or been in physical contact with a dead body, and is perceived and experienced as socially polluting and potentially dangerous. Other kinds of metaphysical entities were mentioned too, including people being seized by jogi and tumi (terms for usually personified ghosts), and Pentecostal possessions were widespread. Like cen, such things were not exclusively viewed as something associated with those returning from the LRA. However, fear of cen was often mentioned specifically in connection with the LRA, because LRA violence was viewed as extreme, and those who had spent time in the bush could not have escaped it. Cen was viewed as something that those returning, including children, might bring with them to the home without wanting to do so. Matters were complicated too in some circumstances by the fact that there were those who openly stated that they had enjoyed their time with the LRA, and even that they felt appreciated and respected when they killed people as combatants.
Overall, it was found that there was considerable diversity in experiences. Some parents stated that they were happy to be reunited and were adamant that their children had been taken by force. In these instances, local rituals were sometimes performed to cleanse them and, more often than not, there was an emphasis on Christian notions of community. However, the numbers participating in such a process were surprisingly small. Just 69 people in the sample (less than 30 per cent of those interviewed) stated that they had been involved in any customary cleansing rites, let alone mato oput. It was also apparent that, even where families seemed to be welcoming back their loved ones and purportedly viewed them as blameless victims who had suffered, there were usually more ambivalent views too. It was known that those who spent time with the LRA were likely to have survived by being willing to witness or to perform atrocious acts. There were certainly those whose detailed accounts of life with the LRA were narrated as if they were entirely passive; and there were extraordinary monologues that would go on for hours, delivered in a monotone, describing what had happened on a day-to-day basis, as if they were recounting an endless dream. But there were others who returned with an acute sense of their own agency. They were open about having made choices in order to survive, and even about having embraced the idea of living in a new moral space.
Those who were located were mostly living in horrible circumstances with their relatives. They may have been struggling and suffering, but so was everyone else. In such circumstances, it was hard to assess the degree to which reintegration had been successful or whether, in general, the experiences of those returning from the LRA were any worse or better than for the majority of the population. Yet, there were a few who stood out. Often, they were those who had obtained a military rank within the LRA or had been a favourite wife of an LRA commander. Some of these people took on leadership roles, often among groups of those who settled in the Gulu urban area, but occasionally in other situations too, such as informally policing life in the larger camps that had turned into sizeable towns (like Pabbo) or even within the UPDF (where their former LRA rank was not officially recognized but was applied informally in the 105 th Battalion).

Long-term Experiences of Returned Children
In addition to the work summarized above, other research occurred during or soon after that early period of large-scale return from the LRA too. Some of it also focused on the experiences of returned children and young adults, although it did not attempt a comparable follow-up of those returning from a random sample of the reception-centre registers. For the most part, the findings reiterated and elaborated the same points, often in insightful ways (e.g. Hovil and Quinn 2005; Children who Returned From the LRA in Uganda 9 Akello et al. 2006;Amone-P'Olak 2007;Borzello 2007;Pham et al. 2007), and there have been numerous subsequent publications addressing similar themes (e.g. Mazurana et al. 2008;Blattman and Annan 2010;Mergelsberg 2010;Akello 2013;Baines 2016;Atim et al. 2018;Dubal 2018). Although the generalizability of the finding is often unclear, the various detailed studies present a more mixed picture of reintegration than, for example, the World Bank (2013) assessment cited above.
All the local-level studies emphasize difficulties faced by those returning, with some authors emphasizing their capacities and resilience, and others focusing more on their challenges, social marginalization and prospects for remobilization. It is additionally fair to say that the more recent literature has become quite repetitive. Indeed, the proliferation of books, articles and human rights reports has prompted one eminent scholar of the region to ask whether they add anything to what is already known (Finnstro¨m 2018). Research fatigue has certainly become an issue, notably in and around Gulu, where interviewees occasionally express exasperation with being repeatedly asked the same questions, without any discernible benefits (Clark 2008;Schiltz and Bu¨scher 2018). Part of the problem is associated with the research methods used. Most samples of former LRA recruits are based on snowballing, or on surveys of those who self-report abduction, or on lists provided by community representatives, or on ad hoc interviews and observations. As a consequence, researchers find themselves being directed to the same individuals, who tend to provide well-rehearsed responses.
The new findings presented here differ in that they offer a long-term assessment of the experiences of returnees from the LRA who passed through GUSCO, based on a random sample of those registered in the surviving original reception-centre records. This is similar to the approach taken by T.A.'s team in 2005, but uses a smaller sample from one reception centre and more detailed interviewing. Neither GUSCO nor similar reception-centre records have been used by other researchers in this way since 2005. One consequence of the method adopted is that the majority of those in the GUSCO sample have had no previous contact with researchers, child-protection officers, social-welfare personnel or aid-agency staff since being reunited with their relatives more than 10 years ago.
The sample was taken from the surviving original GUSCO case files in 2013. The case files of 3,040 individuals registered at the centre had survived. They were found in small piles in different parts of the reception centre, with some half-eaten by termites. Based on previous work at the centre in 2005, it was estimated that about 100 case-file records were missing or were damaged beyond repair (Allen and Schomerus 2006). To select the sample, the records were arranged by the year in which the individual had arrived at GUSCO. A 10 per cent random sample was taken from each year between 1997 and 2012. In total, 304 GUSCO case files were selected. Intensive efforts were made to find, and interview, the individuals named in these records; 103 were females and 201 were males. All but two were children when they arrived at GUSCO. The two that were above 18 years old were males returning in 1997 and 1999, before a reception centre for returning adults had been established (these two adults are included in the data presented below).
As with the research undertaken in 2005, it was far from straightforward to trace those in the GUSCO sample. After almost two years of searching, 75 per cent of the sample were interviewed. Of these 230 individuals, 21 were infants or young children who were born in the bush, ranging in age from 2 weeks to 5 years. No information could be found about those whom we were unable to trace. The most difficult to trace were those returning between 2002 and 2005, when the numbers arriving were highest (181 people returned in these years-63 females and 118 males).
To supplement and contextualize the findings, the team additionally attempted systematic comparison with control cohorts. The first of these was intended to be made up of people who had returned from the LRA without passing through a reception centre, but it proved hard to find such a group. Some people were located who had returned as adults and participated in the amnesty procedures, and there were also a few who claimed to have returned as children before the reception centres were established. However, those who had spent time with the LRA and had returned as children since the late 1990s had almost invariably passed through a reception centre-although there were some who explained that they had intermittent contact with the LRA as children. More successful was comparison with a second control cohort that compared the lives of 10 males and 10 females in the GUSCO sample to individuals who had comparative experiences up to the point of abduction to assess differences in their lives.
All interviews were coded and analysed using MAXQDA. Analysis of the data reveals some striking patterns. One of these concerns access and residence on ancestral land-namely land that is collectively farmed and occupied by an extended patrilineal family (dogola or kaka). Among both males and females who had been recruited by the LRA, the probability that they are currently able to access ancestral land decreases in relation to the length of time they spent with the LRA. Almost all of those who spent less than four years away have access to some family land and they are the majority in the GUSCO sample. Among those who state that they do not have access to any ancestral land (about 25 per cent) are those that were never resettled with their families in IDP camps after leaving the reception centre, because of their relatively senior position in the LRA. In part due to fears about their safety, they were assisted to remain in the bigger towns. Of those cultivating family land, some are living on their father's patrilineal land. Others are allocated plots through their mother's father's clan, or their grandmother's clan, the clan of their mother's current husband or sometimes they are allowed to cultivate family land somewhere by a friend. In these cases, their rights to settle and use the land are based on the changeable attitudes of neighbours. For returned girls, access to land is contingent on their current husband's relatives or their father's and brother's relatives allowing access. Most are given small plots somewhere but depend on goodwill. There are instances in which women were forced to move away before they could harvest their crops (which were then taken by co-wives and relatives) and almost invariably interviews describe movement from one place to another, often with children left with their grandmothers.
In general, those in the GUSCO sample who were found to be living on ancestral land reported ongoing and often acute difficulties. This is particularly the case where the period of time spent with the LRA was relatively short. This group are the most likely to report problems with name-calling, rejection by neighbours and relatives, and cen. Analysis of the interview data shows that the probability of continuing to experience these kinds of afflictions is significantly associated with the number of months spent with the LRA. The less time spent with the LRA, the more likely individuals are to report that they have experienced stigma and cen. This finding partly reflects the fact that length of time with the LRA relates to access to ancestral land. Those who spent shorter periods with the LRA are more likely to be living with patrilineal relatives. It is clear from in-depth interviews that those who are most likely to abuse or reject the returned children-most of whom are now in their late twenties or thirties-are their own family members. In the case of the women, it is the patrilineal relatives of their current husbands or their husband's co-wives. Another factor, which is hard to quantify, is that those who spent longer with the LRA are likely to be more robust, confident and assertive, otherwise they are unlikely to have survived. Problems with being stigmatized are reported by 46 per cent of women and 59 per cent of men. Very often, the antipathies are associated with conflicts over land, but there is also still a genuine worry in rural homes about hosting someone who had spent time with the LRA.
If anything, the threat has increased now that they are adults. It is anticipated that they harm others, due to cen, and that they might become violent at any moment. Instances of such violence happening in practice are probably infrequent, but examples were recorded in interviews. One tragic case involved a returned woman who beat her own child to death when a co-wife complained that the child, who had been brought back from the LRA, had stolen something. Importantly, many of the people with whom the returnees are now living openly expressed fear that cen would pass to them. It was also clear that they were frightened by the idea that cen could pass from one generation to another, including from a mother to her new-born infant. Accepting sons, daughters and grandchildren into the family home thus opened up the possibility of cen becoming a pervasive threat. From the GUSCO sample, 27 per cent of males and 30 per cent of females complained of having experienced or having been accused of cen at some point since their return. These were the people in the sample who were most likely to have performed various customary cleansing rituals, although not the mato oput rites promoted by activists at the time of return (selected cases of cen and spiritual pollution drawn from the GUSCO follow-up interviews are presented in Victor and Porter (2017)). Others said that they had been accused of having cen, even though they did not believe they had been afflicted. As one young woman noted: My family say I should not stay with them because I have cen. They do not want to offer me land for cultivation . . . that is why I am renting in town.
Overall, 15 per cent of females and 12 per cent of males are renting plots near towns and trading centres, and 3 per cent of females and 10 per cent of males were 12 Tim Allen et al. not cultivating at all at the time of interview. In some respects, the 7 per cent of the GUSCO sample who are living in urban settings and the 14 per cent living in trading centres and peri-urban locations can be assessed as generally managing a bit better than many of those who have been allocated ancestral land by relatives and are now living with them. Although those interviewed who are based away from rural locations may speak at length-and often with strong emotionsabout problems in their lives, particularly related to income and accessing education, they are much more likely to be aware of support networks aimed at LRA returnees, such as the Gulu-based, Watye-Ki-Gen (WKG) and the Women's Advocacy Network (WAN), and to highlight particular assistance schemes that they have (or have not) managed to access. It is also common for them to have joined socially supportive Pentecostal churches and to have become 'born again'. In contrast to those living in rural locations, they frequently mention that they had been interviewed by researchers before, often facilitated by WKG or WAN. Several additionally note that they had command responsibilities in the LRA and their support networks tend to replicate LRA hierarchies in that they are led by those who had higher ranks or were the wives of senior LRA officers (Ocitti et al. 2019). This is actually a point of controversy amongst them, with most of those interviewed in towns making accusations about how funding for those retuning from the LRA has been co-opted by a fortunate few-sometimes referred to the 'creamy ones'. Although they express sadness or bitterness about being ostracized by relatives and may face a range of daunting problems, they avoid their families' censorious attitudes and are finding ways to eke out a livelihood. Even if they are aware of urban neighbours saying negative things about them, they tend to be dismissive of such attitudes. It is significant that, from analysis of the GUSCO-sample interviews, the longer someone had spent with the LRA, the more likely they are now not to be living on ancestral land, and the less likely it is that they report experiences of cen or ongoing concerns about being stigmatized.
Within the GUSCO sample, there were also a particularly vulnerable group, made up of those females who returned with babies and young children. Over a third of the females in the sample fell into this category. The fathers of their children are known to be LRA commanders. Anticipating rejection, these women rarely approach their child's patrilineal relatives for support. One who did so reports that her child was seriously mistreated. Rejection by the mother's patrilineal relatives was more complicated. Around 25 per cent of the mothers found that their children were rejected immediately. Others were ostensibly welcomed at first, but gradually faced more hostility. For example, one young woman explained: My uncle's wife . . . sent their son to defile my five year old daughter when I was not at home . . . the boy was caught raping my daughter [and] taken to his mother, but all she said was that my daughter is a . . . child of the bush . . . Formerly abducted people do not deserve justice, they are walking corpses.
Such extreme perceptions were found to be widespread with respect to the 21 children in the GUSCO sample. Born in the bush, they are now mainly in their Children who Returned From the LRA in Uganda 13 mid-to late teens. A further research project addressing this matter is ongoing, using a much larger sample of this subgroup (Atim and Parker 2019;Atingo and Parker 2018). Preliminary findings suggest that they are living in acutely deprived and even dangerous circumstances. Even those few whose mothers have managed to access school scholarships for them (from WKG, WAN and other donors) report persistent abuse at boarding schools in neighbouring districts.

Reflections and Comparisons
Looking back, reception centres have been able to do little to mitigate the challenges facing returned children from the LRA over the long term. Insufficient and irregular funding, limited resources and unrealistic expectations were all parts of the problem, but there were other issues too. GUSCO, in common with other reception centres, succumbed to dualistic and simplistic framing of the issues. Terms such as victim/perpetrator, innocent/guilty, passive/active and good/evil were widely used. Contemplating any kind of responsibility for participating in violent or murderous acts was deemed inappropriate. As one of those in the GUSCO follow-up sample explained: They said we should forget everything that happened in the bush ... [and] we should start a new life . . .. If someone hurts you, you should not be aggressive back . . .. If you want something, you should not take it by force or steal . . .. You should have a forgiving heart. If someone does something wrong, then you should not react with anger. That is all I can remember.
A staff member, reflecting back on the oversimplified counselling pointed out that it was partly to enable the children to fit in and not to make their family scared. Some parents were afraid to take their children back home because they thought 'this child . . . will cut me into pieces'. Importantly, the findings from the GUSCO sample reveal that fears about reintegrating the children have persisted and that is most evident in rural locations, where most of them are currently living. In general, return from the LRA has been most successful where reintegration into rural life has failed or not been attempted. The findings also highlight a limitation with research that has been based on samples linked to urban-based support networks, as well as other snowballing and ad hoc methods. Such methods have meant that the majority of those who returned from the LRA and passed through reception centres have been largely ignored, and that what have been presented as broadly representative insights are actually about particular groups. This helps explain the different conclusions between, for example, Atim et al. (2018) and Dubal 2018. The former elaborates on the struggles of a selected sample who alleged that they had experienced war crimes, while the latter suggests that such research is misleading, and even based on recording lies. Dubal's chosen group are rather nostalgic about the LRA and impressively engaged in making the best of their lives. Both perspectives may be partly correct, but neither is representative of the whole.
14 Tim Allen et al.
It is also interesting to note how the findings from the GUSCO follow-up relate to what has become perhaps the most influential study of reintegration in northern Uganda, with apparent lessons in comparable situations elsewhere. This is the work of Blattman and Annan, which is based on research they carried out in 2005 and 2006 for their doctoral theses (Blattman and Lundberg 2007;Mazurana et al. 2008;Blattman 2009;Blattman and Annan 2010;Annan et al. 2011). The influence of this work is partly related to their methods of analysis, particularly Blattman's use of micro-economic techniques in articles published in leading political science journals. Blattman and Annan surveyed 1,016 households and 741 male youth (defined as 14-30 years of age) who were reported to have been living in those households in 1996. A third of this latter group reported having been taken by the LRA for at least one day. They also found that a sixth of all female youth in their sample households had been taken too. More than half of those who stated that they had been with the LRA claimed to have returned home informally, without passing through a reception centre. Although they noted high levels of what they categorized as post-traumatic stress disorder, their research suggests that, while many of the young people they interviewed initially had difficulties following their return from life with the LRA, they quickly overcame them. Young men returning were found to be twice as likely to be 'minor community leaders' than young men who had not been abducted (Blattman and Lundberg 2007;Blattman 2009). Meanwhile, the majority of female respondents reported positive relationships with their families and the wider community (Annan et al. 2011). The findings have been taken to suggest that the reintegration of the LRA was relatively successful, and also that having been involved in a war or rebel movement provided new opportunities for those willing and able to seize them. There are, however, caveats that need to be considered in assessing this body of work.
First, Blattman and Annan's were collecting data in 2005 and 2006 (like the study led by T.A. at that time: Allen and Schomerus 2006). This was a period of considerable insecurity in the region, with more than a million people living in densely populated internal-displacement camps and towns. These living arrangements may have facilitated networks of support, especially as there were several aid-funded projects promoting such networks. However, the subsequent dispersal of people from the camps and towns may have left many ex-combatants feeling rather isolated and vulnerable. Second, it is possible that those people who avoided the reception-centre arrangements-but were picked up in Blattman and Annan's surveys-actually benefitted in the long term. They may have found it easier to blend in by avoiding official registration. A third possibility is that Blattman and Annan's findings have been affected by characteristics of their sample. They focused on youth, rather than children. Many young adults living in the displacement camps had encounters with the LRA, sometimes helping them move around food and equipment, but they were not forcibly taken from their homes. It not clear from Blattman and Annan's publications how many people in their sample had really been recruited. They recorded what people said, but they did not cross-check with reception-centre registers (or amnesty commission data).

Children who Returned From the LRA in Uganda 15
Their data is all self-reported (or reported by relatives). In 2005-06, it was very common for people to say they had been abducted, because it was a way of being able to claim various kinds of assistance from aid projects. Nowadays, such incentives are absent.
The findings from the GUSCO sample cast doubt on Blattman and Annan's conclusions. The majority of males and females are now living with relatives in rural locations where they remain vulnerable and where they are very unlikely to experience life in the way that Blattman and Annan have described. Few of the females in the sample enjoy caring and supportive marital relations and, while there may well be males (and females) who returned from the LRA and have taken on leadership roles, they are rarely reported among those in the GUSCO sample (less than 1 per cent). Specifically, one explained that he had been offered the position of chairperson of his village council (LC1), but he decided that he was too young and he has taken on the role of secretary. Another mentioned that he has a brother, who had also been with the LRA, who was elected to an LC1 post and one male in the sample was elected to an LC1 post in Lira District (i.e. outside of the Acholi subregion). There are additionally a few in the sample who have responsibilities in savings groups, football teams and in their church. It is interesting to note that several reported that they had been told by reception-centre staff not to put themselves forward for leadership positions to avoid: as one put it: 'a lot of nasty things said about them linked to their abduction.' It is, of course, vital to see the long-term experiences of returned children from the GUSCO sample in context. Results from the comparison of those in the GUSCO sample with paired age-mates who had never been recruited show that experiences can be similar (Brown 2018). The majority of children and young adults affected by the LRA were not recruited. They were brought up in internal-displacement camps-living in abject poverty, with the acute fear of abduction, limited opportunities for cultivation and virtually no opportunities for salaried employment. In addition, infectious diseases (such as cholera and malaria) were endemic. When abducted children and young adults were returned home, this was the world they returned to. The security situation in the region improved from 2006 onwards and this enabled people to leave the displacement camps, with most people either returning to their ancestral lands or seeking plots among their neighbours. There have been large numbers of land disputes and there have been acute strains on social life, partly linked to the breakdown of conventional marital relations in the camps (linked to non-payment of bride wealth and the lack of negotiated agreements between lineages, etc.). The scale of unreported rape has been shown to be high. In most cases, solutions or ways of managing social problems are collectively agreed, often by an endeavour to promote what Porter (2017) has called 'social harmony'. In practice, that means that those abused have no means of attaining much in the way of accountability. In terms of reporting leadership positions, those in the sample of age-mates who had not spent time with the LRA found that they had a slightly higher likelihood of reporting leadership positions, but the majority explained that their interrupted education limited their capacities, not least because they cannot read or write.
There is no doubt that the legacies of life in IDP camps have had far-reaching consequences. Nevertheless, those children who returned from life with the LRA and passed through GUSCO are now usually living in worse circumstances than former classmates living nearby who were not abducted. In rural places, they often remain exposed to outright abuse.

Conclusion
Returned children who resettle with relatives in displacement camps, and subsequently live with them on ancestral land, are likely to be feared and rejected in their daily lives. At the time of reintegration, there was much talk about welcoming children home, and of 'forgiveness' (timo kica). There is little indication that that has happened. On the contrary, antipathy to children returning from reception centres has in many instances become worse as they have become adults, with stigma often coming from their own rural relatives and neighbours. This group is also the least likely to have been contacted by other research teams or to have been followed up in any way since they were returned. Those who have escaped this situation are the minority. They are mainly those who spent longer periods of time with the LRA, are not settled with their families and are living in urban locations.
There are many troubling aspects to all of this, including the fact that humanitarian agencies were given a responsibility to protect very vulnerable people and they have not been able to do so. The expression 'social torture' was used by Dolan in his book about life in the internal-displacement camps when the Acholi people were caught between the Ugandan army and the LRA. The term captures the subsequent experiences of the majority of people in the GUSCO sample. Their presence is tolerated in ways that persistently remind them of who they are. They are, as one local councillor put it, 'just half-loved'. Extended families in post-war rural locations promote mutuality, and do so by regulating moral spaces. That process can involve scapegoating individuals who are perceived as socially contaminating. Those returning from the LRA may fall into that category and what happens to them echoes findings from other research suggesting that witchcleansing, communal containment of spiritual pollution and punitive modes of exclusion are linked to social integration in situations of resettlement after displacement (Allen 1997;Allen and Reid 2015;Victor and Porter 2017). After prolonged investigations, 25 per cent of the GUSCO sample could not be located. Those people may have moved away, perhaps to Kampala, and others may have successfully changed their identities, but a concern is that some among them have been ostracized to the point of self-harm or even secretly killed. Death threats were reported by many of those interviewed. At the time of completing final edits to this article in Gulu town in November 2019, a case is being much discussed of a woman who passed through GUSCO. She has been chased from her farm, accused of being a witch (lajok). It has happened before, but she has not been willing to accept it this time. She has brought her case to the Acholi Paramount Chief, and is asking either for a formal land title or to be taken back to the LRA.
Children who Returned From the LRA in Uganda 17 Save the Children, which played an important role in funding and supporting GUSCO, states on its website that: every child deserves a future . . . we give children a healthy start in life, the opportunity to learn and protection from harm. We do whatever it takes for children-every day and in times of crisis-transforming their lives and the future we share.
Inevitably, this is more of an aspiration, contingent on the context, donor priorities and campaigning. Despite all the international interest in the Ugandan situation, and the concentration of aid projects in Gulu, long-term follow-up research has revealed that instead of fostering processes of social inclusion for children and young adults returning from the LRA, humanitarian agencies have, for the most part unwittingly, participated in rendering most of them invisible and potentially more vulnerable. This is evidenced by a lackadaisical approach to record-keeping; a willingness to support the location of children and young people in what were recognized at the time as insecure internal-displacement camps; and disinterest in following up those children and young adults where they had been sent, even when the situation was relatively stable and presented no exceptional threat to staff. The latter was partly a function of the withdrawal of sufficient donor funding as international interest in northern Uganda declined after the LRA's area of operations shifted to South Sudan, the Democratic Republic of Congo and the Central African Republic. More broadly, experiences of those who passed through the GUSCO reception centre make it apparent that assumptions embedded in international agreements such as the Inter-agency Guiding Principles on Separated and Unaccompanied Children are, at least in this instance, open to question. Where children recruited to rebel groups may have survived after being compelled to perpetrate violence on their own relatives, it cannot be assumed that it is always beneficial for them to be returned to their parents and siblings. In northern Uganda, this was often the worst place to leave them.