II. Typical experiences of child soldiers in the LRA

Child soldiering is not a monolithic phenomenon (Shepler, Reference Shepler2005; Wessells, Reference Wessells2006; Drumbl, Reference Drumbl2012; Gates and Reich, Reference Gates and Reich2010).Footnote ⁶ There is no paradigm case, ‘The Child Soldier’, but only the existence of child soldiers, whose lives, experiences and actions differ in important respects that are relevant to understanding how each goes through the transformation from child to soldier (Shepler, Reference Shepler2005). Notwithstanding these differences, research on child soldiering reveals patterns and shared experiences, out of which we may construct ‘typical’ cases. In this section, I describe shared experiences of child soldiers in the LRA based on fieldwork conducted with former LRA child soldiers, with regular reference to research conducted with former RUF child soldiers. The research reveals that child soldiering in these groups is notoriously extreme, both in the physical and psychological demands it imposes on children and the brutality of which child soldiers in the group have been shown to be capable (Wessells, Reference Wessells2006; Bevan, Reference Bevan2007).

To understand why the LRA uses child soldiers, we must first understand the situation in Uganda out of which the LRA developed. In the mid-1980s, Yoweri Museveni (who is now the President of Uganda) led a military coup and took over the government. In 1986, a civilian resistance movement led by Alice Lakwena mobilised against Museveni's leadership. Lakwena's group, the Holy Spirit Movement (HSM), aimed to protect the Acholis, the northern Ugandan ethnic group that felt particularly disempowered by the government. After Lakwena abandoned the movement in 1987, Joseph Kony formed the LRA out of the HSM but, as popular support for revolution in Uganda diminished, so too did volunteers to fight against the government.

Frustrated by diminishing forces, Kony's LRA began to abduct and forcibly recruit Acholi children to fight for the group, in spite of Kony's claims that the LRA was continuing Lakwena's mission to protect the Acholis. Kony presents himself as a saviour of the Ugandan people and appeals to the religiosity of Acholi children by convincing them that he embodies the spirits of Alice Lakwena and Jesus Christ (Singer, Reference Singer2006, p. 100). Kony allegedly aims to build a Ugandan government based on the Ten Commandments but, under his interpretation, ‘this includes the abduction, torture, rape, and killing of children, the use of sex slaves, and prohibition of living near roads or riding bicycles’ (Singer, Reference Singer2006, p. 100).

The LRA is known to target young and disempowered children for recruitment – a tactic also used by the rebel group, the RUF, which carried out atrocities in Sierra Leone throughout the 1990s. Singer reports that the LRA ‘holds the ignoble record for having the world's youngest reported armed combatant, age 5’ (Reference Singer2014, p. 4), though former LRA child soldiers typically claim they were recruited at thirteen or fourteen (Wessells, Reference Wessells2006, p. 7, citing Derluyn et al., Reference Derluyn, Broekaert, Schuyten and de Temmerman2004; Gates, Reference Gates2011, p. 36). This, however, could just be an estimate, as Acholis do not keep birth records as diligently as Westerners do, and LRA leaders have been reported to lie to children about their ages. Children's age of recruitment has been shown to be a relevant factor in debates about whether child soldiers may be held at fault for their actions inside armed groups.

As Article 26 of the Rome Statute excludes from its jurisdiction persons who were under eighteen at the time of action, child soldiers will not be prosecuted at the ICC for wrongful acts they carry out in armed conflict.Footnote ⁷ Moreover, Article 8 of the Rome Statute identifies as a war crime the ‘conscripting or enlisting children under the age of fifteen years into the national armed forces or using them to participate actively in hostilities’.Footnote ⁸ This effectively sets fifteen years old as the minimum age at which juveniles may lawfully participate in armed conflict, regardless of whether they have volunteered. Most children who come to fight for the LRA do not volunteer, however, but instead are abducted or forcibly recruited, as noted above.

The LRA carries out its abductions through a process known as ‘press-ganging’, which is a ‘form of group abduction wherein soldiers sweep through marketplaces or streets rounding up youths like fish in nets, or raid institutions such as orphanages or schools’ (Wessells, Reference Wessells2006, p. 41). Wessells reported in 2006 that:

Between 1987 and 2006, it was estimated that a total of 20,000 Ugandan children had been abducted by the LRA (Human Rights Watch, 2012). In an attempt to avoid abduction during LRA raids, many Ugandan children (and sometimes entire families) were displaced into camps, but the camps were not well protected and many children were still exposed to the risk of abduction or forcible recruitment (Wessells, Reference Wessells2006; Human Rights Watch, 2012).

During or soon after recruitment, LRA child soldiers endure initiation rituals wherein new recruits are forced to publicly kill a friend or family member at credible threat of immediate execution (Wessells, Reference Wessells2006, p. 14). Under Article 31(1)(d) of the Rome Statute, duress offers a complete defence to crimes alleged under the Statute (Kersten, Reference Kersten2016) and it has been argued that duress would provide a full defence for children who kill innocents to save their own lives during such initiation rituals (Grover, Reference Grover2012, p. 98, citing Happold, Reference Happold2005). After these initiation rituals are carried out, children may then be coerced to drink the blood of their deceased victims or hack their bodies to pieces, in order to desensitise the children to the violence and brutality that will soon become the norm in their lives. Leaders also mutilate the bodies of new recruits in visible ways, on the face, for example, to create stigmatic markers of their association with the LRA. These practices morally sever child soldiers from their previous lives, before they are physically separated from them and brought into the ‘bush’, or isolated jungle, for training.

Moving from recruitment to modes of retention, ethnographic studies recount that many former LRA child soldiers explain that they had learned to follow the rules or consent to being killed (Oloya, Reference Oloya2013). To prevent them from escaping, some children are put into a chain gang, using a chain made out of barbed wire, so that the children have to cut through their own limbs in order to get free (The Reckoning, Reference Yates2009). Others are killed while trying to escape, and some die during raids of villages or clashes with government forces. Children who survive are subjected to a strict regime of training. Like ordinary soldiers, LRA child recruits are put in uniforms and given war names but, unlike ordinary soldiers, among whom camaraderie is encouraged, LRA leaders deliberately undermine the forming of trust among new child recruits. For example, the LRA institutes a policy where ‘talking with other new recruits is a punishable offense’ (Wessells, Reference Wessells2006, p. 63). Wessells further explains that training aims to break the children's wills: ‘Typically the training agenda is not to develop military or survival skills but to break children's will and to achieve high levels of dominance and control’ (Reference Wessells2006, p. 58).

Moreover, Wessells observes that LRA leaders often lie to new recruits in order to further their isolation, heighten their disempowerment and induce them to carry out tasks. He writes:

Engineering child soldiers’ ignorance functions to keep them inside the group, as they come to believe that doing so is their only chance for survival. Life inside the group is also fraught with credible and substantial threats to children's physical security and survival, however. The LRA uses severe corporeal punishments, including torture and execution, for disobedient children. A common torture method used by the LRA is ‘tabay’ where the arms are bound together behind one's back so tightly that the rib cage breaks (Human Rights Watch, 2008). To add to the brutality, some child soldiers report that those who screamed or cried were killed, and onlookers were encouraged to applaud during these routine episodes (Human Rights Watch, 2008).

The lack of trust and constant fear of punishment create an environment where children see one another as competitors for survival. What is more, some child soldiers are made to carry out brutal acts of punishment and torture against one another. Far beyond simply seeing one another as competing for survival, this forces children to be divided against themselves at a very deep level. In light of this brutality, it is natural to think that children subjected to such experiences should know that what they are made to do is wrong (McMahan, Reference Mcmahan2009), but this becomes harder to sustain the longer children survive inside the LRA. The isolation and deliberate undermining of trust, combined with the brutal tactics and psychologically invasive forms of socialisation and indoctrination, can explain why so many children who enter the LRA unwillingly later become willing participants in the groups’ activities.

Indeed, Wessells explains that ‘Children who grow up having learned fighting as their only means of livelihood and survival are likely to continue fighting for more years than adults’ (Reference Wessells2006, p. 30). He adds that:

This quotation reinforces the notion that child soldiers are biologically, psychologically and socially immature. In spite of the brutal tactics used by the LRA against them, child soldiers might reasonably believe that the group is their only source of protection, or that what is done to them, and what they are forced to do, is largely beyond their control.

Gates also finds that many child soldiers become loyal to the LRA as they grow up, despite initial unwillingness to participate (2011, p. 30, citing Vermeij, Reference Vermeij2009 and Beber and Blattman, Reference Beber and Blattman2013). One LRA commander explained in an interview that this is part of the LRA's plan, as the group targets children for recruitment because they are more malleable than adults:

Not all children become ‘good, integrated rebels’, although many do. As Wessells explains:

Counter-intuitively, children who grow up inside the LRA to become ‘hardened perpetrators’ and ‘participate willingly in atrocities no one order them to commit’ may be the most harmed by their experiences. This is supported by research presented in the next section, which shows that the typical experiences of child soldiers in groups like the LRA are known to create a substantial risk of harm to emotional and moral development.

Moving back to the case of Dominic Ongwen for a moment, his defence acknowledges the fact that Ongwen was deprived of ‘any sort of normal development’,Footnote ⁹ but rather than argue that Ongwen's adult development is so harmed as to deprive him of the capacities of responsible agency under the Rome Statute, the defence instead argues that Ongwen should be excused on the basis of duress. Here, I think, the prosecution actually has the stronger argument. For those who spend their adolescent years inside armed groups and become hardened, thus carrying out atrocities no one ordered them to commit, duress is no longer as readily available, as a defence under the Rome Statute. Article 31(d), which contains the defence of duress, states:

1. 1. (i) Made by other persons; or

2. 2. (ii) Constituted by other circumstances beyond that person's control.’

For the defence to apply, it must be true that a person acts reasonably and necessarily to avoid a threat, and not intend to cause more harm than one sought to avoid. It cannot be called necessary to carry out atrocities one has not been ordered to commit; nor could such conduct be interpreted as seeking to avoid more harm than one intends to commit. However, such conduct is nonetheless reasonable to expect of adult soldiers who were abducted or forcibly recruited into armed groups and subjected to the kinds of extreme experiences described above throughout their adolescent formative years. This is because such experiences create substantial risks of harm to adult development. Hence, for an adult soldier recruited as a child who grew into the kind of hardened perpetrator I described above, a capacity-based argument that derives its force from developmental harm is a stronger argument to make over an argument from duress. As I show in the next two sections, this is true in terms of both the defence's psychological and legal foundations.

In the next section, I present recent empirical work to show that the typical experiences to which child soldiers in extreme groups like the LRA are subjected create a substantial risk of harm to adult development. This harm manifests itself as emotional disturbances and harm to moral development, especially to the capacity for moral perception. In Section IV, I then argue that ordinary moral perception is a requirement of responsible agency under the Rome Statute.

III. Developmental risks associated with child soldiering in the LRA

A considerable body of empirical work shows that the experiences associated with child soldiering in extreme armed groups create a substantial risk of harm to emotional and moral development, although former child soldiers have also been shown to be resilient and capable of reintegrating successfully into their communities with proper support (Betancourt et al., Reference Betancourt, Agnew-Blais, Gilman, Williams and Ellis2010a; Reference Betancourt, Borisova, Williams, Brennan, Whitfield, de la Soudiere and Gilman2010b; Reference Betancourt, Brennan, Rubin-Smith, Fitzmaurice and Gilman2010c; Wessells, Reference Wessells2006; Wainryb, Reference Wainryb2011; Boyden, Reference Boyden2003). Much of the existing research on the effects of armed conflict on child soldiers focuses on the psychological impact of child soldiering on youth (Schauer and Elbert, Reference Schauer, Elbert and Martz2010) but, more recently, researchers are looking at the ways in which child soldiering affects moral development (Wainryb, Reference Wainryb2011). Where the former approach tends to view child soldiers largely as passive victims, the latter emphasises the ways in which child soldiers exercise agency inside armed conflict by constructing moral meaning out of their experiences.

This section reviews literature on child soldiering in both the LRA from Uganda and the RUF from Sierra Leone. I consider work conducted with former RUF child soldiers for two main reasons. First, the RUF is demobilised while the LRA is not, so there is simply more existing research with former RUF child soldiers to consider. Secondly, there are important similarities between Ugandan and Sierra Leonean cultures, the conflicts involving the LRA and RUF, and the typical experiences of child soldiers in each group, which makes it reasonable to expect similarities in how child soldiering in these two groups affects children's later development (Peters, Reference Peters2010; Titeca, Reference Titeca2010).

Research conducted with former RUF child soldiers shows a substantial risk of emotional disturbance created by child soldiering in this group. Theresa Betancourt has led much of this research (Reference Betancourt, Agnew-Blais, Gilman, Williams and Ellis2010a; Reference Betancourt, Borisova, Williams, Brennan, Whitfield, de la Soudiere and Gilman2010b; Reference Betancourt, Brennan, Rubin-Smith, Fitzmaurice and Gilman2010c). For example, a longitudinal study conducted by Betancourt et al. (Reference Betancourt, Brennan, Rubin-Smith, Fitzmaurice and Gilman2010c) investigated self-reported internalising and externalising problems and postwar adaptive or pro-social behaviours in relation to war experiences and postwar factors. Test subjects included male and female demobilised child soldiers who ranged in age from ten to seventeen years at the start of the study. Sixty-nine per cent of test subjects remained participants in the study throughout the entirety of the longitudinal course. Interviews were conducted in 2002, 2004 and 2008. If the average age at baseline in 2002 was ten to seventeen, this means that, in 2004, the range was twelve to nineteen and, at the final phase in 2008, the age range was sixteen to twenty-three. Essentially, the longitudinal nature of the study allowed researchers to look at the short- and long-term mental health of former child soldiers.

Internalising problems were measured by reported anxiety and depression. Externalising problems were measured by reported hostility. (Although some level of anxiety, depression and hostility may be ordinary for individuals, higher levels sustained over time contribute to what I call ‘emotional disturbances’ in this essay.) With respect to internalising problems, Betancourt et al. (Reference Betancourt, Brennan, Rubin-Smith, Fitzmaurice and Gilman2010c) found that being raped during the war left former child soldiers at higher baseline levels of anxiety and depression relative to those who were not raped. Increases in postwar internalising problems were also linked with involvement in armed conflict at a younger age and postwar social or economic hardship. Initial community acceptance after demobilisation was associated with lower baseline levels of internalising problems, and decreases in internalising problems were linked with increases in community acceptance over time.

With respect to internalising problems, Betancourt et al. (Reference Betancourt, Agnew-Blais, Gilman, Williams and Ellis2010c) found that increases in externalising problems related to hostility were linked with the experience of killing or injuring others and the postwar experience of stigma. Former child soldiers who were accepted by their communities rather than stigmatised manifested decreased externalisation. This suggests that hostility among former child soldiers who carried out violent acts during the war may be significantly reduced by community acceptance after demobilisation. Moreover, the combination of the war experience of killing or injuring others and postwar stigma were associated with decreases in adaptive and pro-social behaviours in former RUF child soldiers.

Similar findings are rendered from research that includes former LRA child soldiers. A study conducted by Colin MacMullin and Maryanna Loughry (Reference MacMullin and Loughry2004) examined the psychosocial adjustment of escaped former child soldiers from the LRA as well as the RUF. MacMullin and Loughry administered a questionnaire to 209 former RUF child soldiers from May to September 2000 and 567 former LRA child soldiers from July to December 2001. They examined adjustment along four measures: anxiety and depression, hostility, pro-social behaviour and confidence. They tested for relations with the following factors: abduction or not, age, gender, intervention after escape or after having returned home, length of time in armed group, length of time since escape, education and pre-abduction experiences.

MacMullin and Loughry (Reference MacMullin and Loughry2004) found that the most significant factor affecting postwar outcomes was whether the child was abducted or not (p. 467). Accordingly, they found increased anxiety, depression and hostility, and decreased sociability and confidence among abducted children as compared to non-abducted children. Moreover, they found that the higher the child's education level before abduction, the better adjusted the child was after escape, and the more time abductees spent in the armed group, the less well adjusted they were after escape.

Ilse Derluyn et al. (Reference Derluyn, Broekaert, Schuyten and de Temmerman2004) examined PTSD symptoms in former child soldiers who were abducted by the LRA in Uganda. They interviewed 301 former child soldiers, with the average age of abduction at 12.9 years old and the average time spent in the group at 744 days. Seventy-one former LRA abductees also provided information for an impact of event scale and, among them, sixty-nine (97 per cent) reported PTSD reactions with clinical significance.

Other research with former LRA and RUF child soldiers is generally consistent with the findings presented above. Additional factors that have been found to be relevant to postwar outcomes include a child's stage of development before entering armed groups, and the nature and duration of a child's war experiences (Wessells, Reference Wessells2006, p. 129, citing Joop and De Jong, Reference Joop, De Jong, Joup and De Jong2002; Pynoos et al., Reference Pynoos, Steinberg, Goenjian, van der Kolk, McFarlane and Weisæth1996; and Terr, Reference Terr1991).

Overall, the research shows that increases in emotional disturbances among former child soldiers are linked with having entered armed conflict at a younger age, having been exposed to more violent experiences while inside armed conflict, and having spent more time inside armed groups. Other recent research in psychology shows, moreover, that not only can emotional disturbances substantially impair practical reasoning (Damasio, Reference Damasio2005; Kahneman, Reference Kahneman2011), but also such disturbances can impair the capacities of moral agency. This finds support in empirical work recently conducted on the moral development of child soldiers from extreme armed groups.

A theme emerging from the literature on how child soldiering affects moral development is that children's moral development does not stop once they enter armed groups, but that it is constructed in problematic ways. With respect to the continued moral development of child soldiers, researchers have found that children continue to develop moral concepts while inside armed conflicts (Wainryb, Reference Wainryb2011; Posada and Wainryb, Reference Posada and Wainryb2008; Wessells, Reference Wessells2006; Boyden, Reference Boyden2003). The research also shows, however, that former child soldiers apply these moral concepts in problematic ways – that is, consistent with the notions of right and wrong accepted inside armed groups. Taken together, this research shows that, while child soldiering does not destroy the capacities of moral agency, it does create considerable risks of harm to moral development.

In the remainder of this section, I focus on explaining how, and in what sense, child soldiers come to construct their moral agency in problematic ways inside armed groups. This helps set up the legal argument that I present in the next section. My analysis relies considerably on the ethnographic work of Opiyo Oloya, who is an Acholi - that is, a member of the same cultural group as the children that the LRA abducts and forcibly recruits.

Oloya (Reference Oloya2013) finds that the LRA exploits cultural and moral meanings that are familiar to Acholi children, while anchoring them in a larger value system that normalises violence, in order to control the children. The LRA ‘manipulates the children to make these actions [i.e. raping, killing, pillaging] appear normal through liminal repurposing of Acholi culture’ (Oloya, Reference Oloya2013, p. 74). He introduces the concept of ‘liminal repurposing’ to capture the ‘subversion of culture’ whereby the LRA ‘forces child abductees to undergo the rebranding of the mind, a retooling of previously held cultural ideas, and introduction into a new thought process within a repurposed cultural space’ (Oloya, Reference Oloya2013, pp. 74–75). In other words, the LRA ‘mobilize[s] cultural resources’ to normalise the abnormal (Oloya, Reference Oloya2013, p. 75).

Among the many aspects of Acholi culture that the LRA mobilises and exploits is the Acholi notion of manhood. Acholi children go through rituals that mark the transition of boy to man, which aim to challenge children physically, mentally and emotionally. Recall from above that the LRA puts children through initiation rituals that are held to mark the transition from child to soldier. These rituals are meant to mimic the rituals that the Acholis use to mark the transition of boy to man, but the LRA rituals also force children to make a choice between their life and the life of someone else. Thus, the LRA repurposes Acholi rituals by forcing children to make this choice to normalise the violence that will soon saturate their lives.

The LRA also mobilises and exploits Acholi spirituality and their belief in magic. Kony tells the children that he embodies the spirits of Jesus Christ and their former revolutionary hero, Alice Lakwena. To convince them, he uses magic tricks to make it seem that he is impenetrable to bullets. Children are more easily persuaded than adults to begin with, and when this is combined with their spirituality and belief in magic, it is understandable that children would come to believe Kony's stories, especially if showing signs of not believing them is a punishable offence.

Wessells explains that children are just beginning to morally appreciate the world around them, which would support the notion that their moral development is powerfully shaped by their experiences inside armed groups:

This view is consistent with mainstream views in moral psychology on children's developmental vulnerability and with contemporary work in neuroscience on adolescent development (Watson, Reference Watson2004; Wallace, Reference Wallace1994; Wolf, Reference Wolf1993; Steinberg, Reference Steinberg2005; Pizzaro and Salovey, Reference Pizarro, Salovey and Aronson2002).

Ultimately, child soldiers construct moral agency inside armed groups in a manner that is anchored in their own survival, which depends on adopting the rules of the armed group. Children learn the rules of the LRA, and how to use those rules to expand their agency and survive. In doing so, it is natural that they would come to discount suspicions they might have about the nature of their conduct. The mind's inclination to reduce cognitive dissonance can explain this (Festinger, Reference Festinger1957). What is more, in learning how to survive, children also serve to bind themselves to the group because they come to identify group leaders as gatekeepers of their security (Beber and Blattman, Reference Beber and Blattman2013). The well-known phenomenon of the Stockholm syndrome accounts for this process of traumatic bonding through which captives develop empathic identification with their captors (Bejerot, Reference Bejerot1974). The tactics the LRA uses to mould children into soldiers is psychologically powerful and ultimately erodes Acholi cultural and moral norms against violence (Oloya, Reference Oloya2013).

The normalisation of violence that occurs inside the LRA can account for the observation that, while child soldiers continue to develop moral concepts inside the armed group, they also come to apply them consistent with standards of right and wrong established by LRA leaders. As Wainryb observes, child soldiers’ ‘moral concepts of what is just and right are often divorced from what they expect themselves and others to actually do, and are applied selectively to some people but not to others’ (Wainryb, Reference Wainryb2011, p. 274). Moreover, this is not unique to LRA child soldiers. Boothby observed that recently demobilised child soldiers from Mozambique lacked the ability to apply moral concepts consistent with accepted notions of right and wrong outside the group for which they fought (Boothby et al., Reference Boothby, Crawford and Halperin2006).

An inability to apply moral concepts consistent with accepted notions of right and wrong outside of armed groups is evidence of harm done to the capacity for moral perception. This is consistent with the psychological research (considered above) that shows emotional disturbances in former child soldiers, insofar as emotions drive moral perception. Moral perception is the pre-cognitive process by which we apply moral concepts to particular cases (Blum, Reference Blum1991; Reference Blum1994). The process can be captured by the metaphors of ‘seeing’ or ‘feeling’ that something is right or wrong. What is more, moral theorists generally agree that the ability to apply moral concepts to particular cases in accordance with notions of right and wrong accepted by the moral community is an essential ingredient in moral agency (Wolf, Reference Wolf1993; Wallace, Reference Wallace1994; Scanlon, Reference Scanlon1998; Fisher and Ravizza, Reference Fischer and Ravizza2000). The moral community is not a geographic community, but the community of moral agents who have a basic shared understanding of moral concepts and how they are applied in particular cases.

Although finding moral agreement on the international level can be a challenging task, international prohibitions against the killing of innocents and causing wanton harm to civilians capture some of the basic moral concepts shared by the moral community. Indeed, prohibition of these acts, which are defined as crimes against humanity under the Rome Statute, is evidence of these shared moral norms. Inside the LRA, however, child soldiers construct moral agency consistent with the notions of right and wrong that are imposed on them by group leaders, which are radically at odds with these basic standards of right and wrong accepted by the international community. They are encouraged to mutilate innocents and cause wanton harm to civilians and, what is more, they are rewarded for doing so and punished for showing sadness or sympathy at the suffering of their victims, thus inducing further compliance and reinforcing their moral acceptance of such acts as they continue to vie for their survival inside armed groups.

Overall, existing research that examines the moral development of child soldiers shows that, while child soldiering does not stop the development of moral agency, it leads child soldiers to construct it in problematic ways (Oloya, Reference Oloya2013; Honwana, Reference Honwana2011; Wessells, Reference Wessells2006). An important indicator of this is that child soldiers continue to develop moral concepts inside armed groups, but apply these moral concepts in particular cases consistent with the notions of right and wrong that were accepted inside their armed groups. This is evidence that the experiences to which child soldiers in groups like the LRA are subject create serious risks to moral development, and especially to the capacity for ordinary moral perception. This kind of harm can result from the normalisation of violence that occurs inside armed groups, as well as from the emotional disturbances that are found in developmental outcomes of former LRA child soldiers.

With this in view, I now turn to Section IV where I argue that moral perception in accordance with internationally accepted standards of right and wrong is an implicit requirement of responsible agency under the Rome Statute. After locating this requirement in the ‘manifest illegality’ provision, I then argue, in Section V, that it would be unfair to hold ARC soldiers who lack the requisite moral perception to the same standards as ordinary adults under the Rome Statute, provided they cannot reasonably be held to be at fault for having this inability.

IV. Traumatised ARC soldiers and the manifest illegality provision

International criminal law recognises grounds for excusing perpetrators from criminal responsibility under Articles 31–33 of the Rome Statute. The Rome Statute recognises an excuse from criminal responsibility for persons who cannot know what the law requires of them or conform their conduct to it because they are mentally ill. Article 31(a) states that:

This provision requires persons to have the capacity for judgment to be able to appreciate the unlawful nature of their conduct, and the capacity for decision-making (which is the capacity to make the decision to act on their judgments about what the law requires), in order to be capable of criminally responsible choice. This formulation sets the bar very high, insofar as one's capacities must be entirely destroyed for the defence to apply. Even if this defence does not apply to ARC soldiers, in whom practical reason is not destroyed, there is a further provision in the Rome Statute that speaks to the capacities of moral agency required under this body of law.

Article 33 provides a legal basis for thinking that normal moral perception is an implicit requirement of criminally responsible agency under the Rome Statute. Article 33 states:

• ‘1 The fact that a crime within the jurisdiction of the Court has been committed by a person pursuant to an order of a Government or of a superior, whether military or civilian, shall not relieve that person of criminal responsibility unless:

  1. (a) The person was under a legal obligation to obey orders of the Government or the superior in question;

  2. (b) The person did not know that the order was unlawful; and

  3. (c) The order was not manifestly unlawful.

• 2 For the purposes of this article, orders to commit genocide or crimes against humanity are manifestly unlawful.’

This provision contains a doctrine with a long history so, to establish the proper background for my argument, I first lay out a brief history of the doctrine of manifest illegality. Then I argue that, despite the absence of explicit moral language in the Rome Statute's version of the doctrine, it nonetheless assumes that those under its jurisdiction are capable of perceiving where basic moral norms apply in accordance with internationally accepted standards of right and wrong – a capacity that I call ‘ordinary moral perception’.

Before the Nuremberg trials, which prosecuted Nazi perpetrators of the mass atrocity largely against the Jewish people, soldiers could be excused from criminal responsibility for acts they performed in war simply by establishing that they were following orders from a superior. This defence was challenged at Nuremberg. The judges held that some actions are so obviously morally wrong that the fact that a soldier was following an order from a superior could not function as an excuse for performing them. Even if the law did not prohibit such acts, the Court held that they were ‘crimes against humanity’ for which soldiers could be prosecuted, regardless of whether they were obeying orders. Soldiers who were ordered to commit mass atrocity against the Jews, for example, were expected to have the requisite moral perception to realise that what they were ordered to do was wrong and that they were required to do otherwise.

While the Rome Statute's version of the defence does not makes any explicit reference to morality, Larry May observes that ‘there seems to be a partial convergence [of the Rome Statute] with the Nuremberg defence, since ascertaining what is manifest requires use of moral perception’ (2005, p. 197). Mark Osiel, who has also written extensively on the matter, interprets the provision similarly:

May and Osiel both provide convincing interpretations of the manifest illegality provision that show that the Rome Statute assumes persons to have the requisite moral perception that would allow them to perceive the wrongfulness of crimes against humanity and acts of genocide. Because these crimes cover acts such as murder or mutilation of innocent people, it is fair to say that one simply needs a basic moral competence to satisfy the assumption, where this is understood as the ability to see that these acts are wrong. I have called this basic competence ‘ordinary moral perception’ to capture the notion that it is no more sophisticated than that of which an ordinary adult is capable. Simply knowing that others think that certain conduct is wrong is not moral perception as such, though it would be enough (in most circumstances) for a person to know that the behaviour is ‘manifestly unlawful’. It is reasonable to expect that the traumatised soldiers to whom my argument applies lack this knowledge, however, because they have been, and remain, isolated from persons who have ordinary moral perception.

What is more, May argues that the manifest illegality provision is too demanding even as applied to a typical adult soldier with military training who acts in extreme combat situations. He considers the infamous massacre at My Lai, where American troops killed unarmed civilians, including women and children. Where some of these soldiers were prosecuted for the massacre, May questions whether they could reasonably have been expected to perceive that killing the My Lai villagers was unlawful and wrong, after having been engaged in a guerilla war with their enemy and having been notified by their superiors that every person in the village posed a threat to them. May's challenge to the idea that ordinary soldiers can exercise normal perception in the abnormal circumstances of armed combat is based on psychological research showing that hostile circumstances impair judgment. Regardless of whether one agrees with May's claims about My Lai, May's logic is quite strong when applied to traumatised ARC soldiers.

Still, one might object to my application of the manifest illegality provision to ARC soldiers on the grounds that they do not fight for governmental armies, and so are not covered in Article 33. However, the manifest illegality provision creates the obligation to disobey the orders of a lawful superior when those orders are to commit acts of genocide or crimes against humanity, which means that, regardless of the fact that a lawful superior has issued the order, the action contained in the order is itself so wrong that not only can no one can be obligated to obey it, but also there is actually an obligation to disobey it. I argue that, if persons are obligated not to perform particular actions contained in the orders of lawful superiors, then it follows, a fortiori, that persons who are not even ordered by lawful superiors to perform these actions are also under obligations not to perform them. In order to be fairly held accountable for not fulfilling these obligations, however, one needs to have the capacity of ordinary moral perception in order to be capable of seeing the manifest illegality of the conduct one is ordered to engage in or, as I argue below, one must have at least had a reasonable opportunity to develop such perception.

It is a mistake to assume that a person whose practical reason is only impaired, rather than destroyed, has the capacity to perceive, for example, that killing innocent people is always wrong, if such a person has grown up and remains in armed conflict. I argue that ordinary moral perception, furnished by emotional development, is also required for persons to have this ability. As I showed above, there is reason to think this capacity is substantially harmed by the extreme experiences child soldiers are subjected to in groups like the LRA. The next section builds on this idea to argue that, if ARC soldiers who lack ordinary moral perception (or ‘traumatised ARC soldiers’) cannot be held at fault for lacking this capacity, it is unfair to hold them to the same standards as ordinary adults under the law. This section showed that there is language in the Rome Statute that would provide legal grounds for full or partial defence to liability for such individuals.

V. Traumatised ARC soldiers and non-culpable ignorance

This section argues that, if traumatised ARC soldiers are neither at fault for developing or maintaining their defective moral perception, then it is unreasonable to say they are at fault for it. It is relevant whether traumatised ARC are at fault for their defective moral perception because, unlike a mental disease or defect, which must destroy a person's practical reason to qualify as a defence under the Rome Statute, the defence I am developing for traumatised ARC soldiers would apply to individuals in whom the capacity for moral perception is not destroyed, but is substantially harmed. Only those traumatised ARC soldiers who, in light of the Rome Statute, did not culpably contribute to their present incapacity warrant a defence from liability.

First, I argue that we cannot consider traumatised ARC soldiers at fault for the acts they performed during their formative adolescent years that contributed to their defective adult development. Part of the reason for this is that the Rome Statute excludes acts performed by persons under eighteen from its jurisdiction, thus providing no real standards for evaluating their conduct. The scholarly literature on child soldiers typically considers children under fifteen apart from adolescents aged fifteen to seventeen, as some people think the latter are capable of responsible agency. In what follows, I thus examine whether it is reasonable to hold children of either group at fault for their conduct inside armed conflict, under existing standards articulated in the Rome Statute.

The recent ICC case against Congolese war criminal Thomas Lubanga supports the view that all children under fifteen are legal victims under the Rome Statute. Among the international crimes of which Lubanga was convicted at the ICC was the war crime of conscripting and enlisting children under the age of fifteen to actively participate in hostilities. The judges gave the provision a broad construction, interpreting active participation to include serving in auxiliary roles (e.g. spies, porters), not only on the front lines.Footnote ¹¹ The judgment explained that the purpose of the provision is to prevent children from being targeted in armed conflict, and their association with armed groups creates the risk that they will be perceived as legitimate military targets, regardless of whether they directly serve in battle.

The judgment in the Lubanga case did not carve out an exception for child soldiers under fifteen who volunteer for armed groups. If children under fifteen are not mature enough in their agency to decide whether to volunteer for armed groups, then a fortiori they cannot be considered mature enough in their agency to decide how to act inside armed groups, where pressures influencing choice are typically far greater. This provides legal grounds for thinking that existing international criminal law considers children under fifteen insufficiently developed in their agency to be responsible actors under the law, and thus not at fault for their conduct inside armed conflict.

Questions concerning the legal obligations of children aged fifteen to seventeen pose greater challenges for the developing body of international criminal law, however (Acton Thomas, Reference Thomas2013). Children of this age may lawfully participate in armed conflict under international criminal law, but there are no clear legal standards under the Rome Statute for evaluating their conduct. As noted above, Article 26 excludes persons under eighteen from its jurisdiction, but the Statute does not explain whether the exclusion is substantive or procedural. Some hold that the Statute's exclusion effectively establishes eighteen as the minimum age for prosecution of international crimes (Grover, Reference Grover2012, pp. 76–78), while others point out the historical record shows that the drafters of the Statute left the issue of prosecuting child soldiers to the states to ensure ratification of the Statute, and argue that states and other international tribunals thus have discretion to prosecute persons under eighteen (Happold, Reference Happold2005, p. 77; Cohn, Reference Cohn2001, pp. 6–7).

As one scholar argues, the legal invisibility of adolescents who are aged fifteen to seventeen under the Rome Statute renders their conduct neutral before the ICC, thus creating a potentially perilous statutory gap (Chaikel, Reference Chaikel2015). Chaikel argues that the Rome Statute can be amended to resolve this issue, either by raising the age at which children can participate in armed conflict to eighteen or lowering the minimum age for criminal responsibility to fifteen. Advocates of children's rights continue to advocate a ‘Straight 18’ rule that excludes persons under eighteen from the jurisdiction of international criminal courts (Rosen, Reference Rosen2007). Others, recognising the agency exercised by children in armed conflict, propose lowering the age of criminal responsibility to fifteen and creating international juvenile courts where they may be prosecuted (Ursini, Reference Ursini2016).

I offer two reasons in support of the view that the best interpretation of the existing Rome Statute is one that exempts adolescents aged fifteen to seventeen from legal obligation. First, this interpretation is supported by past practice insofar as no international court has yet to hold persons under eighteen criminally responsible for their participation in mass atrocities. Secondly, this interpretation of the Rome Statute is further supported by an international trend towards raising the minimum age of participation in armed conflict to eighteen. This trend is apparent by the adoption of eighteen as the minimum age of participation in the Optional Protocol to the Convention of the Rights of the Child (2000) and in the Paris Principles (2007), the latter of which has been endorsed by over 100 countries worldwide (Chaikel, Reference Chaikel2015).

Based on this, I argue that it is unreasonable to trace the legal culpability of traumatised ARC soldiers to the morally culpable actions they may have performed inside armed conflict, out of which their adult characters developed. Insofar as the developmental harm at issue in the case of traumatised ARC soldiers is substantial harm to moral perception, and harm to practical reasoning, then it is unreasonable to hold traumatised ARC soldiers at fault for these attributes of their adult agency unless it can be shown that they are at fault for not remedying them while inside armed conflict.

Secondly, I argue that we cannot consider traumatised ARC soldiers at fault for maintaining their distorted moral perception insofar as they remain in the forcibly limited and hostile environment of armed conflict. Even if they are now adults and no longer subject to the same demands they were as children, the force of an argument constructed on this derives from the fact that they remain in the larger forcibly limited and hostile environment of armed conflict. This environment greatly burdens the capacity for reflective self-control, the exercise of which would allow them to recognise the distorted nature of their moral perception. Reflective self-control is the set of abilities that allow persons to reflect on constitutive parts of their characters in light of the actions their characters incline them to pursue, to judge whether to maintain or change these parts of their characters and, finally, to act on the conclusions of these judgments (Wallace, Reference Wallace1994).

It is highly problematic to expect traumatised ARC soldiers who remain in armed conflict to be able to exercise reflective self-control to remedy an inability to perceive the manifest illegality of their conduct. First, while they may no longer be subjected to direct threats rising to the level of coercion, it is reasonable to expect traumatised ARC soldiers would experience the situation as threatening as an ordinary person would experience a situation of direct coercion. Secondly, while traumatised ARC soldiers may no longer be forcibly isolated, it is reasonable to expect them to remain isolated from wider society on the basis that they think others pose a threat to them. While one might say that, if this is true, it concedes that on some level they must perceive that what they are doing is wrong, this concession does not necessarily follow. It is simply not the case that all instances in which a person perceives others to be a threat follow from the perception that he or she is doing, or has done, something wrong. (Consider, for example, the perceptions of a paranoid person.) Thirdly, while ARC soldiers may no longer be indoctrinated with highly questionable or patently false accounts that reinforce the values of the armed group into which they were socialised as child soldiers, it is reasonable to expect that this indoctrination has given rise to settled beliefs of ARC soldiers regarding the nature of their situation. Moreover, even if they are capable of reflective self-control, it is unreasonable to expect its exercise will penetrate the interpretation they were taught throughout their lives about their situation, so long as they remain isolated in the forcibly limited and hostile environment of armed conflict.

There is support for the view I have been developing in the Rome Statute's formulation of the defence of duress in Article 31(d). To recall, Article 31 broadens the scope and nature of the threat one acts to avoid, in that it may be a ‘threat of imminent death or of continuing or imminent serious bodily harm’, where the threat one acts to avoid is either ‘made either by other persons’ or ‘constituted by other circumstances beyond that person's control’. There are other provisions that must be satisfied for the defence to apply, which I have reasoned above would not be satisfied in considering the conduct of traumatised ARC soldiers, but the fact that Article 31(d) identifies continuing threats constituted by circumstances beyond a person's control as relevant for duress to apply, is, in effect, to recognise that environments like those of armed conflict place substantial burdens on human moral capacities.

The ways in which the environment of armed conflict places substantial burdens on human moral capacities for perception, deliberation and decision-making is well documented. In light of these burdens, one scholar goes as far as to say it is unreasonable to expect ordinary moral perception in the extraordinary circumstances of armed conflict even from a typical adult soldier trained for the hardships of war (May, Reference May2005). Other scholars maintain that the substantial burdens of this environment create a presumption of non-responsibility for all soldiers who act under such conditions (Doris and Murphy, Reference Doris and Murphy2007). One need not go as far as either of these views to recognise that it is unreasonable to expect traumatised ARC soldiers to remedy their inability to perceive the manifest illegality of their conduct while they remain in the forcibly limited and hostile environment of armed conflict.

The argument that it is unreasonable to expect traumatised ARC soldiers to be able to exercise reflective self-control as adults is especially true for those who are still fully engulfed in active hostilities, but it also applies to those in the wider circumstances of armed conflict. A more precise measure is needed for just how threatening and isolating their circumstances must continue to be for it to be unreasonable to expect that they are able to exercise reflective self-control. I propose a provisional approach that takes into account two measures in determining whether it is unreasonable to hold traumatised ARC soldiers at fault for maintaining or not remedying developmental defects to their moral perception: (1) how traumatised ARC soldiers are, as a result of adolescent experiences for which they cannot reasonably be held at fault, and (2) how objectively threatening and isolating their circumstances are.

VI. Implications

This section considers implications of the foregoing argument on how the international community can justly respond to wrongful, but non-culpable, conduct carried out by traumatised ARC soldiers inside armed conflict. I propose that a provision should be added to Article 31 of the Rome Statute to cover the kind of case I have described in this essay. Such a provision could make it a further ground for excluding criminal responsibility that:

The provision could provide a mitigation of sentence in some cases and a complete defence in others, depending on how well the criteria of the defence are satisfied in individual cases. Where a mitigation of sentence is warranted, the ICC should consider alternatives to punishment in order to offer non-punitive incapacitation and accountability. A special chamber of the ICC can be created into a forum that promotes truth and reconciliation, while also serving expressive functions typically associated with criminal prosecution. The Special Court for Sierra Leone (SCSL) offers a good model here.

The UN and the Sierra Leonean government established the SCSL to bring justice to the most responsible perpetrators of the atrocities carried out by the RUF in the 1990s. Many members of the RUF were child soldiers who were subjected to treatment similar to that described in Section II with respect to LRA child soldiers. The SCSL was granted jurisdiction over persons above fifteen years of age at the time of the offence, although the Statute that established the SCSL emphasised rehabilitation and rejected punishment as an option in sentencing child soldiers (Cohn, Reference Cohn2001, p. 14). Moreover, the Statute left the decision of whether to exercise this jurisdiction to the discretion of the prosecutor, who, after heated international debate, decided not to exercise it. Instead, older adolescent child soldiers were to be held accountable in the Truth and Reconciliation Commission (TRC) that was established at the same time as the SCSL (Cohn, Reference Cohn2001, pp. 25–26).

The joint working of the TRC and the SCSL provides a good model for how the international community might respond to traumatised ARC soldiers who commit wrongful acts inside armed conflict. While traumatised ARC soldiers have aged into legal adulthood, there is reason to think they have suffered substantial developmental harm that leaves them at an immature stage of development, such that it would be unfair to treat them like ordinary adults under the law. By encouraging the creation of a permanent TRC to accompany the ICC's operations, the international community can provide both non-punitive incapacitation and accountability for persons who may pose a threat to others, but whom it would be unfair to hold fully or even partially criminally accountable. Pre-trial proceedings can be used to determine who is eligible for these non-punitive alternatives to formal prosecution at the ICC.

What is more, non-punitive incapacitation and accountability are a desirable alternative to formal prosecution because it would avoid the stigma typically associated with conviction and punishment, which would be barriers to the reintegration process. Reintegration studies show that, despite the psychological and moral risks created by child soldiering, individuals exposed to these risks typically are resilient (Ertl et al., Reference Ertl, Pfeiffer, Schauer, Elbert and Neuner2011; Wainryb, Reference Wainryb2011; Betancourt et al., Reference Betancourt, Agnew-Blais, Gilman, Williams and Ellis2010a; Reference Betancourt, Borisova, Williams, Brennan, Whitfield, de la Soudiere and Gilman2010b; Reference Betancourt, Brennan, Rubin-Smith, Fitzmaurice and Gilman2010c; Bayer et al., Reference Bayer, Klasen and Adam2007; Boothby et al., Reference Boothby, Crawford and Halperin2006; Wessells, Reference Wessells2006). Many former child soldiers successfully reintegrate into the community and desist from violence, hold consistent jobs, get married and have children. The research shows that the reintegration prospects of former child soldiers typically depend on sustained rehabilitative and/or restorative efforts (Ertl et al., Reference Ertl, Pfeiffer, Schauer, Elbert and Neuner2011; Wainryb, Reference Wainryb2011; Bayer et al., Reference Bayer, Klasen and Adam2007; Boothby et al., Reference Boothby, Crawford and Halperin2006). Community and family acceptance has been found pivotal to successful reintegration of former child soldiers (Betancourt et al., Reference Betancourt, Agnew-Blais, Gilman, Williams and Ellis2010a; Reference Betancourt, Borisova, Williams, Brennan, Whitfield, de la Soudiere and Gilman2010b), and there is reason to think this would be true for ARC soldiers as well. Conviction and punishment would bring greater stigma to traumatised ARC soldiers, thereby burdening the prospects for communal and family acceptance.

Non-punitive incapacitation and accountability would also allow the ICC to promote the larger normative commitments identified in the Preamble to the Rome Statute. The Preamble of the Rome Statute declares recognition among Member States that, ‘during this century millions of children, women and men have been victims of unimaginable atrocities that deeply shock the conscience of humanity’. The establishment of a permanent international TRC to complement the work of the ICC could afford the international community a greater role in showing sustained efforts towards the rehabilitation and reintegration of some of the world's worst victims who, for various reasons largely beyond their control, become perpetrators themselves.

VII. Conclusion

In this essay, I have articulated a legal basis for recognising a new defence under the Rome Statute for adult soldiers who are so morally traumatised by their past experiences as child soldiers that they are unable to perceive that what they are doing is wrong while they remain inside armed conflict. The case of Dominic Ongwen motivated my project. Ongwen's defence is advancing a complete defence based on duress, but I have articulated an alternative capacity-based argument that would apply to adult soldiers recruited as children, like Ongwen, who do not satisfy the requirements of responsible agency under the Rome Statute, and who cannot be held at fault, under the Rome Statute, for this.

While insufficient facts are known about Ongwen's case to know whether the defence I have articulated would apply to it, my framework can be used to evaluate the conduct of adult soldiers who grew into hardened perpetrators as a result of spending their formative adolescent years inside extreme armed groups like the LRA. I provided independent grounds for thinking that, for an adult soldier recruited as a child who grew into the kind of hardened perpetrator I described in this essay, a capacity-based argument that derives its force from developmental harm is a stronger argument to make over an argument from duress.

In addition to articulating a legal foundation for this defence in the Rome Statute, I also laid out a psychological foundation for it. I showed that the experiences to which children in the LRA are subject are known to create substantial risks of harm to adult development, and especially to the development of moral perception. This was based on fieldwork conducted with former LRA child soldiers and recent psychological research on moral development.

Finally, I argued that, in individual cases where the defence I have developed applies, it could provide either a mitigation of sentence or a complete defence. Even where a mitigation of sentence is warranted, I argued that non-punitive alternatives should be considered. Non-punitive incapacitation and accountability would allow the ICC to express condemnation of the acts performed by traumatised ARC soldiers without attaching to them the stigma that would hamper the reintegration process. I proposed that a permanent TRC could be established as a special chamber of the ICC to provide non-punitive accountability for the traumatised ARC soldiers who have been found to meet the criteria of the defence during pre-trial proceedings. In this way, the ICC can play an active role in furthering international criminal justice consistent with larger normative commitments of the Rome Statute and the international community as a whole.