Victimhood and the adaptive agency of child soldiers

Abstract

Dominant humanitarian and legal discourses homogenise child soldiers as passive and coerced victims. These narratives obscure the complex ways in which child soldiers navigate conflict. Building on contemporary theories of agency, this article proposes an adaptive conception of agency as a conceptual framework for understanding how children exercise meaningful, though constrained, choices under conditions of oppression. Through analysis of diverse enlistment pathways and relationships to conflict, it argues for a diachronic understanding of agency that recognises both coercion and tactical adaption. Recognising adaptive agency allows for a more sophisticated evaluation of children’s roles in conflict, challenging paternalistic frameworks that silence their voices.

Introduction

Child soldiers ¹ occupy a unique and troubling space in childhood and children’s rights studies. While contemporary scholarship demands the recognition of children as active agents, with a right to have their views and choices respected, legal and policy research follows an opposing path. Child soldiers are predominantly framed as passive victims of conflict who exercise little to no agency over their lives. This pervasive victimhood narrative, however, fundamentally conflicts with reality. The pathways into (and roles within) armed conflict are diverse and manifold. While some children are forcibly abducted, many others join groups for a diversity of reasons, be they ideological or not.

This article addresses this contradiction: on the one hand, childhood and children’s rights researchers are adamant to promote and respect children’s agency and choices; on the other, the constraining conditions that frame the life and options of child soldiers pulls us to emphasise their victimhood, thus, dismissing their diverse roles as agents in their own lives (Derluyn et al., 2015). Part of this dismissal stems from treating agency as a rigid category. To capture the lived reality of these children, this article adopts a more situated definition of it: agency is understood here as the act of navigating one’s social environment, regardless of how constrained or volatile it may be (Vigh, 2009: 420; Beier, 2018: 167). By viewing agency as a dynamic engagement with social forces, we can better understand the ways children mobilise available resources to alter their circumstances within the terrain of armed conflict. This article argues that the prevailing, rigid binary between ‘victim’ and ‘agent’ is insufficient to capture this complex reality, and introduces the concept of ‘adaptive agency’ to provide a more robust conceptual framework for understanding the roles and choices children inhabit, thereby moving beyond the limitations of victimhood discourses.

Dissecting the structural inconsistencies in how child soldiers are conceptualised, this article debunks the passive victimhood discourse. It argues that this conceptual imaginary fails Demobilisation, Disarmament and Reintegration (DDR) programmes, and the broader children’s rights project by validating agency only when it adheres to dominant normative standards. To overcome this, it proposes ‘adaptive agency’ as a conceptual tool to make sense the complex reality of child soldiers. It aims to recognise child soldiers as agents while fully accounting for the oppressive social conditions embedding their choices.

Discourse of passive victimhood

International law and policy frame child soldiers as unambiguous victims. While this ensures protections for them, it conceptually constrains our understanding of their experiences, ignoring agency, accountability, and moral complexity. Children’s agency and choices are not recognised nor respected when it comes to their involvement in armed conflict; they are mere victims of it. This dominant portrayal often forecloses alternative interpretive grammars, and silences more complex relationships between choice, coercion, and agency (Tabak, 2014: 60). Yet legal and institutional narratives persist in reproducing a highly constrained image of childhood vulnerability, often sidelining children’s political subjectivities and agency (Krystalli, 2020).

The participation of children in armed conflict is extensively regulated by international law. Article 38 of the Convention on the Rights of the Child (UNCRC) (UNGA, 1989) prohibits recruitment and participation in hostilities for children under 15, mandating protective enforcement of humanitarian law (Geneva Convention, 1977). The Optional Protocol on Children and Armed Conflict (OPAC) (UNGA, 2000) raises the minimum age for direct participation to 18 (Art. 1), encourages increasing the minimum age for recruitment from 15 (Art. 2), and requires non-state armed groups not to recruit individuals under 18 in any capacity (Art. 4) (Drumbl and Tobin ,2019a). Similarly, the ILO Convention 182 classifies forced or compulsory recruitment of children as among the worst forms of child labour (ILO, 1999: Art. 3.a).

Regarding justice and criminal responsibility, General Comment 24 (CRC, 2019) provides the guiding standard. It stipulates that minors’ criminal acts require lesser culpability than adults’, prioritising restorative approaches and striving to avoid judicial proceedings (CRC, 2019: I.2; III, 8). Crucially, the Comment stresses that children involved in armed conflict, even those who committed atrocities, should be treated primarily as victims due to the unlawful nature of their recruitment (CRC, 2019: IV, F. 100; UNSC, 2018). This view is reinforced by the criminalisation of child recruitment. Both forced conscription and voluntary enlistment are regarded as war crimes, and recruitment under 15 is categorically prohibited (ICC, 2021a). Recruitment is considered “always against the best interests of the child,” regardless of whether it is forced or ostensibly voluntary (Coomarasway, 2010: 542). This legal stance ultimately treats all child recruitment as a fundamentally non-voluntary act.

This consensus extends directly to transitional justice contexts. UN standards for DDR programmes, for instance, prefer terms like ‘release’ or ‘exit’ over ‘demobilisation’ when a child leaves an armed group. This choice strategically reinforces their status as victims, by ensuring the child does not possess an official ex-combatant status (UN, 2021: sect. 5.20). This is mirrored by institutions like the Sierra Leone Truth and Reconciliation Commission (SLTRC) unequivocally stating that “children… do not have the ability or the capacity to ‘volunteer.’ Simply put ‘they have no choice’” (SLTRC, 2004: 234). Civil society organisations, such as COALICO in Colombia, echo this, arguing that “all [child recruitment] is forced even if it wasn’t done using force, even if it wasn’t through coercion” (Pannell, 2025). This illustrates the deep discomfort international actors feel in acknowledging any form of voluntariness (Molloy, 2024). When children articulate reasons for enlistment (e.g., ideology, material need), those expressions are often reinterpreted as proof of oppression and lack of actual choice (Martins, 2011: 437).

The prevailing narrative thus aims to shape understandings of child soldiers by portraying them as passive (Ladisch, 2013). This construction is driven by a conception of ‘childhood’ as non-agentic, vulnerable and threatened (Derluyn et al., 2015: 34). The fundamental assertion is that child soldiers, by virtue of being children, are inherently non-agents in relation to their choices. This claim rests on two arguments: first, that children’s developmental immaturity imposes internal constraints that invalidate their ability to make meaningful choices; and second, that the external conditions of armed conflict are always so oppressive that choices made under those circumstances must be non-voluntary.

The argument for internal constraints specifically denies children have the cognitive, emotional, and moral capacity for decision-making in conflict, and the moral agency required to bear responsibility for their choices (Coomarasway, 2010: 545; Fabre, 2018: 411). They are, thus, victims because they lack the necessary competence for moral agency required for voluntary choice (McMahan, 2010: 32). Distinct from questions of mental competence, the external constraints argument posits that pervasive socioeconomic and contextual factors render children’s choices non-agentic. Child soldiers often live in regions lacking basic needs, education, and security (Boyden and Berry, 2004). Oppressive environments, violence, and power pressures compromise voluntariness, making choice under these conditions suspect (Martins, 2011: 437). A UNICEF report defended in this line that while 57% of child soldiers interviewed claimed to have enlisted voluntarily, this had “to be understood in a context where [they] are not exercising free choice but are, instead, responding to economic, cultural, social and political pressures” (UNICEF, 2002: 74).

The victimhood discourse asserts that child soldiers are necessarily victims, lacking competence for informed choice or negated by oppressive conditions. This strategic framing offers benefits: it secures vital legal protection, enables resource mobilisation, supports DDR, and purportedly facilitates reintegration (Schmidt, 2007: 65). However, it raises three analytical challenges the current literature fails to resolve.

Three problems with passive victimhood

While central to mobilising international protection for children in armed conflict, the victimhood narrative’s homogenising logic carries significant conceptual and practical costs. Framing all child soldiers as mere victims, law and policy obscure diverse realities, introduce structural inconsistencies, and risk perpetuating injustice. This section advances three critiques of the passive victimhood discourse. First, the discourse fails to account for the radical diversity of children’s experiences. Second, it raises claims of unfairness and injustice to communities, adult recruits, and other children. Finally, it reveals concerning inconsistencies and biases in determining who, how, and when a child is deemed “allowed” to exercise choice.

The problem of homogenisation

At the heart of the victimhood discourse lies the assumption that every child’s involvement in armed conflict is fundamentally non-volitional. This follows a “deficit” model of childhood; emphasising developmental immaturity and lack of “adult” capacities, thereby legitimising the construction of children as irrational and passive victims (Uprichard, 2008). This framing wishes away the heterogeneity of children’s pathways into war, disregarding their expressions of agency. As the SLTRC declaration that “children… do not have the ability or the capacity to ‘volunteer'” (SLTRC, 2004: 234) demonstrates, this narrative forecloses any possibility of reflective choice. The child is positioned outside the realm of agency, not as someone needing support to navigate complex terrains, but as someone whose perspective is irrelevant to understanding their own actions.

Crucially, this homogenising logic elides the radical diversity in developmental needs across the age spectrum. By denying agency to all individuals under eighteen, the discourse ignores how an individual’s skills, knowledge, and life experience evolve over time (Terlazzo, 2021). Even the youngest recruits possess some capacity for agency, but the refusal to recognise agential capacity becomes untenable for older children approaching the legal threshold of adulthood. Treating a seventeen-year-old's enlistment as ‘forced’ and an eighteen-year-old's as ‘voluntary’ under identical conditions exposes this arbitrariness (Haer, 2019).

Yet, research across diverse conflict settings decisively shows that enlistment cannot be captured by a simple forced/voluntary binary. Child soldiers are not a uniform group, rather revealing a complex matrix of enlistment pathways and varying degrees of agency (Özerdem and Podder, 2011: 8). Without mapping children’s relationships to conflict and to their choices, assumptions about their capacities and the role of external conditions remain problematic. Empirical data paints a complex reality: research with former child combatants with the Colombian FARC showed only 20% of recruits were abducted, and that most ELN recruits stated to have joined voluntarily, encouraged by promises of wealth, prestige and power (Gutierrez, 2009: 122, 132; Brando and Echeverry, 2022). Similarly, in Liberia, only one-fifth of both adult and child combatants were forcibly conscripted, demonstrating comparable patterns of enlistment (Pugel, 2009: 170). Surveys in Central Africa found two-thirds of respondents claimed to have volunteered (International Labour Office, 2003: 29), while in the DRC only 9% cited fear as their motive, compared to 34% for material gain and 21% for ideology (Lazareva, 2011). Studies document children framing enlistment as a rational response to existential threats or as a pursuit of identity and justice (Vautravers, 2019: 104). Ignoring these dynamics misrepresents reality, infantilising children and denying their tactical agency (Honwana, 2006; more in Section ‘Reframing child soldier’s agency through adaption'). To dismiss such narratives as mere manipulation is to impose an externally defined script that silences children as moral subjects, and fails to account for the fact that choices are never made in a vacuum, but always in specific contexts and in response to specific circumstances (Brando, 2024: 95-101).

Challenging this homogenising tendency is especially critical when gender is considered. Girls constitute up to 40% of child combatants in many conflicts (Haer, 2019: 80) yet remain under-documented and studied. When acknowledged, their experiences are reduced to sexual victimhood, resulting in a double erasure that obscures their diverse roles during conflict (McKay and Mazurana, 2004: 37). Yet fieldwork tells a more complex story, underscoring girls’ political and tactical agency. For instance, none of the girl soldiers interviewed by Hauge in Guatemala were abducted; all joined voluntarily, and many reported positive experiences (Hauge, 2011: 101). Interviews in Sierra Leone documented girls using subterfuge and, at times, lethal force to resist the violence around them (Denov, 2010: 130, 182). Former recruits in the Ethiopian Tigrayan People’s Liberation Front (TPLF), similarly, cited political aspirations, escape from forced marriage, and access to education as key motives (Veale, 2005). These girls did not describe themselves as powerless; instead, they appeared self-confident, independent and politically aware. Flattening girls into a feminised victim category misrepresents their experiences of conflict and obscures their agency (Vaha and Vastapuu, 2018, 235).

The homogenisation of child soldiers as passive victims has detrimental consequences. It denies children recognition as social and political actors capable of choice, reproducing their construction as incomplete beings. Critically, it misrepresents their relationship with their social environments, distorting the realities of war, and undermining the effectiveness of prevention and reintegration policies meant to protect them.

Unfairness to children, adults, and communities

The homogenising victimhood framework, while purporting to protect child soldiers, inadvertently obscures the moral significance of differentiated experiences in conflict regions and can perpetuate injustices against other vulnerable children, against similarly situated adult combatants, and against the communities and reintegration processes it aims to support.

First, by treating every child in conflict as equally coerced, the discourse erases the agency and resilience of those who avoid recruitment, dilutes the specific suffering of the forcibly abducted, and of the children who are victims of crimes committed by child soldiers (Drumbl, 2024, 210). This flattening of voluntariness into non-agency is problematic, given that the majority of children in conflict zones resist enlistment (Schmidt, 2007, 54). Brett observes that even when poverty levels are equal across populations of children “there are many more poor children [in conflict zones] who do not become child soldiers than those who do” (Brett, 2003: 860). Protective factors, such as strong social networks and educational access, enable many youths to resist pressure altogether (Burgess, 2011); assuming all children are equal victims can “minimise rather than accentuate” the trauma of abductees and the resilience of defectors, obscuring the imperative to address their distinct harms and needs (Drumbl and Tobin, 2019a: 1543).

The victimhood narrative, moreover, draws an arbitrary age line that infantilises children while taking adult choices under comparable conditions at face value (Drumbl and Tobin, 2019b: 1685). This raises a critical question: why is the agency of adult combatants not similarly problematised? Drumbl and Tobin (2019b: 1685) note that 18 to 25-year-olds often enlist for the same reasons as minors (i.e., poverty, lack of opportunities), yet their agency is never “subject to the same protectionist review”. If manipulation, coercion, and external constraints can invalidate a minor’s agency, why omit the experiences of young adults who are pressured to enlist under similarly oppressive circumstances? (Haer, 2019: 81). This inconsistent attribution of agency not only infantilises children but subtly legitimises adult violence under similar constraints.

Finally, the victimhood discourse can marginalise communities and undermine sustainable reintegration. By insisting on passive victimhood, many DDR programs ignore local expectations of accountability, which often demand demobilised children to acknowledge wrongdoing as part of social healing and restoration (Akello et al., 2006). The International Criminal Court’s approach to the Dominic Ongwen case, for example, demonstrates how Western-oriented processes can operate outside the demands of local communities, which may prioritise restorative justice and social healing over rigid, top-down models of individual criminal accountability (Manoba and Sehmi, 2021). Ignoring local customs can stigmatise communities and hamper their expectation that local populations should resist militia involvement, and may judge children’s choices harshly (Tabak, 2014, 140). Programmes grounded on victimhood can, thusly, clash with community norms, delegitimising traditional restoration rituals and exacerbating mistrust (Özerdem and Podder, 2011). In short, a discourse designed to protect children can unintentionally harm the social fabrics necessary for their sustainable reintegration.

Inconsistent and biased recognition of agency

A third problem is the inconsistent and biased recognition of agency across different groups of children. These inconsistencies manifest as normative contradictions regarding when children’s agency is acknowledged, and cultural biases judging children’s choices against predetermined conceptions of ‘childhood’. Both fundamentally undermine the coherence of the children’s rights discourse and expose its underlying ideological foundations.

International law affirms children’s “evolving capacities” (UNGA, 1989, Art. 5), recognising their competence in decisions about health, education, and forms of political participation. Agency is often encouraged and praised when children engage in ‘safe and peaceful’ movements like environmental activism, climate change litigation, or gun control reform (Save the Children, 2025). Yet, this recognition is abruptly suspended the moment a child’s choice involves morally ambiguous expressions of agency (Bordonaro and Payne, 2012). When children pursue “undesirable” practices (e.g., working, marrying, joining armed groups), their potential choice is instantly denied; children acting contrary to normative expectations are deemed incompetent, their agency retroactively invalidated (Hanson, 2011, 2016: 471). This selective logic is stark in transitional justice. In Sierra Leone, the TRC required children to “volunteer” to testify as victims or witnesses of conflict, presuming agency for truth-telling (Drumbl, 2012: 184). Crucially, the same volunteering victim or witness children were deemed incapable speaking as ex-combatants. Similarly, the discourse frequently celebrates the agency of children who voluntarily demobilise or desert armed groups, despite evidence that escape is often a far riskier and more conceptually complex choice than the initial decision to enlist (Denov, 2012). That such life-threatening resistance is championed as agentic while the often pragmatic survival tactic of enlistment is dismissed as non-volitional exposes the inconsistency of these judgements. This reveals that agency is affirmed when it serves dominant narratives, but denied when it complicates the moral clarity of victimhood or dominant constructions of ‘childhood’.

Layered onto this is a pronounced minority-world bias. Conceptions of ‘childhood’ as a period of innocence and dependence are historically and culturally contingent, yet international law and dominant imaginaries attempt to universalise this model (Wells, 2015, 175). The result is a striking asymmetry: Western programmes, like the U.S. Junior Reserve Officers’ Training Corps or British cadet schemes, are often praised for fostering discipline and civic responsibility in minors (Tabak, 2025). Structurally analogous enlistment in African or Middle Eastern contexts, however, are immediately cast as inherently coercive and pathological (Happold, 2008, 70; Beier and Tabak, 2020, 286). The OPAC further entrenches this hierarchy by permitting voluntary recruitment under 18 in state armed forces while criminalising the same act for non-state armed groups (Drumbl, 2012: 78). This double standard legitimises dominant military discourses while pathologising potentially legitimate armed struggles, thus, reinforcing established systems of authority and narratives of ‘civilised’ versus ‘lawless’ causes. Taken together, the normative contradictions and cultural bias reveal the selective, instrumental nature of children’s agency. Agency is not an inherently recognised but is conferred or denied based on whether a child’s choices comply with predefined, predominantly Western, normative ideals. This exposes the narrative not as a neutral protective shield, but as an ideological tool that reinforces existing hierarchies and silences complex realities.

The three problems raised above converge on a central claim: the passive victimhood discourse is untenable. By homogenising experiences, it erases agency and diversity. By imposing arbitrary distinctions, it generates structural unfairness. And by applying principles selectively, it reveals ideological and cultural biases. Instead of wishing away children’s agency, we need frameworks that both acknowledge their active agency and recognise the powerful structural constraints framing their choices. I propose adaptive agency as a conceptual tool to achieve this.

Reframing child soldier’s agency through adaption

Having demonstrated the limitations of the dominant victimhood narrative, how can we make sense of child soldiers’ agency under severe constraint? The prevailing discourse falters conceptually, treating agency as binary. To replace it, I draw on the literature on contemporary sociology of childhood and on adaptive preferences, aiming to show how agency can persist, shaped but not wholly determined, within oppressive contexts. Building on this literature, this section introduces the concept of ‘adaptive agency’ (the capacity to adapt to coercive environments while still engaging in purposeful decision-making), and makes three claims. First, theorising agency under oppression requires recognising that adaption can coexist with meaningful choice. Second, this perspective illuminates the plural strategies of agency children employ from enlistment to survival during conflict. Third, agency must be understood across time, as children’s relationship to their choices may shift during conflict, with direct implications for how responsibility and reintegration processes are approached.

Children’s agency: Tactical and adaptive

To conceptualise the agency of child soldiers without falling into the trap of an essentialist binary, where a child is either a passive victim or a fully autonomous subject, it is necessary to first explore the ontological premises of agency itself. Critical childhood studies have increasingly rejected the ‘liberal model’ of agency, which views it as an inherent property or possession of the individual (Valentine, 2011: 356; Greene and Nixon, 2020: 22). Agency is not a static trait but a dynamic process that emerges through interaction with the social world (Greene and Nixon, 2020: 26, 57-8). It is constituted by the child’s engagement with external resources (or scaffoldings) ranging from interpersonal relationships and social norms to material technologies (Osborne and Rose, 2024: 32-4; Brando, 2020: 257-8).

From this perspective, all human agency is adaptive in some form, in that it relies on continuous adjustment to the spaces and constraints of the social environment (Abebe, 2019). In this sense, the question is not whether children have agency; but how, when and where is agency exercised? (Spyrou, 2018: 122; Cavazzoni et al., 2021: 365). In the context of armed conflict, the specific constraints of the child soldier require us to reflect on what happens when this environment does not merely support or constrain agency, but when it actively threatens it; that is, when the social terrain itself becomes oppressive and volatile. To capture this dynamic, it is productive to start by turning to the concept of social navigation (Vigh, 2006, 2009). Vigh argues that in contexts of chronic crisis and conflict, the social environment cannot be viewed as a solid “landscape” upon which actors move; rather, the social environment is more akin to a “seascape”, a turbulent environment where forces, currents, and horizons are in constant flux (Vigh, 2009: 426-7). Agency in this context is defined as “motion within motion”; it is the capacity to act within a social environment that is itself moving and effecting upon the agent (Vigh, 2009: 431).

While Vigh’s concept of social navigation illuminates the turbulent nature of the terrain, Honwana’s concept of ‘tactical agency’ clarifies the specific operational logic of the child soldier within it (2006; 2009). Honwana distinguishes between ‘strategies’ and ‘tactics’ based on control over the environment. A strategy requires acting and decision-making with expansive control over their surrounding social environment (Honwana, 2009: 66). Child soldiers, however, typically lack this control and power over space; they cannot dictate the war’s trajectory or define the environment in which they operate (Vigh, 2006: Ch.6). Instead, they exercise tactical agency: calculated actions aiming at effecting some control over their life and environment. Similarly, Drumbl’s concept of “circumscribed action” (Drumbl, 2012: 98–101) suggests that children retain some ability for choice despite structural constraints and coercive conditions; while having ‘diminished capacity’ children can still make ‘the best out of a bad situation’, and should be recognised as making circumscribed choices (Drumbl, 2012: 100). These frameworks allow us to understand how children, though unable to strategically alter the scope of opportunities in conflict, might still exercise profound forms of action to navigate those available, thereby actively shaping their trajectory without controlling the terrain itself.

However, understanding the child soldier’s experience requires one further analytical step. While the concepts of social navigation, tactical agency and circumscribed action explain how children move through volatile environments, they do not fully account for the internal psychological and emotional costs and effects of inhabiting an oppressive environment that curtails options and forces reforming choices. The literature on ‘adaptive preferences’ provides valuable conceptual support to approach this challenge (Khader, 2011; Terlazzo, 2016). Adaptive preferences describe the phenomenon whereby individuals adjust their desires, aspirations and sense of self to align with the limits imposed by structural inequality and oppression (Cudd, 2015: 150). In oppressive contexts, it might be easy to assume that these types of preferences are pathological and fully non-agentic due to them often being the product of both internal and external constraints on action (Khader, 2011: 20). Internally, individuals may come to believe that certain harms are natural or inevitable. Externally, the available range of actions available is sharply circumscribed by material deprivation, violence, or coercive power structures (Cudd, 2015). These dynamics align precisely with the constraints emphasised by the victimhood discourse, which rightly highlights the effect of structural and psychological pressure on informed choice (see Section 'Discourse on passive victimhood').

The adaptive preferences framework provides a way to account for these constraints and internal shifts, while recognising that agency can persist in adapted forms in oppressive environments. It allows us to acknowledge the weight of coercion on choice without dismissing agency. Understanding child soldiers’ agency as ‘adaptive’ is a necessary conceptual refinement because, while current childhood studies research tends to agree that agency is always situated and relational, their specific predicaments involve more than navigating a fluctuating and constraining environment; it implies profound restructuring of the internal self. The concept of adaptive preferences isolates the mechanisms by which individuals in severely oppressive contexts reduce cognitive dissonance by re-aligning their own aspirations and values within the limited options available to them (Khader, 2011: 122). Agency persists not through mastery of the external world, but through a radical modification of the internal self; a ‘making do’ that can involve adapting to the values of the oppressor or finding utility in violence to preserve a fragile sense of control.

Understanding agency as adaptive sharpens the analytical lens by integrating the internal dynamics of preference-formation with the external influence of volatile environments. Adaptive agency aims to explain this situated form of agency as both effecting on the environment and on the individual, refining our understanding of agency in three distinct ways. First, it emphasises the tactical aspect of agency, explaining how children, lacking strategic control, utilise tactics not to change the warscape, but to secure immediate survival within it; such as a child choosing an armed group perceived as best placed to ensure their survival, even when non-military options seem viable. Second, it foregrounds the moral complexity of action during conflict; recognising that children adapt their internal preferences to oppressive conditions, this framework allows us to view participation in armed groups as a functional choice that may nonetheless entail ‘self-destructive’ or morally injurious outcomes (Bordonaro and Payne, 2012). Finally, it incorporates a temporal dimension, positing that agency is not a fixed attribute but a fluid capacity that mutates diachronically as the child navigates the changing currents of the conflict (more below). It recognises that a child coerced into enlistment might later come to endorse their role, or a volunteer might grow disillusioned and seek to resist from within. In each case, adaptive agency captures how agency evolves alongside changing circumstances, capacities, and self-understandings.

Reframing child soldiers’ agency as adaptive avoids two dominant distortions: first, it resists evaluating agency based on idealised conditions that overlook the structural and psychological constraints shaping choices. Second, it challenges the erasure of agency in the victimhood discourse by recognising that, even in hostile settings, children act strategically based on their own goals. By situating adaptive agency within the continuum of relational, context-dependent action, we gain a conceptual tool that can capture the complex realities of involvement in armed conflict and ground more nuanced assessments.

Adaptive agency and pathways of enlistment (ante bellum)

This subsection examines how adaptive agency, the capacity to recalibrate preferences and pursue instrumental ends within oppressive constraints, operates at the point of enlistment. Rather than treating recruitment as a simple forced/voluntary binary, the evidence shows a plurality of ante bellum pathways and context-sensitive tactical moves by which children respond to constrained choices (Drumbl, 2012). Adaptive agency, as an analytic device, highlights that children often act with purposive intent even when options are severely circumscribed, thereby resisting both romanticising autonomy and reducing all action to passive victimhood.

It is useful to begin with a typology of enlistment pathways. Five types recur across the literature: forcible abduction, conscription, material/economic association, political or ideological recruitment, and familial or community continuity. Each category differs in the degree of immediate coercion and the scope for tactical choice available. Empirical studies consistently underscore this plurality, showing significant variation in self-reported voluntary versus forced recruitment across conflict zones (see Section ‘Three problems with passive victimhood', above). Recognising this heterogeneity is the first step toward assessing adaptive agency accurately. A 9-year-old forcibly abducted is in a different position from a 15-year-old who enlists to secure food for dependent siblings or with aspirations of wealth and power; their room for manoeuvre, scope of agency, and reintegration needs after demobilisation differ (Drumbl and Tobin, 2019a).

While pathways for enlistment are diverse, choices are intelligible only against the background of social, economic, and political structures. Agency is embedded and adapts in relation to these external circumstances, which function as push and pull factors affecting children’s choices (Gates, 2011). Economic insecurity is often decisive. Armed group membership offers quicker material returns (food, pay, looting) than civilian alternatives, making enlistment a rational (if tragic) livelihood strategy (Vautravers, 2019). For many youth, enlistment serves as “both a meal-ticket and substitute education” (Peters and Richards, 1998: 187), offering immediate resources and social recognition. This is not idealised preference formation; it is a pragmatic calculus made to adapt to difficult circumstances.

Second, security and household escape motivate many. Children from abusive or unsafe homes view armed groups as offering protection, functioning as an armed form of emancipation (Drumbl, 2012: 70). Many of the peacekeepers interviewed by Johnson (2023:66) highlighted children frequently considering “escaping gender-based violence” and abuse at home as decisive factors in their enlistment (Johnson, 2023: 66); this is supported in research in Northern Uganda where children stated that joining was a survival strategy to escape abusive households, finding relative safety in the bush (Annan et al., 2006). In these instances, the armed group is perceived as a mechanism for liberation from local patriarchal structures of dominance (Denov, 2012: 284-5). Third, political, economic and cultural logics structure incentives. Where elders monopolise resources, militancy offers a route to adulthood and respect; family histories or liberation narratives also channel choices (Baines, 2009). These drivers are, moreover, gendered: girls may enlist to avoid forced marriage, gain mobility, or pursue political engagement; boys often seek status, revenge, or economic gain (Denov, 2010; Hauge, 2011). Any account of adaptive agency must treat gender as constitutive of the feasible strategy set.

Adaptive agency is especially visible in the micro-strategies children deploy at enlistment, documenting a repertoire of tactical moves intelligible as instrumental responses to constrained option sets. A couple of examples might shed light on how adaptive agency can function. A recurring life-history pattern documented in West African case studies foregrounds the material logic that underlies many enlistment decisions (Utas, 2005a). In these accounts, a young person from a chronically impoverished community, with schooling interrupted, little access to steady work, and often a household that cannot guarantee food or protection, describes the decision to join an armed group not as an embrace of violence but as a pragmatic, tactical, response to immediate scarcity (Honwana, 2009). Ethnographic interviews from Sierra Leone capture this logic succinctly: “the AK47 brings food, money, a warm bath and instant adult respect,” explaining why the immediate returns of soldiering can outweigh the remote, uncertain benefits of civilian life (Peters and Richards, 1998: 187).

Longitudinal studies reveal that decisions for enlistment often begin with a calculus of survival, securing access to food, obtaining access to mobile resources (loot, pay), and gaining social respect denied to youth in civilian life (Vautravers, 2019). In practice this calculus produces a sequence of tactical decisions: the young person chooses the unit perceived as less brutal, seeks proximate, safer roles (like porter or cook), and maintains ties by sending remittances home (Utas, 2005a). While these pragmatic choices eventually routinise into social identities, realigning preferences around combat life’s incentives, their origin remains adaptive and responsive to constrained circumstances, rather than purely ideological.

In contexts where domestic insecurity or political repression renders the home dangerous and impoverishing, another characteristic trajectory appears. Girl combatants in Central America recount joining guerrilla groups to escape threats (rape, forced marriage, state violence) but frame their participation in explicitly political terms. In Guatemala, for example, girls found protection from army terror and a site for political socialisation; they often narrate a process of political awakening and communal belonging that made participation meaningful (Hauge, 2011). These life histories begin with fleeing an intolerable domestic situation, continue with practical adaption inside the movement (learning skills/training), and culminate in a durable identification with the cause, evolving from an initially protective strategy into a reflective commitment (Huynh et al., 2015). Importantly, gender conditions these trajectories, while being contingent upon dynamics and ideologies of particular armed groups (Denov, 2010). In anti-colonial struggles such as the EOKA in Cyprus, girls’ agency was often channelled into ‘covert’ roles, such as hiding weapons or leading school protests, where they exploited colonial gender norms that viewed girls as innocent and apolitical to evade security forces (Hynd, 2021: 6). Contrastingly, in the Liberian civil war, groups like LURD or the NPFL provided a radically different scope for agency, where young women could transcend the vulnerability of the ‘bush wife’ to become high-ranking commanders like ‘Black Diamond,’ securing status and wealth unavailable to them in peacetime society (Utas, 2005b). A girl’s adaptive pathways, whether she navigates through the ‘girlfriending’ tactics of the weak or takes up dominant roles, is fundamentally structured by the specific opportunities and constraints afforded by the group she is associated with (Utas, 2005b: 425-6).

The examples above highlight a central claim: agency is inherently relational, not only in responding to complex environments but also linked to social ties and dependencies. Children do not confront a neutral menu of abstract options; their possibilities are mediated by families, peers, recruiters, and commanders who function as brokers of opportunity. Families may broker enlistment (sending a son for bridewealth or arranging a daughter’s placement) treating armed association as a household risk-management strategy (Drumbl, 2012). These dynamics are, moreover, gendered. Female kin may broker daughters’ placements to avoid forced marriages or worse domestic fates, instrumentally framing enlistment as protective (Denov, 2010; Tabak, 2020). However, commanders may exploit such placements, offering protection in exchange for domestic or sexual labour. While peers recruit boys into combat roles promising loot and status, girls are steered toward auxiliary roles that may nevertheless provide schooling or vocational prospects. Understanding enlistment therefore demands examining the social architectures that enable tactical options, as it is through these networks that adaptive agency acquires concrete shape and moral meaning.

Research on enlistment pathways supports three claims. First, recruitment is heterogeneous, reflecting varied pathways and tactical room available to children (Lazareva, 2011). Second, within these constraints, children routinely deploy a repertoire of instrumental moves (selection, bargaining, escape, and hedging) that instantiate adaptive agency (Honwana, 2006). Third, social networks and gender critically shape both constraints and tactics (Veale, 2005). Read through the lens of adaptive agency, children act tactically under constraint, requiring context-sensitive interpretation.

Adaptive agency and diachronic choices (In Bello)

If adaptive agency explains decisions at enlistment, its true analytical power emerges when we follow it diachronically; that is in how agency develops and evolves through time. The relationship binding a child to an armed group is never static; preferences and identities evolve by in bello structures and practices, not only due to the diversity of children’s own experiences, but also conditioned by the evolving nature of armed groups throughout conflict. In this sense, Terlazzo (2021) argues that agency must be assessed diachronically, not as a fixed snapshot during the initial choice (i.e., enlistment), but across time. Even if the enlistment choice was shaped by oppressive constraint, the relationship to that choice can evolve over time through reflection, endorsement, or transformation. This diachronic approach provides a crucial conceptual tool for assessing the agency of child soldiers: it refuses to fix their moral status at the point of recruitment. Instead, it asks how prolonged in bello exposure to group dynamics, material incentives, and the habituation to violence, alters both children’s preferences and their capacity to assume, repudiate, or renegotiate their roles (Boyden and De Berry, 2004: xvii).

Armed groups are institutions that actively reshape children’s social and psychological worlds. Commanders secure compliance through a careful mix of coercion and incentives (Gates, 2011: 30–31): while beatings and the omnipresent threat of violence enforce discipline, selective rewards, like food, pay, or the promise of promotion make membership both materially appealing and, at times, practically inescapable (Pugel, 2009). Simultaneously, the distribution of roles and ranks draws a child deeper into the group’s web: becoming a messenger, guard, or commander creates new stakes for belonging, shifting their dependency and identity (Gates, 2011: 43–44).

These external structures interact profoundly with internal processes of habituation and moral reinterpretation. Children repeatedly exposed to violence often become desensitised, where the initially unimaginable becomes routinised (Wood, 2008: 546). To reduce cognitive dissonance, they may adopt narratives reframing their actions as necessary or aligned with a higher cause. Emotional bonds to peers and “bush families” generate loyalties that transcend purely coercive retention (Denov, 2010). For many girls, forced marriage or assignment to domestic roles, create distinct forms of belonging, or sometimes, a sense of opportunity (McKay and Mazurana, 2004). These internal shifts in preferences are not pathological capitulations but instances of meaning-making under duress, mediating whether adaption hardens into endorsement or remains a fragile survival strategy.

The interplay of external incentives and internal transformations produces diverse trajectories, where some children move from coerced enlistment into full group identification. Entrenchment often follows prolonged exposure, ritualised violence, or access to authority. The case of Dominic Ongwen clearly illustrates this transformation (ICC, 2021b; Talbert and Wolfendale, 2018). Abducted as a child, Ongwen rose through the LRA ranks to orchestrate the same violence once inflicted upon him. His life history encapsulates the diachronic shift: a child who begins as a victim may, over time, endorse and perpetuate acts that place him in the position of perpetrator (Baines, 2009; Hasselind, 2021). There are also accounts where enlistment leads to a perceived sense of empowerment and self-worth. In Veale’s study of girls in the TPLF, participants emphasised how military life made them more confident, assertive, and politically aware (Veale, 2005). They reported acquiring skills, education, and personal autonomy that distinguished them from civilian peers, and many experienced demobilisation not as relief, but as loss. These accounts complicate moral and psychological assessments: not all adaption to conflict is purely survivalist or psychologically deforming. Some children embrace roles, derive meaning from participation, and even pursue forms of justice through violent struggle. While these experiences are shaped by coercive structures, they are not reducible to passive victimhood.

For many, persistence is less about conviction than pragmatism. Longitudinal studies in Sierra Leone and Uganda show children often remain for the steady provision of food, security, or family support (Annan et al., 2006). Pragmatic adaption can reflect constrained choice rather than deep endorsement of the group’s values. Still others carve out spaces of tactical resistance: they comply outwardly while covertly sabotaging operations or shielding peers, exemplifying adaptive agency that mitigates harm within oppressive systems (Brett and Specht, 2004: 83). At the opposite pole, some experience ruptures that trigger outright rejection of the group. These fractures, catalysed by atrocities, safe exit opportunities, or accumulated moral disillusionment, demonstrate that choices are reversible and that submission is not inevitable (Annan et al., 2006).

In bello trajectories are rarely linear. Many children oscillate between persistence, resistance, and endorsement across time (Veale and Stavrou, 2007). To treat a single moment as definitive risks distorting the moral and empirical picture. Conceptually, this diachronic lens compels a rethink of moral judgment. First, later endorsement of violent roles does not erase the coercion of initial recruitment. A child like Ongwen can be both a victim and morally responsible agent (Talbert and Wolfendale, 2018); diachronic endorsement can alter the meaning of present choices without retroactively annulling past victimisation (Terlazzo, 2021). Second, moral assessment must be graded and temporally sensitive. Agency and responsibility lie on a continuum, not in rigid binaries (Smeulers, 2008; Drumbl, 2012: 69). Agency is not frozen at point of enlistment but adapts and is constantly reshaped during conflict. Children’s identities, preferences, and choices evolve, complicating both the narrative of passive victimhood and the opposite myth of unencumbered agency. A diachronic analysis captures this shifting terrain and illuminates the difficult truth: the same person can inhabit multiple moral positions across time, from coerced child-victim to active participant, and sometimes both at once. Dominic Ongwen’s life stands as a stark reminder that only a temporally sensitive account of adaptive agency can do justice to such complexity.

Conclusion

This article demonstrated that the dominant passive victimhood discourse is morally and analytically untenable, failing to capture the complexity and heterogeneity of children’s involvement in armed conflict. To overcome this, it advances the concept of adaptive agency: the capacity of children to navigate oppressive structures by strategically recalibrating preferences and choices to preserve self-direction. Its analysis, applied to both ante bellum and in bello experiences, shows that agency is not binary and static, but a dynamic, adaptive process. The diachronic lens is essential, revealing that a child can simultaneously inhabit the moral positions of coerced victim and active agent. Protecting children in conflict requires more nuanced categories. It requires recognising their evolving capacities, the diversity of their adaptive strategies, and the moral complexity of their trajectories. By embracing adaptive agency as a conceptual lens, we can move beyond the distortions of the victimhood discourse and build reintegration and justice processes that are empirically grounded, ethically defensible, and true to the lived realities of child soldiering.