Definitions of Child Recruitment and Use in Armed Conflict: Challenges for Early Warning

ABSTRACT

The recruitment and use of children during armed violence is a serious concern in contemporary conflicts, and early warning systems (EWS) can help prevent, not just ameliorate, the resulting harm to children. Effective EWS need to consider recruitment patterns, children’s experiences, relevant legal and political definitions, and data sources. This paper engages with complexities raised by the interaction of legal, experiential, and social scientific dimensions of the recruitment and use of children through analysing the humanitarian and legal definitions of the recruitment of children in conflicts, datasets on armed conflict, and the issue of large-scale violence by organised criminal groups.

Introduction

Improving the protection of children during armed violence, and in particular preventing them being recruited by armed groups or forces, stands to benefit from the use of early warning approaches (Baillie Abidi and Cleave this volume, Lynam et al. this volume). An early warning system (EWS) seeks to monitor indicators that may proceed a negative event such as a natural disaster or armed conflict and provide advance warning to relevant policymakers, practitioners, and organisations so that they can take preventative action (Matveeva 2006, United Nations 2016). EWS tend to use a combination of quantitative and qualitative indicators to produce and verify predictions, and need to be coupled with appropriate theoretical framework and analysis procedures (Naumann et al. 2014). While some EWS only focus on providing warnings, they are more likely to be effective with established communication and action pathways with the relevant stakeholders to ensure that early warnings are acted upon in a timely manner (Wulf and Debiel 2009). At present, there is a lack of consideration of children in conflict-related EWS, and none focused specifically on child soldiering (Cleave and Watkins 2020). As Lynam et al. and Baillie Abidi and Cleave discuss in this issue, such an EWS could greatly improve the protection of children during armed conflict. To be effective, it will need to engage with both the production of knowledge about child soldiering, and the legal and political dynamics involved in responding to this issue, including around how child soldiering is defined.

While it may seem obvious what the term ‘child soldier’ refers to a range of different terms and definitions are used in different fields and by different actors, and interpretations vary based on underlying assumptions and understandings. Early warning-informed prevention of the recruitment and use of children must address these definitions and understandings at various stages, from data collection and analysis through to the politics of taking action. Prevalent definitions of recruitment and use in the international legal and humanitarian fields are rooted in legal definitions of armed conflict. In these fields, two general definitions of child soldiering with different aims are used: a more legally restrictive notion of who is a child combatant under the laws of war, and a broader humanitarian definition adopted with the intention to broadly protect children who are involved with armed forces and groups in a variety of ways not limited to actual combat.

Research on children recruited and used in armed conflict, and research on armed conflict more generally, is a key ingredient of early warning systems. The primary armed conflict databases in this field, ACLED (Raleigh et al., 2010) and the UCDP GED (Sundberg and Melander 2013), use a more expansive definition of armed conflict than that found in international law. Quantitative research on the recruitment and use of children as soldiers has often relied on these datasets, and thus their definition of armed conflict, while they use the broader humanitarian definition of child recruitment and use which draws on the narrower, legal definition of armed conflict. Consequently, children involved in large-scale organised violence, such as that between criminal organisations and the government in Mexico, are considered to be child soldiers in research while the violence in question is not considered by many states or international organisations to be an armed conflict such that the laws of war apply, and hence those children do not fall under the legal definition of combatant (e.g., Haer and Böhmelt 2016). While the lived experience of these children is often similar to those children considered child soldiers in situations that classify as armed conflict, this legal difference leads to different state practice and different international responses, which need to be considered in early warning.

This paper consequently examines the intersection of these varied definitions of child recruitment and use in armed conflict, and children’s experience of armed violence, in light of the early warning process. It examines the legal and humanitarian definitions of recruitment and use and their relation to definitions of armed conflict, and contrasts these with how armed conflict and child recruitment and use are defined in research on this topic. This foregrounds the differences between knowledge production and early response that need to be considered. It then presents a case study of the experiences of children in gangs and criminal organisations as an exemplar of children whose experience resembles soldiering, yet is not included under humanitarian and legal definitions. It concludes by arguing that all these legal, humanitarian, scientific, and experiential perspectives need to be taken into account in the early warning process. In the data collection and analysis stages for generating early warnings, an expansive definition that includes children involved in a range of roles across different forms of organised violence is important for the broadest possible protection of children. However, in the communication of warning and response possibilities to those responsible for taking preventative action, attention to the legal and humanitarian definitions are crucial. This is both due to the politics of ‘naming and shaming’ around child soldiering, and the differences in protection afforded to children inside and outside of situations which are legally classified as armed conflicts.

This analysis contributes to the literature in several ways. First, it draws attention to the importance of considering definitional differences between social science research and state practice on intrastate armed conflict that may affect the practical and policy implications of conflict research. Second, it provides a clear approach to addressing the complications raised by these different definitions in different parts of the early warning process for child recruitment and use that are relevant for other early warning systems. The rest of the paper proceeds as follows: it first examines the definition of children, and the legal and humanitarian definitions of child recruitment and use and their relationship to legal definitions of armed conflict. It then examines the definitions of armed conflict and armed groups used by prominent datasets on armed conflict. It concludes by presenting the case of the experiences of children involved in gangs and criminal organisations outside the legal context of armed conflict, using this example to argue for a flexible approach to these definitions across the early warning process.

Legal and Humanitarian Definitions

The use of the phrase ‘child soldiers’ is common throughout our everyday language; from making headlines in major news sites, to a phrase commonly used at the United Nations. The phrase is used to encompass and refer to children who have been associated with, recruited by, or used by armed forces and non-state armed groups around the world. In the legal and humanitarian fields, however, terms such as child combatant, recruitment and use of children, and children associated with armed forces and armed groups are more commonly used than child soldier. Under international humanitarian law (IHL), the term combatant is used, and when applied to children indicates recruitment and use in hostilities, with the aim to protect children from being targeted unnecessarily. Humanitarian responses to child recruitment and use often use the phrase children associated with armed forces and armed groups, based on a broader definition that encompasses the range of roles children take on. This is done with the intention of not wanting to miss children who are in non-combat roles during humanitarian responses. In both cases, these definitions link back to legal definitions of an armed group or armed force, and legal definitions of armed conflict. These variations reflect the wide range of roles that children carry out for armed forces and groups that include tasks not considered to be soldiering, such as domestic tasks like cooking and cleaning. Consequently, this paper refers to all of these terms in examining their more colloquial versus specific uses. Before considering these definitions in practice and their relation to social science research on child soldiering, it is important to address the definition of children.

The legal definition of a child is different under different international treaties. A country’s definition will be subject to what treaties it has signed onto and also whether those international obligations have been reflected or adopted into domestic law. It is also important to note that the definition of childhood is socially constructed, and the legal context in a given country may not reflect culturally specific understandings of childhood. Globally, the Convention on the Rights of the Child (CRC) sets the age of adulthood at 18, and confers specific rights and protections to those under 18.¹ However, states may set a lower age than the CRC specifies,² and children and adults gain rights and lose certain protections at a range of ages above and below 18 depending on the legal and social context (such as the right to vote, age of enlistment in the armed forces, age of consent, age to hold a certain office, age of criminal responsibility, etc.). Social markers of adulthood commonly do not align with 18 and may in practice be more relevant for what children do than national or international law. Furthermore, legal definitions of childhood and accompanying rights and protections arise from socially constructed and predominantly Western understandings of childhood, which the CRC and international discourse on children’s rights frame as universal rather than contextual (Tabak 2020, Rabello de Castro 2021). While the rights protected by the CRC and most social understandings of childhood acknowledge that people under 18 exercise agency and can make informed choices about their future to some degree, the international response to child soldiering tends to explicitly ignore or deny the agency of children, while neglecting a more nuanced view of both adults’ and children’s agency (Drumbl 2012, Tabak 2020). These considerations are vital for implementation of EWS of the recruitment and use of children to improve both the protection of children in the process, and ensuring their rights to participation and a say in their societies. In this paper, we use the accepted legal definition of childhood lasting until age 18 to advocate for broader protection of children, while also noting the need for their participation.

International Humanitarian Law Definitions

Our interest in this paper is primarily how ‘child soldiers’ are defined in research and practice. As noted above, the definitions used in IHL and humanitarian approaches are rooted in definitions of armed conflict and combatant, and so it is necessary to begin there. IHL is the body of law that governs the conduct of parties to armed conflict, including both regular armed forces as well as non-state armed groups. It provides the legal framework for what means and methods combatants are permitted and not permitted to choose as well as the protection parties to the conflict must afford to those effected by the hostilities. Like all bodies of law, IHL is based upon legal definitions including who is afforded what protections, who is afforded protection from direct attack, as well as any special protections granted, for example special protections when a child is detained. The word soldier is not defined under treaty law. Instead, IHL uses words and phrases such as ‘combatants’ and ‘members of armed forces’. A combatant is the term used to refer to those who are not protected from direct attack as well as indicating that in some circumstances that person will be entitled special protections if the person ‘falls into the power of an adverse Party’.³ A combatant is a lawful military target under IHL, meaning deadly use of force may be lawfully used against them. Key to this discussion is the principle of distinction. This legal principle requires parties to conflict to distinguish between combatants and civilians and between military and civilian objects. Parties to conflict can only lawfully direct attacks against those who are not civilians, or against military objectives.⁴ IHL defines civilian as those who do not belong to an armed force.⁵ Article 50 of Additional Protocol I (API) includes further clarification that ‘[i]n case of doubt whether a person is a civilian, that person shall be considered to be a civilian’. IHL states that civilians must be protected against direct attack, but can lose protection against attack, in other words, they can become lawful military targets, when they directly participate in hostilities.⁶

It is a difficult process to confirm if a person is a lawful target. One can attempt to determine if they are a member of a particular group,⁷ or if the person does not belong to a protected category of person,⁸ or if the person is a civilian directly participating in hostilities. However, there is no overarching agreement regarding what language to use to refer to that person as beyond being a lawful target.

Children are accorded special protection under IHL. In API, article 77(1), ‘Children shall be the object of special respect … . The Parties to the conflict shall provide them with the care and aid they require, whether because of their age or for any other reason’. IHL treaty law clearly states that children are entitled to fundamental guarantees during armed conflict including humane treatment, the provision of and access to medical care, as well as being entitled to procedural fairness.

The law also specifically prohibits the recruitment and the use of children in direct participation in hostilities by both state and non-state armed forces. Additional Protocols I (article 77(2)) and II (article 4(3)(c)) and the Statute of the International Criminal Court (the Rome Statute) all use the minimum age for the participation in hostilities as 15, as does the Convention on the Rights of the Child in article 38 (2–4).⁹ Other international law raises that age to 18. For example, the African Charter on the Rights and Welfare of the Child mandates that States shall take all necessary measures to ensure that no child shall take a direct part in hostilities and refrain in particular, from recruiting any child which is defined at the outset of the Charter as those under the age of 18.¹⁰

IHL addresses certain circumstances and attempts to promote protection. Children as a whole receive special protections under IHL as mentioned previously, including and specifically when they are detained. Building on this treaty law, signatories to the Rome Statute are bound by the legal notion that using children under the age of 15 to ‘participate actively in hostilities’ is a war crime in both international and non-international armed conflicts.¹¹ The Rome Statute’s definition of this war crime leads us to the phrase ‘direct participation in hostilities’ (DPH). As mentioned previously, when a civilian regardless of age directly participates in hostilities and becomes a combatant, the civilian loses protection they have against the dangers arising from military operations (Sassòli et al. 2011) including direct attack and can become a lawful target. To be clear, this applies to children regardless of whether they are forced to directly participate or not.

Governments, legal scholars and practicing IHL lawyers around the world have different interpretations of what it means to directly participate in hostilities because the law is remarkably silent on it. AP1, article 51 (3) simply states that ‘Civilians shall enjoy protection … unless and for such time as they take direct part in hostilities’. Common article 3 to the four Geneva Conventions use the phrase ‘[p]ersons taking no active part in the hostilities’.¹² Case law under different jurisdictions have made different findings regarding this phrase. The ICRC released an interpretive guidance outlining its interpretation. Countries share or hint at their interpretations in their military manuals or in their statements at international and domestic fora. Unfortunately, like many situations which require legal interpretation, there is a patchwork of sources which we must turn to in order to piece together what we understand to be States’ definitions.

In the view of the ICRC, to directly participate in hostilities requires that a given act: adversely affect the military operations or military capacity of a party to an armed conflict or, alternatively, to inflict death, injury, or destruction on persons or objects protected against direct attack; have a direct causal link between the act and the harm likely to result either from that act, or from a coordinated military operation of which that act constitutes an integral part; and must be specifically designed to directly cause the required threshold of harm in support of a party to the conflict and to the detriment of another (Melzer 2009).

IHL does contemplate children becoming soldiers and the drafters took measures to ensure that children who take a direct part in hostilities continue to benefit from special protection when captured or detained.¹³ The ICRC’s study on practice from around the world and customary law revealed that ‘none of the practice supporting the prohibition of the participation of children in hostilities provides that they should be deprived of their special protection if they do participate in hostilities’ (Henckaerts and Doswald-Beck 2009, p. 487).

Because of a general lack of clarity and transparency as well as the broad scope of interpretation allowed by States, a civilian’s chance of being perceived as directly participating in hostilities and losing their protection from attack varies from situation to situation. Due to the extreme consequences of a finding of directly participating in hostilities, which means a person loses protection against direct attack and can therefore be killed in the first instance (so long as the other rules of IHL are complied with), making this determination correctly is of utmost important. By directly participating in hostilities a person becomes a combatant or in the vernacular ‘a soldier’. Does a child helping to carry wood to an armed force’s forward operating base qualify as directly participating in hostilities? Does a child relaying information of troop movements qualify as directly participating in hostilities? What actions turn a child from a civilian to a combatant? What actions turn a child from a child to a child soldier?

When drafting the Rome Statute, the drafters grappled with this issue of definitions. In a report on the Draft Statute of the International Criminal Court, drafters stated that the words ‘using’ and ‘participate’ were adopted in order to:

cover both direct participation in combat and also active participation in military activities linked to combat such as scouting, spying, sabotage and the use of children as decoys, couriers or at military checkpoints. It would not cover activities clearly unrelated to the hostilities such as food deliveries to an airbase or the use of domestic staff in an officer’s married accommodation. However, use of children in a direct support function such as acting as bearers to take supplies to the front line, or activities at the front line itself, would be included within the terminology (United Nations 1998, p. 21).

Consequently, under IHL, the definition of child recruitment and use is restricted to better protect children from becoming lawful targets, even when they may be contributing in some way to the operations of an armed group. When considering what forms of recruitment and use are illegal, there are contextual variations between how children are used, who is recruiting them, and what laws apply that may set the age of legal participation in armed conflict at either 15 or 18.

Furthermore, in practice, it is important to consider how the socially constructed nature of childhood impacts how soldiers apply IHL. In many cases, it can be difficult if not impossible to actually determine the age of a person before a decision is made to use force, detain someone, or take other potentially harmful action. Gender, race, ethnicity, religion, political views, and other factors intersect with age to influence both how childhood is constructed, and how it is perceived. These factors can lead to the denial of some children’s childhood and thus the removal of the protections they are owed (Shalhoub-Kevorkian 2019). It is vital that practitioners understand and consider such politics of childhood in early warning responses.

The other primary definition of recruitment and use comes from the 2007 Paris Principles, which are focused on protecting children from recruitment and use, and ensuring that they are properly demobilised from armed groups and reintegrated into civilian society. It uses the term ‘children associated with armed forces and armed groups’ in order to indicate that children are involved with parties to a conflict, but not necessarily in ways that make them targetable. It defines them as:

any person below 18 years of age who is or who has been recruited or used by an armed force or armed group in any capacity, including but not limited to children, boys and girls, used as fighters, cooks, porters, messengers, spies or for sexual purposes. It does not only refer to a child who is taking or has taken a direct part in hostilities (UNICEF 2007, para. 2.1).

This definition is intended to be broad, so that narrower conceptions of who counts as a child soldier, such as being limited to boys who are armed, does not limit who can receive access to humanitarian support during armed conflict. In particular, it helps to ensure that girls are not neglected in the demobilisation process. This definition though hinges on children being recruited or used by an armed force or armed group. While the definition of an armed force may be clear enough, it references the definition of an armed group in the Optional Protocol (UNICEF 2007, para. 2.3). The Optional Protocol, however, does not provide any more clarity on this matter,¹⁴ and so reference must be made to international humanitarian law and jurisprudence in this matter.

When discussing armed conflict, we refer to parties to a conflict as being a state or a non-state actor. The non-state actor can be any group that is not a state actor and which has several characteristics including some type of command structure, sufficient organisation and capacity to execute military operations within an armed conflict, and must be participating in the fighting of the armed conflict itself among other criteria.¹⁵ Ultimately, a group must be able to implement IHL.¹⁶ IHL speaks to both of these types of groups and gives them certain permissions and instruction as to what it is permitted for each during an armed conflict. These permissions and instructions are more often much less protective than what is given to the individuals under International Human Rights Law (IHRL). The result is that the legal framework afforded to children associated with gangs, not involved in an armed conflict, is more protective even though the reality for these children can be as brutal and violent as situations governed by IHL.

While the existence of an international armed conflict is defined legally as ‘resort to armed force between States’,¹⁷ non-international armed conflicts (NIAC) must meet two criteria to exist: one, the violence has reached a minimum threshold of intensity; and two, the groups involved have reached the threshold of organisation to engage in an armed conflict as a party to the conflict.¹⁸ As legal definitions of armed group and child recruitment, and the rights children engaged in violence have, depend on the legal existence of an armed conflict, this is a final piece of the legal and humanitarian definition that is important to address. Specific definitions are derived from the application of Common Article 3 of the Geneva Conventions of 1949, and Article 1 of Additional Protocol 2. Article 1 of AP2 has a more restrictive definition that requires the state to be one party to the conflict, rather than a conflict among non-state armed groups, and that the non-state armed group involved have some degree of territorial control. However, this restriction only matters to specific application of AP2 and not IHL in general (ICRC 2008).

The commonality between both of these legal and humanitarian definitions of the recruitment and use of children, then, is their reliance on the definition of an armed group, and the existence of an armed conflict. Consequently, both exclude the recruitment and use of children in violence that falls outside the scope of an armed conflict, aside from the recruitment and use by state armed forces. These definitions further ultimately depend on the definition of a child. Many legal, humanitarian, and other practitioners all operate under the CRC definition of under 18, or the relevant national law if different, or in some cases differentiate between the differing laws on recruitment and use at 15 versus 18. Due to the protections this imparts to children, it is reasonable to continue to focus on all children under 18 regardless of whether their recruitment is legal. However, early warning researchers and practitioners should be cognisant of how the social construction and politics of childhood influences the ways in which children can be protected, or participate in their own protection. Next, this paper considers definitions of armed conflict and armed groups used by armed conflict datasets, and the experiences of children recruited and used by organised criminal groups or gangs, to demonstrate the importance of taking legal, humanitarian, and social scientific understandings of conflicts, and the lived experiences of children, into account in early warning systems.

Social Science Research on Child Soldiering, Armed Groups, and Armed Conflict

Researchers from a broad range of fields have produced a rich literature on child soldiering over the past several decades, investigating the prevalence and causes of child soldiering, children’s experiences of recruitment and soldiering, and the mental, physical, and social consequences on them.¹⁹ While much of this literature is relevant to the development and implementation of an EWS of the recruitment and use of children, quantitative studies focused on broad patterns of child soldiering are particularly important for modelling underlying predictive forms of EW. Many of these studies draw on conflict data from the Uppsala Conflict Data Program (UCDP) Georeferenced Event Dataset (GED) (Sundberg and Melander 2013) or its predecessors. These datasets record individual conflict events, involving various state and non-state actors, geolocated to a specific place, and recording as having taken place on a specific day. These datasets provide great flexibility to the researcher to investigate various questions, they can be combined with other temporal or spatial datasets, or serve as a basis for generating new datasets.

The production on conflict data in the UCDP GED relies on social scientific definitions of armed conflict and armed group that are similar, but not identical, to the definitions discussed previously in IHL. This is important to consider in early warning systems as research based on UCDP data, or from similar data programs like ACLED (Raleigh et al. 2010), have a broader definition of armed conflict than that found in IHL. This means that the communication of early warnings and the design of interventions need to consider the legal and political status of a situation of organised violence, and how the application of IHL and IHRL influences state practice with respect to protecting children. This section then analyses key sources on the methodology for the UCDP GED, drawing on its codebook and the journal articles introducing this dataset, as well as the UCDP Non-State Conflict Dataset (Sundberg et al. 2012).

The UCDP GED dataset has global coverage from 1989 to the present (Högbladh 2020, p. 3). It records incidents ‘where armed force was used by an organised actor against another organized actor, or against civilians, resulting in at least 1 direct death at a specific location and a specific date’ (Högbladh 2020, p. 4). Organised actors are broken down into three types:

• The government, which is defined as ‘[t]he party controlling the capital of a state’ (Högbladh 2020, p. 4),

• A ‘formally organized group’ which is defined as ‘[a]ny non-governmental group of people having announced a name for their group and using armed force against a government (state-based), another similarly formalized group (non-state conflict) or unorganized civilians (one-sided violence). The focus is on armed conflict involving consciously conducted and planned political campaigns rather than spontaneous violence.’ (Högbladh 2020, p. 4)

• And an informally organised group, which is defined as ‘[a]ny group without an announced name, but which uses armed force against another similarly organized group (non-state conflict), where the violent activity indicates a clear pattern of violent incidents that are connected and in which both groups use armed force against the other’ (Högbladh 2020, p. 4).

Inclusion relies in part on the intensity of a conflict: for organised actors to be included in the UCDP GED, they have to have participated in a conflict that caused at least 25 fatalities over the course of a year. As noted above, inclusion also relies on the level of organisation of the actor, and that a violent event can be attributed to their actions. Both of these criteria are for theoretical reasons, to provide consistency across the dataset to ensure that it is as uniformly as possible reflecting only armed conflict of a certain level of intensity and organisation, given the limits and biases from the data collection methods (Sundberg and Melander 2013, pp. 526–527).

In an earlier paper on the related UCDP dataset on non-state conflict, Sundberg et al. (2012) discuss considerations for how non-state armed groups should be defined. Their previous dataset on state-based conflict had focused on incompatibilities between states or their political objectives, but it was decided that applying a similar framework looking for incompatibility or dispute over governance or territory would be subjective and biased. Consequently, they instead focused on how well organised the actor is, though compared to the GED they separate armed non-state actors into well-organised armed groups, political parties that engage in violence, and identity-based groups (Sundberg et al. 2012, p. 353). These discussions do not explicitly reference relevant international law on armed conflict and group definition. Thus, in UCDP data there is some complementarity to the requirements of organisation and intensity under IHL for the existence of a non-international armed conflict. However, not all events recorded in this database are likely to be part of a conflict that meets the legal definition of a NIAC.

While research on the recruitment and use of children has not drawn so far as heavily on georeferenced event data, previous studies on child soldiers have made extensive use of other UCDP-based data. A number of papers draw on the Non-State Actor dataset (Cunningham et al. 2013) which draws on UCDP data, including Bakaki and Hinkkainen (2016), Faulkner et al. (2019), Haer and Böhmelt (2018), and Lasley and Thyne (2015). Other papers draw directly on UCDP data, including Haer and Böhmelt (2016), Haer and Böhmelt (2017), Mehrl (2021), Østby et al. (2022) (which used the GED data), and Tynes and Early (2015). These papers largely used UCDP data to identify countries, conflicts and/or armed groups to be investigated for whether or not they recruit and use children. While ACLED data has so far been minimally used in published studies on child soldiering, it has similar approaches to armed conflict and armed group definition to UCDP (see Raleigh et al. 2010, ACLED 2019, n.d.).

Consequently, much of the existing quantitative research on child soldiering primarily uses UCDP definitions of armed groups and armed conflict. Many of these papers refer specifically to the Paris Principles for their definition of who is a child recruited and used by an armed force or armed group, and draw on data sources including reports by human rights groups, which also tend to use legal or humanitarian definitions. Thus, our current understanding of child soldiering in quantitative research, at the multi-country scale, primarily relies on the broader definition of child soldiering that is shared by the humanitarian field, and a broad definition of armed conflict that does not necessarily reflect the definitions used by IHL or in the humanitarian community writ large. As state (and often non-state armed group) practice in these conflicts is influenced by their understanding of the legal nature of the conflict, these differences between scientific and legal definitions of armed conflict, armed groups, and recruitment and use need to be considered in the early warning process. These differences may also be relevant for some research questions about armed conflict drawing on UCDP GED or ACLED.

Next, this paper considers the case of the experiences of children involved in organised violence that are recorded in datasets such as ACLED or UCDP GED due to their level of violence, and the organisation and aims of the actors involved, but are not universally considered armed conflicts from a legal perspective.

Children and Organised Violence Involving Gangs and Criminal Groups

Children involved in organised criminal violence tend to have similar experiences of recruitment, perpetration of violence, victimisation, and constrained experiences of exercising their agency to children who are considered child soldiers. The violence they are involved in tends to be captured by armed conflict datasets like UCDP GED and ACLED, but often fall outside the scope of legal and humanitarian definitions of armed conflict and participation in hostilities. Consequently, in an early warning system focused on child recruitment and use, attention must be paid to both the scope of violence and children’s involvement in it, and legal and humanitarian issues that guide how the state and other actors respond in order to provide contextually relevant warnings that address such concerns as preventing children from being targeted, and political sensitivities around ‘naming and shaming’ of child soldier recruitment and use. This section focuses on the overlap in lived experience of children involved in these different forms of organised violence, and the scale of such violence, to demonstrate the need for their inclusion in early warning frameworks. This serves as an exemplar of the gaps and complexities raised by the previous two sections, both for early warning practice and for social scientific research on children and armed conflict.

There is a growing recognition that the literature on child soldiers can assist in developing a literature on children in organised crime (see Kerig et al., 2013). Scholars, policy makers, law enforcement agencies, and community groups all recognise the reality of children in organised crime, and the threat this presents to the safety of children in their communities as well as the stability in these areas. In the United States alone, there is an estimated 29,400 ‘youth gangs’ with 756,000 members (Kerig et al. 2013); while the number of child soldiers globally is unknown, though they have been used in most post-Cold War conflicts (Haer and Böhmelt 2016). While both children in armed groups associated with armed conflicts and children in groups usually thought of as being part of organised crime²⁰ are put in situations of violence and active danger as victims and perpetrators, international and domestic law classifies the use of children in warfare and the use of children in organised crime differently (see law section in this paper, Kerig et al. 2013). The lived reality of children has resulted in these divisions becoming more and more permeable given that the level of violence and transnational nature of organised crime resembles that of armed conflict in some locations.

Non-state armed groups and groups involved in organised crime tend to be discussed and addressed as separate categories, but they tend to fall on a spectrum of differing degrees to which they pursue political aims, hold territory, contest the government, are organised hierarchically, or pursue financial gain through criminal activity. While there are clear crossovers in levels of violence, recruitment strategies, and motives for children joining both sorts of groups, the way in which these groups are normally differentiated comes down to how we define armed conflict.

The more the actions of organised crime are analysed, the more we can see competing levels of governance within the states that criminal networks are operating within, leading to challenges to the constitutionally demarcated roles of the state (Kingston 2004). While non-state armed groups tend to operate in states with weak territorial control²¹ and are a direct challenge for the political control of the state, organised crime can also create their own spheres of influence within strong states that have experienced softened sovereignty in specific geographic regions. These states-within-states emerge in regions where ‘territorial state control has been voluntarily or involuntarily ceded in whole or in part to actors other than the relevant legally recognized sovereign authorities’ (Clunan and Trinkunas 2010, p. 17). As organised crime grows in these regions of softened sovereignty, they effectively take over as the de facto government for the area. Organised criminal groups tend to have high levels of horizontal legitimacy, as social, economic, and community groups in the region recognise (or even in the most minimal cases tolerate) their authority (Call 2012). Disley and Liddle (2016) argue that while there were significant hierarchical leadership structures in gangs, the structures were highly unstable. While this may differ from more traditional hierarchical armed groups, the majority of current armed groups are more decentralised and fluid, reflecting similar patterns of horizontal legitimacy, unstable leadership, and shifting alliances (Terry and McQuinn 2018, pp. 46–47).

As well, the intensity of violence in some locations may meet the threshold of an armed conflict. Both UCDP and ACLED record violence by criminal organisations, such as cartels in Mexico, as armed conflicts under their definitions. The Rule of Law in Armed Conflicts project, which assesses what it believes to be ongoing armed conflicts based on international law, qualifies the intensity of violence, and the organisation of actors involved, as a NIAC in Mexico between the government and two of the main cartels there, and between those two cartels (RULAC 2021). However, both UCDP and ACLED record a range of such conflicts beyond those identified by RULAC.

A Child Soldier or Children in Gangs?

The specific definition of a child associated with an armed force or armed group, or a child associated with organised crime, will depend on whether or not these forms of groups are believed to be distinctly different. The lines blur in terms of how we differentiate armed conflict and organised crime, as does the way we define children within these groups.

There are significant debates in the literature as to how we can define gang and gang membership. Strocka (2006) discusses that the literature categorises youth gangs as the ‘systematic involvement in violent and illegal activities that distinguish a youth gang from other kinds or juvenile peer groups’ (p. 134). However, for Strocka, the violence and criminal activity is not the key issue to defining gang and gang membership. Strocka argues that what establishes a gang versus another formation is the way in which society has perceived and socially constructed this category. Both a child associated with an armed force or armed group and children associated with organised crime are socially constructed categories. We can see these youth as representing a ‘problem or threat’ to a society that is otherwise stable (Strocka 2006, p. 134). However, Dowdney (2003) argues that the institutionalisation that we are seeing among gangs, as well as their quasi-militarisation is putting the children recruited into these groups outside the category of delinquency. Children in gangs, particularly in areas with drug factions, are given military grade weapons, given consistent salaries, and operate open patrols in the areas in which their gang controls (Dowdney 2003).

Children have become valuable actors within organised criminal groups. As law enforcement agencies have become more adept at identifying traditional members of criminal groups, children do not fit the traditional profile of these actors and can go unnoticed by authorities attempting to disrupt the activities of the organised criminal group (Children’s Commissioner 2019). As such, children are actively targeted for recruitment into the ranks of organised crime groups. Dowdney (2003) believes that there is significant crossover in the use of children both in and out of war zones, including in the practices of recruitment. Young children are initially recruited into these groups, and usually given non-combat/conflict roles. As they become adolescents, their role within armed groups or organised crime groups shifts, and they are armed and actively given roles that put them in situations of violence (Dowdney 2003). Children no matter the group are eventually put in ‘kill-or-be-killed’ realities, due to the orders they receive and punishments for failing to achieve their objectives (Dowdney 2003, p. 9). Research on the recruitment of children by armed groups also discusses many similar underlying factors, reasons for recruitment, and methods of recruitment.

Ward and Bakhuis (2010) conducted a research project which interviewed children in Cape Town about the influence of gangs in their community, and their views on the recruitment of children in gangs. While initiation into the gangs occurs around 12 years old, the level of involvement becomes deeper as they get close to 18. The primary reason the children argued that an individual joins a gang was due to access of material goods, an attractive reward when the context of deep poverty is taken into consideration (Ward and Bakhuis 2010, p. 54). Dowdney (2003) found that children in Rio de Janeiro join drug factions because it is their best option at employment due to the structural socio-economic marginalisation of their community. Similarly, research on recruitment of child soldiers has demonstrated both that children join armed groups in an attempt to access resources offered by armed groups, and that armed groups take advantage of poverty to entice recruits (Tynes 2019, pp. 15–17).

The motivations that lead to joining a gang can differ greatly depending on personal circumstances and environmental factors. In some cases, children can join gangs in order to feel connected within a group, or to redefine who they are (AACAP 2016). Armed groups can provide similar incentives to children, whether through group belonging or involvement in an ideological cause (Gates 2011, p. 37). The insecurity that results from an environment that is controlled by gangs can lead children to believe that the only safe options were to either hide in their houses, or voluntarily join a gang. Children may join armed groups for similar reasons, viewing them as a safer alternative (Haer 2019, p. 76).

Gang leaders use complex techniques to entice children into joining gangs and use violence to keep them compliant once a member (Children’s Commissioner 2019). These recruitment strategies are well planned: they begin with attraction to organised criminal groups through ‘gifts or praise’, in an attempt to develop friendship with children; the criminal group then provides protection to the child from a threat which could either be real or staged; the organised criminal group then provides a child with a small errand to run; finally, during one of the errand runs, the child is mugged to establish a debt to the organised criminal group (Children’s Commissioner 2019). Once children are members of gangs, it is very difficult to leave due to the levels of violence and intimidation that are used to keep them within the group. Similarly, armed groups may entice children to join through promises of future benefits like education or a job (whether or not they actually plan on providing those) (Gates 2011, pp. 34–35), and employ force, coercion, and threats both to recruit children, and prevent them from deserting or demobilising (Gates 2011, pp. 35–40).

Since the establishment of the United Nations Convention of the Rights of the Child, the social agency of children in shaping their own lives has been stressed in the literature (Atkinson-Sheppard 2017). While the concept of choice is emphasised, the UK Children’s Commissioner Anne Longfield argues that these are ‘vulnerable children that are being groomed and need protecting’ (Somerville 2019). The concept of choice presented in the literature is often binary: join a gang or hide at home; use drugs or do not use drugs; and safety or insecurity. By joining a gang, children may feel they have agency and can protect themselves and their family (AACAP 2016). The need to work with organised crime provides protective agency (Atkinson-Sheppard 2017), in which the actual agency of children is ‘thin’ due to the fact that there is no true cost-benefit analysis that can be conducted by children in vulnerable circumstances due to their need to focus on survival (Klocker 2007). Analogous debates on children recruited by armed groups assess how children navigate the perils of armed conflict as best they can, and to what extent this agency should be recognised (Denov, 2011, 2012; Drumbl 2012), while international interventions focused on child soldiering have tended to neglect children’s agency and focus on their victimhood (Tabak 2020). These similarities in children’s experiences of agency, and debates about it in the literature, further indicate the importance of considering early warning across different contexts of violence.

The argument can easily be made that the structures of these organised crime groups, the areas of loose sovereignty in which they operate in, and the levels of violence used all qualify some organised criminal groups to be seen on the same level of armed groups employing child soldiers. When looking at the recruitment of children in war, Singer (2001) differentiated child soldiers from children involved in organised criminal groups simply by adding that the violence that is being committed is ‘of a non-criminal type’ (p. 40). In this case, the differentiation was made through a statement of conventionally accepted divides: war vs. criminal violence. Putting aside the criminal nature of joining a non-state armed group, this distinction is weakened by these commonalities between the recruitment of children by non-state armed groups and criminal organisations.

Throughout the literature, one narrative that is consistently present is the agency children hold over joining groups that use violence (either in the context of war or in the context of criminal groups). Through participation in gangs, children gain protection and affirmation often lacking in their homes (Ward and Bakhuis 2010). However, serious structural constraints are influencing the level of choice that children have in joining armed groups or criminal organisations. Once a member, either as a child associated with an armed force or armed group or as a child associated with organised crime, the actions and traumas that these children endure are similar and cut across any definitions that will be drawn. These violent experiences that children endure both as victims and perpetrators connect children involved in any group that uses violence (Kerig et al. 2013). Therefore, no matter the politicisation or the legal differences that may lead to the differentiation of armed violence in the literature or in practice, all definitions need to consider the experiences of the children within those groups to ensure that early warning processes do not exclude situations where children are at risk of being recruited and used in armed violence regardless of whether that violence is considered ‘war’ or criminal violence.

Implications for Early Warning

Building on these three strands of literature, we argue for a flexible and informed application of definitions regarding child recruitment and use across the early warning process. The heart of such an early warning system will be to identify as early as possible when children may be at risk of recruitment into violence so suitable preventative action can be taken, and thus such a system should take into account the broadest relevant range of situations of concern. The existing literature on children in gangs and criminal organisations points to similarities between risk factors for their recruitment, recruitment pathways, and experiences in these organisations to children recruited by armed groups participating in more traditionally classified armed conflicts. Furthermore, the primary armed conflict data systems currently in use take a broad view of what counts as armed conflict. Consequently, a broad definition of situations of violence that includes those above and below the legal threshold of an armed conflict, and a broad definition of what counts as child recruitment and association, both in terms of roles and actors carrying out recruitment, is more appropriate for protecting the most children in the most contexts when producing warnings. This approach should be taken in the data collection and analysis stages of early warning, with if possible a global scale. However, at the same time, further research on children recruited by gangs and criminal organisations, in light of existing conflict datasets, is warranted for validating this approach.

Response to these early warnings, though, will be carried out by a range of actors with varying political interests, legal requirements, and humanitarian practices. Consequently, an early warning system for child recruitment and use cannot ignore the definitions and regulations of practice arising from IHL, IHRL, and international policy, for three main reasons. First, as noted in the legal section, in situations where IHL applies, children recruited and used by armed groups who directly participate in hostilities lose important protections from deadly force being used against them. In situations short of conflict where only IHRL applies, children involved in gangs and criminal organisations are substantially better protected. Therefore, action on early warning should avoid an approach of broadening the legal application of armed conflict, even if the data collection and analysis phase considers a situation to be an armed conflict.

Second, using the label ‘child soldier’ carries with it significant political sensitivities and legal concerns. The recruitment and use of children formed the original basis of the UN Secretary-General’s listing process for ‘naming and shaming’ armed forces and groups that commit violations of children’s rights and protections. The recruitment and use of children is considered an international emergency that signals the failure of a state to protect its population, necessitating outside intervention (Tabak 2020). Consequently, approaching an early warning from the perspective of ‘child soldiers’ may be counterproductive, triggering a defensive rather than engaging response from the state in question. As well, the term ‘child soldier’ does not align perfectly with the various definitions discussed in this paper, and the association of soldier with targetability under IHL can be problematic when advocating for better protection.

Third, the literature on gangs and organised criminal groups does highlight some differences in the context of state capacity where these groups operate, versus more ‘traditional’ armed groups. These differences may both affect the models used in early warning projections, and matter for what kind of responses states, civil society, and other actors can mobilise to prevent the recruitment and use of children in all situations of violence. Taken together, this approach shows the importance of acknowledging, understanding, and accounting for the different definitions used across fields of research and practice relevant to early warning to ensure that children in need of protection are not missed, while responses to early warning do not result in unintended harm and are properly calibrated to the actors and situations involved.

This analysis also contributes to the literature on intrastate armed conflict in two main ways. First, it highlights how a state’s legal determination on the nature of a conflict affects state practice, as application of IHL versus IHRL leads to significant differences in allowed conduct by security forces. States have clear political interests in actually applying one or the other, and what they communicate externally about this. In addition, security forces do not always follow their legal requirements. Some non-state armed groups also have interests in following, or appearing to follow, IHL in their operations. Consequently, such legal dynamics may be important for conflict actor behaviour and should not necessarily be neglected.

Second, in research on recruitment and use of children specifically, it illustrates the need for scholars to pay more precise attention to the definitions of child soldiering and children associated with armed groups that they use. For instance, a study that considers both state and non-state recruitment and use of children should take into account states’ ability to legally recruit children from ages 15 to 17, while armed groups cannot. Or, when examining child soldier recruitment and armed group legitimacy, it may be important to consider the different levels of criminality assigned to recruiting children under 15, and using children from age 15 to 17 in combat versus support roles.

Conclusion

The recruitment and use of children by armed groups and armed forces in modern conflicts constitutes a serious human rights abuse and a major protection concern; and greater attention is required to preventing and ending recruitment. An important component of more proactive and preventative responses is early warning of recruitment and use. However, current approaches to child soldiering in legal, humanitarian, and academic discourse depend on a range of definitions of armed conflict, armed group, and ‘child soldier’ that overlap but are not congruent. The case of children recruited by organised criminal groups in situations not legally considered armed conflict presents an example of children who face similar lived circumstances to child soldiers in more conventional wars, but are subject to a different legal regime and forms of state and international response. Early warning approaches to child recruitment and use are likely to primarily draw on social scientific definitions and armed conflict databases such as UCDP which take a more expansive view of armed conflict definition and armed groups than international law does. As early warning systems must engage with state, international, and humanitarian actors to prompt effective responses, an understanding of the lived reality of children in violence, the definition approaches used in underlying datasets, and legal and humanitarian norms of responding actors all need to be considered together when identifying, advocating for, and studying responses.

Acknowledgments

We would like to thank the editors and journal manager at Civil Wars, three anonymous peer reviewers, and Dr. Catherine Baillie Abidi for their support on this article which has greatly strengthened it. The views expressed in this paper are those of the authors and not their respective employers.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the Waverley House Foundation.

Notes on contributors

Michelle Legassicke

Michelle Legassicke is a PhD Candidate in Political Science at Dalhousie University. Her research focuses on civil war recurrences and looking at the ways in which post-conflict reconstruction and the reintegration of opposition groups can mitigate a return to conflict. Her SSHRC funded field work took her to Sierra Leone and Ghana in 2018 and 2019, where she also held the position of Visiting Scholar at the Kofi Annan International Peacekeeping Training Centre. At Dalhousie, she designed, taught, and coordinated classes in the Political Science and Sustainability departments. Michelle now teaches at Brentwood College School in the Social Sciences department.

Dustin Johnson

Dustin Johnson is a Research Advisor at the Dallaire Institute for Children, Peace and Security at Dalhousie University, where he currently focuses on how gender impacts child protection operations in peacekeeping, and on early warning of the recruitment and use of children in armed violence. He is a doctoral student in peace and development research at the School of Global Studies at the University of Gothenburg in Sweden. He has worked at the Dallaire Institute since May 2016 after completing his master’s degree and holds master’s and bachelor’s degrees from Dalhousie University. He has previously worked in research and NGO positions in Canada, the USA, and Mongolia.

Catherine Gribbin

Catherine Gribbin is the Senior Legal Advisor, International Humanitarian Law at the Canadian Red Cross where she has worked in a number of policy and legal positions since 2006. As Senior Legal Advisor, Catherine directs the IHL unit which is responsible for IHL dissemination across Canada as well as engaging with interlocutors on IHL implementation in Canada and around the world. From 2017 to 2018, Catherine worked in Nigeria as the Operational Legal Coordinator for the International Committee of the Red Cross (ICRC)’s Nigeria Delegation. Ms. Gribbin received her LL.B. from Dalhousie University; was called to the Bar in Ontario and Nova Scotia; and is a member of the Law Society of Ontario and the Nova Scotia Barristers’ Society.

Notes

1. Convention on the Rights of the Child of 20 November 1989, article 1.

2. Ibid.

3. Protocol Additional to the Geneva Conventions of 12 August 1949, and relating to the Protection of Victims of International Armed Conflicts (Protocol I), 8 June 1977, articles 43 and 44 (Hereafter Additional Protocol I).

4. Additional Protocol I, articles 51 and 52.

5. Additional Protocol I, article 50(1). The Geneva Conventions provided a legal regime for those who are deemed ‘protected persons’ which includes but is not limited to the ‘wounded, sick and aged persons, children under fifteen, expectant mothers and mother of children under seven’ all who must be protected ‘from the effects of war’ (Convention (IV) relative to the Protection of Civilian Persons in Time of War, 12 August 1949, article 14).

6. Additional Protocol I, article 51(2).

7. Convention (III) relative to the Treatment of Prisoners of War, 12 August 1949, article 4.

8. Additional Protocol I, article 50.

9. Listing of numerous additional authorities can be found in Rule 137 of ICRC’s customary law database.

10. African Charter on the Rights and Welfare of the Child, 1 July 1990, articles 2 and 22. Though not treaty law, United Nations Security Council resolutions continue to call for the protection of children during armed conflict which include explicit references to these legal prohibitions.

11. Rome Statute, articles 8(2)(b)(xxvi) and (e)(vii).

12. Convention (I) for the Amelioration of the Condition of the Wounded and Sick in Armed Forces in the Field, 12 August 1949, article 3.

13. Additional Protocol I, article 77(3); Protocol Additional to the Geneva Conventions of 12 August 1949, and relating to the Protection of Victims of Non-International Armed Conflicts (Protocol II), 8 June 1977, article 4(3), hereafter Additional Protocol II.

14. Optional Protocol to the Convention on the Rights of the Child on the involvement of children in armed conflict, 25 May 2000, article 4.

15. See ICTY, The Prosecutor v Duško Tadić, Trial Chamber, judgment, IT-94-1, 15 July 1999, https://www.icty.org/x/cases/tadic/acjug/en/tad-aj990715e.pdf, and ICTY, The Prosecutor v Boškoski & Tarčulovski, Trial Chamber, Judgement, IT-04-82, 10 July 2008, https://www.icty.org/x/cases/boskoski_tarculovski/tjug/en/080710.pdf.

16. For further reading, see ICTY, The Prosecutor v Ramush Haradinaj and others, Trial Chamber, Judgement, IT-04-84-T, 3 April 2008, §60, http://www.icty.org/x/cases/haradinaj/tjug/en/080403.pdf.

17. ICTY, The Prosecutor v. Duško Tadić, Decision on the Defence Motion for Interlocutory Appeal on Jurisdiction, IT-94-1-A, 2 October 1995, para. 70.

18. See ICTY, The Prosecutor v. Duško Tadić, Trial Chamber, judgment, IT-94-1, 15 July 1999, https://www.icty.org/x/cases/tadic/acjug/en/tad-aj990715e.pdf.

19. For a broad overview of the field, see Haer (2019).

20. It should be recognised that such groups are criminal organisations either way since non-state armed groups are illegal organisations.

21. Weak states are characterised by a diminished or, in some cases, non-existent administrative capacity that drastically affects civilian perceptions of the regime’s legitimacy (Kingston 2004).