Another infirmity of a Commonwealth is the immoderate greatness of a town, when it is able to furnish out of its own circuit the number and expense of a great army; as also the great number of corporations, which are as it were many lesser Commonwealths in the bowels of a greater, like worms in the entrails of a natural man (Thomas Hobbes, Leviathan, Chapter XXIX).
Yet, it is hard to say exactly what we mean by ‘warlord’. Part of the difficulty is that the term is used to describe many unlike individuals and military organisations, as Paul Jackson points out:
[that] at the beginning of the twenty-first century, the label warlord has come be used very broadly to cover a wide range of clan and political leaders who use armed civilians to impose power (Jackson, 2003:134).
If we are to accurately analyse contemporary politics, particularly in failed states, a detailed conception of warlord is necessary. By simply calling the leader of any armed group a warlord, we not only loose analytical thoroughness, but also potentially lose effectiveness in our responses to these actors. For instance, insurgencies rely on popular support whereas warlords do not. If an intervention is not able to clearly define which armed groups are reliant on popular support, then attempts to deal with a warlord in terms of his relations to civilians, such as clan elders, may be of littler help in negotiations and only serve to prolong the conflict.
Moreover, much of our understanding about warlords is one-sided; for instance, it only focuses on their economic interactions or is based on assumptions about identity, such as their savagery arises from ‘ancient hatreds’ (e.g. Berdal and Malone 2000, Kaplan 1994, Reno, 1998). Yet it is often held that warlord organisations are akin to corporations in that they are simply a means of making a profit for the warlord and his entourage. But it is not clear why warlords are so concerned with profit or how their concern with economic gains differs from that of an insurgency or nation-state. Similarly, the barbaric nature of warlord warfare is often noted, yet, there is little understanding of why warlords are so savage, except to note that they may have an ethnic grievance or are just plain ‘evil’.
Such ‘greed’ or ‘grievance’ analyses are helpful in understanding aspects of warlordism, but limit our understanding of warlords by limiting the analysis to facets of otherwise complex organisations. Social organisations, such as tribes, nation-states or insurgent groups, are multifaceted in that they have political, social, economic, and military aspects. This paper begins with the assumption that warlords and their organisations are also multifaceted.
Therefore, it is necessary to examine warlord organisations holistically, in a similar manner to the way in which nation-states have been examined. In particular, it is necessary to investigate their social structure and internal ‘micro-politics’. This is not to say that the economic or identity aspects should be ignored, but rather they should be incorporated into a more holistic analysis – as they are in the analysis of nation-states.
This paper will provide such an analysis. It will describe and examine the features surrounding the concept of warlord. Of course, the danger of such an analysis is that it is not derived from any single theory, but rather bases its description on multiple theories. Moreover, this approach does not necessarily provide a model of the dynamics of warlord formation or the destruction of warlord organisations. Rather, it assumes that they already exist. Nevertheless, this conceptualization approach mirrors our approach to the examination of states and, moreover, it will provide the basis for future research into warlords which may address issues of warlord organisation formation.
The paper is organised around defining warlords and in doing this, the political, military, economic, and social structures which underlie warlord organisations will be examined. The paper will begin with an overview of the contemporary warlord debate through examining some of the definitions of warlord which have been put forth. These include those used by analysts of the Chinese warlord period and more recently by Africanists. After tracing this definitional history, the paper will move on to look at more modern definitions and the economic, barbarism, and grievance factors of warlordism that must be taken into account.
The paper will then explore the structure of the warlord organisation. This will involve looking at the warlord as an individual, warlord organisations as political community, the nature of warlord governance and command, as well as the warlord's means of motivation and logistics. The aspects of warlord organisations will be examined using concepts borrowed from the analysis of nation-states, such as political community and Max Weber's notions about authority. Recombining these features of warlords will lead to an inclusive definition of the term. Finally, some implications of defining warlords in this manner and having a more conceptual understanding of their organisational structure will be examined.
The Warlord Debate
In the modern age, the term warlord has been used to describe competing provincial military and political leaders during the period after the fall of the Qing dynasty in 1911 and up to either 1927, when Chiang Kai Shek was able to restore some form of order, or the beginning of the Anti-Japanese war in 1937. The term is a translation of the Chinese junfa which, like the English term warlord, has a pejorative association. The word ‘warlord’ itself is possibly derived from the literal translation of Kriegsherr, a formal title of German Kaisers (Darch, 1989). As opposed to an earlier term, dujun, which simply meant ‘supervisor of military affairs’, junfa connotes military activity without an aim or purpose – an important distinction still applied today (Rich, 1999). The word warlord was used as a translation by foreign correspondents in their sensational descriptions of the violent tactics used by the junfa.
James Sheridan gave a formal academic definition of warlord in his seminal work on the subject of Chinese warlords in Chinese Warlord: The Career of Feng Yu-hsiang (1966). He defines a warlord as an actor who ‘exercised effective governmental control over a fairly well-defined region by means of a military organisation that obeyed no higher authority than himself’ (1966:1). Many later authors have continued to rely on Sheridan's definition.
The staying power of Sheridan's definition is due to its simplicity and applicability. In essence, it demands only two features of a warlord actor. The first is that the actor must exercise control via military power. The second is that the actor is the highest level of a hierarchy. Sheridan's definition was most notably taken up by Africanists in the 1980s, who began applying the concept to African military actors.
African Warlords
The 1989 special issue of the Review of African Political Economy (16, 45/46), which focused on ‘Warlords and Problems of Democracy’, is considered to be a milestone in the warlord debate. The literature these authors were referring to was pioneered by historians like Sheridan in the 1960s and took on growing detail in the 1980s with authors such as Diana Lary's analysis of soldiers in the warlord period (Lary, 1985).
In particular, authors such as Roy May and Roger Charlton applied the concept of warlord to military actors in the protracted conflict in Chad during the 1980s (1989). According to Charlton and May, the Chadian state developed an internal militarism through two related processes:
a process of de-institutionalization and organisational decay at the level of the central government. Second, it involve[ed] a concomitant and progressive growth of regionalism, ultimately emerging as a regionalization of the whole political process (1989:17).
In both cases, these warlords ‘relied upon their personal politico-military skills to establish first, their control over a regional power-base and second, [drew] upon the economic resources of their fiefdoms, to expand, by force if necessary, their domain of effective power’ (1989:17). Put more simply, a warlord had two necessary characteristics, ‘a private army and an area under his control’ (1989:17). From these foundations, the term warlord, and its basic definition, was used to describe numerous political actors throughout sub-Saharan Africa and later the Balkans, Middle East, and Central Asia by both academics and the media.
Contemporary Debate
Since the 1980s there has been a continued interest in warlords. In particular, the Sierra Leonean and Liberian wars in West Africa, clan fighting in Somalia and Afghanistan, and a scattering of other potential candidates including the Democratic Republic of Congo (DRC) and Chechnya have been put forward as examples of conflicts involving warlords. In order to analyse these conflicts theorists have returned to the earlier writings of the Chinese historians as well as of the Africanists of the 1980s in pursuit of an accurate definition and conceptualisation of warlord.
In their analyses, most theorists have kept intact the central tenets of the definitions used by the Chinese historians and the Africanists. In particular, Sheridan's definition is echoed in more recent analyses. For example, Antonio Giustozzi defines a warlord as a ‘particular type of ruler, whose basic characteristics are his independence of any higher authority and his control of a “private army”, which responds to him personally’ in his analysis of Afghan warlords (2003:2). Jackson notes that:
the term warlord has been used … to describe a man who is in control of a particular group or area and who does not answer directly to a higher authority – although they may defer to stronger warlords (2003:134).
Beyond these basic features, two supplementary factors have been noted by commentators on warlords. The first are the economic issues wrapped up in warlord organisations. The second feature is warlords' notorious barbarism.
Economic Factors
Theorists often note the self-serving nature of warlords and warlord organisations. Patrick Chabal and Jean-Pascal Daloz note that warlords are ‘quite literally, businessmen of war, that is, they relied on violence as the main instrument of their economic activity’ (1999:85). Another author notes that the raison d'etre of warlordism should be the ‘pursuit of narrow, commercial self-interest’ (Robinson, 2001:123).
The economic rationale for warfare is nothing new, even Aristotle mentions the use of war fighting for economic gain. In his Politics he remarks, ‘it also follows that the art of war is in some sense … a natural mode of acquisition’ (1958:21). Warlords, however, are thought to make economic exploitation their primary justification for warfare. These accusations are not unfounded. It has been reported that Taylor made hundreds of millions if not billions during his time as a warlord in Liberia. William Reno takes this position to be the essential feature of warlordism and bases his theory of warlordism on their exploitation of economic markets (Reno, 1998). Whether or not it is the defining feature, clearly economic profit is of central importance to warlords and must be taken into account in an analysis.
Barbarism & Grievance
Warlords are almost synonymous with barbarism, savagery, and ‘senseless’ acts of violence. Numerous authors have reflected on the almost unimaginable suffering that warlords have caused, usually for seemingly pointless reasons. For example, Ralph Peters refers to
erratic primitives of shifting allegiance, habituated to violence, with no stake in civil order … [who are] as brutal as ever and distinctly better-armed (Peters, 1994:16).
distinguished by personal leadership, generally weak organisational structures, and still weaker ideological motivation. In those cases where they managed to overthrow incumbent regimes, they generally proved unable to establish effective governments in their place (1996:212).
Often, the explanation for warlord's barbaric behaviour is that they have grievances against specific social groups, based on ethnicity or other identity, and are ‘lashing out’. For example, the Revolutionary United Front (RUF) and its constituency of lumpen proletariat, may be seen as a revolt against the established professional class in Sierra Leone which had failed them (Abdullah & Muana, 1997).
No analysis of warlords can fail to incorporate an explanation of the savagery of the warlord. However, simply noting that warlords are ‘evil’ or ‘psychotic’ is of little analytical value (see Keen, 1998, 2000). At the same time, identity-based grievances are insufficient explanations of warlord behaviour since they often attack their own communities and in general do not seem to be driven by singular ideological concerns of any sort. Thus, it is necessary to explain the role and relationship of barbarism and identity issues with warlords within a broader, more holistic view.
Warlord Conceptualisation
Having looked at some of the different definitions of warlord, we can begin to put together an inclusive definition and explore some of the concepts which underlie the idea of a warlord. Many different concepts have come up in the definitions which others have used, even if they are not noted by name. In addition, there are some underlying concepts which theorists have taken for granted. The following section will explore some of these concepts, including the warlord as an individual, notions of political community, autonomy, and independence. Other issues which must be addressed in regard to warlords include military, motivational, and logistical aspects.
The Warlord vs. Warlord Organisation
A warlord is a man, but we should not analyse the nature of the warlord as only an individual. Warlords include men like Dostum, Taylor, and Mohammed Farah Aidid. They do not act alone, but are instead the highest authority over organisations. While in practice it might not be possible to replace a warlord because he is the central link in the chain of authority, in theory he is replaceable.
There are examples of warlord replacement as well as counter-examples. For instance, Aidid was replaced upon his death in 1996 by his son Hussein Aidid. On the other hand, the União Nacional pela Independência Total de Angola (UNITA) collapsed soon after Jonas Savimbi's death in 2002. The explanation is that in warlord organisations, as with other military organisations, the leader's death weakens the organisation significantly. In some cases, this weakness is enough to cause the complete collapse of the organisation, and in others it is not.
These ‘warlord organisations’ are an emergent entity made up of the sum of its individuals and these individuals combine together into a cohesive unit, which can, for theoretical purpose, be treated as a single actor. In this sense the warlord organisation – which itself is an ‘armed group’ – exists as a separate, definable entity.
Even though the term warlord is usually applied to a single individual, it more rightly refers to a cohesive group. This is the same way that we treat a state run by a dictator as a single group even though we refer to the individual dictator as the instigator of policy. Accordingly, we should treat the warlord as in fact being part of a ‘warlord organisation’, but we may continue the common practice of referring to the entire organisation as simply ‘warlord’.
Political Community
This begs the question, what is the warlord organisation and how is it defined? It is often assumed that authority is territorially based and, by extension, that a definition of warlord should be based on control of a fixed territory. This assumption is made because we define the state in reference to a fixed territory with distinct boundaries. But, as Jeffrey Herbst has shown, territorial control is not always as important as authority over people (Herbst, 2000). While Herbst was referring to Africa, sovereignty as authority over people can also apply to other areas and other groups. For example, we assume it in reference to nomadic tribes in Central Asia throughout history.
Unlike the state, the population that is under the warlord's authority is not based on territory, but is instead based on membership and specific inclusion. With the state, membership is defined by being born within a specified territory, or otherwise attaching oneself to the territorial state. The warlord organisation, on the other hand, is made up of those who are specifically initiated. In effect, this usually means those that have undergone some sort of recruitment process and are part of the patronage system. For example, a person may join and become a paid member of the Somali warlord Mohammed Qanyare's militia. However, initiation into membership can be achieved in other ways, such as forceful conscription. For instance, a child who is forced to fight for the RUF becomes a member of the organisation. This process of initiation and inclusion also differentiates those in the warlord organisation from other forms of compulsive inclusion, such as being part of a family, clan, or ethnic group.
Members are included into a clearly delineated organisation, which is in essence, a political community. Political community is defined by a definite in-group and outgroup distinction (see Simmel, 1955). This boundary determines – to use Carl Schmitt's terminology (1996) – that friends are those also under the authority of the warlord and the enemy is, usually, everyone else. For the state, this ‘friend/enemy’ distinction is based on fixed territorial boundaries or, for a nomadic tribe, it is based on ethnicity and familial relation etc. But, for the warlord, this boundary is based on inclusion and retention into the warlord organisation.
As with the state, this ‘friend/enemy’ distinction must be continually redefined and reinforced. For example, in Liberia, it has been noted that ‘[a]s the overall threat to a faction decreased, as it did in some area after Cotonou [a multifaceted agreement between the factions, including cease-fires], cohesion reduced’ (Alao, Mackinlay, & Olonisakin, 1999:47). As with states, the paramount way to (re)create the friendenemy distinction is through war. This helps to explain the warlord's continual need for conflict; without it, his organisation might simply dissolve. As Weber notes, ‘[t]he charisma of the warlord rises and falls with its efficacy and also with the demand for it; the warlord becomes a permanent figure when there is a chronic state of war’ (1978:1142).
The warlord community is separate, but related to existent political communities, such as clans or states. Often warlords arise out of a sub-national political community which has a defined border. The most common examples of this are clans, as occurs in Somalia or Afghanistan. The leaders of such armed groups, which are by definition militias – since they are the non-professional armed extension of civilian community – are more rightly called warleaders. However, the warlord becomes a warlord in the sense that he breaks away from a dependency on the clan (see Vinci, 2006b). Though he may continue to use the rhetoric of clan, he must base his motivation on other factors, discussed below. This firm line between warlord organisations and clan militia is an important distinction to make because it better reflects reality and is starkly evident in the fact that many warlords do not arise out of clans or other political communities – though they may attempt to adopt such rhetoric anyway or otherwise use such a relationship instrumentally.
Accordingly, we should call the political community which the warlord controls a fiefdom. The term fiefdom refers to a specific piece of territory, which warlords may more or less temporarily control. There is no doubt that warlords sometimes control territory and the people on that territory. But there is also a second meaning of the word which is more apt for this discussion. This is that the word can be used in the sense that it is an organisation which is controlled by a dominant person or group. It is more rightly in this sense that the warlord has a fiefdom over which he is the highest authority.
This notion of membership rather than territorial control aligns with empirical observations of warlords. For example, in Somalia, Qanyare has had control over specific pieces of territory, including, for instance, Dayinle airport. But, it is not that we associate Qanyare's authority with a specific piece of territory, even if he has controlled it for a long period of time. Rather, it is Qanyare's authority over his private militia which is important. A defeat in one place and victory in another will cause Qanyare and his group of fighters to move, but he would still have authority over the group and be no less a warlord. In the same way, for the fighters in Qanyare's militia, their loyalty is not to a territory, but to the organisation, and by extension, Qanyare himself. The men do not answer to any government or authority other than him or their assigned superior.
Understanding this peculiar nature of the warlord political community allows us to better understand the common warlord practice of killing or looting from local populations. It may seem to an outsider that a warlord organisation made up of individuals from the same ethnic group, religion or community of the local population would not want to prey on these people because they are ‘the same people’. Therefore, when warlords kill this local population, they are often considered to be savage. The reason being that we typically think of killing within our own political community as savage and murderous, whereas we can justify the killing of our political community's ‘enemies’.
But this is not how the situation should be understood. The individuals within the warlord organisation have their own separate political community and see the local residents that they prey on as enemies. To members of the warlord political community it is acceptable to kill those outside of the political community in the same way as it is seen as acceptable for citizens from one nation to kill citizens from another nation during wartime, even if they are of a similar ethnic or religious background.
This allows us to understand a seemingly paradoxical organisation like the Lord's Resistance Army (LRA). The LRA is made up solely of Acholi people from northern Uganda and southern Sudan. Yet, the LRA also preys extensively upon the Acholi people – torturing, abducting, and looting from them on a regular basis. To some this may seem contradictory as they would assume that the LRA would not want to alienate the people that they have sprung from and should represent them against the southern ethnic groups, which the LRA professes to actually be fighting. However, when we keep in mind that the LRA personnel have defined themselves separately from the Acholi people, it becomes possible to see their actions as logical. They have their own political community and it is not based on ethnicity, but on initiation and indoctrination. Therefore, to the LRA the Acholi people are as much outsiders as any other political community would be and therefore violence against them is as acceptable as against any other political community in a conflict.
This is only one of the ways in which a warlord political community differs from civil political communities. A related point is that due to the nature of the warlord political community, the idea of ‘public goods’, i.e. those which cannot exclude individuals, is completely alien. Only those initiated into the warlord political community are able to obtain goods from the organisation and these goods are in themselves the glue which holds the organisation together. This is what separates the warlord from the state, which does provide public goods to those within its territory. For, unlike the state, the warlord ‘does not primarily direct the peaceful struggle of man with nature, but the violent struggle of one community against another’ (Weber, 1978:1141, 42). But, the warlord organisation is nonetheless a political community and political communities are dominated in the Weberian sense of the members of the community not just being ruled by force, but of also having an interest in obedience.
Warlord Governance & Command
The warlord's authority over a population implies the existence of governance. However, warlord governance is extremely different from that found in a state or even a guerrilla insurgency. In a state there is a clear citizen base and a military is formed out of this citizen base. Even in a guerrilla insurgency there is a clear base of the people who continue on with their lives while they may be governed by the insurgents or otherwise contribute to the rebel's fight.
The warlord organisation, on the other hand, ‘veers toward a total combination of military and political means’ (Rich, 1999:6). All members of the warlord political community are part of a military organisation, and even the economy is fundamentally wrapped up with the military. This is pure praetorism – i.e. the intervention of the military into political life and breakdown of civilian/military relations.
Consequently, the warlord's praetorian governance structure is the same as its command, control, and communication structure. In the warlord organisation, political organisation is the same as military organisation, and to command the military is to govern. In order to bring about such domination, the warlord organisation must have: a leadership, hierarchical control structure, set of tactics and strategies to carry out, and method for effectively communicating orders.
The leadership of a warlord organisation, and in particular, the warlord himself, can set the rules of governance. For instance, Joseph Kony, the leader of the LRA, has instituted an entire quasi-religious governance and command institution. However, there are certain roots to authority which a warlord may potentially use.
In order to effectively govern a political organisation, the warlord may obtain authority from charismatic, patrimonial, or rational (legal) sources, just as states can (Weber, 1958, 1978). Though Weber admits that authority can come simply from force, this is an unreliable form of authority and therefore the leader must legitimise his authority in some manner. Charismatic power originates in an individual. Patrimonial power derives from direct exchange from the top of an organisation down the hierarchy. Finally, bureaucratic power is instilled in the organisational structure itself. In general, the warlord organisation is a non-bureaucratic form of organisation (Reno, 1998). Rather, the warlord relies on patrimonial and charismatic sources of authority. For the warlord organisation, patronage usually comes in the form of looted goods or other monetary inducements which flow down from the highest levels of the organisation to the lowest fighters in the warlord organisation. The patronage creates a bond between the warlord and the fighters in the political community. This patronage is the incentive which allows the warlord to retain pre-eminent control over the fighters, for it is a mirror of patriarchal domination. As Weber notes,
[u]nder patriarchal domination the legitimacy of the master's orders is guaranteed by personal subjection … The master wields his power without restraint, at his own discretion and, above all, unencumbered by rules … (1978:1006-07).
The reliance on patronage also helps to explain why the warlord as an individual is so central to the organisation. The warlord will almost always attempt to monopolise the economic connections which dominate the warlord organisation's economic system (Reno, 1998). In so doing, the warlord as an individual becomes a necessary link in the chain of organisational control. While Reno stresses the economic repercussions of the warlord's centrality to the organisation, here the ‘political’ repercussions – in the sense of politics being about determining who gets what, when, and how (Lasswell, 1936) – should also be stressed. And, in this sense, patronage is what defines the internal politics of the warlord organisation.
Charismatic power is also important for the warlord. For example, the LRA is governed by the mystic Kony, who ‘uses … spiritualism to maintain control, starting with his overall vision of liberation and destruction and continuing with individual spirits that “guide” specific military tactics’ (Refugee Law Project, 1004:13). Or, for instance, in Liberia ‘success and influence depended more on a commander's power in his own right as a dominating personality in the faction hierarchy than on his capabilities as a military leader’ (Alao, Mackinlay & Olonisakin, 1999:47). This too helps to explain the prominence of the warlord as an individual in the warlord organisation, for it is in him that charismatic authority resides.
Warlords have their own personalised organisational systems for maintaining a hierarchy of domination with corresponding discipline and communication. In general, these systems will be much less efficient than in conventional militaries. Often discipline is low and based on force and abuse. Communication is basic, though satellite phones and other high-tech equipment is regularly used by warlords who obtain it as easily as any other group can through globalised trade networks. Another factor, which Reno has argued, is that warlord organisations lack a structured bureaucracy in the way which most states have (1998). This will also serve to make the warlord organisation significantly less efficient, though not necessarily less effective, than states. Nonetheless, even though warlord organisations are often characterised as nearly chaotic organisations, there is always a system for controlling fighters in place.
With the ability to direct troops in place, a warlord must develop a strategy and set of tactics with which to determine what direction to go in, so to speak. The strategies can be borrowed, such as the guerrilla strategies developed by Che Guevara or the military doctrine of a state's army, or made up by the armed group itself, as the ‘Holy Spirit Tactics’ of Alice Lakwena's Holy Spirit Movement (see Guevara, 1961; Behrand, 1999). In some cases, they seem to fight with essentially no strategy and few tactics other than random shooting, as is often said about warlords of the Liberian and Ugandan conflicts. However, even in these instances, there is in fact a rational set of tactics and strategies (see De Waal, 1997).
With a set of strategies and tactics in place, a warlord must then teach these to the fighters in the organisation. Since such strategic doctrine is, in effect, the praetorian organisation's government, or constitution, by indoctrinating new troops, the warlord reproduces and reinforces the organisation over time. Even the most simple and barbaric warlord still institutes some form of training and indoctrination for the fighters. For example, the LRA gave extensive combat training to its soldiers. One returned fighter noted:
In Sudan they gave us training for three weeks. … I was also trained to shoot, and how to put together guns and handle the weapons – antipersonnel mines, antitank mines, SMG, LMF, PKM, mortars (interview with ‘Sarah’, a returned LRA fighter, reported in HRW, 1997).
Autonomy
Warlords and warlord organisations are completely autonomous from the state. This feature is reiterated in most of the earlier definitions of warlord. The warlord's autonomy makes it the highest authority and this authority extends to both the warlord organisation's internal and external relations.
The warlord's authority over its internal affairs seems paradoxical since almost all the Earth's territory is regulated by one state or another and states are by definition the highest authority over entities within their territory. At the base of the state's authority is its monopolisation over the legitimate use of force (Weber, 1958). Through a process of cooption and force, the nation-state has generally been able to make itself the most powerful actor within its territory (see Tilly, 1995). This authority assumedly extends to non-state actors. For instance, while Microsoft operates multinationally, it is still regulated by US law and is therefore not completely autonomous.
Warlords, however, are not subject to regulation by states because, in some instances, the state ‘fails’ and a non-state actor may become powerful enough, relative to the state, to rival its authority. The state ‘fails’ as it looses control over the population within its territory. Eventually, the state's authority and, lacking that, control, are weakened to the point where another actor can exert its own control. In such instances, the warlord can wrestle away its own autonomy.
Military Force
Warlords are able to rival the state because they have the military ability to do so. As noted, the warlord organisation is praetorian in nature – it is in effect an army. The nature of the warlord military force depends on the particular warlord. Examples range from conventional forces split into companies and platoons and using assault rifles and artillery to irregular units of child soldiers using machetes or ‘technicals’. The military force may not necessarily be enough to topple a state, but it is certainly enough to scratch out some territorial control for a temporary period at least. This military force can not only be used against the state to make them autonomous, but it also used internally to serve as the reservoir of power that allows the warlord to maintain authority.
In particular, the warlord will turn to asymmetric means in order to assure its autonomy. The warlord necessarily exists inside a state and states in general, even ones which are failed, tend to have powerful military capabilities. In contrast, warlords are much smaller organisations and they cannot co-opt society or use international backing in the way that states do. This makes warlords relatively weak. In order to overcome this asymmetry, warlords must turn to asymmetric ways of warfare.
Asymmetric warfare denotes a mismatch between the capabilities of belligerents involved, where at least one of the sides changes its tactics or strategies to exploit the asymmetry. The usual type of asymmetric warfare we refer to is that practiced by guerrillas, however, warlords cannot rely on the civilian population in the way that guerrillas can. Also, unlike terrorists, which also practice a brand of asymmetric warfare, warlords cannot hope to hide in a society.
Therefore, warlords must turn to their own breed of asymmetric warfare, what we might call the warlord way of warfare. This type of warfare is an extremely violent and savage one. It generally involves attacking civilians rather than military targets, using ‘destabilisation’ tactics (De Waal, 1997), such as through committing conspicuous atrocities, and relying on the use of fear as a force multiplication strategy (Vinci, 2005). Yet, it is a highly effective form of asymmetric warfare which allows the warlord to fight extremely protracted conflicts with much more powerful states. This need to turn to brutal forms of warfare helps to explain why warlords are barbaric.
This military ability allows the warlord to remain the highest authority over its external relations. Since the warlord has the military ability to rival the state it can demand its own, autonomous foreign relations for the simple reason that the state cannot stop it from doing so. Thus, warlords interact with the international environment directly rather than having their interactions regulated by a state. For example, Taylor made business deals with various foreign companies, while some Somali warlords have aligned themselves with Ethiopia or other neighbouring states.
Self-Perpetuating Organisations
The warlord organisation is not autonomous for a temporary period, rather, it exists continuously. Put another way, the organisation perpetuates itself. As with states, defence against the state or other aggressors is part and parcel of this selfperpetuation.
Civil political communities, such as states and tribes, are not only separate from other political communities, but are also self-perpetuating organisations. These organisations have people which reproduce at a rate which is at least the level at which people die. They also have a governance structure to rule and direct the members of the organisation and the ability to teach new generations the same rules and norms as previous generations. Moreover, they have economic systems which can provide food and other basic necessities with which to survive.
Like other political communities, the members of the warlord political community can perpetuate itself. In order to be such a self-perpetuating organisation, the warlord must be able to accomplish some very similar objectives to that of civil political communities. In practice this means meeting several requirements (see Vinci, 2006). As has already been demonstrated, warlords have a leadership structure and the ability to indoctrinate new people. They must also have the ability to recruit new members, at least at the rate necessary to sustain its ranks and the economic ability to provide for the survival of its members and to procure a military ability with which to defend the organisation.
Warlord Motivation & Logistics
One of the most difficult problems for a warlord to solve is how to motivate troops to fight for him. Communitarian loyalty, such as to an ethnicity, clan, or ideology very often are used to motivate troops to fight in armed groups. The foundation of warlord organisations are often based on some form of communitarian loyalty for motivation. For example, in Somalia, most warlords had some sort of connection with clans and clan-based loyalty-dominated motivation.
However, warlords generally move away from communitarian forms of loyalty, as their true motivations are connected to the initiation-based warlord organisation, rather than to an external political community. This has disadvantages for the warlord in that it makes it more difficult for him to convince fighters to remain loyal. Nevertheless, it is advantageous to the warlord because it means that no outside authority can usurp his authority.
In cases where they do maintain a pretence of communitarian ideals, they do so cynically, and with an instrumental focus. For instance, in regard to rise of warlord organisations in Liberia, Eboe Hutchful and Kwesi Aning note that:
[t]he leaders of these factions claimed to have built their armies to defend their own ethnic groups against attacks from other armed factions. Ethnicity, however, was employed simply as a façade to camouflage political ambitions and aspirations to maintain power within a small circle of Liberia's elites (2004:209).
Most [fighters] did not receive a cash salary – their food and essential survival needs were ‘found’ from local sources. Looting captured houses was seen as a legitimate reward for months, sometimes years, of extreme hardship (Alao, Mackinlay & Olonisakin, 1999:46).
Although not necessarily a primary means of motivation, the warlord is also likely to turn to coercion as a reinforcement of motivation and as a means to incorporate new members. Brutal violence is an easy way to indoctrinate troops and enforce loyalty. Other forms of psychological manipulation may be used to recruit new members, including offering addictive drugs.
At the same time, warlordism may become a self-fulfilling prophesy as the security they create forces individuals to think in terms of self-help, and in instances join the warlord organisation. This is especially relevent for young men who may only be able to turn to enlistment with the warlord to find any security. As Coker notes, there has been the growth of a ‘neo-feudal security regime in which the only protection against violence is membership of gangs, clans, or allegiance to personal warlords or leaders with their respective feudal affiliations and ties’(Coker, 2001:115). Ultimately, it is often the case that it is safer to be part of a warlord organisation than it is to be a civilian (see Keen, 1998).
More generally, appeasing a warlord may be the only way to remove his threat. For example, when Taylor ran for President of Liberia in 1997, his (unofficial) campaign slogan was: ‘You killed my ma; you killed my pa; I'll vote for you.’ It has been noted that ‘the overwhelming vote for Taylor was in reality a vote for the only leader who could deliver peace’ (Alao, Mackinlay & Olonisakin, 1999:119).
The reliance on patronage only adds to the warlords pressing economic needs. Weapons are also extremely important as are the necesssities of basic food and clothing. In order to meet these economic needs, the warlord must exploit resources, since he cannot easily turn to taxation in the way that more legitimate political actors can. While warlords can turn to external supporters, they are careful about the loss of independence and are notoriously disloyal when their interests change. Instead of being seen as dependents, external suppliers tend to supply warlords when their interests converge rather than when they hope to change the warlord's interests.
Thus, we observe the high prevalence of looting by warlords. This aspect of warlord organisations is explained to some degree by Clapham, who notes,
since it was difficult for insurgents to develop regular structures of production, the export of goods from insurgent-held areas was liable to degenerate into a once-for-all sale of anything that could be carried away… (1996:233).
Warlords can take what they loot and exchange it in the ‘transborder shadow economy’ for goods that they may need (Duffield, 2001). In particular, they can obtain the small arms and other tools of war which they need to perpetuate their conflict (see Cooper, 1999). For this reason, as Mark Duffield notes, ‘today's successful warlords may act locally but they think globally’ (2001:175).
The implication of this analysis is that while it is true that these funds do benefit a warlord or his followers, this is not necessarily to say that this is the reason why warlords exist. Seen from another perspective, economic exploitation serves as a way of perpetuating the warlord organisation, just as national economies and taxes perpetuate states. For example, Taylor was able to perpetuate his warlord enterprise through various informal economic deals with international corporations and even states. Somali warlords use the land that they control for economic purposes; for example, airports are privatised and run by individual clans and ‘transportation taxes’ are an important source of income. The profit from such economic exploitation is then turned around to fund the necessary patronage and logistics which a warlord needs to perpetuate the organisation.
Parasitical Nature
The warlord's independence is an odd and complicated one since the warlord is parasitic off what we may call the ‘host state’. It is in this sense that warlords are, as Hobbes remarks, ‘like worms in the entrails of natural man’ (Hobbes, 1998:221). The warlord for this reason is absolutely illegitimate, for, they have no rightful reason to expect to receive any gains, since they provide nothing, but force others to give without legitimate authority.
It is this parasitical nature of warlord political communities that fundamentally separates it from the state, tribe, or other form of self-sustainable political community. The warlord taps into the people and economic resources of a state, and siphons off what is necessary. It would not be possible for a warlord to exist as an entity without a state of some sort. If such a situation were to occur, the warlord organisation would necessarily have to transform its organisational structure into one that was a de facto state of some sort, which had the organisational structure to rule over a larger, self-sustaining political community. This, in fact, does occur if, for instance, the warlord conquers a state completely, as Taylor did. However, usually the warlord does not do this and simply continues a parasitic existence.
Though the warlord is parasitic, it is nonetheless an independent organisation and truly separate from the state. The warlord is not dependent on the state's public goods, or any other state or entity. The warlord organisation simply loots what it wants. Nor does the warlord organisation need defence from outside actors. Rather, the warlord organisation provides for its own security. When a warlord does make alliances with other states, he is sure to maintain his decision-making authority; as is illustrated by the high rate of warlords cheating on, or otherwise backing out of alliances.
The warlord is also independent from society. The warlord organisation is its own political community and is therefore separate from civilian society. It may exploit local civilians, just as the warlord may exploit any group; however, it is not dependent on local civilians for goodwill. This differentiates him from national resistance movements or traditional guerrillas, which must rely on the support of local civilian communities. The differentiation is important, since by aligning with society, the warlord may have to provide concessions, whereas by looting from society, the warlord does not lose any decision-making authority.
It is the warlord's independence from society which separates it from other types of autonomous and independent armed groups. Some insurgents come to control large swathes of territory as well as large populations. These insurgents, however, usually depend on society. Using standard guerrilla doctrine, they base their power in the ability align themselves with society and use its resources to combat the government. This is how, for example, Yoweri Museveni's National Resistance Movement (NRM) was able to maintain itself, even without turning to significant external support (see Kasfir, 2002). This alignment allows the guerrilla to make the organisation independent from the state, but in turn it limits the guerrillas' independence from society. However, this is acceptable to the guerrilla organisation because they are not attempting to separate themselves from society, but rather they are trying to take power over society from the government or to secede from the state. The warlord does not turn to society for support, nor does he then turn to an external state sponsor; instead, the warlord organisation is an independent, though parasitic, entity.
Proposed Definition
To summarise, warlords are truly different organisations from western liberal nation-state tribes, or other political communities which we are used to thinking about. Nevertheless, the warlord heads a self-perpetuating political community, which is based on membership and inclusion, rather than territorial connection. The warlord is autonomous from the state and, furthermore, independent from all other actors. In order to perpetuate the organisation and remain autonomous, the warlord uses economic or other exploitation as well as military power. Although this conceptual analysis of warlord has brought out many complex ideas, it is possible to simplify the definition into a form which includes the most important concepts and which reflects, though also builds upon, previous definitions.
Thus, a warlord can be defined as the leader of an armed group that uses military power and economic exploitation to maintain fiefdoms which are autonomous and independent from the state and society.
Conclusion
The broad implication of this analysis is that in order to effectively analyse armed groups it is necessary to take a more intensive, conceptual approach. The normative associations of warlord, or any other type of armed group, only confuse the matter. The organisations should not be seen as single-minded, economically motivated machines. Rather, there is a is a complex set of forces which combine to define the warlord, or any other armed group, ranging from international forces, the dynamics of political communities, to basic principles of human nature. In other words, a similar approach must be used to analyse warlords as is used to analyse any other type of human association. In particular, we might draw parallels with the analysis of nation-states. From this conceptualisation, there arise theoretical, comparative, and policy implications.
Conceptualising warlords in depth provides some theoretical benefits. One of the problems with understanding warlord organisations is that they are such a radically different phenomenon from other types of political organisation. This makes it difficult to integrate them into broader theories. Many of the concepts used in this conceptualisation are derived from or applicable to nation-states and other forms of political organisation. From this it is possible to see the foundation for a cross-fertilisation of ideas about warlords and about other types of political organisation. For instance, while the nature of a warlord political community was only briefly discussed in this paper, there is significant literature on this subject and much of it might be applied to warlords. Similarly, warlords were found to be autonomous and independent from the state. Such autonomy is one of the foundational features of the concept of sovereignty and therefore there may exist the possibility of sharing insights concerning sovereignty with analyses of warlords; or even a wholesale integration of warlords into international relations theory.
It is also possible to operationalise this definition to test whether an actor is a warlord. First we must ask whether he has a political community over which he is the highest authority. Is this fiefdom autonomous; i.e. does the group have the military power to do what it wants? We can ask if the group is independent; i.e. does he have the ability to obtain the personnel and equipment and can he direct this force effectively and cohesively enough to retain decision making authority.
In this way, warlords can be differentiated from seemingly similar actors. For instance, some guerrilla insurgencies can in some cases appear very similar to warlords, since they may rely on significant levels of economic exploitation or use highly violent tactics. However, it is possible to tell them apart based on their relationship with society. Warlords are fundamentally separated from society, whereas true insurgencies rely on civilian society for support. Thus, for example, it is better to classify the LRA as a warlord organisation than an insurgency. While it may use the rhetoric of fighting the Ugandan government for political reasons, its complete divorce from the Acholi community of northern Uganda clearly demonstrates its warlord nature.
Such subtle classification can be extremely helpful not just in analysis, but also in response. Going back to the LRA example, the Ugandan government has implemented the standard counterinsurgency policy of ‘protected hamlets’ in northern Uganda. However, this policy – which has effectively cleared most of the northern Ugandan countryside of civilians – has not seemed to weaken the LRA substantially. The reason is clear in light of the warlord conceptual analysis. The LRA is independent from society, except in a purely predatory manner. The implication is that separating it from society will not weaken it in the same way as it might weaken an insurgency which ‘can neither exist nor flourish if it separates itself from [the masses'] sympathies and co-operation’ (Mao, 1961, reprinted in Campbell, 1967:260).
In general, beyond these theoretical and comparative insights, a deeper understanding of warlords and warlordism has many real benefits for policy and, especially, for conflict prevention. Warlords are poorly understood political organisations and often lead to real questions about how and why they act as they do. For instance, it is not clear why some warlords in Afghanistan were willing to give up their weapons and join a national-level government while others were not. This is part of larger issue of the inexplicability of the prolonged nature of warlord fighting which often seems ‘senseless’ or ‘pointless’. Even when the warlords do end their fighting, it is often difficult to demobilise their fighters and, oftentimes, the warlord will simply start another war in another place.
Some of these questions become clearer based on the provided conceptualisation. For instance, the enclosed political community of the warlord organisation helps to explain why it is so difficult to get warlords to (re)integrate into a state or society. Their need to use economic exploitation for both motivational and logistical reasons – in effect their very existence as an organisation – demands continual conflict. Thisrational basis for prolonged war has been noted by analysts such as Keen (1998, 2000), but the particular social and political structures which cause this need have been only generally analysed. The next step in the process is to develop specific ‘counterwarlord’ strategies which are based in a precise understanding of what a warlord is and how a warlord organisation functions, in the same way that counterinsurgency strategies have been developed based on precise understandings of insurgency. In this way, it may be possible to finally alleviate the scourge of warlordism.