International Studies Quarterly (2008) 52, 295314

Anarchy, Failed States, and Armed Groups: Reconsidering Conventional Analysis

Anthony Vinci The London School of Economics
Systemic theories of international politics divide the world into anarchic and hierarchic systems. Conventionally, the boundary of each system is based on juridically dened territorial borders. This article argues that within collapsed and fragmented states there exist autonomous armed groups, which make these states by denition anarchic systems. Unlike earlier accounts of such domestic anarchy, this article argues that the domestic anarchy is open or connected to the international anarchic system. By taking this approach, it is possible to integrate the theoretical understanding of the relations between armed groups and (external) states. Specically, the article illustrates how there can be mixed security dilemmas between states and armed groups, and that other theoretical concepts usually reserved for describing interstate relations can potentially be used to describe state-armed group relations.

For the sake of theory, the world can be divided into anarchic and hierarchic systems. States are hierarchical systems in that a government, or sovereign, structures the relationships between sub-state actors. In contrast, the relations between stateswhich are concerned with maintaining their autonomyare not governed by any higher authority, thus the system is anarchic. This anarchy is, as Hedley Bull notes, the central fact of the international system and the starting place for theorizing about it (Bull 1966, 35). International Relations (IR) theories begin their analysis based on the postulation that the relations of states in the anarchic international system are not chaotic, but rather there is a perceivable pattern, or law like regularities, which can be explained (Waltz 1979, 116).1 This bifurcated view of the world is deceivingly simple, as in some cases it is not an easy feat to decide whether something is an anarchic or hierarchic system. Specically, in fragmented and collapsed states the government is weak or non-existent and therefore cannot structure the relationships of all actors within the state. In particular, the state cannot exert authority over armed groups such as insurgencies or warlords. As such, these groups can maintain autonomy from the government through force, sometimes for decades. In contemporary history there have been numerous such autonomous armed groups. For instance, the Sudanese Peoples Liberation Movement Army (SPLM A) controlled a large portion of Sudan and remained autonomous from the state for decades. In Somalia, warlords and faction militias have independently controlled portions of the state for over a decade. These actors not only are able to structure their own relationships internally, but can also act
1

Also see Buzan, Jones, and Little (1993).

2008 International Studies Association. Published by Blackwell Publishing, 350 Main Street, Malden, MA 02148, USA, and 9600 Garsington Road, Oxford OX4 2DQ , UK .

296

Anarchy, Failed States, and Armed Groups

nancially and politically in the international system without interference from the state in which [they are] based (Dufeld 1997, 18). The fundamental dilemma is that it is not readily apparent whether states containing such actors are anarchic or hierarchic systems and therefore whether the armed groups should be perceived as units within an anarchic system, as with states, or units in a hierarchic system, as with, for instance, political parties. Kenneth Waltz addressed this issue with regard to the Chinese warlord period in Theory of International Politics (Waltz 1979, 88). As Waltz notes, a decision as to whether such states are anarchic or hierarchic systems is necessary because our expectations about the fate of those areas differ wildly depending on which answer to the structural question becomes the right one (Waltz 1979, 116). In particular, if the system is anarchic, we can expect that the actors within the system are likely to act in a manner consistent with insights about relations in the international system. Typically, Neorealists maintain that no matter the presence of armed groups, a state is a state. As such, Neorealism is unable to incorporate armed groups into its theoretical descriptions of the international system and this is a major failing of the theory. In effect, Neorealism ignores armed groups. This is a problem, however, given the increasing importance of armed groups in international politics, for the inuence of armed groups ranging from Hezbollah in Lebanon to the Union of Islamic Courts (UIC) in Somalia to the Taliban in Afghanistan are central to understanding regional and even global politics. As such, this article begins the process of incorporating these actors into systemic theories of international relations, particularly Neorealism, which has proven otherwise to be a very productive theoretical explanation of international relations. Therefore, this article asks the question: are fragmented and collapsed states anarchic or hierarchic systems? Moreover, the article asks, if such states are anarchic systems, what is the nature of this anarchy; in particular, is it similar or linked to the anarchic system? In answering these questions, the article will provide a basis upon which to begin the analysis of armed groups and thereby to integrate or exclude them from models of international politics. The approach to answering these questions will be to present theoretical arguments and to provide empirical illustrations. Constraints on space preclude large-scale empirical studies; however, the article will provide the basis for applying a systemic approach to the study of armed groups as well as testable hypotheses that can be demonstrated in future empirical studies. This article will argue that in cases of collapsed and fragmented states, the hierarchic system breaks down and domestic anarchy ensues. Domestic anarchy arises when the state apparatus loses authority relative to non-state armed groups who are able to become the highest authority over their internal and external relations. In such cases, multiple autonomous actors exist within a dened territory and relate with each other as equal units. Other theorists have examined the explanatory possibilities of domestic anarchy. Barry Posen (1993) rst applied the concept of the security dilemma to ethnic conicts in domestic anarchic situations following the collapse of empires. Since then, Stephen David (1997), Robert Jervis, and Jack Snyder (1999) amongst others have also found insights by conceptualizing the inside of failed states as anarchic systems.2 In particular, these authors have used international relations concepts such as the security dilemma to explain the nature of internal wars. Yet, these authors have not broached the subject of how actors in the domestic anarchic situation relate with international actors such as states, and vice versa. Rather, they have tended to argue that domestic anarchy is different than
2

Also see Kasr (2004).

Anthony Vinci

297

international anarchy. Based on this, they have also made the implicit assumption that actors within a failed state will not, or cannot, relate with actors outside the states shell in a manner consistent with them being part of the same anarchic system. Building on this earlier work, this article will argue that domestic anarchy is not a separate, closed anarchic system, but rather it is open and linked with the international system. It will be argued that assumptions about the differences between domestic and international anarchy are mostly assumed rather than real and the case will be made that autonomous armed groups inside of a collapsed or fragmented state should be seen as being units in the international system proper. The broader implication of this account is that we should see the units of the international system being dened by their de facto autonomy, not by their juridically dened territorial borders. The importance of this is that it allows us to theoretically integrate internal armed groups and external states into a single systemic framework and thereby to describe and predict the relations between these actors. In this way, systemic theory is expanded to incorporate inuential armed groups like the UIC into regional or international theoretical constructs such as the balance of power or security dilemma. The article will begin by addressing Waltzs own musings on the subject of failed states. It will then move on to examine the nature of failed states and autonomous armed groups in order to argue that fragmented and collapsed states are anarchic systems. The article will make the case that the nature of domestic anarchy is no different from the anarchy of the international system and is linked to it. Finally, this approach will be validated by illustrating the applicability of the security dilemma in cases involving states and non-state armed groups. Anarchy, Warlords, and Neorealism Waltz specically addresses the issue of failed states and anarchic systems. After describing anarchic and hierarchic systems, he asks the question: what about borderline cases, societies that are neither clearly anarchic nor clearly hierarchic? Do they not represent a third type? (Waltz 1979, 116). In particular, he addresses the period of Chinese history from the 1920s to 1940s. Waltz notes that,
Nominally a nation, China looked more like a number of separate states existing alongside one another. Mao Tse-tung in 1930, like Bolshevik leaders earlier, thought that striking a revolutionary spark would start a prairie re. Revolutionary ames would start across China, if not throughout the world. Because the interdependence of Chinas provinces, like the interdependence of nations, was insufciently close, the ames failed to spread. So nearly autonomous were Chinas provinces that the effects of war in one part of the country were only weakly registered in other parts. Battles in the Hunan hills, far from sparking national revolution, were hardly noticed in neighboring provinces. The interaction of largely self-sufcient provinces was slight and sporadic. Dependent neither on one another economically nor on the nations center politically, they were not subject to close interdependence characteristic of organized and integrated politics (Waltz 1979, 116).

This reading of Chinese history leads to questions about the nature and boundary of anarchy in general. Waltz goes on to note that [a]s a practical matter, observers may disagree in their answers to such questions as just when did China break down into anarchy, or whether the countries of Western Europe are slowly becoming one state or stubbornly remaining nine (Waltz 1979, 116). Nevertheless, it is clear that it is necessary to make a decision as to whether they

298

Anarchy, Failed States, and Armed Groups

are one or the other, since the decision will lead to very different conclusions about how to analyze the states politics. As Waltz remarks,
The point of theoretical importance is that our expectations about the fate of those areas differ wildly depending on which answer to the structural question becomes the right one. Structures dened according to two distinct ordering principles help to explain important aspects of social and political behavior (Waltz 1979, 116).

The following section will make the case that states containing warlords or other types of autonomous armed groups should in fact be considered anarchic systems. In order to do this, the section will examine the growing literature on failed states and armed groups. Failed States Failed states fail in that they stop fullling the requirements of statehood in whatever way we dene a states requirements. There are many ways in which a state may fail, including the inability to provide welfare, to defend its borders, or to represent a people. In particular, it is often noted that they do not have what Max Weber (1958) sees as the central feature of statehooda monopoly on the legitimate use of forcewithin the territory formally under their control. There are countless theories as to specic factors leading states to weaken and fail.3 For instance, William Reno (1998, 2000) argues that in some cases the rulers of a state will systematically privatize the functions of the state and loot its resources, thereby leaving a hollow shell, something he calls the shadow state. Alternatively, the state may have weak control over specic areas due to geographic reasons. For instance, Jeffrey Herbst (2000) provides a compelling account of how low population densities in sub-Saharan Africa have made it especially difcult for states to maintain authority. A state may not have the economic resources to control signicant portions of its territory. Without economic resources, the institutions and infrastructure of the state, from roads to civil servants, cannot be maintained (Mackinlay 2002, 47). Protracted conicts, begun for different possible reasons, may also eat away at a states ability to govern. No matter the reason for failure, in some instances, the state will fail to the extent that the government loses some or all of its control over its territory. These are the states which are of theoretical importance here. Even Western states may lose some degree of control over their interiors in some instance, such as during natural disasters like the Katrina Hurricane. Nonetheless, such temporary losses of control are quickly subdued. In failed states, the loss of control is much more long term and typically involves not just the temporary diminishing of state control, but also the replacement of control by another armed group. We may refer to states that have lost their authority over denable pieces of their territory as fragmented states and the areas outside the states central authority as fragmented areas. An example of a fragmented state is Columbia, with its large swathe of Revolutionary Armed Forces of Colombia (FARC) controlled territory. In some instances, the state may lose control over most or all of the territory outside of the capital city, as occurred in Liberia during that countrys civil war. In the extreme case of collapsed states the state has failed in that the sovereign government is missing altogether.4 For example, in Somalia all remnants of
3 4

See Rotberg (2004) for more on failed states. See Zartman (1995) for more on collapsed states.

Anthony Vinci

299

a state apparatus had been destroyed by the time of the 1993 UN intervention and no effective government has reappeared since. Rather, there are multiple sub-state groups that control sections of territory or segments of the population.5 The same goes for much of Afghanistan during the early 1990s as well as early twentieth-century China. Autonomous Armed Groups The failure of a state is not necessarily a one-way street however. As the state weakens, armed groupsnon-state organizations that have the capacity for systematic military action6can become relatively more powerful. Within the framework of IR theory, armed groups are non-state actors (NSAs) in that they are, at least in principle, autonomous from the structure and machinery of the state, and of the governmental and intergovernmental bodies above the formally sovereign state (Josselin and Wallace 2001, 3).7 These actors include many different types of organizations ranging from multi-national corporations (MNCs), such as Shell Oil, to nongovernmental organizations (NGOs), such as Oxfam. Armed groups are qualitatively different from other NSAs in that they are completely autonomous and independent of the state and not subject to its authority whatsoever.8 Examples of such armed groups include the SPLM A, which was effectively an autonomous actor for over two decades.9 Another example was the National Patriotic Front of Liberia (NPFL), an armed group controlled by the warlord Charles Taylor, which maintained control over parts of Liberia and Sierra Leone during the Liberian civil war.10 Autonomous armed groups are not subject to regulation by states because the state fails and the armed group becomes powerful enough to rival its authority. As noted, the state fails as it loses control over the population within its territory. Eventually, the states authority and, lacking that, control, are weakened to the point where another actor can exert its own control.11 In such instances, the armed group can wrestle away its own autonomy through the use of military force. In practice this means that the state can no longer expect to regulate the actions of the armed group, as it might regulate the actions of an interest group or other sub-state group through the use of laws or policing. On the other hand, the armed group can expect to control the relations of sub-groups within its own domain, such as villagers who live within a territory that it controls. When interacting together, the armed group and state will use the same means to interact as states use with each other, i.e., the diplomat and soldier (Keohane and Nye 1971, ix). The evidence for such autonomy can be seen in armed groups
See Vinci (2006a) for some examples of the different sorts of armed groups which arose in Somalia. This denition excludes mobs and criminals, which may be violent, but are not systematic or do not have a military ability per se. The types of armed groups contained under this denition include: insurgents, warlord organizations, terrorists, and militias. 7 Also see Huntington (1973) and Keohane and Nye (1971). 8 Some armed groups are not autonomous, as they will be wholly controlled proxies of states. This article does not apply to such armed groups, since they are best studied in terms of the states foreign policy, although it does apply to armed groups which are partially funded and potentially inuenced by external states. States too are provided some aid by other states, and their policy may be inuenced by this aid, however, this is far from assuming that they are proxies. For the same reason, it is acceptable to treat some partially funded armed groups as autonomous and concerned with autonomy. 9 For more on the SPLM A, see for instance, Johnson (2003) and Oystein (2005). 10 For more on the NPFL, see for instance, Reno (1998) and Duyvesteyn (2005). 11 Stephen Krasners (1999) formulation of the denitions of authority and control are useful here. Authority involves a mutually recognized right for an actor to engage in specic kinds of activities (Krasner 1999, 10), whereas control can be achieved simply through the use of brute force with in no mutual recognition of authority at all (Krasner 1999, 10).
6 5

300

Anarchy, Failed States, and Armed Groups

ability to act against the interests of states and in their own interests, most obviously in terms of continuing to make war. For example, the Lords Resistance Army (LRA) has continued to wage its war against Uganda without the blessing of Sudan, while at the same time, maintaining control over the lives of people in its sphere of inuence (Vinci 2007). These armed groups attempt to control the states territory because, like states, they are concerned with maintaining their autonomy.12 Some types of armed groups, such as warlords, are concerned with maintaining autonomy because autonomy is the root of their power and it provides them with signicant economic and other benets. An illustrative example of a warlord is Mohammed Sayid Hersis (better known as General Morgan) and his motives during the period in which he ruled Kismaayo. Clearly Morgan was a warlord in control of territory, a United National Development Ofce for Somalia (UNDOS) report notes that General Morgan has been able to maintain control over the town and prevent the rise of rival civil leaders (UNDOS 1998, 44). He did so for fear that he would lose power to any rival civilian government, since in peace, most observers and local citizens believe that General Morgan would quickly lose his local support (UNDOS 1998, 44). In other instances, warlords have made millions of dollars from the exploitation of natural resources in areas they control. Other types of armed groups, such as guerilla insurgencies, do ght in order to overthrow a government or otherwise cause political change, yet they cannot do so without rst maintaining their autonomy. For instance, Isabelle Duyvesteyn notes that in Liberia the Economic Community of West African States (ECOWAS) Monitoring Group (ECOMOG) wanted to prevent Taylor from taking power and to disarm his troops (Duyvesteyn 2005, 30). However,
Taylor viewed disarmament as a threat to his claims to leadership. The ECOMOG interference was a threat to his almost unstoppable advance and military success. To counter this threat, Taylor decided to attack and strive to expel the ECOMOG forces from Liberia (Duyvesteyn 2005, 30).

This logic is fairly straightforwardTaylor may have had explicit, alternative goals, but the only way he could possibly carry them out was to rst deal with the existential threats to him and his organization. In other words, as with states, the armed group must rst be concerned with survival and maintaining autonomy before it can be concerned with other goals.13 In order to maintain this autonomy, armed groups in effect become functionally undifferentiated.14 They institute economic systems in order to gain the resources necessary to perpetuate their organizationa point amply noted in the war economy literature.15 For example, Taylor made millions of dollars through foreign trade in conict diamonds and timber. These funds were in turn used to fund the weapons, equipment, and paychecks of the NPFL. Armed groups also develop a system of command and control with which to direct their organization and at times an administrative structure to govern populations under their controlin effect, a praetorian government.16 Beyond simple administration, armed groups can institute and maintain fairly complex aspects of
12 See Vinci (2006c). For a specic case study on what may be referred to as existential motivations, see Vinci (2007). 13 This is not exceptional for armed groups, as Martin Van Creveld (1991, 145) notes, over time any war will tend to turn into a struggle for existence, provided only hostilities are sufciently intensive and casualties sufciently heavy. 14 See Vinci (2006b) for an overview of the organizational and other problems which must be overcome in order to perpetuate an armed group. 15 See, for instance, Berdal and Malone (2000) and Reno (1998). 16 See, for instance, Kasr (2002).

Anthony Vinci

301

civilization. For instance, Taylor controlled an area known as Taylorland in Liberia for years, in which he maintained a system of laws, a bank, and even a TV station.17 Armed groups also develop recruitment and motivation capabilities in order to gain and retain members.18 In some cases this may involve popular enlistment, in other cases, such as with the LRA, it involves abduction and traumatic indoctrination.19 Finally, they have the military ability to defend themselves from competitors. This military force can range from poorly trained thugs with machetes and raiding practices to disciplined soldiers with heavy weapons and conventional tacticsdepending on the relative military capacity of the hoststate. These functions are wide-ranging and variable, with some armed groups, like SPLM A, having something very similar to a traditional (if weak) state government in place, whereas some armed groups, like the LRA, have only the most basic services. Of course, some armed groups are very basic and rely on the services of a government to exist; this is especially the case with smaller terrorist cells. Nevertheless, some armed groups can maintain autonomy for long periods of time and are clearly not dependent on any state government for their existence, although they may be funded by external states. Also, in some instances, armed groups may rely on alliances with external states to fund themselves, yet just as with smaller states the armed group maintains control over its actions to some degreeusually to the frustration of the funding stateand can break the partnership if offered a better deal by another state. The armed groups military ability allows it to maintain status as the highest authority over its internal relations. It is able to continuously control territory and or a population and, in doing so, comes to create and maintain its own hierarchical system. For example, in Somalia, some factions maintain sharia law courts in order to mediate disputes. The military ability of the militias allow them to enforce these rulings. Similarly, in the aforementioned Taylorland, Taylor and the NPFL maintained absolute authority over those within its territory because the Liberian state did not have the military ability to interfere. Likewise, since the autonomous armed group has the military ability to rival the state, it can maintain its own, separate, foreign relations for the simple reason that the state cannot stop it from doing so. In this way, armed groups interact with the international environment directly rather than having their interactions regulated by a state. Thus, for instance, the SPLM A was able to maintain its own relations with the Ethiopian state, amongst others, as well as the United Nations vis-a-vis Operation Lifeline Sudan (OLS). ` These relations are reciprocated because of the groups de facto authority. For instance, the SPLM A representative to the African Union (AU) notes: in order to mediate, the AU or [Intergovernmental Authority on Development] IGAD needs a counterpart to Sudan. A government representative.20 The SPLM A provides such a representative. Similarly, border issues with Ethiopia demand representation from the de facto authority, not the Sudanese government, which has no inuence in the border region. Again the SPLM A representative remarks: [Ethiopia] need[s] to know who is responsible. If there is a problem on the border, they need a contact ofce; an authority to deal with the issue.21 Although there is clearly a bias by the SPLM A to demonstrate its importance, the point makes logical sense: if the armed group is in control of a region, that armed group must be dealt with in that region. No matter how much Ethiopia
17 18 19 20 21

See for instance, Reno (1998). See Vinci (2006b). See Vinci (2005). Interview, SPLM representative, Addis Ababa, March 24, 2005. Interview, SPLM representative, Addis Ababa, March 24, 2005.

302

Anarchy, Failed States, and Armed Groups

would like to uphold the sovereignty of its neighbors (since it too relies on the maintenance of a sovereign system of states) it still must face the reality that the government of Sudan did not have any say in what goes on in that particular region. The SPLM A like many other armed groups seeks ofcial recognition because such ofcial recognition is in essence a form of power. By being ofcially recognized, an armed group can maintain easier access to gaining military or other supplies. For example, the former leader of the National Union for the Total Independence of Angola (UNITA) Jonas Savimbis trip to visit Ronald Reagan assured him of continued support. Similarly, by controlling the capital of a state, the armed group assures itself of the same decorum of ofcial recognition which allows it to gain the advantages of traditional states, such as the ability to take part in ofcially licensed international trade or to discredit its own enemies. This was exactly the strategy employed by Taylor, as described in detail by Reno (1998). At other times, the armed group may be recognized by some but not all states, as the Taliban was.22 In other words, juridical recognition becomes a form of power like a strong military or large economy. Nevertheless, such power is a luxury in that the armed group can exist without it; instead what matters is their de facto authority. In a sense, these autonomous actors are empirically sovereign, in that they are the highest authority over a territory or population, although they are not juridically sovereign, in that they are not granted the rights and responsibilities of sovereignty by the international community.23 The idea that being autonomous and the highest authority over internal and external relations engenders sovereignty goes back to the origins of the debate over the denition of sovereignty. Jean Bodin, the French lawyer and political thinker, noted he is absolutely sovereign who recognizes nothing, after God, that is greater than himself (Bodin 2002, 271). While in most cases, empirical sovereignty overlaps with juridically dened borders, this is not the case for autonomous armed groups. As such they are the opposite of Robert Jacksons (1993) concept of quasi states, which possess juridical sovereignty, but not the other markings of sovereignty. Although such exclusivity of authority may not be enough to produce the international recognition of sovereignty, it does have an effect on the decisions of other actors. For example, when a state like Ethiopia needs to solve a border issue, it does not refer to the ofcial representative of the state, but rather to the organization which is in control of the border. Therefore, empirical sovereignty is also a factor in theoretical accounts of international relations because it is the factor which determines the relations of units in the international anarchic system. This point will be returned to below. Domestic Anarchy In order to rectify this disjunction between the states weakness and armed groups autonomy, it is necessary to admit that collapsed and fragmented states are not necessarily hierarchical systems, but rather that they are anarchic. As such, the description of hierarchic units needs to be made at a level below the juridically dened borders of the state. It is clearly the case that collapsed states are anarchic systems. The anarchic system emerges through a positive feedback loop in which armed groups arise in a state and the states authority weakens. In some cases these groups become fully autonomous and the state collapses completely. At this point
22 23

See Rashid (2000). As such, we might refer to these actors as sovereign non-state actors (SNSAs).

Anthony Vinci

303

there is in fact no central authority whatsoever, no leviathan in the Hobbesian sense. The domestic system is thus anarchic by denition, as different self-regarding actors are interacting without a higher authority to structure their relationships. The new unit boundaries should then be dened as being at the level of the armed group and the territory population it controls. In the case of fragmented states, the formulation of anarchy is slightly more complicated. To begin, it is necessary to differentiate the state, which is the entity that is granted juridical sovereignty, from the state shell, which is the entire territorial area which is demarcated as a separate unit in the international system. In a fragmented state the state maintains authority in an area which is less than the entire area of the state shell. Within this area, there exists a hierarchic system, in that the state apparatus is the highest authority and structures the relationships of sub-actors.24 Outside of this area, but within the state shell, one or more autonomous armed groups control some or all of the remaining territory population. These armed groups structure the relationships within their own, separate hierarchic systems. Even though these armed groups exist within the state shell, the state cannot maintain authority over them and therefore cannot structure their relations. Of course, the state will deny the autonomy of any armed group within its boundaries, in particular it will strongly object to any admission of the juridical sovereignty of the armed group, and will in general act to defeat the armed group in any way that it can. Nevertheless, the state will act as if the armed group were an autonomous actor, i.e., in the same terms as other actors in the international system, specically states. Therefore the state and the autonomous armed group interact as equals and thus the system within the state shell is anarchic. The state and armed groups remain differentiated in that the state has both empirical and juridical sovereignty, whereas the armed group only has empirical sovereignty. Nevertheless, the relations between the state and the armed group are determined by empirical, not juridical sovereignty. Thus, as with collapsed states, the system should be dened as anarchic.25 For example, before the recent peace agreement, the SPLM A controlled signicant amounts of territory within the shell of the Sudanese state. It maintained authority within these liberated areas and structured the relationships of subactors, such as tribes and villages. The government of Sudan, which ruled from Khartoum, was forced to relate with the SPLM A as an equal and therefore could only relate in terms of force or negotiationas it would with other states in the international system. The borders between the state and armed groups may shift over time. As with ordinary states at war, conquest may force one side to lose control of an area and another to gain it. Such shifts can occur much faster than the juridical denition of borders. Thus, as with Europe during World War II, we must redraw the map of actors in the system continually. The same goes for collapsed states, except that in such states the map is determined by the interactions of armed groups with other armed groups, not with the state.

24 This is not to confuse the state with the government, for the government of the state may change, just as it does in proper states. 25 To put it metaphorically, it is as if the doors to a house have been open, letting in the ood waters of anarchy but there is still one room, its doors tightly shut, in which the waters have not found their way. It is here that the family continues to live, as other groups slop around the ooded remainder of the house also attempting to shut themselves in and make their own sovereign kingdoms. Yet, while the outside community may still consider the house private property, the neighbors may not practice what they preach.

304

Anarchy, Failed States, and Armed Groups

Test of Anarchy

It is, however, necessary to be cautious in the analysis so as not to see anarchic systems where they do not exist. David is right in saying that anarchy cannot simply be assumed rather, most states most of the time can ensure compliance (David 1997, 559). States regularly face threats to their monopoly over the legitimate use of force. Criminals, for example, are dened by their attempt to break the states laws and often do so violently. However, criminals are not autonomous and their actions do not create anarchy. When the state addresses the issue of crime it does so as a sovereign to an element within its hierarchy. The differentiating factor is that the criminal does threaten the effective control of the state; in fact it relies on the state for providing the hierarchical system in which to exist. As John Mackinlay puts it, the Maa live as citizens of a free society in most cases, and their freedom to move and communicate is not guaranteed by their own military strength, but by the institutions of the state (Mackinlay 2000, 7). In contrast, the autonomous armed group does rival the state and has the military power to maintain its autonomy. In order to identify states that do not maintain a hierarchic system within the entire state shell, this study proposes a simple theoretical test. An area is fragmented if an ofcial of the state government (in the guise of an ofcial acting in an ofcial manner) cannot enter the area due to the presence of an autonomous armed group, which is in control of the area and has the ability to overpower any force the government can muster, in the sense that it can completely defeat it or force it to leave the area. This is not necessarily a one-time event, but a maintainable state of affairs.26 Thus, the state must not only be weak, but there must also be a stronger rival able to confront and overpower it. In this way, the test is able to determine the de facto boundaries of units within the domestic anarchic system.27 For example, at one extreme a taxman would easily be able to go door to door anywhere in Belgium because he would be considered legitimate and the state monopoly on the use of force is strong. At the other extreme, Somalia would clearly be considered anarchic because there is not even a government to take the test. In much of Columbia, the state does structure the relations of sub-state actors, however, in the FARC controlled areas a tax collector would not be legitimate, nor could the state provide the coercive ability necessary to enter the area because FARC has a relative (local) military superiority. This area would therefore be considered fragmented and Columbia a fragmented state. Similarly, Hezbollah has controlled many areas of Lebanon for long periods of time. But, if a taxman was in a Kurdish area of Turkey, where a Turkish state ofcial might not be considered legitimate, he could go in with enough protection from the government to not be under threat because the government is signicantly stronger than any possible Kurdish resistance. Therefore, we should not consider Turkey as a fragmented state.

This long-term state of affairs is what differentiates the situation from short-term victories in battles. In cases where armed groups are more mobile, i.e., they control people but not territory, the test is able to give us an idea of whether or not the government is able to exert authority over individuals in a similar manner. If the state is able to rely on authority over an individual or otherwise control his or her actions, it is sovereign over that individual. If the state is unable to do this, since there is another organization which has authority or control over the individual, it does not have sovereignty over the individual. The organization of individuals which are not under the states authority is, in a sense, another sovereign, and the relations between this organization and the state is marked by anarchy, rather than hierarchy. This would be the case in northern Uganda, for instance, where the members of the LRA are clearly not under the authority of the Ugandan government. Therefore Uganda should be considered a fragmented state. The evidence is in the fact that in order for the government to confront these individuals, it relies on military units, as it would if it were to confront the troops of another state.
27

26

Anthony Vinci

305

Closed vs. Open Anarchic Systems Conventionally domestic anarchy is seen as a separate closed system within a state, in the sense that it is insulated from and different in nature than the international anarchic system. As noted, authors such as Posen and David have drawn the parallel between domestic anarchy and international anarchy. In general, these authors are attempting to explain internal wars, and more specically, ethnic conicts. They conceptualize the state in terms of anarchy in order to use the notion of a security dilemma to illustrate the actions of ethnic or other groups within a failed state. However, they do not link this domestic anarchy with international anarchy. For example, David notes that when central authority collapses in a state a microcosm of the international system is replicated within the state (David 1997, 557). The point to note here is that David considers that the system is similar, but separate from the international system. The implication is that this difference would make integration of domestic actors and international actors within the same theoretical model impossible. This is an important point because without explicitly making that connection, we are still left with two separate theoretical explanations, one that holds inside the state and one that holds outside. The primary reason for this separation is that it is assumed that domestic anarchy is fundamentally different from international anarchy. In particular, the difference is thought to stem from the different origins and history of anarchy within, versus outside of states. For example, Nelson Kasr notes that within the state, anarchy does not have a priori status; rather, it emerges when the state fails and it disappears when state authority returns (Kasr 2004, 60). The implication here being that since anarchy is not the natural state of affairs, actors will make decisions differently and therefore, standard IR theories of their behavior will not work. However, the argument for the exceptional nature of domestic anarchy due to its origins fails for a few reasons. Empirically speaking, it is only contingent that a domestic hierarchical system will return to all areas of a state. In Somalia hierarchy has not returned for over a decade. Moreover, it can be argued that true hierarchy never really existed in some areas of states such as Afghanistan or the Democratic Republic of Congo (DRC). In general, hierarchical systems have been reied, as if it has always and will always exist in the same borders that they have now. However, hierarchy is itself a contingent property of a system and depends on historical circumstances for any particular area. For instance, former Soviet states were once hierarchically aligned, but now have an anarchical relationship with each other, and it is possible to imagine a future in which they are hierarchically associated again. Similarly, in Somalia the assigned borders of the state once did not include the area known as Somaliland and we can imagine a future in which Somaliland is not included again. Therefore, we should not see the juridically dened borders of a state as permanent, but rather accept the possibility that it may settle into separate units and vice versa. Another objection is that neither the past nor the assumed future state of anarchy or hierarchy necessarily impacts on the present actions of groups within an anarchic system. For instance, when an empire collapses or decolonizes, even though the states were once part of a hierarchic system, their actions will convert to reect the anarchic environment. In the same way, even though a state may assume that it will one day be part of a hierarchic system, for instance as a member of the United States of Europe, it will continue to act as if in an anarchic system. The parallel applies to those in collapsed states. Groups may remember being part of a hierarchy, but the immediate necessities of life under

306

Anarchy, Failed States, and Armed Groups

anarchy demand decision-making as if, and only as if, living under anarchy for an indenite and potentially permanent period of time. The hidden assumption made in thinking that domestic anarchy is temporary is that actors want a hierarchical system and will work to create and maintain one; however, this is not necessarily the case. It must be kept in mind that anarchy does not necessarily mean the presence of chaos and violence and therefore, actors may potentially maintain an anarchic system indenitely.28 The assumption is that actors will recreate hierarchy if they can because they do not want to live in an anarchic system. However, anarchy only connotes that there is no central authority, not that there are no authorities whatsoever. Rather, there may be other authorities within the anarchic system. These sub-authorities may control hierarchic systems, as for instance the SPLA controlled its own people and territories within the state of Sudan. These authorities may in fact not be in conict with each other. For example, even at the height of its collapse, there were still local authorities operating throughout Somalia29 and there has been a return to widespread peace in the sense of a lack of overt ghting, even though there is no effective central government. The same is true in the international systemalthough it is anarchic, it is generally not very chaotic. Thus, when a state collapses we should not think of the resultant situation as violent disorder, but as anarchic because there is always another authority to take control of a piece of territory. Therefore, actors could potentially live in domestic anarchy indenitely, as there is not enough pressure to force them back into a hierarchical system. It might also be held that domestic anarchy is separate from international anarchy because the two systems are still separated by the juridically dened borders of the state, regardless of state failure. However, this is not the case either as armed groups, the state, and external states and international actors treat the state shell as open, or permeable, and the two systems as linked. Armed groups clearly treat fragmented and collapsed states as an open anarchic system, which is intimately connected with the international system. They do not respect the authority of the state and treat it as a rival, not as the uppermost authority in a hierarchical system. Therefore these armed groups do not let the presence of any state or state-like authority stop them from interacting with other actors in the international system. For example, the SPLM A made diplomatic connections around the world without taking into account the borders of the Sudanese state or its authority over foreign policy. In fact, armed groups have controlled territory in multiple states, for example the NPFL had effective control of areas in Liberia and Sierra Leone. In a fragmented state, the states government will not legitimize an armed group by calling it autonomous, i.e., sovereign, but its actions against armed groups demonstrate the reality of it being an autonomous actor and potentially having its own separate external relations. (This is not an issue in collapsed states, since there is no government to speak of). For instance, the state will generally try to refer to armed groups as criminals or terrorists, or otherwise de-emphasize the fact that the armed group has de facto authority in a particular area. However, in its actions, the state will have to act in accordance with the reality of the autonomous armed groups authority. It will not try to enter the areas under control of the armed group except to combat it, as, for example, the government of Sudan did when ghting the SPLM A. In other words, the state treats the intrusions into armed group controlled areas as invasions. Accordingly, while the state may wish to regulate the international interactions of the armed group, it cannot possibly do so and must simply accept the reality of its no longer being
28 29

I.e., it is not chaotic in the sense that Robert Kaplan (1994) has referred to it. See, for instance, Menkhaus and Prendergast (1995).

Anthony Vinci

307

a buffer between these actors and the international system. Thus, while the Sudanese government may have objected to the diplomatic connections of the SPLM A with foreign governments, it could not prevent them. For this reason, we regularly see states admit that they cannot control drug dealing or other international actions emanating from armed groups within their dened territory. Actors outside the domestic anarchic system also treat it as openly anarchic. While international actors have attempted to hold up the sovereignty of such states and demand that other states do so as well, the fact is, their actions do not point to such a strong view of juridical sovereignty. It is increasingly common for states to intrude upon the sovereignty of failed states, thereby demonstrating the lack of a states effective control over territory. In particular, state militaries interact in domestic anarchy as if it were international, as for example happened in U.S. snatch and grab operations in Somalia.30 Other states also invade at willnot in order to attack the state, but to attack specic armed groups. This is, for instance, what happened in Africas World War in the DRC, when Uganda and Rwanda (amongst other states) invaded the DRC to attack the Interahamwe and other militias. Even international organizations make the admission that the state shell is permeable in some cases. For instance, the OLS agreement between the UN, Sudan, and the SPLA demonstrated that an actor other than the state could be considered as a legitimate actor to be dealt with within a sovereign state and outside of its permission. Or, in order to enter southern Sudan, a visa is not obtained from the Sudanese state, but from the SPLA.31 International aid organizations regularly accept the necessity and legitimacy of obtaining the required visa. As such, they effectively treat the armed group as the sovereign unit, not the juridically dened state. In addition, both states and international organizations also implicitly admit the de facto nature of armed group control. As noted above, some states excuse drug dealing or terrorism that erupts from areas within their territory, which are out of their control. The international community will generally accept this as reality and not hold it against the state, either legally or morally. This is happening in Afghanistan for instance, where the U.S. continues to have diplomatic relations even though a large percentage of the worlds heroin originates from within its borders. Similarly, the ECOMOG intervention applied economic sanctions to NPFL controlled areas of the Liberian state, but not to other areas, and in doing so, admitted the de facto authority of the NPFL.32 While this is not a formal recognition of the reality of an armed groups de facto control, it is an implicit admission. We should conclude that domestic anarchy is real and linked to the international anarchic system. There is a uid connection between the domestic actors and international actors, in that they interact as part of the same system. Actors may not rhetorically refer to the system as anarchy, and juridically speaking it is not. Empirically speaking, however, international actors do treat actors in domestic anarchy as they would actors in the international anarchic system. Implications for Analysis In a sense, this perspective involves taking the analysis of anarchic systems from the juridical level, in which actors are dened in terms of recognition by the international community, to the empirical level, in which actors are dened by their de facto sovereignty over groups. Although actors remain within the juridi30 31 32

See, for instance, Menkhaus (2004). Authors eld observation, Nairobi, Kenya, and Rumbek, Sudan, April 2005. On the intervention, see Alao, Mackinlay, and Olonisakin (1999).

308

Anarchy, Failed States, and Armed Groups

cal boundaries which are set out by international law and tradition, these actors relate in all ways as if these were not the boundaries. Instead, they relate in terms of the de facto boundaries of power and these boundaries of power can be within or overlapping the juridical boundaries. Thus, in effect, the map of sovereign units is redrawn. Yet it is redrawn with the same exclusivity as the juridical map of sovereigntythereby maintaining separate, autonomous units.33 These autonomous armed groups will act in essentially the same way as states. Armed groups will attempt to maintain their autonomy as a route to carrying out other possible activitiesjust as states will, for the logic of the armed group in anarchy is the same as that for the state in anarchy. As Posen notes, the condition of anarchy makes security the rst concern... It can be otherwise only if these political organizations do not care about their survival as independent entities, (Posen 1993, 104) or as Waltz puts it, self-help is necessarily the principle of action in an anarchic order (Waltz 1979, 111). This reversion to the logic of anarchy has been pointed out by others; for example, David notes that when the state collapses, domestic groups inside a country behave much as states do in the international system (David 1997, 55455). Yet, these authors only consider such similar relations within the state. However, the relations also hold between actors inside and outside of the juridically dened borders of the state, for the relations are not based on the juridical boundaries of power but rather on the empirical boundaries of power which may be drawn within or overlapping the juridical borders of states. The systemic logic of a theory like Neorealism is not based on the juridical boundaries of a state. It is based on the de facto realities of power. For instance, balances of power will form between autonomous actors seeking their continued survival. This logic will hold for essentially similar autonomous actors regardless of what international law dictates to be the boundaries of a state. The implication is that it will hold for certain armed groups that interact with states. The relations between the armed groups and external states will be governed by the same logic.34 Therefore, the actors will be constrained to relate in essentially the same way as two actors in the conventionally conceived international system. Since their decision will be made based on the same logic of action, they should form relationships with states in a similar manner as states form with each other.35 Since these relationships are similar, the analysis of these relationships should be comparable. Thus, the practical application of these insights is that by treating domestic anarchy as a continuation of the international anarchic system, it is possible to integrate the understanding of internal armed groups with external actorsincluding states and other armed groups.36 In particular, these relations can be described using the same systemic theories. Systemic theories such as Neorealism are based on the presumption that self-regarding, autonomous units in an anarchic system will act in predictable ways. Based on this logic, certain predictably patterned relationshipsor law-like regularitiesmay arise in the relations of the actors, and these theories explain and predict those regularities.
33 I.e., there is no overlapping sovereignty as might be called for by, for instance Ruggie (1986) or Cerny (1998). 34 This hypothesis has been theoretically demonstrated here and illustrated to some extent with short example cases. A next step in this research is to demonstrate it in more detailed case studies. 35 It might be held that because armed groups are so much less powerful than states they will not be able to form such relationships. However, some armed groups are more powerful than microstates or failed states. Therefore we should expect that if a theory holds for a microstate or failed state, it should also hold for an armed group. 36 It may even be possible to treat all domestic environments as anarchies, which are more or less under the authority of regimes. This would of course be a fundamental shift in the focus of International Relations theory. While such a shift might help to solve some problems in the relationship between domestic and international politics, clearly it is of a much larger scope than this article.

Anthony Vinci

309

While we typically only look at the relationships of states, the predicted relationships also hold between autonomous armed groups and states. This provides an extremely different analysis of the relations between armed groups and each other and between armed groups and states than is found in the theories used by other theorists who address armed groups, such as Reno. Often, theorists looking at the relations of armed groups will turn to greed or grievance explanationswhich are not applicable to stateswhereas this approach turns to the traditional explanations of Neorealism, such as the balance of power. In doing so, the entire Neorealist literature opens up to provide potential explanations for the behavior of armed groups. For instance, one relationship that theorists have given particular attention to concerning the nature of domestic anarchy is the security dilemma, but as this article has illustrated, it is also possible to posit the existence of security dilemmas between actors in domestic and international anarchy. The Security Dilemma and Armed Group-State Relations The security dilemma arises any time there are one or more units in an anarchical system that are motivated by the desire to remain autonomous (Waltz 1979).37 It arises because states, or other self-regarding units, will attempt to provide for their own security and in doing this they will amass power for defensive purposes. This accumulation of power may be perceived to be for offensive rather than defensive purposes by other actors. Offensive power capability is considered to be a threat and leads the other actors to amass power themselves for defensive purposes but, again, this may be perceived as being for offensive purposes and it is herein that the spiral of the security dilemma begins. Typically the security dilemma is applied to the relations of states, however as already noted, Posen (1993) applied the security dilemma to non-state groups, specically to ethnic conict.
Domestic Security Dilemma

Posen notes that the collapse of imperial regimes can be viewed as a problem of emerging anarchy (Posen 1993, 104). He argues that in the collapse of empires, such as the Soviet Union, the sovereign disappears and sub-state groups, whether ethnic, religious or cultural, will seek security. These actors will then relate in terms of a security dilemmawhich will potentially spiral into conict, just as it does with states. Jervis and Snyder (1999) also use the concept of the security dilemma to explain the relations of actors in a failing empire. They add to this a discussion of the interaction of predatory and security motives. For example, they note that:
The purest type of security dilemma is a situation in which security is the overriding objective of all the protagonists, yet attempts by one party to increase its security reduce the security of the others. At the opposite end of the spectrum some conicts may be driven entirely by the desire of one or both parties to exploit or dominate the other for reasons that would not diminish even if security were not in jeopardy (Jervis and Snyder 1999, 19).

Jervis and Snyder (1999) admit that there are normally mixtures of security and predatory motivations by actors. They nd that the security dilemma itself can give rise to predatory behavior, as well as zero-sum conicts of interest and inimical exclusive identities (Jervis and Snyder 1999, 23). At the same time,
37

See Jervis (1978), for more on the security dilemma.

310

Anarchy, Failed States, and Armed Groups

predatory strategies can create or exaggerate the security dilemma. However, they do not deny the value of the concept and its applicability to armed groups. The point is that although there may be some modication, the security dilemma works in essentially the same way with non-state armed groups as it does with states. Mixed Security Dilemma The security dilemma was applied by Posen and others to domestic armed groups relating to each other. However, as has been argued, the anarchic system which armed groups exist in is open, and therefore external states and armed groups may be involved in the security dilemma. We can dene a mixed security dilemma as one between states and armed groups, or armed groups and other armed groups which exist across state borders. The mixed security dilemma can occur when armed groups exist in both collapsed and fragmented states. In a collapsed state, the armed group may interact with other armed groups and states outside of the state shell. For instance, in Somalia such a mixed security dilemma occurs between Somali warlords and Ethiopia. The Ethiopian state felt threatened by the warlords within Somalia because they are obvious security threats that may have either supported internal Ethiopian threats, as with al Ittihad, or may be direct threats themselves. At the same time, Ethiopias defensive maneuvering could be perceived as a threat to warlords, as for instance, Ethiopias alliance with the Somali National Front (SNF) was perceived as a threat by the Somali National Army (SNA). In fragmented states, the armed group may enter into a security dilemma with the state itself, as well as with neighboring states and armed groups. For example, the LRA presents a threat to Uganda simply by existing as an entity which is not under the authority of the state. The LRA also presents a security dilemma for the Sudanese state, as it may potentially also attack targets inside Sudanas actually occurred after the government of Sudan stopped funding the LRA and it, in turn, began attacking government forces and villages (Integrated Regional Information Networks (IRIN) News 2005). Finally, the LRA has also been a threat to the SPLM A (International Crisis Group (ICG) 2004).
Mixed Security Dilemma and Intervention

An example of the benet of the mixed security dilemma approachwhich is not evident in the traditional security dilemma or the internal security dilemma described by Posen or others theorizing about internal security dilemmasis that it can help explain intervention in a manner logically consistent with a Neorealist approach. When states fail and armed groups arise inside of them, from a traditional Neorealist perspective we might expect neighboring states to see this as a benecial state of affairs because the potentially threatening state is weakened. Therefore it would seem like states should accept the presence of failed states, possibly even promote them, and certainly not intervene in them to end this less threatening situation. At the same time, Posen and others approach to seeing a separate internal and external security dilemma does not provide insight into how an external state might react with an internal armed group, and vice versa, how an internal armed group might react to an external state. Interventions do happen, however, and they make sense if one considers that an armed group can enter into a security dilemma with a state. For example, as Taylor grew in power, he came to represent a threat to neighboring states, and eventually this threat led to an intervention by other states with security concerns in the region. Similarly, Ethiopia considered al Ittihad a threat to its own security, and together with the SNF, invaded Somalia and destroyed the military

Anthony Vinci

311

capability of al Ittihad in 199697. Whether or not these armed groups actions were for defensive purposes, and there is, for instance, reason to belief that al Ittihads were,38 as expected in a security dilemma, the neighboring states would perceive the military build up of these actors as threatening. The action used to obtain an increase in security in the face of the threat was the intervention. This again seems to be happening with the recent invasion of Somalia by Ethiopia in its war against the UIC, where Solomon Abebe, Ethiopian foreign affairs spokesman, said After too much patience, the Ethiopian government has taken a selfdefensive measure and has begun counterattacking the aggressive extremist forces of the (Islamic council) and foreign terrorist groups (Hassan 2006). There are multiple advantages of such a mixed security dilemma approach which arises out of the ability to theoretically integrate our understanding of internal and external relations. Systemic theories tend to be at a loss when dealing with interventions because the units in the system are traditionally held to be at the juridically dened state level. Yet, as noted above, Ethiopia was reacting to and had to defend itself from a sub-state actor. By taking an empirical level of sovereignty into account, however, it makes perfect sense as to why Ethiopia would intervene. Similarly, the domestic anarchy approach fails to provide a framework that incorporates the possibility that not only other armed groups inside of a state would feel threatened by al Ittihad or the UIC, but also that Ethiopia might feel threatened. Similarly, al Ittihad not only felt a threat from the SNF, but also from Ethiopia. From the perspective of the mixed security dilemma, the more recent intervention of Ethiopia into Somalia would easily have been predicted, since the circumstances of the intervention were essentially similar to the earlier intervention from a systemic level. Most importantly, such an approach may potentially provide theoretically predictive insights in the same manner that Neorealism may be able to provide such insights about the structure of state relations. Just as we may predict that balances of power may form between states using war, alliances, or other means, we can predict that armed groups and states may form balances of power through war, i.e., interventions. This may have happened with the recent invasion of Somalia by Ethiopia. Potentially, such an intervention could have been theoretically predicted based on an analysis of the relative power of Ethiopia and the UIC. The route to such a theoretical modeling and prediction would not have been available with a purely internal approach to domestic anarchy as Posen uses or by the greed-grievance approaches used by Reno and others. As such, this approach does initially seem to provide reason for continued exploration and application. Conclusion This article has presented one possible solution to the dilemma of dividing the world into hierarchic and anarchic systems. It has done this through expanding the anarchic system to include actors inside the state shell, but not under the control of the state, and thereby treating the sub-state armed groups as autonomous units. In effect, this study has rearranged the anarchy hierarchy divide along empirical, rather than juridical lines. This is a different approach than some others have taken to theoretically framing the complicated post-Cold War, globalized international system, where armed groups and failed states play a more signicant role. Philip Cerny (1998) has used the model of neomedievalism to describe the overlapping sovereignties
38 UNDOS points to some anecdotal evidence that the al Ittihad movement in the Gedo region did not want any association with terrorist activities inside Ethiopia, and was angry over an Islamist assassination attempt in Addis Ababa, for obvious reasons... (United Nations Development Ofce for Somalia (UNDOS) 1998, 146).

312

Anarchy, Failed States, and Armed Groups

in the current globalized system. Mark Dufeld (2001) has applied Cernys theory to the analysis of Africa and failed states more generally. The implication of these arguments is that the breakdown of the world into exclusive anarchic and hierarchic systems is not possible, and therefore that systemic theories like Neorealism cannot be used to effectively analyze large parts of the international system.39 The approach of this study presents a picture of more subtly dened sovereignties, but no less rigid ones. In this way, it has been possible to continue to assume the strong break between hierarchic and anarchic systems, and thereby to incorporate a seemingly blurry system into the more mechanisticand one might argue more rigorousview of systemic theories like Neorealism. There are practical reasons for taking this approach. This article briey illustrated the use of a mixed security dilemma, but other concepts could also be applied. For instance, the notion of balance of power may be usefully applied to the relations between states and non-state armed groups. For example, such a balance of power seems to have existed in East Africa where Sudan had aligned with the LRA to balance the power of the alliance between the SPLA and Uganda. Another example of the value of applying international relations theory to armed groups is in explaining the war between Ethiopia, the UIC, and to a lesser degree, the United States. Such an example of a mixed actor war that crosses borders makes much more sense when described in terms of actors attempting to maximize their security through alliance formation and making war on threatening neighbors. Systemic theory based analyses of armed groups could draw on many of the insights that have arisen from debates concerning interstate relations. For instance, it would be possible to debate when an armed group would bandwagon and when it would not, or when an armed group might use soft-balancing rather than hard-balancing, just as it is possible to have such a debate with regard to states.40 The next step in this line of research is to address the limits of the application of international relations theory to theorizing about the relations of armed groups. This would involve signicant research into the nature of armed groups. For example, can we treat them as like units and therefore truly apply a Neorealist approach? This line of research would necessitate addressing areas of research which have traditionally been outside the bounds of IR as a discipline. For example, it would be essential to review the Development and Internal Wars literature in order to understand the nature of armed group motivations, i.e., the GreedGrievance debate.41 Such bridges between International Relations and other subjects could be very fruitful and provide increased understanding on both sides. In general, it is necessary to develop a better understanding of the nature of international politics that includes the relations of armed groups. Warlords, insurgents, and terrorists are becoming some of the most signicant strategic threats for numerous states. At the same time, these armed groups are also taking part in diplomacy and other non-military international relations, and these relations must be integrated into our understanding of international politics. Clearly a less state-centric approach to the study of international relations is needed; this article has provided some basis for such a project, one which does not necessitate leaving behind the heretofore helpful systemic level approach.

39 Earlier, Ruggie (1986) made a more general argument that the medieval system had a different balance of anarchy and hierarchy than is present now and that Neorealism was not equipped to understand such a system. 40 See, for instance, Brooks and Wohlforth (2005). 41 See Vinci (2006c) for an example of how this argument might look.

Anthony Vinci

313

References
Alao, Abiodun, John Mackinlay, and Funmi Olonisakin. 1999. Peacekeepers, Politicians, and Warlords: The Liberian Peace Process. New York: The United Nations University. Berdal, Mats, and David Malone, eds. 2000. Greed and Grievance. Economic Agendas in Civil Wars. London: Lynne Reinner Publishers. Bodin, Jean. 2002. Six Books of the Commonwealth. In International Relations in Political Thought: Texts From Ancient Greeks to the First World War, ed. Chris Brown, Terry Nardin, and Nicholas Rengger. Cambridge: Cambridge University Press. Brooks, Stephen G., and William Wohlforth. 2005. Hard Times for Soft Balancing. International Security 30(1): 72108. Bull, Hedley. 1966. Society and Anarchy in International Relations. In Diplomatic Investigations: Essays in the Theory of International Politics, eds. Herbert Buttereld, and Martin Wight. London: George Allen and Unwin. Buzan, Barry, Charlie Jones, and Richard Little. 1993. The Logic of Anarchy. New York: Columbia University Press. Cerny, Philip G. 1998. Neomedievalism, Civil War, and the New Security Dilemma: Globalization as Durable Disorder. Civil Wars 1(1): 832. David, Steven R. 1997. Internal War: Causes and Cures. World Politics 49(1): 552576. Dufeld, Mark. 1997. Post-Modern Conict, Aid Policy and Humanitarian Conditionality. Presented at the United Kingdom Department for International Development (DFID), London. Dufeld, Mark. 2001. Global Governance and the New Wars: The Merging of Development and Security. New York: Zed Books. Duyvesteyn, Isabelle. 2005. Clausewitz and African War: Politics and Strategy in Liberia and Somalia. London: Frank Cass. Hassan, Mohamed Olad. 24 December 2006. Ethiopia Attacks Somalia Islamic Council. Associated Press, Associated Press Online. Herbst, Jeffrey. 2000. States and Power in Africa: Comparative Lessons in Authority and Control. Princeton: Princeton University Press. Huntington, Samuel P. 1973. Transnational Organizations in World Politics. World Politics 25 (April): 333368. Integrated Regional Information Networks (IRIN) News. 12 April 2005. Sudanese Government Troops Attack LRA in the South. New York: UN Ofce for the Coordination of Humanitarian Affairs. International Crisis Group (ICG). 2004. Northern Uganda: Understanding and Solving the Conict. Nairobi Brussels: ICG Africa Report No. 77. Jackson, Robert H. 1993. Quasi-states: Sovereignty, International Relations, and the Third World. Cambridge: Cambridge University Press. Jervis, Robert. 1978. Cooperation under the Security Dilemma. World Politics 30(2): 167214. Jervis, Robert, and Jack Snyder. 1999. Civil War and the Security Dilemma. In Civil Wars, Insecurity and Intervention, ed. Barbara Walter, and Jack Snyder. New York: Columbia University Press. Johnson, Douglas H. 2003. The Root Causes of Sudans Civil Wars. Oxford: The International African Institute in association with James Curry and Indianapolis University Press. Josselin, Daphne, and William Wallace. 2001. Non-State Actors in World Politics. London: Palgrave. Kaplan, Robert D. 1994. The Coming Anarchy. The Atlantic Monthly. 4476 February. Kasr, Nelson. 2002. Dilemmas of Popular Support in Guerrilla War: The National Resistance Army in Uganda 1981-86. Presented at Laboratory in Comparative Ethnic Processes (LiCEP) 6, University of California, Los Angeles (UCLA), November. Kasr, Nelson. 2004. Domestic Anarchy, Security Dilemmas, and Violent Predation: Causes of Failure. In When States Fail: Causes and Consequences, ed. Robert I. Rotberg. Princeton: Princeton University Press. Keohane, Robert, and Joseph Nye. 1971. Transnational Relations and World Politics. Cambridge, MA: Harvard University Press. Krasner, Stephen D. 1999. Sovereignty: Organized Hypocrisy. Princeton: Princeton University Press. Mackinlay, John. 2000. Dening Warlords. International Peacekeeping 7(1): 4862. Mackinlay, John. 2002. Globalization and Insurgency. Oxford, London: Oxford University Press for the International Institute for Strategic Studies. Menkhaus, Ken. 2004. Somalia: State Collapse and the Threat of Terrorism, Adelphi Paper: 364. Oxford: Oxford University Press.

314

Anarchy, Failed States, and Armed Groups

Menkhaus, Ken, and John Prendergast. 1995. Political Economy of Post-Intervention Somalia. Somalia Task Force Issue Paper #3. Oystein, Rolandsen H. 2005. Guerilla Government: Political Changes in Southern Sudan in the 1990s. Uppsala: Nordic Africa Institute. Posen, Barry R. 1993. The Security Dilemma and Ethnic Conict. In Ethnic Conict and International Security, ed. Michael E. Brown. Princeton, NJ: Princeton University Press. Rashid, Ahmed. 2000. Taliban: Militant Islam, Oil and Fundamentalism in Central Asia. New Haven, CT: Yale University Press. Reno, William. 1998. Warlord Politics and African States. London: Lynne Rienner Publishers. Reno, William. 2000. Shadow States and the Political Economy of Civil Wars. In Greed and Grievance. Economic Agendas in Civil Wars, ed. Mats Berdal, and David Malone. London: Lynne Reinner Publishers. Rotberg, Robert I. ed. 2004. When States Fail: Causes and Consequences. Princeton: Princeton University Press. Ruggie, John Gerard. 1986. Continuity and Transformation in the World Polity: Toward a Neorealist Synthesis. In Neorealism and its Critics, ed. Robert O. Keohane. New York: Columbia University Press. United Nations Development Ofce for Somalia (UNDOS). 1998. Study on Governance in Gedo Region. Nairobi, Kenya: United Nations Development Programme (UNDP). Van Creveld, Martin. 1991. The Transformation of War. New York: Free Press. Vinci, Anthony. 2005. The Strategic Use of Fear by the Lords Resistance Army. Small Wars and Insurgencies 16(3): 360381. Vinci, Anthony. 2006a. An Analysis and Comparison of Armed Groups in Somalia. Review of African Security 15(1): 7690. Vinci, Anthony. 2006b. The Problems of Mobilization and the Analysis of Armed Groups. Parameters 36(1): 4962. Vinci, Anthony. 2006c. Greed-Grievance Reconsidered: The Role of Power and Survival in the Motivation of Armed Groups. Civil Wars 8(1): 2545. Vinci, Anthony. 2007. Existential Motivations in the Lords Resistance Armys Continuing Conict. Studies in Conict and Terrorism 30(4): 337351. Waltz, Kenneth N. 1979. Theory of International Politics. Reading, MA: Addison-Wesley. Weber, Max. 1958. From Max Weber: Essays in Sociology. Oxford: Oxford University Press. Zartman, I. William. 1995. Collapsed States: The Disintegration and Restoration of Legitimate Authority. London: Lynne Rienner Publishers.