In the meantime, however, we can note that even if the government improves its service delivery in this particular elder’s area, reduces corruption, does a better job of law enforcement, and creates a more consistent local presence in his area—key objectives of counterinsurgency theory, and exactly what we were trying to achieve across Afghanistan for much of the past several years—this may not help. At this point, having availed himself of the Taliban justice system, the elder’s claim to a valuable disputed asset depends on the guerrillas’ authority, and that makes it extremely difficult for him to support the government or go against the insurgents, whatever his feelings about the state. The elder in this context is a fish in the trap: he’s locked into an incentive structure that is easy and attractive to get into, but hard and painful, if not impossible, to escape. This example, then, shows a real-world competitive control system in action. The next section unpacks the theoretical basis for such a system, and as we shall see, it has broad applicability beyond insurgency.
III. The Theory of Competitive Control
The pattern in which nonstate armed groups create predictability and order as a way of generating popular support, and build incentive systems to corral target populations, is not unique to Afghanistan. Indeed, it’s a near-universal phenomenon, present in insurgencies but also in many other social systems. It is, in fact, utterly characteristic of the relationship between local populations and nonstate armed groups of all types, and is thus an excellent place to begin our search for a set of ideas beyond classical counterinsurgency that may more fully explain the patterns of conflict we see around us.
Before we discuss the theory, I should make a brief comment on terminology. Throughout this book, I use the term “nonstate armed group” to refer to any group that includes armed individuals who apply violence but who aren’t members of the regular armed forces of a nation-state. I draw this formulation in part from Charles Callwell’s late-nineteenth-century definition of small wars, and I prefer it to the more common terms “nonstate actor” and “illegal armed group.”20 Nonstate armed groups can include urban street gangs, communitarian or sectarian militias, insurgents, bandits, pirates, armed smugglers or drug traffickers, violent organized criminal networks, vigilantes and armed public defender groups, terrorist organizations, warlord armies, and certain paramilitary forces.21 The term encompasses both combatants and individuals (for example, facilitators, covert operatives, or political cadres) who don’t personally carry arms or use violence but who belong to groups that do. Terrorists, insurgents, and militants—such as the LeT raiders in Mumbai—who are sponsored but not directly employed by a nation-state form a special subcategory of irregular proxies or state-surrogate groups.
The basis for the control systems applied by nonstate armed groups of all kinds, including the Taliban approach described above, is what I call the theory of competitive control. We can formally express this theory as follows:
In irregular conflicts (that is, in conflicts where at least one combatant is a nonstate armed group), the local armed actor that a given population perceives as best able to establish a predictable, consistent, wide-spectrum normative system of control is most likely to dominate that population and its residential area.
Simply put, the idea is that populations respond to a predictable, ordered, normative system that tells them exactly what they need to do, and not do, in order to be safe. We’ve already seen this system in operation at the hands of criminal groups in the “System” of Jamaican garrison communities, in the actions of clan militias in Somalia, and in the operations of Taliban courts in Afghanistan. The feeling of safety that this predictability creates, in a chaotic and dangerous environment, trumps everything. As a result, even people who would otherwise dislike an armed group (such as the people of Wardak in the kidnapping example, or the people of Mogadishu or Tivoli Gardens in Chapter 2) end up supporting a group because of the order it creates.
Besides explaining one possible mechanism for the support-follows-strength pattern that Stathis Kalyvas observed, this theory suggests a behavioral explanation for the way in which armed groups of all kinds (insurgents and criminals alike) control populations. It also suggests that group behaviors, such as the rule-based Somali swarming tactics I mentioned in Chapter 2, may be emergent phenomena at the level of the population group (rather than rational or conscious choices by the individual), implying that traditional counterinsurgency notions, including “hearts and minds” (the belief that populations can be swayed if their conscious choices are influenced), may need a rethink. Changing the rule set and incentives that define emergent behaviors may be a better option, if indeed such change is possible, rather than attempting to influence conscious choices. Likewise, it may be that there’s a window of opportunity during which such changes can work, but that after this window closes, a population may be locked into an incentive structure that is extremely difficult to shift.22
Sucking the Population In
The Afghan example is just one illustration of an underground control structure (a form of illicit social control, sometimes referred to in classical counterinsurgency theory as a “parallel hierarchy” or “guerrilla government”). Of course, insurgents aren’t the only ones who use systems like this, and nonstate armed groups don’t simply wait until a genuine dispute emerges, then resolve it to gain popular support. On the contrary, they deliberately create disputes, promoting insecurity or fear, precisely in order to pose as saviors and thereby win local allegiance. The very problems that a group inflicts on a community can thus translate into community support. The group itself is the disease it purports to cure.
This way of looking at illicit control structures is far from a new idea. As early as June 1927, the Employers’ Association of Chicago coined the term racketeering to describe a form of social control in which a criminal entrepreneur’s organization gains support from populations (local businesses, residents, or labor unions) by promising to protect them from the very problem of violent crime that the criminal organization itself creates, thereby locking them into a structure that alienates them from the state, while creating incentives for continued support and silence. This is exactly what we saw in Tivoli Gardens, or in the Afghan case just mentioned.23 Insurgents in this sense behave much like gangsters: “In a situation [in 1920s Chicago] where the legal system offered little security, organized crime provided regulatory and mediation services—what some have come to call ‘licensing.’”24 In Afghanistan the overlap between crime, insurgency, tribal or business patronage networks, and government itself is often so extensive that the same individuals play roles in multiple hierarchies.
Analysts looking at underground Communist parties in the 1930s described similar methods of social control by clandestine political movements. Revolutionary cells would absorb and mobilize recruits, blood them in violent street confrontations with rival groups, groom them through a series of increasingly illicit actions, lead them progressively into an ever-greater level of illegality and alienation from society, and thus make it harder for them to betray or leave the movement lest they be punished by the government whose laws they had broken.25 The RAND analyst Philip Selznick wrote in 1951 that in a revolutionary organization, “the emphasis on illegal work creates a conspiratorial atmosphere; this has the dual consequence of disintegrating normal moral principles, thereby reducing inhibitions that might hamper manipulability, and of increasing the dangers (real or imagined) of leaving the organization.”26 Selznick described this process as one of “absorption,” in which an individual is gradually brought further and further into the control system of the underground movement. In our analogy, this is akin to being drawn into the trap.
As part of this process, armed groups often deliberately make local populations or new recruits complicit in acts of violence, as a way to alienate them from the government or other communities, so that they have no choice but to support the dominant group in their area. A 2004 st
udy of former child soldiers in Uganda, for example, found that many new recruits were forced to kill innocent people—in some cases other recruits or even members of their own families—as part of the induction process.27 Compelling children to commit acts that society condemns and the law forbids not only desensitized the children to extreme violence but also separated them from society, bonded them to the insurgent group, and made them feel they could never go back. Some groups cut or otherwise scarred the faces of children who had killed in this way, marking them for life so that they could never go home. Similar rituals are, of course, well known as part of the initiation processes of gangs, organized crime families, and terrorists. Gang tattoos, in this sense, fulfill the same social marking function as the face scarring inflicted on the African child soldiers. In the gangs of San Pedro Sula, discussed in Chapter 1, killing a member of a rival gang or being “jumped in” (beaten by the group for a ritually significant length of time—thirteen seconds for MS13, eighteen seconds for Calle 18, and so on) also became important steps in the induction process.
Diego Gambetta, in his fascinating study of criminal communication, Codes of the Underworld, notes similar initiation behaviors among mafia families, prison gangs, insurgents, and drug trafficking organizations. Gambetta points out that mafia novices are usually required to kill someone as part of their induction process, as a test of loyalty. “The mafia usually does not kill anyone purely for the sake of a test—it optimizes by ‘whacking’ someone who was meant to be whacked anyway and at the same time trying out the determination and bona fides of the novice. The Aryan Brotherhood in prison adopted the same test: to gain membership, candidates ‘had to kill whomever the Brotherhood targeted.’” Gambetta also described cases in which recruits were asked to commit atrocious crimes purely as tests. In youth gangs in Colombia, for example, “it is not uncommon for new gang members to be asked to murder innocent friends or members of their own family, which pushes the test to the extreme.”28 As we’ve seen, though, for many nonstate armed groups, such violence is more than a test of loyalty—it’s a way of separating a new recruit from outside society.
An even starker example of this same technique was the way that nationalist militias in the Balkans in the 1990s forced local people to kill their neighbors. Militias would round up people from another religion or ethnicity, then assemble residents from their own group and force them, sometimes at gunpoint, to massacre their neighbors.29 In 1992 in the northeastern Bosnian town of Brčko, for example, Serb militias rounded up Bosniac Muslim men, women, and children and forced local Serbs—these people’s neighbors, who had known them their whole lives—to kill them, right then and there, in the street. According to eyewitness testimony, given in 1993, the process of ethnic cleansing in Brčko began on May 2, 1992 with the arrival of Serbian “special units” including a group we will return to later, Arkan’s Tigers:
In these first couple of days several hundreds of innocent [Bosniac] civilians were killed in the following locations: in the local police station, behind Posavina Hotel, in the town’s trades center and near the Brka river. Furthermore, people were taken from their houses and killed on their own threshold, yard or on a street. Those Serbs who did not assent to those crimes were executed without any mercy. There are many examples how Brčko Serbs were forced to kill their Muslim neighbors in order to preserve their own lives.30
This happened in dozens of places in 1992–93 and had the effect of forcing the population to become complicit in mass murder and ethnic cleansing in their own villages.31 After the killing, militia leaders would tell people they now had no choice but to join the movement, because they could never reconcile with the families whose relatives they had killed. My soldiers and I saw the horrendous effects of the same thing in East Timor in 1999, as militias forced local people to kill members of their own communities before forcibly marching them across the border into West Timor, thereby assuring themselves that the expelled population could never return.32 Likewise, Shining Path guerrillas in Peru gathered villagers and forced them to stone local government representatives and prominent villagers to death. This made the villagers complicit in illegal, violent, collective action—cutting them off from the state and putting them at the mercy of the movement.33 All these examples represent extreme, and extremely effective, forms of coercion within an overall system of control over the population.
In a similar vein, in a December 1964 lecture at the U.S. Naval War College, the classical counterinsurgency theorist Bernard Fall had this to say about insurgent control:
Any sound revolutionary warfare operator (the French underground, the Norwegian underground, or any other anti-Nazi European underground) most of the time used small-war tactics—not to destroy the German army, of which they were thoroughly incapable, but to establish a competitive system of control over the population. Of course, in order to do this, here and there they had to kill some of the occupying forces and attack some of the military targets. But above all they had to kill their own people who collaborated with the enemy.34 [Italics added]
Fall was talking of the coercive end of a spectrum of incentives—the application of lethal force against those who collaborate with an insurgent’s rivals. Any sensible guerrilla, terrorist, or organized criminal network will of course attack soft targets (the unarmed civilian population that supports a government) in preference to attacking the government itself. Besides being a less risky target than the police or the military, the civilian population is more numerous and easily accessible than are government installations and officials.
But Fall’s point here is larger than simply a comment on the mechanics of violence. Note his language: a competitive system of control. Fall never developed his concept fully—he was killed in February 1967 while accompanying a U.S. Marine Corps patrol near Hue, South Vietnam. But his later writings give a series of examples of this idea of competitive control—an idea that’s not spatial (“insurgent-controlled” or “contested” areas) or structural (“networks” and “movements”) but rather functional. It implies the presence of a range of incentives and disincentives, all of which are used to generate control over population groups—the individual strands of a networked system of control that attracts and then corrals a population, much as a fish trap cages fish. It also implies a competition among several actors who are all trying to control the population in a violent and contested environment.
Normative Systems
I find it helpful to locate Fall’s “system of competitive control” within the broader theoretical discourse of normative systems. The notion of normative systems is long established in sociology and legal theory and, increasingly, in computer science, where developers have found it useful in agent-based modeling (where researchers assign behavioral rules to computer-generated “bots” and then watch how they behave in complex systems).35
For our purposes, we can define a normative system as a set of rules that is correlated with a set of consequences.36 In essence, it’s a system of norms (behavioral rules) that is paired with a set of sanctions (costs and punishments for breaking those rules). This system of norms and sanctions defines the boundaries of permissible behavior for a population. It makes the behavioral space inside its boundaries a safe zone for those who follow its rules, while the space that lies outside the boundaries becomes deeply unsafe. Implicit in this idea is the notion of an actor (an “owner” or proponent of the normative system) who sets the rules, bestows benefits for following them, and inflicts punishments for breaking them, and is thereby able to control a population. According to this theory, the owner of a normative system becomes the dominant actor in a given area (or over a given population) precisely to the extent to which people in that area or population abide by its rules. This applies equally to a wide range of actors: the don of a Jamaican district enforcing his “System,” the leader of a Somali militia, a Taliban court dispensing justice, and a Honduran gang enforcing an extortion racket are
all applying variants of the same norms-based approach.
This actor may be a government or a nonstate group; it may be benevolent or malevolent, legally recognized or illicit, formal or informal. But two characteristics must always be present: the actor must always be armed (that is, it must have the capacity to inflict violence as part of its spectrum of sanctions) and it must always be a group (some form of collective entity), not just an individual. An unarmed actor lacks the capacity both to enforce its own normative system and to resist predation from other armed actors in the violent ecosystem we’ve just described. And enforcing a normative system is fundamentally a group activity, since it involves regulating people’s behavior over a wide area of time and space, a task that lies beyond the capacity of any one individual (as we saw in the Tivoli Gardens example in Chapter 2).
In an environment with only one dominant actor possessing a monopoly on the use of armed violence, we would expect to see an extremely high degree of control over a population or area. This, indeed, is demonstrably the case in areas that are fully controlled by governments. But in the examples we’ve been examining in this book—insurgencies, urban street gangs, informal periurban settlements beyond the direct reach of the state, diasporas subjected to a protection racket by gangs who dominate their town of origin, and so on—we are looking not at uncontested control but rather at a pattern of contested space and at a competition for control among several actors. Each actor tries to create a normative system of competitive control, and the better it does this, the more likely it is to dominate the contested space. Thus we are talking here of competition among organized, armed groups seeking to control populations through normative systems—a construct that applies equally well to insurgency and crime, to state and nonstate actors, and indeed to the actions of states and those of governments.
Within the behavioral space bounded by its rule set, an actor can apply a spectrum of means ranging from persuasion through administration to coercion. At the persuasive end of this spectrum are arguments and inducements to support the dominant rule set. These include propaganda, political and ideological mobilization, social pressure, and identity manipulation. But as we’ve seen, often the most persuasive element is the feeling of security, predictability, order and cohesion (closely related to Ibn Khaldun’s idea of asabiyya, discussed in the Mogadishu example) that comes with adherence to a dominant actor’s norms. In the middle of the spectrum, administrative tools—justice systems, mediation and dispute resolution mechanisms, essential services, social and economic institutions—make it easier for people to follow the rules, and give them tangible benefits for doing so. At the coercive end of the spectrum are punishments that impose costs on people who break the rules. These include punitive violence—up to and including death—as well as expropriation (fines, penalties, or seizure of assets), expulsion and exile, or imprisonment.
Out of the Mountains Page 17
