What Determines the Army's Reaction to an Uprising?

The recent uprisings in the Middle East and North Africa confirmed, yet again, that neither social scientists nor politicians and intelligence analysts are good at predicting revolutions. We may be able to observe that a country or region is "pregnant with revolution," to use Lenin's memorable phrase, but we have no idea when, if ever, a revolution might actually break out. Indeed, it is amazing how autocracy may persist for decade after decade, only to have a seemingly trivial event trigger a massive upheaval that, on rare occasions, might span an entire region: in January 1848, students in Sicily distributed leaflets that criticized the monarch, in reaction to the closing of their university or, in December 2010, the humiliation of a fruit vendor in Tunisia by a low-level municipal official.

Although we might continue to puzzle over what sparks revolutions, we do know one critically important thing about them: once they start, they can seldom succeed without the support of the regime's coercive apparatus, most particularly, the regular army. One of the main arguments of this book is that military responses to uprisings, in fact, largely decide their outcomes. Social scientists generally dislike monocausal explanations with broad applicability; however, I am not suggesting that the military's response to a revolution is the only predictor of whether it will succeed in supplanting the status quo regime or not. Rather, I argue that the military's disposition toward the revolution is the most important predictor of its outcome, and the army's support is certainly a necessary, if not sufficient, condition for a revolution to succeed. Consequently, if we want to contemplate the fate of revolutions, then we must take the next logical step and answer the fundamental question of why armies react to revolutions the way they do? What factors sway their actions? Is it possible to predict a military's reaction to a revolution in a specific context?

Consider the following scenario: you are an analyst at an intelligence agency and your assignment is to advise the president on the action the armed forces are expected to take in Country X, which is experiencing a revolutionary upheaval. (Let's just assume that the president is a bright person who already knows that the way the military will go will probably decide the fate of the uprising.) Where will you start looking for answers? The rest of this chapter will give you the tools — in the form of questions you must ask — that you will need to produce a coherent and logical analysis and provide the most informed report possible.

Which Factors Matter, How Much, Why, and When?

Political science is not a predictive discipline and most, though certainly not all, social scientists would shudder to forecast political events and societal processes given the plethora of factors that would need to be considered. Nevertheless, I believe that we might be able to anticipate — or, at the very least, make a highly educated guess about — the army's response to a revolution if we know enough about that army, the state it serves, the society it comes from, and the international environment in which it exists. If we possess this knowledge, then we ought to be able to suggest — with, to be sure, a degree of accuracy that would vary according to the given context — how the armed forces are going to behave. Given that the army's response to a revolution is critical to determining the revolution's fate, we should also be in an excellent position to offer conjecture about the revolution's outcome.

So, the sixty-four-thousand-dollar question is "How will the generals respond to the uprising?" Will they support the old regime or the rebels, or will they split their support between the two? Deciding which factors best explain the generals' responses to revolutions is a formidable intellectual challenge. One must analyze a potentially large number of variables that interact in complex ways, and all explanatory factors are not created equal: some go much farther in explaining the armed forces' reactions to revolutions than others. The relative significance of these variables can and will differ from case to case. One country's generals might have to take into consideration the prospect of foreign intervention as they decide how to react to an uprising, while in another country, meddling from abroad may not be a factor at all. Similarly, ethnoreligious differences within the armed forces may be critical in shaping one army's response to a revolution in one setting, but in an army that is composed of personnel of the same ethnic origin and faith as the rebels, those issues will have no significance whatsoever. Moreover, these factors might be reinforced or weakened by circumstances that have a bearing on revolutionary outcomes in some states but not in others.

To be able to make an educated guess regarding an army's response to an uprising, one must be familiar with the given context. There is no clever model that can tell us, once we "plug in" all the appropriate variables, what a military will do in a crisis. Ultimately, there is no way around the sobering reality that the weight of each factor is determined by its individual context. There is no shortcut, no substitute for having an in-depth knowledge of the individual case. That said, it is possible to identify some useful generalizations that allow us to distinguish variables that are more important in most cases from those that usually have less impact on the generals' decision-making.

Several scholars have attempted to explain considerations that might enter into the military's decision to respond to a revolution, either listing germane factors or privileging one variable or another in their analyses. In her classic study, Chorley favored military attributes, such as professionalism, political leanings, the latent disaffection of the officers, and the conscripts' willingness to fight as well as their propensity to fraternize with the enemy. Her selection of these criteria was rather haphazard, and her attention was basically limited to military characteristics. Russell, although clearly aware of the importance of identifying which impulses drive the generals' decisions to shoot or not, only lists some potential variables — social-class composition of the armed forces, the proportion of officers to men, recruitment criteria, promotion opportunities — but leaves a more thorough examination to future researchers.

In articles published in the October 2011 and April 2013 issues of the Journal of Democracy, I outlined my initial thoughts about the generals' responses to revolutions: the first concerned the "Arab Spring" specifically with regard to the settings in the Middle East and North Africa, and the second tackled the subject more generally, though discussed the independent variables in far more detail. My work was informed by the important contributions of Chorley, Russell, and a few others and has continued to develop as I have become ever more convinced that only a more comprehensive — if necessarily more difficult and labor intensive — approach allows the would-be analyst to make confident predictions. Nevertheless, I have found the work of scholars who emphasize individual factors, as opposed to a more inclusive argument, quite useful as a test of my framework. The crucial variable for Eva Bellin, and for many military sociologists, is the "professionalization" of the armed forces — in other words, the level of the military's institutionalization. She argues that a highly institutionalized/professional army — that is, one that is rule-bound, in which promotion is based on merit, where there is a clear separation between private and public realms to discourage corruption and predatory behavior, and that has no allegiance other than to the constitution and the national interest — is going to be more likely to support reform and less likely to defend a corrupt regime against widespread popular revolt. The opposite of institutionalization, what Bellin calls "patrimonialism," depicts a military marked by favoritism, corruption, and politicoideological loyalties that negate institutional rules and undermine meritocracy. This patrimonial army would be more likely to oppose reform movements, as it will remain loyal to its benefactor and the status quo regime.

The problem with this explanation is that many highly institutionalized armies tend to exist in democratic polities, where few revolts ever target overthrowing democratic rule. The very appropriateness of the "institutionalized-patrimonial" dichotomy to help explain the high command's loyalties might also be questioned. There are many examples of exceedingly professional militaries that have defended corrupt and fundamentally unjust political regimes — examples can be found in some of the armies of the communist world. What the professional/institutional-versus- personalistic/patrimonial contrast addresses is better explained by the military decision-makers' views of the regime in power or, put differently, their appraisal of the regime's legitimacy. If authoritarian regimes can persuade — a better phase might be "successfully indoctrinate" — their professional officers to view them as legitimate, they are likely to defend those regimes.

Furthermore, as Derek Lutterbeck correctly notes, Bellin's institutionalization-patrimonialism scale also possesses little explanatory power to illuminate cases where low institutionalization might lead to a fracturing of the armed forces. He suggests that another factor — the distinction between armies based on mandatory conscription as opposed to volunteers — is more critical to explaining military loyalties during uprisings. Drafted soldiers, after all, have far more in common with the citizens in revolt than with the professional armed forces that is called upon to suppress them. As Chorley remarks, "Soldiers are under arms for a short period only, and regard their military duty merely as an incident in their lives rather than the exclusive purpose of their working years." Conscripts are less likely to shoot at crowds that might include their friends and relatives than are volunteers. "Inhibitions to the use of coercion," as Tilly put it, "are likely to increase when the coercive forces are drawn from (or otherwise attached to) the population to be controlled." Nevertheless, while the question of whether an army is made up of conscripts or enlisted soldiers is an important factor, it does not offer an infallible method for anticipating a military's reaction to an uprising. No individual factor does, but the conscript- professional dichotomy is undoubtedly a key explanatory variable in many contexts.

Social scientists have long recognized the cohesion of the officer corps and especially the unity among senior officers — the bonding and integration between officers, the length of their military service — as another critically important variable that helps illuminate the generals' responses to rebellions. Cohesion and esprit de corps are the most important parts of what we think of as a military culture that is also composed of discipline, professional ethos, and ceremonial displays and etiquette. Cohesion generally refers to relationships between soldiers on the unit level, while esprit de corps denotes the commitment to and pride in the larger military force to which that unit belongs. A recent study hypothesized that military cohesion depends on two essential factors: the degree of control the regime holds over its citizens and the level of organizational autonomy armies possess that allows them to focus on training and war-fighting. Most experts differentiate between social cohesion, which denotes personal bonding between members of a group, and task cohesion, which sets and imposes behavioral standards, "sustains the individual in the face of stress," and promises to diminish that stress through the assistance, cooperation, or collaboration of other group members. The armed forces' internal cohesion is directly relevant for those studying a military's response to an uprising.

When an army is tasked to put down a rebellion, its soldiers must be prevented from defecting to the enemy. Theodore McLauchlin distinguished between two basic strategies for maintaining the troops' loyalty: individual incentives — that is, rewards and punishments — that are vulnerable to a "cascade" of defections across the army when uprisings erupt, and group-based strategies, such as ethnic preferences that might produce out-group defection but generally make in-group defection less likely. In an individualized incentive system, soldiers' beliefs in the regime's chances of survival — a belief that is largely based on other soldiers' behavior — would either result in their loyalty or, if they thought the regime would collapse, provoke mass defection, "bringing about the very collapse it predicts." In a group-based strategy, on the other hand,

the belief that in-group members have a proregime preference and that out-groups are opposed to the regime helps generate precisely those preferences. Thus preferences can become public by matching public prejudice. The result is a durable cleavage between in-group and out-group, where out-group soldiers are likely to defect but in-group soldiers are likely to remain loyal.

Although McLauchlin might be right that these self-fulfilling prophecies might explain some experts' erroneous assumptions about the expected resilience of regimes under stress, his is, as he concedes, a unidimensional approach that cannot explain the diversity and complexity of the large variety of cases. In many uprisings, different ethnic or even ethnoreligious identities of the soldiers — a key issue in McLauchlin's scheme — is simply not an issue at all. More fundamentally, his reliance on soldiers' beliefs — something that is rather difficult to correctly and reliably ascertain — limits the usefulness of this approach. My hope is that the comprehensive framework I outline below will allow us to more accurately anticipate the army's behavior in domestic conflicts.

To be sure, cohesion is one of the essential explanatory variables, but it, too, has significant shortcomings, and it is certainly not an "omnipotent" variable. First, there are numerous different disruptions that could affect the cohesion of the officer corps, from officers' educational backgrounds to the more common "senior officer vs. junior officer" divide. These splits within the officer corps might be minor, or they might be quite significant, depending on the setting and the particular source of schism. Second, ascertaining the true level of the officer corps' unity often requires a great deal of field research and personal access to military personnel. Taking all this into consideration is necessary because the credibility of official information regarding this particular issue is seldom beyond suspicion: few governments affected by revolutionary upheaval would concede that their armed forces are fraught with internal divisions, lest they should aid their opponents.

Contemplating officer corps cohesion brings up another crucially important issue. Most authors who write on the military's behavior in revolutionary environments restrict their analyses to only one segment of the military, the officer corps. For example, a recent essay by David Pion-Berlin and his coauthors on why armies might or might not stay quartered during civilian uprisings fails to mention the nonofficer component of the armed forces at all. This is a mistake. Elite-centered analyses overemphasize the significance of the officer class while ignoring a major dimension of the armed forces. If officers cannot persuade noncommissioned officers (NCOs) and the enlisted/conscripted soldiers to follow their orders, even the army with the most unified officer corps will fail to reach its objectives.

In other words, we must assess the cohesion and integration among all uniformed military personnel, not just those within the officer corps. Just as those who write on tactics, training, combat readiness, and social origins of the armed forces cannot overlook the NCOs and ordinary soldiers of an army; those who write on armies and revolutions ignore them at their peril, as well. We cannot make an accurate prediction of the army's response to a revolution if we do not give due consideration to those whose job it is to carry out the decisions of the officers or, put differently, to the question of whether soldiers be willing to turn their guns on peaceful protesters or not. Aside from institutional cohesion, the aforementioned article by Pion-Berlin and his colleagues identified several potential causes that might account for military disobedience during civilian crises, such as the army's material interests, professional identity, and what they call "legalities." Some of the problems of this otherwise very useful study are that it is overly deterministic in identifying the causes of the military's behavior and misses some of the important nuances of its numerous cases. Most importantly, the essay overemphasizes military-centric factors at the expense of state-related, societal, and external variables that often have a decisive impact on the generals' decisions and are nearly always seriously considered by them.