Why We Were Surprised (Again) by the Arab Spring


Popular uprisings toppled dictators in Tunisia, Egypt, and Libya in 2011 and (as of this writing) are seriously threatening dictators in Yemen and Syria. There have also been massive political protests in Bahrain, which have been met with considerable violence from authorities, as well as large protests in Morocco. Yet no social scientist or political analyst in either the West or the Arab world itself claims to have predicted these uprisings, so far as I am aware. Nor did any Western or Arab intelligence agency predict them. In fact, the very people who participated in these uprisings, including long-time activists, do not seem to have foreseen them. We were all taken by surprise. The question is, why? Might we have predicted these uprisings if we had been paying closer attention and utilizing the best social-scientific ideas? Were area specialists simply devoting too much time to explaining the persistence of authoritarianism in the Arab world (Gause 2011)?

Before addressing these questions, let me point out one particularly egregious yet illuminating failure to predict the Arab Spring—”egregious” because the individuals involved believe their job consists precisely in providing early warnings of political unrest. For some years now, a U.S. think-tank called the Fund for Peace has published, in collaboration with the journal Foreign Policy, an annual “Failed States Index.” This index, the most recent of which covers 177 countries, is a composite of a dozen demographic, economic, political, and military indicators that are supposed to reveal which states around the globe are weak and failing. By so doing, the index is supposed to give an “early warning of conflict . . . to policymakers and the public at large” (2011: 8). The assumption is that political conflict is more likely to occur where one finds weak and failing states.

Does the Failed States Index that is based on data for 2010 suggest that conflict was about to erupt in the Arab world? Not exactly. If anything, the index predicts an African Spring. Zimbabwe, Nigeria, Kenya, Ethiopia, Uganda, and Cameroon are found among the “top” 25 (i.e., the most weak and failing) states in the index. By contrast, only one Arab country (Yemen, ranked thirteenth) falls into this group. Egypt and Syria barely fall into the top 50, and their scores are similar to rather more stable, semi-democratic countries like Colombia and the Philippines. Finally, three Arab countries that experienced a great deal of conflict and violence in 2011—Tunisia, Libya, and Bahrain—actually fall into the “bottom” half of the index, that is, among the allegedly stronger and more successful states. In short, the Failed State Index might have led us to expect political unrest in Yemen in 2011, but not the serial uprisings that will forever be remembered as the Arab Spring or Arab Awakening.

It is of course possible that the researchers behind the Failed States Index are simply using the wrong measures or that their data are not especially accurate. Or it could be that some or even most popular uprisings occur, when the conditions are right, where one finds relatively strong states. I want to argue here, however, that it is simply impossible for social scientists or others to predict precisely where and when revolutions or mass uprisings will occur—and for some very simple reasons. The Failed States Index failed to predict the Arab Spring because it has set itself an impossible task.

So why do mass uprisings always surprise us? Why do revolutions, to paraphrase Tocqueville, seem so inevitable in hindsight, yet are impossible to foresee?

A few years after the Eastern European revolutions of 1989—which, by the way, no one predicted either—Timur Kuran wrote a compelling article titled “The Inevitability of Future Revolutionary Surprises” (1995). Paradoxically, Kuran accurately predicts in this article our collective failure to foresee anything like the Arab Spring. Kuran’s argument emphasizes the importance of “preference falsification,” that is, the fact that people may not reveal publicly their private preferences, whether out of fear or shame. Kuran is not claiming that preference falsification prevents us from knowing whether a population generally opposes a particular political regime. Indeed, he notes that Václav Havel, an important leader of the Czech opposition in 1989, wrote very eloquently about preference falsification under Communism. And it was certainly no mystery that dictators like Ben Ali in Tunisia, Mubarak in Egypt, and Gaddafi in Libya were widely reviled—although just how widely reviled they were beyond one’s immediate social circle was no doubt difficult for the ordinary person in these countries to say.

Preference falsification, according to Kuran, makes it especially difficult to discern “the precise distribution of individual revolutionary thresholds” (1995: 1538), that is, the likelihood that an individual who opposes a regime would join some kind of movement against it if a certain number of others did so as well. Kuran shows that even a slight shift in this distribution, brought about by “a small, intrinsically insignificant event,” may be sufficient to produce a “revolutionary bandwagon” in which more and more people take to the streets, encouraged by the relative safety and anonymity of large crowds (1995: 1533; see also Kuran 1991). Eventually, as the revolutionary movement overflows with supporters, even people who genuinely supported the regime join the opposition out of fear for their future safety, claiming that they privately opposed the regime all along.

Kuran’s analysis clearly helps to make sense of important features of the Arab Spring. We know that something like a “revolutionary bandwagon” occurred in Tunisia beginning in December 2011, following a seemingly insignificant event, namely, the self-immolation of a provincial fruit vendor after his business was shut down by the local police. The example of the Tunisian uprising, which culminated in the dictator Ben Ali’s precipitant flight from the country, then helped to ignite a revolutionary bandwagon in Egypt, which then soon spread to Libya and other countries where opposition to regimes was widespread and revolutionary thresholds relatively low—although the latter factor in particular could not be known beforehand. The fact that revolution did not spread to Algeria, Saudi Arabia, Jordan, or any number of other Arab countries indicates that the distribution of revolutionary thresholds in those countries was simply not conducive to mass uprisings—although, again, no one could have foreseen just where and how far the Arab Spring would travel.

We also know (as Kuran’s theory again paradoxically predicts) that many people who openly supported or worked for the old regimes jumped ship at the proverbial eleventh hour, claiming in some cases that they privately opposed these regimes all along. In Egypt, Mubarak’s once-loyal generals eventually hustled him off the scene and took power for themselves. They have now acceded, however grudgingly, to a public trial of the former dictator. And many people in or close to Libya’s National Transitional Council—the anti-Gaddafi government established in the city of Benghazi early on in the uprising—were former officials or officers in the Gaddafi regime and army, in some cases until after the uprising began.

A very similar argument has been made by Charles Kurzman in his important book, The Unthinkable Revolution in Iran (2004). The title of Kurzman’s book refers to the fact that all but a very few Iranians thought that a revolution against the Shah was inconceivable prior to 1978. They found it impossible to imagine that a movement could arise that would have the capacity or viability to overthrow the Shah. Like Kuran, Kurzman believes that individuals “were more likely to participate in revolutionary events if they felt that many others would do the same” (2004: 171). And many Iranians invested a lot of energy, he shows, in trying to determine if others would participate in rumored protests. Alas, Kurzman notes, individuals’ estimations of how many Iranians might join a protest varied considerably—in part because of what Kuran calls preference falsification—and so too did the willingness of individuals to join a protest on the basis of these guesses. “These estimations,”Kurzman (2004: 170) concludes,

cannot be known in advance; nor can the willingness to participate. They shift drastically from moment to moment on the basis of amorphous rumors, heightened emotions, and conflicting senses of duty. These shifts make retroactive prediction impossible. Even the most thorough survey would not have predicted very far into the future . . . For this reason, revolutions will remain unpredictable.

Kurzman implicitly suggests here a way in which Kuran’s ideas about preference falsification may actually underestimate the contingency and unpredictability of mass uprisings. For Kuran generally writes as if individuals have a considered and precise idea, which they may then conceal from others, of when they would join a revolutionary movement—for example, when they see that 50 other people have done so, or perhaps 500 people. Kuran (1995: 1548) writes:

If we could know exactly how the members of a given society would react to any given structural shock, we could foretell the society’s political future. In practice, however, such knowledge is unobtainable. And an important part of the reason is precisely that individuals often conceal their political desires and motivations.

But Kurzman rightly suggests, if I understand his analysis, that many if not most individuals simply do not know themselves how they would react if they saw 50, 500, or 5,000 people protesting against a hated regime. Their reaction would undoubtedly depend on exactly who those people were and how the regime was responding to them—and on the “amorphous rumors, heightened emotions, and conflicting senses of duty” which Kurzman mentions. In Kuran’s terms, individual revolutionary thresholds are not simply a given, but may shift radically in the space of a few days or even hours. In other words, even if people are perfectly forthcoming about their private preferences, their actions remain exceedingly difficult for others and even themselves to predict.

This applies not only to people who might join a revolutionary movement but also to those whose job it is to suppress revolutions, namely, military officers and soldiers. As Trotsky once noted, and as many have subsequently repeated (e.g., Nepstad 2011), the outcome of revolutions is ultimately determined by the disposition of the armed forces. Usually, disciplined armies are able to suppress uprisings rather easily. But deep crises and persistent protest may erode this discipline. Still, “even on the very eve of the decisive hour,” Trotsky writes, it is impossible to predict which way the army will swing, for “the revolutionary soldiers do not know how much power they have, or what influence they can exert” (1930: 140). In times of crisis, military cohesion, like popular uprisings, can be impossible to predict.

In the case of the Arab Spring, dictators fell relatively quickly and with comparatively little bloodshed when armies ultimately refused to support them. This is what happened in Tunisia and Egypt. Dictators have held on to power much longer and the conflicts have been much bloodier where armies have generally remained loyal to the regime or have fragmented, with at least some significant factions remaining loyal. This has been the path of Syria, Bahrain, Yemen, and Libya. In hindsight, there seem to be some plausible reasons for this divergence. The more professional and institutionalized armies in Tunisia and Egypt calculated that they could best safeguard their interests by abandoning dictators. On the other hand, where the army is (or was) dominated by the dictator’s clan, tribe, religious sect, or ethnicity, the army (or some fragment of it) remained loyal to the regime. But who at the time could predict how any particular army would react to mass protest? Who could have predicted that the Tunisian army would refuse to fight for Ben Ali? Who could say with confidence that the Egyptian army would abandon Mubarak? The structural characteristics and dispositions of armies often become apparent only after they begin to fight for their survival.

Now, does all this mean that all our fine scholarly studies of revolutions and popular uprisings can shed no light at all on the future? Clearly not. Social scientists know a good deal about what kinds of social organization, economic conditions, political contexts, and even cultural and emotional milieus have been especially conducive to movements and uprisings. A number of recent books have successfully distilled this knowledge (e.g., Sanderson 2010, Snow & Soule 2010, Johnston 2011), which is far too extensive to summarize here. We can say with some confidence, for example, that a future revolution is much more likely to occur in a middle-income country, ruled by a longstanding, corrupt dictator, which suddenly experiences a severe economic crisis than in a rich democracy with a booming economy. What we cannot predict is when or even whether that middle-income country, ruled by a longstanding, corrupt dictator, which suddenly experiences a severe economic crisis will actually experience a popular uprising. Social scientists can talk fairly intelligently about the possible historical trajectories of particular societies. But the highly contingent nature of human behavior makes it impossible to predict accurately the actual or even likely trajectory of societies.