After the Massacre: Mobilization in the Wake of Harsh Repression

Derry, Dublin and St. Petersburg mark Bloody Sundays; Berlin, Vienna and Tehran Bloody Fridays, and throughout the world massacres of even larger proportion and severity occurred during the twentieth century. These events have been consigned to historians in archives and dissident entrepreneurs who seek to keep alive the memory of state brutality. We take up 31 of these singular, cruel events to ask a question seldom posed: What happens after the massacre? What do surviving, uninjured dissidents do in the days after harsh repression? While this is perhaps an intrinsically interesting subject, the paper’s objective is theoretical. Mobilization immediately following a brutal massacre should be the most challenging context for collective action. But if it does occur is collective action theory correct (Olson 1971; De Nardo 1985; Sandler 1992; Marwell and Oliver 1993; Lichbach 1987, 1995 and 1996)?

A massacre ruthlessly and consistently signals the regime’s willingness to repress public protest (Lichbach 1987). It leaves little doubt about intolerance of dissent. There are instances, of course, of unintentional massacres—that is, troop deployment was intended, but the order to fire was never given. Witness the Kent State University shootings in 1970. Even in such an “unintended” massacre, however, dissidents can never be really certain that the shooting was not purposeful. Massive protest occurred in the United States after the Kent State incident, but the United States, a democratic country, tolerated it. Would this be the case in the USSR, Iran, El Salvador, or even British-occupied Ireland? In dictatorial regimes claims of “unintended killings” are dubious.

The hallmarks of collective action theory are leadership, resources, low-risk of action, and the possibility of making a difference through protest. All of these properties have a low probability of supply in the wake of a massacre. While leaders might survive, they might be arrested or unable to communicate; after a brutal state shooting the likelihood of a reservoir of resources is low; risk rises and while martyrdom might make a difference in history, there is little incentive to go back into the street to face a machine gun nest or the muzzle of a tank cannon. Nonetheless, we know that some dictators take violent repression to extremes that occasionally lead to revolution. We also know that inconsistent repression encourages protest (Lichbach 1987) and that while some repression is effective, other repressions hurt the state. The reality is that protest after a massacre is uncharted research territory. We begin a systematic attempt to analyze this arena with a group of harsh repression events.

Repression Events

This investigation of the aftermath of a massacre is based on 31 massacres in the twentieth century. There is, alas, no benchmark list of massacres. Carlton (1994) lists many, but only in general terms and mostly in the context of war. The events for this study were drawn from two sources. First, from my own National Science Foundation grant text coding of protest and repression; second, from Lexis-Nexis and a traditional library search for massacres. For the most part these are well-known events, at least to those familiar with the history of specific regions. Massacres, while numerous, are too rare for true random sampling. The events represented here are the only appropriate ones I could find. The principal qualifying criteria for an accepted event are (1) non-violent unarmed dissidents in (2) a non-democratic country who are (3) shot, killed or injured in a discrete incident by state forces with at least three killed and a total set of casualties over thirty. Genocide is excluded (see Hayner 2001) because it is not discrete, as are most rural massacres since they leave few to lash back. Table 1 provides summary information for all of the repression events in the study.1

The events, occurring on almost every populated continent, range throughout the twentieth century. Events in Europe predominate, but South and East Asia as well as South Africa and Latin America are also represented. They constitute a most-different systems sample in space and time (Przeworski and Teune 1970). While many of these events took place during independence or freedom campaigns, all are discrete repression events. To preclude inconsistent repression as the cause of backlash (Lichbach 1987), there are many consistent harsh-repression dictatorships: Burma, South Africa, occupied Ireland/Ulster, post-World War I Germany, Lebanon, occupied India, Mexico, occupied Palestine, Poland under martial law, Turkey’s repression of Kurds and the former USSR. The events in these countries are simply harsher in scale. Backlash or complete inaction can be documented after each sample event, a criterion that precluded a number of major, serial massacres, e.g., the Armenian, Nanjing and Holocaust genocides.

Insert Table 1 here

But after a massacre, what then? The first problem we face is what really happens after a massacre. Does this kind of brutal event drive everyone from the streets into homes? Information about protest after massacres is available, but not readily summarized. Most books about a single harsh repression focus on the event itself, not its aftermath. Generally I consulted microfilmed newspapers (or Lexis-Nexis for post-1980 events) to ferret out details of what occurred on each post-massacre activity. These searches revealed much backlash against state repression. Table 2 shows first the number of protesters in the massacre, then the number of protesters for three post-event days. The total number of protesters killed and injured is included in each of the backlash events. These data show that a good deal of backlash took place after most of the massacres.

Insert Table 2 here

Figure 1 shows that the mean level of post-massacre mobilization accelerates and dwarfs the original-event mobilization. There is indeed a tremendous amount of backlash. With this foundation, the problem we now probe is how that backlash is mobilized after extremely harsh repression.

Insert Figure 1 here

Collective-Action Theory and Post-Massacre Backlash

How does a dissident leader persuade dissidents to act today and tomorrow after hundreds were murdered by the state yesterday? That is the focus of our problem: can existing mobilization theory survive empirical tests in the case of a cold-blooded massacre? On the one hand, a dissident leader can count on moral outrage from the massacre (DeNardo 1985, 208-9). On the other hand, risk of another massacre lowers the ability of outrage to act. In terms of collective-action theory, let us break the problem into several components: (1) information transmission about the massacre, (2) leadership, and then (3) mobilization values: incentives, risk, efficacy, and pursuit of the public good (Olson 1965; DeNardo 1985; Marwell and Oliver 1993; Lichbach 1995, 1996; and Oberschall 1980). With regard to affective orientation alone, recruitment after a massacre is easy—nearly everyone aware of it hates the state. The problem is rationality. With Lichbach’s (1995) rebel’s dilemma risk at its highest levels, and potential gains largely absent, why would any individual act? We grapple with each of these problems in turn.

Information Transmission after a Massacre

The most basic problem for a dissident entrepreneur who has survived a massacre is how to communicate the repression to other surviving dissidents. Susanne Lohmann (1994) modeled an information cascade that occurred in the German Democratic Republic (GDR) in late 1989; it was restricted to the Leipzig demonstrations. Her model works in its context in part because almost all East Germans watched West German television. They remained well informed about what was happening throughout the country as well as in Leipzig. Lohmann’s (1994) model is a threshold-based signaling representation. It assumes continuity in time. Essentially, people see a repression event, then they monitor backlash turnout. If turnout increases over time, the regime has a higher probability of collapse. The decision mechanism for dissidents is governed by each dissident’s ideal point and the private information flow to each dissident. Those who already dislike the regime are more ready backlash participants. This model is similar to other threshold models (e.g., Granovetter and Soong 1983). Lohmann (1994) is less concerned with how information is transmitted. How do citizens learn about the repression and the backlash turnout? As noted above, it was easy in the GDR.

Heinrich Böll (1981) notes that it was possible to speak against the Nazi regime with relatives and friends, even in the army, as long as one was discreet (see also Huch 1997). But what about a repression context such as Burma, in which the state precludes publicity, denies its repression, and shuts down institutions such as universities? There is, after all, no West Burma and no informative television coverage of events. A slightly different context appeared in Korea. After the Kwangju massacre, the state released brief press statements to newspapers recounting student riots. Deaths were blamed on these riots. The truth lay hidden for more than a year. Of course local citizens knew what had happened; hundreds of thousands of them protested directly after the massacre. Was this due to a localized information cascade? Let us rethink how people find out about massacres.

First, information transmission differs greatly; it depends on the location of the repression event, i.e., an urban versus a rural area. Almost all of the sample events in this paper are urban massacres. In isolated rural areas, massacres may not become known for months. For example, history documents many banana-worker massacres in Colombia (Henderson 2001), but these generated little reaction because they occurred in remote areas. Similarly, the El Salvador El Mozote massacre (Danner 1994) might not have been discovered for months if one woman had not hidden in a field and survived. Rural massacres are isolated and often leave no survivors to lash back.

Second, we can assume that public protesters, who are usually young, have relatives, neighbors, coworkers, fellow students and friends—an extensive network of personal relationships. Within this web there is a high probability that someone is aware of another individual’s participation in a demonstration. Bernard et al. (2001) have modeled this idea, basing it on the size of social networks. Exhaustive survey research indicates that the minimal size of a median U.S. social network equals 290 (with a standard deviation of 232). The median size is 437. How do these figures relate to our events? Dissident and labor communities should be well integrated, particularly in urban areas and around universities.

Whereas Bernard et al. (2001) calculate the proportion of the total population who know someone in an event, we are interested only in the average number of people who might know an activist in an event. Consequently, we simply calculate the assumed size of a personal network times the number of participants in the repressive event. Although Bernard et al. use the minimum 290 figure of personal networks, there are probably many of the same people in a network of protesters. We cannot know about all networks in a century and all of the countries, thus a more conservative multiplier is 100. Table 3 displays the probable number of people in each area who would know at least one protester. It is clear that this personal information transmission alone can account for significant backlash.

From 50,000 individuals in San Salvador and Vilnius to ten million in the East German 1953 event it is apparent that knowledge of the event can travel fast, at least locally, through a myriad of social networks. All of these numbers represent people who might be aware of and concerned about a citizen who participated in a protest event. Information about the massacre is a logical prerequisite for backlash; these numbers are certainly sufficient to explain how dissidents become aware of repression despite state denial and lack of media attention. Clearly, the larger the initial event, and the more dissidents who survive, the more people would know about the massacre. Knowledge of many of the sample events was openly available. One striking example of this occurs in diarist Harry Count Kessler’s (1961) notes concerning the 1918 Spartakus massacre, just two days after the event, while he was in Warsaw. For our assumptions, though, we should not consider communication free in all events. The model above opens a solution to the closed communication problem. The numbers in Table 3 are sufficient to establish a communication flow. Surviving dissidents are likely to provide especially lurid accounts of the state’s repression—accounts that virtually no state can deny to families, friends, and probably even some coworkers.

We should not assume that all people who learn of a massacre will be willing to protest. In this situation, we have Lichbach’s (1995) five-percent limit to mobilization. In other words, the most people who even join a protest are always lower in numbers than five percent of the community. This is particularly likely after finding out about a massacre. Most people are risk averse. Nonetheless, these people form a base of knowledge about the regime and its actions for dissident activists.

Insert Table 3 here

Leadership

State leaders know that dissident entrepreneurs are critical agents of mobilization. In a planned massacre, there is a high probability of an attempt to kill or at least incarcerate the leaders. Dissident leaders in the wake of a massacre are likely to possess coordination power (Hardin 1995); they also can persuade dissidents to act and can form adaptive, risk-averse tactics. What happens when leaders survive and alternatively what happens when leaders die? We have both instances in our sample.

When Father Gapon, leader of the 1905 march to the Tsar, disappeared after the St. Petersburg massacre, labor leaders took over in his stead. After the 1921 Armristar massacre union leaders ordered strikes and religious clerics hartals, a Hindu form of general strike. In tactical contrast, labor leaders in Guinea-Bissau remained in complete control, yet they decided against immediate backlash. Instead, they organized and waged a long-term guerrilla war versus the occupying Portuguese army.

Coordination power after harsh repression is often significant in generating action and maintaining a low-level of repression. Only recognized leaders can generate this power without resorting to force. In Vilnius, the rebels maintained control of the radio station. They were able to play Lithuanian patriotic music and then give detailed instructions: stay in the occupied buildings; come to Kannas to mount a night vigil; erect a barricade at the library (BBC 1991). Perhaps the most dramatic instance of backlash coordination power was after the Derry Bloody Sunday in January 1972. A hastily called pub meeting of old (Official) and new (Provisional) IRA leaders led to an announcement of a general strike. Somehow everyone in the Catholic community got the message. Grocery stores opened for one hour and then closed. Then banks opened for one hour, then they closed. British authorities were astonished that all the Catholic citizens seemed to know precisely when grocery stores and banks would be open (McCann 1992).

Few dissidents seek to demonstrate after the state has killed their comrades in cold blood. To overcome this resistance, leaders generate adaptive tactics. Lichbach (1995) calls this sort of adaptation the choice of a more productive tactic. After a massacre leaders’ credibility of choosing tactics might be low. However if the more productive one obviously is safer, then more dissidents might be willing to venture into action. Table 4 displays the types of tactics used at the massacre and subsequently at backlash. While almost all of the event tactics were in public protest, the backlash tactics mainly consisted of strikes, funeral rallies, and general strikes. The backlash tactics stress security in numbers and in the safety of private homes. Minimally more productive, these adaptive tactics are successful because they reduce risk.

Insert Table 4 here

The adaptive tactics listed in Table 4 successfully reduce risks, most showing zero deaths or injuries in backlash protest. Nonetheless, Caracas, Escalante, Managua, Mendiola Bridge, San Salvador, Soweto, Tehran, Tiananmen, Turkey, Uitenhage, White Bridge/Red Bridge and the Wujek mine massacres all suffered at least one death in backlash. Injuries were especially prevalent after the Soweto, Tiananmen, White Bridge/Red Bridge events and the Wujek mine massacre. Larger scale casualties resulted from the demonstration tactic employed when dissident leaders were absent or misjudged the state’s level of contrition for its repression. These are the events that had the least leadership. While the PKK survived in Turkey, as a guerrilla movement it had little influence in the cities. Soweto students continued to challenge police, largely without direction. Solidarity leaders had been arrested three days before the Wujek massacre, leaving a vacuum throughout the mining community. More wary tactics reduced death and injury greatly; most dissident leaders selected strikes and general strikes. In the 1918 Berlin Spartakus movement demonstrations workers armed with rifles and handguns encircled and protected demonstrators from Free Corps machine guns.

The Risks of Backlash

Before we proceed to backlash mobilization, we pause to reconsider the level of risk in rising against the state (or occupier) after a massacre. Is backlash after a massacre without substantial risk? I argue that it is perilous. Risks are both immediate and residual. The context in which each of our repression events occurred is either anarchic or totalitarian. As Hayek (1979) points out, the act of central economic planning leads to totalitarian control. This means that everyone’s actions can be (and are if one is a dissident) monitored. While DeNardo (1985) rightly notes the importance of power in numbers, after a massacre any number who act will be known by the state. A general strike, for example, is most effective if it is truly widespread. Regardless of its breadth, the state had a high probability to maintain daily lists of all of those who fail to report for work. In martial-law Poland, the GDR, and the former Soviet Union, party cells in factories and mines reported names of absent employees. After the January strikes in 1905 Russia, strikers faced mass dismissals (Suhr 1989, 197). Similar known-identity problems arose in South Africa, South and Central America, as well as in Western Europe and certainly in UK-occupied India and Ireland. At mildest, repression was a warning of demotion, transfer or firing. At worst, residual risk translated into a visit from a death squad. Many in the Berlin Spartakus movement were murdered, including leaders Liebknecht and Luxembourg one month after their backlash. In El Salvador, where even librarians were targeted by death squads, known anti-state activists ranked far higher on the ordinal list for torture and murder (Danner 1994).

The biggest residual risk from backlash, then, is that the identities of protesters are known, making activists vulnerable to state punishment. The punishment risk increases greatly if few act. Leaders assure protesters that their tactics are safe and productive (Lichbach 1995), but no activist can rely on these promises in a non-democratic context. If few turn out, the risks of repression or economic punishment rise even with “safe” tactics.

Buchanan and Tullock (1965, 37) note that rationality of individuals in a political context depends on iterative events. The day after a massacre provides no relevant previous event experience. Therefore, any post-massacre protester faces uncertainty at best, both on the day of action and later on the job or in the university. Table 5 lists the immediate and residual risks in backlash after each massacre. In fact, in nine of our 30 backlash events protesters were killed or injured, yielding a 30 percent empirical probability of immediate death or injury. The aggregate death toll of backlash in Table 2 is 351, with 4,491 injuries.

Insert Table 5 here

Mobilization

At this point we have documented three vital prerequisites for mobilization: (1) established sufficient information transmission of the harsh repression in a local area,; (2) sustained leadership or new leadership in almost all events; (3) and that leaders are able to fashion and communicate tactics that lessen risks. Mobilization in the wake of harsh repression requires all of these, particularly when one considers the risks noted above. Lichbach (1987) proved that repression might either deter mobilization or escalate it. If any kind of repression should deter mobilization, it would seem to be a massacre. Yet we have seen large-scale backlash after almost all 31 harsh repression events. None of this, however, helps to explain why any rational individual would dare to act against the state after a massacre. And therein lies the principal puzzle. To attempt to solve it, let us begin with the levels of mobilization that occurred in the each day after the repression event.

The lowest aggregated level of backlash mobilization (see Figure 1) is post-day 1. With the background information already established, it is not surprising that the first day after harsh repression generates larger mobilization than the event itself. Not all of the backlash protest on post-day 1 is carefully planned, however. Mobilization sometimes erupts before a dissident leader can communicate safer tactics to followers. In Soweto, students were so outraged at the violence directed at them that they continued to riot—as a result many were injured and killed. The first day after harsh repression, then, is the most dangerous by far for injuries and dissident deaths. But the third day after the massacre has the highest level of mobilization. Why then? The principal reason is that burials of the victims typically take place on the third day after a massacre, and funerals usually engender large participation and low repression, even if state forces march lockstep with the mourners. Post-day 3 action does not signal a continuous rise in mobilization. In fact, mobilization sometimes drops precipitously, even in ongoing conflicts such as those in South Africa, Ireland, Northern Ireland and Palestine. Post-day 2 generates middle-level mobilization. To some extent, this is again a function of the fact that funerals happen on post-day 3. Nonetheless, accelerating levels of mobilization underscore the sustained success of mobilization after massacres. It is not easy to keep people from their normal daily lives for three continuous days at any time, much less, after brutal repression. The accelerating pattern of our post-repression backlash meets Trotsky’s (1959) rigorous test: a revolutionary struggle could only succeed in the case that it ascends step by step every day. We know, of course, that revolution was not the dissidents’ first priority after our 31 repression events. Our deeper puzzle, then, remains: how can dissidents be convinced to act after a massacre?

While cold-blooded state repression angers most citizens, a tipping model seems less appropriate for analysis in the sense that such models rarely include risks. The risks of post-massacre mobilization are real and evident. A more appropriate model is Bayesian updating. We can safely assume that many people are outraged by harsh repression, yet simultaneously most are even more risk-adverse. Bayesian updating allows this mixture of outrage and safety attitudes in a single model. The model of choice is Skyrm’s (1990):

where {A_i} is a partition of alternative acts; e is recalculation of expected utilities, and p₂(A) is the updated probability of acting.

The Bayesian updating prior probability is actual repression probability. If backlash dissidents are killed and injured on post-days one and/or two, total mobilization should be dampened. Our first test is graphical: it separates the level of post-day mobilization from repression and non-repression. Figure 2 indicates a distinctly lowered level of mobilization from repression. Figure 3 shows higher levels of post-day three mobilization with no repression.

Insert Figure 2 here

The Bayesian updating test uses post-day 1 and post-day 2 real probability of death and injury for prior probabilities. If there were no post-event repression, the default posterior probability of mobilization on post-day three is set at 0.333, in other words the normal one-day probability of a three-day mobilization. Bayesian updating posterior probability results were set against the actual probabilities of post-day three mobilization. A paired-sample t-test between these two data series should be insignificant if Bayesian updating truly damped mobilization after repression. It is: the means of the two samples are virtually identical (0.3343 and 0.345 respectively) with a low t-statistic of –0.229, indicating that actual mobilization was slightly higher for most (repression free) events on post-day three. Dissidents do consider risks when deciding to act. Repression dampens mobilization significantly compared to truly “safe” tactics and no repression.2

Insert Figure 3 here

Of all the collective action theorists, Lichbach (1995, 1996) provides the best context needed for protest after a massacre. We get little help from Olson (1965), who of course never considered mobilizing dissidents immediately after a massacre. Olson’s (1965) claim that large mobilization requires selective incentives presents problems for our events. Resources were not in plentiful supply after our 31 events of harsh repression—even in documented cases of a million protesters on post-day 3.

DeNardo (1985) considers the effect of repression and posits that only the public good is necessary for mobilization. In the case of heavy repression, however, DeNardo (1985) admits that his mobilization strategy is no longer straightforward; he suggests that it is necessary to narrow the public good and to mobilize few people clandestinely. This, effectively, was the Bolshevik solution—have a small group ready to take over when the regime totters. In none of our cases, however, was any regime close to falling, save of course the Shah’s regime in Iran (see Rasler 1996). In our cases there is large, public mobilization on three successive days after the state shot dissidents dead in cold blood. Indeed, the public good may narrow to no killing or shooting as may the size of the core leadership, but somehow thousands, tens of thousands, and in several cases hundreds of thousands acted to signal their disgust of the state’s action. We need another approach to mobilization, one with greater flexibility, that will encompass this extreme type of context.

To Olson’s (1965) solutions of the collective action problem (selective benefits and the ability to make a difference), Lichbach (1995, 1996) added scores more—with a catch. The catch is Lichbach’s proof that no single group of solutions works alone. Dissident entrepreneurs must combine solution groups in order to persuade the discontented to act. Lichbach (1995, 1996) did not model our situation, attempting mobilizing immediately the day after 400 dissidents are shot dead by the state police or military. But because the approach he designed is both flexible and general, it applies to our 31 events.

To augment the Bayesian updating test, we use Lichbach’s (1995, 1996) solutions as an analytic tool. In doing so, we preclude the assumption of irrationality, the uniqueness of massacres, or any peculiar opportunities of collective action. The task here is to discover whether a second rational approach to mobilization works fully in this extremely narrow and thorny environment. We have already established that dissident leadership either remains or that substitutes appear quickly after a massacre. In addition, we have seen that adaptive tactics were the modal choices of dissident leaders. In mobilization the presence of a leader is a signal event. Sometimes a leader can use coordination power and continually reform tactics to mobilize more followers. In terms of Lichbach’s (1995, 1996) theory, the market and community solutions to mobilization were the most popular approaches in our 31 events; the hierarchy solution was only slightly less common, principally because of the need of a pre-existing organization. The market solution enables an articulate leader, born of a repression event, to transform a public good into a significant public bad. A massacre endangers not only dissidents’, but also most citizens’ lives—a message more potent than a simple plea for more welfare benefits. This sort of public bad is advantageous in preventing a loss, something (Quattrone and Tversky 1988) found to be more important to people than attempting to receive gains. Loss of safety implies Hobbes’s dilemma—an historically important public bad (see Ferrara, 2003).

Salient costs near the time of the massacre are death, injury, or prison. But within the market solution dissident leaders lower costs of protest actions. Opportunity costs persist, but pale against more important life costs. Most leaders chose strikes or general strikes to lower risks (Table 4). Dissident leaders, however, are fallible, and some badly misjudged risks, notably in Burma and after the Tiananmen massacre. Protesters stage post-massacre backlash to signal opposition to the regime, but they use an action with a high probability of safety. While repression may sometimes generate a tipping point over to revolution, blazing guns create a rush to safety or a reluctance to leave home. Strikes and general strikes minimize costs because people stay home, where there is safety because of the enormous spatial diversity of residences. Doing nothing but staying home may be construed as inaction, but if few venture out, there is a high probability of arrest at home. If large numbers strike, there is little any regime can do to respond. In a sense, safety at home in spatial diversity is related to the safety in numbers of a large (at least 10,000) demonstration or rally (DeNardo 1985).

The community solution is the other modal choice dissident entrepreneurs made. Its two principal components (Lichbach 1995) are common knowledge and common values, both of which are critical to mobilization after harsh repression. The first element, common knowledge, leads to communication of the massacre to people who might not know about it, thus increasing the base of the dissident community. We established above merely the probable numbers of people who knew one acting dissident in a massacre. If those people tell others, then mutual ignorance fades rapidly to mutual knowledge. Since these were urban events, it was not necessary that communication be national in scope, but often it is. The West Beirut massacre and its subsequent backlash illustrate this. Few remained alive after the massacre to protest. The backlash occurred in the West Bank and Gaza Strip, where Palestinian Arabs identified with those women and children who were murdered.

The community solution’s second component, common values, leads to a resolution of the selective-incentives problem. Lichbach (1995, 1996) shows that common values can substitute for pecuniary incentives. Protest after a massacre might be articulated by a dissident leader as “costs are benefits” (Lichbach 1995, 123). If the opportunity of working is a benefit, a strike might lower it. However, if a leader points out that everyone’s rights are at issue when the state shoots its citizens, people may accept a brief strike as a benefit that supercedes the cost of lost work.

All of this succeeds if and only if many people act, and this is where pre-existing organizations play a critical role. If labor unions, social movements, and civil rights organizations can show that thousands of people will act, then still more action is likely. This employs two more solution groups that Lichbach (1995, 1996) developed. Hierarchy is a particularly important solution group in tense times such as after a massacre. Pre-existing dissident organizations, part of the Contract solution, gain credibility when the regime kills citizens. Thrust into a dangerous, possibly even Hobbesian-dilemma situation, there is a higher probability that members of existing organizations will unite for action under a persuasive leader. Student and labor organizations in Barcelona, Berlin, Essen, Guinea-Bissau, Kwangju, Mexico City, San Salvador, Soweto, St. Petersburg, Vienna and the Wujek mine were important elements in the design of backlash action. Religious organizations played an equally vital role in Amristar and Tehran, while civil rights and nationalist movements as well as political parties were central to the backlash in the Barcelona, Ciskei, Derry, Dublin, Escalante, Managua, Sharpeville, Tblisi, Temple Mount, Turkey, Uitenhage, Vilnius and White/Red Bridge post-massacre mobilization.

Hierarchy and Contract are most evident in the Guinea-Bissau massacre—the only one with no immediate backlash. Instead of ordering mass protest, the labor organization deliberately decided not to mobilize right away. Rather it began to forge a guerrilla organization against the Portuguese occupying army. In Northern Ireland, IRA hierarchy was the coordinating power in Derry in 1972, as noted above. Interestingly, while one might think that citizens would resent an organization that mobilized its members for a possible massacre. After all, many people could be injured or killed. Quite the opposite is true. In every case, the organization maintained full support of its surviving members, who blamed only the state for repression.

Let us pause and sum up the knowledge about our repression events we now have developed. First, we had more than minimally necessary information transmission. Sufficient numbers of members of every community probably learned about the massacre quickly enough to be able to act. Second, leadership either survived or was immediately and effectively replaced. Sikh leadership gave way unions and then to Mahatma Gandhi and the Derry civil rights leadership to both flavors of the IRA. Both of these replacements gained the movement more national and international recognition. St. Petersburg labor and reform leaders stepped in for Father Gapon and then wrested from a reluctant Tsar the concession of an elected legislature. Third, we found ample use of general mobilization techniques. The Market and Community solutions predominated (Lichbach 1995 and 1996), and supplemented by the Hierarchy and Contract, formed the key methods of mobilization and backlash against the state. It underscores the fact that mobilization after a massacre is so much more challenging that it requires all the solution groups from Lichbach (1995), not the more probable combination of two. Mobilization after a massacre requires extraordinary efforts, but these occur within the bounds of rational choice theory. We have discovered that a single dissident might have good incentives to act after a massacre. Although this notion seems counter-intuitive, a great deal of empirical evidence supports it.

Discussion

Collective action theory operates at the micro level, while this paper uses aggregated data to make theoretical inferences at the macro level. Evidence from a diverse sample of harsh repression events indicates that all of the micro-level necessary conditions for mobilization existed even after a brutal massacre. In addition, the Bayesian updating test shows that dissidents consider repression important in order to backlash after a massacre. While this might not be a definitive test, it is general at least across 31 different repressions in space and time. As noted above, I think findings would differ fundamentally in rural massacres, principally because of the information transmission problem. But in urban areas, people can be activated by dissident leaders to protest safely even after a state massacre.

I have long contended that protest is event driven. It would be difficult to argue the inverse in the cases of mobilization presented in this paper. Post-event mobilization accelerated day-by-day at a level that dwarfed the original repression event. Why would any of these people choose to act if the harsh repression had not occurred? That it occurred accounts for the fact that all four of Lichbach’s (1995, 1996) solution groups had to be used for mobilization.

There is another sense in which events like these provide signal benefits to dissident entrepreneurs. Anniversaries of massacres are great mobilizing devices within the community solution. For decades following the events in Amristar, Derry, and Sharpeville, protest leaders such as Mahatma Gandhi, Gerry Adams, and the collective leadership of the African National Congress and Pan African Congress successfully used the anniversaries of the events to mobilize masses of dissidents.

After the massacre, backlash occurs. It is driven, in terms of Lichbach’s (1995, 1996) solutions, by dissidents enlisting leaders, easy urban information transmission reducing ignorance, leaders shifting public goods to ominous public bads, reducing the risks of action, and invoking pre-existing organizations to order adaptive and “safe” tactics. In other words, there is nothing special about the time after massacres that would require a change in the collective action theory. It remains valid, even in this discomfited context.

References

Baring, Arnulf. 1983. Der 17. Juni 1953. Stuttgart, Germany: Deutsche Verlags-Anstalt.

Bell, J. Bowyer. 1993. The Irish Troubles: A Generation of Violence, 1967-1992. New York: St. Martin’s Press.

Bernard, H. Russell, Peter D. Killworth, Eugene C. Johnsen, Gene A. Shelly, and Christopher McCarty. 2001. “Estimating the Ripple Effect of a Disaster.” Connections 24(2): 18-22.

Böll, Heinrich. 1981. Eine deutsche Erinnerung. Munich, Germany: Deutscher Taschenbuch Verlag.

Botz, Gerhard. 1983. Gewalt in der Politik: Attentate, Zuzammstöße, Putschveruche, Unruhen in Österreich, 1918-1938. Munich, Germany: Wilhelm Fink Verlag.

Botz, Gerhard. 1987. Krisenzonen einer Demokratie: Gewalt, Streik und Konfliktunterdrückung in Österreich seit 1918. Frankfurt, Germany: Campus Verlag.

Brook-Shepard, Gordon. 1997. The Austrians: A Thousand-Year Odyssey. New York: Carroll and Graf.

Buchanan, James M. and Gordon Tullock. 1965. The Calculus of Consent. Ann Arbor: University of Michigan Press.

Carlton, Eric. 1994. Massacres: An Historical Perspective. Aldershot, UK: Scolar Press.

Clark, Donald N., ed. 1988. The Kwangju Uprising: Shadows over the Regime in South Korea. Boulder, CO: Westview Press.

Cockcroft, James D. 1996. Latin America: History, Politics, and U.S. Policy. Chicago: Nelson-Hall Publishers.

Commission for the Defense of Human Rights in Central America. 1990. Condoning the Killing: Ten Years of Massacres in El Salvador. Washington, D.C.: Ecumenical Program on Central America and the Caribbean.

Coogan, Tim Pat. 1992. The Man Who Made Ireland: The Life and Death of Michael Collins. Niwot, CO: Roberts Rinehart.

Cornebise, Alfred E. 1977. The Weimar in Crisis: Cuno’s Germany and the Ruhr Occupation. Washington, D.C.: University Press of America.

Daneshvar, Parviz. 1996. Revolution in Iran. New York: St. Martin’s Press.

Danner, Mark. 1994. The Massacre at El Mozote. New York: Vintage Books.

Davis, Leonard. 1989. Revolutionary Struggle in the Philippines. New York: St. Martin’s Press.

De Nardo, James. 1985. Power in Numbers: The Political Strategy of Protest and Rebellion. Princeton, NJ: Princeton University Press.

De Rosa, Peter. 1990. Rebels: The Irish Rising of 1916. New York: Fawcett Columbine.

Dreetz, Dieter, Klaus Gessner and Heinz Sperling. 1988. Bewaffnete Kämpfe in Deutschland, 1918-1923. Berlin: Militärverlag der Deutschen Demokratischen Republik.

Ettinger, Elżbieta. 1986. Rosa Luxemburg: A Life. Boston: Beacon Press.

Evans, D. Gareth. 2000. A History of Wales, 1906-2000. Cardiff: University of Wales Press.

Fein, Helen. 1977. Imperial Crime and Punishment: The Massacre at Jallianwala bagh and British Judgment, 1919-1920. Honolulu: University Press of Hawaii.

Ference, Gregory C., ed. 1994. Chronology of 20^th-Century Eastern European History. Detroit: Gale Research.

Ferrara, Federico. 2003. “Why Regimes Create Disorder: Hobbes’s Dilemma in a Rangoon Summer.” Journal of Conflict Resolution 47:3: 302-325.

Forrest, Joshua B. 1992. Guinea-Bissau: Power, Conflict, and Renewal in a West African Nation. Boulder, CO: Westview Press.

Furneaux, Rupert. 1963. Massacre at Amritsar. London, UK: George Allen and Unwin, Ltd.

Gilmour, David. 1983. Lebanon: The Fractured Country. Oxford, UK: Martin Robertson.

Granovetter, Mark and Roland Soong. 1983. “Threshold Models of Diffusion and Collective Behavior.” Journal of Mathematical Sociology 9: 165-179.

Grant, Neil. 1993. Chronicle of 20^th Century Conflict. New York: Smithmark.

Hardin, Russell. 1995. One for All: The Logic of Group Conflict. Princeton, NJ: Princeton University Press.

Hayek, Friedrich A. 1979. The Road to Serfdom. Chicago: University of Chicago Press.

Hayner, Priscilla. 2001. Unspeakable Truths: Confronting State Terror and Atrocity. London, UK: Routledge.

Henderson, James D. 2001. Modernization in Colombia: The Laureano Gómez Years, 1889-1965. Gainesville: University Press of Florida.

Huch, Ricarda. 1997. In Einem Gedenkbuch zu Sammelm: Bilder Deutscher Widerstandskämpfer. Leipzig, Germany: Leipziger Universitätsverlag.

Kessler, Harry Graf. 1961. Tagebücher, 1918 bis 1937. Frankfurt am Main, Germany: Suhrkamp Verlag.

Kessler, Richard J. 1989. Rebellion and Repression in the Philippines. New Haven, CT: Yale University Press.

Jackson, Alvin. 1999. Ireland 1798-1998: Politics and War. Malden, MA: Blackwell Publishers.

Laschitza, Annelies. 1982. Karl Liebknecht: eine Biographie in Dokumenten. Frankfurt am Main: Verlag Marxistische Blätter.

Lichbach, Mark I. 1987. “Deterrence or Escalation? The Puzzle of Aggregate Studies of Repression and Dissent.” Journal of Conflict Resolution 31 (June): 266-297.

Lichbach, Mark I. 1995. The Rebel’s Dilemma. Ann Arbor: University of Michigan Press.

Lichbach, Mark I. 1996. The Cooperator’s Dilemma. Ann Arbor: University of Michigan Press.

Lieven, Anatol. 1994. The Baltic Revolution: Estonia, Latvia, Lithuania and the Path to Independence. New Haven, CT: Yale University Press.

Lintner, Bertil. 1990. Outrage: Burma’s Struggle for Democracy. London, UK: White Lotus.

Lohmann, Susanne. 1994. “The Dynamics of Informational Cascades: The Monday Demonstrations in Leipzig, East Germany, 1989-91.” World Politics 47: (October): 42-101.

Lopes, Carlos. 1987. Guinea-Bissau: From Liberation Struggle to Independent Statehood. Boulder, CO: Westview Press.

Mandela, Nelson. 1994. Long Walk to Freedom. Boston: Little, Brown.

McCann, Eamonn. 1992. Bloody Sunday in Derry: What Really Happened. Dingle, County Kerry, Ireland: Brandon Book Publishers.

Marwell, Gerald and Pamela Oliver. 1993. The Critical Mass in Collective Action: A Micro-Social Theory. New York: Cambridge University Press.

Meredith, Martin. 1998. Nelson Mandela: A Biography. New York: St. Martin’s Press.

Ming Pao News Staff. 1989. June Four: A Chronicle of the Chinese Democratic Uprising. Fayetteville: University of Arkansas Press.

Mullan, Don. 1997. Bloody Sunday: Massacre in Northern Ireland. Niwot, CO: Roberts Rinehart Publishers.

Nan Lin. 1992. The Struggle for Tiananmen: Anatomy of the 1989 Mass Movement. Westport, CT: Praeger.

Oberschall, Anthony. 1980. “Loosely Structured Collective Conflict: A Theory and Application.” In Research in Social Movements, Conflicts and Change, ed. Louis Kriesberg. Greenwich, CT: JAI Press.

Olson, Mancur. 1971. The Logic of Collective Action: Public Goods and the Theory of Groups. Cambridge, MA: Harvard University Press.

Payne, Robert. 1969. The Life and Death of Mahatma Gandhi. New York: Konecky and Konecky.

Petran, Tabitha. 1987. The Struggle Over Lebanon. New York: Monthly Review Press.

Przeworski, Adam and Henry Teune. 1970. The Logic of Comparative Social Inquiry. New York: John Wiley & Son.

Quattrone, George A. and Amos Tversky. 1988. “Contrasting Rational and Psychological Analyses of Political Choices.” American Political Science Review 82 (September): 719-736.

Rasler, Karen. 1996. “Concessions, Repression and Political Protest in the Iranian Revolution.” American Sociological Review 61: 132-152.

Ritter, Gerhard A. and Susanne Miller. 1968. Die deutsche Revolution, 1918-1919. Frankfurt am Main: Fischer Bücherei.

Sablinsky, Walter. 1976. The Road to Bloody Sunday: Father Gapon and the St. Petersburg Massacre of 1905. Princeton, NJ: Princeton University Press.

Sandler, Todd. 1992. Collective Action: Theory and Applications. Ann Arbor: University of Michigan Press.

Salisbury, Harrison E. 1989. Tiananmen Diary Thirteen Days in June. Boston: Little, Brown.

Skyrms, Brian. 1990. The Dynamics of Rational Deliberation. Cambridge, MA: Harvard University Press.

Spittmann, Ilse and Karl Wilhelm Fricke. 1982. 17. Juni 1953: Arbeiteraufstand in der DDR. Cologne, Germany: Edition Deutschland Archiv.

Stempel, John D. 1981. Inside the Iranian Revolution. Bloomington: Indian University Press.

Story, Dale. 1986. The Mexican Ruling Party: Stability and Authority. New York: Praeger.

Suhr, Gerald D. 1989. 1905 in St. Petersburg: Labor, Society and Revolution. Stanford, CA: Stanford University Press.

Thompson, Ginger. 2001. “Flashback to Deadly Clash of ’68 Shakes Mexico.” New York Times, December 13.

Trotsky, Leon. 1959. The Russian Revolution: The Overthrow of Tzarism and the Triumph of the Soviets. New York: Doubleday.

Zabih, Sepehr. 1988. The Iranian Military in Revolution and War. New York: Routledge.

“Zum Ruhreinbruch: Tatsachen und Dokumente.” 1976. In Defeat, Revolution and Occupation of the Ruhr: Birthpangs of the Weimar Republic, Part 3. Nendeln, Liechtenstein: Kraus Reprint.

Appendix: Profiles of the Repression Events

Amristar: Citizens of the Punjab chafed under British military rule. General Dyer controlled Punjab and ordered curfews and other limits on the Sikh citizens. 15,000 Sikhs gathered in a park to protest on April 13, 1919. They held hands and stood peacefully in the park. General Dyer and his troops came to the park and saw the 15,000 Sikhs standing and holding hands. General Dyer ordered his troops to shoot. They killed 530 and wounded thousands. General Dyer later said, “It was a horrible duty I had to perform. I think it was the merciful thing. I thought I should shoot well and shoot strong, so that I or anybody else should not have to shoot again” (Payne 1969, 340). Backlash was organized by Punjab union organizations in the form of strikes and sabotage. Religious leaders urged all citizens to perform hartals, a religious ritual of fasting and desisting from daily business; this became an Indian version of a general strike against the British (Fein 1977). And of course Gandhi used the massacre to build momentum for his own nonviolent, hartal-based campaign.

Barcelona: Workers sought a 9-hour day through a strike. Workers protesting on February 18, 1902 were shot by troops: 100 dead, 300 injured and 500 arrested. A general strike ensued throughout Catalonia in response to the repression. The Spanish Cortes suspended the constitution during the conflict.

Berlin: Karl Liebknecht and Rosa Luxemburg led the Spartakus (left-wing communists) rising in Berlin from November 1918 through January 1919, when the two leaders were murdered. On December 6, 1918 a Spartakus demonstration came upon a former-military Freikorps (Free Corps) machine-gun nest. The Freikorps shot and killed 18 and injured 30 as the Spartakus demonstrators ran away. The next two days, Saturday and Sunday, saw huge backlash demonstrations escorted by armed workers for protection (Dreetz et al. 1988).

Caracas: In February 1989 the Venezuelan government imposed austerity and raised bus fares and food prices. Hundreds of thousands of citizens protested these decisions across the country. Our focus is in Caracas, where the principal protests occurred. Troops opened fire on looters and demonstrators, killing 300 in the capital alone. Backlash took the form of rock-throwing demonstrations as well as sniping.

Ciskei: 20,000 African National Congress members demonstrated for democracy in Bisho, Ciskei, a “homeland” in South Africa. Police opened fire, killing 28 and wounding 288. The next day 100,000 ANC members marched to Ciskei, Bishop Tutu led 2,000 in prayer, 12,000 attended a rally led by Nelson Mandela and 1,250 protested in other cities. The following day 2,000 protested in Johannesburg.

Derry: The Northern Ireland Civil Rights Association organized a civil rights demonstration on Sunday, January 30, 1972 in Derry, Ulster. Leaders secured agreement from the Provisional IRA to stay away in order to insure a peaceful protest (Mullan 1997). As hundreds walked toward a phalanx of British soldiers, the soldiers suddenly opened fire, killing 13 and wounding 14. Horrified, the demonstrators broke up and tried to aid the injured and remove the dead. Several of the dissident leaders moved into a nearby vacant pub. An official IRA leader recommended a general strike for the following days. A provisional IRA official seconded the motion and it passed unanimously. Representative of the Catholic press were present at the meeting. On Monday and Tuesday grocery stores and banks would be open one hour each. The coordination power was virtually complete (McCann 1992).

Dublin 1914: A thousand people attended the funeral of an Irish Volunteer (precursor to the IRA) fighter in the Bachelor’s Walk section of Dublin on the eve of World War I. Occupying British troops flanked the mourners. Suddenly, the troops opened fire, mistaking an officer’s command for “Fire!”. In the aftermath, four were dead and 29 injured. Young Irishmen surrounded the British military barracks in Dublin that night, walking continuously around the building. On the following day Irish police resigned from the force, refusing the order to disarm Volunteer fighters. The greatest post-event backlash came on the third day when there were funerals of the four killed.

Dublin 1920: The Irish Volunteers shot dead many of the British military secret agents during the night of November 20-21, 1920. British auxiliary troops, the Black & Tans, were so outraged by the killings that they drove an armored personnel carrier armed with a machine gun into Croke Park stadium where a soccer game between Dublin and Tipperary was going on. They shot and killed one soccer player and thirteen fans as well as injuring hundreds as they sprayed the stands with machine-gun bullets. The next 3 days saw many constables killed, but little public backlash since the UK army was deployed looking for the Irish Volunteers who killed their secret agents.

Escalante: Nation-wide protests on the 13^th anniversary of the imposition of martial law in the Philippines. Troops shot and killed 21 and injured 30. A transport strike paralyzed Escalante so no one could travel. But the strikers were lashing back against the massacre. In a nearby town of Bacolod 30 thousand protested on the third day.

Essen: During the French occupation of the Ruhr area of Germany after World War I, the Germans protested passively, mainly striking. On March 31, 1923, French military officers ordered 12 cars to be provided by the Krupp corporation in Essen. Krupp management sounded the steam siren; workers heard it, stopped working and all 50,000 rallied outside the factory. After a two hour standoff, the French fired into a group of Krupp workers, killing 13 immediately and causing others to flee (Cornebise 1977). Krupp executives were arrested; the French also shut down three newspapers the next day, Easter Sunday. Passive protest, mostly striking, continued and peaked on April 10, the day the 13 workers were buried.

German Democratic Republic workers’ rising: On 17 June 1953 East German workers rose against increased production norms. The rising was general, not just in East Berlin. Police were overwhelmed by as many as 300,000 workers demanding lower work standards and a free trade union. The USSR Red Army intervened, killing 21 workers and injuring 167. Nonetheless, the rising continued through backlash, especially in Magdeburg, for three following days.

Guinea-Bissau: The African Party for Independence of Guinea and Cape Verde ordered a dock strike in Pidiguiti on August 3, 1959. Several hundred workers assembled, refused to work, and demonstrated against Portuguese occupation. The Portuguese police started shooting, killing 50 and injuring over 100. After this massacre, the Guinea-Bissau labor organization decided not to use backlash, but instead started a guerrilla war against Portugal some months later (Forrest 1992 and Lopes 1987).

Kwangju: Students protested martial law in Korea. On May 17, the army burst into student dorms and beat students studying for exams. At 10:00 on the 18^th, a shout of “End martial law!” sparked the demonstration. At lunchtime, business people mingled and joined with the student protesters. Riot police, outnumbered, were pushed back. A sit-in demonstration began. Suddenly armored personnel carriers brought paratroopers with rifles and attached bayonets. First the paratroopers fired tear gas into the demonstration. Then they swung clubs at demonstrators, aiming for the head. They used their bayonets to tear off women’s clothes, then stomped and pounded on the most sensitive parts of the women’s bodies. They chased dissidents who tried to flee and then beat them. The soldiers had removed all identification of their regiment or name. They had uniforms, but no identity. This was a state-adaptive change from a year before, in Pusan, when dissidents had been able to distinguish their repressors’ regiments. The uprising lasted until May 27.

Managua: The opposition Conservative party challenged Somoza, the President and dictator, to have a free election. The National Guard, Nicaragua’s military, fired into the Conservative demonstrators, killing 34 and injuring 100. The battle then shifted to a hotel where the dissidents had sought refuge. The dissidents remained there for 20 hours, then agreed to leave barring additional shooting. Upon their departure, the head of the Conservative party called a general strike.

Mendiola Bridge: Philippine farmers peacefully seeking land reform were shot on a bridge by military forces protecting a palace in Manila (Kessler 1989). President Aquino apologized for the massacre the following day.

Mexico City: 10,000 students rallied peacefully as part of a student strike at 18:00 on October 2, 1968. Hundreds of police and military surrounded the students and began shooting. An estimated 400 were killed, hundreds injured, and over 2000 detained without trial (Story 1986; Thompson 2001). Backlash comprised the continuation of the student strike, burning 3 buses the day after and moving out in brigades (small groups) on the second day. Protest leaders deliberately eschewed public demonstrations as “suicide” (New York Times 10/5/68). The strike continued the next day, a Sunday.

San Salvador: Extreme-leftists kidnapped foreign people and occupied two embassies. Approximately 500 students staged a supportive demonstration in the front of a cathedral. Troops arrived and opened fire, killing 23 and wounding 70 (Commission for the Defense of Human Rights in Central America 1990). The Red Cross was left to collect the bodies and transport the wounded. The next day there was tentative backlash by about 5,000 people. The second day there was a general strike and a funeral with 18,000 people in attendance. The third day saw more funerals, more occupations, and four women dissidents shot dead while putting up a sign.

Sharpeville: the Pan African Congress organized a protest against a new law requiring women to have a pass to walk into other villages or regions. The daylong passive protest occurred on March 21, 1960 all over South Africa; mostly women participated. In Sharpeville, police panicked and starting shooting at the protesters. In all 67 women were killed and 186 were wounded, most shot in the back, running away from the police. The massacre stunned South Africa. Large backlash demonstrations organized both by the Pan African Congress as well as the African National Congress followed the next few days.

Soweto: On June 16, 1976 15,000 school children gathered in Soweto to protest against Afrikaans, the language of their oppressors, used in their textbooks. The students confronted riot police who opened fire at them. Approximately one thousand children were killed and three times as many were wounded. The backlash against this massacre was large, even as riot police continued to shoot at protesters.

St. Petersburg: Father Gapon led thousands of workers seeking more food and higher wages to the Winter Palace on Sunday, January 22, 1905. The protest had been announced; even so, the Tsar’s forces shot at the marchers, killing 175 and wounding 625. Father Gapon was lost, but not killed, in the melee, and most marchers fled the scene. Backlash, mostly in the form of general strikes, was strong against this Bloody Sunday. It led eventually to Russia’s first elected parliament (Duma).

Tblisi: Georgians seeking independence rallied from Saturday through Tuesday on Rustaveli Square in Tblisi. After 21:00 on Tuesday tanks moved in and opened fire, in an attempt to clear the square. The toll: 19 killed and 36 injured. The USSR claimed that 75 soldiers and police were injured as well. Authorities imposed a 23:00-6:00 curfew. In response 15 to 20 activists then distributed leaflets calling for a general strike. Schools and universities closed in a show of solidarity against Soviet repression (Associated Press 4/13/89).

Tehran: In the latter phase of the Iranian revolution, the military imposed martial law and precluded protest on September 7, 1978. A demonstration of 20,000 already planned for September 8 in Tehran was mobilizing on Jaleh Square for the following day. Many demonstrators were unaware of martial law. Troops shot dead over 500 protesters and injured more than 4,000. Many strikes and demonstrations took place the following day, including one in Qum where one demonstrator was killed. There were 100 arsons as well. On the second day there was a large funeral for the demonstrators who were killed, and lesser backlash developed on the third day. All three days saw continuing strikes. According to Shia Islamic law, larger backlash formed on the seventh day after the massacre.

Temple Mount: 3,000 Palestinians massed on Temple Mount in Jerusalem on October 8, 1990. They challenged police with rocks. Police shot dead 20 Palestinians and wounded 140. Palestinian officials organized a week-long general strike in response. On October 9 there were rallies in refugee camps as well as funerals. Hundreds of Palestinian youth challenged Israeli police on October 10. 13 were injured in these clashes. Many groups sought to regain Temple Mount on October 11, but only women and clerics were allowed by the Israeli defense force.

Tiananmen: Under the banner of the Federation of Beijing Autonomous Unions, students led the Chinese democracy movement in 1989. The federation became more active and public as spring turned to summer, and students began to occupy Tiananmen Square, a central space in Beijing. The regime mobilized troops from outside Beijing and brought them in on June 4, 1989. Most students had fled the square before the soldiers’ arrival; nonetheless 2,600 were killed and 3,000 were wounded. Subsequent backlash was active, culminating on the third day as masses of dissidents held a central bridge in Beijing. Backlash casualties resulted from the deaths of 30 dissidents who were crushed as they lay in protest on railroad tracks in front of a locomotive. The state had ordered the engineer to drive forward.

Turkey: Kurds celebrated New Roz (new year) on March 21, 1992 in southeast Turkey. Police and troops in Cizre, Sirnak province and Batman killed 27 demonstrators, while in Istanbul, Ankara and Imir 38 died. On the following days 15,000 PKK guerrillas battled Turkish troops, bombed police stations, and supported Kurdish urban demonstrators.

Uitenhage: On the 25^th anniversary of the Sharpeville massacre South Africans were marching to a funeral of dissidents. Police opened fire on them, killing 19. During the next two days 18 homes of police officers were burned and a demonstration in Port Elizabeth developed. Police came in and clashed with the demonstrators, killing a man and a woman. On the third day, funerals of three dead protesters drew 30,000 mourners (New China News Agency and United Press International 3/24/85).

Vienna: Workers were incensed by the acquittal of police who were tried for killing a worker and a 10-year old child in Burgenland. Thousands of unarmed demonstrators rioted and burned the justice ministry on July 15, 1927. Police and troops shot into the crowd, killing 85 and wounding more than a thousand protesters. A 24-hour general strike followed. Workers occupied several districts of the city; Red Guards from the labor federation controlled main highways, and 500,000 workers marched on the Ringstrasse on July 16.

Vilnius: When Lithuania’s National Salvation committee declared independence at 11:00 pm on January 12, 1991, Soviet paratroops were dispatched to the Vilnius television station and tower. They shot into the crowd protecting the Salvation committee and the television station, killing 13 and injuring hundreds.

West Beirut: On September 14, 1982, Bachier Gemayel, head of the Lebanese Phalangist Party, was killed by a bomb placed in the party headquarters. On September 16 and 17, Phalangist forces under Israeli Defense Force control and command of Defense Minister Ariel Sharon systematically massacred over 3,000 Palestinian and Lebanese women and children in and around the Shatila and Sabra refugee camps. Backlash occurred in Palestine, i.e., the occupied West Bank and Gaza Strip.

White/Red Bridge: On March 16, 1988 dissidents gathered on Rangoon’s White bridge to protest the military dictatorship. Soldiers opened fire and killed 142, injuring 301. Thereafter dissidents changed the name to Red bridge in a symbolic gesture to the blood that flowed that day. Repression continued during backlash. On the third day the regime imposed a curfew and closed all the universities (Lintner 1990).

Wujek: When martial law was imposed in Poland on December 13, 1981, the Solidarity trade union was proscribed, as were strikes and protests. Miners held out in Wujek by remaining underground. Finally, on December 16, troops came to oust the miners and end the occupation. The miners resisted. Troops opened fire, killing 16 miners and wounding 39. Backlash continued in other mines and enterprises for almost two weeks before all protest died down finally dropped to insignificant levels.

Table 1

Repression Events

Event	Date	Number of Dissidents	Dead	Injured	Source

Amristar, India	4/13/ 1919	15,000	530	3,001	Fein 1977; Furneax 1963; Times and Irish Times
Barcelona Workers’ Massacre	2/18/1902	3,001	100	300	Manchester Guardian, 2/1902; Grant 1993
Berlin Spartakus	12/6/1918	1,000	18	30	Ettinger 1986; Ritter & Miller 1968
Caracas Caracazo	2/27/1989	30,000	300	800	Associated Press; UPI
Ciskei massacre	9/7/1992	20,000	28	288	Xinhau; UPI; Reuters
Derry Bloody Sunday	1/30/1972	3,000	13	14	Bell 1993; Mullan 1997
Dublin Bachelor’s Walk	7/26/1914	1,000	4	29	Jackson 1999
Dublin Bloody Sunday	11/21/1920	30,000	14	301	Coogan 1992
Escalante, Philippines	9/20/1985	15,000	21	30	Associated Press
Essen massacre	3/3/1923	50,000	13	31	Zum Ruhreinbruch 1976
GDR Worker rising	6/17/1953	100,000	21	187	Baring 1983; Spittmann & Fricke 1982
Guinea-Bissau dock strike	8/3/1959	1,000	50	101	Forrest 1992
Kwangju, South Korea	5/18/1980	30,500	216	650	Clark 1988
Managua election massacre	1/22/1967	6,000	34	101	Keesing’s Archive, 1967
Mendiola Bridge, Philippines	1/22/1987	10,000	18	301	Davis 1989
Mexico City student massacre	10/2/1968	10,000	400	301	New York Times; Story 1986; http://www.wlu.edu/~jbarnett/212/massacre.htm
San Salvador Cathedral	5/8/1979	500	23	70	AP 5/1979; Cockcroft 1996
Sharpeville massacre	3/21/1960	6,000	67	186	Mandela 1994; Meredith 1977; Pogrund 1991
Soweto riot	6/16/1976	16,000	1,000	3,000	Mandela 1994; New York Times
St. Petersburg Bloody Sunday	1/22/1905	16,000	175	625	Sablinsky 1976; Suhr 1989
Tblisi, Georgia	4/9/1989	3,001	19	36	AP 4/13/89
Tehran Jaleh Square	9/8/1978	20,000	501	4,001	Daneshvar 1996; Stempel 1981; Zabih 1988
Temple Mount	10/8/1990	3,001	20	140	Associated Press; Xinhau; UPI 1990
Tiananmen	6/4/1989	6,000	2,600	3,000	Ming Pao News 1989; Salisbury 1989
Turkey Kurd massacre	3/21/1992	3,001	80	67	Xinhau; AP; UPI
Uitenhage	3/21/1985	3500	19	31	Reuters 3/22/85
Vienna Black Friday	7/15/1927	300,001	85	1,057	Botz 1983, 1987
Vilnius, Lithuania	12/13/91	501	14	170	Lieven 1994
West Beirut	9/16/1982	5001	3,001	0	Petran 1987
White/Red Bridge, Burma	3/16/1988	3,001	142	301	Lintner 1990
Wujek coal mine	12/16/1981	3,001	16	39	http://lark.cc.ku.edu/~ronfran/data/index.html

Table 2

Backlash Mobilization and Casualties

Event	Event N	Post Day 1 N	Post Day 2 N	Post Day 3 N	Dead	Injured	Source
Amristar	15,000	30,000	30,000	60,000	0	0	Fein 1997; Times; Irish Times
Barcelona	3,001	80,000	100,000	100,00	0	0	Manchester Guardian, 2/1902
Berlin	3,001	150000	250,000	0	0	0	Laschitza 1982
Caracas	30,000	60,000	45,000	30,000	41	300	Associated Press; UPI
Ciskei	20,000	103,250	7,008	1,008	0	0	Xinhau; UPI
Derry	3,000	15,000	18,000	20,000	0	0	Irish Times 1972
Dublin 1914	1,000	3,000	6,000	9,000	0	0	De Rosa 1990; Times 1914
Dublin 1920	15,000	301	301	301	0	0	Irish Times, 11/1920
Escalante	15,000	95,000	33,001	600,001	4	3	Associated Press 1985
Essen	50,000	300,000	300,000	300,000	0	0	Cornebise 1977
GDR	100,000	300,000	250,000	200,000	0	0	Baring 1983; Ference 1994; Spittmann & Fricke 1982
Guinea-Bissau	1,000	0	0	0	0	0	Forrest 1992; Lopes 1987
Kwangju	30,500	100,001	200,001	500,001	0	0	Clark 1988
Managua	6,000	3,000	30,000	30,000	1	2	Facts on File 1967
Mendiola Bridge	10,000	300,001	300,001	300,001	3	0	Davis 1989
Mexico City	10,000	150,301	170,000	150,000	0	0	New York Times 10/3-6/1968
San Salvador	500	5,000	38,000	15,346	4	0	Associated Press, 5/1979
Sharpeville	6,000	15,000	30,000	30,300	0	0	Progrund 1991
Soweto	10,000	50,000	70,000	30,000	109	1,100	New York Times
St. Peters-burg	16,000	30,000	300,000	300,000	0	0	Sablinsky 1976; Suhr 1989
Tblisi	3,001	30,001	30,001	30,001	0	0	Associated Press, 4/13/89
Tehran	20,000	30,301	40,000	30,000	1	0	Times 9/1978
Temple Mount	3,001	2,060,000	2,023,001	2,021,150	0	13	Associated Press; UPI; Xinhau, 10/1990
Tiananmen	6,000	30,000	30,000	500,100	30	301	Ming Pao News 1989; Salisbury 1989
Turkey	3,001	15,000	15,000	16,000	8	0	Xinhau; AP; Reuters
Uitenhage	3,500	1,001	1,001	30,000	2	0	UPI, 3/24/85
Vienna	30,001	80,001	400,001	30,001	0	0	New York Times
Vilnius	501	140,001	300,001	30,0001	0	1	BBC 1991
West Beirut	5,001	100,000	100,000	90,000	0	31	United Press International 9/20/82
White/Red Bridge	3,001	2,000	20,000	0	135	1,701	Lintner 1990
Wujek mine	3,001	28,993	315,855	353,127	13	1,039	http://lark.cc.ku.edu/~ronfran/data/index.html

Figure 1: Mean mobilization levels per day

Table 3

Probable Numbers of Citizens Who Knew a Protester in the Repressive Events

Event	Number	Event	Number

Amristar	1,500,000	Mexico City	1,000,000
Barcelona	300,100	San Salvador	50,000
Berlin	100,000	Sharpeville	600,000
Caracas	3,000,000	Soweto	1,000,000
Ciskei	2,000,000	St. Petersburg	1,600,000
Derry	300,000	Tblisi	300,100
Dublin 1914	100,000	Tehran	2,000,000
Dublin 1920	3,000,000	Temple Mount	300,100
Escalante	1,500,000	Tiananmen	600,000
Essen	5,000,000	Turkey	300,100
GDR	10,000,000	Uitenhage	350,000
Guinea-Bissau	100,000	Vienna	3,000,000
Kwangju	3,050,000	Vilnius	50,100
Managua	600,000	West Beirut	500,100
Mendiola Bridge	1,000,000	White Bridge/Red Bridge	300,100
		Wujek mine	300,100

Table 4

Event and Backlash Tactics

Event	Event Tactics	Backlash Tactics
Amristar	rally	hartals (general strike) & rail strike
Barcelona	demonstration	general strike
Berlin	demonstration	demonstration with armed workers
Caracas	demonstration	demonstration & terror
Ciskei	demonstration	rally and occupation
Derry	demonstration	general strike
Dublin 1914	march	rallies and funeral processions
Dublin 1920	watching soccer game	arson
Escalante	demonstration	general strike
Essen	factory rally	general strike
GDR	demonstration	general strike
Guinea-Bissau	dock strike	none; reorganized for guerrilla war
Kwangju	demonstration	demonstration & student strike
Managua	election rally	general strike
Mendiola Bridge	demonstration	demonstration
Mexico City	demonstration	student strike & arson
San Salvador	rally	general strike & demonstration
Sharpeville	demonstration	rallies and strikes
Soweto	demonstration	riots and strikes
St. Petersburg	demonstration	general strike
Tblisi	rally	general strike
Tehran	demonstration	demonstration & general oil strike
Temple Mount	demonstration	general strike, marches, riots
Tiananmen	occupation	occupation
Turkey	demonstration	terror & guerrilla action
Uitenhage	demonstration	demonstration & arson
Vienna	demonstration	general strike, occupation & demonstration
Vilnius	rally	general strike & demonstration
West Beirut	none	demonstration
White/Red Bridge, Burma	demonstration	demonstration
Wujek mine	occupation	occupation

Table 5

The Risks of Backlash

Event	Backlash Risks
Amristar	identified workers striking; later punishment likely
Barcelona	identified workers striking; later punishment likely
Berlin	Free Corps shooting and clash with armed workers
Caracas	open protest repression; police work against terror
Ciskei	open protest repression
Derry	identified workers striking; later punishment likely
Dublin 1914	open protest repression
Dublin 1920	police work against terror
Escalante	identified workers striking; later punishment likely
Essen	identified workers striking; later punishment likely
GDR	identified workers striking; later punishment likely
Guinea-Bissau	none; reorganized for guerrilla war
Kwangju	open protest repression & identified students striking
Managua	identified workers striking; later punishment likely
Mendiola Bridge	open protest repression
Mexico City	identified students striking & police action against terror
San Salvador	open protest repression & identified workers striking
Sharpeville	open protest repression & identified workers striking
Soweto	open protest repression & identified workers striking
St. Petersburg	identified workers striking; later punishment likely
Tblisi	identified workers striking; later punishment likely
Tehran	open protest repression & identified workers striking
Temple Mount	open protest repression
Tiananmen	open protest repression & identified students striking
Turkey	police & military action against terror
Uitenhage	open protest repression & police action against terror
Vienna	open protest repression & identified workers striking
Vilnius	open protest repression & identified workers striking
West Beirut	open protest repression
White/Red Bridge, Burma	open protest repression
Wujek mine	martial law police action against strikers & occupiers

Figure 2: Mobilization after harsh repression with deaths and injuries in the first two days: Caracas, Mendiola Bridge, Soweto, Tehran, Temple Mount, Turkey, Uitenhage, White/Red Bridge, and Wujek events.

Figure 3: Mobilization after harsh repression with no deaths or injuries in the first two days.

Repression Events

The Risks of Backlash

Mobilization

Discussion

Appendix: Profiles of the Repression Events

Event

Backlash Risks