From political jiu-jitsu to the backfire dynamic: how repression can promote mobilization

Published in Kurt Schock (ed.), Civil Resistance: Comparative Perspectives on Nonviolent Struggle (Minneapolis, MN: University of Minnesota Press, 2015), pp. 145-167
pdf of published chapter

Uzbec translation by Sherali Niyazova, 2020

Brian Martin

Go to

Brian Martin's publications on nonviolent action

Backfire materials

Brian Martin's publications

Brian Martin's website

On November 12, 1991, thousands of people joined a funeral procession in Dili, the capital of East Timor. The procession was in honor of Sebastião Gomes, an independence supporter who had died two weeks earlier at the hands of Indonesian soldiers. The procession was a public protest against Indonesian rule, with marchers chanting slogans. Indonesian troops were on hand, walking to the sides of the mourners.

As the procession entered Santa Cruz cemetery, the Indonesian troops suddenly opened fire on the defenseless mourners, killing and wounding many. This event became known as the Dili massacre.

East Timor was formerly a Portuguese colony. It is located in the South Pacific Ocean, north of Australia, half of a small island in an archipelago, much of which constitutes the country of Indonesia. In 1975, Indonesian troops invaded and occupied East Timor. In the subsequent guerrilla war of resistance, tens of thousands of East Timorese died in the fighting or from starvation, perhaps as many as two hundred thousand, nearly one-third of the population.

Given the overwhelming military power of the Indonesian occupiers, it might seem that a massacre of peaceful protesters was simply an indication of their superiority and would send a signal to the East Timorese that they should avoid any form of protest. However, what actually happened was quite the opposite. The massacre provided a powerful stimulus to the international movement in support of independence for East Timor. It was the crucial turning point in the struggle (Kohen 1999; Nevins 2005).

It might seem strange that a dramatic event of vulnerability, weakness, and tragedy for the East Timorese could turn out to serve their cause so powerfully - in other words, that Indonesian troops, by exercising their lethal power, could harm their own cause. Explaining how and why this happened is the purpose of this chapter.

The Dili massacre was not an isolated incident. There are a number of other attacks on peaceful protesters that were counterproductive for the attackers. In 1905 in Russia, then ruled by the tsar, protesters converged on the Winter Palace - the tsar's official residence and seat of government - to present their grievances. The tsar's police shot many of them dead, an atrocity that weakened support for the government and laid the foundation for the 1917 revolution (Sharp 1973).

In 1960 in South Africa, during the apartheid era, black protesters across the country rallied against the pass system that required black people to carry internal passports. In the town of Sharpeville, following a series of small events, police opened fire on the protesters, killing dozens of them. This massacre discredited the South African government around the world and was a significant event in the anti-apartheid struggle (Frankel 2001).


Russia 1905, Sharpeville 1960, and Dili 1991 were individually important for the struggles involved. Together, they point to an important phenomenon: attacks on peaceful protesters can be counterproductive for the attackers, generating more support for the protesters and their cause. Why is this important? The usual assumption, in history books and popular media, is that violence is more powerful than peaceful methods. Peaceful protest is all very well, so the usual thinking goes, but if governments are willing to use force, the protesters will be crushed, as in China in 1989. However, the examples from Russia, Sharpeville, and Dili show that this assumption sometimes is wrong. In cases like these, violence against peaceful protesters might succeed on a superficial level, but rather than crushing the protest movement, it actually makes the movement more powerful. The examples from Russia, Sharpeville, and Dili show that standard ideas about the power of violence are flawed.

In this chapter, I trace the evolution of ideas for understanding how attacks on peaceful protesters can be counterproductive. I start with Richard Gregg's analysis of moral jiu-jitsu and its relevance to Gandhi's salt march in India in 1930. Next I look at Gene Sharp's revamp of Gregg's concept into a broader concept: political jiu-jitsu. Then I point to some limitations of political jiu-jitsu and introduce a yet broader concept, backfire, and show its relevance to some of the campaigns analyzed by Gregg and Sharp. Finally, I spell out some wider implications of the phenomenon.

Moral Jiu-jitsu

Nonviolent action is defined negatively as social action that does not involve physical violence, which means it covers methods such as rallies, strikes, boycotts, fraternization, fasts, and sit-ins (see the introduction by Kurt Schock for elaboration). The study of nonviolent action commonly begins with Mohandas Gandhi, who more than any other individual promoted the use and understanding of this method of social action. Gandhi's primary testing grounds were South Africa from about 1905 to 1916 and India thereafter until his death in 1948.

However, it is important to remember that people have used nonviolent action for centuries. In the 1850s and 1860s, Hungarian nationalists used boycotts and set up their own institutions in a struggle for independence from the Austrian empire (Csapody and Weber 2007). From 1898 to 1905, Finns used a variety of methods against domination by their Russian occupiers (Huxley 1990). These and other examples excited the imagination of pacifists such as Leo Tolstoy.

In these early campaigns, nonviolent action was used spontaneously, without formal theory or training. Gandhi showed how it could be used to pursue social goals. He argued for nonviolence on the grounds that it was ethical: in his view, using violence was a denial of the opponent's humanity. Gandhi began his campaigns by personal appeals to the opponent. Only when these were rebuffed did he initiate methods such as boycotts and strikes. These were intended to encourage his opponents to engage in dialogue, whereas using violence was more a denial of dialogue.

Gandhi was an outstanding practitioner, but his voluminous writings were unsystematic: it is hard to find a succinct account of how nonviolence works and how to use it to achieve a movement's goals. Therefore, understanding Gandhi's legacy has largely fallen to others, who might be called his interpreters (Bondurant 1958; Gregg [1934] 1966; Shridharani 1939). One of the earliest and most important was Richard Gregg, an American, who read about Gandhi's efforts and went to India to learn more. Gregg ended up writing a book titled The Power of Nonviolence, first published in 1934, which drew on what he had learned about the nonviolent campaigns in India led by Gandhi.

One of Gandhi's most famous campaigns was the 1930 salt march. Gandhi came up with the idea of challenging the British salt laws. At the time, the British colonial rulers held a monopoly on salt, a basic commodity used by everyone. There was a tax on salt, and no one was allowed to manufacture salt except under British authority.

The salt march was a planned civil disobedience, with a long buildup to maximize impact. The plan was to make salt from seawater, which was illegal under British colonial law. Gandhi and his supporters designed a two-hundred-mile march to the sea. The protesters walked each day, stopping at different localities for speeches and building support. The march took twenty-four days. The British authorities were caught in a dilemma: nothing illegal was being done during the march. If they arrested Gandhi before the marchers reached the sea, this would be seen as provocative and might stimulate greater resistance. Yet every day that passed, the campaign gained greater strength, with news of it encouraging resistance across the country (Dalton 1993; Weber 1997).

When the marchers reached the sea and began picking up watery mud to make salt, Gandhi and others were arrested. The climax of the campaign came a bit later, when protesters engaged in a set-piece action to enter a salt manufacturing factory at Dharasana. Indian police, who were armed with lathis, hard batons covered in steel, protected the factory. The protesters, called satyagrahis, calmly walked forward toward the salt works and were met by the police, who ruthlessly beat them with their lathis until the protesters fell down, bloody and sometimes seriously injured. Supporters came to pick up the injured protesters and take them to a hospital. Meanwhile, more protesters walked forward to be beaten - and none resisted.

This was a classic confrontation between peaceful protesters and armed police, and on the surface it was an obvious victory by the police. The protesters were stopped from entering the salt works, and so many were injured that eventually the action had to be called off.

Although the police were victorious at a superficial level, the action at Dharasana was incredibly effective at mobilizing support for the Indian independence movement across the country and internationally. Indeed, it could be considered a turning point in the entire struggle, uniting what had been a fragmented and apathetic population into a more unified and committed stance against British rule (Weber 1997).

So how did this work? Gregg drew on psychological theory and came up with the concept of "moral jiu-jitsu." The name is based on a metaphor to the sport of jiu-jitsu, a martial art in which the opponent's momentum and force are used against him or her: a person can use jiu-jitsu moves to beat a larger and stronger but less skilled opponent. Gregg's idea is that nonviolence performs a similar process but in the realm of morals. Activists committed to nonviolence have a moral advantage that, Gregg believed, can throw powerful opponents off balance.

Gregg advanced a series of psychological explanations for the phenomenon of moral jiu-jitsu. In particular, he expected that police beating protesters would become repulsed by their own behavior. Here, I won't go through the details of Gregg's analysis because it has been shown to be inadequate. Decades later, Gandhian scholar Thomas Weber (1993) analyzed the salt march and showed that the police who beat satyagrahis at Dharasana were not deterred from their actions by any moral factors. Indeed, some were infuriated by the passivity of the satyagrahis - by their refusal to fight back - and became more energetic in their beatings. In this case at least, moral jiu-jitsu did not operate at the level of the psychology of the violent attackers.

Although Gregg's explanation for the jiu-jitsu effect had limitations, he was on to something very important, because the salt march was incredibly effective and the beatings at Dharasana were counterproductive for the British. Moral jiu-jitsu was only one aspect of Gregg's book, which was highly influential for decades afterward.

I return to the salt march later in the chapter. Next, though, I continue to follow the trajectory of the concept of moral jiu-jitsu.

Political Jiu-jitsu

U.S. scholar Gene Sharp was initially a pacifist and follower of Gandhi. Then, during the 1950s, he developed his own approach to nonviolence, an approach that has become highly influential.

Gandhi's approach was based on a moral imperative not to use violence. Sharp threw off this moral injunction but arrived at the same conclusion through another route. Sharp argued for nonviolent action on the grounds that it is more effective than violence. This is said to be pragmatic nonviolence in contrast to Gandhi's principled nonviolence.

It is easy to misinterpret Sharp's claim that nonviolent action is more effective. He is not saying that nonviolent action will always triumph over a ruthless opponent. Rather, Sharp is making a claim about what will be more effective against a powerful, ruthless opponent like Stalin, Hitler, or, for example, the South African apartheid government or the Indonesian occupiers of East Timor. Violence can be, and was, used against these oppressors. Sharp says nonviolent action against these oppressors can be more effective than violence.

Sharp's most important work is The Politics of Nonviolent Action, published in 1973. Part one of the book presents Sharp's theory of power, which is that the power of rulers depends on consent of the ruled, namely acquiescence or support from subjects (including functionaries, especially police and military forces). If this consent is withdrawn, then the ruler's capacity to rule dissolves. This so-called consent theory of power has been subject to considerable discussion and critique, but it is not the focus here. The key point is that the idea of withdrawing consent provides a warrant for nonviolent action as a way to undermine rulers.

Part two of The Politics of Nonviolent Action lists 198 different methods of nonviolent action, grouped into three main categories: protest and persuasion, such as petitions and rallies; noncooperation, including numerous types of strikes and boycotts; and intervention, including fasts, sit-ins, sabotage, and parallel government. Sharp illustrates each of the 198 methods with historical examples. One important point here is that nonviolent action can take many forms; it is much more than the public image of rallies and sit-ins.

Part three presents a model of how nonviolent campaigns proceed, which Sharp calls the dynamics of nonviolent action. It has six stages or components: laying the groundwork, challenge that leads to repression, maintaining solidarity and nonviolent discipline, political jiu-jitsu, three modes of success, and the redistribution of power. Of special interest here is the stage called political jiu-jitsu. Sharp says this occurs when oppressors violently attack peaceful protesters, resulting in greater support for the grievance group.

Sharp (1973, 698n1) acknowledges his debt to Richard Gregg's concept of moral jiu-jitsu. However, Sharp says that the jiu-jitsu phenomenon is not based solely on moral dynamics at a psychological level but also involves political and social dimensions. This encompasses the possibilities that witnesses to a violent attack may be repulsed and that entire groups may want to dissociate themselves from the attack. Smithey and Kurtz (1999, 111-13) call this the paradox of repression.

Sharp (1973, 658) says a campaign can often achieve its goals without political jiu-jitsu being involved. For example, rulers may grant the protesters their demands; another possibility is that rulers may find ways to stymie protesters without making a brutal public attack that is counterproductive.

It is illuminating to see where Sharp places political jiu-jitsu within the stages of the dynamics of nonviolent action. A crucial prior stage is "solidarity and discipline to fight repression," which means the protesters must avoid responding to attack or provocation by using violence themselves. When protesters use violence, attacks against them seem more justified, and the jiu-jitsu effect is weakened or nullified.

Sharp presents numerous examples of political jiu-jitsu, such as the 1905 Bloody Sunday massacre in Russia and the 1960 Sharpeville massacre. He gives examples of how attacks on nonviolent protesters can help win over uncommitted third parties, cause splits among the regime's allies, and mobilize support from members of the grievance group. For example, concerning the Bloody Sunday massacre of January 9, 1905, Sharp (1973, 659) says ,"Two days after the massacre, Kokovtsev, the Minister of Finance, reported to the Tsar that not only had the killings impaired morale at home, but that Russian financial credit abroad had been affected," illustrating how uncommitted third parties can be won over. Sharp (679) also writes that "the killing and wounding of hundreds of peaceful marchers who were under instructions to remain nonviolent . . . destroyed [the] alliance of the poor with the Tsar," illustrating how support and participation from the grievance group - those oppressed by the tsarist regime - could be increased.

When Political Jiu-jitsu Does Not Occur

Political jiu-jitsu can be powerful when it occurs. However, Sharp does not discuss examples when peaceful protesters are violently attacked, but there is no jiu-jitsu effect.

The 1991 Dili massacre wasn't the only massacre by Indonesian troops. There had been other, earlier massacres, just as serious (Dunn 2003, 333), but they received little attention and did little to generate greater international support for the independence struggle. Why not?

Another example from Indonesia is telling. In 1965, there was an alleged coup against the government that triggered a massive killing spree, mainly against communists. Over the next year, hundreds of thousands of Indonesian civilians were killed, in what can be called genocide or politicide (Cribb 1990). There was relatively little opposition to the killings and no armed resistance. So why didn't this massive violation of human rights lead to international opposition? One reason was that this was an anticommunist bloodbath, so governments of key countries, especially Australia and the United States, were pleased by the anticommunist takeover in Indonesia (Chomsky and Herman 1979). This was in the middle of the Cold War, when political developments were interpreted through the lens of communism versus anticommunism.

If some violent attacks on nonviolent protesters do not result in political jiu-jitsu, what is going on? There are frequent instances in which police have used force against nonviolent protesters in rallies, apparently without adverse consequences. Does this mean that the external conditions were not right, or that nonviolent discipline had not been maintained sufficiently? Remember that the salt march was carefully planned to mobilize support along the way. The careful preparations for the march, designed to increase popular interest and media attention, meant that the police beatings at Dharasana had a far greater impact than they might have otherwise.


To explain why some beatings and shootings of nonviolent protesters lead to little wider reaction, with minimal political jiu-jitsu, I developed the idea that the perpetrators take action to reduce the reaction. Five types of tactics can be used to reduce outrage.

Cover-up includes keeping actions hidden, using censorship, and falling below the threshold for mass media coverage. Devaluation includes drawing on prejudices such as racism, applying derogatory labels, spreading rumors, and publicizing adverse information, real or manufactured. Reinterpretation includes a range of discursive techniques such as providing false figures, denying that anyone is hurt, giving distorted interpretations, saying others are to blame, and using a perspective that puts perpetrators in a good light. Official channels include grievance procedures, ombudsmen, formal inquiries, and court cases. When perpetrators are powerful, official channels often don't work to provide justice, but they may give the impression that the issue is being dealt with fairly. Intimidation includes threats, job dismissals, and physical assaults - anything that might scare the target. Rewards include payments, promotions, and citations for taking the side of the perpetrator.

Sometimes one or two of these tactics are enough to limit outrage. Massacres in East Timor prior to 1991 were covered up: information about them only trickled out when some survivors told their stories. Survivor stories can be powerful, but for international audiences they have limited news value without photos or firsthand accounts by journalists. Furthermore, the Indonesian government denied the massacres, using the technique of reinterpretation via lying. Western news media commonly defer to government sources, or at least quote them, unless there is overwhelming contrary evidence.

When the methods for reducing outrage are effective, there is little or no public awareness of the attack and seldom much information for documenting the methods. Therefore, ironically, the best evidence of the use of the methods comes from cases in which the methods are unsuccessful - at least in part - and backfire occurs. That is what happened with the Dili massacre, so it is worth recounting evidence about the methods used by the Indonesian military and government to curtail outrage over a massacre of peaceful protesters.

Cover Up the Action

The government cut off all electronic communication out of East Timor.

Devalue the Target

There was already devaluation inherent in attitudes toward the East Timorese: from the point of view of Javanese rulers, the East Timorese were an inferior group. There is some evidence of more active efforts at devaluation. For example, General Try Sutrisno, commander-in-chief of the Indonesian armed forces, referred to the protesters as "scum" and "delinquents" (McMillan 1992, 162).

Reinterpret What Happened through Lying, Minimizing, Blaming, and Framing

The Indonesian government initially reported that 19 people had died; later this was raised to 50. An independent inquiry found that 271 were killed. The Indonesian government figures are an example of lying. In this case the lies served to minimize the death toll, so this could also be classified as minimizing.

The Indonesian government claimed that protesters were carrying weapons and were responsible for triggering the shooting. This is an example of lying and blaming.

The Indonesian government assumed the legitimacy of its occupation, presenting the protest as unacceptable and the cause of problems. This constitutes a framing of the issue from the occupier's point of view.

Use Official Channels to Give an Appearance of Justice

Both the Indonesian military and the Indonesian government initiated inquiries. There had never before been any official inquiries into massacres in East Timor. The obvious reason for the inquiries into the massacre at Santa Cruz cemetery was the international publicity that it received.

The inquiries gave an appearance of justice: a few individuals were found guilty and given short prison sentences. However, these were token outcomes. No senior commanders were tried, and the prison sentences were light compared to those given to East Timorese who were guilty of nothing more than peaceful protest.

Intimidate and Reward People Involved

The massacre itself was a potent form of intimidation of the East Timorese population. Also, Indonesian troops subsequently searched out independence supporters in Dili, beating and killing them.

Rewards refer to inducements to support or keep quiet about the massacre. This mainly applied to Indonesian troops, especially commanders, who had something to gain by endorsing the official line.

The Dili massacre, a lethal attack on peaceful protesters, had the potential to backfire against the attackers. But the Indonesian military and government were not going to just let this happen. They took numerous measures to deter and limit popular reaction. These can be classified into the categories of cover-up, devaluation, reinterpretation, official channels, and intimidation and rewards.

However, the Dili massacre turned out to be a public relations disaster for the Indonesian government. To understand why, it is useful to look at countertactics that can be taken by protesters and their allies to increase outrage from the massacre. These can be classified into five categories, each one corresponding to one of the five methods of reducing outrage.

The key method at Dili was "expose the actions." Western journalists, from Australia, Britain, and the United States, were present throughout the march. They witnessed the massacre, and some of them were beaten. When they returned to their home countries, they were able to provide vivid, credible testimony about what had happened.

Furthermore, the journalists had images. Max Stahl, a British filmmaker, took video footage of the massacre. He expected that he would be searched on departure, so he hid the videos in the cemetery, later returned and collected them, and entrusted them to a friend who got them out of the country. When Stahl arrived in Darwin, he was searched by Australian officials seeking to confiscate any video evidence. At that time, the Australian government recognized East Timor's incorporation into Indonesia - the only government in the world to do so.

Stahl's experience is an illustrative episode in the tactics aimed at increasing or decreasing outrage. The Indonesian government sought to confiscate evidence. Stahl anticipated this, giving his videos to a friend who would be less likely to be suspected. Australian officials, serving the interests of the Indonesian government, searched Stahl in Darwin. This was an engagement over one aspect of outrage management: cover-up versus exposure. Stahl and the other journalists were successful in taking their story to international audiences.

The Dili massacre illustrates the two core requirements for backfire to occur. The first requirement is that significant numbers of people believe that something is unjust, unfair, wrong, or inappropriate. The typical response is outrage, anger, disgust, revulsion, concern, or some other emotion indicating a negative attitude. Shooting unarmed protesters in cold blood certainly satisfies this requirement, but not everyone responded negatively. The troops, commanders, and government officials apparently thought the massacre was acceptable; there is no evidence that any of them were upset.

The second requirement for backfire to occur is that information about the event is communicated to receptive audiences. The key audience for information about the Dili massacre was international civil society groups. The eyewitness testimony, photos, and Max Stahl's video were crucial communication tools. Portions of the video were broadcast in several countries. John Pilger and David Munro used Stahl's video material in a 1994 television documentary, Death of a Nation, about East Timor that was screened widely and used to gain support for East Timor solidarity groups.

It is useful to think of several potential audiences. One was people in East Timor. They certainly learned about the massacre. However, they were already largely united against the occupation; news of the massacre probably did not make a big difference. Another audience was the Indonesian military. Leaders and troops involved in the massacre knew about it, of course, but they did not think it was wrong - they were involved in an ongoing war - so there was little prospect of troops turning against the occupation. Foreign government leaders and officials were also unlikely to be greatly influenced by information about the massacre. For example, Australian foreign affairs officials and politicians had supported the Indonesian takeover of East Timor from the beginning and continued to be apologists for Indonesian actions in East Timor.

The most receptive audience was composed of individuals who already had some awareness or connection with Indonesia. For example, some members of the U.S. Congress were outraged. They knew about the U.S. government's support for the Indonesian military, and they were outraged to see this aid being used to kill civilians. They sponsored resolutions calling for all U.S. military aid to Indonesia to be cut off.

The massacre provided an enormous boost for East Timor solidarity groups in several countries. Members were given renewed incentive to take action, and new people were brought into the movement. Solidarity groups were composed largely of people with no vested interest in the conflict: their members were committed to human rights and self-determination, so they were very receptive to information showing abuses in East Timor.

In light of the struggle between cover-up and exposure concerning East Timor, political jiu-jitsu can be rethought. Rather than being a predictable outcome of a violent attack on peaceful protesters, it is better conceived of as an outcome of an event that is contingent on efforts by various parties to manage outrage. Cover-up is commonly the initial method used by perpetrators; when it is successful, little or no outrage occurs. Activists hence need to expose what happened to receptive audiences.

Cover-up versus exposure is the first of a set of competing tactics. The next set is devaluation versus validation. When the targets of attack are of low status or stigmatized by allegations, such as being called terrorists or criminals, then what is done to them may not seem so serious. On the other hand, if the targets are highly valued, then outrage will be greater. It is easier to devalue violent protesters: they are seen as deserving or inviting violence in return. Protesters typically have higher status when they behave in a dignified fashion, dress respectably, and have prestige because of their occupations or accomplishments. Police know it looks much worse to beat and shoot business executives, Buddhist priests, or Catholic nuns than protesters who look scruffy and dress strangely.

Reinterpretation is the technique of explaining things in a way that justifies an action or makes it seem not so important. This can be done by lying (for example, about who initiated the shooting), by minimizing (for example, by underestimating the number of people hurt or the degree of harm), by blaming (for example, by saying just a few rogue troops were responsible rather than commanders), or by framing (for example by assuming that protest was illegitimate). Of these techniques, framing is most likely to be legitimate, in the sense that it is a way of viewing the world that is sincerely held. The techniques of lying, minimizing, and blaming involve deception or denial of responsibility. Reinterpretation can be challenged by providing accurate information, assigning responsibility where it is deserved, and using a frame that emphasizes the injustice involved.

Official channels include courts, official inquiries, ombudsmen, expert panels, and any other procedures or processes that appear to provide a fair treatment of the matter. Official channels are a powerful method of defusing outrage, because many people presume that the formal procedures provide justice. Some official channels do this, but when perpetrators are powerful, official channels often provide only an illusion of justice.

At the best of times, official channels dampen outrage. They are almost always slow, which means people's passions die down as weeks, months, or years go by. Courts and other agencies are highly procedural: they follow technical rules that most people do not understand or care to explore, so outsiders lose interest. Official channels rely on experts (for example, lawyers) and move an issue from the public arena to a narrow arena where only specialists are involved. Sometimes inquiries give sound, sensible recommendations, but governments fail to act on them. Courts sometimes rule on behalf of protesters, but then prosecutors appeal. To be effective, using official channels can require staying power and a lot of money, which means powerful perpetrators have a great advantage.

The power of official channels is based on people's belief that they work. Courts, for example, often have great credibility. Only those directly involved have a good idea of how the system can be used and abused.

For those who understand the political uses of official channels and seek to promote outrage over injustice, there are two main approaches. One is to try to discredit the official channels, such as by pointing out vested interests. The other approach is to mobilize support - in other words, to keep campaigning. The main thing is not to rely on official channels to provide a satisfactory solution.

The final technique for perpetrators is intimidation. A violent attack on peaceful protesters is itself a form of intimidation. Intimidation includes all methods to discourage action due to outrage. For example, witnesses of a massacre might be afraid to protest because they too might be assaulted. Local journalists might be afraid to report the story because of possible reprisals. To challenge intimidation, at least some people must be able and willing to resist, namely to protest or otherwise act against the perpetrators.

There is another technique by perpetrators, grouped with intimidation: rewards. It can also be called offering bribes, benefits, or advantages. The basic idea is that those who support or do not oppose the attack may gain rewards. For example, commentators who justify the attack might be given new opportunities or promotions. Resisting rewards means being willing to expose and oppose the attack despite losing possible benefits.

Analyzing the struggle over outrage shows that Sharp's political jiu-jitsu can be a complex process, involving tactics on both sides in fierce and ongoing attempts to dampen or promote responses to a violent attack. Throughout this discussion, I refer to a "violent attack" or a "massacre." Using this sort of language is already to have made an interpretation and taken a side. Perpetrators use methods of reinterpretation to say that something else was involved, perhaps a riot or a terrorist attack on police. Taking the side of nonviolent protesters is natural for Gregg, Sharp, and protesters themselves, but it is important to remember that others may have different perspectives, and therefore a struggle over perceptions and reactions is crucial to the outcome.

The model of tactics for managing outrage is called the backfire model because when the perpetrator's tactics are inadequate or unsuccessful, the attack is counterproductive: it backfires. As an outcome, backfire is the same as political jiu-jitsu: a violent attack on peaceful protesters actually benefits the cause of the protesters. As a process, the backfire model involves assessment of the five methods of reducing outrage and the five corresponding methods of enhancing outrage (Martin 2007, 2012).

Other Issues

If perpetrators use cover-up, devaluation, reinterpretation, official channels, and intimidation to reduce outrage over violent attacks on nonviolent protesters, perhaps the same methods are used in relation to other sorts of injustices. This is exactly what is observed.

In 1991 - the same year as the Dili massacre - a man named Rodney King was arrested by Los Angeles police. During the arrest, King was twice hit by tasers, whose 50,000-volt shocks normally make people collapse, but King was still active. The police then struck King with metal batons dozens of times before he was fully subdued.

King was not a peaceful protester. He was drunk and tried to avoid arrest, leading police in a high-speed chase before his car was brought to a halt. The beating King received was severe, but Los Angeles police had done much worse to others. The difference was that in the commotion over the arrest, a man named George Holliday, who lived nearby, recorded the arrest on video. When Holliday's video was broadcast on television, it ignited massive uproar over the brutality of the arrest.

There is extensive documentation about the beating of King, from both supporters and critics of the police (e.g., Koon 1992; Owens 1994). From all this information, there is ample evidence that the police used all five methods for limiting outrage: cover-up, devaluation, reinterpretation, official channels, and intimidation. However, these efforts were insufficient: the beating backfired, greatly damaging the reputation of the police (Martin 2005).

Backfire analysis can be applied to a wide range of issues, including sexual harassment (McDonald, Graham, and Martin 2010), censorship (Jansen and Martin 2003; Yecies 2008), treatment of refugees (Herd 2006), torture (Martin and Wright 2003), war (Martin 2004; Riddick 2012), and genocide (Martin 2009). Remarkably, perpetrators in all these areas have used the same sorts of outrage-management techniques. Opponents can use the same sorts of countertactics: exposure, validation, interpretation, mobilization, and resistance.

The backfire model shows the potential of taking ideas from nonviolent action and adapting them to apply to other arenas. Gregg and Sharp assumed that nonviolent discipline was crucial to the jiu-jitsu effects they observed, but this was a special case of a more general phenomenon. Outrage will be greater when violence is used against protesters who are nonviolent, but outrage is still possible even when protesters use some violence - as long as the encounter is widely seen as unfair, which usually requires that the police violence be serious, unprovoked, indiscriminate, or otherwise unjustified. The keys to backfire, as an outcome, are that people perceive an action as unfair and that credible information about the action is communicated to wider audiences.

The lessons concerning backfire, drawn from numerous issues ranging from censorship to genocide, can then be fed back into cases involving violence versus nonviolence. In 2010, a flotilla of ships, with protesters, attempted to deliver supplies to Gaza and break the blockade imposed by the Israeli government. Before the flotilla reached Gaza, Israeli forces attacked, arrested the protesters, and impounded the ships. The Israeli military and government used all the usual methods to inhibit outrage but were unsuccessful: the attack on the flotilla was a massive public relations disaster internationally (Martin 2010). What is interesting here is that not all the protesters remained nonviolent: some of them fought against the Israeli attackers. However, the Israeli violence was much greater, with nine protesters killed.

This case illustrates that a perception of injustice can occur even when protesters use violence. However, it is far easier to discredit protesters when they use violence, so maintaining nonviolent discipline remains important. The point here is that the outcome is the product of a struggle over information and interpretations rather than being determined purely by the interaction itself.

The classic case is Dharasana, the engagement that was the culmination of the 1930 salt march and one of the cases that inspired Gregg to conceptualize his concept of moral jiu-jitsu. The events at Dharasana had a psychological impact, but in a different way than Gregg realized. A United Press journalist, Webb Miller, was present and wrote moving accounts of the beating of nonresisting protesters. Readers in Britain, the United States, and elsewhere were incredibly moved by Miller's accounts, which were reprinted in hundreds of thousands of copies. The incident at the salt march, via international communication at the time, namely the printed word, undermined popular support for British rule over India within Britain and other Western countries and thus laid the basis for independence less than two decades later (Weber 1997).

There was no armed struggle for independence, and the total death toll of the independence struggle was modest. This was not due to the kindheartedness of the British. In a very different context, Kenya, where there was armed resistance, the British rulers used killings, torture, and concentration camps to quell the Mau Mau Rebellion (Elkins 2005).

At Dharasana, the British used all five methods to inhibit outrage. For example, they attempted to block distribution of Miller's dispatches (cover-up), denigrated the satyagrahis (devaluation), claimed that protesters were not really hurt and indeed were faking their injuries (reinterpretation), offered a conference that promised much more than it delivered (official channels), and imprisoned numerous members of the independence struggle (intimidation). The beatings at Dharasana thus did not cause a jiu-jitsu effect simply because violence was used against peaceful protesters. The events were the subject of a struggle over information, interpretations, and emotions in which each side used tactics to foster attitudes and actions serving its own goals.


For most people, it seems obvious that violence will prevail over nonviolent protest - all that is required is the political will to use force. From this point of view, the only guaranteed way to overcome violence is by counterviolence. To claim that nonviolent action can be more effective than violence can appear counterintuitive at first glance.

Political jiu-jitsu is one of the processes that enables nonviolent action to be more effective than violence. Struggles involving violent or nonviolent methods, or both, can be thought of as efforts to win support. Violence attempts to win support through destruction and domination: opponents succumb or acquiesce. Nonviolent action attempts to win support through pressure and demonstration of commitment without physically threatening opponents. The power of nonviolent action comes through a greater opportunity for participation in action and through a greater capacity to engage with and win over opponents and third parties. (In the next chapter, Véronique Dudouet elaborates on the role of external third parties in nonviolent conflict.)

The concept of political jiu-jitsu helps explain the power of nonviolent action. The opponent has a weapon, namely violence, that can be risky to use because it may create greater support for nonviolent protesters. However, there is more to it than just violence versus nonviolence. Attackers can use a variety of methods, from cover-up to intimidation, to inhibit outrage. In order to make an attack backfire, protesters need to be aware of these methods of inhibition and be prepared to counter each one of them. The countertactics that promote outrage - exposing what happened, validating the target, interpreting events as unjust, mobilizing support (and avoiding official channels), and resisting intimidation - provide a toolkit for campaigners. These tactics can be used in designing actions to increase the likelihood of backfire and therefore to deter opponents from violent attacks in the first place.

The backfire model is a way of looking at a struggle as a strategic engagement. Each side has various possible tactics and can use them in various ways. That means struggles may evolve as each side becomes familiar with the opponent's tactics. For example, attackers may seek to hide their actions through various forms of mass media censorship. Protesters can respond by using social media, such as cameras and YouTube. Governments can then try to censor the Internet, and protesters can use proxies and other means to get around the censorship.

The evolution of ideas from moral jiu-jitsu to backfire shows the potential insights to be gained from studying the dynamics of nonviolent action. The phenomenon of backfire can be observed in various situations, such as censorship, but is perhaps most striking in violence-versus-nonviolence engagements. The step to backfire involves starting with observations - such as that some attacks are counterproductive - and examining them closely to ask why they develop the way they do.

As discussed earlier, Sharp, in his classic book The Politics of Nonviolent Action (1973), proposed what he called the dynamics of nonviolent action, a series of stages or features of nonviolent campaigns, with political jiu-jitsu as one of them. The backfire framework developed from an analysis of why political jiu-jitsu does not always occur, namely as a result of tactics used by the perpetrators. This same approach can be applied to other stages of Sharp's dynamics.

The first stage in the dynamics is "laying the groundwork," which involves developing ideas, networks, plans, organizations, and everything else that can provide a foundation for taking action. However, in many circumstances involving ongoing oppression and repression, the groundwork seems never to be laid: opposition groups are squashed or splintered or go down unproductive paths. The normal focus is on the opposition forces themselves. To develop Sharp's framework further, it would be worth examining the actions taken by rulers to inhibit opponents or cause them to work in unproductive ways, as Sharon Nepstad does in chapter 4. There are some well-known examples, such as the U.S. Federal Bureau of Investigation's COINTELPRO program beginning in the 1960s that aimed to disrupt social movements, including by stimulating distrust between them (Blackstock 1976). However, cunning governments usually keep their surveillance and disruption operations hidden, so not enough is known about the techniques involved. There is much to learn by further study of the groundwork-laying stage. From the point of view of those who advocate nonviolent action, it might be argued that government interventions at this stage can encourage movements to adopt violence, thereby undermining their capacity to mount a potentially more powerful nonviolent campaign.

The final stage in Sharp's dynamics is the "redistribution of power." Of all the stages, this one is the least well articulated, yet it is crucially important for the success of campaigns. There are numerous examples of effective campaigning that leads to some changes, especially in government leaders, but no structural change. In other words, the change is more symbolic than substantive. This is reminiscent of the role of official channels in dampening outrage. One of the reasons for the neglect of the stage of redistribution of power is that it mostly occurs out of the public eye, namely through negotiations inside political parties, government bureaucracies, corporations, militaries, and other hierarchical systems. Civil resistance theory and practice have mainly focused on activities in the public sphere, such as rallies and strikes, and have given little attention to struggles inside organizations. Civil resistance theory could provide the stimulus for developing a framework for understanding the politics of transforming organizations.

One of the striking features of civil resistance theory has been its close connection with practice. Gregg's concept of moral jiu-jitsu grew out of observations of Gandhi's campaigns. Gregg's work was highly influential for decades, making more people aware of nonviolent action and thus indirectly influencing campaigners. Sharp drew on studies of hundreds of nonviolent actions and campaigns. His work has been used by activists across the world. This suggests that further developments in the theory and practice of nonviolent action are most likely to involve linking theory and practice in some way.

Reflecting on Theory Development

By examining the evolution of ideas from moral jiu-jitsu to backfire, it is possible to derive some suggestions for future theory development in civil resistance. Gregg, in developing the concept of moral jiu-jitsu, drew on his observations and understandings of nonviolent campaigns in India and used ideas from one discipline, psychology. Sharp built on moral jiu-jitsu in formulating his concept of political jiu-jitsu. Sharp thus had the advantage of being able to look for jiu-jitsu-type events. He explored several. In addition, he drew on several disciplines, including psychology, politics, and economics. Sharp did not disown the psychological processes postulated by Gregg but rather supplemented them.

In addition, Sharp embedded political jiu-jitsu into a wider framework, his dynamics of nonviolent action, which included preconditions for the jiu-jitsu effect. For example, "maintaining nonviolent discipline" ensures that attacks on protesters are seen as unfair because of the qualitative difference between violence and nonviolence.

Sharp thus used four main theory-building techniques. First, he identified moral jiu-jitsu as an important phenomenon, worthy of further attention and development. Second, he gathered more information, so he had a wider body of material from which to examine moral jiu-jitsu. Third, he broadened the concept, expanding it from a psychological phenomenon to one involving psychological, political, and social factors. Fourth, he placed political jiu-jitsu into a larger framework of stages or facets of a civil resistance campaign.

Sharp's political jiu-jitsu was the inspiration for developing the backfire model. My first step was to ask why political jiu-jitsu sometimes does not occur. Rather than looking solely at successful nonviolent action, the scope is broadened to include failures. This led me to look at actions by the opponent, namely what the attacker was doing that might be reducing the possibility of the attack being counterproductive. This broadened the model into a strategic engagement involving tactics on each side. I then looked at more cases by going outside the violence-versus-nonviolence template used by Gregg and Sharp, addressing numerous topic areas, such as censorship, police beatings, and torture.

I used four main theory-building techniques in developing the backfire model. First, I identified political jiu-jitsu as an important phenomenon, worthy of examination and elaboration. Second, I expanded the range of observations to include unsuccessful cases. Third, I treated the interaction as a strategic interaction, involving tactics on both sides, as well as looking at backfire as an outcome. Fourth, I examined cases in areas outside violence-versus-nonviolence, looking for parallel processes and using observations from different sorts of cases to generate ideas for others.

The theory-building techniques used in developing the concepts of moral jiu-jitsu, political jiu-jitsu, and backfire can be used as guides for stimulating yet further development of civil resistance theory. The first implication is that observations are crucial. Researchers can learn from what happens in a wide variety of situations. Civil resistance researchers have mainly focused on successful cases. To broaden the theory, more attention could be given to unsuccessful struggles, to ambiguous cases (where it is not obvious that a struggle is occurring), and to less visible cases requiring much more digging to obtain information.

Detailed accounts of struggles are valuable, but for theory development the key is noticing something that is crucial in the process. So observations need to be guided in some way.

One way is by looking at prior theory. Gregg didn't have this advantage, as nonviolence theory was rudimentary at the time. Gregg linked observations with his understandings of relevant psychology. Drawing on theory remains the way forward. This can either be prior nonviolence theory or theory from various fields or entirely new theory. For future development of civil resistance theory, it is worth considering all sorts of frameworks and perspectives from different fields and seeing whether and how they apply. Examples include social movement theory (Schock 2005 and this volume) and communication theory (Martin and Varney 2003).

Another way to develop theory is to move toward microscale interactions or to far broader frameworks. The backfire model looks at tactics used by both sides, but it is possible to look much more closely at events. For example, discourse analysis could be used to examine the uses of words used in campaigns or in attempts at fraternization. Semiotics could be used to examine the symbolic dimension of struggles, in what might be called satyagraha semiotics.

Sharp developed a framework for nonviolent struggles from laying the groundwork to transforming power relationships. This is a much bigger picture. Yet it is possible to think yet more broadly, such as at international political dynamics, regime types, technological transformation, and large-scale patterns in education and communication. The challenge is to use insights from social dynamics to throw light on actual struggles.

Civil resistance theory has largely been about collective action, when groups of people act to bring about a change. However, few people in their daily lives engage in rallies, strikes, sit-ins, or fasts. Therefore, civil resistance theory so far does not have a strong link to daily activities. This could be developed, for example, by addressing ordinary conflicts in families and workplaces. The backfire model applies in some cases, such as in bullying and sexual harassment.

Forging a link between nonviolence at the collective level - rallies, strikes, and the like - and at the small-scale interpersonal level would accomplish two things. At the level of theory, it would stimulate fruitful interactions. At the level of practice, it would make possible a powerful learning process. If people could see the power of nonviolence in their daily lives, they would be much better able to understand the world using a nonviolence lens and become more effective contributors to wider social change.

Gregg and Sharp appear to have developed their ideas largely on their own. In contrast, in developing the backfire model I entered many collaborations (e.g., Jansen and Martin 2003; Martin and Wright 2003; McDonald, Graham, and Martin 2010). This had two advantages. First, it drew on expertise outside the small but growing band of nonviolence researchers: the experience and insights of people outside the field was brought to bear on nonviolence-related issues. The second advantage was that people in other fields were brought in greater contact with nonviolence theory and practice.

This survey of the trajectory from moral jiu-jitsu to backfire points to several areas for further theoretical development, including studying unsuccessful and ambiguous cases, looking at smaller-scale or larger-scale processes, forging links to people's personal experiences, and fostering research collaborations. As well as suggesting possible directions for future research and practice, this suggests how much is yet to be learned from the study of nonviolent action indicates the ongoing crucial importance of action for developing theory.

Acknowledgements Thanks to Peter Gibson, Les Kurtz, and Majken Sørensen for valuable comments on drafts.


Blackstock, Nelson. 1976. Cointelpro: The FBI's Secret War on Political Freedom. New York: Vintage.

Bondurant, Joan V. 1958. Conquest of Violence: The Gandhian Philosophy of Conflict. Princeton, N.J.: Princeton University Press.

Chomsky, Noam, and Edward S. Herman. 1979. The Political Economy of Human Rights. Volume 1: The Washington Connection and Third World Fascism. Montréal: Black Rose Books.

Cribb, Robert, ed. 1990. The Indonesian Killings of 1965-1966: Studies from Java and Bal?. Melbourne, Victoria: Centre for Southeast Asian Studies, Monash University.

Csapody, Tamás, and Thomas Weber. 2007. "Hungarian Nonviolent Resistance against Austria and Its Place in the History of Nonviolence." Peace and Change 32:499-519.

Dalton, Dennis. 1993. Mahatma Gandhi: Nonviolent Power in Action. New York: Columbia University Press.

Dunn, James. 2003. East Timor: A Rough Passage to Independence. 3rd ed. Sydney: Longueville Books.

Elkins, Caroline. 2005. Imperial Reckoning: The Untold Story of Britain's Gulag in Kenya. New York: Henry Holt.

Frankel, Philip. 2001. An Ordinary Atrocity: Sharpeville and Its Massacre. New Haven, Conn.: Yale University Press.

Gregg, Richard B. [1934] 1966. The Power of Nonviolence. 2nd ed. New York: Schocken Books.

Herd, Andrew. 2006. "Amplifying Outrage over Children Overboard." Social Alternatives 25:59-63.

Huxley, Steven Duncan. 1990. Constitutionalist Insurgency in Finland: Finnish "Passive Resistance" against Russification as a Case of Nonmilitary Struggle in the European Resistance Tradition. Helsinki: Finnish Historical Society.

Jansen, Sue Curry, and Brian Martin. 2003. "Making Censorship Backfire." Counterpoise 7:5-15.

Kohen, Arnold S. 1999. From the Place of the Dead: The Epic Struggles of Bishop Belo of East Timor. New York: St. Martin's Press.

Koon, Stacey C., with Robert Deitz. 1992. Presumed Guilty: The Tragedy of the Rodney King Affair. Washington, D.C.: Regnery.

Martin, Brian. 2004. "Iraq Attack Backfire." Economic and Political Weekly 39, no. 16 (April 17-23): 1577-83.

- - - . 2005. "The Beating of Rodney King: The Dynamics of Backfire." Critical Criminology 13:307-26.

- - - . 2007. Justice Ignited: The Dynamics of Backfire. Lanham, Md.: Rowman and Littlefield.

- - - . 2009. "Managing Outrage over Genocide: Case Study Rwanda." Global Change, Peace and Security 21:275-90.

- - - . 2010. "Flotilla Tactics: How an Israeli Attack Backfired." Truthout, 27 July.

- - - . 2012. Backfire Manual: Tactics against Injustice. Sparsnäs, Sweden: Irene Publishing.

Martin, Brian, and Wendy Varney. 2003. Nonviolence Speaks: Communicating against Repression. Cresskill, N.J.: Hampton Press.

Martin, Brian, and Steve Wright. 2003. "Countershock: Mobilizing Resistance to Electroshock Weapons." Medicine, Conflict and Survival 19:205-22.

McDonald, Paula, Tina Graham, and Brian Martin. 2010. "Outrage Management in Cases of Sexual Harassment as Revealed in Judicial Decisions." Psychology of Women Quarterly 34:165-80.

McMillan, Andrew. 1992. Death in Dili. Sydney: Hodder and Stoughton.

Nevins, Joseph. 2005. A Not-So-Distant Horror: Mass Violence in East Timor. Ithaca, N.Y.: Cornell University Press.

Owens, Tom, with Rod Browning. 1994. Lying Eyes: The Truth behind the Corruption and Brutality of the LAPD and the Beating of Rodney King. New York: Thunder's Mouth Press.

Riddick, Brendan. 2012. "The Bombing of Afghanistan: The Convergence of Media and Political Power to Reduce Outrage." Revista de Paz y Conflictos 5:6-19.

Schock, Kurt. 2005. Unarmed Insurrections: People Power Movements in Nondemocracies. Minneapolis: University of Minnesota Press.

Sharp, Gene. 1973. The Politics of Nonviolent Action. Boston: Porter Sargent.

Shridharani, Krishnalal. 1939. War without Violence: A Study of Gandhi's Method and Its Accomplishments. London: Victor Gollancz.

Smithey, Lee, and Lester R. Kurtz. 1999. "'We Have Bare Hands': Nonviolent Social Movements in the Soviet Bloc." In Nonviolent Social Movements: A Geographical Perspective, ed. Stephen Zunes, Lester R. Kurtz, and Sarah Beth Asher, 96-124. Oxford: Blackwell.

Weber, Thomas. 1993. "'The Marchers Simply Walked Forward until Struck Down': Nonviolent Suffering and Conversion." Peace and Change 18:267-89.

- - - . 1997. On the Salt March: The Historiography of Gandhi's March to Dandi. New Delhi: HarperCollins.

Yecies, Brian. 2008. "Planet Hallyuwood's Political Vulnerabilities: Censuring the Expression of Satire in The President's Last Bang (2005)." International Review of Korean Studies 5:37-64.