I gave a talk at TEDxBoulder on September 21st. It was a great event, and I shared the stage with over a dozen terrific speakers and a number of talented musicians. We shared our ideas with a sold-out audience of about 2,200 people, and I’ve never been more nervous giving a talk!
Here’s the video:
I anticipated that people might have questions about some of the claims I make in this talk, as well as some of the specific references I make. As you can imagine, when you have 12 minutes to tell a story, present some counterintuitive information, and try to make it engaging, there’s no time to fully reference your points. So I decided to post the script and expand it with a variety of links to sources, references, and resources for those interested in pursuing the topic further. There may be a few deviations/word changes here and there because, well, I didn’t deliver the script verbatim. But you’ll get the gist.
Feel free to leave your remaining questions in the comments section. I will collect them and answer them in a follow-up post.
I’d like to ask you to imagine that you live in a very repressive country—there are elections but they are fake. The leader wins 100% of the vote each time. Security forces beat up opposition leaders with impunity, and they harass everyone else. This is a country where being in this room right now would get you on a list. Now let’s say you’ve had enough, and so have many other people that you talk to in low whispers. I’m not talking about the Hunger Games although that would be awesome. Unfortunately I’m talking about real world conditions that many people find themselves in right now.
Assuming you’ve decided to act, what would be the best way for you to challenge the system and create major change?
My own answer to this question has changed over the past five years. In 2006 I was a PhD student in political science here at CU-Boulder, and I was finishing my dissertation about how and why people use violence to seek political goals. As for the scenario I just described? Well, back then I bought into the idea that “power flows from the barrel of a gun.” I would have said that although it was tragic, it was logical in such cases for people to use violence to bring about change.
But that June, I was invited to an academic workshop put on by the International Center on Nonviolent Conflict. They were giving a week-long primer on nonviolent resistance to encourage people like me to teach about it in our courses. Now, my view of all this stuff was that it was well-intentioned, but dangerously naïve. The readings they sent me[1] argued that the best way for people to achieve political change was through nonviolent or civil resistance. The authors described civil resistance as an active form of conflict where unarmed civilians used tactics like protests, demonstrations, strikes, boycotts, and many other forms of mass noncooperation to confront oppression. They brought up cases like Serbia, where a nonviolent revolution toppled Slobodan Milosevic—the butcher of the Balkans—in October of 2000, or the Philippines where the People Power movement ousted Ferdinand Marcos in 1986.
At the workshop, I said things like, “Well, for every successful case you guys mention, I can think of a failed case like Tiananmen Square. I can also think of plenty of cases where violence worked pretty well, like the Algerian, French, and Russian revolutions. Maybe nonviolent resistance works if you’re seeking labor rights, gender rights, or environmental reform, but it generally can’t work if you’re trying to overthrow a dictator or become a new country. Serbia and the Philippines–they were probably exceptions. And there’s no way nonviolent resistance can work against a ruthless opponent.”
By the end of the week, as you can imagine, I wasn’t too popular.
My soon-to-be co-author—Maria Stephan–came up to me and said something like, “If you’re right, prove it. Are you curious enough to study these questions empirically?”
Believe it or not, no one had systematically done this before.[2] Although I was still skeptical, I was curious. If they were right and I was wrong, I figured somebody had better find out. So for the next two years, I collected data on all major nonviolent and violent campaigns for the overthrow of a government or territorial liberation since 1900. The data cover the entire world and include every known campaign that consists of at least a thousand observed participants, which constitutes hundreds of cases.[3]
Then I analyzed the data, and the results blew me away. From 1900 to 2006, nonviolent campaigns worldwide were twice as likely to succeed outright as violent insurgencies. And there’s more. This trend has been increasing over time—in the last fifty years civil resistance has become increasingly frequent and effective, whereas violent insurgencies have become increasingly rare and unsuccessful. This is true even in extremely repressive, authoritarian conditions where we might expect nonviolent resistance to fail.[4]
So why is civil resistance so much more effective than armed struggle? The answer lies in people power itself.
Researchers used to say that no government could survive if five percent of its population mobilized against it. But our data reveal that the threshold is probably lower. In fact, no campaigns failed once they’d achieved the active and sustained participation of just 3.5% of the population—and lots of them succeeded with far less than that [5]. Now, 3.5% is nothing to sneeze at. In the U.S. today, this means almost 11 million people.
But get this: Every single campaign that did surpass that 3.5% threshold was a nonviolent one. In fact, campaigns that relied solely on nonviolent methods were on average four times larger than the average violent campaign. And they were often much more representative in terms of gender, age, race, political party, class, and urban-rural distinctions.
Civil resistance allows people of all different levels of physical ability to participate—including the elderly, people with disabilities, women, children, and virtually anyone else who wants to. If you think about it, everyone is born with an equal physical ability to resist nonviolently. Anyone who has kids knows how hard it is to pick up a child who simply doesn’t want to move, or to feed a child who simply doesn’t want to eat.
But for lots of people, violent resistance is much more physically demanding. You have to train to be good at it. When I was in college, I took military science classes because I wanted to go through the ROTC program and become an army officer. I liked the rappelling, the uniforms, map-reading, and shooting at the range. But I wasn’t stoked about getting up in the wee hours of the morning to run until I vomited. I quit–and chose the far less strenuous career of professor.
Not everyone wants to take the same chances in life, and many people won’t turn up unless they expect safety in numbers. The visibility of many civil resistance tactics, like protests, helps to draw these risk-averse people into the fray. Put yourself back in that repressive country for a minute. Say your neighbor comes to you and says, “We’re going to have a demonstration in the main square down the street at 8pm tonight. I hope you can make it.” Now, I don’t know about you all, but I’m not the person who is going to show up at 7:55 to see what’s up. I’m going to wait until about 8:30 or so, check out my window, and see what’s going on. If I see only 6 people assembling in the square, I’m probably going to sit this one out. But if I see 6,000 and more coming down the alleyway, I might join them.
The point here is that nonviolent campaigns can solicit more diverse and active participation from ambivalent people. And once those people get involved, it’s almost guaranteed that the movement will then have some links to security forces, the state media, business or educational elites, religious authorities, and civilian bureaucrats who start to question their allegiances. No regime loyalists in any country live entirely isolated from the population itself. They have friends, they have family, and they have existing relationships that they have to live with in the long term, regardless of whether the leader stays or goes. In the Serbian case, once it became clear that hundreds of thousands of Serbs were descending on Belgrade to demand that Milosevic leave office, policemen ignored the order to shoot on demonstrators. When asked why he did so, one of them said: “I knew my kids were in the crowd.”
I’ll bet some of you are thinking, “Is she insane? I watch the news, and I see protestors getting shot at in the streets all the time!” Sometimes crackdowns do happen. But even in these cases, nonviolent campaigns outperformed violent ones by two-to-one. When security forces beat up, arrest, or even shoot unarmed activists, there is, indeed, safety in numbers. Large and well-coordinated campaigns can switch from concentrated methods (like protests) to dispersed methods, where people stay away from places they were expected to go. They do strikes, they do stay-at-home demonstrations, they bang on pots and pans, they shut off the electricity at a coordinated time of day — these tactics are much less risky. They’re very hard or at least very costly to suppress, while the movement stays just as disruptive.
What happens in these countries once the dust settles? It turns out, the way you resist matters in the long run too. Most strikingly, nonviolent campaigns were far more likely to usher in democratic institutions than violent insurgencies. And countries where people waged nonviolent struggle were 15% less likely to relapse into civil war.
The data are clear: When people rely on civil resistance, their size grows. And when large numbers of people withdraw their cooperation from an oppressive system, the odds are ever in their favor.
So. Many people in my field[6] had largely ignored the millions of people worldwide who were skillfully using civil resistance in favor of studying things that blow up. I had a few questions about the way I used to think. Why was it so easy and comfortable for me to believe that violence works? And why did I find it acceptable to simply assume that violence happens—almost automatically—because of circumstances, or by necessity—that it’s the only way out of some situations? In a society that celebrates battlefield heroes on national holidays, I guess it was natural to grow up believing that violence and courage are one and the same—and that true victories can’t come without bloodshed on both sides.
But the evidence I’ve presented here today suggests that for people serious about seeking change, there are realistic alternatives. Imagine now what our world would look like if we allowed ourselves to develop faith in them. What if our history courses emphasized the decade of mass civil disobedience that came before the Declaration of Independence, rather than the war that came after? What if Gandhi and King were the basis of the first chapter of our social studies textbooks, rather than an afterthought? What if every child left elementary school knowing more about the Suffragist movement than they did about the Battle of Bunker Hill? And what if it became common knowledge that when protests become too dangerous, there are many nonviolent techniques of dispersion that might keep participants safe and keep movements resilient?
So here we are in 2013 in Boulder, Colorado. Maybe some of you are thinking, “OK, I get that civil resistance is the best bet, but what can I do?”
Encourage your children to learn about the nonviolent legacies of the past two hundred years and explore the potential of people power. Tell your elected representatives to stop perpetuating the misguided view that violence pays by supporting the first groups in a civil uprising to take up arms. Although nonviolent campaigns can’t be exported or imported, it’s time for our officials to embrace a different way of thinking—that in the short and long term, civil resistance tends to leave behind societies in which people are able to live more freely and more peaceably together.
Now that we know what we know about the power of nonviolent conflict, I see it as our shared responsibility to spread the word so that future generations don’t fall for the myth that violence is their only way out.
Thank you.
[1] These readings included Gene Sharp’s From Dictatorship to Democracy; Kurt Schock’s Unarmed Insurrections; Zunes, Asher, and Kurtz’s Nonviolent Social Movements: A Geographical Perspective; Peter Ackerman and Jack Duvall’s A Force More Powerful; and Peter Ackerman and Christopher Kruegler’s Strategic Nonviolent Conflict, and a handful of article-length readings. Resources like this can be found here.
[2] As note 1 suggests, many people had theorized about civil resistance and developed case studies analyzing its effectiveness, but no one had systematically compared how well it worked compared with armed struggle over time and space.
[3] Many of these are discussed in detail at the Swarthmore Global Nonviolent Action Database.
[4] For answers to questions about how we counted the campaigns, coded success and failures, etc., see here, here, and here.
[5] This figure is based on the highest number of observed participants directly confronting the opponent during the campaign. It does therefore not represent an aggregate number of participants, but rather that the maximum number of people the campaign involved in peak events.
[6] There are exceptions. See note 1.
Great TED talk Erica, I really enjoyed it : )
I postulate that we as human animals are more likely to resort to violence in the most extreme cases (example: Operation Valkyrie and La Résistance Française in opposition to Nazism as opposed to a sit in at Hitler’s bunker, etc.) and that the inherently more oppressive, organized nature of regimes against which violence is used may be skewing your results as these regimes will always be harder to topple. I’m new to your research, so could you please direct me towards data relevant to this postulate?
Did you somehow control for the possibility that nonviolent uprisings are more successful merely because people only resort to violence when the state control system is the most oppressive (which would explain why they eventually fail), or some similar effect?
Fascinating talk nonetheless.
Coming from India, the land of Gandhi, the original apostle of “Ahimsa” , non violent protest, and knowing how we did turn out into a pretty strong democracy, it was very interesting to see your data and the conclusions drawn.Thank you.Hope more of my countrymen realize what we have and cherish it.
Reblogged this on My Blog.
What about ‘non-violent’ protest that is specifically designed to incite violent response from the oppressive regime and thus increasing the sympathy for the victim/protestors? Did this reality of many ‘non-violent protests’ figure into your data classification?
As we have seen with the sorry outcome of the color revolutions and more recently with the Arab Spring non-violent resistance has its drawbacks. One of its main drawbacks is its intolerance. Having 3.5% resist a leader whom everybody dislikes is one thing. But what about the case where half of the population believes he is a great guy – or at least better than the protesters? Syria is an example where we see the latter play out.
In India too one saw the drawbacks. The same “my way or the highway” attitude that brought independence also brought the most deadly partition of a country the world has ever seen.
Real non-violent resistance is capable of seeing the human in everybody and deal with him as such – even when he is a detested dictator.
The partition of India wasnt a result of the non violent agitataion. Why link the two together?
This talk was fascinating, but I think there is a selection bias problem in your argument. You write, “from 1900 to 2006, nonviolent campaigns worldwide were twice as likely to succeed outright as violent insurgencies,” and then assume a causal link. Populations don’t start protests at random; they consider the costs involved before doing so.
If we assume that leaders of a militarily weak dictatorship know (1) that they are unable to withstand violent popular uprising and (2) that if they respond violently, they will be punished by the population once it succeeds in taking power, then they have an incentive to relinquish power peacefully if the population begins protesting. The protesters, knowing all of this, understand that they have a low chance of being hurt while protesting and that in order to overthrow the regime, all they have to do to is protest peacefully, so as to make the regime understand that they are prepared, if need be, to rise violently.
They are more likely to protest in a nonviolent way and to be more successful. But power still flows from the barrel of a gun.
Philippe, you’re right that the initial results suggest that there may be selection effects involved. This was my primary concern as well. In my book with Maria Stephan, we devote an entire quantitative chapter (where we use multiple two-stage models identifying both the choice of NV/V resistance and the link between this choice and the outcome) and four cases studies to this possibility. We find that, while it’s true that people consider the costs before acting, the information environments in which they are operating are highly uncertain and that the selection process isn’t really influencing a majority of the cases. Activists do not know when they go into the streets whether the regime is going to crack down indiscriminately or whether the regime will ignore them. And some of them disagree about which method is most effective, resulting in some people within the campaign choosing nonviolent resistance and others using violent resistance under identical circumstances. Indeed, we found no structural factors (including violent repression) that systematically influenced whether people resorted to nonviolent or violent resistance. In other words, it is not the case the the probability of success influences the choice to use nonviolent resistance. I’ll write more on this in a follow-up post, but just wanted to quickly mention that we thought of this and that it isn’t driving the results as much as you’d expect.
Thank you for your reply. Testing whether structural factors influence the type of protest (violent/nonviolent) indeed corrects for selection bias. However I think the most important factor would be the military strength of the regime, to see if populations are not just acting according to an underlying balance of power. Looking forward to reading that follow-up post (and your book).
Gosh, I find this fascinating research. I watched your talk and will read your article later. I feel relieved to know that it only takes 3.5% to topple a regime. I enjoyed your examples of dispersion, and your suggestion to include nonviolent movements in our education. Thanks!
I agree with the premise, and would submit that the Tea Party movement is, so far, the most successful movement, in that it civilly resisted Barack Obama and his statist policies. Of course, that’s not viewed as a “successful civil disobedience” cause in the traditional sense, but then, voting itself can be construed as the result of the greatest act of civil disobedience in history: the American Revolution. While the conflict was armed, it would never have succeeded without correspondingly equal peaceful change.
One can apply Erica’s empirical research to US climate policy.
1. California has adopted the world’s most effective and comprehensive set of laws/regulations to protect the climate (AB32 – 2006 Global Warming Solns Act, Cap & Trade, vehicle GHG standard, low carbon fuel standard, Governor’s executive orders, renewables, solar roofs, efficiency, regional transport, recycle, reduce waste, water, agriculture, forests.) This represents an historic government accomplishment featuring startling cross-departmental state staff collaboration and visionary legislative leadership.
2. In contrast, the US congress is gridlocked for the foreseeable future. There is no ability to enact game-changing legislation/regulation. Obamacare is incremental (versus universal Medicare would be game-changing). Enacting effective climate policy requires a game-change akin to combining New Deal, WWI, and Civil Rights. As far as the UN climate effort, there was no progress at “COP19” and the US was blamed for foot-dragging.
3. Applying the 3.5% rule: If 3.5% (11 million Americans) participate in a vigorous citizen movement, then there is an 80% chance of achieving game-changing federal climate laws/regulations akin to those in California. And there is then a further argument that the end of US foot-dragging will immediately spur pro-climate policies in the rest of the world.
People tend to feel they can do nothing about the climate, so I have found that the message about 11-million Americans is received as very hopeful and motivating to “the movement.” Not to pick on any one individual, but Bill McKibben is a virtuous climate activist and professor who has somewhat blindly deployed a series of random tactics without predicting success or failure of those tactics based on a rudimentary, professorial literature review. There may be a more efficient narrative for how to achieve US climate laws/regulations, but Erica’s word provides one strategic path backed by empirical evidence.
In this instance, Robert Reich’s book Supercapitalism illuminates the gridlocked “opposition” as the US crony capitalism political/economic system. The staunchest members of that opposition are fossil fuel companies that could experience reduced profits as a result of pro-climate policies.
4. Books such as Berkeley Professor Andrew Guzman’s Overheated provide a mass-market explanation that climate is the biggest challenge humanity has ever faced. In Guzman’s optimistic 2C 21st century scenario, his chapter on African explains how climate change is a very dangerous 2020 threat multiplier to “precarious, subsistence peoples.” Hence there is a moral case for a pro-climate American citizen movement to overcome federal gridlock. This requires a similar level of movement power as a regime change would.
5. I find it frustrating that conferences such as The New Economy, DeGrowth, and Radical Emissions Reduction fail to invite Erica to share her insights on first world economic justice and climate protection movements. The critique of McKibben applies to these disappointing festivals of hand-wringing. Everyone can agree on what to do, but only Erica can tell you how to bring about those policies.
This Talk is timeless, but so needed again right now! I would love to spread this message – that we need action (and we do!) but real success will come if this is done in a non-violent manner.
More timely than ever, but a lot of the hyperlinks are broken! Could someone update?
Also would love thoughts on how the “safety in numbers” principle applies (or doesn’t) when the crowd is primarily people of color.
It is always dangerous for marginalized communities to dissent. This research discussed in this talk is based on cross-national studies where the POC frame does not necessarily apply; hence the immediate application of this research to your question is not straightforward. However, in an analogous situation, it is true that secessionist campaigns (where ethnic minorities are often the claimants) are less likely to succeed than those targeting removal of the incumbent govt (often comprised of diverse social groups), likely because (1) ethnic minorities are unable to build the #s required, on their own, to elicit change; and (2) identity-based differences between themselves and the opponent security forces, who may be less likely to defect as a result. However, there are cases of successful self-determination using NV resistance when claimants were able to expand their active supporters to include those in the ethnic majority (e.g. East Timor). The Civil Rights Movement in the US may be another case of this. Conversely, in some cases (e.g. South Africa), people of color were the majority, but they were suppressed by a minority-white government. It would have been too dangerous there to appeal directly to white supremacist security forces. In that case the anti-apartheid movement was instead able to mobilize sufficient numbers to seriously disrupt the economic status quo in a way that made white South Africans appeal to the apartheid regime to reform. They did this by mobilizing black townships, generating connections with liberally-minded South African whites, and appealing for international assistance via divestments and sanctions.
In the US case, some people have done research on the way that race affects attributions of blame and perceptions of the legitimacy of protest policing. For example, Davenport, McDermott, and Armstrong have a working paper that shows that there is a strong racial dimension to whether observers sympathize more with police or protestors (http://politicalviolenceataglance.org/2014/08/17/seeing-contention-in-black-and-white-protest-and-protest-policing/). That said, Omar Wasow has written that white observers were much more likely to empathize with / join the Civil Rights Movement in the south as black protestors maintained nonviolent approaches, compared with later phases of the Civil Rights Movement when protests were perceived as more violent (http://www.omarwasow.com/Protests_on_Voting.pdf).
Regardless, more anecdotally, I’ve heard from many people of color in the U.S. that they feel safer in large groups that isolated / alone – even if that means gatherings related to dissent. Hence I believe the logic often holds despite the vulnerability of various marginalized communities.
Done!
If I am being honest with myself and putting it out there, I will divulge that violent methods simply feel more satisfying (despite rationally understanding that non-violent methods are more efficient). My wife says it is a guy thing. Perhaps, but assuming I am not an isolated case (e.g. football/fans), then one shortcoming of a NV approach is that it is not engaging enough (physically, psychologically) for a good portion of the population. That’s not rational, but human.
Another potential shortcoming of NV approaches pertains to its reliance on large numbers. If you have a controlling regime that is strategically using shotgun misdirection, then the resistors are spread too thin across multiple crises to be effective, or at least it might take much longer than three years to achieve critical mass. In that case perhaps focused tactical push-backs etc., are more efficient.
If I am being honest with myself and putting it out there, I will divulge that violent methods simply feel more satisfying (despite rationally understanding that non-violent methods are more efficient). My wife says it is a guy thing. Perhaps, but assuming I am not an isolated case (e.g. football/fans), then one shortcoming of a NV approach is that it is not engaging enough (physically, psychologically) for a good portion of the population. That’s not rational, but human.
Another potential shortcoming of NV approaches pertains to its reliance on large numbers. If you have a controlling regime that is strategically using shotgun misdirection, then the resistors are spread too thin across multiple crises to be effective, or at least it appears like it would take much longer than three years to achieve critical mass. In that case perhaps focused tactical push-backs are more efficient.
Sorry if this appears as a reply to your comment, joseph, but i wanted to add a post and couldn’t see how – maybe posting comments has been shut off now?
Anyway, my comment is about the 3.5% figure. I’ve written to both Maria and Erica, but i guess they both get a pile of email in which mine has gotten lost. So maybe Erica will reply here – though anyone else who knows the answer is also welcome!
I rented the book for a month. I was curious about the 3.5% figure, but couldn’t find it using a search. I skimmed through the book looking for something similar, but couldn’t locate anything.
The closest i’ve come so far is Figure 2.1 on page 78. This certainly shows that a rising participation rate increases the chance of success, but doesn’t even get to 3.5% (and %s below that certainly don’t achieve 100% success).
So i’d love to find out:
1) where this figure comes from – what pages in the book provide the information?
2) same thing for this: “But get this: Every single campaign that did surpass that 3.5% threshold was a nonviolent one. In fact, campaigns that relied solely on nonviolent methods were on average four times larger than the average violent campaign.”
3) does this mean “reached” rather than surpassed, or do it actually mean that campaigns had to get *over* 3.5% of involvement to be nonviolent?
4) Wikipedia says that no violent campaign reached the 3.5% threshold. Is that correct, and again, which pages confirm that?
Rico, here are responses to your questions:
(1) It’s not in the book. I calculated the 3.5% rule based on the data in the book for a workshop for activists a few years after the book was published. I cited this figure in my TED Talk, which is the only published source for the claim.
(2) See #1.
(3) Reached or surpassed.
(4) See #1. It’s in the data, not in the book.
Thank you very much, Erica.