he power of nonviolence: Confirming and explaining the success of nonviolent (rather than violent) political movements S. Nima Orazani & Bernhard Leidner

EJSP RESEARCH ARTICLE The power of nonviolence: Confirming and explaining the success of nonviolent (rather than violent) political movements S. Nima Orazani & Bernhard Leidner University of Massachusetts Amherst, Amherst, Massachusetts, USA Correspondence Seyed Nima Orazani, Department of Psychological and Brain Sciences, University of Massachusetts Amherst, 135 Hicks Way, Amherst, Massachusetts 01003, USA. E-mail: orazani@umass.edu Received: 13 February 2018 Accepted: 21 July 2018 https://doi.org/10.1002/ejsp.2526 Conflict of Interest Statement The authors declare that there are no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. All studies conducted in the current article have been approved by Institutional Review Board of University of Massachusetts Amherst. Abstract Drawing on literatures on social movements and collective action, mentalization, and morality, four studies investigated whether a social movement’s use of nonviolence can increase people’s willingness to support and join the movement. In a correlational study with a nested design, across 23 movements perceived use of nonviolence predicted participants’ willingness to support and join the movement (Study 1, n = 203). This effect was also found experimentally, with Americans supporting nonviolent movements more than violent ones, in hypothetical and real foreign countries (Study 2 and 3, ns = 606 and 373). Study 4 (n = 247) replicated the effects in participants’ own country. The effects were transmitted by attribution of mental states to nonviolent movements and subsequent greater perceived morality (Study 2–4). This research demonstrates that nonviolence can benefit social movements in terms of support and mobilization potential, and that these benefits are rooted in perceptions of mental capacity, humanness, and morality. Keywords: collective action, nonviolence, mentalization, morality, underdog effect Data Archiving and Accessibility Statement Data files for all studies are publicly accessible on data.world and can be found through the following link: https://data. world/orazani/the-power-of-nonviolence. Since 2011 several Arab countries have been going through unprecedented social change. In Tunisia, Egypt, Libya, and Syria, among other countries, citizens protested and demonstrated for fundamental changes to their political system. These protests turned into popular social movements and in a short period of time spread across many Arab countries, now known as the Arab Spring. Yet, not all movements succeeded in effecting the social change they strived for. While the movements in Tunisia and Egypt did induce social change, the movements in Libya and Syria fell short. Understanding such differences in movements’ effectiveness is tantamount to developing a comprehensive theory of collective action, social movements, and change (see Louis, 2009), as political struggles and resistance always incur costs for all sides of the struggle as well as society at large (Dodd, 2011; Saleh & Werr, 2011). Many observers, pundits, and scholars in political science, sociology, and psychology have attributed the difference in success across social movements to a 688 difference in strategy (i.e., use of violent vs. nonviolent strategies). These strategies are believed to have differential effects on people’s willingness to support the movements (Schock, 2013). While all aforementioned movements started out with nonviolent strategies, the movements in Libya and Syria eventually turned to violent strategies, whereas the movements in Tunisia and Egypt remained nonviolent. This distinction between violence and nonviolence is akin to distinctions between normative and non-normative collective action (Tausch et al., 2011; Zaal, Van Laar, St ahl, Ellemers, & Derks, 2011), radicalism and activism (Moskalenko & McCauley, 2009), and moderate versus militant political action (Barnes & Kaase, 1979; see also Thomas & Louis, 2014). Yet, people attribute real life differences in movements’ success to these dichotomies often based on their personal opinions and beliefs rather than on conclusive empirical evidence for cause–effect relationships. Another reason for the need for conclusive empirical evidence of cause-and-effect is that no strategy, be it European Journal of Social Psychology 49 (2019) 688–704 ª 2018 John Wiley & Sons, Ltd. S. N. Orazani & B. Leidner nonviolent or violent, may always be successful. Instead, a strategy’s success may depend on specific circumstances (see e.g., Stieren, 2001). While sociological and retrospective analyses of social movements indicate that nonviolence could be relatively more successful than violence, under certain circumstances violence may be deemed as a legitimate strategy even in cultures with strong norms against violence. Research has shown for example that it is more likely for those who hold moral convictions to use violence and perceive it as a legitimate strategy (Skitka, 2010). In the context of social movements, for example, social justice in its broadest meaning is a value that people, especially those who sit on the left side of the political spectrum, hold dear. But the dilemma poses itself when people think about how far they should go in terms of adopting violence when it comes to waging social change considering that pacifism, like social justice, is a virtue, and it is a widely accepted norm in our society. Moreover, political ideology in general and right-wing authoritarianism in particular are associated with perceiving the use of force as justifiable (Benjamin, 2006; Gerber & Jackson, 2016). Group membership too can distort ingroups’ attitudes toward violence such that they perceive it as necessary or downplay its otherwise perceived morality (Leidner & Castano, 2012). It is therefore important to clarify that in the present contribution we were interested in comparing the likelihood for success of nonviolent and violent movements in the context of political struggles against state authorities. We consider this context a particularly rigorous test for the hypothesis that nonviolence can be more successful than violence, because in this context people may tend to be especially forgiving for “underdog” movements using violence against repressive and/or violent state authorities. Past studies showed that even though nonviolent movements usually face violent repression1 —any attempts by groups, individuals, or the state to maximize the costs of participating in collective action (Ritzer, 2007)—they are significantly more successful than violent movements. Literature on civil resistance has echoed the importance of nonviolence and the cost of violence by introducing concepts such as “moral jiu-jitsu” (i.e., losing moral balance due to violence; Gregg, 1934) and backfire (i.e., violent repression of a movement that feeds back into further (non)violent resistance; Martin, 1998; Sharp, 1973). However, this scholarship relies almost exclusively on archival data that is retrospective in nature (Chenoweth & Stephan, 2011; Stephan & Chenoweth, 2008). Consequently, the findings are subject to many possible confounding factors. At the same time, there is a widespread assumption, especially among political scientists as well as laypeople, that violence is more effective than nonviolence in 1 Throughout the paper we consider corruption and repression as two separate yet related constructs. The power of nonviolence creating social change (Stephan & Chenoweth, 2008). And indeed, research on intergroup conflict has demonstrated that people tend to see the use of violence in the context of politics and foreign relations as more tolerable than in other contexts, and even at times advocate for it as a legitimate and moral strategy (Giner-Sorolla, Leidner, & Castano, 2011; Leidner & Castano, 2012; Skitka, 2010; Skitka & Mullen, 2002). These past findings from research on intergroup conflict and the over-reliance of research on movement strategies on correlational data make it hard to pinpoint the cause of success or failure of movements. Only few papers provide causal evidence for the claim that nonviolence increases a movement’s success. Two studies showed that adopting violent strategies decreases popular support for the movement because of a perceived violation of broader ingroup norms (Becker, Tausch, Spears, & Christ, 2011) or reduced identification with the movement (Feinberg, Willer, & Kovacheff, 2017). But they did not answer the question of whether the use of nonviolence can increase popular support. While there is emerging experimental evidence for Americans’ increasing support for Palestinians in response to their nonviolent (as opposed to violent) resistance (Bruneau, Lane, & Saleem, 2017), other research has shown such an effect to be limited to situations where nonviolent movements oppose a system that is not seen as corrupt; when the system is seen as corrupt, however, the effect disappeared and violent movements received the same level of support (Thomas & Louis, 2014). Thus, it is still an open question if movements can garner more support by using nonviolent tactics and, if so, why and through what mechanisms. To answer these questions, we conducted four studies testing whether the use of nonviolence indeed causes an increase in the likelihood of success of a movement that faces an oppressive regime. Specifically, our studies examined to what extent people who belong to third-party groups support a movement depending on whether the movement uses violent or nonviolent strategies, and whether this support is driven by people’s perceptions of the movement’s members as moral agents sensitive to experiencing pain and suffering when harmed by opposing groups (i.e., by people’s attribution of mental capacity to, and perceived morality of, the movement). In doing so, we contribute to the emerging social psychological literature on nonviolent movements in particular, and the literatures on collective action and social movements in social/political psychology and sociology/political science in general. Predictors of a Movement’s Success While there is considerable debate over which strategy —violent or nonviolent—is more effective, there is consensus that public opinion is crucial for movements European Journal of Social Psychology 49 (2019) 688–704 ª 2018 John Wiley & Sons, Ltd. 689 The power of nonviolence S. N. Orazani & B. Leidner to succeed. Bringing public opinion on one’s side increases domestic and international support and hence gives movements more room to maneuver in their demands for social change (e.g., Burstein, 2003; Burstein & Linton, 2002; Louis, 2009). Therefore, social movements seek their success through convincing the domestic and/or international public to support them. Indeed, Stephan and Chenoweth (2008) found that support was the main predictor of movements’ success (defined as reaching their stated goals). This support can take many forms: transformation of a neutral third party into a sympathetic one (Oegema & Klandermans, 1994; Simon & Klandermans, 2001), transformation of a sympathetic person into an engaging member of the movement (Oegema & Klandermans, 1994; Van Stekelenburg & Klandermans, 2013), encouragement of regime supporters to stand by or even defect (Nepstad, 2013), encouragement of third parties that support the regime to withdraw their support (Stewart et al., 2015), or encouragement of third parties to help the cause of the movement (Stewart et al., 2015). Faced with an oppressive regime, third-party support to help a movement’s cause has been deemed especially important since movements that received external support were found to be more than three times as likely to succeed (Stephan & Chenoweth, 2008). During the Arab Spring, for example, many international parties engaged in sympathetic collective action in solidarity with the movements, particularly the ones in Tunisia and Egypt (Stewart et al., 2015; Strenger, 2011). Besides our question whether nonviolent (rather than violent) strategies are critical in eliciting this support, a closely related and equally important question is why that would be. Here, drawing on research on the theory of mind (mentalization), morality, collective action, and the “underdog effect,” we suggest that in the context of political struggles, the use of nonviolence increases third parties’ support since they attribute greater mental capacity to members of a nonviolent rather than violent movement, and therefore perceive members of a nonviolent rather than violent movement as more moral. Predictors of Third-Party Support: Mentalization and Morality The best humanoid robots are still robots, not human. Being considered human by others requires mental states such as desires, motivations, intentions, thoughts, and emotions (Haslam, Bain, Douge, Lee, & Bastian, 2005); it requires mind (Waytz, Gray, Epley, & Wegner, 2010). While mental states are thus necessary for a person to be considered human, they are not necessarily sufficient. Other people, too, should acknowledge that the person has these qualities—or, in other words, “a mind.” Otherwise, the person’s humanity would be insulated from the outside world (Waytz et al., 2010). More importantly, if others do not attribute mental states to the person, they will 690 likely treat the person as nonhuman, regardless of what the person thinks of him- or herself. Mentalization refers to the acknowledgment that other people have mental states, to the attribution of mental capacity to others. It also refers to subsequent attempts to predict others’ mental states in order to regulate our behaviors toward them accordingly (Frith & Frith, 2006). The attribution of mental states has thus crucial implications for human morality. Indeed, the more people attribute mental states to an entity, the more they perceive it as moral (Gray, Gray, & Wegner, 2007; Gray, Knobe, Sheskin, Bloom, & Barrett, 2011; Gray & Wegner, 2009; Gray, Young, & Waytz, 2012; Waytz et al., 2010) and the more moral sensitivity they attribute to it (for a review see Haslam, 2006). The reason is that when an individual is capable of abstract thinking (i.e., has cognitive mental capacities), it means that (s)he can distinguish between what is morally good or bad. This capacity is fundamental to be seen as a moral person, and to be treated as such by others in return (Haslam, 2006). Furthermore, attributing emotional states to a person makes him/her morally sensitive; either as a moral agent capable of engaging in moral behavior and refraining from immoral behavior, or as a “recipient” capable of enjoying and appreciating others’ moral behavior, or suffering others’ immoral behavior. In this way, mentalization and being perceived as a moral being are closely intertwined. The greater our attribution of mental states to others (i.e., mentalization), the more we see them as moral beings and the more likely we are to grant them status as human beings (i.e., humanize them). Conversely, denying others’ mental states excludes them from moral consideration. Leyens et al. (2000, 2003) showed that one way in which people dehumanize the outgroup is to deny that its members experience specific emotions that are typically believed to be unique to human beings. In this case, neither are otherwise immoral acts considered as such when visited upon dehumanized/non-mentalized others, nor are such people seen as moral agents (for a review see Li, Leidner, & Castano, 2014). It is easier and more justifiable to inflict pain upon others when the perpetrator believes that the victim is not capable of experiencing emotions such as pain in the first place. In this sense, there is a proximity among (de-)humanization, mentalization, and perceived morality. To our knowledge, literature on collective action has yet to explore and fully understand the potential benefits of nonviolence and their foundation in the attribution of mental states and morality. Applying this literature to the context of social movements, we hypothesized that adopting nonviolent strategies increases the extent to which thirdparty observers attribute mental states to nonviolent movements because nonviolent movements commonly suffer more violence (often at the hands of government) than they perpetrate—especially relative to violent movements. Consequently, people should European Journal of Social Psychology 49 (2019) 688–704 ª 2018 John Wiley & Sons, Ltd. S. N. Orazani & B. Leidner The power of nonviolence Figure 1: Hypothesized model of the effect of nonviolent (vs. violent) movement strategy on support for the movement through attribution of mental states to and perceived morality of the movement’s members perceive nonviolent movements as more moral and, ultimately, be more supportive of and more willing to join them (see Figure 1). This hypothesis is also in line with the research on moral typecasting and the “underdog effect.” According to the “underdog effect” victimization by a high-power entity such as a government leads people to perceive the victim as an underdog, which people generally see as more moral and deserving of support (Kim et al., 2008). Further, research on moral typecasting showed that people tend to perceive moral situations as dyads, where one side of a given conflict will be perceived as the perpetrator and consequently immoral, and the other side as the victim and consequently moral (Gray & Wegner, 2009). Overview of the Studies Four studies examined the effect of movements’ strategy on people’s willingness to support and join the movement. Using a nested design and 23 social movements in the United States, Study 1 tested whether perceived violence of a movement’s strategy can predict likability, support, and willingness to join the movement beyond the perceived morality of the movement’s goal. The next three experimental studies tested whether adopting nonviolent tactics leads people to attribute greater mental states to the movement and consequently perceive it as more moral, and if this greater perception of morality leads them to be more willing to support and join the movement. Study 2 tested this hypothesis among Americans learning about a hypothetical movement in a hypothetical foreign country. To test whether the findings of this study were specific to movements fighting for democracy or would generalize to movements with other goals, we manipulated not only the movement’s means (i.e., strategy: nonviolent vs. violent vs. unknown) but also its ends (i.e., goal: democracy vs. authoritarianism). This study also ruled out the underdog effect (i.e., perceived powerlessness of the movement) as an alternative to mentalization in explaining the effect of nonviolence. Study 3 conceptually replicated the previous study in the context of a real country (Myanmar) and a real movement. In order to extend our findings of Study 2 and 3 to people within the country of the movement, Study 4 tested our model among Americans in the context of the Black Lives Matter movement. Last but not least, we combined those parts of our data from all experimental studies (Studies 2–4) that could be subjected to a cumulative meta-analysis, confirming the results that emerged from the three individual experiments (see Appendix S1). Study 12 In this study we included real movements in the United States with (im)moral goals and (non)violent strategies, ranging from the U.S. civil rights movement to the Ku Klux Klan (for a complete list of all movements see Appendix 1). In order to have a more representative sample of social movements we used both widely known and less-known movements (e.g., the Fat Acceptance movement). Familiarity with the movements varied from 3.60 to 6.90 in a scale of 1, Not at All Familiar, to 9, Completely Familiar (see Table 1). This variety of movements maximized the ecological validity and generalizability of our results. Further, we aimed to test whether perceived strategy of a movement as (non)violent predicts support for the movement above and beyond the perceived morality of the movement’s goals. Method Participants. We recruited 203 participants via Amazon Mechanical Turk (MTurk). After excluding four participants who were not born in the United States, and one participant who used significantly more time to complete the survey (univariate outlier analysis: 3 SD above the mean; Tabachnick & Fidell, 2007), 198 participants were retained for data analysis (120 female, age M = 39.81, SD = 14.55, range = 18–78). Procedure. We asked participants to rate 23 social movements that were or are active in the United States (e.g., civil rights movement, Tea Party) on 2 All data and materials for all studies are publicly accessible on data. world and can be found through the following link: https://data. world/orazani/the-power-of-nonviolence/workspace/dataset. European Journal of Social Psychology 49 (2019) 688–704 ª 2018 John Wiley & Sons, Ltd. 691 The power of nonviolence S. N. Orazani & B. Leidner Table 1. Participants’ perception of morality and use of violence regarding the 23 social movements in the United States Perceived morality of the movement 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. Movement M (n) SD M (n) SD Familiarity with the movement M (SD) American Independent Movement Anti-Tax Movement Modern Anti-Slavery Movement States’ Rights Movement Women’s Rights Movement Labor Movement U.S. Civil Rights Movement Anti-War Movement Environmental Movement Occupy Wall Street LGBT Movement Animal Rights Movement Anti-Abortion Movement Tea-Party Movement Gun Control Movement Anti-Gun Control Movement Immigrant Rights Movement Economic Justice Movement Anti-Whaling Movement Pro-Choice Movement Neo-Nazi Movement Ku Klux Klan Anti-Gay Movement 6.06 (136) 5.22 (158) 7.63 (143) 5.54 (161) 7.45 (193) 6.64 (174) 7.77 (190) 7.11 (177) 7.03 (182) 5.87 (185) 6.50 (184) 7.31 (193) 5.42 (194) 4.04 (192) 5.85 (194) 4.29 (191) 6.44 (187) 6.21 (159) 6.81 (168) 6.33 (194) 1.91 (174) 1.74 (195) 2.91 (183) 1.91 2.29 1.74 2.28 2.04 2.16 1.71 2.05 2.12 2.50 2.76 2.01 2.80 2.48 2.58 2.59 2.32 2.22 2.28 2.68 1.95 1.82 2.61 5.32 (136) 6.33 (158) 6.01 (143) 6.36 (161) 6.49 (193) 5.47 (174) 4.80 (190) 4.82 (177) 5.22 (182) 5.06 (185) 6.37 (184) 5.03 (192) 3.54 (194) 5.45 (192) 6.21 (194) 5.39 (191) 6.22 (187) 6.44 (159) 4.79 (168) 6.06 (194) 2.56 (174) 2.07 (195) 3.79 (183) 2.56 2.30 2.41 2.29 2.34 2.44 2.56 2.53 2.50 2.71 2.35 2.56 2.58 2.63 2.48 2.69 2.31 2.15 2.47 2.69 2.22 2.03 2.75 3.60 (2.73) 4.03 (2.62) 3.89 (2.73) 4.25 (2.57) 6.73 (1.95) 5.07 (2.53) 6.61 (2.20) 5.42 (2.55) 6.11 (2.32) 6.16 (2.40) 6.22 (2.49) 6.50 (2.03) 6.56 (1.90) 5.90 (2.44) 6.50 (2.00) 6.22 (2.14) 5.70 (2.34) 4.01 (2.52) 4.67 (2.61) 6.90 (1.92) 4.29 (2.46) 6.29 (2.07) 5.26 (2.59) different measures using 1–9 visual analog scales. For each question, participants saw a list of 23 movements and were asked to indicate their answer to the question for each movement. Unless noted otherwise, the scale endpoints were labeled Not at All and Very Much. Materials. Perceived morality of the movement. Participants indicated to what extent they perceived each movement’s goals as moral (“For each of the following movements, to what extent do you think the goals of the movement are moral?”). Perceived use of nonviolence. Participants indicated to what extent each movement used strategic violence (“For each of the following movements, to what extent do you think it has used extreme actions (e.g., throwing stones at demonstrations, doing property damage, bodily harm, etc.) to achieve its goals?”). This item was reverse scored in order to reflect nonviolence, or lack of violence. Perceived likability of the movement. Participants indicated to what extent they liked each movement (“For each of the following movements, how much do you like the movement?”). Willingness to support the movement. For each movement, participants rated their support for it (“How much active support would you give to the movements below (e.g., sign petitions, donate money, participate in events, etc.)?”). 692 Perceived nonviolence of the movement Willingness to join the movement. Participants indicated to what extent they were willing to join each movement (“For each of the following movements, if you could, how much would you like to join the movement (i.e., become a member)?”). Attitudes toward repression of the movement. Participants indicated to what extent they would understand if a government were to suppress the movement (“For each of the following movements, imagine that the movement exists right now in another country whose government feels challenged or threatened by the movement. How understandable would you find it if the government were to use repression or violence against the movement?”). Results Descriptive statistics of perceived morality and perceived nonviolence for all 23 movements are reported in Table 1. Since the data from this study design (i.e., participants’ opinions about different movements) was nested within participants, we transposed the data so that each row represented a participant’s opinions about a given movement on the various issues we had measured. Since each of the 198 participants gave his or her opinions on 23 movements, the transposed data set had 4,554 rows. In a mixed model with participants as random factor, we entered perceived nonviolence and perceived morality of the movement as predictor variables, and, respectively, likability of the movement, willingness to support and join the movement, and attitudes toward repression of the movement as European Journal of Social Psychology 49 (2019) 688–704 ª 2018 John Wiley & Sons, Ltd. S. N. Orazani & B. Leidner The power of nonviolence Table 2. Prediction of outcome variables by perceived nonviolence and perceived morality of movements’ goals simultaneously Perceived morality of movement Likability Support Willingness to join Attitudes toward repression F p 4605.49 2196.01 1727.36 383.97