On September 16, 2019, the Democracy, Conflict, and Governance (DCG) program at the Carnegie Endowment for International Peace and the Bridging Divides Initiative at Princeton University convened fifty scholars, practitioners, funders, and elected and government officials to discuss political violence in America in the context of what we know about political violence abroad.

Participants were chosen for their capacity to increase understanding of and/or influence the prevention of potential violence. We sought to include a broad range of perspectives on the problem and means of addressing it. Participants were not expected to agree. In fact, some may have seen others in the room as a potential problem. Coming from the field of international conflict, where armed groups previously at war must come together to negotiate peace, the organizers accept that it requires engaging divergent perspectives for a country to make progress on reducing political violence. To solve our problems, we must begin by acknowledging that all of our identities are complex and that we must work together to achieve durable solutions to preserve our democracy and deliver more equal rights and freedoms.

We are grateful for the support of Stand Together/CKI for underwriting the workshop. Ongoing program funding for the Carnegie Endowment’s DCG Program is supported by the Hewlett Foundation.

Download the full text


Rachel Kleinfeld, Senior Fellow, Carnegie Endowment for International Peace (CEIP)
Nealin Parker and Shannon Hiller, Bridging Divides Initiative, Princeton University
Ashley Quarcoo, Democracy Specialist, USAID; Visiting Scholar, CEIP
Sadia Hameed, Founder and Executive Director, Thought Partnerships          

We see preventing targeted violence not only as an end in itself, but as part of a larger set of needed interventions to improve U.S. democracy. The United States has much work to do to rebuild healthy democratic institutions and communities while addressing our polarization and long-standing divisions over inequity and injustice. These efforts take time. Violence truncates the time and space needed to address challenges while deepening polarization.

Political violence can’t be predicted perfectly, but international conflict prevention practitioners have determined a series of risk factors and warning signs that can begin long before violence occurs. While targeted violence often appears spontaneous, it results from years of groundwork. Across countries, similar patterns of fear, division and threat catalyze violence, discrimination, and social segregation along identity lines. These markers can be monitored ahead of time.

We know that violence is more likely in countries where it has happened before. It thrives on polarization and begins with the dehumanization of opponents. Opportunistic politicians test the system, seeing how people react to violent language to determine the potential costs. Based on such warning signs, many who study political violence overseas believe the United States is at risk.

America’s history of political violence spans our civil war, decades of lynchings, and the assassinations of a president, presidential candidate, and national community leaders in the 1960s and 1970s. Thomas Carothers and Andrew O’Donohoe find in their new book that while most polarized countries are divided along just one or two dimensions, America has three fissures—ideological, ethnic, and religious—that overlap and augment one another. As for dehumanization, scholars Nathan Kalmoe and Lilliana Mason found last year that 20 percent of Republicans and 15 percent of Democrats believe that members of the other party “lack the traits to be considered fully human – they behave like animals.” The FBI found hate crimes spiked 17 percent in 2017, with a 37 percent increase in anti-Semitic incidents. Based on such risk factors, the 2018 Fragile States Index ranked America among the top five “most worsened countries” for political stability, alongside Yemen and Venezuela. Moreover, the U.S. represents just one front in what is now a global movement – violent white supremacists from New Zealand to Canada are interacting online, reading each others’ manifestos, and inspiring one another.

We have the tinder for political violence. Impeachment, the 2020 elections, or any number of possible local occurrences could provide the spark. Some among us worry about significant increases in violence targeted at visible minorities. Others are concerned because great harm to our democracy does not require mass bloodshed, but could involve more targeted tragedies. A pipe bomb such as the ones sent two years ago could, this time, kill an intended recipient, such as a Supreme Court justice, further politicizing the court especially throughout the nomination process. A shooting such as the one at a Congressional baseball game in 2017 could, this time, tilt the balance of the Senate. Even implied violence could derail our democracy in this precarious time. More than forty states allow open carry of guns into voting places. What if the 2020 election is close, and many Americans believed that armed intimidation tilted the balance?

Yet there is still time for America to avoid escalating violence. Our diversity of viewpoints is critical to this discussion, but in those differences, we are also unified. We believe that well-functioning democracy depends on non-violent politics. And we believe there is something we can do about the current trajectory of animosity in the United States. Our workshop’s goal was to facilitate greater understanding of the current state of targeted violence in America and of the broad range of interventions used in other countries to interrupt violent trajectories. Our hope is that we can amplify the work being done to address this problem in order to catalyze further action on the wide range of interventions needed to collectively short-circuit future targeted violence.

10 Takeaways:

  1. The time to stop violence is before it begins. Targeted violence builds on itself due to reprisal, but also because seeing others act on latent desires makes those desires feel more acceptable. Support for violent groups rises immediately after incidents of targeted violence, and when it declines it re-levels at a higher rate than before.
  2. Philanthropy in this arena is thin and concentrated in a few areas. It has so far overlooked a number of arenas the workshop highlighted as vital, particularly:
    • Altering in-group norms that can normalize and lead to violence, for elites and political leaders, rank-and-file partisans, and those at risk of committing violence;
    • Preparation for local officials in pre-violence prevention and planning;
    • Training law enforcement and Attorneys General in de-escalatory tactics and laws to reduce violence;
    • Community resilience work for targeted communities;
    • Helping perpetrators and those at risk of perpetration leave violent groups; and
    • Improving data on communities at risk of targeted violence in the U.S., gaps in assistance, and incidents of targeted violence with bipartisan agreement.
  3. Interventions need to focus on multiple points of influence to reduce violence, including political leaders and elites who normalize violent rhetoric and actions as well as individuals who might directly commit violence. Promoters and perpetrators of violence are not the same, though they can be mutually reinforcing. Therefore, neither can be addressed alone.
  4. Interventions work best from the local level upward, not the top-down. Violence happens in a locality and people will draw on the assets in their locality to prevent it, so resilience and prevention are both highly localized – though with the right resources, local interventions can be replicated, adapted, and scaled. Mapping warning factors can help predict where violence is most likely to occur so that resources can be directed ahead of time toward prevention.
  5. Local communities need assistance to plan ahead of time to prepare for and respond to targeted violence and potential violence. Separating protesters and counter-protesters, swiftly arresting perpetrators, deterring militias, and training police in respect and de-escalation dampen ardor for confrontation. Legal challenges against violent groups can also reduce their momentum.
  6. We must all speak against violence, but party leaders and elite influencers from both parties are particularly important in speaking against incitements to violence. Because people respond to in-group norms and violence is stronger on the far right, politicians and leaders who support President Trump play a particularly strong role in shaping or censoring violent behavior.
  7. Moderates willing to work across communities and temper their own groups are the first to be intimidated and silenced as extremism grows; they need particular support.
  8. Media interventions should train journalists to complexify their story lines. Coverage that counts every issue as a win or loss for one side increases the temperature. Build on journalists’ incentives to offer surprising or positive stories that emphasize ambivalent attitudes and focus on multifaceted identities to humanize fellow Americans.
  9. People seeking to engage with violent far-right groups are also far more likely to click on mental health ads. We should increase resources to off-ramp potentially violent individuals who are seeking help and belonging and finding it in violence.
  10. Based on historic and overseas trends, we should focus on violence just after elections and intimidation beforehand rather than election day itself. In the U.S., violence spiked for the two weeks after the 2016 election. Research suggests that winning makes supporters feel more justified in using violence, while losing might provoke anger, particularly if elections are contested. Preparation now can help mitigate these risks in 2020.

Further Resources

Anti-Defamation League H.E.A.T Map and “Murder and Extremism in the United States in 2018

Nichole Argo Ben Itzhak, Rachel Brown, and Heather Hurlburt, “Building U.S. Resilience to Political Violence: A (Globally Informed) Framework for Analysis and Action,” New America Foundation and Over Zero.

Bridging Divides Initiative

Rachel Hilary Brown, “Defusing Hate: A Strategic Communication Guide to Counteract Dangerous Speech,” U.S. Holocaust Memorial Museum

Thomas Carothers and Andrew O’Donohue, “How Americans Were Driven to Extremes: In the United States, Polarization Runs Particularly Deep,” Foreign Affairs, September 25, 2019.

Communities Overcoming Extremism, “The After Charlottesville Project Final Report,” October 2019.

Moonshot CVE, “Charlottesville: The Aftermath.”

Nathan Kalmoe, “Yes, Political Rhetoric Can Incite Violence,” Politico, October 30, 2018.

Nathan P. Kalmoe and Lilliana Mason, “Lethal Mass Partisanship: Prevalence, Correlates, and Electoral Contingencies,” Prepared for Presentation at the January 2019 NCAPSA American Politics Meeting.

Rachel Kleinfeld, “A Short Primer on Preventing Political Violence,” Washington Post, October 4, 2019.

Rebecca Littman and Elizabeth Levy Paluck, “The Cycle of Violence: Understanding Individual Participation in Collective Violence,” Advances in Political Psychology, 36 no. 1, 2015.

Elizabeth Levy Paluck and Michael Suk-Young Chwe, “Confronting Hate Collectively,” American Political Science Association, October 2017.

“Prohibiting Private Armies at Public Rallies: A Catalog of Relevant State Constitutional and Statutory Provisions,” Georgetown Law February 2018.

Amanda Ripley, “Complicating the Narratives: What if journalists covered controversial issues differently – based on how humans actually behave when they are polarized and suspicious?” The Whole Story Solutions Journalism, June 27, 2018.

Cathy Shufro, “Everyday Justice: Law School professor Tracey Meares is helping police to build trust with civilians,” Yale Alumni Magazine, Jan/Feb 2018.

Session 1: Targeted Violence in America: Different Lenses on the Problem

Oren Segal, Director of the Center on Extremism, Anti-Defamation League (ADL)
Nathan P. Kalmoe, Assistant Professor, Louisiana State University
Liliana Mason, Assistant Professor, University of Maryland
Nichole Argo Ben Itzhak, Director of Research and Field Advancement, Over Zero

These four scholars as well as other experts in attendance provided a snapshot on targeted violence in the U.S. today and its psychological motivating factors. Domestic terrorism is down since the 1970s but has been rising again since 2010. The ADL H.E.A.T Map, which monitors violent extremist activity in the U.S. over the last decade, finds that although Islamist and right-wing terror plots are about equal, 78 percent of extremist murders are committed by right-wing groups. While 87 percent of Americans believe political violence is never okay, according to Kalmoe and Mason’s survey research, 3 percent of respondents believed that if violence advanced partisan goals, it was very justified, and an additional 5 percent felt such violence was moderately justifiable. Thus, a vanguard of likely perpetrators exists (in nearly equal amounts in both parties), surrounded by a larger community willing to excuse and normalize their violence.

Nichole Argo Ben Itzhak provided an overview of the psychological literature on norms and hatred. As social beings, people’s behavior is strongly influenced by what we believe other people think. Indeed, our perceptions of the norms of our in-group matter as much in shaping our behavior as our own attitudes and beliefs. So cues from in-group members, such as language normalizing violence from within one’s group, are especially dangerous, because they suggest that supporting violence is part of belonging.

Targeted violence relies on both those who commit it and those who normalize it. It plays on the human tendency to favor in-groups and denigrate out-groups. Motivated cognition means all people seek out information that reinforces our preexisting beliefs and in-group norms. Threats—whether real (such as competing for jobs) or symbolic (such as threats to one’s identity and status, which are fairly easily conjured by opportunistic leaders)—exacerbate these tendencies and can lead people to deepen in-group affiliation. Dehumanization or violence against an out-group is often asserted in positive terms as protecting the in-group. Commonly, hateful rhetoric and behavior will paint the in-group as the “real” victim, justifying dehumanization, discrimination, or even violence by claiming that it is essential to protect the in-group.

People who have more status or popularity within an in-group are particularly influential. Thus, we must all speak against violence, but party leaders and elite influencers play a particularly strong role in shaping violent behavior by signaling what behavior is allowable or important to the in-group. For example, by not condemning inflammatory or dangerous speech by President Trump, Trump-supporting politicians play an outsized role in signaling to broader political partisans that such language—and ultimately, behavior—is allowed. Crucially, remaining silent is taken as a signal of acceptance by violent actors.

Violence is best stopped before it begins. This is because partisan violence deepens political polarization. In its aftermath, more people justify political violence. Violence builds on itself not only due to reprisal but also because seeing others act on latent desires makes those desires feel more acceptable. Kalmoe and Mason’s research found that both the mass shooting in a Pittsburgh synagogue and the mailing of pipe bombs to prominent Democrats increased peoples’ belief that using violence was justified for partisan purposes. Similarly, after a white supremacist attack, there is a spike in engagement with white supremacy content. Following the spike, the baseline resets at a higher level of engagement than before the attacks. Moonshot CVE found that the week after the 2017 Charlottesville Unite the Right rally, internet searches across Google, Yahoo, and Bing registered an 1,800 percent increase in searches indicating a desire to kill Jewish Americans, an increase of 200 percent for killing ethnic minorities, and 40 percent for murdering African-Americans. Searches to join the KKK increased by 800 percent, and there was a 22,000 percent increase in people wanting to donate.

Within the ensuing conversation, two themes emerged:

The Role of the Media:

  • Rachel Brown explained that journalists must be particularly careful to avoid strengthening in-group/out-group norms. Amanda Ripley emphasized that the media should stop treating every issue as a win or loss for one partisan side, which increases polarization. Instead, they should emphasize the complexity of policy outcomes, the multifaceted nature of our identities, and the ambivalence of many people towards the varied outcomes of most policy—all of which complexify media story lines.
  • Repeating hate speech, even to condemn it, actually normalizes it. Instead of repeating hateful/stigmatizing/or dehumanizing language, journalists can allude to the concepts while highlighting positive speech and action – for example, in the wake of the synagogue shooting in Pittsburgh, highlighting the mosques that supported the Jewish community, rather than focusing on the attacker.

The Profile of People Who Support Violence:

Kalmoe and Mason’s research found that the most correlated traits with normalizing violence are people with aggressive personalities, strong partisanship, and “moral disengagement”—that is, they rationalize that the out-group is lesser than themselves.

However, in the case of internet trolling of the family members of a victim of a white supremacist attack, several of the trollers turned out to be 13-15 year olds who thought it was funny or simply fun to be transgressive. This is also the crucial age for joining gangs of all types. Supremacist groups are targeting recruitment at this demographic. People just slightly older, from 16 to 25, are the most likely to commit all types of violence globally. There is strong demand from school boards and K–12 institutions who feel ill-equipped to tackle these issues amongst their students. Helping them can have long-term impact.

U.S. government programs focused on targeted violence (largely cut) had focused on assigning those identified as white supremacists to mental health professionals. Yet there is significant debate around the issue of mental illness and acts of violence. Individuals are attracted to violent groups for many reasons, particularly a desire for belonging and feelings of marginalization, as well as socioeconomic disenfranchisement. Since not all violent people are mentally ill, future government programs should make the first means of engagement more of a sorting process, possibly conducted by trained social workers.

Nevertheless, Moonshot CVE found that people seeking to engage with violent far-right groups are also 115% more likely to click on mental health ads. So ensuring that more mental health resources are available to those who are suffering from depression, social isolation, and loneliness would be a useful intervention.

Researchers are concerned about how loneliness might be driving a desire to engage with problematic groups or greater partisanship to create a sense of belonging. The significant expansion of loneliness among younger adults, especially young men, may be leading to deep reductions in social well-being across a variety of indicators.

Session 2: Interventions Overseas: What Do We Do, Why, & What Have We Learned

Jason Ladnier, State Department Bureau of Conflict and Stabilization Operations, Inter-agency Fellow at U.S. Institute of Peace
Rachel Brown, Executive Director of Over Zero, and author of Defusing Hate: A Strategic Communication Guide to Counteract Dangerous Speech
Heather Hurlburt, Director of New Models of Policy Change, New America
Mike Jobbins, Associate Vice President for Global Affairs and Partnerships, Search for Common Ground
Vidhya Ramalingam, Founder and Director, Moonshot CVE

Mass violence results from a confluence of factors. While various constituencies may have grievances, it often takes elites to mobilize people to violence. Political elites who do this believe that inciting violence will help them gain or maintain power or resources, and they tend to be opportunistic, making decisions about whether to escalate violence based on how a society (or the international community) reacts to smaller violent events, such as a protest that turns violent. Mass violence results from a pathway of decisions by the elites who incite it for personal gain, the people who supply vehicles or housing, the media who amplify, and those who carry it out.

Overseas, we take a whole-of-society approach. We need to address elites altering the norms of violence, but also people committing violence. Law enforcement strategies are important, as are strategies that build the resilience of communities to withstand violence. What’s done at the elite level has high visibility while civil society action is more easily felt at the local level. Both are crucial. Grassroots and local action that asserts norms of non-violence can enable political elites to take similar stands. Philanthropy has played a crucial role in Northern Ireland, Colombia, and elsewhere in fostering dialogue between groups and within groups to enable this change at the grassroots and elite levels. There is no single intervention that fixes political violence; instead, prevention involves managing stresses to the whole system and improving the health of the whole system.

When we intervene overseas, we begin by mapping all avenues for disrupting targeted violence, then look for the weakest links in the chain. For instance, it might be hard to influence political elites. So can targeting those actually committing the violence deter others? Or are there ways to target those less invested in the outcome? At the same time, we look for positive change agents, trying to identify people in all communities who can engage their own sides in positive ways.

Rachel Brown and Heather Hurlburt noted that it is hard to build strength around what we don’t have. We should instead take an asset-based approach that cultivates resilience before a problem erupts. For example, the fact that mosques came out to protect synagogues in the wake of the Pittsburgh shooting was not just a lucky happenstance – it had been cultivated through years of building bridges across those communities. In the wake of the Boston marathon, calm voices ensured that Muslims as a whole weren’t portrayed as violent perpetrators. This response was the result of years of investment in resilience activity and concerted efforts by local leaders.

A few particular lessons we can draw from addressing violence overseas:

  • We need to help targeted communities build resilience, explained Rachel Brown and Heather Hurlburt, helping them develop early warning monitoring, responses to protect themselves, communicators that can bridge communities, and methods to deal with the trauma of an attack and its aftermath. This has intrinsic worth in helping those most targeted by violence. It also plays an instrumental role in reducing tensions: unaddressed trauma can impede engagement across divides when positive outreach does begin.
  • Pre-violence planning at the local level can make a huge difference, even in the worst of conditions. In the Bosnian war, the city of Tuzla stayed peaceful because of strong coordination. When nationalists started spreading propaganda, civic organizations pushed back to show there was political will to maintain norms of nonviolence. This led to the elites’ ability to take action to prevent the city from erupting into violence and to further coordinate with city officials and civic organizations. This resilience was the result of strong coordination beforehand with the mayor, civic leaders, and elites.
  • Build on assets: People tend to focus on preventing and counteracting negative messaging within groups. But equally important is finding creative and constructive ways to channel angry but positive energy (for instance, white people who feel that white nationalism is “not them”) into productive messaging and campaigns in solidarity with targeted groups. This helps renormalize anti-violence and delegitimize violence against targeted communities.
  • Finding the right messengers to reach different constituencies is crucial. In Burundi, one person explained to Mike Jobbins, “the people leading during the daytime weren’t leading at night.” Both are important to target. Helping elites use messages that cut across identities and built norms of non-violence has been important overseas: for example, in Kenya, Rachel and Heather explained that emphasizing neighborhood identities helped normalize tensions between ethno-linguistic groups. Meanwhile, the “people leading at night”—that is, those committing the violence—may not be engaged with mainstream media and may need to be targeted in other ways.

Lunch, Session 3: Focus on Charlottesville and Lessons Learned

Michael Signer, former mayor of Charlottesville during the Unite the Right rally in 2017 and recent director of Communities Overcoming Extremism, a program to work with law enforcement, politicians, and the private sector to reduce the potential for violence
Shanna Batten, Charlottesville native and recent director of the Community Resilience Initiatives Program at the University of Maryland Center for Health and Human Security.

The first rule for preventing violence during a protest or public event is to separate groups likely to be confrontational. This requires clarity on the roles of local actors: someone must have the authority to make the decision to separate groups, for instance, there needs to be a pre-agreed plan between elected and appointed officials and police for how to do so, and they must possess the legal authority to do so.

Legal authority can be difficult. Groups hoping to spark violence are using free speech events as cover, because free speech is heavily protected legally—protestors have the right to group near the object of their protest (that is, a civil rights memorial or a courthouse in the center of town). Elected officials needed to overcome a court’s presumption in favor of protecting free speech by showing there is overwhelming evidence that a protest would be violent and that the violence would be targeted—that is, that there is “credible evidence” of an “imminent attack.” The constitutional standard for an imminent attack makes it harder to overcome the free-speech objections, and thus harder to use re-location to de-escalate tensions.

There are, however, laws that can be used against organized militias. Mary McCord and Georgetown University have pioneered lawsuits based on the requirement that militaries be under civilian control to prevent militias from returning to Charlottesville, and both are teaching other local communities about how to use these laws.

Ultimately, pre-planning can reduce the likelihood of violence, and the violence in Charlottesville resulted primarily from errors in planning. There were no fixed barriers across the streets, no game plan for the post-rally period, and no planning for how to proceed if intermixing of protestors got violent. It was this failure to plan ahead that allowed violence to swell. The event could not be successfully relocated to a safer location due to disagreements between an elected but largely disempowered mayor and city council and unelected city manager and police leaders about strategy. Providing accountability for mistakes was also made more difficult because of those same factors.

Whether targeted violence comes from within a community or outside actors come to a community, in the aftermath of an event like Charlottesville, a community faces trauma. For instance, Shanna’s African-American mother was so concerned about being targeted for violence that she scraped beloved partisan political bumper stickers off her car and hid in her home all weekend.

Communities can build resilience to this type of violence/trauma, but resilience and holistic prevention efforts should be developed locally because violence occurs at the local level. Vague federal policy for preventing political/targeted violence is of limited usefulness for localities. A range of resources establishing preventative violence interventions can be supported by state or federal levels, but should operate from the county- or local-level upward.

Shanna described her program’s mixed methodology drawn from principles of collective impact, public health, and emergency preparedness. Since the program is community-led in its prevention of targeted violence, it helps communities overcome feelings of helplessness. Through awareness education and intentional engagement, it creates a sense of shared community norms and a common denominator in prioritizing early prevention of targeted violence. Community members then are able to turn to each other for help and create intervention resources across sectors, drawing on public health concepts and approaches.

Legislation and litigation are essential for immediate solutions, but these are stopgaps. A positive solution to build community resilience is equally necessary. Local buy-in is essential to the development and sustainability of prevention resources. Responsibly informed and collaborative pre-violence prevention activities within communities and preparation activities for local governments will help both localities and communities reduce the likelihood of violence, suppress violence in the moment, and rebuild afterward.

Session 4: Interventions in the United States: Typology, Maps, Gaps

Neil Aggarwal, Assistant Professor of Clinical Psychiatry, Colombia University
Samar Ali, Co-founder and President, Millions of Conversations
Brent Decker, Chief Program Officer, Cure Violence
Tracey Meares, Walton Hale Hamilton Professor of Law, Yale Law School, and former committee member, Task Force on 21st Century Policing
Shannon Hiller, Non-resident Fellow, Bridging Divides Initiative, Princeton University

Session 2 showed the importance of a whole-of-society approach to addressing targeted violence, one that considers separately, in broad strokes:

  • Political elites who may incite, normalize, or speak against violence;
  • In-groups whose fringes commit violence but whose more moderate members can choose to either normalize or build stronger norms against violence;
  • Civil society groups that can bridge divides and strengthen norms against violence or ostracize those trying to reach across differences;
  • Those who commit or are likely to commit violence;
  • Targeted communities who need help building resilience before an event and empowerment afterward;
  • Law enforcement, who play a decisive role in preventing violence and de-escalation; and
  • The media, whose coverage can exacerbate polarization or help communities complexify their identities

This session discusses some of the types of work being done in the US in these arenas:

Mapping Actors and Gaps: The Bridging Divides Initiative has created a beta map to analyze where risk factors for targeted violence and extremism are most present, where there are civil society resources addressing violence or polarization, and where the gaps exist. Their maps require further work but already show large gaps in the Mountain West, parts of Texas, and the Great Plains. With more data and analysis, these maps can help local leaders better understand and mitigate these risk factors in their communities.

In-Group Persuaders: Cure Violence works with those who are close to perpetrators but have chosen a different path, such as former gang members talking to current ones.. These in-group persuaders can detect the most high-risk individuals and situations in a way that outsiders simply cannot, and they can raise the alarm when trouble is brewing. In-group persuaders can be the most influential messengers to reach those at risk and dissuade them from using violence. However, they often face difficulties from outsiders who distrust their closeness to the problem. In the world of political violence, in-group persuaders might be extremely far-right individuals who nevertheless disavow violence.

Millions of Conversations partnered with M&C Saatchi to study how to take people persuaded by hate and extremism and move them to being neutral to positive against the “other.” The study found that Americans who espouse views prejudiced against Muslims are looking for common American values that are shared across groups, messaged by news sources they already trust and conveyed by messengers they can relate to. Millions of Conversations is careful to do this in a way that doesn’t reinforce negative tropes.

Helping Perpetrators: Engaging with perpetrators and would-be perpetrators is crucial. Isolation is an important variable, so it is important that worried family or friends who see someone they care about getting involved in hate have somewhere to go—such as a group like Life After Hate, trained social workers, or other programs at the local level that provide a way out.

Targeting the people who have committed or are at risk of committing violence may be done through innovative online platforms such as the online-offline work of Moonshot CVE, because people searching for hateful content online can be located and addressed at scale online but then moved offline. Once people are found who are searching for words indicating interest in white supremacy or other hateful activity, they can be targeted with repurposed ads and alternative content to keep them from going down a rabbit-hole of hate, to plant seeds of doubt, and to help them engage with local services that move their interactions offline. The best technological interventions ironically start online but must get as personal and individualized as possible, enabling the scaling up of personal, one-on-one highly individualized interventions with social workers, for instance.

The USG programs targeting extremists tend to take a mental health approach. Moonshot CVE’s data explains why this has merit, based on the number of people who both search for white supremacist content and click on ads for mental health services. However, not everyone who is violent has a mental health issue. We should both increase mental health resources and have a more multidisciplinary approach to off-ramping those with propensities towards violence.

Law Enforcement: Law enforcement personnel are unique in their ability to exacerbate the problem of political violence if they take, or are perceived as taking, a side. In the United States, many communities targeted for violence also have high levels of distrust for law enforcement. Disrespectful policing practices, racist chat-groups, and movements such as the Constitutional Sheriffs and Peace Officers’ Association further drive a wedge between law enforcement and the people they must help. This distrust has consequences: studies of violent radicalized individuals in northern Africa and the Middle East show that a violent or negative encounter with a security professional (affecting them or someone they knew) was the tipping point that moved people from having grievances to engaging in violence among 71 percent of the nearly 500 radicals interviewed. Security personnel can increase the likelihood of violence through poor riot- and crowd-control techniques, as in Charlottesville where the planning did not entail separating groups of protestors.

However, law enforcement are also uniquely positioned to lower the temperature and reduce violence through swift arrests of perpetrators, crowd control separation, and other techniques that can be taught. Interventions to help police avoid bias, de-escalate situations, and serve the law rather than a partisan or ideological goal are therefore essential. Working ahead of time at the local level to build trust and foster mutual respect between police and communities pays significant dividends, because public safety is ultimately co-produced by citizens and police who are viewed as legitimate. These law enforcement interventions are an under-used but decisive element in reducing the likelihood of spiraling violence.


The group worked through four scenarios detailing ways that violence might plausibly emerge from America’s current situation. After the scenarios, groups convened by professional area to determine what that set of actors could do to reduce violence. Ideas that emerged included:

Government and Parties

The group agreed that political parties and the federal government are likely to be followers rather than leaders in reducing targeted violence in the current political climate. But there are actions that could be facilitated and would be helpful, for instance:

  • Non-Violence Agreements: Party leaders (and/or major party funders) could issue a joint political party statement on guardrails against violence for 2020. The statement would draw a line at some level of violence past which both parties would agree that violent rhetoric or behavior was unacceptable and both parties could proactively agree that “this is too much”. It might include a rejection of deep fakes, doxxing, or depictions of violence being committed against other candidates or officials, as well as a more general commitment to non-violence. The group discussed whether a national security argument could unify parties against efforts by foreign countries to interfere in our elections.
  • Local and State Coordination: Mayors/governors/other state and local leaders could be convened for a consensus-building process on how to maintain the rule of law and work with law enforcement in this moment of high partisan tension, forging agreements between the state and local levels for red lines. These efforts should also highlight positive models.
  • Block Grants: Grant funds could be created at the local level to support groups discouraging or preventing violence, ideally incentivizing groups to come together through the grant structure.
  • Temporary Legal Actions Around Elections to Reduce Risk: In open carry states, time and date laws around polling places could be re-examined by local officials to prevent intimidation. The group discussed whether this type of action would be feasible or politicized in the more than 40 states where guns can be carried into polling places.

Within the scenarios, participants also discussed how to disincentivize violence by lowering the stakes – for instance, depoliticizing institutions such as the Supreme Court through term limits or various other suggests to reduce the existential fight that takes place around each appointment. Reducing the extremism of political parties through actions such as Ranked Choice Voting was suggested. Some participants felt, conversely, that strengthening political parties at the state level and helping them build more community presence could aid their position of gatekeeping against extremism, others were more skeptical given the current reality of parties being driven by more extreme base voters in primaries.


Media should prioritize “do no harm” reporting that reduces polarization. It should balance positive and negative stories, looking for stories on helpers and other positive interventions. Media training should help journalists within their incentive structure by finding novelty in re-setting norms and framing helpful interventions as “surprising” news. A toolkit for journalists and media producers to help them report with greater complexity would be useful – so would prizes and financial incentives.

Media interventions should focus intentionally on different forms of media to reach different groups. For example, social media and partisan media may be the most likely to reach potential perpetrators. Local news may be most useful for speaking to targeted communities. Bystanders who must be encouraged into action may be best be reached through social media, entertainment, and even commercials. More thought should be given to how media can reach “influencers” – religious, political, and celebrity, among others.

Civil Society

The civil society group felt their sector could play two crucial roles: resetting norms within groups and building bridges between groups.

Finding credible individuals who can engage within an intra-group dialogue may be one of the more important roles of civil society. For instance, conservative, Evangelical pastors are best positioned to talk to their own flocks. Helping these intra-group conversations reinforce other cross-cutting identities that unite them with fellow Americans could be useful—particularly if built around a common existential or large threat that pulls people together. To shift norms and narratives within groups, civil society could help communities develop a storyline of where a group is and where it could be that does not involve stigmatizing or dehumanizing other groups. People within groups who are willing to outreach to “the other” are the most vulnerable and need protection. Conciliation is dangerous and civil society must confront extremism within itself to protect those individuals.

In building bridges, the civil society group ran into some division. Civil society groups exist on both sides of the partisan divide, and need to wrestle with how to engage parts of civil society from the “other side” as well as groups on their side willing to work with people or organizations some deem “outside the pale”, determining whether it is most useful to marginalize or to bring them into the conversation. In working across divides, the civil society group suggested looking for particular cross-cutting issues of opportunity that would enable working together across partisan lines to build relationships and trust, such as criminal justice reform work did.

Finally, civil society can help paint a picture of the risks America faces, making it credible and believable to people of all sides so that they can see the abyss before they are in it.

Law Enforcement

The law enforcement group felt that rebuilding trust in law enforcement among communities targeted by violence required:

Training: Interventions that emphasize de-escalation and other tactics useful in potential violent conflicts, as well as norm-setting within forces regarding how to build respect and legitimacy. The group felt there was a strong need to actively influence in-group norms of law enforcement to reduce violence, racial profiling, and other precursors to violence. Such interventions would cultivate positive messengers and isolate negative messages and those perpetrating them. Messengers can be within law enforcement, but also among others who would influence law enforcement, such as Chambers of Commerce, the local judiciary, the governor, etc.

Reform-Minded Affinity Groups: Discussants thought there was a need for strategy sessions to determine how best to support reform-minded law enforcement officials, building their numbers and amplifying their voices. They also suggested forming and calling on affinity groups to speak out when there is a legal issue and to emphasize accountability.

Recruitment changes: Discussants felt that it was important to make police look more like those communities not through quotas, but by making it more attractive for a variety of minorities to serve. However, they also noted that demographic similarity was not enough to prevent violence and perceptions of violence.

For the particular problems with the Constitutional Sheriffs cohort choosing to follow their own interpretation of the constitution and potentially ignore or overrule their state laws, as with the 2nd Amendment “Sanctuaries” in states with gun control laws, discussants felt the answer required supporting at the electoral level more reform-minded sheriffs.


Philanthropists discussed the issues areas they were already funding. These included:

  • Civil society network building initiatives, which promote a culture where openness to the other is possible;
  • Litigation, including: voting rights, rule of law, and immigrant rights;
  • Legal defense for immigrant communities;
  • Free press work;
  • Government oversight, including funding watchdogs;
  • Groups rebuilding confidence in democratic institutions;
  • Voter engagement;
  • Groups addressing anti-semitism and other forms of hate;
  • Democracy protection across the ideological spectrum;
  • Analysis of census failure and preparation for legal issues with census;
  • Strengthening Congress as an institution;
  • Improving campaigns and elections through addressing big money in politics;
  • Election systems protection; and
  • Voter registration and turnout.

Philanthropists saw their main roles as legitimating democratic institutions, such as voter engagement, voting rights, and media freedom; using litigation to strengthen norms and buy time; and supporting efforts to bridge divides and build bipartisan coalitions.

The conversation illuminated that much of the funding came from a more progressive viewpoint, and that extremely little existing funding addressed some of the key areas essential to preventing violence that were mentioned throughout the workshop, particularly:

  • Norm-resetting among politicians, government leaders, and other influencers;
  • Intra-group norm change that could reach those likely to rationalize violence and thus normalize it;
  • Training law enforcement and prosecutors in de-escalation and use of available laws to reduce violence;
  • Community resilience for targeted communities;
  • Pre-violence planning at the community level to reduce violence;
  • Pathways to off-ramp those at risk for violence, such as social work, mental health, and internet targeting approaches;
  • Institutional changes and long-term work with communities. De-polarization efforts were largely focused on building conversations across divides, but did not address electoral structures (primaries and safe seats) or media structures (consolidation, etc.) that may be contributing to polarization. Most de-polarization efforts were individually focused or short-term and also failed to include longer-term community reconciliation efforts.

Philanthropists discussed whether it would be useful to use philanthropic networks to start a national conversation among funders and influencers to promote democratic, nonviolent practices. There was tension between funders wanting to respond to emergencies and those who felt that philanthropy was best positioned to catalyze long-term structural change. A second tension emerged from whether philanthropy should prize de-escalating polarization, building bridges, and rebuilding healthy democratic institutions, or focus on pursuing a policy agenda (including a policy agenda believed to be strengthening democracy, such as reforms to Congressional procedures) that may only be achieved through greater partisanship and practices that could further polarize the country, such as close single-party votes.

CLOSING KEYNOTES – Tony McAleer and Samar Ali

Tony McAleer, Co-founder and Board Chair, Life After Hate
Samar Ali, Co-founder and President, Millions of Conversations

The workshop closed with an acknowledgement that even among the assembled group, there was a wide divergence of opinion on the scope and urgency of the problem and lingering questions on whether and how to apply lessons from international experience to the U.S. context. While some called for more facts to establish whether there actually was a problem of targeted violence in America, others thought that stories would be more useful in getting around issues of motivated-cognition and helping people set new norms. In any case, agreeing on what red lines would be regarding targeted violence and its normalization could form the start of a useful conversation that could bring more actors into dialogue about how to prevent such lines from being crossed.

Attendees adjourned to hear Tony McAleer’s personal story and his founding of Life After Hate. He shared how being middle class and from a comfortable, if cold, home did not prevent his teenage self from searching for belonging within a white supremacist organization. Tony described how he became a bully to prevent getting bullied and how the internal belonging of the white supremacist movement pulled him further in. He then described how crucial it is to not close the door, but instead provide pathways out for those who may be ashamed of their violence and desire to leave but fear social stigma on the “outside.”

We then heard from Samar Ali, Director of Millions of Conversations, about the hate campaign against Muslims that has targeted her for years. Her core message began with the idea that if one wishes to disrupt political violence, they must begin with changing the narrative of “the other.” This is because the cycle of political violence begins with the labeling of the other. Millions of dollars today are flooding into demonization and dangerous labeling techniques that have the possibility of leading to violence. This happened to her while working at the White House and again while working for the Governor of Tennessee.

The slippery slope follows a pattern from labeling to animosity, progressing to fear, blame, and anger, then transforming into hate and political violence. She discussed how she and her supporters were able to slow the negative momentum against her by working across partisan lines to build an accurate narrative that successfully countered the demonizing one.

Having learned that disrupting the cycle of hate required changing the narrative of “the other,” Millions of Conversations is now catalyzing a media campaign to do that work through in-person conversations reinforced by a robust online digital strategy that builds empathy. They have found that success requires having infrastructure ahead of time to counter negative campaigns once they begin. That’s because, as she described, local hate campaigns are facing well-funded, well-established, and carefully managed campaigns, often catalyzed by individuals at the national level using hate to advance political goals.

After a full day considering how close America might be to greater violence, one participant proposed that we take a different perspective. Perhaps, she suggested, the level of polarization and violence we see today might be a sign of a long-overdue reckoning with issues our country has long swept under the carpet. Change is always accompanied by backlash, and the change our country is experiencing now is no exception. If we can avoid targeted violence and keep our democracy whole through this moment, our country may end up in a far better place.