Before opening fire in a Bend, Oregon, grocery store in August 2022, Ethan Miller expressed his hatred toward “EVERYONE & EVERYTHING.” Bearing this out, his rambling online posts displayed racial animus and extreme misogyny. He also railed against the U.S. government’s coronavirus response, police, religion, and technology. To Miller—whose attack killed two people—the 1999 Columbine High School shooters were icons and role models. But though Miller’s journal and online posts exhibit a range of extreme and targeted sentiments that seem to qualify the attack as a form of ideologically motivated violent extremism, he explicitly rejected being labeled a white supremacist or incel (involuntary celibate). Given the range of disparate sentiments that Miller expressed, if he were indeed a violent extremist, what kind should he be understood as?

Miller’s complex ideological profile is not unique, nor are the questions and confusion surrounding his motivations. Over the past few years, the United Kingdom and the United States have expanded their counterterrorism efforts to include individuals like Miller. British counterterrorism officials created the mixed, unstable, or unclear ideology category to include extremists who did not fit other counterterrorism categories (for example, the extreme right). The U.S. Department of Homeland Security’s 2019 counterterrorism strategy addresses targeted violence alongside terrorism, in recognition of the fact that many acts of targeted violence bear the hallmarks of terrorism even if they may not be categorized as such. The FBI has also begun to employ the phrase “salad bar extremism” to describe a trend of ideological mixing. In recent congressional testimony, for example, FBI Director Christopher Wray described extremists who hold a “weird hodgepodge blend of ideologies,” noting that this trend is producing challenges in “trying to unpack what are often sort of incoherent belief systems, combined with kind of personal grievances.” Other government officials and private-sector researchers have used a variety of different terms to discuss the same phenomenon, including ideological mixing and ideology à la carte. While government officials and experts have highlighted the particular challenges that this phenomenon poses for law enforcement and prevention practitioners, the trend as a whole is insufficiently conceptualized and lacks a framework for understanding distinct elements within. 

Recognizing the reality and urgency of this challenge, we introduce the term “composite violent extremism” (CoVE) and provide an accompanying typology as a mechanism for more rigorously conceptualizing violent extremists who defy neat categorization.

The CoVE framework addresses a number of shortcomings in current efforts to identify and label ideologically complex violent extremists. First, the metaphor of the salad bar does not accurately characterize all individuals who fit the trend and perhaps is not even accurate for a majority of them, as the salad bar metaphor implies intentionality on the extremists’ part. At a salad bar, people select ingredients with some level of care, only in rare instances ending up with food or other objects in their salad that they never intended. But it is possible that for many extremists, the mixing of ideologies and grievances is less intentional than the term suggests. The deluge of information in today’s online environment, and particularly in the social media space, may mean that people are shaped by the information rather than intentionally selecting certain ideas. For instance, ideological mixing might occur in part because of the illusory truth effect: When people are exposed to something over and over, they are more likely to believe that it’s true. Exposure to multiple toxic online cultures and message repetition that plays on previously held grievances or ideas might facilitate mixing in a far less deliberate way than the salad bar metaphor implies.

Second, and more important, salad bar extremism lacks a widely accepted definition or a publicly available conceptual framework for understanding cases that may be categorized this way. The same is true of other synonymous terms and concepts, such as ideological mixing. Absent such conceptualization, these labels become catch-all concepts that are poorly defined and thus encompass myriad individuals who are, in reality, not alike. The lack of ability to disaggregate individuals who fit the salad bar extremism concept presents difficulties for analytic, prevention, and detection efforts. Better categorization might reveal different patterns of behavior, unique risk factors, or trends in radicalization that would be invisible if every extremist who held any shade of ideological mixing was simply lumped together under one term.

Due to these intellectual gaps, we undertook a broad examination of ideologically ambiguous violent extremists, arriving at CoVE terminology as a replacement for the salad bar metaphor. But whatever term is employed to describe violent extremists that do not fit old ideological patterns, our more important contribution is the typology we provide to classify today’s “weird hodgepodge” extremists. This conceptual framework provides a coherent model for understanding cases that might otherwise be poorly categorized or under-examined and lays a foundation for further research into this evolving threat.

The Composite Violent Extremism (CoVE) Framework 

Composite violent extremists are violent extremists whose worldview appears to be composed of multiple distinct ideologies, sentiments, grievances, and fixations. The word “composite” denotes something made up of various parts or elements, thus encompassing the core of the salad bar metaphor without suggesting particular characteristics of the radicalization process (for example, that the mix of ideologies and grievances that the extremists have are selected with intentionality). 

In creating this typology and situating cases within it, we evaluated discernible ideologies, prejudices, grievances, and fixations that an attacker expressed, and further attempted to evaluate how central an expressed belief was to an attacker’s worldview. We assessed the centrality of a belief by analyzing how frequently an attacker expressed that belief, the attacker’s connections to groups or individuals adhering to a given ideology, individual self-identification with a particular ideological movement, and whether the attacker explicitly expressed ideological motivations. 

The work we undertook in categorizing attackers led us to conclude that ideological amalgamation happens in multiple ways and that four primary buckets can be used to classify extremists under the CoVE framework: ambiguous, mixed, fused, and convergent (Figure 1). 

 

Figure 1. CoVE typology.

The first category of CoVE, which we term ambiguous, encompasses individuals who draw on a variety of disparate prejudices and grievances but do not adhere to a discernible ideological framework. Violent extremists who fit this category include Frank James, who wounded 29 people in a shooting on the Brooklyn subway in April 2022, and Ethan Miller, the Safeway attacker in Bend, Oregon. Although Frank James posted hundreds of videos airing a variety of prejudices, grievances, and conspiracy theories—including expressing anti-U.S. sentiments, political grievances, and bigoted views about white people, Black people, Jews, and Latinos—SITE Intelligence concluded that none of his views “reflect any distinct extremist movement.” Similarly, Safeway shooter Ethan Miller claimed inspiration from the 1999 Columbine school shooting, expressed a range of misogynistic and racist sentiments, and aired various grievances related to the coronavirus. But he does not neatly fit into any ideological category. For both men, there was no clear expression of cohesive, discernible ideologies, though they articulated a range of sentiments that motivated their attacks. Ambiguous cases raise the thorniest questions about whether they should be categorized as violent extremism at all, as they are by definition the cases most lacking in discernible ideology. While one might question the framework’s ambiguous category for this reason, we think this point underscores the category’s importance: Inclusion of the ambiguous category in the framework reduces the risk of overlooking violent extremist cases that closely resemble non-ideological mass killers.

The second and third categories, mixed and fused, describe attackers who fit into existing ideological categories, such as white supremacist or jihadist, but also express other ideologies or sentiments that complicate placing them into only a single category. The mixed category applies to violent extremists who hold multiple different ideologies on relatively equal levels. For example, Zale Thompson, who attacked a group of New York City police officers with a hatchet in October 2014, drew motivation and inspiration from both black separatism and jihadism. Both Black separatism and jihadism are coherent ideologies. We classify Thompson as fitting CoVe’s mixed category because it was not clear that one of these ideologies primarily influenced him. Rather, both apparently had a significant effect on his worldview and identity. Including the mixed category in the CoVE framework reduces the risk that such cases will be inappropriately grouped into a single ideological category and also allows for the identification of potential synergies between different extremist ideologies.   

The fused category describes violent extremists who are oriented primarily around a single core ideology but also express sentiments associated with other ideologies. For instance, Jack Reed, convicted in Britain in November 2019 of six terrorist offenses, fused satanism and interest in Columbine with his neo-Nazi ideology. Reed’s embrace of neo-Nazism was clear from his journal, where he drew Nazi symbols and expressed admiration for Hitler. He also supported the British neo-Nazi group National Action. However, he also showed interest in satanism, describing his satanic beliefs on an online forum and referencing the esoteric satanist group Order of Nine Angles in his journal. He also repeatedly searched online for content about the Columbine shooting and repeatedly viewed content related to Anders Brevik, the perpetrator of the 2011 Norway terrorist attacks that claimed 77 lives. Compared to Thompson, Reed is less ideologically complex and possesses a core ideology. However, his interests in satanism and Columbine complicate the picture and differentiate him from your average neo-Nazi. We thus coded him as fused. The fused category ensures that the role of extremists’ dominant ideology, when they have one, is taken into account. This reduces the risk, which exists in grouping such incidents into catch-all categories like salad bar extremism, of inadvertently playing down the threat posed by broader violent extremist movements like white supremacism.

Convergent, the final category of CoVE, is for individuals who embody the adage “the enemy of my enemy is my friend.” This category applies when individuals or groups espousing different ideologies choose to cooperate or collaborate but do not adopt one another’s outlook. Each party maintains its own beliefs but agrees to work with those who hold fundamentally different beliefs but possess overlapping interests. One example of convergence in our dataset is the case of Michael Solomon and Benjamin Teeter, two self-proclaimed members of the Boogaloo Bois, a U.S.-based anti-government movement, who agreed to sell weapons to an individual they believed was a member of the terrorist organization Hamas. They did not convert to Islam nor did they join Hamas. Rather, the partnership was forged over shared anti-U.S. sentiments and a pragmatic money-making scheme. While the convergent category applies only in cases where individuals holding different beliefs choose to work together, it is indicative of a broader phenomenon of extremists expressing support for attacks carried out by violent extremists with different ideologies. As one small example of this broader trend, an Islamic State supporter who was planning to attack a synagogue claimed inspiration from the 2018 Tree of Life synagogue shooting in Pittsburgh, which was carried out by a white supremacist. The attack was inspirational for him based on shared antisemitic beliefs. Including the convergent category in the CoVE framework ensures that cases of cross-ideological cooperation are not ignored in the examination of the broader trend we describe simply due to their lack of ideological blending.

While CoVE can help to explain and categorize a broad range of violent extremists, it does not account for attackers who do not appear to adhere to an ideology, a prejudice, or a grievance, even if they otherwise seem to fit the characteristics of ideologically motivated attackers. For example, the July 4 Highland Park shooter displayed extreme interest in violence but did not express a specific extremist ideology, prejudice, or grievance.

The Implications of the Composite Violent Extremism Framework

The CoVE framework is intended to help practitioners and scholars conceptualize a change in ideologically motivated threats that the United States’ top law enforcement officials have identified as concerning and difficult to understand. CoVE fills a definitional and conceptual gap, providing greater clarity on a topic that has gained significant attention in recent years. This typology has implications for the full range of domestically focused counterterrorism efforts, including tracking overarching trends, prevention, deradicalization/disengagement efforts, the posture of law enforcement, and research efforts that can inform policy.

First, this typology will enable efforts to track important shifts in idiosyncratic violent extremism over time. The simple observation that, for example, “salad bar extremism is increasing” is not sufficiently granular to allow practitioners to understand specific radicalization dynamics. The four categories we delineate make clear that there are significant differences among the various extremists who have been classified under the “salad bar” label and enable analysts and scholars to track trends over time. These important differences among violent extremists have practical consequence:

• Prevention efforts. A range of ideologies, sentiments, grievances, and vulnerability factors can influence people’s propensity to conduct an act of violence across the four categories of CoVE. Given that prevention efforts are apparently most effective when they are as individualized as is practical and can speak directly to relationships, grievances, and the individual’s ideologies, a more granular understanding of the evolving radicalization cocktail could render prevention efforts more effective. Further, CoVE’s categories could allow for greater tracking of the most effective methodological approaches for each subtype. 
• Deradicalization and disengagement. As is the case for attempts to prevent people from succumbing to violent extremism in the first place, a more granular understanding of applicable drivers is important for disengagement and deradicalization efforts for individuals already involved in extremist activities. Previous deradicalization efforts have struggled to account for individual motivating factors and the ways they interact with an extremist’s beliefs. Given the complex way it conceptualizes ideological makeup and the unique accounting it provides of the full range of a person’s grievances, the CoVE framework could be fruitful for methodological development and evaluation for different types of attackers. For example, ideological factors play a very different role across the range of CoVE subtypes. The lack of a clear core ideology makes an extremist in the ambiguous category substantively different from a fused extremist, who in many (though not all) important ways resembles more traditional conceptions of an extremist with a defined core belief system. As such, effective deradicalization/disengagement efforts for individuals in these two categories will not look the same. This framework importantly provides a means of program evaluation based on how different individuals hold ideologies, prejudices, and grievances, not just what ideology they hold.
• Law enforcement. The manner in which the CoVE framework facilitates consideration of the totality of an individual’s ideological and grievance profile intersects in meaningful ways with law enforcement efforts, including efforts to protect potential targets of violence. For example, a mixed extremist who holds both radical environmentalist and extreme misogynist views might consider attacking both a gas pipeline (a potential eco-terrorist target) and a women’s college (a potential target for misogynistic violence). Further, holding composite extremist views may influence the scope of events that could precipitate violent acts. The greater the number of ideologies, prejudices, or grievances, the more potential triggers for violence. Thus, the ideological confusion of an ambiguous extremist might broaden the range of triggers or targets considerably. In contrast, a fused extremist with one core ideology combined with other sentiments might have a narrower range—albeit potentially still broader than if that person held just one ideology, with the extremist’s sentiments generally aligning with that ideological perspective. For instance, if a neo-Nazi also is an adherent of QAnon, that individual presents two vectors for attacks: The individual can be triggered to violence in response to a motivation rooted in neo-Nazism or a QAnon motivation. That individual could perpetrate violence in response to an instance of perceived government overreach—a QAnon trigger—presenting an additional vector for attack than if the person held only neo-Nazi beliefs. For an ambiguous attacker holding four or five different distinct prejudices, there could be even more possible triggers. 
• Policy research. As noted, the CoVE framework can drive research efforts that can inform policy. For example, the framework can assist efforts to understand how changes in the information environment are altering ideological formation and giving birth to new radicalization patterns. If cases involving ambiguous extremists are increasing but fused attackers seem to be decreasing, it would be important to understand that phenomenon specifically and conduct targeted research into potential causes. Resulting research will have implications for a wide range of stakeholders, including policymakers, practitioners, social media companies, mental health professionals, and civil society leaders.

As instances of ideologically motivated violence continue to make headlines, the concept of composite violent extremism highlights the continuing evolution of the face of violent extremism. CoVE can be a framework for understanding and classifying a broader range of ideological actors, and it can aid efforts to tailor prevention and detection efforts to combat this growing threat.