Introduction

An increasing number of politicians, media outlets, and members of the public have grown alarmed that digital social media may be amplifying the frequency and severity of urban violence (Crane 2018; Darville 2022; Tarm 2018). These voices call on law enforcement agencies to leverage social media to more aggressively surveil, arrest, and prosecute (potential) perpetrators, placing already-marginalized communities of color at even greater risk of harmful legal entanglements (for a review and critique, see Lane 2019; Patton et al. 2017; Stuart 2020a, 2020b).

Unfortunately, despite the central place of urban violence in sociological research, the discipline has been surprisingly slow to disentangle the empirical relationship between violence and social media (see Moore and Stuart 2022). Among the small body of existing scholarship there is consensus that social media provide new and additional avenues for challenging rivals, displaying toughness, and building street cred—processes long theorized to precipitate and perpetuate violence. Whether such online behaviors lead to offline violence, however, is subject to significant debate. Some scholars suggest that social media play a causal role in exacerbating violence. Given the capacity of social media to increase individuation, disinhibition, and anonymity, (potential) combatants may be less likely to feel the risks and dangers that once prevented them from escalating conflicts or engaging in retaliation (Lapidot-Lefler and Barak 2012; Patton et al. 2013). Other scholars are more skeptical that social media amplify offline violence. Studies suggest that although violence may spill into the streets in some instances, the majority of challenges stay confined to online space, in part because social media allow users to build coveted reputations for violence without engaging in the physical violence that may have once been required (Lane 2019; Stuart 2020a, b; Urbanik and Haggerty 2018).

Much of this ambiguity and disagreement can be traced to the relatively narrow scope of existing research, which has focused primarily on the individuals most immediately involved in conflicts, or on inflammatory attributes of social media content, seeking to pinpoint which types of online challenges are most likely to provoke retaliation (Patton et al. 2019; Stuart 2020b). Yet, as research on informal social control and violence intervention programs indicates, the mechanisms influencing violence often involve people, networks, and institutions beyond those directly involved in confrontations (Garot 2010; Sampson 2012; Slutkin et al. 2015). This broader “ecology of violence” necessarily shapes how (potential) combatants perceive and act on any online content they encounter. Indeed, the proliferation of social media may actually enhance the influence this ecology exerts on residents, lowering the barriers of contact between various intervening parties, enabling novel methods of informal social control (Hampton 2010; Lane 2019; Stuart 2020a; Villamizar-Santamaría 2022).

We advance this emerging body of research by examining the mechanisms by which social media mitigate urban violence. Drawing on ten years of fieldwork among Black youth in Harlem and Chicago, we argue that social media afford a historic level of what new media scholars refer to as “communication visibility” (Leonardi 2014; Treem et al. 2020). This allows onlookers not only to observe others’ relationships, exchanges, and actions, but to exert direct and indirect influence over subsequent relationships, exchanges, and actions. In Harlem and Chicago, this takes the form of friends, families, community members, and other third parties exploiting communication visibility to intercede into violent conflicts, preventing them from occurring or limiting their reach and intensity once they begin.

Our findings carry important theoretical implications. By introducing the concept of communication visibility into urban violence scholarship, we draw much-needed attention to the broader, digitally mediated ecology of urban violence. This contains actors and processes that are otherwise overlooked by theories of violence, which focus on dyadic exchanges between conflicting parties or the expectations of physically co-present audiences. Our analysis demonstrates how third parties can serve as a vital protective factor against exposure to violence. This provides a potential answer to longstanding questions about why certain individuals, despite their presence in violent cultural contexts and risky networks do not engage in or experience violence at the same rate as peers under similar conditions. Methodologically, our approach to digital urban ethnography offers a framework for examining social media activity as a historically new vehicle by which community members exert informal social control.

Theorizing (Non-)Violence in the Social Media Age

Understanding the capacity of social media to reduce and even prevent urban violence requires updating traditional theories of urban violence with emerging insights from communication and technology scholarship.

Expanding the Code of the Street

The predominant theory of urban violence was developed well before the proliferation of social media. Throughout the twentieth century, sociologists, criminologists, and other scholars explained urban violence as the manifestation of campaigns for respect, which Anderson (1999) famously termed the “code of the street;” amid insufficient economic opportunities and inadequate police protection, disadvantaged residents cultivate violent reputations to resolve disputes and attain dignity. “The code,” writes Anderson (1999, 72), “revolves around the presentation of self. … A person’s public bearing must send the unmistakable…message that one is capable of violence, and possible mayhem, if the situation requires it.” Successfully building a violent reputation requires publicly challenging others, discrediting their violent reputation, and claiming their status for oneself.

The street code theory has been instrumental for understanding disparate rates of violence between neighborhoods. Yet, the theory struggles to account for divergent behaviors between residents living within the same neighborhood cultural context. Why do only certain individuals engage in violence, and why in only certain situations? What leads others to seek respect and resolve feuds through non-violent means? Despite Anderson’s (1999, 97) claim that “the culture of the street doesn’t allow backing down,” non-violence is, in fact, the modal response to affronts, even in the most violent settings (see Garot 2010). Unfortunately, by focusing primarily on instances of violence, much research is unable to identify the intervening factors that mediate the relationship between the street code and physical confrontations, particularly those that re-direct potential violence toward peace.

The most notable exception to this trend is the work of Garot (2010), who analyzes instances that, according to the street code theory, should have resulted in violence but did not. Garot (2010) finds that even individuals who proclaim strict adherence to the street code frequently leave disputes unresolved, opt for non-violence, and accept the loss of “street cred” in order to maintain relationships and commitments that exist beyond the dispute at hand. “The necessity to preserve social ties dominates,” Garot (2010, 160) writes, “concerns with ‘masculinity’ and ‘respect’ are important only insofar as they affect such social ties. Whichever option is chosen…is ultimately based on an individuals’ calculus of which might threaten his or her social integration less or enhance it more.”

The mediating role of non-violent social relationships indicates that researchers must extend the scope of analyses to the broader ecology of violence, comprised not only of potential combatants, but also of non-combatant third parties, who can exert a de-escalating influence. Within sociology and criminology, three related yet distinct lines of research leverage this perspective to explain (and address) violence as the result of overlapping networks and interactional structures. The first line is research on informal social control and disorganization (Bursik and Grasmick 1993; Sampson 2012). Almost a century of literature suggests that residents in impoverished neighborhoods are less willing and/or able to “prevent or sanction disorderly and criminal conduct through informal surveillance of the streets and direct intervention into problems, such as questioning persons about suspicious activity, admonishing individuals who are misbehaving, and informing parents about their children’s misconduct” (Kubrin and Weitzer 2003, 375 − 76). This formulation has been subject to numerous critiques. Ethnographic research demonstrates that while neighborhoods may appear “disorganized” to outsiders, they are in fact characterized by dense social networks that afford high levels of solidarity, trust, and durable systems of informal control (Pattillo 1999; Suttles 1968). Critics attribute such blind spots in informal control scholarship to a number of factors, including the overreliance on self-report surveys that may miss the diversity of residents’ ground-level practices of intervention; the near exclusive focus on social control efforts that occur within public, face-to-face scenarios; and the increasing fixation on formal control efforts by nonprofit organizations, police departments, and other external institutions (Stuart 2016; Vargas 2019; Villamizar-Santamaría 2022).

The second line of scholarship examines how network structures and characteristics influence individuals’ likelihood of violent perpetration and victimization (Green at al. 2017; Papachristos 2009). For example, leveraging arrest data to construct networks between co-offenders, Papachristos and Wildeman (2014) find that homicides in Chicago are densely concentrated within very small but high-risk networks, and that exposure to violence diminishes with each social tie removed from a victim. Similarly, the shorter one’s network distance to other individuals in possession of firearms, the greater the likelihood of exposure to gun violence (Roberto et al. 2018). Although this shift in focus from dyads to multiplex relationships has produced key findings on the transmission of violence, questions similar to those raised by the street-code theory remain: Why, for example, do individuals located in comparably risky networks differ in their involvement in and exposure to violence? What are the protective factors that buffer the influence of risky networks? Answering these questions has proven difficult, as the methodological reliance on arrest data to construct co-offender networks necessarily omits individuals’ relationships with other non-violent individuals throughout their multiple, overlapping social networks.

A third line of research emerges from recent innovations in violence intervention philosophies that embrace an epidemiological approach typically directed toward limiting the spread of infectious diseases, like AIDS/HIV (see Slutkin et al. 2015). Interventions such as Illinois’ Cure Violence, Chicago’s Institute for Nonviolence, and Boston’s Ceasefire employ trusted neighborhood residents—often former violence-involved individuals—as street outreach workers. When violent events occur, outreach workers act as concerned third parties to “interrupt” subsequent retaliation. Outreach workers gather intelligence on the social networks of the victim and offender to determine potential pathways of future violence. Then, they make contact with the involved parties’ social ties to steer them toward peaceful resolutions, social services, and material and health resources (Cheng 2017; Moore and Stuart 2022). Studies find that the Cure Violence program reduces shootings in targeted communities by 41-73%, with program participants reporting significant reductions in disposition to use violence in both petty and serious disputes (Delgado et al. 2017; Milam et al. 2016; Slutkin et al. 2015). Although this approach recognizes the importance of third-party influence, it is similarly limited in its capacity to systematically identify and account for the de-escalating influence of the expansive network of informal and intimate neighborhood social ties, which may play out beyond the purview of outreach workers. Such ties may not only enhance the de-escalatory capacity of outreach workers, they may also be operating further “upstream,” so to speak, preventing some number of violent events before they occur, and thus reducing the need for interrupters’ involvement.

These three lines of research suggest the need to expand methods, data sources, and analyses to more completely capture local ecologies, particularly the otherwise “invisible” relationships and interventions reducing and preventing violence. Overreliance on self-report surveys and observations of public spaces and community organizations produced theories that overemphasized the social control maintained in physical locations where people can observe, monitor, and interact face-to-face. These frameworks also privilege the actors (e.g., victims and perpetrators) and temporal moments that are most proximate to violent events as most determinative.

Times, however, are changing

The ubiquitous presence of social media in urban neighborhoods presents new possibilities for violence mitigation as individuals’ social ties and networked behaviors become increasingly visible and manipulable among the entire ecology of violence. This visibility similarly extends to researchers seeking to analyze this process.

Social Media and Communication Visibility

With the recent proliferation of digital social media, scholars have begun to examine their influence on neighborhood networks and social ties (see Hampton 2016; Lane 2019). Arguably the most important transformation wrought by social media is their capacity to render communication visible to others in ways that were previously impractical or impossible. This increase in what communication and technology scholars refer to as “communication visibility” makes relationships and interactions between individuals visible to existing ties and new and wider audiences. It allows users to view more readily monitor and intervene in each other’s respective communication networks. In turn, this expands the ease of surveillance and expansion of informal and formal social control mechanisms. The visibility of online systems affords accountability and “onlooker effects,” i.e., the capacity of third parties to exert influence over other users’ behavior through their explicit or implicit presence. As Treem et al. (2020, 49) describe, “Onlookers can directly influence user behavior by actively intervening into users’ activities, but often their influence is indirect—through users’ assumptions and inferences about third-party actors’ presence and judgements.”

Communication visibility has been applied primarily to organizations to understand how workers and managers use enterprise social media to integrate knowledge within firms, streamline work practices, and maintain managerial control (Ellison et al. 2015; Leonardi 2014). Some neighborhood studies of social media use also implicitly discuss the role of communication visibility in enabling new forms of monitoring and accountability within neighborhood contexts (Lim 2017; Patton et al. 2016, 2018).

Our focus here builds on unprecedented levels of communication visibility across different social relationships within neighborhood-level networks and direct and indirect onlooker effects on neighborhood violence. Research on the integration of social media in outreach work has shown how youth workers use communication visibility to bridge intergenerational gaps and gain a more holistic picture of at-risk youths that attends to risk factors, expressions of emotional state, and prosocial interests (Lane 2019; Lane and Marler 2020; Lim 2017; Patton et al. 2016; Stuart et al. 2020) discuss how social media have enhanced crisis intervention in Chicago neighborhoods by allowing youth workers to identify and respond to problematic situations by directly following the feeds of youth or relying on youth informants. By witnessing in real time or learning through informants about provocative content, such as posts of youth gang members posing in rival territory (what Stuart [2020b] terms “calling bluffs”), youth workers can dispatch to problem areas and talk privately to key involved individuals. This online access to mounting conflict provides an opportunity to redirect emotions and de-escalate tensions through various communicative strategies. These include emphasizing the consequences of violence and pointing to its potential visibility to police, family, employers, and others. Youth workers, therefore, use communication visibility to enable established mechanisms of nonviolence by helping potential combatants to reflect on “good reasons” not to fight (Garot 2010).

The theory of communication visibility suggests at least three important shifts that social media use enables within local ecologies of violence. First, social media expand and deepen the social ties that dampen violence. Friends, family, and other third parties responsible for pulling individuals away from violent interactions are now more privy to the social lives of potential combatants. Second, social media make monitoring neighborhood youth and surveillance easier. Non-violent third parties are much less dependent on physical co-presence to observe the potential for conflicts to develop in real time and can now track problematic situations with their “eyes on the screen” (Villamizar-Santamaría 2022). Third, social media expand the available means of intervention. Just as surveillance has been untethered from physical co-presence, third parties can now sanction (potentially) violent behavior digitally by using mass or interpersonal communication. Communication visibility also means that these online sanctions and efforts by young people to address their own role in violence (including desistance) will be witnessed by a wider range of onlookers, who can join the ongoing surveillance and intervention efforts.

Data and Method

To empirically examine the relationship between communication visibility and non-violence, we employ a digital urban ethnographic approach in which the ethnographer simultaneously position themselves “on the ground, in the social media feeds, and in the online networks” of neighborhoods and their residents (Lane 2019). This approach is better positioned to reveal the variable relationship between (potential) conflicts and neighborhood ecologies of violence than self-report surveys, arrest data, or even traditional ethnographic methods of shadowing those directly involved in violence. Methodologically, communication visibility is a vital affordance for fieldworkers to extend their data collection to incorporate both the physical and digital streets of urban neighborhoods, and to identify key protective factors that help de-escalate and prevent physical conflicts.

We draw on two multi-year digital urban ethnographies in Harlem (2009–2014) and Chicago’s South Side (2014–2019). Both projects involved the ethnographer (in Harlem, a White man in his thirties, and in Chicago, a mixed-race, Black-identifying man in his thirties) working closely with Black teenage residents in the context of outreach work and after-school programming to develop multifaceted relationships that evolved over years. These relationships required unique forms of trust building and acceptances of enduring social differences. In both field sites, we shadowed participants individually and in small groups throughout their daily lives. We spent between twenty and fifty hours per week conducting direct observations in offline spaces (e.g., in their homes, their neighborhoods, to and from school, at work, in criminal justice settings) and online spaces (e.g., profiles on platforms such as Facebook, Twitter, and Instagram). To better understand the relationship between online and offline behaviors, we enlisted participants to review daily social media activity with us. This included their various posts, content uploads, and private messages. These debrief sessions focused attention on key themes and surprising findings. They offered prompts and discussion topics for follow-up conversations and interviews, in which we asked participants about the origins, aims, meanings, and consequences of particular online behaviors. To enhance reflexivity, both authors engaged in a number of explicit fieldwork strategies and orientations. This included shadowing participants across multiple settings and contexts to capture their complex personhood—that is, the multiple and even contradictory social identities that emerged with various others. This holistic approach to fieldwork mitigated exoticizing narratives, as did conducting “member checks” to verify data accuracy (see Stuart 2020a, 212–216). Throughout the fieldwork process, we recorded notes by hand using small notebooks, smartphone note-taking applications, or laptops. Immediately following fieldwork observations, we expanded notes into detailed narratives. We inserted screenshots of social media into our field notes and annotated them using our own observations and the accounts of participants.

The current analysis emerged from an abductive process whereby we compared a striking finding that had emerged in our respective field sites—namely, that residents commonly used social media to mitigate violence (see Tavory and Timmermans 2014). After identifying this theme, we both independently revisited and re-coded field note and interview data to better identify the mechanisms by which social media afford the ability to de-escalate conflicts. We then combined the data, developed common codes, and compared incidents across contexts to find commonalities and differences. This stage directed our attention to the role of third parties, not only as direct agents of intervention, but also as an imagined audience among potential combatants. Recognizing a key parallel between our empirical findings and theories of communication visibility, we reanalyzed our common dataset once more, asking three key questions of each piece of data: First, who is the third party, and what is their relationship to the potential combatants and the larger ecology of violence? Second, how does this third party act on communication visibility? Third, how does this third party’s intervention alter or affect the broader ecology of violence and the likelihood of physical violence? Following this process, we selected examples that most clearly illustrate the primary de-escalation processes of communication visibility in each site. These serve as the basis for the empirical narrative to follow.

Neighborhood Third Parties and Direct Onlooker Effects

Below, we discuss two types of direct onlooker effects observed in Harlem and Chicago—namely, third-party interventions by women protectors and a street pastor. These onlooker effects are prominent and powerful mechanisms in the prevention and de-escalation of neighborhood violence.

Sisters, Girlfriends, and Women Protectors

Privileging violent incidents and feuding parties, prevailing theories of violence overlook the manner in which third parties commonly intercede to prevent such incidents from occurring in the first place. As in many violent neighborhoods, young women in Harlem and Chicago frequently assume the responsibility of keeping the men in their lives (usually boyfriends and brothers) safe from violence. They envision themselves as protectors with the unique ability and obligation to steer their loved ones away from relationships, interactions, and spaces that are likely to expose them to violence. This takes the form of three common strategies.

The first strategy—preventing network overlap—entails third parties monitoring loved ones’ online communication partners to better hinder friends and family from entering into the social networks of potentially dangerous individuals. Danielle, a seventeen-year-old Chicagoan, exemplifies this technique. Danielle’s fifteen-year-old younger brother, DJ, had recently been spending time with the local gang faction, called CBE. At the time, DJ’s affiliation was relatively peripheral and tenuous, but in his eagerness for their approval, he postured heavily online, using gang symbols and exaggerating gang-related behaviors. Danielle was fearful that DJ’s online behavior would draw aggression from CBE’s nearby rival faction, called Murderville. Fortunately, like many of the young women in our fieldwork, Danielle maintained online relationships with the men affiliated with Murderville. They had attended the same school, pursued romantic relationships, and remained mutual friends and social media followers. This connectivity allowed Danielle to regularly “lurk” on the profiles of young men associated with Murderville to continuously assess whether they shared friends and followers with her younger brother, DJ. Danielle’s biggest worry was that DJ might romantically pursue a young woman without realizing her intimate involvement with a man affiliated with Murderville. This might stoke jealousy and retaliation. A related worry was that Murderville would enlist one of these women to seduce and ambush DJ (see Patton et al. 2020). As Danielle explained:

She [a woman closely tied to Murderville] ‘gon set him up, maybe.. .. They get that girl to send you a text or something, all like, ‘Come meet me over here, I wanna see you.’ Stuff like that. Then when he go over there, they’re all outside waiting for him. That’s how you get caught up. They do that. They use these females to get you alone. You think you’re getting a piece [sex], but then they jump out on you.

On one occasion, Danielle noticed that DJ had posted photos on his social media profile with a young woman whom she had previously seen on the profiles of young men affiliated with Murderville. Danielle did not divulge information about the young woman’s relationship with Murderville to DJ; she feared this might encourage her brother to challenge his rivals, putting him into increased danger. Instead, she covertly undermined DJ’s relationship with the young woman by sharing false rumors about her. She told DJ that the woman was known to be very promiscuous (a “thot”) and was currently sleeping with several people. Discouraged by the thought of the young woman’s multiple sexual partners, DJ severed his ties with the woman.

Danielle is far from alone in this strategy. In Harlem, for example, a young woman named Kelice prevented network overlap by printing out her younger brother’s inflammatory social media posts to give to her mother, who summarily grounded the young man, forestalling future run-ins with those who might do him harm.

A second strategy deployed by third parties is geolocating threats. As protectors monitor the social media activity of their loved ones’ rivals, they note the background photos and location tags to discern common travel paths and current whereabouts. Armed with such real-time information, they can then steer loved ones away from these locations, thereby preventing feuding parties from sharing the same physical space. This strategy proved particularly important for Danielle, who was aware that, given DJ’s desire to prove his mettle to his friends associated with CBE, if he encountered Murderville in public, he was unlikely to back down from a challenge. On one occasion, Danielle noticed Murderville posting photos of themselves hanging out at a popular basketball court bordering their territory. Knowing that her brother, who is an avid basketball player, might travel to the court at some point that day, she provided a preventative and more attractive distraction. Danielle invited DJ to spend the day with her and her girlfriends, several of whom DJ had been “crushing on.” The group spent the entire day together, grabbing lunch and shopping downtown.

Some protectors were far more transparent than Danielle in their use of locational intelligence. For example, nineteen-year-old Shawnae advised her cousin to avoid a nearby public bus stop based on the knowledge—gained through social media monitoring—that a rival group of young men frequently hung out outside an adjacent fast food restaurant.

A third strategy deployed by third parties is disciplining digital communication. Communication visibility affords protectors the ability to monitor and discreetly regulate loved ones’ social media content creation and conversations with subtle rewards and punishments. Danielle, for example, dangled the prospect of spending time with her girlfriends in front of DJ as a kind of “carrot” to encourage him to cease his inflammatory exchanges online. On one occasion, two of Danielle’s girlfriends had arrived by car to pick her up for the afternoon. DJ, who had been standing outside of their apartment, pleaded with them to tag along. Danielle used the opportunity to convince him to remove two Facebook posts in which he had indirectly insulted a slain Murderville member. She insisted to DJ that he was “too much drama,” and that she and her friends wanted to enjoy a “drama-free day.” Danielle made DJ a deal: delete the two posts immediately, and she would allow him to get into the car. She watched closely as he pulled up his Facebook account and deleted the two posts in question. With a smile, DJ climbed into the car. In a later conversation, Danielle reported that she also finds ways to reinforce DJ positively when he posts non-gang-related content on social media, even enlisting her friends to send him “fake congratulations” when he posts about his accomplishments in basketball or school. According to Danielle, she intends both positive and negative reinforcements to encourage DJ to take a moment to think more carefully about the potential consequences of his online communication.

Disciplining digital communication is particularly common among friends and family members who possess key social and material resources desired by the young men in their lives. In Chicago, a mother of four, named Tasha, successfully persuaded her eldest son, Junior, to delete particularly inflammatory social media posts as a condition for providing him rides to school, court, and other places that are difficult for him to travel by public transportation. Several romantic partners reported “playing hard to get,” withholding intimacy when they felt young men had unnecessarily escalated online conflicts. Concerned third parties also use online spaces to discipline behavior. In Harlem, older teens call out younger neighbors’ incendiary posts, publicly casting them off as artifice and bluster, in the hopes of pressuring them to delete such content.

The Street Pastor

Street pastors depend on complex understandings of dating dynamics and adolescent social life (Lane 2019). Lane studied a street pastor, an ordained minister and Black man in his fifties, known simply as Pastor, who was a key outreach figure in Harlem until his death in 2018. For Pastor, social media provided a visualization and point of entry into the concatenating networks of local youth. Pastor’s intervention methods relied and acted on the communication visibility of teen networks and evolving neighborhood drama.

Pastor’s Twitter account distinguished Pastor as the only (known) adult in many teenagers’ Twitter networks. This digital connection enhanced his capacity to locate youth and build on relationships that were rooted in his presence and history in the neighborhood (his ubiquitous SUV, drop-in space for youth, and street credits as a former drug dealer and incarcerated individual). The very question of how he knew where to address his outreach efforts, i.e., which neighborhood young men and groups, situations, and locations to prioritize, depended upon communication visibility and gendered assistance from young women in roles equivalent to Danielle’s, with networks that crossed and spanned rival groups and territories.

As a third party to neighborhood violence in an established intervention role, communication visibility afforded Pastor the opportunity to digitize street outreach. When Lane joined Pastor and a group of youth and adult volunteers at a college basketball game, Pastor discussed with Lane how he had connected his outreach efforts to the ecology of violence in Harlem. Pointing to two high-school senior girls with us at the game, “They represent [access to] fourteen crews,” he said. The girls and their Twitter accounts were the “common denominator” between boys and boy groups in conflict. Pastor, who was following 1,723 accounts that day, picked from the girls’ follower networks the boys he wanted to follow, drawing also on the fact that many of the boys referenced their neighborhood groups in their profile details. Pastor added that he was initially surprised that the “crew members” he requested accepted him, taking this as a positive indication that they were open to his help.

The visibility of communication within the youth Twitter networks provided a system for monitoring network overlap and geolocating threats that he used to intervene and thwart (potential) occurrences of physical violence. Based on Twitter (and direct communication with the teens), Pastor made his rounds in the neighborhood to be physically present when and where he believed violence might unfold, such as at parties in unsupervised apartments and community rooms where rival groups were likely to converge. Pastor drove his familiar SUV to park outside the party and draw attention to his presence by playing music and providing a place for teens to socialize in the seating rows or around the vehicle while charging their phones. In the event of rival groups gathering and posturing in front of each other, Pastor stepped in the middle and yelled “Pastor on deck” or “P.O.D,” which initiated a script for groups to walk away and wave off the fight.

Pastor used the visibility of potential conflict on Twitter to broker information about threats, using text-message “blasts” that he sent to parents and other neighborhood adults to mobilize them to locations where a fight or shooting might occur. Using technologies that were more intergenerational, namely text messaging and Facebook, he quoted threatening communication and reported on brewing conflicts for the adult segments of the community. Pastor sometimes cajoled the adults to take action when he felt that they were not doing enough. For instance, after two fights among large groups of girls in the St. Nicholas Houses, Pastor texted, “Word on the street they will be at it again tomorrow,” before adding: “Won’t be long before the guys get involved. I wonder what parent[s] are waiting for.” After a separate shooting in the same general area a few days later, he connected these events to stir outrage, texting, “I warned a shooting would be next when the boy[s] got involved. NOW I’m SAYING THERE WILL BE RETALIATION SHOOTING! PARENTS IT’S OUR KIDS.” That afternoon, on Facebook, he identified by their initials the young people shot and informed the community that they “will make it.” Then he warned, “If we don’t get involved to end this, the second wave is coming.”

Pastor acted not only in concert with the adult world to intervene through a physical presence in the neighborhood but also established an online presence in unfolding teen drama. He interceded by making intervention visible within youth Twitter networks, for instance by reframing conflict to deter actual participation in physical violence. After Pastor had read Tweets from youth that anticipated mob violence, such as “2mm [tomorrow] shit gonna be craazii wrdd,” Pastor, affected the same writing style to discourage violence. He Tweeted, “Tomorrow Friday! Itz gonna be a crazy night! I’m praying no one getz body bagged! Holidayz comin an all don’t wanna be meeting any new grievin momz!”

Pastor used the communication visibility of social media to connect and amplify his relationships and efforts in the neighborhood. He drew on access to young people and their peer interactions through social media, especially girls’ social media. He altered the ecology of violence by changing the visibility and scope of conflict to inform parents and neighborhood adults and bring them into the unfolding social action to mitigate violence. The physical presence of Pastor and the adults served to deter and disrupt violence (or at least delay and displace it elsewhere) by giving young people who were considering violence a credible “out.” That youth agreed to be monitored in the first place and shared information with Pastor directly or indirectly implied that they were open and willing to de-escalate. Pastor both acted on and created his own communication visibility to motivate direct and indirect onlooker effects that prevent and reduce incidents and levels of neighborhood violence.

Indirect Onlooker Effects: Arranging Truces on Social Media

We have focused attention on direct onlooker effects and demonstrated how third parties exert influence over potential combatants’ behaviors, relationships, and communication practices. Next, we turn to indirect onlooker effects—that is, the process by which potential combatants proactively alter their own communication practices in anticipation of third parties’ potential monitoring and intervention. As exchanges become increasingly visible to onlookers throughout the broader ecology of violence, potential combatants endeavor to predict and manipulate third-party behaviors. Previous interactions implicated in urban violence take on new significance in the context of heightened visibility. This is perhaps nowhere more apparent than during online truces, which afford a number of functions and benefits previously unavailable offline. First, online truces preclude the need to enter rival territory for face-to-face interaction, allowing rivals to reach amity without the fear of ambush and despite low levels of mutual trust. Second, feuding parties can more readily demonstrate to concerned loved ones and community members that they are responsive to their desires for de-escalation. Third, by reaching a truce in full view of others within the ecology of violence, potential combatants are able to more rapidly and widely disseminate news of a détente, allowing onlooking rivals and allies to alter their relationships and behaviors.

Each of these affordances is well illustrated in an online truce carried out in Chicago by CBE and a rival gang faction known as Go Hard Block (GHB). The CBE-GHB feud had begun rather abruptly, when GHB formed an unexpected alliance with one of CBE’s long-standing neighborhood rivals, a faction known as Crown Town. As in many cities, Chicago gangs adhere to the unwritten rule that when one faction “cliques up” with another faction, it is expected to enter into its new ally’s ongoing rivalries and active wars. As AJ, a nineteen-year-old associated with CBE, explained, “Let’s say you and me clique up. Then your opps [rivals] is my opps. If you’re blowin’ [shooting] at them, then I better be blowin’ too. If I ain’t, then I ain’t really your mans [ally].” Indeed, young people routinely use their allies’ willingness to aggress against their own rivals to measure their alliance’s strength.

Given the politics of alliances, the new relationship between Crown Town and GHB caused CBE considerable concern; CBE had attracted yet another rival. Within a matter of days, young men affiliated with GHB joined in on Crown Town’s ongoing Twitter activity, insulting CBE with derogatory tweets about their slain friends. CBE responded online and in person. On two occasions, young men associated with CBE attacked young men affiliated with GHB, when they unexpectedly encountered them in public. As the rivalry intensified, third parties began stepping in. Mirroring the direct onlooker effects described earlier, CBE’s friends, girlfriends, families, and neighbors expressed their disappointment on Twitter about what they considered an unfounded and unnecessary rivalry. A young woman and cousin of one of the men affiliated with CBE tweeted, “#CBE + #GHB better WORK SHIT OUT mfs need to quit all the noise.” Another young woman from the neighborhood tweeted, “All Over Twitter I See #CBE #GHB Into More Bullshit. Grow Up!”

One afternoon, a young man in CBE, named Adam, received an unexpected tweet from one of the young men in GHB, named Salim, intending to put an end to their rivalry. As up-and-coming rappers in Chicago’s “drill music” scene (Stuart 2020a), Salim and Adam were the most publicly recognized affiliates of their respective factions. Salim opened the door to reconciliation, writing, “ay @[Adam’s user name] dont even know why we into it dis shit pointless.” Adam immediately replied, albeit defensively, acknowledging that GHB had forced CBE’s hand, “yall started this. you see how we comin.” Now that they were in dialogue, Salim extended an olive branch in the form of a musical collaboration: “lets link on a song. fuck this beef shit.” Adam agreed to the request, “bet.”

In the words of Jabari, another young man affiliated with CBE, who had been seated next to Adam and helped him craft his response, “That was it. We squashed [reconciled] the beef right then.” Although they never did collaborate on a song, from that day forward, physical violence and online insults between CBE and GHB ceased. In a later interview, Jabari detailed the young men’s thought process throughout the short, albeit effective exchange. Described by Jabari, the two acted on behalf of their gang faction, seizing the opportunity for an online truce. Jabari cited several key motivations. First, they recognized that an online truce was easier and certainly less dangerous than coordinating in person. “I ain’t about to go over there [to GHB territory]. I ain’t trusting no invite to come squash it. After what we been doin’ to them, they come out the blue with a tweet inviting us over there? Hell naw!”

Second, Jabari, Adam, and others had grown tired of the pressure from friends and family to end the rivalry. “Everybody was getting in our ass, talking ‘bout we ain’t got no real beef with GHB,” Jabari described. This pressure was particularly acute from girlfriends and sisters, who apparently threatened to withhold affection and various support. “The females quit fucking with us,” he complained, “talking’ ‘bout they don’t want to see us die. They got babies and stuff. Don’t nobody want their baby daddy dying.” By completing the truce online, they created verifiable evidence to satisfy those who aimed to hold CBE accountable. Rather than merely attesting to the fact a truce had been reached, they could show loved ones the series of publicly available, time-stamped tweets.

Third, despite the seemingly fleeting nature of social media conversations, the rapid pace of online dissemination serves to increase the potential success of truces, particularly in their nascent, most fragile moments. Contemporary gangs are more fractured and horizontally organized than in the past, with fewer systematized communication channels. With their high levels of communication visibility, social media platforms overcome this dilemma. As Jabari described:

There ain’t no leaders no more to tell everyone to squash it. The leader today is whoever got the most [social media] followers. They the ones who gotta broadcast it. If they squash a beef, everybody gonna see it right away. It’s about what your followers and entourage and the people you cliqued up with do. You might have squashed it, but they don’t know. You don’t want none of them to get into it with them [GHB] because they don’t know that you ain’t into it with them like that no more.

As Jabari indicates, a truce can only be upheld if most, if not all, of the connected parties are aware of it. Without up-to-date knowledge about their allies’ truces, former hostilities remain firmly intact. If an ally acts on these antagonisms, those more closely involved in the truce may interpret that behavior as a violation of the ceasefire. Junior summarized this point succinctly, “If it’s your entourage [committing violence], it’s basically you. … So the bigger your entourage, the more important it is to squash it online.” Indeed, Jabari attributes CBE’s current, durable peace with GHB to the truce arranged via Twitter.

Similarly, in Harlem, three groups involved in gun conflict solidified a truce by playing a basketball game in a previously contentious park, photographing and uploading the resulting images of peaceful recreation to social media. Online circulation of these images to the entire ecology of violence suggests that truces need not originate online to become visible to third parties; however, social media greatly enhances combatants’ capacity to shape the reactions and behaviors of onlookers.

Discussion

This paper employs digital urban ethnography to expand the scope of neighborhood fieldwork, allowing us to examine the evolution of informal social control within an increasingly digital age. Despite a long history of research about informal social control taking place on the physical street, we can no longer limit our lens to this interactional setting. Young people have increasingly gravitated toward social media to engage with one another, often in ways intended to de-escalate violence. Ignoring this space and its interactional affordances threatens to reify stigmatization and criminalization. Youth may be perpetrators of violence, but they are also (perhaps even more often) peacemakers.

Using Harlem (2009–2014) and Chicago’s South Side (2014–2019) as our case, we argue that social media regularly provide protective and diversionary affordances, “rooted” (Flyverbom et al. 2016) in the visibility of communication on these platforms and direct and indirect onlooker effects (Leonardi 2014; Treem et al. 2020). Communication visibility offers concerned parties tools to manipulate and influence behaviors, relationships, and encounters that otherwise accelerate violence. We presented several types of direct onlooker effects, in which third parties leveraged social ties and norms of non-violence (Garot 2010) to “pull” potential combatants away from violent encounters, despite their ostensible commitment to the street code (Anderson 1999). We also examined indirect onlooker effects as would-be combatants act on their own communication visibility to diffuse violent tension spreading across the network by enacting online truces that validate the concerns of their loved ones.

Of course, we are not Pollyannaish enough to suggest that onlooker effects are necessarily or inherently positive. Indeed, past research has identified at least two contexts—namely, prosecutions and vigilantism—in which onlooker effects harm community and individual wellbeing. As we contend, however, whether onlooker effects enhance or reduce community vitality depends largely on who acts as a third party (e.g., community insiders versus outsiders) and the goal of third-party monitoring and intervention (e.g., community inclusion versus exclusion).

Scholars have documented the affordances of social media for prosecutors who use the visibility of associations to bring enhanced charges against youth suspected of gang-related crime (Lane et al. 2018). As community outsiders, prosecutors seek to use communication visibility to remove suspected “gang members” from the community. The resources brought to bear are punitive, compulsory, and highly exclusionary. They include felony charges, incarceration, and criminal records. Although some argue that such surveillance and punishment can reduce neighborhood violence, these reductions come at steep costs to individual defendants, uninvolved community members, and other public safety mechanisms (Bell et al. 2021; Lane 2019; Levinson-Waldman 2019). When police detectives surveil social media and prosecutors use young people’s content in criminal indictments, these youth may respond by avoiding the exact forms of online communication that community insiders previously used to intervene and provide support (Lane 2019; Stuart 2020a). In contrast to law enforcement, the goal of the neighborhood onlookers discussed here is not to incapacitate (potential) combatants but rather to further integrate them into prosocial neighborhood networks, thus mitigating their risk and steering them toward nonviolent contexts (Stuart et al. 2020). These community members mobilize interpersonal, strategic, and affective resources that are based on voluntary compliance.

Scholars have also described a variety of onlooker effects that operate at the level of neighborhood watch programs and “digital vigilantism,” defined as “a process where citizens are collectively offended by other citizen activity, and coordinate retaliation on mobile devices and social platforms” (Trottier 2017, 55). As in the case of law enforcement, the aim of a neighborhood watch is the removal and exclusion of a (potential) perpetrator of violent crime (and property crime) treated not as a family member or community member but a stranger (Ahmed 2000) by neighbors and homeowners who “participat[e] in surveillance [that] is often based on the exclusion and suspicion of others” (Kurwa 2019, 113). Digital vigilantes reproduce racial narratives of exclusion (Kurwa 2019) and rely on intimidating, spiteful behavior, such as public shaming and doxing (Loveluck 2020; Trottier 2017). We see in these anti-social examples a critical theoretical distinction: not all onlooker effects are related to communication visibility as documented in our analysis. Digital vigilantes do not position themselves within the communication networks of their targets to influence their target’s behavior positively. Quite the contrary. Digital vigilantes communicate among themselves in opposition to those they deem “the other.”

Our analysis demonstrates the role of a wide range of actors involved in everyday anti-violence efforts that serve to integrate the community and repair its cleavages. These caring forms of lateral surveillance and intervention, we believe, merit further investment and coordination with violence intervention programs to provide more productive responses than prosecutions and vigilantism. Programs like Ceasefire and Cure Violence, in particular, can draw on this more proactive approach. Rather than wait until after a shooting to discern network rosters and intercede into the ecology of violence, outreach workers might instead leverage communication visibility on a more continual basis, moving further “upstream” and across the network to organize earlier, more community-based interventions (e.g., Stuart et al. 2020). In the corporate firm context in which communication visibility was originally theorized, firm workers use this affordance to both build on existing expertise and avoid work redundancy. Similarly, intervention programs might make more strategic choices about which community processes to further engage and where their efforts are inadvertently duplicative.

Finally, our study highlights that communication visibility and onlooker effects are examined most effectively through ethnographic study designs that consider the complex meanings of online interactions that surround offline violence. Our findings suggest that the same digital spaces of incendiary, online communication may also support lines of intervention, caring, and de-escalation with peer and intergenerational networks. This lends support to a growing sentiment among violence scholars that research should consider violent online content not in a vacuum, but in relation to positive, counteracting forms of engagement that iteratively co-evolve across both online and offline contexts (Lane 2019; Stuart 2020a).