Coordinated Dis-Coordination

Dissidents mobilizing against a repressive regime benefit from using public information for tactical coordination since widespread knowledge about an upcoming event can increase participation. But public calls to protest make dissidents’ anticipated activities legible to the regime, allowing security forces to better stifle mobilization. I examine collective action during Sudan’s 2018–19 uprising and find that mobilization appeared to be publicly coordinated through social movement organizations and internet and communicative technology, consistent with common channels identified by existing literature. Yet embedded field research reveals that some dissidents independently used public calls to secretly organize simultaneous contentious events away from publicized protest sites, perceiving that their deviations would make the regime’s repressive response relatively less efficient than the resulting efficiency losses on the movement’s mobilization. These findings push future work to interrogate more deeply the mechanisms by which dissidents use coordination channels that are also legible to the regime they are mobilizing against.


INTRODUCTION
P rotests against autocracy are more likely to achieve regime change when they draw large turnout (Chenoweth and Stephan 2011): if mobilization cascades, then regime elites-including officers in the security apparatus tasked with repressing mobilization-may switch their allegiance and stand with the population. Mobilization, however, requires overcoming multiple collective action problems including that of tactical coordination (Little 2016): the actual how, what, when, and where of contention.
Past work has highlighted the importance of public information for tactical coordination in at least three ways. First, formal civil society groups can act as social movement organizations (SMOs) and disseminate coordinating information to members and nonmembers alike (McCarthy and Zald 1977). Second, societally implicit focal points or repertoires of contention can coordinate collective action without explicit communication (Schelling 1960;Tilly 2010). Third, information and communication technology (ICT), and especially many-to-many platforms like Twitter and Facebook, can facilitate the quick and public dissemination of information about anticipated or ongoing collective action (Christensen and Garfias 2018;Clarke and Kocak 2019;Enikolopov, Makarine, and Petrova 2020;Steinert-Threlkeld 2017).
Although strategies that rely on public information can make mobilization more efficient by increasing the visibility of upcoming protests to a larger number of potential participants, these strategies also make dissidents more legible (Scott 1998) to the very regime they are trying to undermine. Public coordinating information allows the regime to better anticipate dissidents' actions and counter mobilization in the first place.
The effects of public information for protest coordination must, therefore, be evaluated against the concurrent repression effects (Carter and Carter 2020;Chau, Hassan, and Little 2022;Dragu and Lupu 2021). This trade-off is well established within research on SMOs: organizations facilitate mobilization (Olson 1965), but are also prime targets for repression (Berman 2021;Davenport 2015;Sullivan 2016), cooptation (Collier and Collier 1979), and infiltration (Mattingly 2020) precisely because they are known loci for action. Recent events have shown that this puzzle extends to other forms of public coordination, beyond SMOs. Many contemporary movements have been less formally organized and have instead relied on focal points of action or ICT. In turn, research has shown how mobilization organized through public coordination is repressed using that very same public information-for instance, repression is preemptively increased before focal days (Truex 2019); and autocracies censor online calls for collective action (King, Pan, and Roberts 2013), coerce online opinion leaders (Pan and Siegel 2020), and at times, turn off the internet entirely.
To what extent and in what ways do activists use public coordinating information to address the tactical coordination problem in repressive environments? This paper finds that dissidents may engage in coordinated dis-coordination whereby they manipulate publicly available coordinating information to purposefully introduce friction and disorder into their activities with the belief that doing so will negatively affect the regime's repressive response to a relatively higher degree than the inefficiency introduced onto the movement. Specifically, I describe tactics in which dissidents use widespread public information about upcoming "main" events-whether created by SMOs, drawn from implicit focal points of action, or widely circulated through ICT-to independently and secretly plan other "parallel" mobilization activities away from the event expected through public information. 1 These simultaneous events retain most parameters of public coordinating information (time, date, and type) but change one element (place). Public coordinating information plays a critical role, even if it is not fully followed.
At its broadest, then, coordinated dis-coordination is an example of the ways in which social movements must perpetually innovate (McAdam 1983). The initial "liberation technology" of ICT caught many autocracies flat-footed, but these same regimes are continuing to adapt their strategies of political control amid our notso-new digital environment, and as a result, ICT has proved to be a double-edged sword for dissidents coordinating contention (Earl, Maher, and Pan 2022;Morozov 2012;Weidmann and Rod 2019). I build on these insights to show that dissidents have innovated as well, devising novel uses for technology of their own so as to become less legible to the regimes they are trying to overcome. More specifically, I empirically find that one strategy is to purposefully not follow the coordinating information that ICT helps solve-that is, to make more disorganized the very tactical coordination problems that dissidents are thought to use ICT to address (Little 2016)-precisely so as to catch the regime off-guard.
I suggest two scope conditions for this strategy. First, coordinated dis-coordination is more likely within protracted movements. Evading repression is less of a concern in movements that have cascaded because security forces often defect and refuse to repress civilians. Second, the usefulness of coordinated dis-coordination to dissidents is inversely correlated with the capacity of the security apparatus. Secretly planned, simultaneous events are more disruptive when security forces are limited, under-funded, and where surveillance capacity to pick up on these deviations is low.
I examine the dynamics of coordinated discoordination during Sudan's 2018-19 popular uprising against autocrat Omar al-Bashir. For much of the uprising, the Forces for Freedom and Change (FFC)-an opposition coalition comprised of unions, political parties, and other formal groups-publicly announced upcoming collective action activities through schedules that were widely circulated through ICT. The FFC's public coordination during this period led to modest levels of participation as well as high levels of regime repression (Dahab et al. 2019).
In response to persistent repression at main events, exacerbated by public coordination, I find that informal groups used the FFC's public announcements to independently organize simultaneous protests that fall under the category of coordinated dis-coordination. This occurred in two distinct ways. First, some groups used advance knowledge of main events to plan parallel protests in areas away from the location of the main event in a process I call jittering. Dissidents retained most of the publicly disseminated coordinating information but jittered location to hold an event in a place where security agencies were not expecting. Second, and separately, some groups would plan contingent takhf īf (meaning "lightening" or "reducing" [the repressive burden] in Arabic) protests. Some dissidents not participating in a scheduled main event would remain on the lookout for high security presence at main events. If violence was heavy, they then launched a parallel protest in another locale to attract security forces, distracting them away from and reducing repression on the main site.
On the surface, it is theoretically puzzling why dissidents committed to contesting a repressive regime through popular mobilization would intentionally organize parallel events that compete with, and reduce the size of, publicly advertised collective action. Yet organizers of jittered and takhf īf protests explained that they engaged in these tactics on behalf of the long-run growth of the movement at the expense of the short-run attendance at any one main event. The simultaneity of parallel events was thought to make regime repression less effective by giving participants additional time to evade security forces as officers shuffled from the main site to that of the secretly organized parallel one as well as to wear officers down over time.
From above, the Sudanese uprising appears to be a case that exemplifies the importance of SMOs and ICT in the way that existing literature expects: this protest movement was coordinated by an SMO that utilized ICT to spread details of its events to large numbers of dissidents. Yet I find that the way in which these coordinating tools were sometimes used differs from what past research would predict, and from how outsider observers described 2 -namely, that the FFC's publicity of the specific details of each protest event through ICT facilitated some informal groups to independently coordinate other events in their attempt to rattle the security apparatus.
This research is able to shed light on these hidden processes of mobilization by collecting data with ethnographic sensibilities (Fu and Simmons 2021). Many dissidents actively working against an autocracy purposefully do so away from the view of the state, and thus also outside the view of many of the systematic forms of data that scholars studying contention traditionally use. Scholarship hoping to understand how 1 For ease of exposition, I term events coordinated through widely known public information as "main" activities and coordinated discoordination protests as "parallel" or "simultaneous" activities.
2 For media coverage on conventional reliance on SMOs and ICT during the uprising, see "Sudan Trade Unions Call for March to Presidency as Protests Grow," Al-Jazeera, December 24, 2018; "The Revolutionary Force behind Sudan's Protest Movement? Doctors," New York Times, April 20, 2019; "'Facebook Protesters' Helped Sudan Drive Out Bashir," Wall Street Journal, April 12, 2019; "Sudan's Omar al-Bashir Mocks 'Facebook Protesters'," BBC, January 31, 2019; and "Tasgut bas' to #SudanUprising: How Social Media Told the Story," Al-Jazeera, August 12, 2019. dissidents oppose repressive regimes must continually determine new ways to analyze data that is systematically missing or intentionally inaccurate (King, Pan, and Roberts 2013;Mitts 2022) or learn from in-depth, real-time engagement from dissidents themselves about their subversive mobilization tactics (Fu and Simmons 2021). The latter approach also helps uncover "unidentified political objects" that other research methods have overlooked "but which are nonetheless meaningful for local political actors" (Jourde 2009, 201). I therefore ground my analysis in qualitative and ethnographic data collected before, during, and just after the uprising. I carried out more than one hundred interviews and focus groups with leading activists and dissidents, and I engaged in participant observation of the movement as it occurred. The majority of interviews and focus groups were with members of the informal neighborhood groups across Greater Khartoum that engaged in dis-coordinated mobilization. I also interviewed many of the elites in the FFC whose widely circulated calls for collective action were the basis for improvisation.
The data collectively paint a rich picture of the mechanics of mobilization under subversive environments, including the reliance on parallel, microprotests that datasets and media coverage of the Sudanese uprising missed. 3 Indeed, while other ethnographic accounts of contention under autocracy examine how state repression pushed dissidents to innovate completely new forms of collective action (Fu 2018), I contend that parallel protests are not a novel protest tactic that was invented and refined during the Sudanese uprising. Instead, coordinated dis-coordination may be considered a class of strategies that are common in protests movements, thus giving this paper broad external validity, however, one that is hard to see if we rely only on publicly available data.
Ultimately, a core implication of this paper is that we may be mis-attributing the role of any tactic that we assume can publicly or consistently organize disparate individuals under repressive environments. I find that dissidents implicitly recognize that public information meant to organize civilians against the regime is liable to be used by that regime to preempt mobilization. In turn, dissidents likely act on public information in more indirect and complex ways than unembedded researchers relying only on public information can observe. This finding has important methodological implications for research on ICT in particular. Technological advances now allow researchers to scrape large amounts of data to study the relationship between public information and mobilization. But much of this existing work does not acknowledge that some of the tweets and posts analyzed are purposefully written to be wrong and misleading or that dissidents may not follow coordination instructions as publicly conveyed. The role of public information in repressive environments is, unfortunately for unembedded researchers, more complicated than is convenient to assume.

THE COORDINATION PROBLEM UNDER REPRESSIVE ENVIRONMENTS
Coordinating mobilization under repressive environments is difficult. Tactical coordination conveyed through public channels or through common knowledge is liable to be repressed. Yet, while repression is meant to demobilize (Davenport 2007), research from these environments has shown how dissidents can continue to coordinate in anticipation of, or after innovating responses to, repression.
Perhaps most obviously, dissidents may coordinate secretly or informally. In contexts where the state tries to infiltrate or repress formal civil society, dissidents may tap into informal groups or interpersonal networks to organize (Opp and Gern 1993). These networks allow for the transmission of information internally without broadcasting to the regime. 4 And, when dissidents at the edges of different networks coordinate collectively, they can activate large swaths of society (Clarke 2014). In other cases, civil society organizations may facilitate disguised collective action of individuals making similar demands but acting alone when traditional collective action is too dangerous (Fu 2017;. Other research suggests that instances of collective action under repressive regimes may not be due to coordination at all, but instead, spontaneity. For instance, Pearlman (2021) examines "mobilizing from scratch" in Syria during the Arab Spring. Collective action could not be planned in advance because the Assad regime would have preemptively repressed it: instead, randomness was preferred because "the formulaic organized thing, it can be figured out" (1797). This stream of literature rests on the assumption that collective action requires only a minimal degree of structure that even heavily repressed societies still possess, whether that is "some sense of common identity; some sense of shared definitions of grievances and antagonists; some ability to communicate .. that [does] not depend on […] dense and enduring lateral relationships" (Piven and Cloward 1991, 443-4). Others document instances of collective action that rely on demonstration effects (Barrie and Ketchley 2018;Bishara 2018). Binding these ideas together is that contentious events are "not planned, intended, prearranged, or organized in advance of their occurrence" (Snow andMoss 2014, 1123).
Relatedly, research on the repression-backlash nexus suggests that mobilization after state repression of collective action is often a spontaneous, reflexive response. For instance, we see evidence of spontaneous backlash during the Hong Kong Umbrella Movement, when "[i]mages of police brutality spread through television and social media … after which thousands of protesters spontaneously marched" (Cheng and Chan 2017, 228). One's spontaneous response after observing repression on others may be due to the triggering of new emotions that, by nature, are immediate; or spontaneous mobilization can be the result of reflexive emotions that give way to more enduring affective orientations (Pearlman 2013;Young 2020). Other times, repression creates opportunities for collective action at "emotionally charged focal events" such as the funerals of movement participants (Almeida 2003, 354).
The shape of mobilization in recent years shows how widespread grievances might motivate mobilization without traditional mobilizing structures. Take the Hirak movements across the Middle East where there is no central SMO or charismatic individual directing participants in a vertical manner. Instead, the internet facilitates horizontalism which manifests as small, informal groups independently deciding to plan mobilization events in their locality that are under the broad goals of the movement (Schwedler 2022;Yom 2014;.

SITUATING COORDINATED DIS-COORDINATION
I argue that another way in which dissidents under repressive environments coordinate mobilization is through coordinated dis-coordination whereby groups of dissidents use public information-often distributed through SMOs or ICT, or implicitly known due to cultural repertoires or focal points-meant to coordinate widespread collective action to secretly coordinate other mobilization events. Coordination in the term "coordinated dis-coordination," then, refers to the organization of distinct collective action among dissidents outside the public, main event. Discoordination refers to dissidents' purposeful introduction of friction into the movement by eschewing public, widely expected parameters of coordination.
The friction-generating behavior in acts of coordinated dis-coordination have analogues to other contexts outside of mobilization, perhaps most notably to those in Zomia. Scott (2009) describes how "hill people" in Zomia have purposefully eschewed some technological innovations available within their society and that the state assumes people will unflinchingly adopt-for example, purposefully destroying bridges and cutting telephone poles (166), choosing swiddening instead of padi rice cultivation (193), and moving toward more "elementary" forms of social organization as opposed to adopting more complex ones (208).
Though unembedded outsiders may assume that these decisions are backward or primitive, Scott (2009) argues that these actions are a rational strategy by hill people to increase their distance from the encroachment of the state. The actions of people in Zomia teach us that we should not expect the linear or whole-scale adoption of efficiency-gaining technology; what hill people give up in efficiency they reap in other realms such as political autonomy.
I describe two distinct types of collective action events that fall under the banner of coordinated discoordination, though there may be others. Both are types of simultaneous events whereby dissidents use widespread information about the coordination parameters of a public protest (chiefly: date, time, type, and place) to privately organize other parallel collective action activities that occur elsewhere. First, some dissidents may "jitter" plans for an upcoming protest. Dissidents rely on public coordination information about upcoming collective action to plan events that explicitly change one parameter of coordination but retain the others. Given my scope conditions (see below), I speak specifically about jittering place such that we see a jittered event that is simultaneous to, but away from, the main event. Whereas information about the main site of collective action is broadcast widely, location information about a jittered event is only shared within a closed network of dissidents to prevent leakage to the security forces. The end result is collective action that is planned to be simultaneous to the original main event, but which occurs at a site away from it to prevent the immediate state repression of the jittered event.
Second, some dissidents may organize takhf īf protests to "lighten" repression on collective action that the regime is currently repressing, should they become necessary. Dissidents who intend to hold a takhfīf protest explicitly decide not to participate in an upcoming, widely advertised main event. Instead, they make plans for a contingent protest that they stand ready to launch if they receive information that violence is high at the main event. The goal of a takhf īf protest is to create a sufficiently large and disruptive simultaneous event that the regime must transfer some of the security presence from the heavily repressed site, in effect reducing the repression on dissidents there.
The main theoretical difference between these types of parallel events is their position toward expected regime repression. Dissidents planning jittered protests do so to alleviate expected repression on themselves whereas those planning takhfīf protests hope to draw fire on themselves and away from others. As such, we may expect differences in how information about parallel protests, once launched, is spread. Those organizing a takhf īf protest are apt to publicize their protest in real time so as to advertise it to the regime such that they require a response whereas, in theory, those organizing a jittered protest might shy away from concurrently posting about the event on social media. That said, the line between jittered and takhf īf protests can be blurry in practice. For example, we might encounter instances in which organizers of a jittered protest post about their event to draw more people but in effect, attract regime repression at the expense of the main event such that it acts as a takhf īf protest.
Despite their theoretical differences, jittered and takhfīf protests are similar in key respects. Most importantly, they are instances of dis-coordinated protest tactics that are only made possible by the public coordination of other mobilization activities; jittered activities explicitly rely on the public coordinating information of other events, so as to keep the most coordination parameters but choose a place away from the main site. And, public knowledge of upcoming events is necessary for organizers of contingent takhf īf protests to know when and where it might be necessary to lighten repression on others.
Another reason to group jittered and takhf īf protests together underneath the banner of coordinated discoordination is that the decision to hold each type of collective action is a bottom-up undertaking. There is no central SMO coordinating parallel protests, and different groups of dissidents planning parallel events may not be in contact with each other. Instead, different groups independently discern the pitfalls of public coordination-namely, increased legibility to the state and thus more effective state repression-so they introduce friction into their actions. One implication is that we may see multiple, separate jittered and/or takhf īf events organized independently of each other in response to a singular main event. Another implication is that we might see an overlap in those involved in each type of parallel protest, as those organizing a jittered protest one day may decide to organize a takhf īf protest the next.
Coordinated dis-coordination differs from the past accounts of mobilization under repressive environments described. Secret mobilization through informal groups is thought to be flexible on all coordination parameters. However, coordinated dis-coordination requires some public information about upcoming mobilization events for dissidents to secretly improvise some coordination parameters away from. And, though both coordinated dis-coordination and disguised collective action are tactics used in repressive environments, dissidents of coordinated discoordination are not hiding their mobilization from the regime as in Fu (2017;: indeed, those organizing takhf īf protests, in particular, are intentionally trying to attract the attention of the regime. Further, the advanced planning of both jittered and takhf īf protests differentiates them from spontaneous collective action and backlash repression. Even though takhf īf protests are responses to repression, they are not reflexive as literature on the repression-backlash nexus would suggest, but instead explicitly premeditated. At the same time, what I describe has parallels to recent leaderless movements in that small groups of dissidents are organizing events under the broader banner of a movement. But overall, leaderless movements are orthogonal to what I describe since they are differentiated by their lack of a central leader or SMO, whereas the tactics I describe are less about the chain of command and instead more concerned with the tactics that grassroots activists engage in.

Coordinated Dis-Coordination Scope Conditions
I stipulate two scope conditions for coordinated dis-coordination. First, coordinated dis-coordination is limited to movements in which mobilization has not cascaded. Mass mobilization, once triggered, reduces repression by leading to the defection of the security forces (Chenoweth and Stephan 2011); however, dissidents come to expect regime violence in movements that have not (yet) cascaded. Dissidents who do not expect sufficiently large crowds at the main event such that security forces may potentially switch their allegiance away from the regime are more apt to perceive benefits from parallel events.
Second, dissidents confronting a weaker security apparatus are more likely to think that acts of coordinated dis-coordination are useful. Takhfīf protests will be perceived as more disruptive when the regime does not have sufficient officers such that it is forced to shuffle officers from protest to protest, whereas jittered protests rest on the assumption that the regime does not have the capacity to lay an entire city under siege. 5 And security forces that can infiltrate society, including the informal groups where parallel protests are planned, can pre-emptively repress simultaneous protests and their organizers. We should expect dissidents in regimes with stronger security apparatuses to follow public coordination to a greater extent (Carter and Carter 2020;Chau, Hassan, and Little 2022) or innovate other forms of collective action altogether (Fu 2017;.

DATA AND METHODOLOGY
This paper examines coordinated dis-coordination within Greater Khartoum during Sudan's 2018-19 uprising. Though the movement began in the periphery of the country, mobilization in Greater Khartoum, a city with an estimated population of as many as 8 million by 2018, was the most intense. 6 I focus on mobilization from December 25, 2018 to April 6, 2019, a period I call Phase 2 of the uprising (see Figure 1). These months saw near daily small-and medium-scale protests with no event larger than five thousand people, and most involving substantially fewer (generally in the hundreds). I date the beginning of this phase to the first coordinated protest of December 25 and it continues until a milyūniyya (million-person protest) on April 6 of hundreds of thousands and a sit-in outside the military's headquarters. In response to this mobilization cascade, security forces ousted al-Bashir on April 11 and began negotiations with the FFC for a transition to civilian rule. Mobilization in Phase 2 differs from Phase 1 as it was coordinated, and no longer sporadic and unorganized. Phase 2 differs from Phase 3 as this is when protests cascaded into mass mobilization.
Given Greater Khartoum's importance for this paper, I provide basic information about its geography. "Greater Khartoum" actually refers to three cities split by the Blue Nile, White Nile, and confluence of the two. These three are Omdurman, Bahri, and Khartoum proper. I refer to the tri-city capital as "Greater Khartoum" and the portion of Greater Khartoum that is in-between the Blue and White Niles as "Khartoum." As in any large city, Greater Khartoum has individual neighborhoods that are well defined by large cross streets and important landmarks. Figure 2 provides an overview of Greater Khartoum, coloring in different neighborhoods that are referenced in this paper.
I draw on more than one hundred formal interviews and focus groups, as well as participant-observation data that include informal conversations and participation in anti-regime mobilization. The majority of the data was collected during five months of in-person field research before, during, and after the uprising. Due to the COVID-19 pandemic, additional field research was conducted over the phone. Quotations are either from recorded interviews or verbatim transcriptions taken in the moment. Unless specified, in-person interviews and focus groups occurred in Greater Khartoum. Participant-observation data is drawn from dissident meetings about coordination, formal and informal opposition group meetings, as well as participation in and observation of large-scale mobilization activities. Interviews and focus groups were semistructured in nature and found using snowball sampling. The vast majority of interviewees were active participants in the uprising. Others were elites who coordinated action including leaders of the largest opposition parties, labor

FIGURE 2. Map of Greater Khartoum
Note: This map gives the geography of the capital, Greater Khartoum. Shaded are the neighborhoods from which interview and/or focus group data were collected, or explicitly referenced.
unions, the FFC, and other prominent civil society groups opposed to the regime.
I also interviewed and conducted focus groups with individuals who took the lead in dis-coordinatingleaders of nearly a dozen neighborhood resistance committees, or lijān, from across Greater Khartoum (see below). I used a convenience sample to choose lijān for two reasons. First, not all neighborhoods were equally active during the uprising (el Gizouli 2020). By choosing on more active areas, as opposed to sampling neighborhoods based off pre-existing characteristics, I was better able to understand the dynamics of mobilization by those who played an outsized role. Second, and given Sudan's authoritarian history and dissidents' hesitancy to speak candidly, I chose lijān in which I had pre-existing contacts and thus could expect more open discussion. I attempted to interview at least two people separately from each lajna. Since these interviews were meant to gauge behavior during more routine events and processes, verification was less about timelines and specific events but instead about the standard operations of a lajna and individual or group justifications for those actions. Other interviews are from participants in the uprising who were not members of formal SMOs or informal resistance committees.
Given that interviewees and focus groups emerged from pre-existing contacts, there is a bias toward middle-and upper-class parts of the city as well as residents who have spent more time in the city. That said, I was particularly cognizant of sampling participants from a wide range of social classes and demographics (age, gender, and region of origin) for individual interviews.
I provide additional fieldwork procedures and an ethics statement in the Supplementary Material, including a discussion of interviewee safety in light of the October 2021 counterrevolutionary coup. 7 I note that the evidence below omits identifying information outside neighborhood name (and then only provides it if interviews expressly allowed it). Further, recorded interviews only took place after the uprising cascaded.

Al-Bashir's Autocratic Regime and Past Attempts to Mobilize Against It
Omar al-Bashir took power in a 1989 Islamist-military coup and ushered in the Inqāz ("salvation") regime. This autocracy tried to reform Sudanese society through the imposition of formal laws (Massoud 2013) and informal norms that reflected only the traditional cultural norms of the country's Arab Muslim population (Salomon 2016, 64-5). The regime largely failed at transforming society into its ideal, but it did succeed in creating real discontentment among the population, including across Greater Khartoum. Despite widely held grievances, however, nonviolent organized resistance was limited, in part, because the regime had succeeded in dismantling many of the formal civil society groups that had led mobilization during prior popular uprisings (Berridge 2015) including independent NGOs (Kadoda and Hale 2015, 221), professional unions and associations (Massoud 2013, 121), and the civil service (Mann 2014, 569-70). Further, the regime sowed discord among the formal groups that managed to persist, preventing their collective coordination. 8 That said, the 2005 Comprehensive Peace Agreement set into motion factors that would eventually change Greater Khartoum's civil society landscape. Alongside ending the Second Sudanese Civil War, the agreement mandated a nation-wide opening of the country's civil society space. To be sure, changes among formal civil society were not felt immediately, and most organizations initially remained weak. Most new organizations created during this period were fronts for the regime that attempted to capture the influx of western money (Massoud 2013, 165). The minority that were independent were small and, since they relied on western resources, often tailored programming to broader western goals instead of the particulars of the Sudanese situation (Massoud 2011, 17).
The weakness of civil society continued even in the midst of widespread grievances and regional protest waves nearly a decade later. The most threatening episode of nonviolent mobilization against al-Bashir prior to 2018-19 occurred in September 2013. Yet these protests showed a "lack of coordination" (Langlois 2022(Langlois , 1315; they were spontaneous and unorganized, amounting to "loosely coordinated rioting and demonstrations" (de Waal 2013) that were fairly limited in size. Further, the 2013 uprising showed the limits of public-facing, many-to-many ICT for anti-regime mobilization. Though social media increased connectivity among dissidents, the regime had "penetrated cyberspace and [used] various tactics to monitor and divide the opposition. Intelligence agencies [turned] social media against its practitioners" (de Waal 2013); "an overreliance on social media had provided an entry point for government infiltration" (Branch and Mampilly 2015). Protests quickly fizzled out as the regime responded with severe repression that left more than 200 dead.
During the last few years of al-Bashir's regime, however, the openings in the wake of 2005 began to bear fruit for civil society. In light of the repression against political mobilization in 2013, some of the most impactful organizations were informal and advocated for seemingly nonpolitical issues such as seasonal Nile flooding (Marovic and Hayder 2022, 6). 9 At the same time, many of these informal groups continued to build linkages with each other and formal civil society (Medani 2021, 215;2022, 258).
Most important for the analysis below, independent unions (e.g., for doctors, journalists, and lawyers) joined together to create the Sudanese Professionals Association (SPA) in 2016. The regime engaged in relatively less repression against the SPA than its predecessors because the group's initial aims centered squarely around economic grievances. 10 For reasons of organizational survival, and similar to the not explicitly political demands made among informal civil society, the SPA's initial expressed goal was a raise to the minimum wage. 11 In sum, by late 2018, Sudan's civil society landscape looked much different than it did in the run-up to prior periods of contention. While civil society was still relatively weak in that it was unable to begin widespread collective action against the regime, denser bonds between informal groups and across civil society allowed for this sector to quickly mobilize once contention was sparked elsewhere.

Initial Coordination of Mobilization in 2018-19
A new protest wave began in December 2018 (see Figure 1), with spontaneous and uncoordinated protests initially. 12 To both on-lookers and participants alike during the study period, it was clear that mobilization had not cascaded until April 6, 2019.
The SPA became integral for coordinating collective action events for the period between late December 2018 and the cascading of mobilization some 4 months later. This body's prominent role was fortuitous. The SPA had long planned to hold a rally against the country's low minimum wage on December 25, 2018. After protests in the periphery gained traction, however, they converted this planned and already advertised economic protest into a political one that demanded the end of the Inqāz regime. 13 The protest was advertised through social media and spread through ICT. 14 This public coordination helped make it the largest rally to date against the Inqāz regime with some five thousand participants. For these same reasons, it was also violently dispersed by regime forces. The SPA publicly coordinated another rally on December 31, and saw similar turnout-and regime violence-as before. Though they were quickly put down, the December 25 and 31 protests proved important in solving the initial political collective action problem: a large portion of Greater Khartoum felt similarly about the regime and was willing to participate in collective action. 15 Further, the substantial, if not overwhelming, turnout at these two events helped streamline coordination of the movement behind the SPA. By January 1, 2019, the SPA and other formal organizations signed the Declaration of Freedom and Change. Signatories joined together to form the FFC, a negative revolutionary coalition (Beissinger 2013) dedicated to bringing about the end of the Inqāz regime. 16 For the remainder of the uprising, the FFC attempted to solve the tactical coordination problem by publicly coordinating mobilization through manyto-many technologies. The FFC would plan and publicize collective action by posting flyers of scheduled events on social media, at times even posting route maps for marches (left-and right-hand sides of Figure 3). By January, this planning was done through the FFC's Field Committee which created weekly schedules that laid out planned uprising activities (middle image of Figure 3). Knowledge of these protests was ubiquitous across the capital.
Aside from the coordinating activities of the FFC, much of the mobilization in Greater Khartoum was done through informal groups. Individuals within some neighborhoods independently organized their neighborhood's residents through a lajnat muqāwama (resistance committee), or simply a lajna (plural: lijān). 17 They are similiar to the tansiqiyat behind Syria's uprising (Pearlman 2021) and local hirak groups in Jordan (Yom 2022). Many of these grassroots organizations emerged organically after the uprising began, or reconstituted themselves after forming in 2013 but had laid dormant since. Each lajna looked different, reflective of the socioeconomic and ethnic heterogeneity of each neighborhood (el Gizouli 2020, 4). Yet there were some broad commonalities. They emerged due to the inherent quotidian ties that exist within neighborhoods. These quotidian ties often crossed organizational ties and it was not uncommon to find lajna members who belonged to different political parties or formal civil society groups. They tended to have various subcommittees (e.g., social media and security) that were in charge of different mobilization elements. And, they worked secretly and tended to be cautious with extending membership so as to prevent infiltration by 10 Interview with Central Committee of Sudan Doctors member, February 9, 2020, Khartoum and by phone. 11 Interview with civil society leader, December 2, 2019; Discussion with SPA Spokesperson and Leader, October 1, 2019, Washington, DC; Interview with FFC Coordination Committee member, October 11, 2019, Washington, DC. 12 Field notes, December 2018. The first event during this period of contention is contested. Some point to al-Fashir on December 12, 2019, others to ad-Damazin on December 13. However, perhaps the most important of these initial collective action events occurred in Atbara on December 19; what began as an economic protest quickly turned political as locals burned down the NCP's local headquarters. Images of this act were spread widely across social media helped mobilize collective action in Greater Khartoum (Field notes, December 2018). 13 Interview with FFC Coordination Committee member, May 19, 2019; Interview with SPA leader, October 6, 2019, Berkeley, CA. 14 Field notes, December 2018. 15 Ibid. 16 The FFC collectively decided on major mobilization actions; however, the SPA was considered the main mouthpiece of the uprising. As indicated by some interview evidence, many citizens referred to main protests as organized by the SPA, not the umbrella FFC. I refer to mobilization as coordinated by the FFC unless I quote an interviewee who specifically cites the SPA. 17 On the importance of lijān muqāwama during the Sudanese uprising, see, for instance, Berridge et al. (2022, 25), Hassan and Kodouda (2019, 99), Marovic and Hayder (2022, 10), and Medani (2022, 256). undercover regime agents. For example, members would perform background checks on less well-known neighbors who wanted to join. 18 Despite their informal nature, it is useful to consider each lajna as a subgroup of the larger protest movement under the FFC. In fact, each lajna member I interviewed reiterated that they saw their committee as a local, self-appointed, independent branch of the movement that followed the FFC's lead: though their efforts to mobilize were determined locally, each lajna did so in service of the uprising and took their larger cues from the FFC. For instance, many lijān would print flyers with SPA logos and use SPA chants. Different lijān had different relationships with the FFC based on pre-existing ties, but even those with the strongest ties operated fairly independently with regard to mobilization. Instead, stronger ties generally manifested in more material resources (e.g., spray paint and paper) for the lajna. 19

Regime Repression
By 2018, the Sudanese security apparatus was weak. The regime's precarious financial situation had perceptibly lowered material capacity and the patronage links between the regime and officers. 20 Yet, despite this weakness, the FFC's public coordination of protest activities facilitated repression by the regime's security apparatus, eventually leading to the death of one hundred to two hundred protestors and the detention of around two thousand protestors by April 2019 (Dahab et al. 2019;Medani 2021, 208).
Dissidents internalized this high level of violence as connected to the public coordination of mobilization. One activist explained that the use of schedules and flyers meant that, "the security apparatus would know to close [a scheduled protest] area in advance." 21 Numerous protestors discussed the high levels of repression they faced at planned protests since their actions were so predictable: "just getting in, your appearance [as a protestor] is clear … Even before the protest starts, they are arresting people." 22 One activist discussed how he attended a protest in which the security forces began arresting people in the planned area if they were checking their phone too often before the protest start time. 23 As one protestor explained: "The thing is, [mobilization coordination] wasn't secret anymore. [The FFC] would say we should protest at 1pm. But then the police would be waiting for us at that place." Recalling a specific incident, he continued: "when I got there, the area was full of [officers] who were apprehending anyone who was walking there." 24 FFC elites realized that publicly coordinating protests made regime repression more likely. But they were committed to this strategy. On principled grounds, and in line with the logic of political jiu-jitsu (Sharp 1973), one of the SPA's leaders explained that "the idea was [to] challenge [the regime in public]. We are not hiding. We are calling for everyone to come out and protest out in front." 25 But more practically, and while there were many differences in context between the uprising of 2013 and that of 2018-19, participants and observers alike had internalized a narrative of the failure of the previous uprising as due to the unorganized nature of contention and the disunity of the movement (Berridge et al. 2022, 8). The importance of coordination was firmly believed across the movement even as some individuals and groups intentionally dis-coordinated action. 26

Dis-Coordination through Jittering
Jittering involved activists relying on the FFC's widely circulated calls to preemptively organize protests that changed the locale from that of the main protest. In this way, though the FFC coordinated specific types of collective action in specific places at specific dates and times, activists would sometimes mobilize their neighborhood for parallel events that resembled an FFC call in date, time, and type-but not in place. Jittered protests thus improvised off publicly coordinated events and FFC public coordination was necessary for their existence even if the calls were not fully followed. For example, different lijān explained that, as soon as the FFC released a new calendar for the upcoming week's protests, they would strategize their neighborhood's mobilization activities which often included a mix of following some of the FFC's calls or deciding to hold a parallel jittered (or takhfīf) protest. 27 One FFC leader who helped plan the uprising's main protests confirms that jittered protests were done without direction from above: referencing an early protest during the uprising, "when that [main] protest was hit … Burri had a protest on their own … they thought of this on their own. There was no coordination with the FFC." 28 Similarly, one protestor observed: "Sometimes, neighborhoods would hold their own protests … whenever there was a big protest in the center there would [also] be parallel ones in infamous neighborhoods." 29 Another protestor recalls: "I remember one time I went to the center of Khartoum [for a main protest but] … it was repressed by the police and they arrested a lot of people. When we got back into our cars, amid the chaos and the teargas, we found that there were huge protests [in various neighborhoods across the capital]." 30 Instead of direction from above, jittered protests were coordinated through small groups of activists with dense ties. For instance, a leader of a lajna in Bahri recounted jittering the place of FFC-planned protests and only communicating the details to others in the neighborhood (who were not in the lajna but whom he trusted). 31 Other times coordination spanned neighborhoods: we would agree with neighboring lijān. Say the [main] protest is at 1pm … then all of the neighboring lijān agree to meet somewhere and march together elsewhere … this way, there would be two large protests. One we made, and since many people went to the [FFC's advertised site] on their own, there would still be a protest there. 32 Similarly, a member of another lajna described a particular instance: "[The FFC] asked us to go to Qurashi Park [for a central protest]… we decided that we would not meet at Qurashi Park but instead by Zinc [restaurant] … We decided this with the other lijān in the area." 33 Jittered protests were not spontaneous. They were the result of careful planning that relied on, but improvised away from, public information and were "outside the realm of the SPA." 34 Dissidents shared plans for jittered protests across dense bonds through a variety of methods including word-of-mouth, one-toone ICT, and private WhatsApp groups. However, even if ICT were used to spread information about jittered events, it was done as much as possible away from the watchful eye of the regime. For instance, dissidents might share the information via WhatsApp only among a small group of neighbors whose antiregime status had all been verified as opposed to the larger WhatsApp group for the entire neighborhood. 35 When asked why they organized parallel protests, dissidents explained that jittering was a rational response to expected repression at main, publicly advertised events. As one lajna leader recounted: When the SPA said [to protest at] a place [like the Central Transport Station], you would go and find the security apparatus already there. After a while … we took it to mean areas next to the Central Transport Station, not the Central Transport Station itself … because you're not going to walk straight into a fire, right? 36 Another local activist stated, "if the SPA calls for a protest, we would organize other ones because the ones that are announced are the ones that are most probably going to [get repressed]. So our idea was, we should always, for the day something had been announced …. explained that the site of a jittered protest was often done with repression in mind as well. Members of a lajna's security subcommittee might scope out potential areas for a jittered protest days in advance to gauge the presence and routines of security forces in the area.
Further, there was a belief that jittered protests would help increase participation. Dissidents expected the lower levels of repression to attract individuals who happened to be in the area. And separately, individuals whose risk tolerance was not high enough to face regime repression head on-that is, "walk straight into a fire"-but who received private information about a jittered protest beforehand might be willing to attend the parallel event. In the long run, leading activists expected these new recruits' willingness to engage in higher-risk activism to grow over time, similar to processes laid out in McAdam (1986). 38 Some lajna leaders discussed jittering the date or time of a protest, though these tactics were described as less common than jitters of place. Members of different lijān confirmed that it was within their repertoire to alter the time of a local neighborhood protest by pushing the protest back a few hours or converting an FFC-publicized daytime protest to an evening protest. 39 Changes of date occurred too: Let's say the schedule said that there is supposed to be a neighborhood protest [in my neighborhood] today … but instead, we [as a lajna] might decide to go out tomorrow or the day after …. for instance, if the schedule says that we should have night protests on Sunday then the lajna would say, we're not going to go out on Sunday. Not even on Monday, but Tuesday. 40 Another commented, "say the SPA advertises a protest for a Thursday. There is no advertised protest on Wednesday and Tuesday … But then we would agree among ourselves … to hold a protest instead on that Tuesday or Wednesday." 41 Jittering date or time was thought to make a protest safer since the security forces could not plan for this mobilization. Yet, since these protests were not concurrent with main events, they did not benefit from the reduced security presence of a simultaneous protest that only jitters place.

Dis-Coordination through Takhfīf
Takhfīf involved activists preemptively planning protests or contentious actions that they might launch in order to reduce the realized repression on widely advertised protests should this distraction become necessary: activists who were not planning on attending a main protest would instead organize contingent simultaneous protests that drew security forces away from an ongoing protest. As one activist explained, "when a protest starts and fails because of repression, then people in other places and other neighborhoods go out instead." 42 In this way, takhf īf protests were explicitly meant to reduce repression on other areas: "the idea was that instead of the security forces all concentrated in [the main site of an FFC planned protest], the security forces have to split up." 43 Takhfīf protests looked different from the FFC's planned protests as well as jittered events. To begin, takhf īf protests were often small and short-lived. In some cases, they amounted to minor acts of disruption at well-trafficked intersections or near important focal points that would require an immediate police response: One protestor explained, "If a protest in, say, al-Abassiya started, and you could tell that it was becoming brutal, then people would start random acts of [traffic] obstruction in their own area." 45 At first blush, these may seem likely instances of distraction that arose spontaneously. And, indeed, there were instances of spontaneous protests that were launched to distract officers. But the takhf īf events I describe are different in that dissidents planned for their possibility in advance. Dissidents used the FFC schedule to anticipate when a takhf īf protest might be necessary and sketch out organizational details for these contingent events. For example, the majority of a lajna might attend a scheduled protest while delegating some members to stay behind and mobilize a backup takhf īf protest should the main protest fail, thus providing cover for lajna members at the main site to escape. 46 Or some members of a lajna might collect tires and bricks to build a roadblock in anticipation of a takhf īf protest or diversion. At the same time, these protests were contingent: the day's FFC-sanctioned protest might not see much repression, whereas a secretly coordinated simultaneous jittered protest might. As one protestor explained, "you never knew if the other protests that day would be a success or not, if they would need our help or not. The contingent nature of takhfīf protests meant that ICT were integral to their launching as they allowed for the quick transmission of information about the current protest environment and the level of repression across the city. Multiple organizers explained that some of this advanced planning involved someone on the administration or social media subcommittee of their lajna staying home during a main protest to relay information to other lijān via one-to-one ICT should the level of repression make a takhf īf protest necessary. 48 Another activist laid out the use of many-to-many ICT for this purpose: Say there was a neighborhood protest and the police came, and the crack-down on them was violent. When the violence is really high in an area … then people from the area go on Facebook … Speaking about a protest in his neighborhood that was particularly repressive, an activist explained: "when we realized the pressure was really intense, the admins posted on the Facebook page [of our lajna]. And almost immediately, people in [another neighborhood] went out." 51 In turn, activists explained that they expected neighborhoods to ask for help in reducing repression, so they themselves were often looking for calls for help on social media to know where or not to launch their contingent event: "The thing is, on days when you know there are [FFC scheduled] protests, you don't just follow the SPA page. You follow the pages of other lijān and prominent people [in the uprising], and you look for [Facebook] Lives." 52 Takhfīf protests tended to occur near the repressed protest, and were often spearheaded by lijān in neighborhoods abutting the repressed site. This was, first, because of social reasons and the high level of social capital in adjoining neighborhoods. 53 These were individuals who they cared deeply about and for whom they were willing to risk repression. Second, there were also strategic reasons to hold protests in nearby areas since participants believed takhfīf protests would be more effective if they drew security forces away from an on-going event. If a takhf īf protest is launched sufficiently far from the repressed site, then, dissidents described, the regime might send a new squadron of officers instead of redirecting officers from the main site. Consider this description by one member of the Abrov lajna: These dissidents believed that they benefited from keeping dissidents in nearby neighborhoods safe so that these other neighborhoods could provide them cover when their protests were repressed. For many dissidents, taking on the short-term risks of drawing regime repression during a particular takhf īf protest was worth the potential long-run effects of sustaining mobilization in nearby areas, and thus the uprising as a whole.

Regime Adaptation
There is evidence that the Sudanese security apparatus began to adapt its repressive response in light of dissidents' usage of parallel protests, in-line with ideas conceptualized by McAdam (1983). 55 For one, different security apparatuses began to increase the number of officers deployed at any one time, and assigning some officers who were not on active duty to camp out in front of important intersections and buildings. This helped the regime in two ways. First, with more personnel available, officers did not have to shuffle from one site to another and could better repress simultaneous events. Second, even if security forces were still surprised by a parallel protest, there would be less lag time in responding as there were officers ready to be deployed from major cross streets. 56 48 Focus group with members of Wad Nubawi neighborhood committee, December 9, 2019. 49 Phone interview with Abrov lajna member, August 24, 2020. 50 Interview with civil society activist, December 20, 2019. 51 Phone interview with Bait al-Mal lajna member, April 13, 2020. 52 Ibid. 53 Focus group with members of Wad Nubawi neighborhood committee, December 9, 2019. 54 Phone interview with Abrov lajna member, August 24, 2020. 55 I cannot empirically validate the extent to which these regime adaptations "worked" in stopping coordinated dis-coordination protests; however, dissidents conjecture that they were not broadly effective. Instead, I aim to demonstrate that just as protesters innovate within repressive environment, tactics of repression gradually adapt in turn. 56 Field notes, July 2019. Separately, some dissidents noted that periods in which their lajna put on numerous parallel protests or parallel protests that happened to grow substantially large tended to be followed by periods of repression in which security agencies preemptively sent officers to their neighborhood in anticipation of a simultaneous protest. 57 When the regime could predict where discoordinated protests were going to take place, then dissidents saw substantially less benefit in engaging in this tactic.
Further, different activists explained that they perceived subtle shifts in the regime's attempt at infiltration, indicative of a growing recognition about the informal organizational nature of this uprisingwhereas the security apparatus was invested in infiltrating formal civil society groups during the first few weeks of the uprising, the regime also began trying to infiltrate lijān as time progressed. 58 CONCLUSION This paper describes tactics used to coordinate mobilization under repressive environments. Using in-depth evidence from the Sudanese uprising, I find that dissidents at times relied on public coordinating information -released by the main SMO and disseminated widely through ICT-to instead plan other collective action events that occurred simultaneously but in a different locale. In this way, independent organization of these parallel protests was dependent on publicly available coordination information, even though this coordinating information was intentionally not fully followed.
While past work has not devoted much space to these tactics theoretically-likely because of the difficulty in observing hidden processes of mobilization (Fu and Simmons 2021)-that does not mean that we should consider parallel protests as first innovated by Sudanese dissidents. Instead, the specific tactics described here likely have analogues elsewhere. For instance, take descriptions of mobilization in Egypt in the decade prior. While tens of thousands followed publicly advertised calls, some activists separately "[began] a 'secret protest' which would not be publicized, two hours earlier; this [allowed] them to … confuse and divert the security forces" (Clarke and Kocak 2019, 12). Kadivar and Ketchley (2018, 10) focus on collective action of a different type (unarmed collective violence such as throwing rocks); however, one benefit of the parallel events that they describe is that they "can protect peaceful protestors by diverting a state's repressive forces away from frontline protest policing duties while disrupting the coercive capacity of an authoritarian regime." We should expect coordinated dis-coordination across movements outside Sudan, however, not across all instances of anti-regime mobilization. Returning to this paper's second scope condition, dissidents organizing against a strong regime will likely find that the resulting friction from dis-coordination is not worth the gains because the regime can quickly adapt to dissidents' innovations (Carter and Carter 2020;Chau, Hassan, and Little 2022). But the illegibility and unpredictability of parallel protests can yield numerous positive outcomes for a movement in cases like Sudan where the security apparatus is weak. Ping-ponging between protest sites may "turn around the security forces," 59 make them "exhausted," 60 "trick the security forces," 61 and "tire the officers, their energy and morale" 62 -eventually leading to officer burnout. 63 Or the unknown location of parallel events may facilitate recruitment: individuals who happen to be at the site of a simultaneous event find the costs of joining this protest much lower than those at the main site. 64 I leave it to future research to examine these extensions since the qualitative and inductive nature of the methods I employ cannot systematically test different mechanisms by which coordinated dis-coordination may succeed in cascading mobilization.
Though beyond the focus of this paper, I note that future movements in contexts that have used parallel protests may continue to have coordinated discoordination within their menu of tactics, but we should expect further innovations by dissidents on-and regime adaptions to-this tactic. The external validity of this tactic should, therefore, not be measured solely on the extent to which movements rely on jittered and takhf īf protests alone, but instead, tactics that use publicly advertised or widely held expectations of upcoming collective action to instead engage in other mobilization activities meant to reduce repression on the movement as a whole. Coordinated discoordination should be viewed as a larger class of strategies that dissidents use to introduce friction into their mobilization efforts in the hopes that this purposeful eschewing of efficiency makes the regime's repressive response relatively worse off than the movement.
In conclusion, the broader theoretical point I draw attention to is that we should expect dissidents to balance the ability of any strategy to streamline coordination against the expectation of making their actions more legible to the regime they are trying to evade or undermine. Specifically, with regard to 57 Focus group with members of Wad Nubawi neighborhood committee, December 9, 2019; Phone interview with Bait al-Mal lajna member, January 12, 2021. 58  research on collective action in our digital age, that so many dissidents rely on ICT for mobilization means that the negative legibility costs of new technologies are surmountable (Chau, Hassan, and Little 2022); however, the way in which ICT is used may be distinct from their original, assumed purpose or different from what is easily observable from the outside. 65 While we may observe correlations in reliance on public communication platforms and protest activity, we should interrogate this relationship before assuming that mobilization occurs solely through processes that are readily observable and channels that are easily researchable.

SUPPLEMENTARY MATERIAL
To view supplementary material for this article, please visit https://doi.org/10.1017/S0003055423000291.