How critical junctures shape secessionist movement cohesion: Strategies, framing processes, and interorganizational relations before and after the 2017 referendum in Catalonia

Gunzelmann, Hans Jonas Article — Published Version How critical junctures shape secessionist movement cohesion: Strategies, framing processes, and interorganizational relations before and after the 2017 referendum in Catalonia Journal of Peace Research Journal of Peace Research Provided in Cooperation with: WZB Berlin Social Science Center Suggested Citation: Gunzelmann, Hans Jonas (2025) : How critical junctures shape secessionist movement cohesion: Strategies, framing processes, and interorganizational relations before and after the 2017 referendum in Catalonia, Journal of Peace Research Journal of Peace Research, ISSN 1460-3578, Sage, Thousand Oaks, CA, Vol. 62, Iss. 4, pp. 961-975, https://doi.org/10.1177/00223433241258368 This Version is available at: https://hdl.handle.net/10419/310917.2 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ https://doi.org/10.1177/00223433241258368 Journal of Peace Research 2025, Vol. 62(4) 961 –975 © The Author(s) 2024 Article reuse guidelines: sagepub.com/journals-permissions DOI: 10.1177/00223433241258368 journals.sagepub.com/home/jpr Introduction The comparative literature on secessionist conflicts often treats pro-independence movements as unitary actors (Griffiths, 2016; Griffiths and Wasser, 2019; Siroky and Cuffe, 2015; Sorens, 2012), but it has also been acknowledged that movements in fact are often comprised of a multitude of collective actors, ranging from political parties and civic organizations to armed militias and terrorist groups (Cunningham, 2014). Whether these actors work together in the pursuit of independence is an important question. Previous research found that whether a movement is united or divided is decisive for the trajectory and outcomes of self-determination conflicts (Cunningham, 2011, 2014; Krause, 2017; Pearlman and Cunningham, 2012). What drives the cohesion and fragmentation of secessionist movements? Given the importance of the internal structure of opposition movements, research on its causes has been growing (Fliervoet and Seymour, 2023: How critical junctures shape secessionist movement cohesion: Strategies, framing processes, and interorganizational relations before and after the 2017 referendum in Catalonia Hans Jonas Gunzelmann Center for Civil Society Research, WZB Berlin Social Science Center, Germany Abstract What drives the cohesion of secessionist movements? Previous research emphasized the role of internal and external factors but produced mixed results regarding their effects. This article advances scholarship on this question by examining the role of critical junctures as periods of heightened contingency that can shift movements towards fragmentation or cohesion. It focuses on independence referendums and how states respond to them as important critical junctures, and on how they shape interorganizational relations as a key dimension of movement cohesion. Empirically, it explores the effects of the 2017 referendum in Catalonia using a mixed-methods research design that combines qualitative inquiry with network analyses of protest event data. The network analyses showed that the movement was notably less cohesive in the protest arena after the referendum than during the referendum campaign. Qualitative materials were employed to inductively identify strategy framing processes as key mechanisms to explain this development. Frame alignment around the referendum as a shared goal led to more cohesion during the campaign. After the event, a frame dispute over the meaning of the referendum led to diverging strategies and fragmented the movement, as state repression limited the movement’s room for maneuver. The findings suggest that research on secessionist movement cohesion should pay more attention to critical junctures and how secessionists make sense of them. Keywords Catalonia, fragmentation, frame alignment, independence referendum, networks, secessionist movements Corresponding author: Hans Jonas Gunzelmann, Center for Civil Society Research, WZB Berlin Social Science Center, Reichpietschufer 50, 10785 Berlin, Germany. Email: [email protected] 1258368JPR0010.1177/00223433241258368Journal of Peace ResearchGunzelmann research-article2024 Regular Article 962 journal of Peace Research 62(4) 168). Previous research has highlighted that cohesion is the product of ‘dual contests’ within the movement and with the host state (Bakke etal., 2012; Seymour etal., 2016) but has produced mixed findings on the precise effects of these contests (e.g. Cunningham, 2014; Fjelde and Nilsson, 2018; Lawrence, 2010; Mosinger, 2018). Examining these dual contests during moments of critical contingency promises to shed more light on this question. While the analytical importance of critical junctures and transformative events has long been acknowledged in comparative politics and elsewhere (Capoccia and Kelemen, 2007; Collier and Collier, 2002; Mahoney, 2002; McAdam and Sewell, 2001; Sewell, 1996; Wagner-Pacifici, 2017), the literature on secessionist conflicts has rarely considered them as an explanatory factor. Seymour etal. (2016), for instance, found that the onset of civil war increased movement fragmentation. McLauchlin and Pearlman (2012) showed that ‘repression shocks’ can both enhance and decrease movement cohesion, depending on the organizations’ satisfaction with the movement’s internal structure. Zooming in on such critical junctures has the potential to reveal the mechanisms that drive cohesion and fragmentation. This article focuses on independence referendums and how states respond to them, which arguably represent the most important critical junctures in secessionist conflicts besides the onset of civil war and the declaration of independence. Referendums are a key strategy for secessionist movements (Cortés Rivera, 2020; Sanjaume-Calvet, 2021). They open up opportunities for civil society actors to mobilize independence supporters and organize protests to put pressure on the host state and attract the attention of the international community (Della Porta et al., 2017). As such, they may foster cohesion among protest groups. However, referendums in liberal democracies rarely lead to independence (Dion, 1996). Instead, they often produce a backlash from the host state (Basta, 2021; Ferreira, 2021). The failure to achieve independence may also demobilize supporters (Lecours, 2018). Counter-secessionist repression and demobilization may in turn create conflict within the secessionist movement (Gade, 2020; Gunzelmann, 2022; Krause, 2023). Independence referendums thus represent moments of heightened contingency during which the actions of secessionist and state actors may have lasting consequences for movement cohesion. Empirically, the article focuses on the 2017 independence referendum in Catalonia, which was unilaterally called by secessionists and met with severe opposition by the Spanish state. Although the referendum was suspended by the Constitutional Court, two million Catalans cast their votes largely in favor of independence on 1 October and the Catalan parliament declared independence 4 weeks later. This led to a severe institutional crisis, which involved the suspension of the region’s autonomy and a trial that ended with long prison sentences for the movement leaders (Ferreira, 2021; López and Sanjaume-Calvet, 2020). I develop a novel approach to movement cohesion, which highlights interorganizational relations as a key dimension. Whether secessionist organizations collaborate in collective actions represents a crucial indicator for the movement’s internal structure, but it has rarely been scrutinized (Gade etal., 2019b). Furthermore, research on movement cohesion has predominantly focused on armed groups and violent action (Fjelde and Nilsson, 2018; Gade etal., 2019a, 2019b; Mosinger, 2018). In contrast, this article focuses on collaboration in primarily nonviolent protest, which is fundamental for most secessionist movements (Chenoweth and Stephan, 2011; Griffiths and Wasser, 2019). I employed a mixed-methods research design that combined protest event data with qualitative materials to address the research question empirically. I elaborated an original event data set of self-determination protest in Catalonia from 2015 to 2019. These data allowed for a more fine-grained analysis than previous data at the yearlevel and were used to generate networks of interorganizational collaboration in collective action before and after the referendum. Retrospective semi-structured interviews with key organizers were used to identify the mechanisms explaining the results of the network analyses. I found that the 2017 referendum was the cornerstone for how secessionist actors framed their strategies. Before the event, the referendum as a shared goal provided the basis for cohesion, but conflicts over the meaning of the referendum led to fragmentation afterwards. While these findings emerged inductively from the qualitative materials, they are best captured by the concepts of ‘frame alignment’ (Snow etal., 1986) and ‘frame dispute’ (Benford, 1993). They highlight the role of critical junctures and strategy framing as key mechanisms driving secessionist cohesion. Conceptualizing cohesion: A relational approach A central dimension of the structure of a secessionist movement is whether it is cohesive or fragmented.1 Fragmentation and cohesion can be conceptualized as Gunzelmann 963 opposite endpoints of the same spectrum (Bakke etal., 2012; Pearlman, 2011: 25). The most widespread approach is to conceptualize and measure cohesion and fragmentation as the number of active organizations within a movement (Fliervoet and Seymour, 2023: 166). In other words, the more organizations a movement has, the more fragmented it is (Cunningham, 2013a, 2014; Fjelde and Nilsson, 2018; Seymour etal., 2016). However, this approach has been criticized as too one-dimensional (Krause, 2013; Pearlman, 2011: 8). In response, scholars have developed more complex conceptualizations. Dowd (2015) included not only the number of organizations but also the degree of violence they use. Bakke et al. (2012) added institutions and power as important dimensions. On the other hand, Krause (2013, 2017) and Pearlman (2011) dismissed the importance of the number of organizations. While Krause emphasized the role of power within movements, Pearlman conceptualized movement cohesion as the combination of leadership, institutions and the constituencies’ ‘sense of collective purpose’ (2011: 23). This article suggests that there is another key dimension of movement cohesion and fragmentation: interorganizational relations. Interorganizational relations may range from conflict to cooperation.2 Organizations within the same movement might see each other as partners, competitors or even enemies. The relational dimension is distinct from other elements of cohesion and fragmentation. First, unless a movement is fully unified in a single organization, interorganizational relations are independent of the number of organizations. A movement with a large number of organizations that have friendly relations should be considered more united than a movement that is divided into only two polarized blocks. Second, cooperative or conflictual relations are different from power relations. Movements with dispersed power can be quite cohesive if organizations have friendly relations and cooperate. Vice versa, a movement in which most power is concentrated in one organization may be less cohesive if smaller organizations challenge the hegemonic organization. The relational dimension of movement cohesion has best been approached through the notion of institutionalization. Bakke etal. (2012) even stated that ‘the degree of institutionalization of a movement characterizes the ties between organizations that it comprises’ (2012: 270) but restrict it to ‘formal and informal rules’ (2012: 268). Likewise, Pearlman (2011: 24) defined institutions as ‘the structures and norms that govern social interaction’. However, this structuralist focus overlooks that interorganizational relations can take a myriad of forms. Organizations may collaborate in the absence of rules and norms. Also, the notion of institutionalization largely neglects the conflictual side of interorganizational relations. This is why it is necessary to conceptualize and study interorganizational relations in secessionist movements as a key dimension of cohesion and fragmentation. The most promising steps in this direction have been made by Gade etal. (2019a, 2019b) in a pair of complementary articles. Drawing on data from the Syrian civil war, they use network analytical tools to study collaboration and conflict within rebel groups. However, their focus is predominantly on the empirical question of with whom rebels cooperate or infight rather than on conceptual concerns. I propose a relational conceptualization of movement cohesion, which rests on an understanding of social movements as networks of collective and individual actors that share a common identity and engage in conflictual relations with their opponents (Della Porta and Diani, 2020). The central building block at the meso level are collaborative ties. Collective actors may exchange resources and knowledge or may even have overlapping membership (Baldassarri and Diani, 2007). The denser these collaborative ties, the more cohesive the movement. It is an aggregate quality located at the movement level and should be thought of as a spectrum from 0 (absence of collaboration) to 1 (permanent collaboration). The form of collaborative ties may vary with regard to their organizational form (formalized or not), their geographical scope (local, national, transnational), and their durability (single events, campaigns, or long-term alliances) (Brooker and Meyer, 2018; see McCammon and Moon, 2015; Van Dyke and Amos, 2017). Critical junctures and movement cohesion Previous research on movement cohesion has emphasized the importance of interactions between secessionists and the state as well as among secessionists themselves – what Cunningham et al. (2012) and Seymour et al. (2016) called ‘dual contests’. With regard to internal dynamics, the ideological proximity of the organizations matters (Gade etal., 2019a, 2019b), as well as the diversity of their demands, and whether they use violent tactics (Seymour etal., 2016). There is less agreement on the role of the state. Seymour etal. (2016) and Fjelde and Nilsson (2018) argue that accommodation increases movement fragmentation, whereas Cunningham (2014) found that it led to cohesion in the long run. Cunningham (2014), Lawrence 964 journal of Peace Research 62(4) (2010), and Seymour etal. (2016) suggest that (violent) repression leads to fragmentation, while Fjelde and Nilsson (2018) found no effect. These mixed findings could be the product of moments of heightened contingency that tip the movement towards cohesion or fragmentation. Such critical junctures are fairly short periods of choice during which the course of history is decided (Capoccia and Kelemen, 2007; Collier and Collier, 2002; Mahoney, 2002). They may have long-lasting consequences and become ‘transformative events’ (McAdam and Sewell, 2001; Sewell, 1996). Recent studies have demonstrated the consequences of these events for coalition building in social movements (Ciordia, 2021; Portos and Carvalho, 2022; Wood etal., 2017). This article builds theory on how critical junctures drive movement cohesion by looking at independence referendums. Independence referendums represent the most important critical junctures in secessionist conflicts besides the onset of civil war and the declaration of independence (Qvortrup, 2014a). Not all referendums are actual decisions on independence, which ultimately depends on international recognition (Coggins, 2014; Griffiths, 2017). A key factor for recognition is whether a referendum is agreed upon with the host state or pursued unilaterally by the secessionist movement (Sanjaume-Calvet, 2021). Theoretically, the effects of an agreed referendum on movement cohesion are straightforward: if the pro-independence camp wins the vote, secession is very likely and the movement achieves its goal. If the vote is lost, blame attribution and the lack of options going forward is likely to increase movement fragmentation. The effects of unilateral referendums on secessionist movements are less clear. On the one hand, unilateral referendums represent a crucial opportunity for secessionists. By setting a date for the referendum, secessionists dare the central government to react and thus self-create a window of opportunity (Cortés Rivera, 2020). Bringing in the population of the seceding region can be used to leverage the legitimacy of secession (Sanjaume-Calvet, 2021). Most importantly for the protest arena, referendums can spark the mobilization of supporters (Della Porta etal., 2017). Calling a referendum may also set a short-term goal around which different organizations and parties may rally (Della Porta etal., 2019). On the other hand, unilateral referendums also carry risks for secessionist movements. Most unilateral referendums do not lead to independence (Sanjaume-Calvet, 2021). In established democracies in particular, secession appears to be virtually impossible (Dion, 1996; Qvortrup, 2014b). The failure to achieve independence may demobilize and exhaust supporters (Lecours, 2018). Similar to agreed referendums, blame attribution and the lack of options going forward may lead to internal conflicts. Finally, unilateral referendums normally lead to a backlash from the host state, including violent repression (Basta, 2021). Repression, in turn, has been found to contribute to internal conflicts within the movement (Gade, 2020; Lawrence, 2010; Seymour etal., 2016). The remainder of the article focuses on the case of the October 2017 referendum in Catalonia to explore these contrary expectations empirically. I inductively build theory on the mechanisms connecting critical junctures and movement cohesion. The next section describes the data and methods that were used for the empirical part. Methods and data To study the impact of independence referendums on movement cohesion, I employed a mixed-methods research design that consists of two parts. The first part built on previous studies combining protest event and network analyses to analyze movement cohesion over four phases of contention (Bearman and Everett, 1993; Ciordia, 2021; Pirro et al., 2021; Wang and Soule, 2012): the phase before the announcement of the referendum plan in October 2016, during the referendum campaign, during the counter-secessionist backlash right after the referendum, and the contraction of the conflict in the second year after the conflict (see Table 1). The second part drew on qualitative data and in particular on semi-structured interviews with key organizers from the independence movement. These materials were used to explore the impact of the 2017 independence referendum on movement cohesion. Protest event analysis In this study, cohesion was conceptualized as the density of collaborative ties among collective actors in a secessionist movement. Collaborative ties were operationalized as co-participation in contentious events, that is, when two actors take part in the same protest event (Ciordia, 2021; Wang and Soule, 2012). The more often two actors protest together, the stronger the collaborative tie between them. Actors are usually very conscious of showing themselves with other actors in public, which is why co-participation is a good indicator of collaborative ties (Ciordia, 2020: 72; Diani, 1995: 99). Gunzelmann 965 The data collection covered self-determination (SD) protest events and their participating actors in Catalonia from October 2015 through September 2019. I follow Tilly and others in defining protest as the collective, non-routine act of public claims-making (McAdam etal., 2001; Tilly, 2008; Tilly and Tarrow, 2015). Protest events are claims-making acts that are bounded in time and space. The data collection included the larger category of SD protests rather than just secessionist contention. SD protest refers to all public acts making claims about a group’s control over its own affairs in relation to the state (Cunningham, 2014). These demands can be defensive ones, such as the protection of minority rights, but most will be expansive ones such as greater fiscal or institutional autonomy, as well as calls for outright independence. The reason for including a broader set of claims was that SD protests beyond independence claims provide important opportunities for secessionist actors to collaborate. Data were collected from two Catalan newspapers: El Periódico and El Punt Avui. The selection is justified by two criteria. The first is geographical proximity, which has been found to increase the propensity to cover protests (Danzger, 1975; Fillieule and Jiménez, 2003; Ortiz etal., 2006). El Periódico is a national newspaper based in Barcelona, while El Punt Avui, is a regional newspaper based in Girona. Being based in Catalonia and having an explicitly regional focus, the two newspapers were preferable over Madrid-based newspapers such as El País or El Mundo. The second criterion is ideological balance with regard to the secessionist conflict in Catalonia. El Punt Avui openly supports Catalan secession from Spain, while El Periódico is known to be opposed to Catalan independence. I suggest that their selection biases should be complementary. Whereas El Punt Avui should be more likely to cover peaceful events, especially in response to counter-secessionist repression, El Periódico can be expected to focus on violent and disruptive actions. The starting point for the time frame is justified with the regional elections in September 2015, which resulted in a clear pro-independence majority in the Catalan parliament and led to a new phase in the territorial conflict (Martí and Cetrà, 2016; Orriols and Rodon, 2016). At the heart of this phase is the referendum on independence on 1 October 2017. The end of the trial of the Catalan leaders in October 2019 represents the endpoint of the study. Geographically, the project focuses on protests in the four provinces (Barcelona, Girona, Lleida and Tarragona) of the Autonomous Community of Catalonia. I used a combination of action forms and SD claims as keywords to search the archives of the two newspapers through the online platform Factiva (see Online Appendix 3). I followed Ciordia’s (2020, 2021) strategy for coding the articles. This strategy involves two separate codebooks: one for events and one for event participants (see Online Appendix 2). I used the Discourse Network Analyzer software developed by Leifeld (2016) to perform these codings. The main part of the data analysis was devoted to reconstructing network patterns among secessionist collective actors before and after the 1 October referendum. This required moving from protest event and participant data to network data, for which I created attendance lists for each of the four time periods. Only secessionist actors that attended at least two protest events were included in the network analyses for each period. To control for the fact that some organizations participate in many more events than others, I followed Ciordia (2020, 2021) in using Jaccard coefficients as a normalized indicator for co-participation. Raw and Jaccard matrices were exported for all four periods separately. Network statistics were calculated in Excel and Visone and are displayed in Table 3 in the following section. Networks were visualized with the basic Visone algorithm, which uses metric multidimensional scaling and stress minimization. Network graphs are available in Online Appendix 4. Qualitative analysis The qualitative part of the research design is based on 10 months of fieldwork in Catalonia between May 2018 and March 2019 (see Online Appendix 5). The primary database consists of 30 interviews with key organizers Table 1. Phases of contention in the secessionist cycle of contention in Catalonia. ID Time period Phase of contention 1 October 2015–September 2016 Pre-campaign 2 October 2016–1 October 2017 Referendum campaign 3 2 October 2017–September 2018 Post-referendum I 4 October 2018–September 2019 Post-referendum II 966 journal of Peace Research 62(4) from the most important organizations from all sectors of the Catalan independence movement: Assemblea Nacional Catalana (ANC), Òmnium Cultural, Committees for the Defense of the Republic (CDR), and Universitats per la República (UxR). To increase the diversity of organizational backgrounds, the interviewees varied with regard to organizational level (local and regional), activist experience (from several years to a few months), and geographical location (rural areas, small towns, and several neighborhoods of Barcelona). The names of the interviewees were pseudonymized. All interviews except one were transcribed for analysis, which produced 555 pages of raw data. These materials were supplemented by field notes, seven expert interviews (activist-scholars and journalists), and documents produced by social movement organizations. The additional materials were used to interpret the context and timing of interorganizational relations. Part of the interview guide comprised questions about the changing relations with other major movement organizations as well as with political parties. The analysis of their answers followed the logic of grounded theory (Charmaz, 2006; Glaser and Strauss, 1967; Mattoni, 2014), trying to identify changing patterns in the relationships as well as how they were connected to the 1 October referendum. Four analytical phases were performed in MaxQDA. First, the interviewees’ statements were coded into three categories: relations within umbrella organizations, among organizations, and between organizations and other actors. Second, I constructed summary grids and tables from which to abstract individual statements and provide an overview for each organization. Third, using the summary tables, I identified patterns for each organization as well as overarching themes. Fourth, I selected representative quotes for each of the themes, which structure the narrative in the empirical section of this article. Secessionist protest and the October 2017 referendum The 2015 regional elections marked the beginning of a new phase in the secessionist conflict. It was proclaimed a ‘referendum on independence’ by the secessionist parties Convergència Democràtica de Catalunya and Esquerra Republicana de Catalaunya (ERC), which ran on a single platform called Junts pel Sí (Together for Yes) (Martí and Cetrà, 2016). During the campaign, the new coalition committed to an 18-month process of unilateral secession from the Spanish state in the case of electoral success, but failed to obtain a majority of seats in the Catalan parliament (Orriols and Rodon, 2016). A motion to start secession from Spain shortly after the election was quickly suspended by the Spanish Constitutional Court, which meant that the unilateral strategy promised before the elections had reached an impasse. I consider the year after the regional election to be the first phase under study (pre-campaign, see Table 1 in the previous section). After surviving a vote of confidence in the Catalan parliament on 28 September 2016, Puigdemont changed course and vowed to call a binding referendum on independence. This can be considered the beginning of the extended referendum campaign, which represents the second phase of the late secessionist conflict. On 6 June 2017, the referendum was officially announced for 1 October of the same year. The Catalan government tried to achieve an agreement with the Spanish government, which rejected the referendum arguing that it was against the Spanish constitution. At the beginning of September 2017, the pro-independence majority in the Catalan parliament passed the ‘Law on the Referendum on Self-Determination of Catalonia’ in a controversial session boycotted by the unionist parties. The Spanish Constitutional Court immediately suspended the law after an appeal from the Spanish government and declared the referendum illegal. On 1 October, riot squads of the Spanish National Police and the Civil Guard attempted to prevent the referendum by entering voting stations and confiscating ballot boxes. The crackdown resulted in more than 900 people being injured and some voting stations had to be closed because of police intervention (Barceló, 2018; Della Porta etal., 2021). More than two million Catalans (about 43% of the electorate) turned out to vote overwhelmingly in favor of independence (more than 90%), while the unionist camp did not participate in the referendum. The first year after the referendum can be considered the third phase of the late secessionist conflict. On 10 October, Puigdemont stood before the Catalan parliament to declare independence – only to suspend his declaration a minute later. A few days later, the leaders of ANC and Òmnium Cultural were arrested for their role in organizing a protest against the police raids prior to 1 October. On 27 October, the pro-independence majority in the Catalan parliament voted to declare independence. That same day, the Spanish Senate voted in favor of applying Article 155 of the Spanish Constitution, which imposed the temporary suspension of Catalonia’s autonomy and snap regional elections. In the following days, Puigdemont and other members of the Catalan government left Spain to avoid legal prosecution, while others, including Vice President Oriol Junqueras, were arrested. Nevertheless, the secessionist parties defended Gunzelmann 967 their parliamentary majority in the elections on 21 December. At the end of March 2018, several former members of the Catalan government were arrested, while Puigdemont was briefly detained in Germany. The fourth phase of the late secessionist conflict began in October 2018 and was less eventful. It was marked by the judicial trial of 10 Catalan politicians, including former Vice President Oriol Junqueras and the activist leaders Jordi Sànchez and Jordi Cuixart. The trial started in February 2019 and ended with lengthy prison sentences for nine of the accused in October 2019, which is also considered the endpoint for this study. The protest event data revealed the trajectory of SD protests during the four phases of conflict. Figure 1 shows how the number of SD protests evolved over time. In the first phase, there were fewer SD protests and fewer actors engaging in these protests in comparison to the following phases. The secessionist conflict unfolded mainly in the institutional realm and less so in the protest arena. During the referendum campaign (Phase 2), movement actors organized about a hundred events more than in the previous phase. Most of them took place in September 2017, right before the referendum. This included the protests against the police raids immediately before the referendum, but also the massive mobilization to occupy voting stations and ‘defend’ them against police intervention on the day of the referendum. The first year after the referendum (Phase 3) represented the peak in the number of events. It accounts for about half of the events of the whole period under study. Most of these events took place in the most contentious phase right after the referendum. For instance, the general strike on 3 October turned into a massive protest event against the police violence on the day of the referendum. After the ineffective declaration of independence on 27 October, the level of mobilization dropped until the end of March 2018 when Puigdemont was briefly arrested in Germany. Less than a third of the events in this phase were collaborative (see Table 2). During the second year after the referendum (Phase 4), the level of mobilization rose initially but dropped after the onset of the trial of the Catalan leaders in February 2019. Only when the end of the verdict came closer in September 2019 did the activists initiate a new wave of protest. However, the number of collaborative events dropped remarkably in comparison to previous phases. Figure 1. Trajectory of self-determination protest in Catalonia 2015–2019 (N = 1,043). Table 2. Types of SD protest events. Phase 1 2 3 4 Time Oct 2015–Sept 2016 Oct 2016–1 Oct 2017 2 Oct 2017–Sept 2018 Oct 2018–Sept 2019 Total events 80 180 506 277 Event types 0 org. events 13 76 199 97 1 org. events 31 37 163 135 Collab. events 36 67 144 45 Org.: organization(s), collab.: collaborative event with two or more organizations present. 968 journal of Peace Research 62(4) The data illustrated how SD protest mobilization was closely connected to the 2017 referendum and the dynamics of conflict in the institutional arena. While the level of mobilization was rather low for the first two years after the regional elections, the referendum triggered a wave of protest that lasted for almost one and a half years. Protesters also responded to counter-secessionist actions by the Spanish state, such as the numerous arrests of Catalan politicians and activists. The next section turns to the networks of collaboration in the protest arena to show how movement cohesion evolved over the four phases. Networks of protest collaboration over time The secessionist push for a binding referendum and the response by the Spanish state triggered an episode of intense contention in Catalonia. This section focuses on the networks of collective actors in the protest arena before and after the referendum. Network analyses showed that there was less collaboration in the protest arena after the referendum despite this massive mobilization. The protest event data showed that the two largest social movement organizations, ANC and Òmnium Cultural, played a significant role in the first phase after 2015 but were far from being the only ones (participating actors are displayed in the network graphs in Online Appendix 4). Interest groups such as the Associació de Municipis per la Independència (AMI) or the Associació Catalana de Municipis (ACM), pro-independence trade unions such as Coordinadora Obrera Sindical (COS) and CSC-Intersindical, and smaller organizations such as ERC’s youth organization Joventuts’ Esquerra Republicana (JERC) also collaborated in the protest arena. Institutional actors engaged in the protest arena, too. The pro-independence parties frequently participated in protests as well as the then President of the Generalitat, Artur Mas, and the then President of the Catalan Parliament, Carme Forcadell. These actors formed a network of 17 nodes (see Table 3). This network changed slightly during the second phase. The network grew from 17 to 20 nodes. The main new actor was the platform UxR, which was formed by the youth organizations Arran, JERC, Joventut Nacionalista de Catalunya, and Sindicat d’Estudiants dels Països Catalans (SEPC), and emerged specifically to mobilize students for the referendum. These groups had particularly tight relations through their participation in UxR. Despite the larger size, the movement was better connected than before, as the density value had risen from .57 in the previous phase to .70. Collective actors also collaborated more often than before: the average Jaccard coefficient increased slightly from .11 to .12. This shows that the secessionist movement became more cohesive during the referendum campaign. The data indicated that the referendum marked a turning point for the cohesion of the independence movement. In the third phase, the network grew substantively to 31 nodes. However, the protest arena was divided into four key groups. First, most political parties, institutions, as well as ACM and AMI formed one group (on the right side in the network graph in Online Appendix 4). Second, the left-wing actors COS, Alerta Solidària and Endavant aligned with the student organizations through Arran and SEPC as brokers. Third, ANC and Òmnium continued to have tight relations with each other and were at the center of the network, keeping some relations with most other actors in the network. Fourth, the CDR emerged as the most Table 3. Descriptive network statistics. Phase 1 2 3 4 Time Oct 2015–Sept 2016 Oct 2016–1 Oct 2017 2 Oct 2017–Sept 2018 Oct 2018–Sept 2019 Total events 80 180 506 277 Network statistics Nodes 17 20 31 25 Isolates 0 0 1 2 Edges 680 1,246 1,760 1,098 Average degree 9.18 13.20 11.87 12.88 Density 0.57 0.70 0.40 0.54 Jaccard coeff. Range 0–1 0–1 0–0.89 0–0.75 Average 0.11 0.12 0.05 0.07 Gunzelmann 975 Social Movement Theory. New Haven, CN: Yale University Press, 133–155. Snow DA, Rochford EB, Worden SK, etal. (1986) Frame alignment processes, Micromobilization, and Movement Participation. American Sociological Review 51(4): 464. Sorens J (2012) Secessionism: Identity, Interest, and Strategy. Montreal, Canada: McGill-Queen’s University Press. Staniland P (2014) Networks of Rebellion: Explaining Insurgent Cohesion and Collapse. Ithaca, NY: Cornell University Press. Tilly C (2008) Contentious Performances. Cambridge: Cambridge University Press. Tilly C and Tarrow SG (2015) Contentious Politics. 2nd revised edn. Oxford: Oxford University Press. Van Dyke N and Amos B (2017) Social movement coalitions: Formation, longevity, and success. Sociology Compass 11(7): e12489. Wagner-Pacifici R (2017) What Is an Event? Chicago: The University of Chicago Press. Wang DJ and Soule SA (2012) Social movement organizational collaboration: Networks of learning and the diffusion of protest tactics, 1960–1995. American Journal of Sociology 117(6): 1674–1722. Woldemariam M (2018) Insurgent Fragmentation in the Horn of Africa: Rebellion and Its Discontents. Cambridge: Cambridge University Press. Wood LJ, Staggenborg S, Stalker GJ, etal. (2017) Eventful events: Local outcomes of G20 summit protests in Pittsburgh and Toronto. Social Movement Studies 16(5): 595–609. HANS JONAS GUNZELMANN, b. 1990, PhD in Political Science and Sociology (Scuola Normale Superiore, 2021); Postdoctoral Researcher at the Center for Civil Society Research, WZB Berlin Social Science Center (2023–present); research interests: political sociology, protest, social movements, nationalism.