Introduction
On 12 July, tens of thousands of members of the Orange Order – a Protestant fraternal order styled on Freemasonry – parade through the streets of Northern Ireland to commemorate and celebrate the 1690 victory of Protestant King William of Orange over the Catholic King James II. While ‘The Twelfth’ is the peak of the parading season, it represents just a fraction of the more than 2,000 parades that take place across the north of Ireland each year, the vast majority of which are displays of allegiance to the Union with Great Britain and to the Protestant faith. Of these, approximately 100 parades are deemed ‘sensitive’ by the Parades Commission – a body set up to determine whether restrictions should be imposed on parades – because they have the potential to raise concerns and community tensions between the two major communities in Northern Ireland, Protestants and Catholics. The marching season was a catalyst for sectarian street violence during the thirty-year-long conflict known as ‘the Troubles’. While the armed conflict officially ended over two decades ago with the Good Friday Agreement (1998), tensions around parading persist today.
Parades in Northern Ireland are a prominent example of a larger phenomenon: contested cultural symbols and rituals that are not explicitly political but are seen as making political claims and, as such, have the potential to maintain, or even deepen, divisions between identity groups. These ‘contentious rituals’ (Blake Reference Blake2019 a) often take place in ethnically divided societies – including conflict-ridden or post-conflict societies in which the conflict itself has pitted identity groups against one another. For example, in 2023, the annual ‘Dance of Flags’ parade by Israeli nationalists through Muslim parts of the Old City of Jerusalem led to clashes with resident counter-protesters and ultimately the storming of al-Aqsa mosque (The Times 2024). In 2022, a procession of Hindu nationalists armed with swords and pistols marched through the Jahangirpuri district in Delhi, shouting anti-Muslim slogans (BBC News 2022). Similar marches are often the catalysts of communal rioting and violence in India. In the United States, too, there are contentious parades, such as the Columbus Day Parade in Denver, with opponents to the parade seeing it not as a celebration of Italian-American culture but, rather, a celebration of the genocide against Native Americans (Hitchmough Reference Hitchmough2013). In this paper, we focus on the question of how contentious rituals can contribute to perpetuating divisions in post-conflict societies, using parades in Northern Ireland as a case study.
We argue that while contentious rituals cultivate social or collective memories that foster in-group solidarity and identity, they are detrimental to intergroup relations through both a pride and provocation mechanism. Parading and other forms of collective performance can encourage participation and commitment by ‘in-group’ members while fostering polarizing attitudes against the ‘out-group’. We expect this negative effect of contentious rituals to be relevant for intergroup relations in divided societies generally, including conflict-ridden and post-conflict societies in which the main cleavage has pitted identity groups, particularly ethnic groups, against one another. In our case, we expect that members of the parading in-group in Northern Ireland – predominately members of the Protestant community – that are exposed to parades have more negative attitudes towards the out-group because their in-group pride is elevated. In contrast, out-group onlookers – predominately members of the Catholic community – are provoked by parading, and the provocation further generates negative intergroup attitudes. Thus, overall, we expect that parades as political symbolism contribute to the deterioration of intergroup relations in Northern Ireland.
We use a multilayered mixed-method research design to assess the effect of parades on intergroup relations and to explore key mechanisms. First, we gather fine-grained geolocated data on over 55,000 parades in Northern Ireland, from 2002 to 2022. We combine this novel Irish Contentious Parades (ICP) dataset with election surveys to assess if living in constituencies that experience contentious parades is associated with declining support for intergroup mixing, such as interethnic marriage and schooling. Second, we use an original survey fielded in Protestant and Catholic urban communities to leverage the plausibly exogenous variation in the survey interview dates to estimate the effect of parade exposure on attitudes towards intergroup mixing. Finally, we conduct fieldwork in Northern Ireland around ‘The Twelfth’, including participant observation of contentious parades and interviews with residents and key community leaders, which, in combination with qualitative secondary sources, enables us to gain insights into the mechanisms of in-group pride and intergroup provocation.
Pooling publicly available surveys over time, we find that parades are associated with negative intergroup attitudes, specifically towards intergroup marriage and schooling. The analysis of an original survey conducted in 2022 suggests that this general trend is due to negative short-term effects of parades on intergroup attitudes. Our fieldwork in Northern Ireland, in summer 2023, indicates that out-group members refer to provocation when discussing parading, while members of the in-group highlight the pride they feel as members of their community. We also find, unexpectedly, that members of the in-group identify provocation as an important mechanism, which may explain escalation dynamics – that is, the in-group perceives contention around their parades as a provocation tactic, leading to more intergroup confrontations and a subsequent worsening of relations.
We make two core contributions. The first is empirical; namely, a novel geolocated dataset on contentious rituals in Northern Ireland that can be used for other analyses and provides insights into the legacies of the conflict known as ‘the Troubles’. Second, our substantive findings make theoretical contributions to literatures on intergroup relations and contentious rituals. Our findings show that collective social memory practices, such as the parades in Northern Ireland, in Jerusalem, or in Delhi, are important ways of fostering a sense of community but, simultaneously, have the potential to undermine intergroup relations. Drawing on research on social identity and group processes, we describe two plausible mechanisms through which intergroup relations may be undermined during and immediately after contentious rituals. Furthermore, we quantify the extent of this undermining effect and provide a rich empirical examination through a multilayered research design.
Overall, we demonstrate that memory practices and rituals can activate group processes – well-known to the social identity literature – that ultimately deteriorate intergroup attitudes. Our study also matters for policy, suggesting that policymakers walk a tightrope between allowing groups to engage in a traditional practice that fosters strong communities while ensuring that the same practices do not undermine intergroup relations.
Intergroup Relations and Political Symbols in Divided Societies
While ethnic group divisions can foster conflict (e.g. Horowitz Reference Horowitz1985; Hewstone and Greenland Reference Hewstone and Greenland2000), most conflicts, particularly in ethnically divided societies, also reinforce or produce strong identity divisions in the population (e.g. Wood Reference Wood2008). These legacies can be persistent across time and manifest in post-conflict political behaviour, institutions, and attitudes. For example, studies have shown that many ethnic groups in post-conflict settings continue to vote based on their ethnicity (e.g. Hadzic et al. Reference Hadzic, Carlson and Tavits2020). Shayo and Zussman (Reference Shayo and Zussman2017) show that ethnic biases in the judicial decisions of Israeli Arab and Jewish judges showed no sign of attenuation during times of relative stability. A meta-analysis of the effects of violence on prosocial behaviour finds that wartime violence increases co-operation and altruism but fosters positive in-group bias rather than generalized positive attitudes (Bauer et al. Reference Bauer, Blattman, Chytilová, Henrich, Miguel and Mitts2016). Thus, central to the long-standing conflict resolution literature, which is rooted in research on social identity theory (Tajfel and Turner Reference Tajfel, Turner, Jost and Sidanius2004), is the emphasis on the removal of the negation of ‘the other’ in people’s identities (Kelman Reference Kelman, Nadler, Malloy and Fisher2008) and programmes that foster intergroup contact and reconciliation (e.g. Hewstone et al. Reference Hewstone, Cairns, Voci, Hamberger and Niens2006; Noor et al. Reference Noor, Brown, Prentice, Nadler, Malloy and Fisher2008; Bar-Tal Reference Bar-Tal2000).
The wide literatures on the effects and legacies of violence suggest various mechanisms through which conflict identities shape and are maintained in post-conflict societies. Traumatic individual experiences can increase fears of future threats and, hence, lead to hostility and exclusionary attitudes towards the out-group(s), which can be passed on to descendants (e.g. Beber et al. Reference Beber, Roessler and Scacco2014; Hirsch-Hoefler et al. Reference Hirsch-Hoefler, Canetti, Rapaport and Hobfoll2016). Indeed, a growing body of work shows that individual wartime experiences can, through socialization, have intergenerational effects on identities and political preferences (e.g. Balcells Reference Balcells2012; Lupu and Peisakhin Reference Lupu and Peisakhin2017; Rozenas et al. Reference Rozenas, Schutte and Zhukov2017; Walden and Zhukov Reference Walden and Zhukov2020). Others emphasize that political institutions – such as parties and candidates – rely upon group-based strategies and patronage, which reinforces divisions rather than encouraging group assimilation over time (e.g. Parenti Reference Parenti1967). While power sharing has long been considered an institutional ‘solution’ to peace in ethnically divided societies (e.g. Lijphart Reference Lijphart1990; Jarstad and Nilsson Reference Jarstad and Nilsson2008; Cederman et al. Reference Cederman, Hug and Wucherpfennig2022), critics of power sharing highlight that these institutions can lock in conflict-era identity cleavages (e.g. Jarstad and Sisk Reference Jarstad and Sisk2008; Roeder and Rothchild Reference Roeder and Rothchild2005). In addition, informal wartime institutions can persist over time and maintain group differences because civilians are socialized to rely on their groups’ informal institutions (e.g. Bateson Reference Bateson2017; Rickard and Bakke Reference Rickard and Bakke2021).
Identity groups in divided societies also draw on and develop practices, rites, and symbols to activate their identity and draw boundaries against ‘the other’. Central to understanding ethnic war is the myth-symbol complex – particularly chauvinist myths – that underpins and fosters fears of the out-group, which can be manipulated by political elites (Kaufman Reference Kaufman2001). In their classic contributions, Horowitz (Reference Horowitz2000) shows that many ethnic riots worldwide are preceded by processions and other symbolic events, and Brass (Reference Brass1974) shows how political elites seek to foster consciousness by manipulating symbols of group identity. An abundant literature highlights that religious authorities can be powerful motivators for collective and political action, as Qur’anic verses motivate people to donate in Afghanistan (Condra et al. Reference Condra, Isaqzadeh and Linardi2019), Sabbath pamphlets in Israel increase national sentiment (Freedman Reference Freedman2019), and Pentecostal messages in Nairobi increase political participation (McClendon and Riedl Reference McClendon and Riedl2015).
We home in on contentious rituals, which are socially standardized and repeated actions that are considered ‘contested because of the political claims [they are] seen to make’ (Blake Reference Blake2019 a, p. 7–8). There are tangible objects like national symbols – flags and banners, for example – that can accentuate people’s identification as members of an ethnic or national group by visualizing group identities (Schatz and Lavine Reference Schatz and Lavine2007). There are also ritualistic-ceremonial activities such as parades (Blake Reference Blake2019 b), celebrations (Rohava Reference Rohava2020), and songs (Schumann Reference Schumann2013) that behaviourally communicate group belongings and cultivate social, or collective, memory. These behaviours require group members to actively express their identity, typically in plain sight of members of other ethnic groups, with consequences for intergroup relations. For example, Allie (Reference Allie2023) shows that incompatible ritual holidays in which a religious ritual clashes with other practices explain why Hindu–Muslim rioting emerges during holidays in India. These practices are symbolic acts to reinforce the in-group’s values and identities, but can, for that reason, come to be seen as contested or inherently political by out-groups.
Importantly, contentious rituals are distinguishable from other forms of ‘contentious politics’ (Tilly Reference Tilly1978), such as demonstrations and strike actions, through two defining features. First, contentious rituals do not make direct demands (for example, calls for policy change or elections) but are nevertheless perceived as contested in a politicized context. Second, organizers of contentious rituals attach the practice to the in-group’s wider culture, religion, or traditions and claim that the rituals serve the expression of the in-group’s history or social memory. Hence, contentious rituals particularly emerge in what Lijphart called ‘divided societies’ with clear cultural, ethnic, or racial identity cleavages (Lijphart Reference Lijphart1977). This can include conflict-ridden or post-conflict settings.
We focus on the question of whether and how these symbolic practices shape intergroup relations in the context of Northern Ireland. Following Blake (Reference Blake2019 a), we consider the predominately Protestant parades as an example of contentious rituals (also see Blake Reference Blake2019 b). The aim of this paper is not to provide an in-depth account of the importance or social relevance of parading in Northern Ireland, as this is already well documented (e.g. Kelly and Fraser Reference Kelly and Fraser2000; Jarman Reference Jarman1997; Blake Reference Blake2019 a; Bryan Reference Bryan2000). Rather, this project aims to assess the impact of these parades, as an example of a larger class of contentious rituals on intergroup relations over time, and identify causal mechanisms.
How Contentious Rituals Shape Intergroup Relations
How do contentious rituals contribute to the persistence of divisions in post-conflict and divided societies over time? Drawing on social movement theory and political psychology, we argue that contentious rituals, such as parades in Northern Ireland, cultivate social memory, generate in-group favouritism, and reinforce antagonistic intergroup relations by increasing pride in the in-group and provoking the out-group.
Activating Pride in the In-Group
In post-conflict societies, wartime agendas often prevail (e.g. Keen Reference Keen2012; Lake Reference Lake2017). Actors who have previously been directly involved in combat or indirectly mobilized for a conflict party may have electoral incentives (e.g. Hadzic et al. Reference Hadzic, Carlson and Tavits2020; Parenti Reference Parenti1967), economic motives (Nitzschke and Studdard Reference Nitzschke and Studdard2005), or social and political reasons to uphold their wartime social control and relevance (e.g. Themnér Reference Themnér2013; Carey and González Reference Carey and González2021; Rickard and Bakke Reference Rickard and Bakke2021). One way to do so is through cultural expressions (McEvoy Reference McEvoy2011), including encouraging ritualistic practices such as marches and parades. The advantage of traditions and ritual events are that they can be subtle and ‘cheap’ forms of mobilization (Allie Reference Allie2023), as opposed to activities that may violate peace agreements and commitments to disarmament, or cause direct violent backlashes in polarized environments. These behaviours to uphold political divisions can emerge bottom-up from ordinary people as ‘symbolic’ involvements with the in-group but can also be captured by utilitarian actors, such as former conflict parties, as ‘instrumental’ involvement with ordinary citizens (Schatz and Lavine Reference Schatz and Lavine2007). Contentious rituals delineate the in-group from the out-group; only the in-group takes pride in and can participate in this practice, while the out-group watches a demonstration of the mobilizing capabilities of the other side.
What is the effect of contentious rituals on the attitudes of in-group members? Rituals remind in-group members about their joint history, collective memory, and past political successes, which are central to their identity (e.g. Horowitz Reference Horowitz1985) and, in social identity theory terms, ‘heighten[s] identification with, and positive attachment to, the in-group’ (Tajfel and Turner Reference Tajfel, Turner, Jost and Sidanius2004, p. 277). Marches, processions, and parades are also powerful because already-mobilized individuals can draw in less-mobilized in-group members, who are encouraged to engage and participate in a celebration of history and pride. As noted by Jarman (Reference Jarman1997, p. 10) on the specific case of parades in Northern Ireland, ‘it is difficult to remain on the side-lines of such events, to be present physically but not emotionally’. In line with this, Blake (Reference Blake2019 a) finds that people take part in parades not because they are ideologically driven but because they enjoy participation.
We argue that social activities and engagements, such as rituals and marches, foster pride in the in-group and that this pride, in a divided society, can diminish positive attitudes towards the out-group. While the effect of group-level or collective pride on intergroup attitudes is less established than research on individual pride (Sullivan Reference Sullivan2014 b), Harth et al. (Reference Harth, Kessler and Leach2008) find experimental evidence that group-based pride predicts greater in-group favouritism in a resource distribution task, and Wagner et al. (Reference Wagner, Becker, Christ, Pettigrew and Schmidt2012) use longitudinal data to show that in-group pride predicts ethnic prejudice. Moreover, De Figueiredo Jr and Elkins (Reference De Figueiredo and Elkins2003) show that nationalistic pride, as opposed to patriotism, is a cause of pride-based prejudice. There are several mechanisms through which in-group pride generated by public rituals could decrease positive attitudes towards other groups. First, because pride is typically linked to achievements (for example winning in a historic battle) – and public rituals celebrate those achievements – it sets up group members in direct opposition to the out-group (Ross Reference Ross2007). Second, increased pride in their own group may make in-group members more sensitive to the need to maintain group differences to continue achieving their goals, directly hindering intergroup mixing and yielding self-assertive, defiant, or aggressive behaviour (Sullivan Reference Sullivan, von Scheve and Salmela2014 a). Third, individuals with high pride in or intense emotional identification with their in-group may respond negatively to criticism by the out-group, further increasing group divisions (Weisel and Böhm Reference Weisel and Böhm2015; DeLamater et al. Reference DeLamater, Katz and Kelman1969). Individuals who strongly associate with their group may be unable to see validity in criticism directed at their in-group due to their increased belief in the morality of their own group (Weisel and Böhm Reference Weisel and Böhm2015; Parker and Janoff-Bulman Reference Parker and Janoff-Bulman2013). As such, we suggest:
HYPOTHESIS 1: Contentious rituals foster negative intergroup attitudes among in-group members.
A Provocation for the Out-Group
If the in-group is motivated by pride, how does the out-group respond to displays of the tradition and history of the in-group? We follow the argument of Blake (Reference Blake2019 a) that out-group members do not perceive ritualized practices as merely apolitical commemoration that encourages celebration and participation. Instead, out-group members see a parade or march as ‘rituals of provocation’ (Blake Reference Blake2019 b) and as political statements.
First, when observing a ritualistic practice, members of the out-group receive a credible signal of the still-existing and potentially growing mobilizing power of the other group. This may induce fears of their own safety or frustration with the lack of reconciliation between the groups. Second, out-group members see these habitual behaviours as strategic attempts by elites from ‘the other side’ to dominate the public space, and as a consequence their lives, thereby reconfiguring the civic space (Bryan Reference Bryan2015). The perception that the opposing group continues to mobilize may be seen as a provocation (Blake Reference Blake2019 b) and as an attempt to re-mobilize and move the status quo in favour of their group. This may cause both fear and resentment directed towards the out-group – feelings that can foster divides and conflict (cf. Petersen Reference Petersen2002). For example, Guo (Reference Guo2020) shows in a formal model that provocative actions cause the out-group to worsen their discrimination against the other side and to engage in more in-group favouritism that is, in turn, reciprocated and leads to escalation.
Finally, compared to more strategic activities, such as in-group policing or co-opting of governments, contentious rituals are more loosely organized. In the case of Northern Ireland, a mix of ordinary people, low-level ex-combatants, and strategic leaders co-organize these ‘The Twelfth’ parades. The looser and less strategic organization of these celebrations leaves room for individual spoilers who make political statements or engage in anti-social behaviour that further provokes the out-group (going back to Stedman Reference Stedman1997). This negative group contact contributes to the alienation of both groups. As such, we suggest:
HYPOTHESIS 2: Contentious rituals foster negative intergroup attitudes among out-group members.
Overall, we expect that contentious rituals have a negative effect on intergroup attitudes through two distinct mechanisms – pride for in-group members and provocation for out-group members. After briefly discussing the history of parading in Northern Ireland, the remaining parts of the article aim to assess the effect of parading on intergroup relations and to provide evidence for the main mechanisms at play.
Northern Ireland and Parading
The Conflict in Northern Ireland
Northern Ireland has been affected by episodes of conflict since the seventeenth century, which corresponds to the beginning of ‘the Plantation’, the organized settlement of the region by Scottish Presbyterians (Wormald Reference Wormald, Ciardha and Siochrú2012). Divisions became even more pronounced after the partition of Ireland in 1921, with the creation of the Republic of Ireland, which was independent from British rule, and Northern Ireland, which remained part of the United Kingdom. Remnants of the Irish Republican Army who fought in the Irish War of Independence (1919–21) were not satisfied with the ‘unfinished revolution’ and continued their military activity with the goal of unifying Ireland (Bowyer Bell Reference Bowyer Bell1989). Animosity between the Nationalist and Unionist communities came to a head between 1968 and 1969, Footnote 1 when the civil conflict between paramilitary groups belonging to both communities and the British state began. The conflict, known as ‘the Troubles’, officially ended with the Good Friday Agreement in 1998, although armed dissident factions exist to this day and relations between the two communities remain tense (McKearney Reference McKearney2011; Horgan Reference Horgan2013; Hamill Reference Hamill2018).
History of Parading in Northern Ireland
Parading in Northern Ireland traces its beginning to the 1770s, when Protestant Irish communities used parades as a form of political action to obtain more autonomy for Ireland under the British Crown. The Orange Order, which is today the main organizer of Protestant parades, was founded at the end of the eighteenth century. Since then, parading has become an integral aspect of culture and politics in Northern Ireland, as well as some parts of the Republic of Ireland (Jarman Reference Jarman1997). ‘The Troubles’ brought an increase in violent incidents connected to parades. A prominent example for a contentious parade that is linked to outbursts of violence between Republicans and Loyalists is the Drumcree Parade. Since the first violent incident in 1972, this parade has remained highly controversial and has not been allowed to enter areas inhabited by a Catholic majority (Bryan Reference Bryan2000).
Since the beginning of parading, the Protestant community in Northern Ireland regularly organize parades around 12 July to celebrate the 1690 victory of the Protestant King William of Orange. Parades have become much more than the expression of collective history and memory. Ethnic and sectarian identity dynamics in Northern Ireland keep changing, and the meaning and importance of parading fluctuate and respond to the contemporary political situation (Jarman Reference Jarman1997). For example, new parades are added while others no longer take place, parade routes may change, and the classification of parades as sensitive Footnote 2 is renegotiated depending on counter-demonstrations and public concerns. This adaption of historical rituals to the current political situation highlights that parades are now a form of competitive commemoration (McQuaid Reference McQuaid2017).
Parades in Contemporary Northern Ireland
What do contemporary parades look like? Based on secondary sources and our fieldwork, we construct a schematic representation of two types of Protestant parades: the ‘church parades’ and the parades on 12 July. ‘The Twelfth’ parades are grand parades and tend to be many times bigger in size than smaller church parades. The largest parades, according to the parade organizers, included 15,000 participants and 40,000 supporters. Footnote 3 Many parades occur in Belfast, where bands start in different parts of the city and converge towards the centre, ultimately passing by Belfast Town Hall. The paraders are mostly comprised of Orange Order members and band members. The musicians’ appearances vary in uniform style, banners, and musical instruments. A notable feature of the parades is the presence of impressively big shoulder-mounted drums, which are played with great energy by band members (cf. Blake Reference Blake2019 a, pp. 3–4, 29–30). There are many onlookers, especially closer to the city centre in the Belfast parade, where large crowds come to cheer on the participants. Once the paraders have passed through the city centre, they move to ‘the field’ on the outskirts of Belfast. Here, a religious ceremony is held and distinguished figures in the Orange Order give speeches. Afterwards, the paraders move back to the city on their return routes. Return routes were described by interview participants as the most problematic moments of ‘The Twelfth’ parades due to high levels of alcohol consumption and associated disorder.
In contrast, ‘church parades’ are smaller. As their name implies, church parades end with the participants attending a religious service in a place of worship (Jarman Reference Jarman1997, pp. 99–100). They generally begin at the church where the service will be held. Participants mostly comprise of members of the Orange Order dressed in formal black attire. At the front of the parade, there is usually a band dressed in ceremonial, almost military-like, uniforms. The band usually includes a flag bearer. The flags we witnessed when attending included the Union Jack, the Orange Order flag, and the band banner. Marching is orderly and there tend to be few onlookers. Parades usually pass in front of an Orange Order lodge, with the route potentially passing through or near predominately Catholic residential areas. The parade then returns to the church for the service. During the parade, police officers may be present on the sidelines to record the parade and ensure that the Parade Commission’s determinations are respected, if there are any.
Research Design
To study contentious parades, we created the Irish Contentious Parades dataset, containing over 55,000 parades registered in Northern Ireland between 2002 and 2022. The dataset is based on publicly available data which we web-scraped from the Parades Commission (2023) website. The dataset contains the date of the parade, organizers, proposed start time, route, number and list of bands, and expected number of participants and supporters. We also code whether the Parades Commission judged the parade as ‘sensitive’ due to its potential to raise concerns and community tensions. When making this decision, the Commission seeks to balance the conflicting rights of different groups within the criteria laid down in the Public Processions (Northern Ireland) Act (1998) and engages with the relevant communities. To determine whether a parade is likely sensitive, the Parades Commission evaluates the passage of parades through or next to areas with a ‘majority of inhabitants with different tradition’, the timing of the parade (parades in the evening are usually more likely to be considered sensitive), and plans for loud noises (for example certain songs or the use of drums). The Parades Commission may impose conditions on sensitive parades, for example by altering their route, the number of participants, or the types of songs played. In practice, parades that are likely to affect areas of out-group majority are usually deemed sensitive, which is why we focus on these parades for our tests. Finally, we code whether parades were organized by Protestant or Catholic organizations based on words used in the names of the parade organizers (see Appendix A.1 for more on the dataset and coding of relevant variables).
Figure 1 shows temporal variation in total number, Protestant, Catholic, and sensitive parades between 2002 and 2022. The number of parades drops drastically in 2020 due to the COVID pandemic, when lockdowns and restrictions on gatherings were imposed across the United Kingdom, although there were some parades in 2020 with restrictions on participants (BBC 2020).
Figure 1. Temporal variation in the total number, ‘sensitive’, Protestant, and Catholic parades.
The Parades Commission (2023) provides the town in which a parade takes place and the parade route (in the form of a list of road names), but they do not provide geolocated spatial data. We automated the geolocation of parades using the Google Maps API algorithm to identify the starting point Footnote 4 of each parade and save this point as spatial information in our final dataset. Footnote 5 The spatial distribution is shown in Figure 2. Further details on the procedure to detect parade routes can be found in Appendix A.1.
Figure 2. The start points of parades and processions across Northern Ireland (2002–23).
We use the ICP dataset in a mixed-methods design with qualitative fieldwork and regression analyses to understand the effect of parading on intergroup attitudes and to trace our main mechanisms. Table 1 provides an overview of the different components of our research design.
Table 1. Overview of the mixed-methods research design
First, we combine the ICP dataset with multiple rounds of the British Election Study (BES) survey to assess if respondents in constituencies that have experienced contentious parades are more likely to have negative attitudes towards the out-group and to establish if there is a temporal downward trend in these areas. To capture attitudes towards the out-group, we use respondents’ views on intergroup mixing as our outcome variable, as assessed by their attitudes towards mixed marriage and mixed schooling. Mixed schooling, in particular, is considered important for overcoming conflict-era cleavages, but most schools in Northern Ireland remain segregated.
After establishing the association between parading and attitudes towards intergroup mixing over time, we turn to survey data collected in the summer of 2022 around ‘The Twelfth’ to provide a quasi-causal estimate of the short-term effect of parades on intergroup attitudes, this time focusing on attitudes towards living in mixed neighbourhoods. We exploit the fact that whether a respondent was interviewed before or after 12 July is plausibly exogenous and allows us to estimate the short-term effect of parades during one ‘marching season’ on intergroup relations. Footnote 6 While we cannot fully identify a causal effect with our survey data, we understand the different design elements in this paper as a way to triangulate the impact of parading on intergroup relations in Northern Ireland. In a last step, we turn to qualitative evidence from participant observations and interviews during the 2023 parade season in Belfast, which, in combination with qualitative secondary sources, we rely on to trace our proposed mechanisms. We discuss the methodological details of each research design component when we present the results.
We pursue this mixed-methods research design for several reasons. We recognize that we are studying complex social phenomena and thus employ a range of research methods to combine their strengths (Johnson and Onwuegbuzie Reference Johnson and Onwuegbuzie2004) at different levels of analysis (Lieberman Reference Lieberman2005; Brady and Collier Reference Brady and Collier2010). The location and routes of parades are not random and, indeed, often coincide with demographically contested areas. As such, estimating the causal effect of parades is challenging. We provide a quasi-causal effect estimate by exploiting the plausibly exogenous timing of interviews in the 2022 survey. However, this captures only the short-lived effect of parades on attitudes, while parades are a much bigger phenomenon reoccurring every year and stretching from the negotiation of parades routes, over the actual parading days, to the clean-up and media discussion after the parades. It is important to put this analysis into context and understand how parades affect attitudes over time, which we do in correlational longitudinal regression analyses, and what the mechanisms are that drive deteriorating group attitudes, which we do through our qualitative research.
Analysis and Results
Pooling of Northern Ireland Surveys over Time
Our first aim is to establish whether there is a general correlation between exposure to parades and negative attitudes towards intergroup mixing. We also want to investigate whether these associations become stronger over time, indicating greater affective polarization (Iyengar et al. Reference Iyengar, Sood and Lelkes2012; Mason Reference Mason2015).
We combine the geolocated parade data from the ICP dataset with survey data from the BES (UK Data Service 2024), a long-running social science study on political and social values that takes place immediately after every general election in Great Britain and Northern Ireland. For our study, we combine nationally representative surveys fielded in 2003, 2010, 2015, 2017, and 2019, which amounts to a sample size of almost 7,000 respondents. Footnote 7 The time period is chosen because (1) it is the post-conflict period and (2) the BES is conducted at general elections.
We explore the association between exposure to the number of sensitive parades in a respondent’s constituency and views on intergroup relations. To measure the independent variable (number of sensitive parades), we use the constituency level as spatial aggregation because this is the lowest administrative level for which spatial information is consistently provided for analysis from the BES across all years.Footnote 8 We count only parades deemed ‘sensitive’ by the Parades Commission, as these are parades that are likely to lead to intergroup tensions and likely to pass through or by out-group areas. As an outcome measure of intergroup relations, we focus on responses to two questions asked in every survey since 2003. First, respondents are asked whether they would mind a lot, a little, or not at all if one of their close relatives were to marry someone of a different religion. Second, they were asked whether they would prefer to send their children to a school with only their own religion or a mixed-religion school. These are two salient intergroup attitudes in Northern Ireland, which we discuss in more detail in Appendix A.5.
We use linear regression models in which we regress attitudes towards mixed marriage and mixed schooling in two separate regressions on the count of sensitive parades that occur in respondents’ constituencies. As shown in Figure 3, parading varies over time and space. Our regression models include the total number of parades and respondent demographics as controls. The respondent demographics are age, gender, level of education, and religion. All models include year fixed effects.Footnote 9
Figure 3. Spatial and temporal variation of ‘sensitive’ parading during the years of the British Election Study surveys across constituencies of Northern Ireland. Darker shading represents a higher number of parades considered sensitive by the Parades Commission.
The results of the analysis are shown in Table 2. Footnote 10 Our key independent variable is negatively associated with attitudes towards mixed schooling (column one) and mixed marriage (column two). Generally, respondents living in areas with high numbers of contentious parading are more likely to mind if they or a close relative married someone of a different faith and would prefer to send their children to schools with their own religion only. We uncover a negative association between contentious parades and attitudes towards the out-group.
Table 2. Association between the number of contentious parades and attitudes to mixed schooling and intermarriage between Catholics and Protestants
*** p < 0.001; **p < 0.01; *p < 0.05; p < 0.1.
To explore the relationship between contentious parading and intergroup attitudes over time, we reran both models separately for each year. Again, the models include demographic controls and the total number of parades per constituency. We only remove the year fixed effects. The results are shown in Figure 4. The coefficients for the number of contentious parades (in 100s) are shown as points with 95 per cent confidence intervals. We conclude that there is a negative association between sensitive parades and intergroup attitudes that appears to have gotten stronger over time for mixed schooling.
Figure 4. Modelling contentious parading and intergroup attitudes. Coefficients for the number of contentious parades (in 100s) are shown.
2022 Survey
The findings of the longitudinal analysis demonstrate a negative association between parades and intergroup attitudes. This analysis is limited in several ways; key among them is the endogenous nature of intergroup attitudes and contentious parading, limiting our ability to estimate the effects of parade exposure. To overcome this limitation, we measure the quasi-causal short-term effect of parade exposure at a more granular level.
We use an original survey that was fielded between May and September 2022 in 32 urban neighbourhoods in Northern Ireland (N = 1,024) as part of a project on paramilitary social control that also captures intergroup relations. These areas were selected because they are urban, relatively deprived, and segregated – and these are areas where we would expect paramilitary groups to still be active. Footnote 11 They are often the spaces where both communities meet (so-called interface areas) and where sectarian divides are most pronounced. Based on the planning of the survey firm implementing the survey, we assume that whether a respondent in these communities was interviewed before or after 12 July is ‘as-if random’ and not related to parading or our outcomes, and so we can compare the average treatment effect of the peak of the parading season on attitudes towards the out-group. We do not claim that this is a causally identified estimate, but this evidence helps to understand if there is a short-term effect of one parading season on intergroup attitudes.
We consider respondents as treated if they were interviewed after 12 July. Footnote 12 Just under 25 per cent of respondents are interviewed before ‘The Twelfth’ (229 respondents) (top pane of Figure 5). Our main outcome variable is the answer to the following question: If you had a choice, would you prefer to live in a neighbourhood with people of only your own religion, or in a mixed-religion neighbourhood? Lower scores indicate a preference for living with people from the respondents’ own religion. The bottom pane of Figure 5 visualizes this approach. Descriptively, those interviewed after ‘The Twelfth’ are less likely to report wanting to live in mixed neighbours (58 per cent) compared to those interviewed before (63 per cent).
Figure 5. The top pane shows the number of respondents interviewed before (darker spots) or after (lighter spots) ‘The Twelfth’. The bottom pane shows the average support for living in mixed communities before (darker spots) and after (lighter spots) ‘The Twelfth’.
Comparing respondents before and after ‘The Twelfth’ is a quasi-experimental setup because we assume that the date on which respondents were interviewed is exogenous to ‘The Twelfth’ celebrations. While our sample is relatively balanced without explicit randomization, balance tests indicate that those interviewed before ‘The Twelfth’ in our sample were less likely to be Protestant and reported lower household income. While statistically significant, the income differences between our control and treatment sample are substantially small. Footnote 13 In our regression model of the effect of ‘The Twelfth’ on intergroup attitudes, we therefore include demographic controls (age, gender, education, and income) to account for potential confounding that could bias the estimated effect. We cannot rule out seasonal trends in people’s attitudes over the course of the summer data collection and hence we limit our sample to respondents in the month before 12 July and compare them to respondents interviewed after with an increasing window approach. Footnote 14 Specifically, we compare the same control group with respondents interviewed in temporal windows after 12 July increasing by weekly steps. The results of this analysis are visualized in Figure 6. The estimates should be interpreted as associational or quasi-causal rather than full causal evidence given the slight imbalances in our data.
Figure 6. Effect of ‘The Twelfth’ celebrations on support for living in mixed communities with an increasing windows approach.
Being interviewed after 12 July is negatively associated with views on intergroup mixing. When we subset the analysis to only those who identify as Catholic (shown in the middle pane) or Protestant (shown on the right), the coefficients remain negative, but the effect size decreases in more inclusive time windows. For Protestant respondents, the effect is no longer significant just one month later, while the effect on attitudes for Catholic respondents appears to persist beyond the end of the study period. Combined, these findings provide suggestive evidence that ‘The Twelfth’ celebrations are associated with a short-term decline in intergroup attitudes, which may help explain the over-time polarization observed in our longitudinal analysis.
Fieldwork during ‘The Twelfth’: Participant Observation and Interviews
Our quantitative analyses do not allow us to test whether pride and provocation are the mechanisms that drive our findings; hence, we supplement these analyses with a qualitative case illustration. To empirically examine the mechanisms through which contentious parading affects intergroup relations, we rely on both secondary qualitative sources and fieldwork in Belfast from 7–17 July 2023. Over this period, we conducted eight interviews with relevant stakeholders and participant observation at four parades (Table 3). We conducted interviews and observations in and around Shankill, Falls Road, Donegall Street, Tiger’s Bay, and the city centre. We chose these areas based on the concentration of parade routes, incidents of past sectarian violence, and vicinity to out-group dominant neighbourhoods (Figure 7).
Table 3. Interview details
Figure 7. Map of sensitive parades in July 2022 in Belfast (black lines). We added the rough location of our interviews with dots and diamonds corresponding to CNR (Catholic, Nationalist, Republican) and PUL (Protestant, Unionist, Loyalist). The shade represents whether the dominant religion in an area is Catholic or Protestant, based on the 2021 Census (NISRA 2022).
The interviews focused on how participants experience contentious parades and what they believe is the aim and effect on in-group and on intergroup dynamics. The interviewees were selected through snowball sampling. Beyond this, we had ‘informal chats’, or shorter ad hoc interviews, with several individuals who were informed of the purpose of our talks and verbally consented. With these informal chats, we aimed to identify the discussions surrounding parades among individuals living in parading areas, enabling insights into locals’ raw reactions to the parades. We then analyse the notes from the interviews (not including informal chats) using thematic analysis. We identified themes from the notes, while being guided by our theoretical mechanisms. In this way, we looked for the presence (or absence) of themes of provocation and pride in the interviews, but we allowed ourselves to identify other themes inductively, based on our coding of the notes (Strauss Reference Strauss1987; Saldaña Reference Saldaña2021).Footnote 15
We complement this interview strategy with our attendance at contentious parades in the lead-up to and on the day of 12 July and engagement in participant observation. Footnote 16 There are multiple parades in Belfast on ‘The Twelfth’. We selected contentious parades that were accessible without imposing high levels of risk for the researcher and fulfilled the following criteria: the parade (1) was deemed sensitive by the Parades Commission, (2) has led to counter-protests by local resident groups in the post-conflict period, and (3) planned to march through predominately Catholic areas. Footnote 17 During the parades, we observed the unfolding of events, including the behaviour of participants and onlookers (both from the in-group and out-group), symbols on show, and songs played. We employed a passive form of participation, following the parades and observing their unfolding but without directly engaging in the festivities or demonstrations. This participant observation allowed us to gain a first-hand understanding of the dynamics surrounding contentious parades in Northern Ireland, to observe this practice without the need for respondents to articulate their culture, and to assess if the theorized mechanisms find some accordance with the qualitative evidence (cf. Kapiszewski et al. Reference Kapiszewski, MacLean and Read2015; Fox Reference Fox2004). We triangulate evidence from multiple qualitative methods during our fieldwork – including formal interviews, informal discussions (Patton Reference Patton2002), and participant observation – and secondary qualitative sources. These components provide metadata that contextualize the interviews and deepen our understanding of the case (Swain and King Reference Swain and King2022).
The pride mechanism was evident among Protestant respondents throughout our interviews and participant observation. Interviewees claimed that paraders feel a strong sense of pride in participating in activities that build a sense of community and celebrate their shared British identity. This relationship – pride builds community – is what we expected theoretically. According to parade organizers, pride comes from participating in the parades but also from other activities during the parading season, such the building of Loyalist arches and massive bonfires in their neighbourhoods. Footnote 18 ‘The Twelfth’ is seen as a ‘demonstration of a people, of a way of life, of an ideal’ (Interview 6). This feeling is often related to the support felt by family and friends, who usually cheer the paraders at the sidelines (Interview 6), or the memories of past participation in the parades (Interview 8). In fact, being part of the Orange Order is passed down generations, which is exemplified by the popular song ‘The Sash My Father Wore’ (Interview 6).
That our interviews and participant observations revealed parades as moments of great joy and celebration within the Protestant community is consistent with other sources: one of Blake’s (Reference Blake2019 a, p. 91) interviewees expresses that to them, Protestantism is more about ‘just the pride and the understanding of being a member of it’. Blake (Reference Blake2019 a, pp. 91–4) also reports Protestant interviewees boasting about the fame of the parading band they play in and asserting that parades are a way to cherish and confirm their group identity. In fact, these individuals consider parading to be an integral part of being a Protestant – which is also apparent in our own interviews and other sources (see, for example, Devine-Wright Reference Devine-Wright2001; Graham Reference Graham2022; Rainey Reference Rainey2023).
We found that Protestant interviewees invariably mentioned that relations between the Protestant and Catholic community are damaged when parading is frustrated by imposed restrictions and contestation. They stressed the overwhelming support for parades among residents of traditionally Protestant areas. When following Loyalist parades through Belfast, we saw variation in levels of support. Some residents of predominantly Protestant areas would look onto the parades from their homes, and sometimes they would cheer and support on the roadside. In certain areas (for example city centre and Shaftesbury Square), the number of supporters was massive. Here, many onlookers were decorated with Union Jacks and sometimes flags and symbols associated with Loyalist paramilitary organizations. In sum, we find evidence of a pride mechanism, but only among members of the Protestant community.
The theme of provocation was present in many of our interviews. Catholic interviewees – the out-group of ‘The Twelfth’ parades – argued that some Loyalist parades aim to provoke a reaction. One interviewee claimed that a reaction ‘helps [the] narrative of culture under attack’ espoused by parts of the Protestant community – the in-group (Interview 1). As evidence of the link between a provocation mechanism and greater in-group favouritism, one respondent claimed that contestation around Loyalist parades creates rallying points for Catholic mobilization – the examples of the Ormeau Road and Drumcree parades, which were banned by the Parades Commission, are important (Interview 1). Footnote 19 In the case of Drumcree in particular, the Catholic community vehemently opposed the passing of such parades in Catholic-majority areas, while the Orange Order would attempt to circumvent the rerouting of the parades, often leading to standoffs with the police and Nationalists rioting (Bryan Reference Bryan2000). Another interviewee noted that if parades were to run through his immediate area, ‘the people [would] unite to defend the children and the community’ (Interview 3). One interviewee believed that although some parade bands aimed to intimidate, and some bands and lodges retain controversial ties to paramilitary culture, ultimately the contention is performative and choreographed on both sides (Interview 5). Another believed that parades are not sectarian per se but are often seen as such by the Catholic community (Interview 4). They also noted that the return routes become rowdy, as the participants are often intoxicated (Interview 2 and 4). In informal chats, Catholics referred to bonfires burning images of Sinn Fein politicians and the participation of ex-paramilitary members in the parades as provocation.
The use of ‘provocative/illegal flags’ is another example of provocation tactics that are being contested by the Catholic community (Parades Commission 2007, p. 12). Moreover, each year a sharp increase in sectarian incidents is registered around 12 July (Police Service of Northern Ireland 2024, p. 15). The informal chats also highlighted the role of social media in heightening indignation; for example, participants showed us a video of a scuffle between a disruptive individual and the paraders in Ballycastle (Scott Reference Scott2023). Although these Catholic participants saw the video as evidence of parades’ sectarian attitude, Protestants saw this as an orchestrated attempt to provoke a reaction from the paraders. The sharing of social media posts may disseminate the effect of parades widely beyond the immediate vicinity of where the parades take place (Reilly Reference Reilly2021). In sum, we found evidence of a provocation mechanism, sometimes linked to a sense of a need for in-group mobilization, among members of the Catholic community (especially those living near the parade routes).
However, this mechanism was also – perhaps even more – apparent among the Protestant in-group, which we did not expect. For instance, one interview claimed that mobilizing the residents against parades is a way for Republican activists to attack ‘Britishness’ and provoke a negative response from the Protestant community, which would then cast the community in a bad light (Interview 6). Interviewees also saw this as a strategy deliberately adopted by Sinn Fein and other ‘dissident Republicans’ once the armed struggle ended (Interviews 7 and 8). Footnote 20 This theme is closely linked to the theme of ‘choreography’ – interviewees expressed the view that contestation is often orchestrated as a means to launch political attacks on the Protestant community. These findings confirm what was already reported by Jarman and Bryan (Reference Jarman and Bryan1996) on the issue of contestation surrounding the parades. This emerged as a common narrative among Protestant interviewees: the Catholic community, having almost entirely abandoned the armed struggle save for a few dissidents (McKearney Reference McKearney2011), is moving to limit Protestant power in the political arena, and contestation around parading is part of this strategy. Finally, we also uncovered a link between provocation and in-group favouritism among the Protestant community. Specifically, one respondent claimed that protests and attacks against parades created a rally-around-the-flag effect in defence of their identity, customs, and values (Interview 8). An interviewee conceded that ‘our people are only too willing to react negatively to the provocation [by Nationalist protests]’, and expressed their disapproval of sectarianism (Interview 6). Footnote 21 Evershed (Reference Evershed2018) reports similar findings, demonstrating that members of the Protestant community feel that the contentiousness of parades derives from the conscious efforts of the Republican movement to foster a cultural conflict and mobilize Catholics in support of their policy agenda. According to this view, intergroup tension is the product of the mobilization of communities by ideological entrepreneurs (Staniland Reference Staniland2010; Blake Reference Blake2019 b) (c.f. Benford and Snow Reference Benford and Snow2000; Adarves-Yorno et al. Reference Adarves-Yorno, Jetten, Postmes and Haslam2013; McLamore et al. Reference McLamore, Levi and Leidner2019). Protestant interviewees admitted that members of their own community were easily provoked into a response. The existence of a provocation mechanism within both groups, as opposed to just among the out-group, is an important finding because it may help explain how parades lead to escalation dynamics and tension between the groups.
We note two recurring themes, supremacy and domination, which shed light on why parading is central to contestation between both groups. Catholic participants (both in interviews and informal chats) drew parallels between the Orange Order and other prominent organizations involved in Loyalist parading and white supremacist organizations such as neo-Nazis and the Ku Klux Klan. They claimed that these ideologies have negative effects on intergroup relations. The Orange Order as an organization does not allow membership for individuals who are not of the Protestant faith. This, along with the history and culture surrounding the group, was considered proof that the organization is supremacist. Loyalist parades into predominately Catholic areas are, therefore, seen as promoting Protestant supremacy and triumphalism (Ross Reference Ross2007; Blake Reference Blake2019 b). During the larger contentious parade that we attended, an onlooker standing outside the Ardoyne shops – a highly contested stretch of the Crumlin Road – told us that the paraders were all ‘bigots’. Previous incidents witnessed by the participants first-hand – such as the planting of Union Jacks in Catholic families’ backyards and the shouting of sectarian insults and songs outside of St. Patrick’s Church on Donegall Street (Interviews 1 and 2) – were presented as proof of the sectarian character of the parades.
The theme of historical (but fading) domination was also apparent. Some Catholic participants claimed that because the Protestant community is losing its dominance in the political and economic landscape in Northern Ireland, it uses parading as an attempt to conceal the group’s waning power. The parades, then, are a way to continue reaffirming Protestant dominance and to demonstrate an aspiration for the Protestant community to regain its lost power. Therefore, members of the Catholic community see the increasing number of parades as a sign of the out-group’s waning influence (Interviews 1 and 2). Some Catholic interviewees opposed parade routes that went through and around Catholic areas. Footnote 22 These interviewees generally claimed that they were not inherently against parading, but they opposed them when the routes included Catholic areas (Interviewees 1, 2, and 3). They stated that parades were used by the Protestant community to intimidate members of the Catholic community and to reinforce Protestant dominance, which – according to these same interviewees – is waning. Therefore, parades are seen as inherently promoting ideals of domination over Catholics. Existing research also reports that Catholics view the Orange Order and the parades as triumphalist reminders of past domination (Jarman and Bryan Reference Jarman and Bryan1996; Ross Reference Ross2007).
In sum, both our fieldwork and secondary sources suggest that our proposed mechanisms – provocation and pride – are plausible. Pride in participating in the parades and related activities, as well as in seeing relatives and friends cheering on the sidelines, seems to bolster one’s identification with the in-group and one’s sense of being part of a community. Conversely, the parades themselves generate fear or resentment among Catholics, who see them as an attempt to provoke a reaction and reproduce themes of domination. Similarly, some Protestants view the contestation surrounding parades as an attempt to provoke a response in the Protestant community, ultimately with the goal of suppressing the community’s cultural identity. Indeed, all Protestants (and even some Catholics) interviewed see Catholic opposition to parades as manufactured contestation aimed at attacking their traditions and identity. Both communities accuse each other of attempting to incite a negative response from the other side and view the other community’s position as regressive and uncompromising. The themes of supremacy and domination are important here. While Protestant participants consider parades an expression of their cultural identity, some of the Catholic interviewees see them as an expression of supremacist views. This is important because it is the perceived supremacy and domination that makes these rituals contentious, as opposed to simple rituals.
Discussion and Conclusion
Public rituals – ceremonies, commemorations, or parades filled with symbolism – help individuals make sense of their place in the world by drawing links between the past, present, and future. Taking part in rituals can help create solidarity and a sense of belonging even when participants do not have uniform beliefs (Kertzer Reference Kertzer1988). However, in trying to foster social unity, rituals also reveal or even create contrasts to ‘the other’, particularly in divided societies, including post-conflict societies such as Northern Ireland. This seems to be the case for parades in Northern Ireland, particularly ‘The Twelfth’ parades organized by the Orange Order. To members of the Unionist community in Northern Ireland – primarily of Protestant background – these parades are important community-building rituals. Yet to many members of the Nationalist community – primarily of Catholic background – these parades are seen as rituals of provocation and domination. Orange Order parades were a source of sectarian tension during ‘the Troubles’ and, for some, remain so today, more than 25 years after the conflict officially came to an end.
Anchored in long-standing bodies of research on social movements and social psychology, we demonstrate in a fine-grained quantitative test that parades in Northern Ireland are associated with negative intergroup relations. Using qualitative evidence, we show that while parades in Northern Ireland create community cohesion among Protestants by fostering in-group pride, which in itself can contribute to scepticism towards the out-group, they also have a deteriorating effect on intergroup relations through a provocation mechanism. Therefore, we find that group processes defined in the social identity literature play a role in ritualistic commemorations. While this is a common interpretation of parades and interethnic relations in Northern Ireland amongst researchers of this case, we bring to the table a wealth of systematic data that enables rigorous empirical analyses.
First, we create a novel dataset of geolocated data on the routes of over 55,000 parades from 2002 to 2022 in Northern Ireland. By combining these data with election surveys, we can assess how residents’ proximity to parading routes is associated with intergroup attitudes over time, focusing particularly on people’s attitudes to mixed schooling and marriage. This pooled survey analysis allows us to assess the effect of parading between areas with low and high parading activity in Northern Ireland. Second, we use an original 2022 survey from predominately Protestant and Catholic urban communities, leveraging the plausibly exogenous variation timing of interviews – some conducted prior to 12 July and some after – to estimate the effect of parade exposure on attitudes towards living in a mixed area. This analysis provides evidence on the temporal effect of parading as identified in the 12 July celebrations. Finally, based on participant observation of contentious parades and interviews with residents and key community leaders, we further assess the plausibility of the hypothesized mechanisms of pride and provocation. We find suggestive evidence that exposure to parades fosters scepticism towards interethnic mixing, driven by our hypothesized mechanisms, though further work can disentangle how in-group pride shapes attitudes to the out-group – particularly in light of our finding that the negative effect of parades is driven primarily by members of the Protestant community.
While the combination of our empirical approaches indicates a negative effect of parades on intergroup relationships, the individual analyses cannot conclusively test the mechanisms through which parades shape attitudes, in particular due to lacking systematic quantitative data on perceived pride and provocation. Most importantly, we do not claim that parades are the only or the most important factor shaping politics in Northern Ireland, as the tensions between Protestants and Catholics are multifaceted and have long historical roots. But because symbols and rituals are often central to conflict processes, particularly in conflicts in which ethnic identity groups mobilize against one another, understanding how rituals may contribute to the endurance of interethnic cleavages in post-conflict societies is important for both theory and practice.
Our study can speak to similar phenomena in other divided societies. As mentioned in the introduction, in Israel, the annual ‘Dance of Flags’ is a parade attended by far-right Israelis to celebrate the annexation of East Jerusalem in 1967. This parade takes place in Jerusalem’s Old City (which is predominantly Arab) and often sparks violence with local inhabitants and protesters (Ben-Amos Reference Ben-Amos and Bartov2021; The Times 2024). Marches in India are often catalysts of violence between Hindu nationalists and Muslims (BBC News 2022), and the potential tensions extend beyond parades within India. At the 18 August 2024 India Day Parade in New York City, the inclusion of a float with a large-scale model of the newly built Ram Temple – which was built on the site of the 500-year old Babri Mosque that was destroyed by Hindu nationalists in 1992, leading to episodes of violence killing thousands – was highly controversial. Whereas those organizing the float emphasized how it celebrated Hindu culture, opponents from the Muslim community said it was a symbol of Islamophobia and anti-Muslim ideology meant to spread division (Keleshian Reference Keleshian2024; Stack and Bayya Reference Stack and Bayya2024). In Bosnia and Herzegovina, the ‘Republika Srpska Day’ military parade, which has been deemed unconstitutional since 2015, has been a source of tension between the Serb and Bosniak communities.
Given the similarity of the social context and the one-sided commemorative nature behind these examples, our findings speak to ‘contentious rituals’ beyond Northern Ireland – both in other post-conflict societies, particularly where the conflict itself pitted identity groups against one another, and in divided societies more generally.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S0007123425100835
Data availability statement
Replication data for this paper can be found in Harvard Dataverse at https://doi.org/10.7910/DVN/47SXKJ.
Acknowledgements
We thank Dale Pankurst and Andrew Thomson for their invaluable input during our fieldwork. We would also like to thank participants of the annual conferences of the Swiss Political Science Association and International Studies Association for their helpful comments, as well as members of the Conflict & Change research cluster at University College London. The original survey data is part of the project ‘Street-Level Autocrats: Individual Decisions with Collective Consequences’, which is funded by the Research Council of Norway (PI: Kristin M. Bakke; co-PIs: Marianne Dahl, Scotte Gates, Håvard Mokleiv Nygård, and Pavel Baev). Our fieldwork was made possible by a Small Research Grant from the British Academy/Leverhulme, and we thank Andrea Ruggeri for support in the application process.
Author contributions
We consider Kit Rickard and Giovanni Hollenweger to be joint first authors.
Financial support
This work was supported by the Research Council of Norway (FRIPRO) (grant no. 275404) and the British Academy/Leverhulme (grant no. SRG2223\231259).
Competing interests
None.
Ethical standards
The research was conducted in accordance with the protocols approved by University College London (no.4931/002).
Open access
