Reclaim, occupy, pillow fight!: movement continuity in the Urban Playground Movement’s Budapest scene

ABSTRACT Flash mobs are commonly portrayed in the media as light-hearted events intended to surprise bystanders. Meanwhile, academic literature features a number of explicitly political interpretations of flash mobs. However, to date, no empirical study has addressed why individuals organize and take part in the internationally spread mass events organized by the most popular contemporary flash mob incarnation, the Urban Playground Movement (UPM). In this paper, we place Budapest’s UPM on an entertainment practice versus social movement mobilization spectrum. Our interviews with the leaders of the largest Hungarian UPM organization show that the functioning of the UPM in Budapest heavily relies on social movement continuity, where organizers see their events as ways to reclaim public spaces and to prefigure more social cities. However, among participants surveyed at the most notable yearly UPM event, the International Pillow Fight Day, we found fun as the primary motivation for participation, thus showing movement discontinuity. We conclude that the UPM in Budapest is a collective action, but not a social movement. Secondly, we find that the UPM exhibits ‘thin’ diffusion both in its reliance on individual activists over coalitions and in its usage of commercial social media as its diffusion infrastructure. Lastly, our results show that UPM collectives can act as ‘containers’ of social movement frames, implying that these playful groups could function as vehicles in the spread of protest in movement milieus. Our key finding that UPM is not a social movement contradicts the numerous academic interpretations of flash mobs as outright political events.

2005, p. 12). However, researchers have attached a range of values and socio-political motivations to the acts, and numerous political interpretations appear in sociological literature. Sociologists have identified flash mobs as Dadaism-inspired critiques of capitalist spectacles (Nicholson, 2005, p. 12), expressions of the key values of anarchism (Mattern, 2016), and as spatial resistance against the neoliberalization of urban spaces (Molnár, 2013;Plyushteva, 2009). While a handful of small-scale studies address certain aspects of the phenomenon (e.g., Grant et al., 2015;Molnár, 2013;Nicholson, 2005), the motivations of organizers and participants of the most popular contemporary flash mob incarnation, the Urban Playground Movement (UPM), remain empirically unexplored. We intend to fill this gap by presenting the results of a mixed-method inquiry that locates the motivations of both the leaders and the participants in UPM's largest annual event, the International Pillow Fight Day, in the context of the event in Budapest, Hungary. The research study's aim is placing Budapest's Pillow Fight event on an entertainment practice vs. social movement mobilization spectrum, through which we can better understand to what extent the Urban Playground Movement falls into the category of social movement.
We explore the motivations behind UPM activities from the perspectives of the city's current UPM collective, Budapest UP!'s organizers and as well as those who joined the Pillow Fight Day as event participants. We assess whether they are involved in conflictual relations with clearly identified opponents, as this would mean that the UPM would be best categorized as a social movement, following the current mainstream understanding of the term (Della Porta & Diani, 2006, pp. 20-23) that we refer to in our Theoretical Framework section. Our investigation is also informed by theories of diffusion, with special reference to the spread of protest in the Global Justice Movementa large network (also called the movement of movements) that was fighting against corporate/ neoliberal globalization and for a world with less social inequality (e.g., Della Porta 2007)and in the movements of the post-2008 crisis, such as Occupy Wall Street, Indignados, and the Arab Uprisings, because of the cultural similarities that our case shares with these mobilizations. We proceed with a short introduction to social movements in Hungary and the UPM in Budapest. Subsequently, we present our mixed methods research design in the 'Exploring the Budapest Scene of the Urban Playground Movement' section. Both our qualitative and our quantitative fieldwork took place in Budapest in 2014, prior to and during the biggest annual UPM event in the country, the International Pillow Fight Day. We examine the motivations of the Budapest UP! leadership through a thematic analysis of two paired interviews where we concentrate on organizational self-definitions and conceptions of public space. Then, in our 'Findings' section we present the results of a survey of the Pillow Fight participants (N = 111) conducted during the 2014 International Pillow Fight Day event in Budapest. By juxtaposing the leadership interviews with the participant surveys, our results show that individuals have a range of motivations from political activism to pure fun. Organizers present social movement mobilization frames, which are absent for participants, who are fun oriented. Confronted with contrasting images of politically-motivated leaders and fun-motivated participants, we conclude in our 'Discussion' section that the UPM in Budapest is not a social movement as a whole. Our results indicate that the Urban Playground Movement can only be partially described by the political interpretations presented in flash mob literature. While the collective action's functioning relies a great deal on movement continuity on the level of the left-libertarian, activist-minded leaders, its actual execution relies on participants who simply want to have fun.

Theoretical framework
Social movement and other conceptualizations of the Urban Playground Movement Due to their visual and surprising nature, there is an abundance of media coverage of UPM events; however, there are few scholarly investigations focusing specifically on them, and academic accounts of similar phenomena largely lack empirical grounding. The closest conceptual roots are found in the critic and writer Howard Rheingold's (2002) concept of smart mob, which describes events characterized by highly cooperative, autonomous networks of people empowered by digital technologies like SMS messaging or mobile internet-based communication. Smart mob examples included teenagers' pass-time meet ups in Tokyo, anti-G8-protests in the Battle of Seattle, and neighborhood gift-economy initiatives.
The smart mob concept influenced Sean Savage, whom Judith Nicholson credits with coining the notion of flash mob (Nicholson, 2005), to name the by now well-known social media-driven events as such. Savage saw them as events intended to disrupt everyday urban functioning for short time periods by performing surprising actions publicly, without any goal. However, artistic and political flash mob interpretations also emerged among observers and participants. Some of these accounts perceive flash mobs as Dadaism-inspired critiques of capitalist spectacles or as spatial acts of resistance against post-9/11 fear (see Nicholson, 2005, p. 12). Mark Mattern even portrayed them as artistic expressions of core anarchist values like autonomy and horizontality (2016). Mattern (2016) believes that flash mobs, alongside DIY punk music, poetry slam, graffiti, and street art, prefigure the anarchist dream of decentralized power. He claims that, together with similar artistic forms, they represent a new kind of social movement (p. 138). In Mattern's view, flash mobs are involved in 'prefiguring anarchist resistance to domination, flash mobs challenge the established, conventional order that is structured and controlled by state and corporate power ' (p. 113).
This political line of flash mob interpretation is also prevalent in Virág Molnár's work, but she links it specifically to the Urban Playground Movement (2013). She is critical of flash mob accounts that emphasize their novelty and their reliance on digital media. Instead, Molnár explores their historical roots in analog forms of 'urban pranksterism', including the city-focused activities of the early 20 th century Dadaists and Italian Futurists, and of the mid-century Situationist International. She builds a contemporary typology starting with atomized flash mobs, the events that 'hardcore flash mobbers' or 'flash mob purists' consider the most authentic. Molnár accepts the New Yorker journalist Bill Wasik's narrative portraying these events as his own intellectual children (2006). Wasik aimed to create surprising 'l'art pour l'art' gatherings representing a 'cynical in-joke on scenesterism,' a kind of 'anti-expression.' 1 These popular, at times dramatic, visual events quickly attracted the attention of corporate and political campaign designers, giving birth to advertising flash mobs and political flash mobs. 2 Furthermore, performance flash mobs also appeared, which are artistic actions that typically celebrate urban lifestyle and are carried out by professionals. Molnár asserts that interactive flash mobs, which involve interaction among both participants and bystanders, represent the most popular type. They typically take a children's game, like a pillow fight or capture the flag, and make it a large urban event by mobilizing hundreds of people through social media.
Molnár mentions that some regard interactive flash mobs as apolitical, but those observers and organizers who use the Urban Playground Movement term to refer to them argue that interactive flash mobs 'combine whim and serious social commentary: they want to create fun but also to "reclaim" public space that is otherwise often choked with tourists and overtaken by commercial uses' (Molnár, 2013, p. 50). The leading Urban Playground manifesto, published by the New York-and Toronto-based Newmindspace group, asserts that 'the urban playground [movement] is growing around the world, leaving more public and more social cities in its wake' (Newmindspace, n.d.). In line with the ideology-focused interpretations described by Molnár, Plyushteva's claim (Plyushteva, 2009, p. 91) is that pillow fights are the closest applications of Harvey's 'Right to the City' call-to-action (Harvey, 2008). This view places them conceptually in an international struggle against urban neo-liberalization that commodifies the city and subordinates it to corporate interest.

Distinguishing characteristics and diffusion of social movements
Based on the manifesto of Newmindspace and in line with Plyushteva's account, when UPM leaders create their events to transform urban functioning, their goals are simultaneously demonstrative and prefigurative. They are demonstrative because they convey a message about social isolation of citizens that is reproduced by a dysfunctional public space, and they are prefigurative because they temporarily create an urban space conducive to interaction. Reclaim the Streets might be the most apparent precursor to the leaders' experience of UPM both in terms of its urban space criticism and its fun-focused occupation tactic (Carmo, 2012). This also entails a close resemblance to the larger 'protestival' scene characterized by a mix of serious social commentary and deliberately fun-centered forms of action most often discussed in relation to the Global Justice Movement (St. John & Graham, 2008). Furthermore, the discourse and practice of spatial reoccupation emerges from a rich international social movement tradition from the American Civil Rights Movement, the 1968 French student movement, and the Reclaim the Streets (RTS) collective in the UK all the way to the recent city-square movements. These different initiatives had their own separate causes, but spatial reoccupation became an integral part of these movements, both as a tactic and as an ideology. In the words of RTS activists: 'It's about reclaiming the streets as public inclusive space from the private exclusive use of the car. But we believe in this as a broader principle, taking back those things which have been enclosed within capitalist circulation and returning them to collective use as a commons' (Fourier, 2003, p. 54). The theoretical starting points of our empirical research are the elements of Della Porta and Diani's widely accepted definition, which conceptualizes social movements as 'a distinct social process, consisting of the mechanisms through which actors engaged in collective action are involved in conflictual relations with clearly identified opponents; are linked by dense informal networks; [and] share a distinct collective identity.' (Della Porta & Diani, 2006, p. 20, based on Diani, 1992. As we will see, the UPM consists of actors of collective action connected by dense informal networks that exhibit a distinct collective identity. But faced with the contrast between the main movement manifesto and political interpretations on one side and the overtly fun-oriented interactive flash mobs on the other, we seek to identify mobilization frames that would signal involvement in conflictual relations with clearly identified opponents. We set out to identify these conflictual relations to determine to what extent the UPM falls into the category of social movement. The latter tradition distinguishes social movements from other collective actions by their effort to advance or hinder social change by addressing specific political or social targets with political, economic, or cultural claims. It argues that, if a social movement's goals were achieved, it would damage the interests of the movement's adversaries. Conflict thus becomes the marker of social movements, separating them from consensus movements where 'sustained collective action does not take a conflictual element. Collective goods are often produced through cooperative efforts that neither imply nor require the identification of specific adversaries [. . .]. Prospected solutions do not imply redistribution of power nor alterations in social structure, but focus instead on service delivery, self-help, personal and community empowerment' (Della Porta & Diani, 2006, p. 23).
Social movements do not start from scratch when they begin articulating the conflictual relations in which they see themselves involved. They rely a great deal on existing frames, strategies, and repertoires of action that they creatively adapt to their own situation through the processes of diffusion. In this vein, scholars have also turned to analyzing processes that can sustain movement continuity between cycles of contention (e.g., Taylor, 1989). The ways in which diffusion occurs have long been subject of the study of social movements (Givan et al., 2010). With the advent of resource mobilization and political process theories, diffusion models moved from the micro-level process approach of early contagion-models towards a focus on connections between movements, organizations, and political environments (Soule, 2013, p. 349). Two large international protest waves of the recent past, the Global Justice Movement (GJM) mobilizations of the late 1990s and the post-2008 movements of the crisis, inspired the construction of new models of diffusion.
Contrasting these two international protest waves, Mattoni and Della Porta (2014) found that the GJM's 'thick' diffusion processes were replaced by 'thin' ones, which is a development that can be seen on multiple levels. Firstly, this distinction means that social movement networks and coalitions played a smaller role in recent protests compared to the GJM mobilizations, while individual activists gained more importance. Secondly, the dominant infrastructure of diffusion changed from activist mailing lists and alternative websites to commercial social media platforms. Finally, while GJM mobilizations were centred on transnational social fora, the movements of the crisis (like the Spanish Indignados or the Occupy Wall Street) functioned, to a large extent, through national and local city-square occupations. The open-air camps were autonomous communities set up by anti-austerity and pro-democracy protesters to prefigure the kind of society that activists wished to live in: those characterized by participative and deliberative decision making and inclusivity (e.g., Castañeda, 2012;Pickerill & Krinsky, 2012). Given the centrality of these occupations to their movements, research on diffusion tended to concentrate on the spread of the camps.
Our quest for traces of conflictual relations in the UPM is also a search for adapted frames that would link collective action to social movements. David A. Snow distinguishes collective action frames from everyday interpretation frames, stating that 'collective action frames not only perform an interpretive function in the sense of providing answers to the question "What is going on here?", but they also are decidedly more agentic and contentious in the sense of calling for action that problematizes and challenges existing authoritative views and framings of reality' (Snow, 2006, p. 385). In the following, we examine the framing of UPM leaders and participants seeking continuation withor difference fromother movements in the Hungarian context.

Social movements: the Hungarian case
When Pieter Vanhuysse analyzed rates of worker involvement in strikes and lockouts in post-transition Central and Eastern Europe (CEE), he showed that countries of the region in the early 1990s were more peaceful than Latin American states had been in the 1980s, and argued CEE countries were comparable to stable Western democracies (Vanhuysse, 2004). Béla Greskovits highlighted that it was a mix of structural, cultural, and institutional features of post-socialism in Eastern Europe that led to this political stability despite the deficiencies of both democracy and market economy (Greskovits, 1998). Grzegorz Ekiert's and Jan Kubik's protest event analysis of former East Germany, Hungary, Poland, and Slovakia in post-transition years until 1993 showed that Poland was the period's most contentious state and Hungary was the least eventful one (Ekiert & Kubik, 1998, pp. 555-556). They also demonstrated that all of these countries except for former East Germany exhibited a decidedly non-violent protest repertoire (Ekiert & Kubik, 1998, pp. 556-557). Máté Szabó (1995) identified a set of general trends of protest in 1993 in Slovakia, Slovenia, and Hungary, showing that protests were predominantly organized events, spontaneity played a marginal role. He also found NGOs to be the dominant actors that organized and financed the protests and showed that economic issues and political conflicts of major players were the typical mobilizing factors. Finally, he concluded that the emergence of protest culture should be perceived as a sign of democratic consolidation and the spread of a culture of participatory democracy.
Some of the movements of the late Kádár era proved to be virulent enough to live through the democratic transition and remained active ever since either in their original movement form or blended into parties. Even twenty-five years after the transition, in 2014, three of the four parties in the Hungarian parliament had deep roots in the Kádár-era opposition movements. Dániel Mikecz showed that the trajectories of the farright movement and the ecological movement, both present in the 1980s, involved partially transforming into parliamentary parties by the early 2010s (D. Mikecz, 2015).
The Hungarian green movement around the millennium found a great deal of inspiration in the international activity of the Global Justice Movement (GJM) that linked environmental problems to the functioning of the globalized economic system. While the GJM had reached Hungary too, it did not become as strong as elsewhere in the world, notably in Western-Europe and the Americas (D. Mikecz, 2015;Piotrowski, 2013Piotrowski, , 2017. Piotrowski describes the Global Justice Movement in Central and Eastern Europe to be much smaller than in other parts of the world and he explains this with the postpolitical, anti-ideological nature of the political culture in the region (Piotrowski, 2013, pp. 416-417). Furthermore, Piotrowski claims that the global justice movement has fewer currents in the region, which makes it more radical (in terms of discourse and protest repertoire) than elsewhere. He characterizes the movement as closer to subcultures and countercultures than in other regions, where participants concentrate on the feeling of 'orthodoxy,' which they supposedly brought from the 1980s anti-system activism that created the image of the pure, grassroots activist. Finally, Piotrowski mentions that GJM in the region is also in a peculiar position, as leftist slogans and arguments are much harder to articulate due to the post-socialist heritage that associates left with the earlier political systems.
Ágnes Gagyi, who gained most of her insights into the Central European functioning of the GJM from observing the Hungarian and Romanian GJM groups, also revealed some consequences of the different meaning of 'the left.' She concluded that while leftist slogans resonated very differently in the region, the local activists choose not to amend them. 'European activists were motivated to stick with the global agenda because it promised them an equal position within the movement, unlike the mainstream hierarchical discourse of post-socialist transition, which framed any local social problem as a shameful mark of Eastern inferiority' (Gagyi, 2012, p. 146). However, Eastern European activists were in a peripheral position in the global movement due to a number of reasons, including less experience and less financial means. Gagyi's analysis of the Romanian context concludes that 'unable to feed back on local conditions, alterglobalism remained the imaginary movement of a small activist elite, while the antiglobalist anger of the population got channeled by the extreme right' (Gagyi, 2012, p. 147).
Thus, by the 2010s Hungary, similarly to other countries of the region, had a wellestablished, although not particularly virulent, mostly peaceful protest culture. Contentious politics was rooted in the local experience of Kádár-era resistance, but was also connected to international social movement currents, albeit in an asymmetric way. While the mission statement uses the international movement's vocabulary, the local scene's evolution is by no means reducible to a straight-forward adaptation process. Budapest UP! represents a second generation of UPM organizers, and despite the succession of some key people, ideas, and techniques, it is rather different from the first group. The predecessor of Budapest UP! was the Fourth Republic Movement, which started their UPM activity in 2007. The collective, commonly known as 4K!, slowly transformed into an oppositional youth organization active in anti-government demonstrations and performances during the second Orbán government (2010-2014). As a final step of politicization, they became an official political party in Mikecz, 2013 and expressed sympathies towards the Party of the European Left (4K! -Negyedik Köztársaság, 2013). 4K! failed to gain seats in the 2014 elections and continued to function as an extra-parliamentary party until its dissolution in 2016. The founders of Budapest UP! were activists in 4K! before its transformation and, when 4K! became explicitly political, decided to start a separate group exclusively focused on urban playground activities. This ensured the continuation of the Pillow Fight and similar popular events. This 'succession' allowed Budapest UP! to build a purely UPM profile devoid of association with that of 4K! as a political party.

Methods
We first explored the motivations of the 'leadership' of Budapest UP! through interviews that took place before the annual Pillow Fight in 2014. 5 To identify participants who could be considered 'leaders,' we took into account the judgement of the other participants, a formal element (who the main organizer of the Pillow Fight was in terms of liaison with the police), and included our own assessment of roles within the group. This led to the identification of four leaders with whom we conducted paired interviews, allowing for the direct observation of in-group communication style (Rubin & Rubin, 2011, pp. 25-31).
We perceived that leadership positions were not institutionalized in the group. Our observation is that leadership is mostly determined by the amount of time and energy an individual spent on advancing the group's goals. This functioning creates a differentiated and transparent power structure. The organizer's group is informally led by leading organizers, consisting of the four people we interviewed (Organizer A paired with Organizer B, Organizer C paired with Organizer D). They are embedded in the more encompassing category of active organizers, a group composed of around 10 people, while the largest unit is that of the organizers, making up a group of around 25 people in total. These positions show a degree of fluidity based on (1) the events that the group is focusing on, (2) the variation over time in how active the organizers are, and (3) the in-group dynamics, including the recruitment and exit of members. The degree of fluidity is reduced as we move upwards in this informal hierarchy.
Additionally, out of the four leaders interviewed, we considered Organizer A to play the role of the overarching leader in a sense similar to Gerbaudo's notion of the 'reluctant leader' (Gerbaudo, 2012). Organizer A reflected on his particular situation in the group during our interview, expressing multiple times that he is uncomfortable with the quasimain-leader role he plays. He claims that the group spontaneously evolved into its current form; he was the only organizer taking up some key tasks, which others did not object to, but he would prefer a more horizontal functioning of the group.
We chose to perform a thematic text analysis of our interviews, coding each utterance that contained an explicit claim with one of the two focal themes: (1) organizational goaldefinition or (2) public space conception. Note that an 'utterance' is not a grammatical category, but a unit based on meaning, and can be a word, short clause, or lengthy thought connected by the same substantive point. Because our units of analysis could be one or many words, the quantification that will follow is only indicatory. This is in line with the thematic analysis tradition that uses coding, but where conclusions do not stem as directly from numeric results as in quantitative research (see Guest et al., 2011). This interpretive process ensures we are less likely to fall into the trap of downplaying the importance of rare or singular leads that Biernacki warns about when advocating against the coding of text samples and for the practice of humanistic interpretation in textual analysis 6 (Biernacki, 2014).
To compare the interview results with the knowledge and attitudes of participants, we conducted a survey during the 2014 International Pillow Fight Day in Budapest. Around 450 participants showed up, and they fought for around one and a half hours. The survey interviewers were trained in advance in how to choose their respondents, using a simple randomized pointing process employed in protest surveys to avoid biases discussed by Walgrave and Verhulst (2011). We instructed them to select every third person they saw when they looked in a given direction as a mechanical way to counteract the natural tendency to approach some people more than others. The pillow fighters were a chaotically moving crowd. We prepared a sectorial division of the space, but because the turnout was smaller than expected, our sectors provided less orientating guidance than intended. However, we did orientate the surveyors in real-time with redefined areas according to the used space. We asked the interviewers about their experience during and after the data collection as a quality control effort, including how they managed to follow the procedure and how they perceived of their peers' practices. Based on the feedback and our own impressions on the spot, we identified two interviewers who systematically disregarded the sampling instructions and cherry-picked respondents. We decided to exclude these respondents (23 cases altogether) from the dataset to preserve the randomness of our sample, resulting in a final sample of 111 valid cases.
The questionnaire contained three main sections, the first two directly paralleling our interview themes: (1) whether the respondent had taken part in similar events before, how they would describe these events, and what they considered the events' goals to be; (2) their opinion about public space; and (3) demographic information (see Supplemental Online Material: Urban Playground in a Social Movement Context -Questionnaire). Conducting each survey took an average of seven minutes, and the nonresponse rate at 3% was outstandingly low compared to usual in-person surveys.

Findings
Interview results: leading organizers motivated by and engaged in conflictual social relations In analyzing the interview results, we focused on two questions: how the organizers viewed their own organization and how they viewed public space. We coded the claims relating to the overall self-assessmentsuch as self-definitions and selfinterpretationsof the group into the two categories of non-conflictual and conflictual definitions. Following the already mentioned social movement definition of Della Porta and Diani (Della Porta & Diani, 2006, p. 23), we consider frames to be conflictual when they identify adversaries or imply a redistribution of power or alteration in social structure. To better understand how conflict appears in the selfnarrative, we split the conflictual category further, identifying mildly and pronouncedly conflict oriented self-assessments, the former representing cases where the conflictual element is present more as a hint than an explicitly formulated conflictual stance. Doing so, we find that non-conflictual and conflictual (mildly or pronouncedly) utterances occurred at similar rates, where we coded 14 instances of the first and 13 of the second category.
Having fun was the most frequently mentioned (7 instances) across all categories: Interviewer: What does Budapest UP! do in your view?
Organizer C: If we want to put it very simply we could say . . .
Organizer D: We want to feel good . . .
Organizer C: Yes.
Organizer D: And for all the other people who come [to the events] to feel good too.
Fun-focused self-definitions are often given in the discussion without any definitionfacilitative push from the interviewer, and they are at times expressed in such straightforward, essentializing formulas as in the final sentence here: Interviewer: How do people react coming across your events in action?
Organizer C: With a big smile! Organizer D: Yes. Surprisingly, it is exceedingly rare for us to get a negative comment.
Organizer C: 'Oh, that is so great, I want to go too!' is what my friends usually say.
Organizer D: If I must pick a negative comment, once a person rejected their invitation to the Facebook event writing that 'If you have so many pillows, why don't you give it to those who don't have one?'. [. . .] I didn't want to write back 'Because we are not a charitable association [for the poor], we entertain people.' Other non-conflictual self-definitions emphasized the goal of continuing popular events that were already part of the urban culture of Budapest (4 instances), being a child or being young again for the duration of the events (2 instances), or following a lifestyle (1 instance). The middle group in terms of frequencies is composed of claims coded as 'mildly conflictual': communications of a less-directly conflictual nature (10 instances).
Interviewer: Are your events flash mobs?
Organizer D: Flash mobs in the contemporary sense . . . no, the flash mob notion has simply ceased to exist. Nowadays only marketing-oriented lame things (are organized) on which they write with letters this big [he shows it] that they are flash mobs, but they have nothing to do with it. The whole thing is ridiculous. We simply organize community events.
Here flash mobs are denounced for their commercialization and contrasted with Budapest UP!'s community-oriented activity, but it is not clear whether countering commercial entertainment and fighting against the individualized nature of urban lifestyle are positioned as the goals of the group, or whether the organizers only pronounce a value judgment between two forms of entertainment.
Claims expressing a clearly conflictual orientation were the less recurrent (3 instances). The exact 'reoccupying public space' formulation only appeared once directly and once as a reference, but substantially similar claims were formed, and its components were spread across the interviews, either using a non-conflictual tone or a mildly conflictual one.

Interviewer: Does Budapest UP! have a political message?
Organizer D: This circles back to what we talked about at the beginningthe political message is the one about public space. Also, I don't know, I think we could categorize into that the one about [inclusion and] exclusion. There are people from very diverse backgrounds among the organizers and among the participants. [. . .] When there, they are all just people. Everybody becomes just like each other, and everybody feels free to pillow fight with everyone else. Meanwhile, the following example of the unequivocal occurrence of the main conflictual claim also shows the ever-present tension between having a mission and staying away from party politics: Interviewer: And how do you relate to that?
Organizer D: You mean to politics? Budapest UP! is really just about having fun, being stupid, etc., so it doesn't have a political message. Maybe only what the Urban Playground Movement stands for, so that we should reoccupy the public spaces, but apart from that, we wouldn't like to assert any kind of political influence on people.
Our other interest was in how the interviewees conceived of public space. They mostly emphasized social aspects (8 instances), and we only identified one mention of architecture. When asked about the meaning or use of public space, they articulated similar problems to those mentioned in the UPM manifesto.
Interviewer: What does public space mean for you? What is public space 'good for' for you?
Organizer D: A place to play pillow fight at. (laughs) Organizer C: Indeed! For me, it should be used as a public and community space. [. . .] These days they create like little corridors instead, so that you can go through them.
Organizer D: Yes, they transform all of them into shop windows. So they are not there to be used, but just so that people can say: 'This is so pretty!' What can I do with it if it is so pretty, but I can't sit on the grass? These sentences reverberate the passivity/individualization-type of public space criticism in line with the UPM manifesto and observers' conflictual accounts of the UPM. On a subtler level, we can recognize a populist-toned argument considering that in this discourse 'they,' the elites, are despised for designing and maintaining spaces with the goal of satisfying their own upper-class aesthetic preferences and the needs of businesses, instead of striving to create a city usable for the people.
The articulation of conflictual frames is positively related to the levels in the internal hierarchy. Organizer A, the leader, is the one most motivated by conflictual thoughts. He would even move the group's activity to explicitly political grounds, but he respects the majority stance. He overtly links these political ambitions to his previous socialization in 4K!. While he is alone in having directly political aspirations for the group, he has allies in embracing the mainstream UPM ideology; some other leading organizers also actively bring in conflictual claims into discussion, and those who do not still show some signs of agreement when their interviewee pair brings up such claims. We did not systematically research outer layers of the organizer group, but inferring from the interviews and the fact that leading organizers are open about their conflictual motivations to the larger group of organizers, we assert that there is a passive acceptance of the existence and role of conflictual frames on the part of all organizers. Social conflict-motivated people represent a small minority in the whole group, but as they are the most active, the total time and energy spent organizing Budapest UP! events is 'conflict-charged time' and 'conflict-charged energy.' Summarizing our interview findings, we spotted the conflictual patterns we were expecting to find based on the UPM manifesto and on the political interpretations in the literature. Numerically, the explicitly conflict-committed self-definition statements were few, and they were only actively articulated by two of the four interviewees (including the 'reluctant leader' of Budapest UP!), but they are quietly endorsed by all main organizers. Interviewees also tended to reproduce the conflictual discourse of the international UPM when asked questions related to public space. Thus, while the main organizers' activity is partly based on a large set of non-conflictual motivational factors, social conflicts seem to play a key role in their participation.

Factual knowledge
In gauging the degree to which the participants are motivated by social conflict compared to the larger population, the ideal design would be a direct comparison to them. However, to our knowledge, there is no representative data available on Hungarians' normative approach to public space. Consequently, we had to take a different route. Hungarian media reporting on the Pillow Fight tends to use an overwhelmingly entertainment-focused frame (e.g., Blikk.hu, 2015;Tények.hu videó, 2015), therefore our assumption is that those visiting the event for the first time are most likely to engage in the activity without social conflicts in mind. The survey therefore measures how much becoming a regular participant alters people's UPM-and public space-related attitudes and knowledge as compared to first-time participants. Hence, we explored whether the leaders' social conflict-oriented frames 'trickle down' to participants through longer exposure to the events.
The sample of 111 participants divides roughly into one third who previously took part in a UPM event (35 people) and two thirds who had not previously taken part in an event (76 people). The distributions of key demographic variables (gender, age, highest educational attainment, occupation) are broadly similar between the two groups, as seen in Table  1 (see in Appendix). When applying a logistic regression to estimate whether any of the demographic characteristics have a predictive power in explaining whether an individual is a returning participant or a newcomer, we find that only education does so, where we find that those with an undergraduate degree are less likely to be a returning participant than those with a high school degree. However, the trend might be explained by the overall sample being overwhelmingly young (see results in the Appendix in Table 2 for the logistic regression, and Table 3 for descriptive statistics on the opinion questions of the survey). The fact that new and returning participants are largely similar in their demographic characteristics reassures us that returning participants are not strongly self-selected based on these demographic characteristics (that is, for example, teenagers are not more likely to decide to come back than other generations).
Our survey was divided into two blocks of substantive questions; the first block asked about the event and organizers, and the second asked about perceptions regarding certain behaviors and social connections. As can be seen in Figure 1, when asked to name the group who organized the local Pillow Fight, most returning attendees gave 'Budapest UP!' as an answer, while two mentioned the title of the international UPM event instead (International Pillow Fight Day), and one named 4K!, the former organizer. Only 17% of the experienced participants didn't give any answer, and nobody gave other answers than the three response options mentioned above. Meanwhile, among newcomers, only 35% named the event's creator correctly, 51% didn't give an answer, and the rest of the respondents gave various incorrect answers. Thus, the factual knowledge gap between the two groups is wide. Using Fisher's exact test, we find that the proportion of correct responses given is significantly different between returning participants and newcomers at the 1% level.
When asked how they would describe the pillow fight and 'similar events,' only four respondents knew that the activity belonged to the Urban Playground Movement; those respondents were all returning participants (constituting 11.4% of those returning), and the difference in knowledge between the two groups again was significant at the 1% level. Similarly, when people were also asked if they had 'heard of the Urban Playground Movement,' more returning participants said that they had.
The above results highlight that organizers are a source of factual information to participants. Engaging with the organizers through participation transmits knowledge to participants, and factual information regarding the event does trickle down. The finding also suggests that a newcomer versus returning participant comparison appears to be valid, event exposure does influence how much knowledge participants have about the event's organizers.

Attitudes and values
To test whether participants conceived of the Pillow Fight with the same social conflictoriented frame as the organizers, we asked them about values related to the use of public spaces. Survey interviewers asked participants an open question about what they thought the intended uses of public spaces were. As seen in Figure 2, these responses were coded into the pre-existing categories provided for the interviewers as follows: 'pass through,' 'be there,' 'meet others,' 'enjoy oneself,' 'spend free time,' and 'spend there as little time as possible.' When testing the differences, we find that in none of these categories was there a difference in the number of mentions. We conclude that newcomers and returning participants are homogenous in terms of the attitudes towards public space, suggesting that the two groups thought of the purpose of public space similarly.
Returning to a survey question discussed in the factual knowledge section, we noted that participants were asked, 'What broader name could you use to describe the pillow fight and events similar to this?'. While, as noted, we found only four respondentsall returning participantsnamed the Urban Playground Movement, the lack of factual knowledge prompted the vast majority of both new and returning participants to provide creative, descriptive answers to how they perceived the event. These answers in turn prove to be highly insightful in understanding the two groups' ways of thinking about similar events and about using public space. To do so, we categorized these descriptions of the Pillow Fight and similar events based on whether they had a social conflictoriented tone or a fun-oriented one. From a total of 105 answers, 34 responses were pronouncedly fun oriented (with descriptions such as 'party,' 'crazy,' 'game,' and 'physical activity'), while only 2 were conflictual in any sense ('independent' and 'guerilla'). The rest of the responses often emphasized the community aspect of the events, while 12 responses contained the expression 'flash mob.' Crucially, social movement frames were almost completely absent across all participants, and new and returning respondents associated the Pillow Fight with group entrainment (as well as with it being a flash mob) equally.
In conclusion, there are two clear patterns emerging from the survey data. First, those attending for the first time and those with previous experience exhibit a significant difference in factual knowledge, but not in their personal value set; participating in multiple events does not change opinions. Second, non-conflictual, fun-oriented frames motivate those taking part in UPM events in Budapest; participants primarily associate playfulness and entertainment with the UPM. Therefore, we conclude that the leaders' values do not trickle down, the vast majority of the participants are not conflictmotivated, and the participants do not recognize social conflict-oriented frames to be present in the group.
Discussion: continued social movement frames in a collective action that is not a social movement As we juxtapose the thematic analysis of the leaders' interviews with the results of the participants' survey, we are faced with a contrasting picture. The main organizers in Budapest's UPM scene are partly motivated by social conflict-oriented ideas and are organized around the master-frame of 'reoccupying public space.' Their involvement relies on a sense of social mission in line with the core beliefs expressed in the movement manifesto. The leader level shows the pattern of an identity politics-oriented social movement with a clearly spelled out conflict over the ways to use public space, a reference to those seen as responsible for the grievance (urban policy makers influenced by corporate interest), but without pointing at policies or aiming to drive specific political actors out of power.
The main organizers have an activist past and present, and the UPM is one of the activist projects in their lives. While we did not systematically retrace the route of the ideological elements, most probably it is through the involvement in 4K! that the leaders internalized the core beliefs, as is spontaneously pronounced multiple times by Organizer A during the interview. These thoughts are deeply-rooted, not simply the products of a one-way learning process from the international UPM scene.
We are facing a particularly contrasting result of diffusion: social movement continuity on the highest level of the Budapest UPM scene, but discontinuity on levels below. While we did not analyze the further social movement activities of the main organizers, our results show that the UPM in Budapest acts as a 'container' of social movement frames, skills, and networks, so it might assure movement continuity between cycles of protest, similarly to social movements.
The fact that social movement frames function at the organizer levels, but not at the participant levels, might be due to various factors. To the best of our knowledge, as ours is the first empirical study of this nature, we cannot judge whether the Budapest case is special in this respect. Still, we might hypothesize that the lack of movement frame diffusion in the Budapest scene can be related to the peculiarity of the post-communist region and the Hungarian context, where movements and protests are relatively rare. Thus, it is possible that the Urban Playground Movement's diffusion from New York and Toronto to Budapest was hindered in a similar way as in the case of the spread of Global Justice Movement from the US and Western-Europe to Central Europe. We might witness a similar process of a movement appearing as a subculture in the Central European context. In this case, what we observe might be the outcome of a successful diffusion of the repertoire of action (pillow fight, Capture the Flag, etc.), but a failed diffusion of frames except from the very thin layer of main organizers.
On the other hand, we should not exclude the possibility that what we see in the Budapest scene is the standard pattern of this collective action. Our study suggests that besides the well-known central role of social fora and protest camps in the diffusion of GJM and recent movements of the crisis, respectively, UPM might have also helped the spreading through the continued frames and practices of urban reoccupation, but without a transmission of frames to the participant level. It would be a fruitful avenue of future research to explore whether UPM did play an intermediary role between GJM and the post-2008 mobilizations on the international level similarly to its 'container' function in Hungary.

Conclusion
Our contribution is twofold. First, the motivations of participation in Urban Playground Movement events, the most popular type of contemporary flash mobs, have not been addressed empirically before, despite the abundance of theoretical texts. We explored the functioning of social conflict-oriented frames in the UPM mobilization in Budapest to answer the question to what degree UPM is a social movement type collective action. Our interviews show that the main local organizers use strongly conflictual statements, with the 'reclaiming of public space' functioning as a master-frame. Thus, we can spot the 'conflictual relations with clearly identified opponents' elements of Della Porta and Diani's social movement definition, which was the only component of the definition we questioned the existence of. The conflict that the organizers articulate is that policy makers (as opponents) take the city from the people and transform spaces of possible interaction into 'corridors' with only aesthetic value, designed and regulated according to the needs of businesses (also as opponents). In turn, organizers fight urban isolation by prefiguring a more social city.
On the one hand, we conclude that the functioning of the UPM in Budapest heavily relies on social movement continuity. The social movement-experienced leaders are the motors of the initiative, and their motivation originates in the international UPM's conflictual frames rooted in both earlier and contemporary social movements, signaling that diffusion mechanisms took place successfully. The relationship with social movements is also visible in the biographies. The main organizers were activists in the leftwing 4K! and belong to a leftist sub-culture. The ideas presented by the leading organizers bear close resemblance to the capitalism-critique of both the Global Justice Movement's Reclaim the Streets collective and of the recent movements of the crisis. The organizers espouse an urban space-focused repertoire of action that was characteristic of both Reclaim the Streets and of the recent city square movements. Furthermore, Reclaim the Streets' emphasis on playfulness and the city square movements' focus on inclusivity and participation are important features of UPM events.
Conversely, the participant level shows movement discontinuity; our survey exhibits a clear pattern of non-ideological, fun-seeking motivation. Thus, we conclude that the UPM in Budapest is not a social movement. There are social movements where leading activists are more conflict-oriented or use different frames than followers, but in the case of Budapest UP! participants completely lack conflictual motivation. We witness strong social movement continuity that enables a collective action that is not a social movement to function by motivating its leaders through 'continued' social movement frames.
Our second contribution is to the studies of the diffusion processes of social movements, notably those of the recent past, including the Global Justice Movement and the post-2008 movements of the crisis. Our endeavor follows in the footsteps of Mattoni and Della Porta (2014), as it tackles diffusion cross-temporarily and focuses on an entity that is not a social movement. Our results show that the Urban Playground Movement, a widely spread collective action, is not a social movement in itself, but functions as a 'container' of social movement frames, networks, and skills on its main-organizer level.
While we only analyzed the 'incoming' frames and biographies, UPM's outwardpointing diffusion processes could be likewise salient, so they might ensure movementcontinuity between cycles of protest, especially in the movement milieus of the Global Justice Movement and the post-2008 movements of the crisis. This outward-going dimension should be addressed in further research. Furthermore, we could not address the specificities of the Budapest scene due to the lack of empirical studies on UPM motivations in other contexts. However, we formulated two hypotheses for further research. One possibility is that it was the Hungarianand in general, Central Europeancontext that hindered the diffusion of conflictual frames from New York and Toronto to the participants in Budapest (except for main organizers) due to the post-communist heritage of low movement activity and a mistrust of politics. This is the phenomenon that is known to have hindered the development of a strong Global Justice Movement scene in the country.
Alternatively, what we see in the Hungarian case could apply to the core, Western UPM scenes too, thus we would not observe a social movement in the US and Western-European cases either, but a fun-centered collective action. In this case, UPM might have played a role in preserving movement frames, skills and energies in an entity that is not a social movement, thus contributing on the long term to social movement continuity in-between the GJM and the post-2008 movements of the crisis. In any case, the international Urban Playgound Movement is most probably among the largest collective action networks relevant for the field of social movement studies either as a link in the diffusion of social movements (as is the case of Hungary), or potentially as a social movement itself (possibly in other cases to be explored).

Notes
1. Characteristic events of this type are 'freezing' in public spaces, when flash mobbers suddenly stop their motion in the middle of moving crowds, and silent raves, when they dance publicly to music only they hear through their headphones. Flash mobbers gather in one place and perform these acts briefly, typically for less than 10 minutes, then they disperse as if nothing happened. 2. Political flash mobs are used instrumentally for political purposes, while our study explores the UPM's inherent, potentially political qualities. 3. Description in its original is provided in Hungarian. Translation by the author. 4. The full set of events by Budapest UP! that occurred at least once in 2014 are as follows: Pillow Fight Day, Capture the Flag, Last Team Standing, MP3 Experiment. Other events by the same organizers in preceding or following years included Water Fight, Snow Fight, and Delighted Downtown. 5. Mihály Gyimesi, the author of this article, was among the founding members of this local UPM organizationhowever not of the local UPM scenebut was not anymore part of it when the research took place. It was a professionally trained interviewer, having no previous connection to the organizers in any form, who interviewed them. 6. Although he is also skeptical about mixed methods.   Note: Logistic regression with robust standard errors. Baseline categories are: Age: 18 to 20; Education: High school degree; Occupation: Employed full-time. Note: Four responses are mechanically excluded from the regression due to these being singular observations in one of the demographic categories (for example, we only observe one person reporting to be a Seasonal worker, and also one reporting to be on Parental leave). Standard errors in parentheses + p < 0.10, * p < 0.05, ** p < 0.01 Note: Values expressed in averages. Differences are significant at the 5% level for average number of events attended in past two years. Differences are not significant about opinions about others' behaviour in public, about the various purposes of public space, and about statements on public space.