Cultural clusters and social interaction dynamics

Analysing the territorial dynamics of culture, particularly its tendency to form clusters, has become a study area that draws the attention of various social science disciplines. It has grown steadily in importance over the last 20 years alongside the increasing interest in creative industries and cultural institutions as factors in socio-economic development and urban regeneration. Most of today’s literature on the subject takes cultural clusters as a single type and focuses on aspects linked to their urban planning or economic impact. However, there is a growing awareness of the importance of the social dimension of cultural clusters. This paper aims to differentiate between various cultural clusters in the city of Barcelona by constructing models or types of cluster, taking into account the predominant interaction dynamics and the type of social ties generated between the cultural agents participating in these groupings. Following these criteria we distinguish three types of cultural cluster according to whether a bureaucratic, associative or community dynamic predominates. These social dynamics enable us to understand the success or failure of a cluster policy.


Introduction
Over the last few decades a great many European cities have seen heavy investment in cultural facilities and policies designed to encourage the development of companies and institutions characterized by the generation of cultural goods and services. These strategies fall within two clear directional groups: (1) those aimed at developing economic sectors linked to the production of cultural goods by encouraging activities related to the so-called cultural or creative industries (film, radio, television), design (web, textile, graphic, industrial), fashion, advertising, photography and architecture (Scott, 2000); and (2) those aimed at generating cultural services to attract tourism and business by recovering the heritage that already exists in towns and cities (urban regeneration), creating cultural institutions and facilities (such as museums and cultural centres), organizing events, etc. (Evans, 2001;García, 2004a;Landry and Bianchini, 1995). Both strategies have a specific urban setting characterized by a concentration of companies, institutions and agents in what are known as cultural clusters or cultural quarters (Cooke and Lazzeretti, 2008;Mommaas, 2004;Montgomery, 2003). In our study -which deals with how these concentrations of specialized cultural actors work -we chose to use the term cultural cluster because it is the one that refers more clearly to the actors' level. This term, already present in the academic literature -as mentioned above -as well as increasingly in the administrative discourse (e.g. in the classifications of the European Cluster Observatory) has been subject so far to a limited theoretical elaboration. In this regard, our piece of research tries to contribute to the development of a notion of cultural cluster that goes beyond a merely descriptive approach and also can embody a greater analytical potential as compared with the current one.
Development, regeneration and cluster strategies are found in the framework of a new urban cultural scenario. This scenario has emerged as a result of the long process of change that has taken place since the 1970s in the economic and political (public and cultural) spheres and in the ways in which work in the art world is organized and carried out. The main changes include the following: (1) the reorientation of cultural policy towards a new paradigm linked to the generation of cultural value and urban regeneration (Landry and Bianchini, 1995); (2) the configuration of a new socio-economic scenario for the economic development of cities in which culture occupies a central place (Scott, 2000); and (3) the emergence of new forms of organization in the art world that have an effect on the types and forms of organization in the world of work (Boltanski and Chiapello, 2002).
As a result of these changes, the first decade of this century saw the appearance of a number of papers in academic and professional circles that aimed to explain or create specific tools in order to understand or encourage this type of transformation through the creation and use of ambiguous notions such as creative cities (Landry, 2000), creative industries (DCMS, 1998) and the creative class (Florida, 2002), which provided little explanation. Interest in this area has continued to increase in line with the growing interest in cultural institutions and creative industries as a factor for socio-economic development, urban regeneration and social inclusion.
The aim of this paper is to analyse cultural clusters in Barcelona from a sociological perspective. Unlike the traditional literature on the subjectwhich tends to take a one-dimensional approach to clusters, looking at aspects linked only to either policies and cultural management or the economic exchanges and transactions between companies and workers in the sector (omitting any other type of dimension) -here we consider the existence of three different types of cultural cluster in Barcelona depending on the predominant means of social interaction: 1. the cultural cluster as a bureaucratic organization: these are clusters of cultural institutions whose interaction dynamics appear to be conditioned by the system of cultural policy, the system of relationships between cultural administrations, the patterns of interaction between political leaders and the directors of the cultural institutions, and the cultural and legal framework governing public-private sector relations; 2. the cultural cluster as a market-oriented association: these are clusters of cultural production and/or consumption in which a shared professional culture and/or common interests based on fairly flexible projects predominate; 3. the cultural cluster as a community dynamic: these are creative clusters in which community ties (based on a common sense of belonging) and non-formalized creative relationships predominate.
This new perspective enables us not only to better understand the social interactions within each cluster in Barcelona, but also to construct a typology to help clarify and differentiate between them. In addition, this analytical clarification enables us to understand the success or failure of different cultural clustering strategies. The city of Barcelona has been chosen because it is a particularly rich, advanced case due to its support of culture as an urban development strategy.
The methodological approach used in our research was a qualitative one, the data collection techniques being basically two: on the one hand, semi-structured and in-depth interviews; and on the other hand, an analysis of public documents related to government projects (such as the case of the Plan 22@ or the Ciutat del Teatre) or to cultural institutions (such as the Instituto del Teatre or the Teatre Lliure). The semi-structured interviews were undertaken with social actors belonging to the artistic sector (artists, creation and exhibition centres, gallery owners), the neighbourhood sector (neighbours and neighbourhood associations) and the administrative sector (staff of the Barcelona's City Council related to the creation and development of the Plan 22@ and the Ciutat del Teatre). About 50 interviews on the transformations undergone by the Raval and the Poblenou district were carried out in the framework of the doctoral research by two of the authors of this article The analysis developed by field information obtained through the project Neighborhoods, museums and art: public art, artists, institutions funded by the Spanish Ministry of Economy and Competitiveness (-HAR2012-38899-C02-01).
The paper is divided into two parts. The first part runs through the debate on cultural clusters. The second part analyses the different types of cluster in the city of Barcelona according to their predominant social dynamic.

Analysis of cultural clusters: from buildings and figures to social interactions
The concept of cluster emerged as a neoliberal alternative to the type of centralist planning policy carried out by the nation states, reviving the old model for developing the Marshallian industrial district of the early 20th century (Porter, 2000). This concept connects to an extensive literature on the phenomenon of industrial clustering, its origins dating back to the Marshallian notion of 'industrial district' (Marshall, 1920). This approach, attempting to understand industrial clustering on account of the competitive edge it allegedly provides, gained credibility and has become popular since the 1980s due to the new view on the post-Fordist industrial development, based on flexible specialization and vertical disintegration -concepts introduced by Piore and Sabel (1984). Within this context, the concept of cluster -as coined by Porter -highlights the importance of the cross-relations among the companies within the cluster, their common elements and complementarity being the keystone of their joint competitiveness (Porter, 2000: 15). Porter's reasoning has been highly influential, especially throughout political spheres. It has also received significant attention within the academic environment, while also being subject to widespread criticism. It has been criticized, for example, for its excessive vagueness, both in regards to geographical limits and industrial outlines (Martin and Sunley, 2003); it has been noted, too, that not all the competitive advantages of clustering are located within the cluster boundaries: the urban and regional environment, as well as the national and international connections are usually very relevant (Simmie, 2004;Turok, 2004).
Notwithstanding its ideological connotations as well as the weakness and analytical problems entailed by its original formulation, the cluster concept seems particularly suitable for conveying the spatial organization of cultural production. On the one hand, this is characterized almost without exception by the territorial concentration in distinct areas of large cities, as many researchers have pointed out (Cooke and Lazzeretti, 2008;Menger, 1993;Mommaas, 2004;Musterd and Murie, 2010;Pratt, 2008b;Scott, 2010). On the other hand, the agents, companies and institutions involved keep intensive (either competitive or cooperative) relations with each other; for them, face-to-face exchanges become essential (Currid, 2007). Cultural clusters have, in this respect, a specific empirical consistency. However, also in this case the original cluster concept has proved to be insufficient, because the broader urban framework is heavily involved in its dynamics (Scott, 2006(Scott, , 2010 and often the supra-local and international relationships have been demonstrated to be decisive (Bathelt et al., 2004;Coe, 2001;Grabher, 2002;Pratt, 2008b). Nevertheless, for the purposes of our research, these imbalances turn out to be relatively marginal. Here we will take the cultural cluster approach, in that sense, not only because it fairly consistently represents the socio-spatial organization of cultural production, but also because the cluster scale, bringing together the highest diversity of face-to-face relations, also allows for a better differentiation of the social dynamics of creativity -and that is precisely our intention.
As regards the analysis of cultural clusters, there is a wealth of academic studies, developed from different approaches: economic, urban, political or social (Karlsson, 2010;Lorenzen and Frederiksen, 2008). Nevertheless, most of these studies provide for the existence of only one type of social cluster, while acknowledging its wide diversity in terms of spatial configuration, levels of interaction or governance dynamics (Cinti, 2008). Some authors have developed classifications on the basis of its location within the urban pattern (Frost-Kumpf, 1998), the government's level of intervention (Wen, 2012), its organizational context (Redaelli, 2008) or its evolution over time, as is the case of the cultural districts examined by Zukin and Braslow (2011). There are also studies taking into account differences between types of clusters according to the economic output they produce (Santagata, 2002) or the economic and spatial configuration, as well as the governance structure they are endowed with (Markusen, 1996). However, despite the value of these contributions, they do not include any typology allowing one to differentiate cultural clusters according to the type of social link between actors and organizations and its relationship with the processes of urban creativity.
One of the perspectives from which cultural clusters are analysed -in this case linked to creative cities -originates not only from consultants connected with the public administration, but also from the academic sector (Landry and Bianchini, 1995). Its aims tend to involve regulations and its authors usually come from urban planning and public and/ or cultural policies (Evans, 2001). Most of these papers focus on the positive or negative impact that this type of strategy has on urban transformation. In some cases they also propose new tool kits or perspectives to take into account when applying certain urban regeneration strategies. This viewpoint mainly analyses the administrative dynamics of cultural clusters based on the concentration of cultural industries (e.g. museums and cultural centres) (Mommaas, 2004). A second line of research springs from economic geography, economics and business analysis. Its main interest is strategies for local economic development within the framework of the socalled cultural or creative economy (Scott, 2000). This type of research analyses the economic dynamics of cultural clusters in which there is a predominance of companies involved in the production and/ or consumption of cultural goods (Cooke and Lazzeretti, 2008). Finally, there is a third perspective that centres on the analysis of culture professions and professionals in urban contexts and the ways of organizing work in the cultural world and how the field of art is structured. This type of analysis is carried out by social scientists, either sociologists or researchers connected to cultural and/or art studies. Some of the recent research into this perspective, using dubious concepts such as the creative class (Florida, 2002), has given this type of analysis a certain public importance. Two clear lines of research can be distinguished in this area: those that focus on traditional artistic and cultural professions (Markusen, 2006) and those involved in analysing the so-called creative professions (Florida, 2002). This literature tends to analyse certain neighbourhoods that are characterized by a concentration of artists and creators.
These three approaches have a reductionist viewpoint that links a cluster's cultural dynamics to particular dimensions. Hence, the literature on creative cities tends to hypostasize cultural processes to a particular urban infrastructure. In this sense culture is reduced to material resources (accessibility and technical, urban and cultural infrastructures) and symbolic resources (design of the urban landscape, architectural icons and intensity of urban life), with the social processes from which they originate being removed. The literature on the creative economy, meanwhile, tends to limit its analysis to the processes of economic innovation brought about by the exchange generated between workers and transactions between companies and industries connected to the so-called creative economy (e.g. cultural, creative and knowledge industries). Finally, the literature on the creative class tends to ignore the social relationships upon which the cultural processes in cities are based. Hence, this type of analysis focuses on the idea of individual talent and its classification within a wider concept, such as the creative class (a new class made up of people who have this talent). 1 The social dimension of cultural clusters is taken into account by these disciplines and also by sociology. However, this is a minority dimension and, when dealt with, it is used with somewhat inefficient interpretative frameworks. In this paper we maintain that the tools of general sociology and the sociology of culture and the arts, so far little used for this type of analysis, have great potential for understanding cultural clusters.
The social dimension of cultural clusters has been dealt with from various approaches. From the point of view of Economic Geography or Economics, social relationships and face-to-face exchanges are confined to transactions between companies and workers as maximizing economic agents. The learning and knowledge exchange processes assumed in that context are usually a black box (Glaeser, 1999). On the other hand, when one tries to unravel those processes -departing from the neoclassical parameters -utilizing concepts from more qualitative-oriented disciplines (Scott, 2000(Scott, , 2010Storper and Venables, 2003), an exclusive focusing on the actors' economic interest -by discarding other dimensions of the social relationships crucial to the maintenance and prosperity of cultural clustersdeprive these concepts of much of their heuristic potential. Furthermore, from the urban and cultural planning approach there have also been warnings about the importance of paying attention to the social dimension when there comes the time to conceive and develop urban regeneration polices through culture. Still, the analysis on those relationships is not very effective in terms of conceptual clarification (Comunian, 2011). Finally, there is a whole range of work related to cultural studies and social sciences that deals with the creative potentialities of the faceto-face social interactions taking place in cultural districts.
One of the dimensions identified in the framework of these exchanges is the informal meetings between creators in the so-called 'Third Space' (Lloyd, 2010). These spaces are where information is exchanged and collaborations and projects can take shape (Currid, 2009), which some authors term buzz (Storper and Venables, 2003). Other authors who try to go more deeply into the subject settle on the idea of an 'art scene' (Molotch and Treskon, 2009) in which the players collaborate to bring value to cultural products and attract the general public, who may later become impulse buyers or regular customers. These papers understand social dynamics as interactions between equals with shared interests and common cultural codes that enable productive exchange to take place. However, the social interactions that take place in cultural clusters are structured and occur in defined institutional contexts. For this reason it is useful to adopt the Sociology approach -and particularly the Sociology of Culture and Art approach -going beyond the casual use of some of its concepts or models in order to understand the dynamics of cultural clusters.
However, it is very seldom that this approach appears in debates on the subject. One reason for this is that sociological models on cultural dynamics in urban contexts are either constructed within sectoral disciplinary parameters of an implicitly modern character (Bourdieu, 2002b) or because the research most sensitive to contemporary cultural transformations generates macro-social explanations that lose sight of the social mechanisms that define the cultural processes in each situation (cfr. Lash, 1990).
Some trends within the Sociology of Art and Culture have examined and defined the field or realm of cultural production as social spaces invested with relative self-sufficiency and working as mediators between the creators and the broader social reality. Two currents can be discerned within these studies, according to whether this field is viewed from a valorative approach (as a sphere of values with its own beliefs, hierarchies and conflicts) as is the case in Bourdieu's artistic field model (2002a), or technical (as a space defined by a cooperative chain between different activities put together on the basis of particular conventions) as is the case of Howard Becker's art world notion (2008). Nevertheless, sociological approaches tend not to pay enough attention to the spatial determinants on the processes of cultural creativity.
A way to solve this issue is to make use of some significant conceptual tools from Sociology that will allow us to identify different types of social interactions in order to relate these with the territory. For this purpose, the Goffmanian notion of frame is pivotal (Goffman, 1974). Goffman uses this notion of frame with the aim of understanding certain aspects of a strip of activity from a symbolic (what the actors perceive of a situation), organizational (what premises or rules conform those) and limiting or statusdependent (the place and type of relationship each particular activity establishes with its environment) point of view.
The Goffman framework (Goffman, 1974) is highly suitable for understanding and distinguishing the type of social interactions that exist in different cultural clusters of Barcelona. Social frameworks delimit the social forms or dynamics in which the intentions and motivations of specific social actors take shape. The urban, economic and organizational dimensions that predominate in analyses of creative cities, the creative economy and the creative class, despite their reductionism, make it possible to distinguish three dynamics that frame the main social interactions of cultural clusters: bureaucratic, utilitarian and community. These three dimensions make it easier to understand the predominant social dynamics in cultural clusters of Barcelona. However, this paper argues that, in order for them to acquire greater consistency and clarity, they need to be included within a more general framework connected to the field of culture.
There has been a weakening of the opposition between the commercial sector and the autonomous culture production sector and a generation of creative dynamics between sectors and different artistic areas (Rodríguez Morató, 2007. However, the creative dynamics and the interactions that sustain them still originate in cultural sectors (Rodríguez Morató, 2007 and are therefore based on the shared codes typical of the creators in this field (Becker, 2008). Firstly, it can be said that in the field of art the social actors, despite sharing the same sense of play, do not start from the same position and may therefore have conflicting interests (Bourdieu, 2002a). For this reason the structure of the field of art should be seen as conditioning these interactions. Secondly, as regards the idea of possible interactions within cultural clusters in neutral or informal spaces, it should not be forgotten that interactions between individuals are conditioned by the role they represent within a general organization of a hierarchical nature. Finally, this perspective does not exclude the possibility of conflict between either individuals or groups due to different interests or projects in the urban configuration of the cluster or on the cultural production stage.
Hence the setting of the bureaucratic, utilitarian and community dimensions within a broader reference framework, connected to the rules governing the field of culture, enables us not only to distinguish the predominant social dynamics in Barcelona city with a certain degree of consistency, but at the same to differentiate between cultural clusters and explain the reasons for the success or failure of different clustering policies. It is therefore possible to distinguish between social interactions driven by the bureaucratic logic of institutions involved in the provision of cultural services, social interactions driven by the associative-utilitarian logic of companies involved in the production of cultural goods, and social interactions driven by artistic community logic. The first kind of tie predominates in clusters characterized by a concentration of cultural institutions; the second in clusters where companies involved in the production and/or consumption of cultural goods predominate; and finally the third in neighbourhoods characterized by a concentration of artists and creators.
In the city of Barcelona today there are six cultural clusters, four of which have been selected for analysis ( Figure 1) 2 because they are representative of the typology described earlier: (a) clusters based on bureaucratic organization dynamics or clusters of cultural institutions (Montjuïc); (b) clusters based on association dynamics or cultural industry clusters (22@ and Consell de Cent); and (c) clusters based on community dynamics or cultural neighbourhoods (Raval).
The typology introduced in this study is not intended to be a closed categorization but a proposal of ideal types (cf. Weber, 1978), not exactly reflected in the empirical reality but working as models that the examined cases get close to, as a way of distinguishing and classifying cultural clusters according to the predominant interactions or the social dynamics in every case. In addition, there are also urban transformations -either planned or not -that modify the predominant type of interaction or the social dynamics within a cluster, as well as samples sharing characteristics of the different types. This situation would allow us to speak of hybrid forms of cultural clusters. Notwithstanding, this article is not aimed at examining these kind of processes and the strain caused by the coexistence of different social dynamics. On the contrary, it is focused on building ideal types helping to understand -in future research -the predominant social dynamics delimited by certain urban spaces.
The methodology used to analyse the cases is based on the construction of models or types of cultural cluster, taking into account the predominant social interaction dynamics generated between the cultural agents (e.g. creators, producers, intermediaries, managers, consumers, etc.) that participate in this type of grouping. The task involves looking at: (a) the socio-genesis of the cultural clusters, distinguishing between top-down and bottom-up strategies; (b) the predominant type of interactions -relationships between individuals, collectives and institutions; and (c) the type of interactions -relationships between individuals, collectives and institutions and the context at district, neighbourhood and city levels (according to the degree of integration, separation and transformation of material and symbolic resources and the pre-existing socio-cultural dynamics). The following section presents a series of concepts from sociology for analysing the cultural clusters of the city of Barcelona.

Analysis of the types of social of social dynamics of cultural clusters in Barcelona
Barcelona has generated a model of urban development to a large extent based on culture, which has tried to combine the attention to the local population's needs with an enhancement of its self-image among the citizens through internal promotional campaigns in order to create a social consensus on the city project (Marshall, 2004;McNeill, 2001), with an increasing effort to project the city's image internationally. In this sense, Barcelona has become a brand (Balibrea, 2004). This brand is focused on emphasizing the city's Mediterranean temperament and the figure of Gaudí and his creative character while dismissing the memory of the city's industrial economy and its history of political rebelliousness (Balibrea, 2007). The selective historical memory that Barcelona's branding has involved has caused certain academics and political activists to describe Barcelona as 'The Liar City' (Delgado, 2007;Espai en Blanc, 2004).
However, there is a consensus that the Barcelona brand image has enjoyed remarkable international success as a high technology city, a conference location, and a city of trade fairs and arts festivals (Degen and García, 2008;García, 2004;González, 2011;Majoor, 2011;Trullén, 2001;Walliser, 2004). This process has been the outcome of the combined efforts of Barcelona's local government and the Catalonian regional government, who from the 1980s have generated a paradiplomacy that has resulted in their leading of city or region networks, with a special tendency to employ culture as a resource both to build up a network of contacts and complicities and to gain visibility in the international arena (Zamorano, 2012). The results of these actions can be seen in different facets of the city's economic activity, but perhaps the clearer indicator of this success is reflected in the steady increase in the number of international tourist visits to the city (Casellas et al., 2010), which increased from 2.4 million visitors in 1993 to 7.13 million in 2011 (Turisme de . Barcelona is a city with considerable cultural activity and heritage. Aware of this situation, the local elites have made culture a central element when redefining strategies for the city's future (Rodriguez Morató, 2008). These strategies revolve around: (1) the generation of big cultural events as an element of the city's symbolic and material transformation (Subirós, 1998); (2) the conversion of cultural and knowledge sectors along with tourism into vectors for developing the local economy (Trullén, 2001); and (3) the aesthetic transformation of the city following certain principles indicative of class (Julier, 1996(Julier, , 2005. These strategies quickly led to Barcelona City Council's cultural policy being geared towards joint governance between private and public cultural sectors. This type of governance is aimed at boosting the creation of cultural value (Rius, 2005), combining urban regeneration and cultural planning (Subirós, 1999), and facilitating the conversion of different city spaces into cultural clusters.

The cultural cluster as a bureaucratic organization or clusters of cultural institutions: Montjuïc
The Ciutat del Teatre (City of Theatre) began in 1997 with Barcelona City Council commissioning the prestigious theatre director Lluis Pascual to lead a project involving various institutions that either already existed or were under construction in the area of Monjuïc (Pascual, 2001). The project coordinated three different theatre spaces: the Mercat de les Flors (dance), the new Teatre Lliure (contemporary theatre) and the Institut del Teatre (further education in theatre and dance). The project was implemented using a top-down strategy originating in a request from the independent theatre sector, which was seeking to consolidate itself in the Teatre Lliure, then based in Gràcia (Rius, 2005). The aim of the initial project was to create synergies between the institutions: The Ciutat del Teatre is a cultural project conceived in all-inclusive strategic terms. It aims to give shape to and boost the energy generated by the three facilities located in the same area of Montjuïc so as to create a centre for the scenic arts that will also act to revitalize this area of the city (Antón, 1999: 158). The Ciutat del Teatre will also act as a training, research and dissemination platform for professionals in the various scenic disciplines. (Gual, 2003: 23) Social interactions in Montjuïc are strongly structured by rules and hierarchies connected to stable, regular organizations (Crozier and Friedberg, 1982). It can be said that the cultural cluster of Montjuïc is dominated by a bureaucratic social dynamic. This type of dynamic defines the social interaction of those cultural clusters characterized by a spatial concentration of national cultural institutions and facilities due to their relatively large size and organizational and economic complexity (Rius and Rubio, 2013). Since the beginnings of cultural politics, cultural facilities have been one of the main executive arms and showcases of public policy. In the case of France -since culture was designated a public action category in 1959 -cultural policy has centred on spreading high culture as far as possible via cultural facilities. To a great extent this focus on spreading artistic excellence still continues today (Dubois, 2010). However, since the 1980s a new component has been included: the will to improve and brighten up the city while at the same time revitalizing less privileged areas (Urfalino, 1994). This type of project gave rise to the idea of clustering cultural facilities, with the dual aim of increasing the impact of the facilities and generating collaborative synergies between them.
Projects aimed at clustering cultural institutions manifest themselves in the construction of large infrastructures that require relatively large, functionally specialized management teams with a centralized internal hierarchy. These are therefore what the sociology of organizations characterizes as bureaucratic organizations (Crozier, 1964). The interaction dynamics between different cultural institutions towards the exterior are conditioned by political-institutional confrontation (between parties or different public administrations), by different definitions as to the mission of each cultural institution, by rivalry between their artistic directors (who consider collaboration an attack on artistic freedom) and by the legal-juridical framework (which in the case of Spain is very restrictive).
The Institut del Teatre and the new Teatre Lliure were opened in 2001. However, the Ciutat del Teatre project as a whole, which was intended to become a cluster of private agents from the theatre world, was abandoned by those in charge of the cultural administrations. The main reason for this was that they did not want to give up control of the cultural institutions. Therefore, although the Montjuïc cluster project initially integrated participative dynamics from the cultural sector, this clashed with the bureaucratic dynamic at work in the way it was run.
Nevertheless, the cluster's interaction dynamics are stable although not very intense: the three institutions share exhibition spaces, and students from the Institut del Teatre present their creations in the other two institutions as part of the Assaigs Oberts cycle (a showcase for students' work). The collaboration dynamics, despite shared common ground in the shape of the theatre, are based on agreements between the managements of the two institutions and are formalized in administrative law or under the inter-administrative agreement. Collaboration, far from being spontaneous, is a result of decisions made by the politicians in charge and is implemented from the management down to the organization (top-down).
One of the objectives of the Monjuïc cluster is the revitalization of an area, Mount Montjuïc, densely covered by cultural and sports facilities. However, the theatre cluster has been implemented using a basically top-down dynamic. This, along with its orientation towards professional theatre, with no community theatre line, explains why it has had little effect on the urban areas surrounding it.

Cultural Industry Clusters:
Social interactions in cultural industry clusters take place within a framework of associative-utilitarian logic typical of economic transactions. Therefore, the predominant type of social tie in the framework can be characterized according to the idea of society put forward by Tönnies (2011): a type of interaction in which economic-professional relationships take priority over friendships. In associative interactions the ties are defined mainly according to interactions based on projects aimed at the production and sale of cultural goods and services (Cooke and Lazzeretti, 2008;cf. Krätke, 2011). Within this type of dynamic there is a clear separation between spaces set aside for work, leisure and home. These factors condition the type of interaction with the social context and vary according to whether the cluster is geared towards cultural production or consumption.
The cultural production cluster: District 22@. This project officially began in 2000 and involved the renewal of approximately three-quarters of the spaces used for production activities. As Martí-Costa and Pradel (2011) point out, 22@ is part of the urban renewal strategy for the old manufacturing, working-class neighbourhood of Poblenou, initiated in the late 1980s by the municipal government. In this respect the type of clustering strategy is top-down based on encouraging industries linked to creativity and knowledge, and transforming and renewing neighbourhood infrastructures and facilities.
The cluster was planned with the intention of developing projects linked to the new economy based on the setting-up of various thematic subclusters grouping together 56% of the companies located there. These companies are related to information and communication technologies, medical technologies, energy, media and design (Ajuntament de Barcelona, 2010). Taking into consideration only the two types of sub-cluster that could be qualified as cultural industries -media and design -this involves just 63 companies of widely varying sizes and legal structures (from universities to the selfemployed). These cultural sector companies include publishers (5) and audiovisual producers (4). As far as the theatre sector is concerned, there is the Fura dels Baus theatre company and Focus, the most important production and exhibition company in the theatre sector. As a whole, the weight of the cultural sector in itself within the framework of 22@ is minor and split between different sectors, unlike in other cultural production clusters in other European cities (Scott, 2000).
The sub-clusters were conceived as having a public-private governance structure. The media sector is led by the Centro de Innovación Barcelona Media, which is intended to be a 'neutral meeting point for collaborative R&D processes, a link between academic and industrial research' (Barcelona Media, 2012). The cluster's other coordination point, the Parque Barcelona Media, plans to allocate 12,000 m 2 for developing the media sector. This is currently under construction. In this case it involves a publicprivate cooperation project with the participation of Barcelona City Council, the Pompeu Fabra University and the audiovisual producer MediaPro. 3 The cultural production cluster at 22@ appears as an integration strategy for the creative production chain characteristic of the art world (Becker, 2008). Interactions between its members are based on the fact that they share certain conventions originating from within the discipline itself, conventions that are no doubt subject to debate and which generate cultural innovation when they are adopted by some of the profession. The actors in cultural production clusters develop strategies to create or destroy oligopolies using certain stylistic innovations (Peterson and White, 1979). This tendency favours the institutionalization of new fields in cultural industries and their agglomeration in one space, such as in the case of country music (Peterson and Di Maggio, 1975).
Despite being an example of a cultural cluster designed and implemented using a top-down method, 22@ has tried to generate cross-sectional association dynamics between companies and to this end has an association to encourage networking. However, none of its members are from the cultural sector. When asked why they had relocated to 22@, the director of entertainment group Focus stressed that it was 'in response to an offer made to us by the City Council' (interview with the director of Focus, 2011). He added that Focus did not collaborate with other companies in the cluster because 'its spaces for creating and exhibiting are outside 22@, in our theatres in the city centre' (Peterson and Di Maggio, 1975). The lack of any critical mass of cultural companies, their heterogeneity and the lack of crosssectional collaboration has meant that, in cultural terms, 22@ has not seen any of the expected benefits of agglomeration. Its presence in the district is explained by its participation in a project designed by the administration and by the facilities offered in terms of infrastructure rather than by any expected benefits from the interactions that normally justify clustering strategies for cultural industries.
The relationship with the social and urban environment can range from partial integration into the pre-existing social fabric to a tendency towards complete substitution (Martí-Costa and Pradel, 2011). In this case there is an absence of social relations with people from outside the cluster (apart from those needed to provide services to its professionals) and an appreciation of isolation as a condition for the development of the professional dynamic that can be seen reflected in the promotion of a 'one-size-fitsall' urban planning approach unrelated to the local environment (Muñoz, 2010). In addition, as Martí-Costa and Pradel (2011) point out, the construction of 22@ has meant the disappearance of most of the spaces that contained artists' studios, thereby dismantling the pre-existing artistic community (Martí-Costa and Pradel, 2011).
The cluster as an association oriented towards cultural consumption: Consell de Cent. Unlike cultural production clusters, the cultural industry clusters geared towards consumption have a more intense although ambivalent relationship with their surroundings. These groupings are becoming more and more aware of the weight of the brand associated with the social space. Therefore, there are strategies to find locations close enough to prestigious cultural institutions so they can fall within their aura of modernity (Moulin, 1997). Another strategy is to associate their brand with a particular piece of heritage or creative community (Zukin, 1995). However, this strategy has a clear limit. Cultural consumption clusters are conceived with a single objective: to attract casual visitors and turn them into loyal customers. Their relationship with the urban space is therefore divided between a discoursively community strategy and a concealed instrumental dynamic. Their creation is generally implicit and bottom-up. This is the case of the gallery cluster of Consell de Cent.
The first art galleries came to Consell de Cent in the early 1960s, forming an initial nucleus for exhibiting and selling modern art in the city of Barcelona (Torres, 1993). Since then the Consell de Cent galleries have promoted local and international artists and grown in prestige to become the centre of the art market in the city, grouping together 22 art galleries, 24% of the total (Rius, 2002). However, the predominance of the Consell de Cent cluster cannot be explained by numbers alone, but also by the importance of the galleries concentrated in these four blocks of the Eixample. Figures from the Departament de Cultura (2006) show that, while the average turnover of an art gallery does not exceed 320,000 euros, in Consell de Cent the average is 545,000 euros, with three galleries turning over in excess of 2 million euros. As a whole, the galleries in Consell de Cent account for 58% of the total turnover of art galleries in Barcelona. Unlike the clusters analysed so far, the Consell de Cent cultural cluster is generated by a dynamic from within the gallery sector itself, without the intervention of the public administration (a bottom-up strategy).
Cultural consumption clusters in this sense are to be found either in places where potential customers live nearby or in places normally used by the uppermiddle classes for consumption and leisure. The absence of social relationships beyond the circle of the cultural cluster along with the high disciplinary specialization of its work means that the social and urban environment is not important as an element for creative activity. On the contrary, being isolated from the surroundings is seen as an asset when it comes to creating an artistic atmosphere to attract buyers (Molotch and Treskon, 2009).
When asked why they decided to open in Consell de Cent, art dealers have a dual discourse consistent with their being double agents participating in both the artistic and economic fields (Bourdieu, 2002a). First of all they have a notion of community based on an idea of the common good: I think it was good for everyone that I came to Consell de Cent. Good for the other dealers because I was bringing new life to the street, and good for me because I was excited at the thought of opening in a street where there are people really dedicated to the art world.
(Interview with gallery owner in Consell de Cent, 2003) Collaboration between galleries has turned Consell de Cent into the benchmark for modern art for the Barcelona art buyer, together creating a brand of artistic quality associated with the area. However, when asked about practical collaboration, the gallery owners limit themselves to talking about possible joint openings organized by the professional associations. In this sense it can be said that, on a discoursive level, there is a principle of fictional solidarity at work among the dealers based on an idea of community that, in practice, is in contrast to the interaction dynamic based on rational-instrumental interest, in the end forming an associative dynamic within which all their interactions are framed.
As analysed by Moulin (1983), the gallery owners are concentrated in one place in order to attract buyers who are geographically scattered. However, in the case of Consell de Cent there is another reason, which is its proximity to the Passeig de Gràcia, placing them close to the main luxury shopping area in Barcelona (Mars, 2006). The clustering can also be interpreted in a less disinterested way: all the dealers point out that Consell de Cent has succeeded in creating a quality brand that, by being associated with the cluster, increases sales. By grouping themselves together in a limited area with few available premises and high rents, the result is that not many dealers can enter the cluster. This is the way in which we should understand the dealer who said that: Crises are times of big opportunities. I had to wait for a gallery to go bankrupt and free up a space before I could move here. It was lucky that I had a dealer friend who told me about the opportunity. (Interview with Consell de Cent gallery owner 2) Clustering is also a way of creating an oligopoly in attracting customers: demand is strongly concentrated in this street and therefore there are many dealers who want to move to Consell de Cent. However, one of the functions implicit in clustering is to prevent the arrival of art galleries of a low artistic level, which may endanger the quality brand enjoyed by the whole street.

The cultural cluster as a community or creative cluster: Raval
Since the beginning of the 20th century the Raval has been well known locally and internationally for being a bohemian, marginal neighbourhood (Villar, 1997). From the start of the urban regeneration process in the early 1980s there has been an artistic population centred around what was then called the Barrio Chino. This pre-existing substratum has been boosted by the arrival of new creators attracted by the area's dual condition of bohemian neighbourhood and cultural cluster (Subirats and Rius, 2008). There has been no census of the artistic population, but its presence can be felt in the fact that almost 50 artists participate in the Tallers Oberts de Ciutat Vella, when artists' studios are open to the public (Foment de les Arts Decoratives, 2011). This artistic presence has been visible for years through the widespread presence of urban art on walls in the neighbourhood (Rius, 2008).
From the 1990s in the Raval we see the emergence of institutions that reproduce a bohemian subculture outside the 'official' Barcelona, which attracts creators from all disciplines (Aisa and Vidal, 2005). Indeed the neighbourhood inherits from the Barrio Chino its tradition of bohemian spaces, some of which still survive (El Cangrejo, Café Teatro Llantiol, Teatro Riereta). This legacy is combined with fringe cultural spaces designed by young creators (Espai Mer, Areatangent, Forn de Teatre Pa'Tothom, Miscelanea, Almazén) that function as a 'Third Space' (Lloyd, 2010) in which they can rehearse and exhibit their alternative shows to other creators and a restricted public outside the commercial circuit. This alternative Raval scene is self-run and self-financed. This situation gives artistic life in the neighbourhood a dynamic character, although with ups and downs, and efforts have been made to consolidate and project this beyond the artistic community itself through various attempts at occupation, such as the Teatre Arnau and the Teatro El Molino in 2006, although without success.
Unlike cultural industry clusters, creative clusters such as Raval are framed within communitytype relationships. Within this framework the most important social ties are affective and personal ones (Tönnies, 2011). The community dimension of creative clusters has been described on numerous occasions in the history of art and literary studies (Franck, 2003;Lottman, 1981) and sociology (Lloyd, 2010;Simpson, 1981). This kind of literature characterizes the relationships between creators based on ties of friendship, disinterested relationships, 4 a mixture of work and leisure, the joining of workspace and homespace (in the shape of the artist's studio, later known as the loft), the relationship with the social environment within the cluster and the appreciation and transformation of the environment.
Economic analyses and analyses of the artistic professions usually highlight the extraordinary concentration of artists in big cities and give reasons for their location there: the greater abundance of artistic work and spaces for training and exhibition (Rodríguez Morató, 2001) and more pay and prestige (Menger, 2009). However, economic and professional factors do not completely explain the phenomenon of cultural clusters. Literature on the bohemian life, from its beginnings in the mid-19th century, highlights the unconventional lifestyle which, according to Chiapello (1998), forms part of the artistic criticism of bourgeois life and, according to Bourdieu (2002a), forms part of the habitus typical of the artistic field. In this sense the concentration of creators in one area makes it easier for a bohemian subculture to emerge (Fischer, 1995). Generally speaking, this type of cluster is conceived implicitly, on the basis of collaboration and mutual help networks of a bottom-up type. Grouping together creates a 'critical mass' powerful enough to generate institutions to reproduce this subculture outside the community. Artistic neighbourhoods in this sense seem like a stage where bohemian life is played out, separate from the urban spaces of the middle-culture (Lloyd, 2010).
Various sociological studies reflect this sense of community in artistic neighbourhoods. These papers generally focus on the type of relationship that the artists' community establishes with the neighbourhood: their capacity for urban transformation (Simpson, 1981); the social changes they bring about (Zukin, 1989); the type of affective and dedifferentiated tie established (Chalvon-Demersay, 1999); the common meeting places (Lloyd, 2010); the use of the social and urban environment as material for symbolic production (Lloyd, 2010); their identification as a creative brand (Rius, 2008); their involvement in the political struggle against renewal projects that could mean their expulsion due to gentrification phenomena (Zukin, 1989); and in order to build institutions to enable them to project their creations to the rest of society and claim their status as social agents (Fischer, 1995).
Unlike the other clusters analysed, in the Raval we see that there is a community of creators, especially in the scenic arts and music, who develop intense collaboration guidelines within the framework of the neighbourhood as a shared social and creative space. Living together in a single space facilitates collaborations, fusions and common projects, such as the case of the compilation CD the Barcelona Raval Sessions. In addition, the interaction dynamics between musicians at first take on a spontaneous and disinterested form until they transform into a commercializable format. At the CD presentation it was explained that: It all began in early 2002 (…) with the start of a recording studio project open to the neighbourhood and especially its musicians (…) and this became studio 08001, a centre of pilgrimage through which would pass, in the course of ten months, over twenty musicians who spontaneously and disinterestedly day after day created what was first a rough cut and which today is released under the name Raval ta Joie. (Radiochango, 2003) Creative clusters generate community-type social interactions on the basis of a common lifespace and workspace, on cooperation in early career stages and on the exchange of cultural references and artistic skills in order to generate new cultural products.
The Raval appears to be an urban space in which a large number of interactions between creators take place. However, it is not only a question of the number but also the quality and intensity of these interactions that turns the Raval into a creative laboratory. Also, the Raval artistic community's level of commitment and emotional attachment to the urban space is very intense. An investigation into social and symbolic change in the neighbourhood recorded dozens of cultural productions (cf. Subirats and Rius, 2008) in various formats inspired by or set in the neighbourhood (novels, poetry, theatre, comics, film, documentaries, music, etc.). This shows that the artistic community converts the urban space into not only a creation space but also an object of symbolic production.

Conclusions
Cultural clusters are very important phenomena at a cultural and urban level. They emerge as a response to the challenges of the new economy and as a strategy for urban development (Karlsson, 2010;Lorenzen and Frederiksen, 2007;Mommaas, 2004). Analysis of their emergence has focused on economic and urbanistic aspects, despite the fact that there is ever greater agreement as to the importance of the social interactions developed within them (Comunian, 2011;Currid and Williams, 2010;O'Connor, 2004). Sociology can make important contributions based on classical sociology concepts, such as community and association (Tönnies, 2011), urban sociology (Fischer, 1995) and sociology of the arts (Becker, 2008;Bourdieu, 2002a;Moulin, 1983Moulin, , 1997. This paper has used these theories and concepts to debate the usual cultural cluster analyses centred on economic flows and urban planning and their one-dimensional conception of cultural clusters in order to suggest that there are three types of cultural cluster and that in each one there is a predominant interaction logic. These forms of interaction have been classified as being of three types: bureaucratic, associative and community. While cultural clusters are usually analysed from a single dimension and understood as being of the same type, the hypothesis defended in this paper is that it is precisely because of the different nature of their organization and social interactions that differences between cultural clusters can be established. These differences are not due to contextual or contingent factors but involve the organizational dynamic of social action in general (Crozier and Friedberg, 1982), the structure of the artistic field in particular (Bourdieu, 2002a) and the interactions that come about in creative processes on the basis of certain shared conventions and rules (Becker, 2008). This explains why, despite the discourse of positive aspects on cultural clustering, creative dynamics may or may not develop within them, depending on the social characteristics of their participants and the logic that leads to this participation (disinterested or instrumental). To some extent this explains the failure or success of different types of cluster.
In the case of Barcelona, we have seen how the creation of clusters has been one of the phenomena that have contributed to redesigning and defining the identity of large areas of the city. Four cases have been identified that are representative of the cultural cluster typology presented: a cluster with bureaucratic organization dynamics (Montjuïc); two following the association logic of cultural industries, one geared towards cultural production (22@) and the other towards cultural consumption (Consell de Cent); and a final case that can be classified as following the logic of the creative community (Raval).
As far as the logic characterizing the different clusters is concerned, in general they match the characterization described earlier. Cultural institution clusters come about following a top-down logic that places the definition of the project and a good proportion of its development in the upper levels of government. However, it is true that in some cases there was a more participative cultural sector dynamic, but in general these clusters are highly conditioned by the political-administrative structure, inter-elite conflicts, their rigidly defined mission definitions and a hierarchical dynamic in which decisions are made by the directors. This gives a context that explains why few collaborations among different sectors come about in this type of cluster. As regards association clusters, here we find greater differences in the genesis and the resulting dynamic: in the case of 22@ the design is clearly top-down and, although there is an attempt to generate part of its development through autonomous management (sectoral bottom-up), the cultural industries are disjointed and have generated no cooperation or solidarity dynamic. This contrasts with the official discourse of 22@, which justifies itself on the basis of the benefits of agglomeration. The art gallery cluster, on the other hand, is a bottom-up phenomenon in which an apparently community discourse conceals an instrumental logic: to create a quality brand, attract potential customers and prevent unwanted competition from accessing the market. This is therefore a case of instrumental association. Finally, in the Raval we see a community dynamic in which there is cross-sectional and initially disinterested cooperation between equals that succeeds in creating a different bohemian subculture which proclaims itself as such.
The relationship with the urban space is very conditioned by the type of cluster. Many bureaucratic organization clusters have been geared towards urban regeneration. In some cases the starting point for this has been the idea of turning the area into a tabula rasa (Montjuïc), while in other areas the neighbourhood has been treated with greater sensitivity, which has been of benefit to the cluster itself. In all cases, however, their bureaucratic organization dynamic geared towards a mission defined by politicians and implemented from an elitist viewpoint of culture has involved barriers in relation to the urban space and its inhabitants. A similar phenomenon has happened with 22@: a redefinition of the productive space has swept away the residential dynamics and dismantled the weak fabric of artistic studios that used to exist. The Consell de Cent cluster, on the other hand, does not aim to transform the urban space but to take advantage of the resources it provides (proximity to luxury shopping streets) and its configuration as a compact urban space in which to create a dense group of artistic spaces, enabling it to gain its label of central gallery district. Finally, in the case of the Raval, the artistic community's relationship with the urban space is intense: it forms a dense space for living, working and consuming that enables a subculture to be produced that in turn is material for the symbolic production of the community's own cultural productions.
Finally, cultural clusters with community-based social dynamics -as is the case of the Raval and Poblenou-Distrito 22@ -have undergone substantial transformation processes on the basis of culturedriven regeneration policies and urban development. In both cases the community-based social interactions, for the most part related to the artistic sector, have been eroded and partially replaced by a new type of relationships -of a more bureaucratic nature -between cultural institutions (such as in the case of the Raval's cultural institutions) or of a more associative-utilitarian nature between creative companies (as in the case of the companies of Distrito 22@). This situation causes on many occasions strain or conflicts among actors performing different types of dynamics of social interactions, and this generates a significant effect on the urban creativity processes. However, this article was not targeted at examining this type of process and the conflicts it creates, but to showcase ideal types allowing for a classification on the basis of the predominant social dynamics of every case.
In conclusion, it can be said that there are three very different types of cultural cluster and that, in the case of Barcelona, their performance from a creative and artistic point of view differs greatly depending on the type of cluster and therefore the type of social dynamic that maintains them. Institutional cultural clusters are very efficient tools for redefining urban spaces and certainly make it easier to attract the public, but the level of cooperation is very limited, at least up to now. Neither can it be said that association clusters geared towards the market are very productive in this respect: without greater sectoral coordination or within a cultural production chain, their substantive performances in cultural terms are rather mediocre. Artistic community clusters, however, are shown to be truly productive spaces on a cultural level because they make it possible to generate locally different subcultures that express themselves in a number of different artistic ways. In this respect we maintain that one of the elements favouring the creativity of cultural clusters is their relationship with community-type dynamics. Analysing how and under what conditions these interactions develop between different types of cluster is a challenge to be undertaken in future research.