Talking about the same but different? Understanding social movement and trade union cooperation through social movement and industrial relations theories

In the context of continuing political, social, and economic crises, trade unions in most European countries are seen as weakened, protest as having little influence, and the solidarity between European populations as damaged. Under these circumstances, scientists and practitioners have placed hope in the cooperation of social movements and trade unions in order to revitalise trade unions and achieve common goals. However, the conditions for forming alliances between trade unions and social movement organisations have, thus far, been primarily researched, and partially theorised, from the point of view of different disciplines. In doing so, approaches that emphasise the strategic alignment with resources and context diverge from approaches that highlight the relevance and necessity of shared ideologies or identities. This contribution aims to bring existing approaches from movement and industrial relations research into dialogue with each other and calls for a further integration of both perspectives. It contributes to a more holistic understanding of joint movement and trade union action current concepts such as social movement unionism cannot provide. It uses two examples to show that instrumental, strategy-driven modes and identity-based, culturally-driven modes of cooperation are not contradictory. There are situations in which strategic decisions on resources and political influence are more decisive than ideological proximity and vice versa. These findings are of social and scientific relevance for understanding mechanisms of solidarity construction and processes of bridging differences even in increasingly fragmented and unequal societies.

Why do trade unions and social movement organisations work together?Social movement organisations and trade unions have organised joint protest events, shared campaigns, or worked together using a variety of strategies of political influence.These often have been labelled social movement unionism (Ford, 2009;Mathers, Upchurch & Taylor, 2019;Nowack 2017;Zajak, Piper & Egels-Zanden, 2017).Still, there have also been occasions where they have competed with each other; organising separate events (e.g. the May 1st trade union demonstrations or alternative social movement demonstrations) or reaching out to the same target group (e.g.precarious workers or the unemployed) (Choi & Mattoni, 2010).For the most part, however, they ignore each other.They operate in different spheres (the industrial sphere versus the public sphere) and usually have neither incentive nor necessity to work together.They differ in their organisational structure (bureaucratic and hierarchical versus networks based on informal and decentralised participation), underlying democratic principles (representative democracy versus undertakings in direct democracy), or motives of collective action (material and employment related concerns versus postmaterialist values) (Grote & Wagemann, 2019).Indeed, both social movement and industrial relations literature have stressed that trade unions and social movements represent distinct organisational forms.It is against this background that social movement and industrial relations theory developed rather independently and have followed "their own avenues" (Grote & Wagemann, 2019, preface), despite having shared roots in Marxist thinking (Yon, 2016).
Industrial relations scholars take the inherent conflicts of interest between employers and employees as a basic starting point for their analyses.They study the diverse institutional arrangements that shape the employment relationship, ranging from norms and power structures on the shop floor to collective bargaining arrangements at different levels (Dunlop, 1993(Dunlop, [1958]]; Jackson, Kuruvilla, & Frege, 2013;Pries, 2010;Müller-Jentsch, 1995).Social movement scholars, on the other hand, explicitly distance themselves from explanations grounded in class relations as a sufficient condition for mobilisation and protest, arguing that factors such as resource mobilisation, political opportunities, framing, and identities are paramount (Della Porta & Diani 2010;Rucht, 2017;McAdam, McCarthy, & Zald, 1996).Touraine (1984) linked the decline in prominence of Marxist theories to the emergence of new post-industrial social movements in Europe and to the fact that trade unions had, to a great extent, been incorporated in the (neo-) corporatist structures of the state.This led to a decreasing interest in trade unions in social movement theories.Despite this, empirical studies began to explore the relationship between social movements and trade unions, but they did so from the perspective of either social movement studies or industrial relations (IR), or to be more precise from the field of labour studies within IR. 1  For social movement scholars, the temporary neglect of trade unions changed when researchers re-discovered and re-included capitalism in their analysis (della Porta, 2015;Barker, Cox, Krinsky, & Nilsen, 2013).Especially in the context of the European economic crisis, studies began to re-consider trade unions as relevant actors (Kanellopoulos, Kostopoulos, Papanikolopoulos, & Rongas, 2017;della Porta et al., 2012). 2 Social movement scholars started looking into the development of cooperation across organisational differences from the cultural perspective, which has become dominant in social movement theories, focussing on processes of brokerage, bridge building and the construction of a joint identity (Baumgarten, 2014;Flesher Fominaya, 2010) Among industrial relations scholars, the interest in social movement-trade union cooperation developed in the context of the trade union revitalisation debate (Schmalz & Dörre, 2013;Rehder, 2008;Frege & Kelly, 2003).Taking into account the decline of unionisation and collective bargaining, social movements have increasingly been considered as strategic partners who could compensate for the loss of structural, associational and institutional 1 A lot of work on trade unions has been done in the "sociology of work", "industrial sociology", "political economy" and "labour history", which could also add valuable perspectives, concepts, and empirical evidence, especially concerning cultural aspects.However, this is beyond the scope of this paper (for example Bieler et al. 2015;Gallas 2015;Jansen 2013). 2 Yon speaks of a "long-awaited homecoming" of the labour movement in social movement studies (Yon, 2016, p.82).Empirically, trade unions have (historically) always been part of social movements.For example, in a longitudinal study of protest events in Germany, Rucht showed a significant amount of trade union participation (2007).
The term was introduced to capture social movement and trade union cooperation with the overall aim of democratizing all social relations (Waterman, 1991). 3Thus, in contrast to cultural approaches, which are key in social movement studies and stress the importance of shared identities and worldviews, this perspective predominantly focused on strategic considerations in order to gain additional modes of political and economic influence (Schmalz & Dörre, 2013;Webster, 2015;Brookes, 2013).
In sum, to answer our question on why trade unions and social movements cooperate, the literatures on industrial relations and social movements provide two explanations: one perspective stresses the strategic considerations on how to increase power, influence and resources; the other perspective emphasises the relevance of shared world views, common identities and other cultural similarities.What we do not know is how these factors relate to each other.Which combination is more likely to result in the initialisation or disintegration of cooperation?By cooperation we mean working together, for example organising a campaign or protest event on more than one occasion.We use the word cooperation or alliance instead of collective action to indicate: firstly that we are looking at organisations not individuals; secondly, that the actors still keep some autonomy and distinctiveness -they do not merge into a single entity (see Rucht, 2004, p. 203); and thirdly, avoidance of the normative bias inherent in concepts such as social movement unionism or networks of labour activism (Zajak et al., 2017).The contribution of this paper is two-fold: first this article follows Grote & Wagemann's (2019) call "to take a position of analytical equidistance with respect to both forms of collective action".By looking at how cooperation is conceptualised within social movement studies and IR theories, we show that both research fields do consider similar factors when explaining cooperation between collective actors -the politicaleconomic context, resources and culture -albeit in different ways.By including factors from both research fields, we contribute to the development of a more holistic model of cooperation between these collective actors.Second, based on the combination of these three explanatory dimensions, we identify and differentiate between various pathways that lead to cooperation.We choose two examples that contrast one case where shared roots and understandings made cooperation possible, despite limited strategic chances to reach certain goals or increase organisational resources (the Alter Summit network), with another case where movement organisations and trade unions cooperated in spite of significant ideological differences (Anti TTIP network).Our results suggest taking a closer look at the pathways and drivers that lead to cooperation or its break-up.Next, we give an overview of the key theoretical concepts from both research fields -social movement studies and IR -before presenting our two examples.

3
The term has been used widely and across a broad range of countries (see also Köhler et al.'s and Fink's contribution in this volume).However, the concept is based on a normative view, claiming that cooperation with social movements is necessary for trade unions not only to revitalize but also to become more democratic and less hierarchic (Waterman 1991, Nowak 2017).The concept has little value for explaining why cooperation empirically emerges in some contexts but not in others and why sometimes cooperation is sustained while in other times it dissolves.

Paths to cooperation between trade unions and social movements
Both social movement and industrial relations research use similar dimensions to explain collective action, but give differing names to specific theories or analytical concepts.In social movement studies, resource mobilisation theory, political process theory, collective identity, and framing concepts have been most influential in the study of collective action (Rucht & Roth, 2007;Hellmann & Koopmans, 1998).In labour studies, the resource dimension has been dealt with in the power resource approach, context is considered in the political-economic perspective, and cultural aspects play a role in concepts of ideology and class consciousness (Bieler et al., 2015;Gumbrell-McCormick & Hyman, 2013).Table 1 summarises the main dimensions.The question, then, is which combination of factors best explains cooperation.Movement research long ago stressed that looking at only one theoretical paradigm (political opportunities, resources or framing/identities) is not enough to understand mobilisation or collective action (McAdam et al., 1996).This can also be said for trade union actions.In the following section, we cross-fertilize both disciplines and point out various indicators in all three dimensions, which will help us to understand joint actions involving social movements and trade unions.

Dimension I: Context
Social movement and IR scholars have both stressed the crucial importance of institutions for explaining the strategies and strength of movements and trade unions (Tarrow, 1996;Müller-Jentsch, 1995).In social movement studies, the most prominent approach to conceptualise how context affects the development and outcome of social movement is political process theory, with its key concept of political opportunity structure (POS).Activists do not choose their goals at random.The frequency and intensity of protest actions depends to a large extent on the openness of the political system (Meyer, 2004)."Political opportunities are understood as signals that encourage or discourage actors to form social movements" (Tarrow, 1996, p. 54).Political pluralism, divisions of elites, institutional access, or decline in repression all count as favourable opportunities (Meyer & Minkoff, 2004). 4The presence of potential allies has also been treated as an opportunity dimension, as finding allies is "critical for a movement's survival, particularly when it is in an outsider position" (Rucht, 2004, p. 197).From POS theory we can derive two broader aspects that define incentives for social movement-trade union cooperation.Firstly, one basic premise of the POS approach is that the political context shapes the grievances around which movements organise, thereby shaping social movements.Groups tend to cooperate with each other and formulate joint tactics, goals, and claims based on those shared grievances (Goldstone & Tilly, 2001).Grievances can emerge as a response to the immediate political environment, but they can also be more deeply rooted in societal cleavage structures (Kriesi, 2004).Going back to Lipset & Stein (1967), social cleavage theory argues that political action is deeply rooted in key conflicts within society.Key divides include class conflict, religious divides, core and periphery cleavages, and, as more recently stated, cosmopolitan or transnational cleavage (Hooghe & Marks, 2018;Eggert & Giugni, 2015).Kriesi et al. (2012) argued that the more prominent the class cleavage in society would grow, the narrower the space for cultural protest, associated with so called "new social movements", would be.In terms of cooperation between trade unions and social movements, the cleavage argument suggests that the strong relevance of class cleavage would weaken so called "new social movements" and make them less appealing as cooperation partners for trade unions.This strongly resonates with research in IR, which found that strong unions are less likely to cooperate with social movements.Gumbrell-McCormick & Hyman (2013) found that the decline of traditional participatory power influences trade unions to search and use alternative power resources through coalition building.In countries with stronger and more coordinated institutions, cooperation with business and with state agencies is considered a core strategy and, thus, strategizing with social movement organisations may be less necessary (Frege et al., 2004).
Still, this assumption could also be challenged.As della Porta et.al. ( 2017) state, the austerity crisis in Europe has transformed the cleavage structure in affected countries by enhancing the precarisation of labour and a proletarisation of the middle class.A strengthening of the class cleavage, however, would imply a weakened interest from trade unions in cooperating with "new social movements".Nevertheless, the fact that European societies are under severe social stress also means that many different groups are affected and can become mobilizing actors and therefore potential cooperation partners.Being exposed to similar threats of economic depression, austerity policies, and arcane, high-speed political decision-making might create a particularly fertile ground for alliance formation (Roose, Sommer, & Scholl, 2018;Zajak, 2018).With this in mind, Alter Summit is an interesting example of a European wide coalition of trade unions and social movement organisations in response to the European crisis.We show below that these contradictory aspects are not mutually exclusive, but represent two tendencies that organisations are exposed to in the aftermath of the economic crisis in Europe: on the one hand, fragmentation due to the national search solutions, and, on the other hand, the welding of organisations through a new European cleavage structure.

4
The term structure has been largely abandoned in political opportunity research, as most scholars have agreed that opportunities have to be perceived, interpreted, and framed, and depend on the specific historic context and situation (Gamson and Mayer, 1996, p.283).
Secondly, the availability of institutionalised access of one organisation can increase the interest of others in cooperating with that organisation.We can assume that activist organisations that wish to increase their influence through improved institutional access will develop an interest in joining forces with trade unions.Insider organisations might also be able to strengthen their position by participating in a broader network, for example during consultation procedures with DG Trade in the European Union (Fogarty, 2013).This assumption is challenged by the perspectives of IR scholars who study the participation of trade unions and co-determination in diverse institutional arrangements (Berger et al., 2018;Pries, 2010:).Trade unions that have institutionalised sources of influence (e.g.Germany or Nordic countries) (Bernaciak, Gumbrell-McCormick, & Hyman, 2014) are expected to be less likely to cooperate.Trade unions that are not well integrated in corporatist structures, countries like Greece, Spain and Portugal, are more open to cooperating with social movements.
Overall, using the industrial and political context as explanatory context suggests that different organisational forms might work together due to strategic considerations, even if they do not share the same specific goals, identities, or frames.Closely linked to theories dealing with context are approaches that take the relevance of resources into consideration.Resource mobilization theory (RMT) in social movement studies and the power resource approach (PRA) in labour studies both share the perspective of strategic choice with context theories. 5

Dimension 2: Resources
According to resource mobilization theory (RMT), a group with a minimum of resources must start the mobilisation process and reach out to attract more resources from others (Jenkins, 1983;Polanska & Piotrowski, 2015).Social movement resource mobilisation theorists remain relatively vague when it comes to specifying resources.Tilly (1978, p. 10), for example, lists: "labour, power, goods, weapons, votes and any number of other things, just so long as they are usable in acting on shared interests".
Predefining resources is difficult because resources can vary across contexts and depend on the goals that the collective actors strive to achieve and the tactics and strategies they adopt (Etzioni, 1968).Despite the variety of potential resources, it is of course possible that organisations strive to gain the same resources.In cases in which these organisations have different interests, it is likely that they will be in a competitive relation and cooperation will not occur between them.A helpful differentiation was made by Rogers (1974) and Jenkins (1983), who suggested differentiating between types of resources by asking whether they enable the exertion of power or whether they serve the mobilisation process itself.Overall, social movement research has been more interested in resources that serve the mobilisation process than with power resources (in contrast to the approaches in IR).From this point of 5 Resource mobilization theory has also inspired the development of the power resource approach e.g. in Kelly's work (1997Kelly's work ( , 1998)), also Darlington (2012).
view, organisations as such are considered a key resource, because they contribute to the emergence, growth, and maintenance of movements over time by providing infrastructure for communication between the different components that form a movement, as well as facilitating communication with organisations from other movements or with institutions and individuals (Kendall, 2005).This perspective suggests that cooperation is more likely to occur between organisations that share similar organisational characteristics in terms of income, number of paid staff and internal committees, chapters, and number of members (Andrews, Edwards, Al-Turk & Hunter, 2016). 6Following these insights, we assume that the more similar the movement and trade union organisations are, the more likely it is that cooperation will occur.In other words, social movements which have developed, or integrated institutionalised social movement organisations are more likely to work with trade unions than movements that do not have this type of organisation in their constituency.The latter often relies on more disruptive power resources (cf.McCarthy & Walker, 2010;Caniglia & Carmin, 2010).Similarly, unions that employ disruptive strategies have a decentralised structure, have limited organisational characteristics, and depend directly on their ability to mobilise (e.g.anarchist unions), could be more prone to cooperate with movements that prioritise the same resources.
As movement scholars tend to be vague about the type and nature of resources, we suggest adding the power resource approach developed within union revitalisation studies, as it is more specific.This strand of research tried to identify new power resources that could help to revitalise trade unions.Despite some variation between different approaches, structural, associational, institutional, coalitional, and discursive/symbolic power resources count as the main pillars of the power resource approach (Brookes, 2013;Schmalz & Dörre, 2014;Webster, 2015;Zajak, 2017).The basic assumption emerging from this perspective is that trade unions seek alliances with social movements when they lose associational power (the ability to organise workers and to mobilise them to take collective action) (Silver, 2003;Wright, 2000), structural power (the ability to disrupt), or institutional power (exerting leverage through social partnership institutions) (AK Strategic Unionism, 2013).
The assumption is that unions that lose members and the ability to disrupt production or that experience a decrease of their institutional access will cooperate with social movements under two conditions: movement organisations either have the ability to leverage institutions (Brookes, 2013) or the ability to win over the public and thus possess symbolic power, through their ability to influence what is publicly perceived as right or wrong (McGuire, 2013;Chun, 2009;Webster, Lambert, & Bezuidenhout, 2008).However, not all trade unions view symbolic power as a necessary or relevant resource, and thus do not perceive cooperation with social movements to be a good strategic choice.As Ibsen & Tapia (2017) have shown, many unions have responded to the collapse of collective bargaining and membership decline through a strategic re-orientation, organising new members and mobilising the existing members, thereby shifting focus from institutionalised power resources to trade unions' original associational power source (Urban, 2010;Frege, 2004).It is therefore possible that unions that strongly emphasise membership will only cooperate with movement organisations if they expect new membership through activism.Unions will not cooperate with movement organisations if they consider the organisation unable to leverage symbolic power (i.e. because they consider the organisation or their action repertoires as illegitimate).
Overall, PRA has been developed to explain cooperation from the trade union perspective but does not really consider which resources the potential cooperation partner, that is not a trade union, needs.It is possible that organisations, even in a favourable political context and in the presence of opportunities to increase influence through cooperation, might not cooperate because of a different assessment of what would count as an appropriate resource.Similarly, unfavourable opportunity structures could be ignored or altered when the alliance structure is able to acquire sufficient resources to overcome obstacles in the political-economic context.Still, what is viewed as a power resource (or not) is a matter of shared frames and ideology -in other words we need to include the cultural dimension to understand cooperation.

Dimension 3: Culture and shared identities
The last dimension refers to cultural factors like frames, shared identities, classconsciousness, and ideologies as key factors in explaining collective action.The hypothesis resulting from this perspective is straightforward: alliance formation can hardly be understood "without reference to cultural processes of identity formation and group solidarity" (Buechler, 1993, p. 230).The aspect of cultural influence is reflected in both social movement and industrial relation research, but with different intensities regarding theoretical relevance and empirical application and preferring different concepts and approaches.Obviously, there is no such thing as "the one" cultural approach, just like there are many definitions of culture (for a recent overview in social movement studies see Baumgarten et al., 2014).While some cultural approaches define any strategic choice as a cultural practice (Hardy & Robin, 2014; for a broader discussion of the strategy concept see also Kryst in this volume), we maintain that we can differentiate between choices made based on strategic considerations -we call that instrumental alliance or strategy-driven cooperation -and cooperation based on shared identities, consciences or world views -we call that identitybased or culturally-driven cooperation.
"Cultural differences in organisational culture, in language, education, or traditions, can create misunderstandings between groups that undermine trust and therefore decrease the likelihood of cooperation between actors with similar strategic goals" (Flesher Fominaya, 2010, p. 17).
However, movement scholars have also highlighted different mechanisms in dealing with cultural differences and settling disagreements (e.g.Polanska et al., 2015).Most importantly, organisations need brokers or bridge builders (Tattersall, 2010), which help to connect organisations across such differences, for example by developing mutual respect and recognition of the issues and grievances of others (Beamish & Luebbers, 2009;Brookes & Zajak, 2018).These people help to overcome certain barriers and to construct a new sense of collective identity or, at the very least, a similar interpretation of the problem and its solution (framing).Within social movement studies, a shared identity is considered the most crucial factor for collective action.Cooperation between groups and organisations would not be possible without a collective actor that helps to construct imagined communities that "are able to define themselves and their relationship with the environment" (Melucci, 1995, p.43).A collective identity is perceived as being essential for cooperation, as it facilitates the formation of trust and feelings of mutual belonging (della Porta & Diani, 2010;Melucci, 1995Melucci, , 1996;;Touraine, 1977).
The addition of these factors helps us to further specify our assumptions for cooperation between social movements and trade unions as follows: Firstly, we can assume that organisations that stress differences and conceptualise social movement organisations as being different, as "others", might not cooperate, even when opportunities are present, or resources needed (Flesher Fominaya, 2010;Snow, 2001).In contrast, identities that are "highly elastic" in nature seem to facilitate cooperation, even with groups that, on the surface, do not seem to share goals and commonalities.A particularly relevant example of this is the Global Justice Movement (Flesher Fominaya, 2010).Similarly, we could speculate that organisations that stress solidarity, understood as 'perception of commonalities of interest and purpose which extend, but do not abolish, consciousness of distinct and particularistic circumstances' (Hyman, 2001, p. 170) as a strong norm, are more likely to cooperate with each other.
Secondly, there are also autonomous activist groups that reject the idea of a shared collective identity or representative politics and emphasise the role of individual expression. 7 We can assume that organisations stressing autonomy and self-realization reject cooperation with most trade unions, which are the organisational expression of representative politics and collectivistic decision making.
Thirdly, groups that share similar frames are also more likely to cooperate.Frames are "interpretive schemata" used to simplify reality by "selectively punctuating and encoding" actions and events "within one's present or past environment" (Snow & Benford, 1992, p. 137).Following Benford & Snow's (2000) differentiation of prognostic and diagnostic framing, we assume that organisations are more likely to cooperate when they are able to define a social grievance as a problem and identify its causes ("diagnostic framing") and if they have similar strategies of problem handling ("prognostic framing").
Conversely, divergent frames, particularly when connected to identity and belief systems, 8 can be a barrier in the construction of alliances between organisations.Such differences are difficult to overcome because they are based on experience and routinized social practice (Baumgarten et al., 2014), but also because values shape the understanding of legitimate and appropriate actions.As Polletta explains, "culture sets the terms of tactical choice" (Polletta, 2008, p. 86), emphasising the culturally rooted nature of more radical and moderate tactics.
Cultural differences have been researched by labour scholars in regards to trade union ideologies and, from a Marxist perspective, class-consciousness.In Industrial Relations: A Marxist Introduction, Richard Hyman defines class-consciousness as: "the recognition that all who sell their labour in order to live, by virtue of that very fact, have common interests in opposition to those who own and control capital" (Hyman, 1975, p. 178).

7
This is why some scholars like McDonald (2002) began to argue against the concept of collective identity.8 In contrast to frames that are constructed strategically and are defined as being a response to the context, instead of one's own ideology or belief systems.
It is important to underscore that there is a difference between class-consciousness and class-identity, as the latter is understood as the recognition of a role as an agent of class (Hyman, 2001), but this recognition is not politically conscious enough "to challenge a class-based society" (Anderson, 1977, p. 334).This is particularly relevant when discussing the role of trade unions.Workers may acknowledge that they need to cooperate with other workers to defend their interests against the interests of the employers, but this cooperation does not necessarily lead to a questioning of capitalist society and its mode of production as a whole (Hyman, 2001).Trade unions may, but do not necessarily have to, arise due to an emerging class-consciousness (Moore, 2011) and therefore reflect different positions vis-àvis capitalism.Whether or not a union is likely to seek cooperation with social movements seems to be related to the role that class-identity plays in its ideological orientation.Hyman describes an "eternal triangle" (2001, p.3) of three ideological orientations within which all trade unions can be positioned: class opposition (where the union seeks to mobilize its membership to challenge the existing social and economic order); integrationist unionism (where the union conceives of itself as a social partner that accepts the broad constitution of society and a plurality of legitimate interests but seeks reform in the interest of its class); and business unionism (where the union defines its sphere of representation in the market) (Hyman, 2001).Frege et al. (2004) used Hyman's (2001) distinction to explain the cooperation of trade unions with social movement organisations in different countries.They observed that trade unions with class opposition background tend to prefer militancy to cooperation with capital and thus tend to build coalitions of protest with social movement organisations.Furthermore, integrationist unionism leads to a strong tendency to build coalitions, particularly coalitions of influence, where alliance partners have multiple access points and opportunities to exert influence, while business unionism oriented trade unions tend not to cooperate with social movements at all (Frege et al., 2004, p. 13).In a recent study, Andreta, Bosi, & della Porta (2016) came to a similar conclusion on trade union and social movement cooperation in Italy.We consider the specification of the nature of organisations and their ideological positions as crucial for understanding which types of organisation tend to cooperate and which do not.Movement and trade union organisations might refrain from cooperating with each other, even if the organisations think that cooperation would increase their power or leverage, due to ideological differences or perception of the other as distinct or illegitimate.The interesting question is then, if and how such differences can be overcome through 'cooperation work', such as by reconstructing, aligning, or amplifying frames (Benford & Snow, 2000), developing more abstract master frames (Gerhards & Rucht, 1992) or collective identity work (e.g."the enemy of the enemy is my friend" logic).

Pathways to cooperation
Different combinations of factors can lead to cooperation.We can, for example, assume the following.It is highly likely that organisations will cooperate if there are incentives set by the institutional environment, if resource gains can be expected, and the organisations share similar goals and world views -thus all factors (institutional incentives, resource gains and shared identities) are present (path I).On the other hand, if all these factors are absent it is plausible that organisations will tend to ignore each other and not cooperate (path II).However, more interesting are those cases in which we see contradictory incentives, for instance it is possible that a strong shared belief system leads to cooperation, despite an unfavourable context or expected loss of resources (path III).In contrast, the context can be open for new mobilisation and resource gains could be expected, all of which should make the strategic choices of working together likely, even if the organisations differ in terms of their world views, identities, or underlying ideologies (path IV).This reasoning follows Olson's argument on the "logic of collective action" (1965), in which he stresses that individuals make a rational decision to cooperate when they might benefit from a certain action or good which they alone could not achieve or.This is why in our example we will focus on the paths III and IV. 9 We will use the example of Alter Summit to show that developed shared identities, the belief in transnational solidarity, and shared frames against austerity politics facilitated cooperation ("identitybased, culturally-driven cooperation"), in spite of how unlikely it seemed that organisations would increase their access or influence through the network.We will then use the case of protests against the Transatlantic Trade and Investment Partnership negotiations in Germany to illustrate how approaches drawn from social movement theories and labour studies can help to understand the impact of strategic choices on cooperation ("instrumental or strategy-driven cooperation").

Two examples of paths to cooperation
In the following section, we provide a brief example of each of the two cases of cooperation.The first case (Alter Summit) exemplifies cooperation primarily occurring as the result of prior shared experiences, shared interpretations, and framing of the problem ("identitybased, culturally-driven cooperation"), despite a significant lack of influence or expected resource gains.The second case, anti-TTIP mobilisation, highlights how cooperation between trade unions and social movement organisations occurred regardless of the fact that they had very different positions in terms of their overall goal and their ideological differences in relation to free trade ("instrumental or strategy-driven cooperation").These examples should provide a brief illustration of two significantly different paths to cooperation.The illustrations are largely based on desk-research and 10 qualitative interviews with involved organisations conducted as part of a larger research project. 109 Also interesting are cases where all factors are favorable and still no cooperation emerges and cases where everything seems to speak against cooperation but it still happens.However, these cases also seem very rare and exceptional and therefore we focus only on III and IV in this article.Other combinations of factors are also possible.For example, closed opportunities, a lack of shared identities, but the involvement of resource strong organizations willing to share those resources.Future studies could further explore different variants of these pathways and the contextual conditions under which they become relevant.10 Research group "Transnational Alliances between Social Movements and Trade Unions.In depth case studies are still work in progress and will be presented at a later date on our project website http://www.isb.ruhruni-bochum.de/forschung/graduiertenkollegs/transnationale-allianzen/index.html.en.

Example 1: Alter Summit
Alter Summit is a network of trade unions and social movement organisations from various European member states, founded in 2012/2013 in order "to strengthen the people's struggle against austerity and for genuine democracy". 11In the run-up to the network formation, the initiating organisations 12 had known each other and worked together for many years in the European Social Forum and had subsequently built the Joint Social Forum, before forming Alter Summit as the latest attempt to consolidate their cooperation.These previous joint experiences facilitated the establishment of personal ties and gave participating organisations the time and place to develop a shared identity, attempting to build "a united movement for a democratic, social, ecological and feminist Europe". 13Alter Summit is an interesting case that highlights how collective identities developed over time, based on and a further strengthening of the belief in transnational solidarity, and shared frames opposing austerity politics.This facilitated cooperation, contrary to what could be predicted using resource approaches (resource mobilization theory and power resource approach alike).Firstly, the development of the new network was resource consuming for all organisations: representatives had to be sent to the European meetings and financial contributions were required.For smaller and resource-poorer organisations (e.g.SAT-Andalucía) investing money, knowhow, and working hours was a challenge they had to overcome.Secondly, the network offered few opportunities for national trade unions to increase institutional, associational, or discursive power resources.Trade unions engaged the European Union, where neither national unions, nor social movement organisations, have significant access to the political processes.Some social movement organisations might have gained material resources through cooperation with big national trade unions in the Alter Summit network, but they could not increase their institutional access or discursive visibility either.On the contrary, cooperating with unions is difficult for some movement organisations (e.g.some 15-M initiatives), as some perceive trade unions as an auxiliary part of the capitalistic system.For trade unions, it was unlikely that they would gain more members or discursive power, as Alter Summit was not well known and did not significantly engage in public campaigning or the organisation of public protest.On the contrary, trade unions could lose public legitimacy by using an alternative channel of cooperation beside the European Trade Union Confederation (ETUC).Thirdly, the joint framing against austerity politics presented a challenge for some of the organisations involved, as austerity-measures affected countries differently and, for the most part, not at all.Of the 22 countries from which the Alter Summit organisations come from, only Spain, Greece and Portugal 14 were directly affected by austerity policy.
To understand why organisations still invested their resources in the Alter Summit network, we must take a closer look at the cultural and contextual factors.The founding organisations of the network had years of joint work in which to converge their ideological background and to develop shared framings regarding the causes of the crisis (which started in 2007/2008, years before Alter Summit was created), interpreting it as a systemic crisis of capitalism.During international gatherings of leftist movements, networks, and organisations such as Florence 10+10 in 2012, founding organisations of Alter Summit had the opportunity to develop shared prognostic framings, looking for collective strategies to assisted the fight against the austerity crisis.These interactions helped to construct the shared understanding that building a European network where having diverse organisations "support [ing] and strengthen[ing] each other's struggles" was an important step en route to stopping austerity and challenging the European balance of power. 15The European debt crisis and the prevalent austerity measurements gave them plausible opportunity to consolidate their cooperation, rooted in the belief that a European network between "different types of organisations and movements, especially trade unions, social movements and networks of researchers and institutes" is necessary, with the goal "to effectively influence and exert all its weight on the EU policies". 16owever, a shared perspective on the crisis, ideological closeness and a belief in transnational solidarity were not enough to maintain cooperation, at least not for German trade unions like GEW, IG Metall, Verdi and DGB youth.Although involved in the social forums prior to and at the beginning of Alter Summit, they left the network after a few meetings as they began to view it as being ineffective.They were not only dissatisfied with the meagre media attention that the new network attracted, they also perceived the other participating organisations as being insignificant, meaning that the symbolic power resources they apparently anticipated were lacking.Additionally, for strategic reasons and in order to maintain their associational power resources among their members, they perceived the ETUC to be a more legitimate European channel for political action.This suggests that integrationist unions (like GEW, Verdi & IG Metall) prefer to cooperate with social movements in instrumental coalitions, while cultural factors continue to hold the rest of the network together.

Example two: Anti-TTIP Mobilization
To illustrate how approaches from social movement theories and labour studies can help to understand strategy-driven cooperation we will now look at the protest march against the Transatlantic Trade and Investment Partnership (TTIP) in Berlin, Germany on October 10 th , 2015.
Negotiations regarding a trade and investment agreement between the European Union and the United States started in 2013.The European Commission believed that TTIP would not only provide a boost for the economy, including growth in jobs, especially of those in export-oriented industries (European Commission, 2013), but also that it would "set the standard -not only for our future bilateral trade and investment but also for the development of global rules" (European Commission, 21.02.2013).Numerous organisations, networks, and groups of the civil society in several European countries shared the belief that TTIP would set a standard but were not in agreement with the European Commission's positive assessment.In Germany, an exceptionally large and heterogeneous alliance emerged, criticising the lack of transparency during the negotiations and, in part, the agreement itself.The alliance mobilised in the form of a European Citizen Initiative (ECI) and several large protest marches, against the passage of TTIP and for transparent and democratically controlled negotiations (European Initiative against TTIP and CETA; Trägerkreis Großdemonstration am 10.Oktober 2015 in Berlin, 29.08.2015;Trägerkreis der bundesweiten Großdemonstrationen "CETA & TTIP STOPPEN! -Für einen gerechten Welthandel!").The alliance in Germany was particularly surprising as Germany, the forerunner in export, stood to reap significant benefits from TTIP and German trade unions, supporters of the export orientation of the German industry, seemed unlikely to join mass protest against TTIP.And yet the alliance that organised the largest protest in Berlin on October 10 th , 2015 consisted of, among others, the German Trade Union Confederation (Deutscher Gewerkschaftsbund, DGB), Attac Germany, Brot für die Welt (a developmental aid organisation affiliated with the Protestant church), Naturfreunde Deutschland (an organisation with roots in the socialist and environmental movement), Greenpeace, Mehr Demokratie (which lobbies for more direct democracy elements in German politics), and Arbeitsgemeinschaft Bäuerliche Landwirtschaft (ABL -the German member of the global network of the peasants' movement, Via Campesina).These organisations hardly shared a collective identity or classconsciousness and the cooperation also did not seem to build on common ideologies or worldviews.Even the frames regarding TTIP itself are heterogeneous as some of the organisations renounce free trade outright and therefore aim to prevent TTIP (e.g.NaturFreunde Deutschland; Attac), while other organisations, for example the DGB, do not reject trade agreements in general, but demanded a modification of TTIP (Gegenblende das DGB-Debattenportal, 2015).The lowest common denominator might be found in the shared belief in the need to intervene in the TTIP negotiations.To understand why and which organisations cooperated based on such a small common ground, it is necessary to take context and resources into account.
At first glance, no cooperation would be indicated based on the contextual factors from both social movement and industrial relations studies: German trade unions of the DGB continue to be embedded in social partnership institutions and Germany still counts as a coordinated market economy, where trade unions would seek influence through institutional channels rather than through public protest.Also, from a resource perspective, it seems unlikely that they would gain more members, increased access, or any kind of material resources by cooperating with social movement organisations in joint protest events.Even their legitimacy and discursive power could be at stake given that the DGB is involved in Labour20, a Network of trade unions trying to represent trade unions' demands within the G20 and supports the Trade Union Advisory Council (TUAC) of the OECD.Therefore, in other trade related contexts the DGB seeks to be perceived as a legitimate voice within consultations, as opposed to protesting and questioning the legitimacy of these consultations.
To understand why trade unions and social movements cooperated nonetheless, the concept of shared grievances can offer possible explanations: the original TTIP proposal included an Investor-State-Dispute-Settlement (ISDS) mechanism and regulatory cooperation which were both perceived to be an extraordinary threat to democracy not only at a European, but also at a national, level (European Initiative against TTIP and CETA).These mechanisms could affect every area of legislation, from labour law to consumers' rights and environmental protection, and created an opportunity for a large alliance of various social movements and organisations.For the German trade unions this threat may have had significant weight, as it could have potentially undermined their institutional power resources at the national level.Another shared grievance can be seen in the negotiation process itself, which started out as a very secretive process, taking measures to keep documents of the negotiation content confidential (arrangements on TTIP negotiating documents, 05.07.2013), thereby making it almost impossible for trade unions, as well as for social movements, to gain access to detailed information.These specific circumstances of trade negotiation may have made it impossible for the German trade unions to rely on their institutional power resources, thus creating an incentive to ally with social movements in order to strengthen their discursive power resources.For the social movement organisations involved, cooperation with the German trade unions meant benefitting from a huge increase in resources in terms of outreach and infrastructure.For example, the organisation of busses and mobilisation of union members contributed to the size of the TTIP protest in Germany.
The position of the unions may have been influenced by a third factor: though they can be perceived as following the integrationist model, their link to the Social Democratic Party (SPD) has weakened (Dribbusch & Birke, 2014).Given that the SPD had been part of the governing coalition during the negotiations, and that their chairman, Sigmar Gabriel, had been heavily in favour of TTIP, German trade unions may have had another incentive to search for new allies outside the party system.At the same time, it is possible that social movements may have speculated that old ties between German trade unions and the Social Democratic Party could help to influence the European Parliament and German government, thus perceiving unions to be, partially, insiders.
Finally, cooperation mainly took place between "resource similar" organisations, just as resource mobilisation theory predicts.Trade unions cooperated with large and relatively formal social movement organisations like Attac, BUND, and Campact.The organisational similarities probably facilitated bridge building across the ideational differences of social movements and trade unions vis-à-vis globalization, global governance and trade.
These brief illustrations of two cases show that the factors that explain cooperation can be very different, depending on the specific actor constellation and setting -an aspect which receives little consideration in conceptualizations of social movement unionism.

Conclusion
This paper examined prominent theories within social movement studies and industrial relations in order to theorise and hypothesise factors and conditions that influence cooperation between social movement organisations and trade unions.We departed from earlier studies, which looked into cooperation through the particular viewpoint of either industrial relations or social movement theories.Our paper contributes to increasing the understanding of cooperation between social movement organisations and trade unions by further integrating various analytical dimensions of both social movement and industrial relations theory (Grote & Wagemann, 2019).We used concepts and indicators from both theories to distinguish between different pathways to and drivers of cooperation, resulting from different combinations of these factors.We chose two cases to highlight the differences between instrumental, strategy-driven cooperation and identity-based, culturally-driven cooperation.Our results suggest that we cannot simply assume that a shared identity or shared strategic goals are necessary conditions for cooperation; both aspects are important, but to varying degrees under different circumstances.Thus, different paths to cooperation are possible and more systematic and comparative research, within and across countries, is needed in order to be more precise about the specific combination of factors leading to cooperation or the breaking of alliances.The outline of the different pathways suggested in this paper could be a first step towards such a comparison, which future studies can explore.Further research is imperative as it increases our understanding of how different organisational forms can cooperate with each other, even across cleavage structures or deep ideological divides.Such cross-organisational and cross-ideological cooperation is paramount in ostensibly increasingly fragmented societies.

Table 1 :
Explanatory dimension of cooperation