Introduction

Numerous studies in industrial relations have for several years sought to identify the challenges faced by trade unions in both the North and South (^{Anner et al., 2006}; ^{Silva, 2013b}). A first body of research has emphasized the structuring conditions of globalization, making the perspectives of union renewal obsolete. Whether involving the internationalization of production or the loss of effectiveness of collective bargaining, the sources of trade union decline are countless. The dual movement of the financial concentration and decentralization of production among transnational companies has made the latter practically elusive, for both labor law based on the integrated firm-model (^{Martin, 2015}; ^{Martin, Dufour-Poirier and Villanueva, 2021}) and for trade unions. A second body of literature has attempted to demystify the paths to union renewal. For these authors (including ^{Frege et al., 2004}), globalization appears to have brought constraints for trade unions anxious to push through the boundaries of their traditional zone of action (^{Haiven, 2006}). This body of research seeks, among other things, to examine the variety and complexity of external networking initiatives, of which trade-union cross-border alliances are at the forefront (^{Bieler et al., 2015}; ^{Erne et al., 2015}; ^{Bieler and Erne, 2015}; ^{Lévesque and Murray, 2010}), also referred to in the literature as global company networks (^{Croucher and Cotton, 2009}).

These studies have the merit of outlining the potential for alliances to achieve the social regulation of employment and transnational companies and counterbalance their growing power (^{Routledge and Cumbers, 2009}). As mentioned in ^{Dufour-Poirier and Hennebert (2015)}, “these alliances, which bring together unions from different countries representing workers from a single multinational corporation (MNC), have emerged as platforms for exchange, dialogue and the coordination of union action.” (p. 1). There is a consensus, however, among several classic authors in the field that identities are at the core of trade union projects (^{Offe and Wiesenthal, 1980}) and the source of their renewal (^{Hyman, 1994}), even transnationally (^{Frangi and Zhang, 2021}). Moreover, the results emerging from more recent analyses (^{Sarkar and Kuruvilla, 2020}; ^{Smale, 2020}) call for researchers to conceptualize more clearly what labor transnationalism is, and better grasp the meaning of success in labor transnationalism. By extension, other authors (^{Brookes, 2019}; ^{Frangi and Zhang, 2021}) also converge in pinpointing the need to better contextualize the debate on union identity (^{Dufour-Poirier, 2011}; ²⁰¹⁶; ^{Dufour-Poirier and Lévesque, 2013}).

In this context, the article intends to answer the two following questions. First, what are the impacts of trade union identities on the internal functioning of a cross-border alliance? Second, what are the contingencies (Wright's concepts of structural and spatial power) (2000) shaping the organizational identities at stake in this alliance? This article draws on ^{Gahan and Pekarek’s (2013)} idea of collective action frames and on Greer and Hauptmeier’s (2012) notion of identity work, by which they refer to the processes through which an organizational identity is created, sustained and modified. Their studies provide tools for analyzing transnational union solidarity through the lens of collective identity framing processes. In keeping with ^{Hunt and Benford’s (2012)} work, collective identity framing processes are understood here as the raw material of union representation and an essential requirement for overcoming problems, reframing differences of opinion, and facilitating commitment prior to engaging collective action. Given that trade unions frame their identities through their interactions, but also in response to the multi-faceted contingencies they face, this article contends that collective identity framing processes (agency), such as transnational cross-border alliances, intervene prior to the emergence of new modes and levels of action pursued by unions (structure). Therefore, the way in which the opportunities for action are perceived and used depends on these organizational identities, which are meaning makers: as social and institutional actors, unions (as an amalgam of rank-and-file members and leaders) are considered here as organizational identities’ holders (^{Whetten and Mackey, 2002}). For us, collective action (in this case, cross-border alliances) depends, first and foremost, on the actors identifying as a group and identifying with the goals and the project that such a group pursues, thus fostering a sense of belonging. Without such a collective consciousness, committing to, and endorsing, a collective project will be considered irrelevant (^{Hunt and Benford, 2012}). This argument also takes into account the downstream contingencies that come into play in order to better understand the internal dynamics of transnational alliances.

To substantiate the case, this article considers an in-depth intra-case analysis of a North-South alliance that came into being in the Americas, where the absence of any state power regulating such initiatives made their emergence more complex and disorganized (^{Bieler et al., 2015}). This alliance brought together local unions and national trade union organizations from Canada, Chile and Peru operating in the mining industry. The lifespan (from 2000 to 2008) of the alliance under study involved three major phases of development. In the end, the alliance could appear to be a failed transnational solidarity initiative. However, we think that wider lessons can be learned from this case. This paper is divided into five sections. The first section describes the research method and is labelled a longitudinal methodology. The second section recalls that Collective Identity Framing Processes at the Heart of Cross-Border Union Solidarity. The third section describes the 3-phase process of building the North-South Alliance at Stake. The fourth section presents the contingencies that undermined the development of this alliance. The fifth section, Conclusion, discusses broader considerations in regard to the effectiveness of cross-border trade union alliances.

A Longitudinal Methodology

This project's originality lies in the extended duration of the observations on which it is based. The alliance emerged as a result of six local union initiatives (two refineries and one mine in Quebec, Canada; two mines in Peru; and one refinery in Chile), initially without the assistance of a Global Union Federation (GUF). The three local unions based in Quebec were affiliated with a major nationwide industrial union in Canada (referred to hereinafter as the Industrial Union) that operates in the mining industry. Right from the start, the local unions in Quebec, which launched the alliance, benefited from the support of the Industrial Union. At the time of this study, the Industrial Union was the largest private-sector union in North America and the most active in the metal and mining industry, notably due to its strong activism at the transnational level. Over the years, the Industrial Union invested all kinds of resources in the alliance, such as funding the travels of union delegations abroad. It also coordinated some follow-up meetings, and ensured, for a few years, the presence of an international project manager based in Canada, responsible for building relations between local trade unions and national federations and confederations in the North and South. Lastly, the Industrial Union coordinated contacts and followed up on the initiatives developed during these meetings. From 2000 to 2006, the Industrial Union in Canada, through its international project manager and until his retirement, more or less formally took on the responsibility for coordinating and monitoring the alliance's activities.

Some 77 one to two-hour, semi-structured field interviews were conducted between 2004 and 2018 with 60 key union representatives working in local, national and international labor organizations in Canada, Chile, Peru and Switzerland. Each of the interviewees was recognized as playing a key role in creating and coordinating the cross-border alliance under examination. These interviews were conducted in the four countries during formal or informal meetings. All the interviews were taped in full and transcribed for analysis. The data collected was analyzed using a method drawn from studies by ^{Yin (2003)} in which the researcher is asked to describe the case study as accurately as possible. The following (Table 1 in Appendix 1) provides an overview and chronology of the fieldwork conducted for this study.

Our approach was mainly descriptive and involved an intra-site analysis: this means that the type of unionism, the technological and the economic changes along with the specificities of the mining industry (and its growth) prevailing in one country in particular were not targeted in this study. The alliance as a whole was considered the preferred unit of analysis throughout our research process.

In order to analyze the data effectively, we first transcribed the interviews, recorded with participants’ consent, in minute detail. We also took into consideration the notes, comments and impressions resulting from the discussions that were conducted informally, often at the end of the interviews. Close reading of these transcriptions enabled us to put the still-raw data that had been recorded into perspective with the necessary distance. Secondly, we carefully reviewed all the documents from a critical standpoint through attempting to identify the obvious facts they contained and decipher what they possibly alluded to or implied. Thirdly, the verbatim of the interviews conducted with the union actors were broken down into units of meaning and compared with one another to bring out the similarities and differences on a North-South basis. This coding into conceptual units was a way to identify the analysis conducted by assigning a code or key words to a word, line or paragraph. In doing so, we had a dual objective: 1) to allow the codes to emerge from the data, and, especially, 2) to gradually extrapolate from the recorded material a general, meaningful pattern on the basis of our analytical framework. The codes used with the Atlas.ti software confirmed the presence of a number of patterning levels and stages in our analysis process. Following that, we carried out an iterative pruning of data deemed equivalent, recursive or unnecessary to prioritize other data considered essential for properly understanding our subject. This constant process of manipulating and revising the codes assured us of the relevance, conceptual validity and accuracy of the encoded passages: the identification of regularities and possible co-occurrences result from a first, second, and even third review of our data, refining the level of analysis and abstraction developed as a consequence of our contact with the field during the various research phases. This organization took the form of synoptic tables that also provided an overview of the quotes used to substantiate our argument.

The sources of information used include interview data, corporate documents, articles from the print media and union publications. All such materials, as well as our field notes and various travel reports, were coded in their entirety. Triangulating the data from these different sources and methods of data collection allowed us to achieve data saturation, further validating our findings. This wealth of written information and verbal accounts helped to deepen our understanding of an alliance built in a particularly unstable corporate context. Visits to plants and meetings with workers at several production sites in the three countries took place at different phases of the study. Our status as an observer and a researcher, which included accompanying delegations of Canadian union representatives during visits to Chile and Peru, was always clear, in all phases of the research process.

The accounts collected revealed the respective organizations' overall viewpoint in terms of identity. Our analysis focused nevertheless on the internal functioning dynamics of the alliance as a whole. It should be stressed that we did not look at the dynamics punctuating the organizational life of each of the participating unions. Last, but not least, ethical constraints completely prevent us from revealing the identity of the company and the representatives consulted. Any breach of anonymity could potentially affect the safety of some of these representatives, particularly those in the South. The interviews will simply be referred to by number and the year in which they were conducted. Our results are presented chronologically, based on the changes that occurred within the transnational employer, CanMin1, which later became CanMin2, and lastly EuroMin. Table 2 in Appendix 2 enables an assessment of the magnitude and scope of the fieldwork that took place over a period of 10 years, which not only allows us to maintain the validity of our analyses and of the concepts used, but also the reaching of the saturation point desirable in any research initiative.

Collective Identity Framing Processes at the Heart of Cross-Border Union Solidarity

Transnational alliances are a complex phenomenon, still in the midst of a period of growth and experimentation. However, their study is not new. A first body of literature has explored them from the perspective of cross-border trade union alliances. These alliances are composed of groups of trade unions representing workers in the same transnational corporation and are established in a more or less formal, ad hoc and structured manner in order to militate for the safeguarding of trade union rights, to denounce union repression or to support a struggle therein (Greer and Hauptmeier, 2012). They create opportunities for actors to coalesce around common issues to challenge current power structures and intervene on multiple levels, despite the distances (physical, political, linguistic, etc.) separating them (^{Dufour-Poirier, 2011}; ²⁰¹⁶; ^{Dufour-Poirier and Hennebert, 2015}; ^{Dufour-Poirier and Lévesque, 2013}). Studies such as those by ^{Fairbrother et al. (2013)} have documented the process leading to the creation of such alliances, the forms of action emerging from them, and the obstacles they face. However, ^{Brookes and McCallum (2017)} “suggest that this cacophony of case studies on labor transnationalism from diverse disciplines offers more trees than forest” (p. 208). ^{Anner et al. (2006)} also claimed, 15 years ago, that more systematic and contextualized comparisons were needed to synthesize all these very diverse case studies.

A second body of literature has examined the variety and complexity of external networking initiatives, this time under the cover of heterogeneous alliances (^{Routledge and Cumbers, 2009}; ^{Tattersall, 2010}; ²⁰¹⁸). These networks of activists entail the establishment of relatively durable, flexible and inclusive relations between union organizations, and sometimes other civil society actors, in order to work together on broad socio-political issues that go beyond the strict framework of traditional union demands (e.g., wages) in workplaces. These alliances differ from one another in terms of the variable involvement of the actors participating in them, the fluctuating intensity of relations between them, the collective actions stemming from them and the level of intervention favored. Through them, trade unions open up new channels of communication that are likely to rebuild the effectiveness of their actions and their political leverage on the societal scene (^{Dixon and Martin, 2012}). Without exhausting all theoretical avenues, studies on cross-border alliances (^{Bieler et al., 2015}; ^{Dufour-Poirier and Hennebert, 2015}; Greer and Hauptmeier, 2012) and social movements (^{Gahan and Pekarek, 2013}; ^{Melucci et al., 1989}) have acknowledged the importance of organizational identities in the process of launching collective action, and the need to deepen our understanding of the micro forces shaping them (^{Dufour and Hege, 2010}; ²⁰¹³; ^{Fox-Hodess, 2020}; ^{Frangi and Zhang, 2021}; ^{Smale, 2020}).

Similarly, ^{Kay (2011)} sees labor transnationalism “as a process of creating a transnational union culture based on cooperative and complementary identities defined as shared recognitions of mutual interests coupled with a commitment to joint action” (p. 27). By extension, a growing body of both theoretical work (^{Hunt and Benford, 2012}) and empirical studies (^{Kirton and Healy, 2013}) has also claimed that solidarity gives rise to social and moral cohesion (^{Morgan and Pulignano, 2020}) which, in turn, depends on an awareness of the group and on identification with the group or with the collective conscience and goals that it disseminates. This identification process refers to the delineation of the we-feeling which, in turn, entails a sense of mutuality and solidarity that enables organizational actors to share a sense of loyalty, morals and emotional meanings in time and space and to endorse a project based on a perceived common cause and fate. Building on Melucci's previous work (1995), actors thus get involved in collective action if they have the ability to maintain these identities that make undertaking and justifying strategic action legitimate and possible. Collective identities act then as a sort of strategic compass, guiding the way actors frame the problems with which they are confronted, interpret situations and give meaning to their relations with others and their environment.

However, reinstating the importance of collective identities does not discount the multi-level nature of the challenges confronting trade unions. Collective identities are socially constructed and contextually anchored, as are trade union behaviors, which are highly sensitive to context (Greer and Hauptmeier, 2012). Even though identities are far from being determined only by external contingencies (^{Marginson, 2016}), trade union responses to transnational solidarity are nevertheless shaped by the latter and subject to constant change: identities can only be considered as the outcomes of external opportunity structures and self-awareness. We thus contend that the driving force behind a cross-border trade union alliance lies in collective identity framing processes; that is, trade union participation in a transnational alliance is based on a union’s ability to frame and upstream, in relation to the contingencies facing it, the relevance of uniting around cross-cutting causes in order to engage in a collaborative process. Transnational solidarity has to become a meaningful priority for the trade unions involved and they need to identify with it before opting to act at this level (^{Föhrer, 2015}).

A first level of constraints refers to union practices and routines. The level of inventiveness required here is often difficult for local unions to mobilize since most of them, especially in the South, are not always properly equipped to take their struggles to the transnational level due to a critical lack of resources (financial and human), whereas in the North, without generalizing, the temptation to rely on advocacy tools and micro-corporatist strategies, which have been tried and tested in the past, can be strong (^{Vogler, 1985}; ^{Marginson, 2016}). Moreover, deep-rooted traditions of collective representation, a dense relational past and a strong regional identity can also arouse suspicions and lead trade unions to solve their problems locally (^{Silva, 2013a}; ^{Stevis and Boswell, 2008}). ^{Bernaciak (2010)} and ^{Fetzer (2008)} developed interest-based frameworks to explain why some unions may be less inclined to internationalize their action, especially if they have access to options for acting locally and national solutions no longer appear viable, restraining the possibility for them to act and establish a shared sense of we-ness at the transnational level. Yet, participants need to have the ability to adhere to the importance of transcending the diversity of their practices, their possible rivalries and their socio-economic realities for the sake of “making meaning” (^{Tarrow, 2011}).

A second level of constraints relates to the employer as well as to the structural factors (^{Wright, 2000}) and existing competitive dynamics in the mining sector. On the one hand, the instability of the corporate actor, the fluidity of its industrial capacities, the complexity of the inner structure of multinational companies and their ever-changing nature appear to severely erode the capacity of trade unions to develop collaborative ties on a transnational basis. On the other hand, in the mining industry, a producer-driven and particularly volatile sector (^{Gereffi, 2001}), unions could possibly be tempted to renew their repertoires of action through local and national strategies or fall back on microcorporate strategies. These options raise the possibility for them to exploit the advantageous balance of power afforded by their position within the value chain of their home multinational company, even though spatially fixed mining operations often make the use of coercive comparisons, or threats to offshore production, fruitless. The outflow of jobs to the South, successive merger and acquisition operations and virulent anti-union strategies have disrupted the development and the sustainability of cross-border alliances in recent years. By constantly threatening to suspend their investments and withdraw their infrastructure, such corporations force workers to compete among themselves and suspend any effort to transnationalize their action. Furthermore, this instability also appears to negatively affect the ability of union representatives in the North, but even more so in the South, to convince their members that alliances are relevant and to ensure that their involvement at the transnational level will not end up costing them their positions (^{Anner et al., 2006}; ^{Anner, 2011}).

A third level of constraints, which we will not dwell on, relates to the institutional level. Cross-border alliances face complex challenges, especially since they seek to align heterogeneous labor relations regimes, union traditions and socio-economic realities (^{Croucher and Cotton, 2009}; ^{Vogler, 1985}). Furthermore, the limited effectiveness and the hostility of institutional channels, which are particularly hostile in Southern countries where union activists can expect day-to-day reprisals for their activism, can also undermine the feasibility of transnational action (^{Silva, 2013a}). Table 3 in Appendix 3 enables comprehension of the extent of the differences between Canada, Chile and Peru identified on the institutional level in regard to the framework for collective labor relations; such challenges certainly complicate the possibility of creating interunion ties transnationally. Finally, in addition to this are the linguistic, cultural and geographical distances separating them and the coordination costs that establishing such relations can entail for actors.

The North-South Alliance at Stake

Contingencies Shaping the Collective Identity Framing Processes

Our study establishes that organizational identities are a fundamental element in understanding how transnational trade union alliances function and develop. Organizational identities are conceptualized here as social constructions that are dynamic, that are defined by their situation in time and space and that evolve through exchanges between the actors. As true strategic compasses (^{Dufour and Hege, 2010}; ²⁰¹³; ^{Dufour-Poirier, 2011}; ²⁰¹⁶; ^{Dufour-Poirier and Hennebert, 2015}; ^{Hunt and Benford, 2012}; ^{Melucci, 1995}), they determine how actors define problems, consider solving them and identify sufficiently unifying incentives to pool their efforts in a transnational alliance. On the one hand, the results emerging from our analyses confirm that a key to the success of transnational alliances lies in the actors’ ability to develop, upstream and based on their exchanges and reflections, a community of risks and fate (Greer and Hauptmeier, 2012), in short, a transnational trade union culture (^{Vogler, 1985}; ^{Kay, 2011}) built around a strong sense of belonging to the group (we-feeling) (^{Föhrer, 2015}), in opposition to opponents and in the name of clearly defined project and scale (^{Tattersall, 2010}; ²⁰¹⁸). On the other hand, this study highlights the plastic, relational and situational nature of the collective identity framing processes underlying the decision by a union to participate in a transnational alliance and adhere to the ideals of solidarity it conveys (^{Dufour-Poirier, 2011}; ²⁰¹⁶). These organizational identities, generated by the actors based on their exchanges and experiences, influenced their choices regarding union solidarity.

However, three main levels of contingencies shaped the alliance's internal dynamics and development. The first level of contingencies relates to the dynamics that punctuated the internal life of the unions participating in the alliance, especially as regards the allocation of resources (budgetary, human, logistical, etc.) relating to alliance building, particularly in the North (^{Marginson, 2016}). We refer to this as the dual process of centralization-decentralization. In our case, this opposition refers, on the one hand, to the gradual reduction of resources invested in the project by the Industrial Union in Canada and, on the other hand, to the progressive centralization of the work carried out by its decision-making structure over the years. This process prevented the local trade unionists from participating in the discussions undertaken by the alliance, which would have helped them better identify with it and grasp the relevance of the cross-national understandings, synergies and strategies it intended to develop: these considerations confirm Sarkar and Kuruvilla’s work (2020) that states the importance of allowing local concerns to find voices in global campaigns in order for them to result in concrete gains at the local level and to be considered relevant. In our case, a number of the local leaders were unable to maintain the contacts they had sought to establish, which decreased their motivation to continue participating in the alliance.

As noted by ^{Spalding (2013)}, “differentiating local actors, [national] and transnational ones as well as delineating their respective roles can be particularly challenging”(p. 26), in any alliance-building effort. In addition, the retirement of the international project manager and the failure to replace him at the Industrial Union, as well as the non-re-election of several of the local trade unionists most heavily involved in the project, also weakened their ability to continue negotiating their differences, to frame the relevance of those contacts and to strategize around the collective cohesion developed over time. Another very critical aspect was the lack of multiple contact nodes of coordination (^{Diani, 2019}), which could have helped the too-soon retired and never-replaced international project manager at the Industrial Union ensure the development of the alliance. This “branded knowledge broker” (^{Agarwala, 2014}) acted within the alliance as an “imaginer” (^{Spalding, 2013}), meaning that he was able to prompt all the trade union representatives involved to connect with one another, to facilitate the circulation of new ideas, and to articulate connections across differences and levels, both horizontally and vertically (^{Routledge and Cumbers, 2009}). These roles were in addition to those he played as a translator (responsible for diffusing knowledge), coordinator (in charge of organizing the distribution of resources and responsibilities) and articulator (able to bring together actors) (^{Von Bülow, 2013}), facilitating discussions. These roles also eased tensions among the participants and united them around sufficiently cross-cutting and mobilizing identities to overcome their distinctive features. Such key individuals can act as the glue for an alliance to adopt horizontal and vertical operational logics at the same time: their absence here was found to be critical (^{Routledge and Cumbers, 2009}) and did not prevent the alliance from becoming overtly unsuccessful over time.

These findings also critically reiterate the importance for umbrella organizations in the North, in this case the Industrial Union in Canada, to ensure follow-up and cooperation and to invest a minimum of indispensable resources, in particular human resources, in a spirit of labor empowerment rather than the philanthropic labor spirit traditionally associated with the protectionist policies of Northern labor (^{Agarwala, 2014}) or limited corporate internationalism (^{Vogler, 1985}). Moreover, such issues corroborate Brookes and McCallum’s (2017), and Sarkar and Kuruvilla’s (2020) findings that belonging to multinational companies does not constitute a sufficiently unifying bond for establishing minimum working relationships. In our case, facing a common and highly unstable enemy did not represent a sufficiently clear project of resistance (^{Hochstetler et al., 2013}) to lead beyond the search for the lowest common denominator.

Based on Diani’s work (2019) on networks and also on that of ^{Sarkar and Kuruvilla (2020)}, these issues substantiate the need to multiply mediating and coordinating contact nodes, decentralize the decision-making process, and enable all the actors involved to launch initiatives, strategize around the incentives established through their exchanges and nourish, from a multi-level perspective, the internal dynamics of the alliance. We are referring here to the concept of “centralized decentralization” developed by ^{Gallin (1994)}, which emphasizes the need for a back-and-forth dynamic between the interventions conducted by the top tier of the union hierarchy, at the national and transnational levels, and those conducted by the local activist base to ensure the efficiency and sustainability of all kinds of in-breeding initiatives. To give rise to increasing trade union counter-coordination as well as new fields and foci of action (^{Lillie and Martinez Lucio, 2012}; ^{Sarkar and Kuruvilla, 2020}), alliances need to fuel themselves with multiple poles of upward, downward and lateral influences (^{Routledge and Cumbers, 2009}), actions and knowledge, across different levels (^{Marginson and Sisson, 2004}). Maximizing the depth and fluidity of these connections is the only way to provide union alliances, such as the one under study, with vital “shared meanings” (^{Tarrow, 2011}) or a “common agenda” (^{Tattersall, 2010}; ²⁰¹⁸), to deepen knowledge-sharing processes and to allow the associational power (^{Wright, 2000}) they underlie to survive and grow, despite the contingencies they face. More generally, such changes call for a profound decompartmentalization of union work and the inherent and harmful divisions that still exist therein.

A second level of contingencies now relates to the very rapid evolution of the employer's nebulous nature. In this case, the optimal version of the studied alliance would have consisted of keeping its progress in step with the changes taking place within the transnational employer and matching the latter’s spatial scope (^{Marginson, 2016}; ^{McCallum, 2013}). Moreover, the instability of the corporation and the increasing difficulty, for all participants, including the Canadian Industrial Union, in accessing the corporation’s decision-making center made it more complicated for the participants to develop joint action strategies. Participating in the alliance ended up being seen by some members and local union representatives, who could not take part in the exchanges, as a privilege, a kind of union tourism, rather than a necessary contribution to the building of an overarching community of coping. Following his non-re-election, one Canadian local unionist decried the fact that “it was always the same people who participated in the trips and interacted among themselves [making it difficult to transfer knowledge and ensure the next generation’s involvement in the project and institutionalize its survival]” (Interview no. 45, personal communication, 2007, November 11^th). Without this legitimation, the project ended up costing the local leaders most heavily involved in the alliance their positions, simply because they were not able to justify their involvement to their members and faced highly repressive tactics deployed by their local corporative counterparts, such as through the proliferation of yellow unions or new competing union teams at their workplaces. In Peru, such upheavals meant that relationships of trust that had taken years to forge had to be rebuilt. The process was quite threatening for local unionists, who were only elected for a 2-year mandate. This explains, in part, why trade unions in Latin America retreated on a subregional basis, making it easier to develop common strategies. Following the alliance's retreat, the Southern unions decided to continue exchanging information among themselves. A similar withdrawal was also observed in Quebec, where the local affiliates of the Industrial Union returned to their initial routines. The leaders of the Industrial Union were also severely criticized for their lack of leadership within the alliance and their inability over the years to deliver concrete local results on the ground. Such subregional withdrawals confirm the heterogeneity of the North and South, and the non-existence of what one might be tempted to call the Global North or Global South (^{Lindell, 2009}; ^{Silva, 2013a}).

These volatile positions reaffirm the need to multiply mediating and coordinating nodes and/or bodies, such as, but not solely, the Canadian Industrial Union, and to install disciplined dialogue when interests diverge among participants in order to maximize the “actionability” of the associational power at play (^{McCallum, 2013}). Nevertheless, like ^{Marginson (2016)}, we do not support the idea that such alliances necessarily gain from developing within highly structured institutional, hierarchical and vertical frameworks on a transnational basis, which is anyway not the case in the Americas. Those institutional solutions do not necessarily guarantee the future of this type of initiative, much less the decision of actors to espouse it (^{Bieler and Erne, 2015}). Global Union Federations are certainly in a better position to pick up the threads of the frayed networks and loosely coordinate them. By extension, our study also implicitly questions the viability of initiatives that are supposed to apply models of union efficiency in the North to the realities of the South, without validating them beforehand in democratic terms (^{Silva, 2013a}).

Lastly, these challenges bring out a third level of contingencies, this time linked to the need to act across scales (^{Hochstetler et al., 2013}; ^{Lindell, 2009}) with the local and global being interconnected, mutually co-constructing rather than opposing one another, as is nevertheless often the case. The tensions emerging from these contingencies can be seen in the difficulty, for the trade unions involved, in bringing out strong connections between the actions pursued at the local level and the global demands put forward by the alliance, without, however, overlooking the local dynamics. In our case, conciliating local and global priorities turned out to be all the more contentious for the unions involved since the alliance had, over the years, brought together a growing diversity of inevitably heterogeneous actors. Indirectly, this local-global distancing appeared to have undermined the trust capital and associational power (^{Wright, 2000}) that had been tentatively created within the alliance. In fact, it intensified the dissimilar paths and asymmetries of power and resources between the North and South (^{Spalding, 2013}), rather than creating unifying bonds around common strategies. The accounts collected herein confirm that, although capitalist globalization appears to be calling for trade unions to act transnationally, it should be understood that any transnational alliance must continue to focus on its local bases, without which it will have to take on substance, legitimacy and justification with regard to its members, to whom it is accountable (^{Dufour and Hege, 2010}; ²⁰¹³; ^{Fox-Hodess, 2020}). This result shows that trade union participation at the transnational level remains a deeply relational phenomenon, not only grounded in the exchanges between actors, but also in the local and national realities within which they are situated (^{Silva, 2013b}). It also reiterates the need to keep local foundations at the center of any alliance-building process, especially for the trade unionists promoting it.

Conclusion

Though this study, our aim was to gain insight into how and under what conditions a North-South trade union alliance could bring together a diversity of actors. Within this framework, three levels of constraints were identified, complexifying the opportunities for a heterogeneous group of actors to work together. At the sectorial level, the possibilities of making the alliance work and sustainable were hindered by the employer's instability over the years. Within that context, the alliance participants, despite their a priori enviable spatially fixed location within the production chain (^{Wright, 2000}), found it increasingly difficult to identify who their common employer was. The absence of such identity markers had the effect of ossifying the internal dynamics of the alliance under study. The situation led the employer to make full use of coercive comparisons, forcing the production sites to compete with one another, and made the North-South rapprochement more complex as well as fueled the fears sown by the anti-union repression underway in Chile and Peru. These shortcomings consequently prompted the participating actors to withdraw subregionally in order to achieve concrete results and strengthen trade union action through the sharing of information and expertise.

These findings apply to both the North and South and provide important conclusions on two levels. Identity markers continue to be the basis for any evaluation and calculation of the costs and benefits of collective action, prior to engaging in it (^{Hunt and Benford, 2012}; ^{Melucci, 1995}). Trade union transnationalization must take into account, first and foremost, what trade unionism is, i.e., a process of defending local interests and identities. In that sense, the strength of the alliance, and the trade union participation in it, lies in the solidity of its local foundations. Furthermore, in the absence of legitimacy for the transnational alliance, such a strategy becomes a threat to the legitimacy of actors on the ground at the local level (^{Dufour and Hege, 2010}; ²⁰¹³; ^{Fox-Hodess, 2020}). These findings run counter to the argument that the global level is always necessary for creating a base at the local level as long as the latter is part of a globalized economic strategy (^{Silva, 2013a}). It is only when trade union representation is strong at the local level that more distant strategic alliances can be supported by strong local representatives. If the latter are weak locally, their contacts at the transnational level will always be seen as privileges.

Similarly, trade unions’ volatile positions with regard to transnational solidarity demonstrate that such alliances suffer from a great paradox, since the local union leaders involved are forced to make trade-offs between time spent fostering transnational work and time spent on their own work, leaving themselves vulnerable. This finding reveals that a powerful determinant of a union’s capacity to adhere to transnational solidarity and act on that level is linked to a deeply entrenched national leadership and its desire to invest the time and resources needed to legitimize such a file among its affiliates. As such, the North is all the more called on to act as guarantor to the South (^{Gallin, 1994}), which is more fragile in terms of resources, and owing to the virulent anti-union repression that often bears down on representatives and activists. In the same vein, only local leaders with a strong base are in a position to invest time and energy in transnational work, whether that involves spending their own time or shifting resources to it. This is particularly true for leaders who are striving to achieve longer-term goals or launch initiatives that will be slow-burning in terms of the material impact on their members, as opposed to leaders whose unions may see more immediate benefits from transnational solidarity. Arguably, in this case Chilean and Peruvian local and national unions, federations and confederations had much to gain immediately, notably in terms of resources and legitimacy, although as the alliance went into decline, even those organizations had little to gain, hence the deep frustration among their leaders, who complained about their selfish Canadian counterparts.

Last, but not least, this study confirms that the actors’ origin, that is, whether they came from the North or South, did not overwhelmingly influence their ability to frame the relevance of participating in a global solidarity process. A trade union’s decision to overhaul its repertoire of action at the transnational level and become a relevant “broker of ideas” (^{Tarrow, 2011}) appeared to depend on its ability to inhabit the new space. This finding not only dissolves the analytical relevance of the North-South divide (^{Hochstetler et al., 2013}), but also emphasizes that trade union sensibilities regarding solidarity are relationally and contextually anchored (^{Morgan and Pulignano, 2020}). From that perspective, such sensibilities need to be examined on a case-by-case basis in order to be fully understood. Similarly, the expressions Global North and Global South appear to us to be overused conceptually. Indeed, they seem fundamentally incorrect, given the plasticity of the union responses in terms of transnational solidarity, which are strongly rooted in space, time and the contingencies of the moment, as well as dependent on the links taking place in the heart of activist work in the field. Such features and findings open up an almost limitless field of possibilities for future research and validation in both the North and South, so as to better comprehend how the moral and political solidarity sought transnationally through union alliances are grounded in the concrete and the real.