Practicing Intersectionality: Brazilian Domestic Workers’ Strategies of Building Alliances and Mobilizing Identity

Domestic workers, as essentially black women in highly precarious conditions, lacking social protection, and often not recognized as proper workers, are considered particularly hard, or even impossible, to organize compared to the “standard” male industrial worker. Yet they have been mobilizing for decades in several countries, including Brazil, where they won new labor rights in 2015. Through an ethnographic study of domestic workers’ unions in Brazil, I argue that what has made their mobilization possible is precisely the intersectional dimension of their oppression. While gender, race, and class have produced multiple forms of exclusion, these vectors of oppression have also enabled domestic workers to build alliances with women’s, black, and workers’ movements, thus giving domestic workers more visibility and more resources to organize their members. I have identified three forms of alliance building, each determined by domestic workers’ unions’ particular framing of gender, race, and class issues: rigid autonomy, critical alliance, and encompassing unionism.


Introduction
Domestic workers' struggle in Brazil has to do with gender, race and class. There is still a remnant, or a historical process, from the movement against slavery. We, black women and black men, we were trafficked from Africa and brought here to Brazil. The majority worked in the plantations, the others worked at the master's house. -Creuza Maria de Oliveira, August 2015 In this quotation, Creuza Maria de Oliveira, leader of the National Federation of Domestic Workers (FENATRAD) in Brazil, explicitly connects domestic workers' intersectional oppression to colonial history. Indeed, more than a century after the abolition of slavery, black women are still the main providers of domestic work. In 2013, paid domestic work, which includes any personal service provided to a private household such as cleaning, cooking, and caring for children or the elderly, employed 14 percent of all economically active women and 22 percent of all economically active black women, being the largest sector of employment for the latter. Of the estimated 6.4 million domestic workers, 93 percent are women and 61 percent are black women. Domestic workers earn on average 60 percent less than other workers, and only a third of them are formalized (DIEESE 2013). 1 The perpetuation of the racial and gendered division of labor inherited from the colonial past, which I will call the coloniality of labor (Gutiérrez-Rodríguez 2014), has led some Brazilian scholars to describe contemporary domestic work as a legacy of slavery (Ávila 2016;Bernardino-Costa 2015;Santana Pinho and Silva 2010). I argue that this specific intersection of gender, race, and class oppressions within the (post) colonial and capitalist economy has produced domestic workers as an underclass of servants. Indeed, it is because they are poor black women descendants of slaves that domestic workers were not recognized as proper workers until the approval of the 2013 Constitutional Amendment (PEC 72/2013), and its decrees of application (Law 150/2015), which declares them equal to other workers. This means that until 2013, domestic workers were not deemed worthy of having equal labor rights. In addition, the very denomination of the professional category is a site of struggle. The most common term used in Brazil, and reproduced in the legislation, is "domestic employee" (empregada doméstica), which would be better translated as "maid." In contrast, the unions prefer "domestic workers" to insist on the professional dimension of their activity. In respect for their self-denomination, I will use the term "domestic workers" throughout this article.
As a result of their marginalized position, domestic workers are deemed as lacking political consciousness (Britto da Motta 1999) and "notoriously hard to organise" (Cox 2006, 125). They are informal, work isolated in private homes, lack access to legal protection, and remain underrepresented or simply absent from established workers' organizations. In fact, industrial relations scholarship quite unanimously predicts that "nonstandard" workers are harder to organize than "standard" industrial (male) workers (Anner 2007;Bonner and Spooner 2011;Vandaele and Leschke 2010). Indeed, there is no shop floor to reach and organize them against a common employer. Domestic workers' activity is often not recognized as valuable work, and their multiple vectors of oppression make it harder to create a collective identity within the sector (Kabeer, Sudarshan, and Milward 2013).
Yet domestic workers do organize, both at the national and transnational level, and have created their own labor unions in several countries (Ally 2009;Bonner 2010;Fish 2017). In Latin America, domestic workers have been mobilizing for decades (Blofield 2009;Chaney and Castro 1989), and, in 1988, they founded the Confederation of Domestic Workers of Latin America and the Caribbean, which is now the oldest and more structured regional network of domestic workers in the world (Pape 2016). In Brazil, the largest employer of domestic workers in the region, they are organized into one of the only national federations of domestic workers worldwide. As a result of their decades-long mobilizations, Brazilian domestic workers gained recognition in 2015 with Law 150/2015, which extends most basic labor rights to this professional category, while at the international level, the International Labour Organization (ILO) adopted Convention 189 in 2011 guaranteeing domestic workers the right to decent work (ratified by Brazil in 2018). Thus, despite cumulating multiple forms of exclusion, Brazilian domestic workers successfully mobilized and gained recognition from the state.
Drawing on ethnographic study conducted between 2015 and 2017 with six unions affiliated to the National Federation of Domestic Workers (FENATRAD) in Brazil, I argue that what has made domestic workers' mobilization possible is precisely the intersectional dimension of their oppression. While gender, race, and class have produced their marginalized status, these vectors of oppression have also enabled them to build alliances with women's, black, and workers' movements, thereby gaining more resources and visibility. I understand intersectionality as both a critical inquiry and a political praxis (Hill Collins and Bilge 2016) that can sustain domestic workers' collective action. More than an analysis of their multiple vectors of oppression, intersectionality is what enables domestic workers to contest and transform this oppression by building a collective identity and political alliances.
Thus, I aim to discuss how nonstandard workers can practice intersectionality and demonstrate how this praxis relates to their organizational success. More specifically, I find that unions' ability-or willingness-to form alliances is informed by how intersectional they are in practice. The more social identities are incorporated into their analysis and discourses, the broader their alliances are, and the more successful the unions can be. Therefore, an intersectional praxis not only can support movements' coalitions (Bernardino-Costa 2015; Nadasen 2015; Tilly, Rojas-García, and Theodore 2019), but I further argue that this praxis also explains domestic workers unions' successes and failures. I have identified three different modes of alliance building: rigid autonomy, in which a union refuses sustained partnerships with other organizations and is consequently weaker than the other unions; the critical alliance, which is a noninstitutionalized and fluid form of alliance with other movements, giving unions only a limited success; and encompassing unionism, wherein the union is conceived as a broad platform for organizing black women and as a result is the most successful of the six unions observed.

Toward an Intersectional Analysis of Social Movements
Coined by the black feminist scholar Kimberlé Crenshaw (1989), the term "intersectionality" has been developed to capture the struggles of black women and to theorize the particular forms of oppression they face. Black feminists have argued that their experience of oppression cannot be understood in terms of either being black or being women separately; instead, it results from interlocking systems of oppression (Combahee River Collective 1986;Davis 1982). Using more specifically the definition of Patricia Hill Collins (2007), I understand intersectionality as an analysis of matrices of oppression. In this framework, social identities are not only individual or subjective elements but are the product of structural inequalities that place groups of people into certain positions of power or oppression.
I contend that an intersectional lens is indispensable to understanding the situation of domestic workers; indeed, studies quite unanimously present them as cumulating all the vectors of oppression, both nationally and on a global scale (Anderson 2000;Glenn 1992;ILO 2013;Pérez and Llanos 2017). In Brazil, it is the specific combination of gender, race, and class oppressions in the (post)colonial and capitalist economy that has constituted domestic workers as an underclass of servants, excluded from labor laws for decades. Their extreme vulnerability also exposes them to sexual assault and violence in the workplace, revealing the depth of gender and racial oppression in Brazil (Ribeiro Corossacz 2019). As black Brazilian feminists argue (Carneiro 2018;Gonzalez 1984), domestic work continues to be perceived as the "natural" place of black women, thereby perpetuating the legacy of slavery and justifying their lower social status. This position of marginality further reflects the devaluation of care work in general, framed as a natural feminine task rather than an actual valuable work, which translates into lower wages, lack of social recognition and adequate labor regulations for care workers (Hirata and Guimarães 2012;Kowalchuk 2017;Parreñas 2001).
In this context, intersectionality provides an analytical tool to understand and challenge structural inequalities. Indeed, domestic workers are placed into a particular position of marginality grounded on their gender, race, and class oppression, but they can also contest this situation through collective action. As Hill Collins and Bilge (2016) suggest, intersectionality is both a critical inquiry into power and inequality and a praxis deployed by individuals and movements; it allows us to simultaneously analyze the formation of marginalized social identities and the ways in which actors contest those. By recognizing and naming their common experience of oppression, domestic workers can turn them into a driver for collective action; they use this experience to build a collective identity as well as alliances with other movements. In fact, domestic workers' mobilization for equal rights challenges gendered and raced hierarchies by affirming the value of black women's reproductive labor.
Within social movements and industrial relations scholarship, the inclusion of multiple social identities is usually included under the concepts of "community unionism" or "social movement unionism," which emphasize the links between the workplace and other dimensions of workers' lives (McBride and Greenwood 2009;Fairbrother 2008;Murray 2017). This scholarship looks at how established labor unions, as part of a revitalization strategy, try to organize the so-called nonstandard workers by activating their social identities and networks beyond the workplace. Typically, in the UK and US contexts, studies show how conventional unions form alliances with migrant and community-based associations in order to organize precarious migrant workers (Alberti, Holgate, and Tapia 2013;Engeman 2015;Yu 2014). However, quite problematically, the literature on the organization of nonstandard workers-women, migrant, informal, low-paid, and racialized workers-tend to assume that precarity is a new phenomenon, associated with contemporary neoliberal reforms that would pose new challenges to labor unions (Standing 2014;Milkman and Ott 2014;Kalleberg 2009). Furthermore, focusing on the actions of established (declining) unions to organize the unorganizable erases these workers' agency. What do nonstandard, or intersectional, workers think and do about their condition of exploitation?
In Western-centered case studies, precarious work is opposed to standard work, usually understood as industrial, male-dominated, permanent work. Yet standard work is not the standard everywhere, and in Latin American, as for an important segment of the population worldwide, precarity has always been the norm (Mosoetsa, Stillerman, and Tilly 2016;Braga 2016). In particular, if we look at female-dominated activities, such as domestic work, insecurity and low wages are the standard. Domestic work is embedded in a historical gendered and racial division of labor, which enables some to access good secure jobs while outsourcing the burden of social reproduction to the precarious others (Duffy 2007;Glenn 1992;Ehrenreich and Hochschild 2003). The very existence of protected male industrial jobs has only been possible through such a division of labor whereby some were protected while others were precarious or unpaid (Fraser 2016).
Thus, conventional approaches to nonstandard workers' organizing cannot fully account for domestic workers' mobilization in Brazil; these approaches seem to neglect the coloniality of labor intrinsic to the sector and the intersectional dimension of oppression that has made it a marginalized group. In fact, nonunionized domestic workers tend to reject the very denomination "domestic worker," which is socially stigmatized and devalued on the labor market (Acciari 2016). As a result, their worker identity is not necessarily the most relevant compared to their experience as women, black, or poor. Union leaders are then faced with the difficult task of addressing domestic workers' multiple forms of oppression while also creating a common identity they can relate to.
In addition, for domestic workers, the workplace is also a private home, which challenges the assumed boundaries between workplace and community, public and private spaces. As a result, their modes of organizing have always included the formation of alliances with other social movements and the recruitment of their members within their communities (Bernardino-Costa 2015; Cornwall, Oliveira, and Gonçalves 2013;Nadasen 2015). What seems to be a renewal strategy from established unions in the global North to reach out to precarious migrant workers (Birdsell Bauer and Cranford 2017; Lazar 2017) has been a long-standing strategy in the sector of domestic work. For marginalized workers, class and other vectors of oppression cannot be separated, rendering the distinction between the workplace and the community artificial (Alberti 2016).
Feminist scholars have shown that an intersectional approach can be a tool to uncover points of similarities between different groups of women, thereby facilitating the formation of coalitions (Chun, Lipsitz, and Shin 2013;Cole 2008;Carastathis 2013). According to Chun Lipsitz, and Shin (2013), illuminating common forms of oppression can enable women from different backgrounds to organize collectively against those. Similarly, Nadasen (2015) and Bernardino-Costa (2015) have demonstrated that domestic workers in the United States and Brazil, respectively, were able to form alliances with other movements by strategically mobilizing their multiple identities as women, black, and workers. For instance, the framing of domestic work as a legacy of slavery has helped in both cases to legitimize the place of domestic workers' organizations within black movements.
Drawing on these insights, I propose to evaluate more systematically the link between intersectionality (as a theory and praxis) and mobilizing strategies. I argue that the more intersectional unions are in their discourses and practices, the stronger they can be. Intersectional praxis not only gives unions access to diverse allies, but it is also the condition to build unity, create a positive collective identity, and to reach domestic workers in multiple social spaces. Indeed, in the absence of a common workplace to organize domestic workers, and the difficult self-identification with their own professional category, unions have to reach their rank-and-file through intersectional strategies. It is sometimes easier to find a black woman or a poor woman than a doméstica. Thus, I show that the unions that have the ability to mobilize domestic workers' multiple social identities are relatively more successful. I assess the relative success of domestic workers' unions on the basis of their membership rate, their "network embeddedness" (Lévesque and Murray 2010), and their capacity to reach out to the rank and file. The more successful unions are those that have a higher number of members, more participation in their meetings, more visibility within both the rank and file and with external stakeholders, and more established alliances with other movements. By comparing the strategies of six different local domestic workers' unions in Brazil, I further aim to explain the variations and tensions in the process of forming alliances. Although I argue that an intersectional praxis can be a factor of success, I also show the challenges and difficulties of such a praxis.

Methods
This article draws on fieldwork conducted with six local unions affiliated to FENATRAD in the cities of São Paulo, Franca, Campinas, Rio de Janeiro, Volta Redonda, and Nova Iguaçu between 2015 and 2017. Taken together, the states of São Paulo and Rio de Janeiro employ 40 percent of the national contingent of domestic workers (CUT-RJ and Sintell-Rio 2013;DIEESE and FES 2015). I conducted in-depth qualitative interviews with thirty-two leaders from both the local unions mentioned above and the national directorate of FENATRAD (in Salvador and Brasília). I asked them questions on their experience as women, black, and domestic workers; their relationships with other movements; and their explanation for the devaluation of domestic work. I also interviewed forty-five nonunionized domestic workers coming to the unions for a labor dispute (unions are obliged to inform and represent their professional category even though the workers are not members). With them, I discussed their perception of their jobs, their knowledge of and involvement with the unions, and the new legislation. The average age of my sample is fifty-five, ranging from thirty to eighty-five years old, and an absolute majority of 92 percent is nonwhite, while 70 percent self-identified specifically as black.
The interviews are complemented by nine months of participant observation in which I attended unions' daily activities, monthly meetings, and some events organized by their partners. These events provided occasions to confirm the absence or presence of domestic workers from broader coalitions. Lastly, I collected internal documents such as pamphlets, minutes from meetings, and photographs. This helped me map unions' history and the evolution of their discourse. I paid particular attention to local unions' perceptions of gender, race, and class and traced their connections with women's, black, and workers' movements. Through this qualitative analysis, I have identified three modes of alliance building: rigid autonomy, critical alliance, and encompassing unionism.

Domestic Workers' Movement in Brazil
The first association of domestic workers was created in 1936 by the black activist and domestic worker Laudelina de Campos de Mello (1904-1991 in the city of Santos. At the time, labor laws were being passed for standard industrial workers, such as the right to unionize (Law 19.770/1931)-except for domestic workers (art. 11)-or the right to a minimum wage (Constitution of 1934, art. 21). The Labor Code (Consolidação das Leis do Trabalho, or CLT) was adopted in 1943 to systematize workers' rights, while explicitly excluding domestic workers on the grounds that domestic labor is a "noneconomic service" (CLT 1943, art. 7). In this way, domestic work was not recognized as proper work; consequently, domestic workers did not qualify for labor rights like regular workers. This exclusion from the Labor Code must be understood as part of the coloniality of labor (Gutiérrez-Rodríguez 2014) that is co-constitutive of the Brazilian state; indeed, as white male industrial workers brought from Europe to "whiten" the population gained labor rights, black women descendants of slaves were disqualified as nonworkers.
Local associations of domestic workers gained force in the 1960s with the support of the Youth Catholic Workers, a branch of the church linked to the theology of liberation, which provided them with financial, logistical, and human support and helped them survive through the military regime. In fact, the first law giving domestic workers the right to have their CTPS signed was passed in 1972 (Law 5.859/1972). But it was only with the return of democracy in 1988 that they were given the right to unionize, even though they remained excluded from the CLT in virtue of the article 7 of the Constitution (changed in 2013). In 1997, local unions gathered into the National Federation of Domestic Workers (Federação Nacional das Trabalhadoras Domésticas, or FENATRAD), and became affiliated with the Unified Central of Workers (Central Única dos Trabalhadores, or CUT), a national union central associated with the Workers' Party (Partido dos Trabalhadores, or PT). This close relationship with the PT enabled FENATRAD to negotiate directly with the left-wing government between 2003 and 2013 and to press their ally for a new legislation.
In 2013, Dilma Rousseff's government approved Constitutional Amendment 72 (PEC 72/2013) which rectifies article 7 of the Constitution and declares domestic workers equal in rights to other workers. The "PEC das domésticas" was then adapted into Law 150/2015 to regulate the constitutional changes and specify which rights domestic workers are entitled to. These include the obligation for employers to sign the CTPS, the national minimum wage, remuneration of night work, an eight-hour working day, compensation for extra time, weekly rest, thirty days of paid annual leave, compensation for unfair dismissal, unemployment benefits, pension, and the recognition of collective agreements. However, Law 150/2015 still provides a list of rights inferior to that of other workers and maintains domestic workers' status of exception within the CLT. As a result, domestic workers' unions do not enjoy the same rights and benefits as unions in the other sectors; for instance, they do not receive the union tax, and their elected officials are volunteers. 2 Even though they are recognized unions, they remain underresourced, have no paid staff, and display lower affiliation rates compared to other sectors-only 2 percent nationally (Costa Furno 2016). In these circumstances, building partnership proves quite crucial to their material survival.

Mobilizing Gender, Race, and Class to Form Alliances
At the national level, FENATRAD has implemented quite an extensive strategy of alliance building. In addition to being part of the CUT, the federation receives support from well-established groups such as the Feminist Centre for Research and Advice (Centro Feminista de Estudos e Assessoria), the Unified Black Movement (Movimento Negro Unificado), and international organizations such as UN Women and the ILO. Creuza, cited in the introduction, developed a strategy of alliances corresponding to her vision of domestic work. The location of domestic work at the intersection of gender, race, and class oppression and its strong historical ties with antislavery movements form the basis of alliances with feminist, black, and workers' movements. This expansive strategy of alliance made FENATRAD particularly visible during the PT government, which was more attentive (at least in in its discourse) to the poor. Domestic workers' ability to frame their demands simultaneously as a question of workers', women's, and black people's rights gave them some leverage to influence policy making (Acciari 2019). However, I observed diverging strategies among local affiliates in the states of São Paulo and Rio de Janeiro, revealing the challenges of building an intersectional praxis.

Rigid autonomy
The first mode of alliance building discussed here is the refusal to build alliances, which I call rigid autonomy. This strategy is implemented by the unions of the cities of São Paulo and Franca, both in the state of São Paulo, which have made the choice to remain autonomous from other movements at the risk of being isolated. 3 They receive very limited financial or human support from other organizations and do not take part in broader coalitions. This strategy results from their understanding of domestic workers' oppression as being primarily a class-based issue, and their decision to emphasize almost exclusively labor concerns. These two unions focus mostly on individual and legal casework for domestic workers, and at the time of my fieldwork, they did not have a clear strategy of base mobilizing.
The local association of São Paulo was founded in 1962 by a group of empregadas who were part of the Young Catholic Workers and became a union in 1988. Franca's association was also created by a progressive branch of the Catholic Church in 1986 and became a formal union in 1996. Both unions are quite isolated and have developed distrust toward the CUT. Maria Lima, the president of São Paulo's union, does not want to get involved with the CUT because, she says, "they only take advantage of us. They only invite us to their own thing and don't take any interest in us." 4 Similarly, Rosa da Motta Jesus, the president of Franca's union, mentioned a conflict she had with a member of the PT who tried to set up his own domestic workers' union, bypassing her, in order to use this structure as a fundraiser for his political campaign. 5 Nonetheless, the unions' connection with progressive religious groups has enabled the articulation of a strong class identity, justifying their affiliation to the CUT. A document written by the former president of the union of São Paulo states that their main purpose has always been the inclusion of domestic workers within labor laws: "Domestic worker is a profession as worthy as any other, therefore, what we want is to reach full equality of rights." 6 The current leadership confirmed this position; class is the most salient element in their discourses and forms the basis for their collective action. They explain discrimination against domestic workers as being essentially based on their lower social status and prioritize a discussion on labor rights at every monthly meeting. In this sense, being part of the CUT and allied to the PT is a way to formally inscribe domestic workers into the Brazilian working class. Posters of domestic workers holding a CTPS are exhibited in the unions' offices, together with a banner from the CUT claiming: "Rights cannot be reduced, they can only be amplified." This current strategy of isolation and single-issue focus in the union of São Paulo is distinct from that implemented in the 1990s, when the union seemed more integrated within other local movements. An information bulletin from 1999 mentions the union's participation in a black women's seminar and a local church led by a black priest. 7 However, under the current leadership, the union does not have any established partnership with black organizations and does not consistently integrate race into its discourses on labor and value. Bete, a leader of São Paulo's union, claims that "the feminists and the black movements don't come find the domésticas. They don't really care about us." 8 She further affirmed that when she started volunteering for the union, domestic workers would look down on her and refuse to take information from her because she is black, even though they were black women themselves. Thus, even though Bete recognizes the manifestations of structural racism, the identities black or black woman seem to be carrying the weight of social stigma rather than representing a driver for collective action.
An ambivalent dynamic also occurs with gender. Some aspects of gender oppression are recognized indirectly, but gender does not form part of local leaders' discourse on labor and value. Being a woman is never really acknowledged or discussed; nonetheless, Silvia, the general secretary, implicitly recognized domestic work as a women's job: "You see, nowadays, women work more often outside; they don't have time to do the housework, so they need someone. And who better placed than the doméstica to do this? Therefore, they need one. They go to work, and when they come back home they do not want to clean, cook, take care of the kids, their homework, this kind of thing. They come back home and everything is ready, why? Because there is a doméstica doing it all." 9 Here, Silvia seems to make a difference between women and domésticas. Class mediates her experience of being a woman or her idea of what being a woman means: some women have a job outside the house and can hire an empregada, while others, the poor, are the empregadas themselves. This illustrates perfectly the argument made by black feminists that not all women face the same type of oppression (Carneiro 2018;Crenshaw 1989;Hill Collins 2007), thereby making intersectionality a relevant tool of analysis. However, gender was never explicitly acknowledged in the leaders' discourses or at the unions' meetings and does not seem to be consistently connected to class or race. When the leaders talked about discrimination, it was always in relation to the law and their rights as workers. In addition, or perhaps because of this lack of identification with the category "women," local leaders have a relation of distrust with feminist movements. Maria sees feminist groups as "messy" and "gossipy" and believes it would be "a waste of time" for her to attend their meetings. Thus, the (partial) lack of identification with the social categories "women" and "black," combined with the impression that these groups do not care about domestic workers, explains the relative disconnection of the unions from other social movements. And even though the unions do identify with the working class, they have an ambivalent relationship with the CUT and the PT, preferring to remain isolated than having to compromise with them.
This relative isolation limits the unions' capacity to recruit members and mobilize domestic workers. São Paulo has only five active members, while Franca has three. Local leaders estimated that they had about fifty paying members in the case of São Paulo, and seventy in the case of Franca. This number is particularly low for a city of the size of São Paulo, which has half a million domestic workers. The unions are quite unknown to their rank and file; indeed, the majority of nonunionized domestic workers I interviewed did not know they had come to a union and thought this place was a public service. They were also unaware of the unions' history and the struggles behind the 2015 legislative victory, which they attributed exclusively to the good will of the PT. Furthermore, the union leaders' narrative on domestic work remains limited to class and labor issues, while the identity "domestic worker" is precisely the one being rejected by the nonunionized. As a result, if the union does not try to mobilize other aspects of domestic workers' social identity, it becomes harder to recruit and organize them.

Critical alliance
The second strategy observed is what I have characterized as the critical alliance. This consists in having a noninstitutionalized relationship with key stakeholders, forming occasional alliances but remaining distant and critical of these partners. This mode of alliance building is implemented by the three unions of the state of Rio de Janeiro in the cities of Rio de Janeiro, Volta Redonda, and Nova Iguaçu. These unions, too, have a strong focus on casework, but they appeared as more inclined to be part of some institutional processes, using their occasional alliances to pressure decision-makers.
The three unions were created by local progressive branches of the Catholic Church, and they all kept a close relationship with religious groups. Nova Iguaçu and Volta Redonda gained access to a space to establish their offices through members of their local church, and most of the leaders joined the domestic workers' movement through the church. The city of Rio de Janeiro used to have a Pastoral da Doméstica, a religious group entirely dedicated to and run by domésticas. 10 Domestic workers could more easily negotiate a day or an evening off with their employer if it was to attend religious meetings. These local groups would also provide a space for training and socialization for domestic workers, progressively turning them into political activists. Pamphlets from the 1970s and 1980s also suggest active solidarity links between the Workers' Catholic Movement and domestic workers' associations in the periphery of Rio de Janeiro.
The unions' strong connection to the revolutionary branch of the church has influenced their discourse on class and their modes of organizing. They use a language inherited from the Ecclesiastical Community Bases (Comunidades Eclesiais de Base, or CEBs) of "base organizing" and "nucleus of consciousness raising." In an information bulletin from 1996, they explain the importance of having "base groups" in each city and region whose role is to inform, raise consciousness, and organize the political struggle of "the class." 11 This class-consciousness oriented their choice to join the CUT and the PT, and the three local unions seem to have a relatively good relationship with the workers' central. Through this partnership, some of the leaders have a seat in local councils (women's rights, elderly people, and health), and all three unions are close to the PT congresswoman Benedita da Silva, a black woman and former domestic worker herself. However, local leaders complained about a certain lack of support from the CUT and protested against the fact that they have to pay an affiliation fee of R$30 (US$6) per month when their resources are so scarce.
Apart from class-based organizations, the three local unions do not have any established partnership with black or women's movements. In 2017, Nova Iguaçu had just started working with the black women's organization Criola but has no formal ties with it, and feminist groups seem mostly absent from unions' everyday struggles. Although Cleide Pinto, the president of Nova Iguaçu, defines herself as a feminist and has a quite pragmatic approach to building partnerships in order to capture financial resources, she remains a relative exception. One day when I was having lunch with Carli Maria dos Santos and Noeli dos Santos at the local union of Rio de Janeiro, they asked me if I was a feminist. I answered that I am in favor of gender equality, and that for me this was feminism. They laughed and replied, "Oh no, that's not feminism! Feminism is a white movement, they do not share domestic workers' problems and do not pay attention to us. They are all employers. While they demonstrate in the streets, who takes care of their houses and children? Now there's even those women who march naked and say they are sluts and that it's cool!" 12 The feminist movement is seen as something belonging to upper-class white women, distant from domestic workers' preoccupations and reality. These critiques have been addressed to white feminists for decades, both in Brazil (see, for instance, Gonzalez 1984) and elsewhere (see, for instance, Lorde 1984), and appeared very distinctly in the case of domestic workers. In addition, domestic workers' religiosity makes some feminist demands, such as the legalization of abortion, quite controversial. The comment from Carli and Noeli regarding the Slut Walk conveys a discrepancy between their moral values and the tactics used by some feminist groups.
Bernardino-Costa (2015) argues that domestic workers' distrust toward the feminist movement comes from the understanding that white feminist women rely on the exploitation of black women's labor to emancipate themselves from housework. A leaflet produced by the local union of Rio de Janeiro in 1995 states: "As long as a woman, to free herself, oppresses another woman, there will be no liberation nor equality between women" (cited in Bernardino-Costa 2015, 125). Class and race mediate domestic workers' experience and understanding of gender relations; domestic workers deny the universality of women's liberation since they are maintained in a lower social position so that richer white women can emancipate themselves. Race is instrumentally mentioned by local leaders to mark a distinction with other groups; the employers, the feminist movement, and some elected politicians. It draws a line of demarcation between different social spaces: the rich and the poor, the domestic workers and those who can employ them.
The local unions of the state of Rio de Janeiro have a quite weak activist base. Nova Iguaçu has three active leaders and twenty-two paying members, while Volta Redonda has three active leaders and only ten paying members. The union of the city of Rio de Janeiro has six active leaders, and although they could not provide me with an exact figure for their paying members, they benefit from the concentration of upperclass employers in the city. As a result, they see more women coming to the union on a daily basis. None of these unions is well known to their rank and file, who think they are going to a public service rather than an activist space (Vidal 2009) and come to the union out of necessity rather than because of interest in the movement. They are, nonetheless, quite well known to other national and international stakeholders, 10 Minutes from the National Coordination of FENATRAD on June 2, 1999, include a short report of activity for each region. In the Southeast section, the union of Rio de Janeiro reports attending activities from the Pastoral da Doméstica. 11 Report from the First Regional Workshop, with the presence of the unions of Rio de Janeiro, Minas Gerais, and Espírito Santo, September 14 and 15, 1996. 12 Group discussion, Carli Maria dos Santos and Noeli dos Santos, May 2016. They are referring here to the Slut Walk.
including the International Domestic Workers' Federation and the ILO, and Noeli is the representative for Brazil within the Confederation of Domestic Workers of Latin America and the Caribbean. These partnerships give them some leverage to be heard in institutional spaces.

Encompassing unionism
The last strategy I identified is what I call encompassing unionism, exemplified by the union of Campinas. There, local leaders see the union as a means to organize black poor women rather than an end in itself. They are not only concerned with domestic workers' labor rights but also with black women's empowerment more broadly. This could be characterized as a form community unionism or social movement unionism-a union reaching out to its rank and file beyond the workplace, including more dimensions than just class and building broader alliances (Holgate 2015;Lazar 2017;Fairbrother 2008). However, in the case of domestic workers, this social movement dimension is not linked to a revitalization strategy-a concept that does not apply here since the unions were always weak compared to other sectors-nor is it a tactic implemented by established unions to reach out to precarious workers. Given the specificities of the sector, domestic workers' unions have always been forced to reach out to domestic workers outside of their workplaces. Furthermore, I chose the term "encompassing" to reflect the intersectional praxis deployed by the leaders: they fully integrate gender, race, and class into their discourses and practices.
The Campinas' union was founded in 1962 by the leader Laudelina de Campos de Mello, who also created the first ever association of domestic workers in 1936. 13 She was an activist of the black movement (Frente Negra Brasileira) and had the support of the Communist Party, which shows how race and class were already connected since the very beginning of the movement. Dona Laudelina is widely known nationally, and Campinas benefits greatly from this history. Campinas is also the place where FENATRAD was created in 1997, and the first national president (Anna Semião) came from this union. Building on the legacy of Laudelina, the union has implemented a strategy of community organizing and insertion in multiple social movements. As explained by Regina Teodoro: "We do base mobilizing, leafleting in communities, in gated communities, every day we leave the union with a pack of leaflets and we go on distributing them. We have a meeting every third Saturday of the month, open to everyone. And all the movements we are part of, we talk about it, we always have a newsletter in our pocket." 14 Thus, there is a clear strategy to be part of other movements, using this multi-positionality to reach domestic workers and gain visibility. The more social spaces leaders are embedded in, the more they can disseminate information about the union and recruit new members. Many leaders are also part of leftleaning religious groups inherited from the CEBs, which are usually connected to the PT and the CUT. Anna Semião, for instance, was involved in her local church and the PT before joining the union, which shows how well embedded the union is with other movements. When remembering her first years in the union, she describes an extremely active organization: "Obviously, they say we, the domestic workers, we are at the forefront of every struggle; we are in the periphery, we are in the communities, we are in the CEBs, in the Movimento dos Sem Terras, in the homeless movement, in the black movement, in the women's movement, in the PT, in the CUT … so we did cover all the spaces." 15 This citation illustrates quite powerfully what intersectionality as a praxis can be (Hill Collins and Bilge 2016); domestic workers' leaders are part of all the other movements that share some of their struggles and social identities. As argued by Chun, Lipsitz, and Shin (2013), intersectionality helps illuminate common forms of oppression between social actors and enables them to organize collectively against these oppressions. In the case of domestic workers, this means that they can find potential allies in almost every other coalition. Regina is part of a black organization called Black Communitarian Zumbi Festival (Festival Comunitário Negro Zumbi), as well as of the Popular Legal Promoters (Promotoras Legais Populares), a feminist group that teaches women in poor areas about their rights, with a specific focus on gender-based violence. For Regina, race and gender are organically connected, and she puts it into practice in her activism. In her perspective, domestic workers are at the same time poor, black, and women, and all these vectors of oppression must be addressed simultaneously. A domestic worker is also a woman who could be facing violence in her home or racist discrimination in her daily life.
The leaders of Campinas have crafted a distinctive black women identity, which became almost a synonym for domestic worker and is used to mobilize and empower their members. Anna systematically links the fact that she is a domestic worker to her blackness and explains exploitation in the sector as being the result of interconnected oppressions. According to her, it is because domestic workers are black women that they are devalued. At the same time, she conceives the black movement and her union as places where these oppressions can be discussed and challenged, thus leading to a process of empowerment: "And I was going to all these meetings, it was amazing! I heard things that I had never heard before! The racial question was always something … let's put it this way, it is part of your daily life but you don't know what the dynamics of racism are…. I was fascinated by the movement, it was so good to see those women empowered, beautiful, wearing different clothes and authentic hair, without any straightening product or wig." 16 Members of the Campinas union would all talk about race quite spontaneously, mentioning their blackness and their engagement with black groups. Race works here both as an analytical tool to explain structural racism and as something to be repossessed and used in the movement. Most interviewees would insist, for instance, on the importance of being proud of their African heritage and reclaiming black beauty. At their monthly meetings, leaders often propose activities (dinâmicas) to value their members and improve their self-esteem, asking them to say something positive about themselves or reminding them of their struggles and trajectory in the movement. By sharing their experiences of oppression and discrimination, individual members become more aware of the collective and structural aspect of inequalities and are more inclined to mobilize in order to change this situation. This shows the potential of an intersectional praxis. While it provides the tools for social actors to analyze and understand the dynamics of their oppression, intersectionality can also lead to certain forms of mobilization and empowerment.
The Campinas union appears quite skilled at using all the different vectors of oppression faced by domestic workers as resources for collective action. The union has developed a diverse range of actions beyond legal casework, including festive events, trainings for members, active participation in other movements, leafleting outside of workplaces, and monthly meetings turned toward self-esteem and empowerment. This strategy opens up the possibility to recruit beyond strictly defined domestic workers' spaces and issues and enables members to turn their vectors of oppression into an organizing strength. This union stands out compared to the others in terms of membership and vitality; it has 180 paying members and over 2,000 affiliates who contribute irregularly. It can also count on a pool of fifteen active leaders, and the union is well known to other social movements, officials of the city hall, and the city public university (UNICAMP). These alliances make the union an unavoidable actor locally and a key actor of the movement nationally.

Conclusion: Intersectionality as Constantly in the Making?
Across the six local unions, the class dimension appears stronger and more articulated than other vectors of oppression, followed by race and last by gender. Gender is the least salient dimension, and this can be explained by tensions with feminist movements and their partial failure at including domestic workers. The black movement has proven to be a better ally, at least in some cities, and has supported the creation of the first association in 1936. Nationally, FENATRAD also insists on the racial dimension of domestic work and creates a direct lineage between domestic work and slavery. If not all women are poor, poverty is distinctly racialized in Brazil, which could explain the easier identification with race than with gender in this sector.
Domestic workers' close ties with the PT led them to transform their local associations into proper trade unions in 1988 and to be affiliated to the largest national central of workers, the CUT. The labor movement has always been the first place from which to ask for rights, and indeed, the main struggle of domestic workers since 1936 has been their inclusion within the CLT. Only equal labor rights will ensure that they are recognized as workers and that their labor is valued. However, interviews with nonunionized domestic workers revealed that they do not easily identify with the professional category; they understand this is a devalued activity, and feel somehow ashamed of their profession (Acciari 2016). Unions are then faced with the task of building pride in being a domestic worker, while also using the other social identities and lived experiences of domestic workers in order to bring them to the movement. As a result, the capacity of local unions to produce a discourse on domestic work that goes beyond class is what leads them to build broader alliances and expands their ability to recruit members.
Each union defines a particular balance between autonomy and access to resources, while developing their own vision of domestic work. Depending on their articulation of gender, race, and class, local unions have a different conception of unionism and what the role of their organization should be. São Paulo and Franca, on the one hand, and Campinas, on the other hand, exemplify two opposite strategies in this regard (Figure 1 illustrates these distinct strategies). While the former conceive the union as an end in itself, only 16 Interview, Anna Semião (Kota). and exclusively turned toward domestic workers and their most pressing labor issues, the latter considers the union as an encompassing platform to organize poor black women. In the case of São Paulo, autonomy is privileged over access to resources, and members are expected to first identify as domestic workers. Other movements are only useful as long as their agenda explicitly overlaps with that of the union. By contrast, in Campinas, union leaders use all the other vectors of oppression their member can face to bring them to the union, which then acts as a means to organize the category rather than an end. The union becomes an umbrella organization to mobilize black women.
Existing scholarship discusses quite extensively the difficulties in organizing precarious workers that result from their fragmentation, lack of rights, and isolation (Murphy and Turner 2014;Milkman and Ott 2014;Standing 2014). Yet the sort of fragmentation created by the intersection of multiple vectors of oppression is less often debated (see, for instance, Birdsell Bauer and Cranford 2016), although it appears as quite central to the organizing of marginalized groups. Indeed, domestic workers' unions face the challenge of creating unity in a sector where workers are also poor, black, and women. Gender, race, and class are embedded, but one element can be perceived as more salient in a given context and prioritized differently by each individual. The ways in which social movements interpret and practice intersectionality is fundamental to their mobilizing strategies; it determines their capacity to organize workers, who are not only workers but also gendered and raced subjects. Recognizing and using the intersections of gender, race, and class has enabled Brazilian domestic workers to build alliances with other movements and organize their members in the various social spaces they navigate. The Campinas union, for instance, successfully created an encompassing identity, "black women," that serves as a basis of recruitment and empowerment for domestic workers.
In this article, I have confirmed that domestic workers were able to strategically use their multiple social identities to build alliances (Nadasen 2015;Bernardino-Costa 2015); and I have also demonstrated that the more intersectional unions are in their discourses and practices, the more successful they can be. The strongest union in my sample, which has more active members, more resources, and more visibility, is also the one with the most comprehensive intersectional strategy. Therefore, the case of Brazilian domestic workers provides some insights regarding how to build an intersectional practice as well as the challenges and opportunities associated to it. International comparisons could help to further refine and test this framework, in order to detect to what extent intersectionality explains the successes of domestic workers' movements across the region. Domestic workers remain one of the most marginalized and precarious groups worldwide, yet they have achieved organizational and legislative victories in Brazil and in Latin America. Their forms of mobilization and recent achievements can be an inspiration, showing a way forward to organize the so-called nonstandard workers.
Indeed, workers are always at the intersection of various and simultaneous forms of oppression, based on their gender, race, occupation, religion, citizenship status, and so on. In Latin America, more often than not, workers are nonstandard; they are informal and precarious, and their labor condition is deeply embedded with race and gender inequalities. They rarely experience only class-based issues and tend on the contrary to face a combination of multiple vectors of oppression and exploitation. These factors shape their material conditions of existence, as well as their subjectivities, and should therefore comprehensively be included in any strategy of collective action and resistance.