Rethinking governance through Samarco’s dam collapse in Brazil: a critique from the STS perspective

ABSTRACT

This paper is based on a study of the Fundão dam collapse in Minas Gerais, Brazil, and it aims to analyze a widely used and rarely problematized concept in the corporate and, increasingly, in the public policy world: governance. Considering, in this case, the environmental governance that originated the disaster and the subsequent remediation governance, source of profound inequity, I try to extract its main problematics. In fact, the critical literature on governance systems points to the forced depoliticization of political decisions related to the transfer of public responsibilities to the private sector as the main issue. Based on the case, I intend to argue that the hegemony of techno-scientific solutions in eminently political contexts founds the aforementioned depoliticization, adding to and therefore aggravating the asymmetries of power already configured in terms of race and social class.

RESUMO

Este artigo é baseado em um estudo sobre o rompimento da barragem de Fundão em Minas Gerais, Brasil, e tem como objetivo analisar um conceito amplamente utilizado e raramente problematizado no mundo corporativo e, cada vez mais, no das políticas públicas: governança. Considerando, no caso, a governança ambiental que origina o desastre e a governança da subsequente reparação, fonte de profunda iniquidade, procuro extrair suas principais problemáticas. De fato, a literatura crítica sobre sistemas de governança aponta para a despolitização forçada das decisões políticas relacionadas à transferência de responsabilidades públicas para o setor privado como o principal problema. Com base no caso, pretendo argumentar que a hegemonia das soluções técnico-científicas em contextos eminentemente políticos fundamenta a referida despolitização, somando-se e, portanto, agravando as assimetrias de poder já configuradas em termos de raça e classe social.

RESUMEN

Este artículo se basa en un estudio sobre la rotura de la represa de Fundão en Minas Gerais, Brasil, y tiene como objetivo analizar un concepto ampliamente utilizado y poco discutido en el mundo empresarial y, cada vez más, en las políticas públicas: la gobernanza. Considerando, en este caso, la gobernanza ambiental que da lugar al desastre y la posterior gobernanza de la reparación, fuente de profunda inequidad, trato de extraer sus principales problemas. De hecho, la literatura crítica sobre los sistemas de gobernanza apunta a la despolitización forzada de las decisiones políticas relacionadas con la transferencia de responsabilidades públicas al sector privado como el principal problema. A partir del caso pretendo argumentar que la hegemonía de las soluciones técnico-científicas en contextos eminentemente políticos subyace a la mencionada despolitización, sumándose a y, por tanto, agravando las asimetrías de poder ya configuradas en términos de raza y clase social.

KEYWORDS:

• Governance
• Samarco’s disaster
• knowledges
• State

PALAVRAS-CHAVE:

• Governança
• desastre da Samarco
• saberes
• Estado

PALABRAS CLAVE:

• Gobernanza
• desastre de Samarco
• saberes
• Estado

1. Introduction

“Governance” is a term widely employed to address social and environmental conflicts in Latin America. That is why governance policies prove to be a privileged category when the goal is to pragmatically observe power dynamics. Montoya-Domínguez and Rojas-Robles (2016, 302) trace its origins back to the questioning of the State’s role during the post-cold war social and economic crisis. They observe that the term can be now understood under three approaches: one in which it is a synonym for government, another that assumes it as a normative framework mobilized to achieve the political objectives pursued, and a third that takes it as a non-hierarchical mode of government, with the participation of civil society in the management of common goods and territories, compatible with a new view of democracy. I focus on the third approach. This kind of governance comes as a response to Nation-States legitimacy crisis and bureaucratic inefficiency (Sousa Santos 2005). The sociology of governance is, therefore, “associated with the study of how diffuse forms of power and authority can guarantee some degree of order in the absence of a state” (Losekann and Milanez 2021, 3)¹ and has to do essentially with the “weakening of the bureaucratic-hierarchical model and its replacement by market systems or networks.”

In Latin American countries, the governance model was implemented through Structural Adjustment Programs (SAP) promoted by the World Bank and the United Nations Development Programme (UNDP). For these transnational organizations, “good governance” would be characterized by participation, transparent public decision-making, building consensus on topics of general interest, efficient use of public resources, and accountability. It would, furthermore, aim to overcome the Nation-State’s mismanagement by improving the technical capacity of governments, fighting corruption, and creating new public policies (Montoya-Domínguez and Rojas-Robles 2016, 304). Cristina Zurbriggen (2014, 355) considers that this implementation occurred as a merely technical-administrative issue, without considering Latin American countries’ historical processes, State capacities, and power relations. In the region, the culture of patronage and corporatism between States and society prevails, which undermines the democratic purposes of governance systems. It is precisely by focusing on the false impression that governance systems would be more participatory that Boaventura de Sousa Santos formulates his critique. For him, in global or neoliberal governance, the State assumes a meta-regulatory role: as the owner of national wealth by law, it manages the sale of public goods. Historically, its role has shifted from one of distributing everything that is public to another, always strong and significant, of wealth administration. The gradual predominance of transnational power makes State participation only possible as long as it does not take place precisely as a sovereign State, but as a business mediator. What at first seemed to be an attempt to combine social movements’ demands for participation with market requests for self-regulation ended up fully functioning in the second frame. Thus, verticality is dissimulated in horizontality and laissez-faire eventually looks like participation (Sousa Santos 2005).

When governance systems intend to rule “nature,” it enables international finance capital to directly interfere in decisions on territories inhabited by small communities (Bebbington 2007; Svampa 2011). Environmental governance is fundamentally linked to the processes of defining, regulating, and implementing rules that decide on access to “natural resources” (Martínez-Alier 2004; Castro 2015). As in the SAP’s model, in the studied case, this control is exerted by conceiving social and environmental problems as technical-administrative difficulties that can be solved through the use of new technologies and rational planning. The decisions are restricted to a regional corporate elite that would have enough social capital to interact with transnational power: technical expertise, mastering of other languages, the domain of specialized vocabulary, etc. (Zhouri 2008, 99). Liberal States also tend to deal with environmental issues under the “rationalist assumption that the environment can be controlled and that certainty can be achieved via prediction” (Boulot and Sterlin 2021, 2). “Nature,” thus, in addition to being an abstract object, is passive and has no agency. But the objectifiable and controllable “natural resources” approach is often not shared by those small communities regarding the places where they live.

What I want to argue in this article is that the exclusion of populations inhabiting territories affected by extractivist projects, whose knowledge about these territories is often based on their own subjectivities and experiences with and within them, is fundamentally entangled with the overestimation of technical-administrative and rational planning models. This overestimation is, therefore, a fundamental problem to be considered when dealing with the forced depoliticization of governance and the consequent transfer of public responsibilities to the private sector. Thus, Science and Technology Studies (STS) are an interesting angle to critically analyze governance policies. To formulate my argument in this regard, I rely on the Fundão dam failure case, considered one of the major environmental disasters in Brazil, and on the process of damage repair for the affected people. I draw, more specifically, on the period preceding the dam rupture in the 2010s toward the second reparation agreement, signed in 2018 and named TAC Governança. Some of the elements present here were collected in semi-structured interviews carried out in 2018, in several cities located on the Doce River Valley. In addition to the interviews, I mobilize online legal documentation, journalistic articles, the company’s and public agencies’ websites, and specialized bibliography. These considerations were developed in my master’s thesis and other discussions during my Ph.D.²

The paper is structured in three sections. The first one outlines a brief regional history based on the mining presence in the upper Doce River, highlighting Vale S.A.’s expansion. The second deals with the environmental governance system in Minas Gerais and how it relates to the construction of the disaster. The third is dedicated to the governance of the remediation process for the affected people, carried out mainly through out-of-court settlements.

2. The presence of Vale S.A. on the upper Doce River

Around 3:00 pm on November 5th, 2015, Fundão dam collapsed in a district of Mariana town, Minas Gerais, Brazil, letting escape at least 34 million cubic meters of mining tailings composed of mud, iron ore, and silica.³ This dam, at Samarco’s mining complex, belongs, in equal parts, to the Australian company BHP Billiton and its Brazilian partner, Vale S.A. I focus, however, on Vale S.A., since Samarco’s activities in Brazil constituted a “non-operated joint venture,” that is, Vale S.A. assumed its sole operational and legal responsibility⁴ (Milanez et al. 2016, 23). The dam breach resulted in 19 deaths and directly or indirectly affected the entire Doce River basin. The impacts on the interviewed people’s lives were diverse: the first villages were buried in mud and its residents lost all their goods and belongings, fishermen and small farmers further on the riversides and at its mouth lost their income sources, many have had difficulties in accessing water. The Krenak indigenous people, in addition to the economic impact, say that the river, with which they consider to have a kinship, is dead. Despite these disparities, all of them felt wronged by the remediation process.

In the Mariana region, mining precedes the Brazilian State itself: it was the main reason for colonization expeditions in the sixteenth century and the source of the first economic gold cycle in the eighteenth century. After the decline of this first cycle, large-scale mining was only retaken at the beginning of the twentieth century, with the foundation of Vale do Rio Doce Company (CVRD in its Portuguese acronym). Created in 1942 during the dictatorial period of Getúlio Vargas’ nationalist government, its goal was to supply raw materials for the military industry during the post-war period (Corsi 2008, 75). From its inception until the 1960s, the company grew within a context of Import Substitution Industrialization (ISI) policies, which ended up leaving the country indebted and susceptible to international pressure to implement SAP in the 1990s. In 1997, CVRD was privatized, becoming Vale do Rio Doce S.A. and, as of 2007, Vale S.A. (Dalla Costa 2009).

The relationship between the new company and the public sector persisted throughout its privatization in 1997: Vale’s sale was considered to be one of the most outrageous in the country’s history due to the fact that it was exceptionally undervalued (Rodrigues and Jurgenfeld 2019).⁵ Bradesco, a private bank, was responsible for its evaluation and one of the directors of the same bank was named the company’s first president. Since then, an important block of its shares has remained in the power of senior government officials through pension funds. The National Bank for Economic and Social Development (BNDES) has always financed Vale’s expansion (Milanez et al. 2018). Today, the company is the fourth largest mining company in the world in terms of value⁶ and the fifth in terms of revenue.⁷ Tax subsidies and the flexibility of various control mechanisms are also expansion factors (Marshall 2015, 149). According to its own website, the company operates in over 30 countries.⁸ As of 2017, Vale underwent a restructuring that implied a significant participation increase of foreign institutional investors. This process meant a redistribution of power in favor of owners and to the detriment of managers and the State (Santos 2019).

One of the company’s growth strategies is bolstering its institutional influence by lobbying. In the National Congress, the so-called mining bench, formed by deputies and senators financed by or supporting the sector, became official in 2011 with the Frente Parlamentar da Mineração Brasileira and eventually with the Frente Parlamentar Mista da Mineração in 2019.⁹ Shortly after the dam collapse in Mariana, the Minas Gerais mining bench blocked the approval of strengthening control measures proposed to prevent such tragedies at the federative unity level.¹⁰ Another example of the confusion between private and public is the National 2030 Mining Plan – meant to provide the basis for drafting the bill referring to the new code – debated in 2010. Vale mobilized its lobbying forces to participate in part of the workshops alongside government organizations. In 2014, there was another debate on the mining code. Of the 27 members participating in the special committee, 19 received Vale’s donations. In the same year, Vale donated R$22.6 million to political party committees. Dilma Rousseff, of the Worker’s Party (PT), received R$12 million, and her opponent, Aécio Neves, of the Brazilian Social Democracy Party (PMDB), received R$3 million (Milanez et al. 2018). “Revolving doors,” the coming and going of certain people between top positions in companies and in civil service, is another common practice (Milanez et al. 2018). Since the mid-2000s, the mining bench, together with other conservative groups in Congress have been pressing for greater flexibility in environmental laws, which are still those of the 1988 Constitution, considered by them an “obstacle to development.”¹¹

SAP’s emphasis on corporate self-government, justified by the distrust of State institutions in “Southern” countries, opened space for transnational companies to develop their own Corporate Social Responsibility (CSR) policies as a response to environmental conflicts and as a means of local social insertion. CSR includes sustainability and transparency policies and risk management plans (Bebbington 2009; Perelman 2019; Leclerc 2017). Vale’s policies place the company as a “sustainable operator,” a “local development catalyst,” and a “global sustainability agent.” In explaining the last point, it commits, among other things, to “promote complete transparency in terms of governance, policy, procedures, practices, and the company’s performance to global stakeholders.”¹² Vale undertakes also several educational, cultural, social, and environmental projects. At the local level, it establishes partnerships with city halls to carry out infrastructure works, as its revenue is much greater than the municipal ones. These discourses and practices help consolidate and expand companies in the environment where they operate, influencing “emotional and cognitive patterns and the agency of civil society”¹³ (Milanez et al. 2018, 28). The company monitors communities’ “level of criticality” in relation to its presence by developing mediatic strategies, and employs communication and socio-environmental management agencies for the planning of strategic approaching, also monitoring communities and potential “social risks.” Carrying out “political and economic diagnoses,” organizing “dialogue meetings,” and financing social projects, the company stabilizes and legitimizes its presence. In addition to insertion and control policies, in 2013 the press denounced that Vale hired intelligence services in order to obtain information and infiltrate certain groups and social movements (Milanez et al. 2018, 29).

Vale S.A. is a source of labor and socio-environmental conflicts in several countries around the world, especially Mozambique (Amisi et al. 2015, 98). In Mariana, the main forms of disrespect for labor legislations by Samarco were illegal outsourcing, non-payment of commuting time (considering that the areas of operation can be quite far from urban centers), and non-inspection of working conditions to be in compliance with labor standards by service providers (Milanez et al. 2016, 26). Even so, the Samarco complex was elected one of the “Biggest and Best Companies of the Year” by the national magazine Exame¹⁴ in 2015, reaching, according to economic experts, the best profitability in the mining sector, despite the economic and financial crisis and the iron ore price fall. Additionally, Samarco won several awards in 2014 for its high standards of environmental responsibility,¹⁵ and BHP Billiton was ranked by the Corporate Human Rights Benchmark group among the most responsible companies of the year in 2017.¹⁶

As in other parts of the world,¹⁷ the installation of mining companies in the Mariana region generated local economic and social dependence on this activity. According to an IACHR document (cited by Borges and Maso 2017, 77), this occurs because the product and service sectors focus their activities on meeting the company’s direct and indirect demands. The villages in the region soon end up naturalizing the dependency and considering it a privilege to have the resources from it, disregarding that, in fact, the lack of other sources of revenue is an effect of the presence of mining companies. Poverty and social inequality make negative impacts more easily acceptable. The regional income inequality indicators are particularly alarming: among the 853 municipalities in the Minas Gerais State, Mariana is the one that collects the most royalties and, at the same time, is the 226th most unequal (DATASUS cited by Milanez et al, 2016, 192). Thus, poverty and company dependency feedback on each other. Furthermore, the region is heir to the social relations established in colonial slavery times, with racial inequality adding to income inequality. Black and brown groups were disproportionately exposed to the dam failure, the black population representing between 80% and 60% of the first affected villages, that is, those who suffered forced displacement. The percentages are much higher than the State average of 54% (Milanez et al. 2016, 33). This multiplicity of issues and scales makes the disaster debate very complex. This is why it is important to draw some lines on its construction; subject of the next topic.

3. Environmental governance in Minas Gerais

In Brazil, as in many countries, environmental issues are regulated by a specific structure due to their multi-sectorial complexity. In Minas Gerais, the highest bureau in charge of managing environmental policy is the State System for the Environment and Water Resources (SISEMA). Within the SISEMA, the State Secretariat for the Environment and Sustainable Development (SEMAD) is responsible for environmental licensing, having two divisions: one corresponds to the State Council for Environmental Policy (COPAM) and to the State Council for Water Resources (CERH). These two councils regulate and develop technical standards for environmental and water policy. COPAM and CERH are collegiate, normative, consultative, and deliberative bodies.¹⁸ These councils are composed of public power and “civil society” ¹⁹ representatives. However, being able to participate in the councils as “civil society” depends heavily on the cultural and social capital that each actor can mobilize, which eventually means that popular participation is extremely limited. Agribusiness and mining companies are the sectors taking control of the councils (Carneiro 2007). The other division corresponds to the State Foundation for the Environment (FEAM),²⁰ an executive bureau responsible for granting environmental licenses and for observing compliance with regulations, which must take into account what is defined in the two councils. The monitoring itself is not exerted by FEAM, but by outsourced companies.

Let us summarize the essentials regarding the licensing related to Fundão dam’s building and expansion. Its construction authorization process began in 2005, with the submission of an Environmental Impact Study (EIA)²¹ to the FEAM. In 2007, during the first spike of iron’s international price, the license was granted and in 2008 the dam started operating. In 2011, there was a second international increase of iron prices, which was followed by a growth in Samarco’s production. In 2012, an overload was detected in the dam as a result of it. Despite the problem, Samarco submitted an EIA requesting an operating license renewal in 2013. This license was not granted until the dam’s breach. In 2013, another EIA was presented requesting the dam’s raising and its unification with a neighboring dam to increase its capacity. Iron prices began to fall, so the proposed raising model was the cheapest and riskiest, linked to several dam failures around the world: the upstream one. The license for the upstream raising was granted in 2014 and construction started in June 2015. In November 2015, the dam collapsed (Milanez and Santos 2017, 136).

Residents of Bento Rodrigues, the first community to be affected by the dam tailings, were never consulted about the construction or alteration of dams. Eliene dos Santos, former resident,²² reports that, between 2010 and 2011, a Samarco outsourced worker warned them that they were at great risk. Frightened, the residents’ association asked for a meeting with Samarco. According to Santos, during that meeting:

Samarco’s representatives told us that there was no risk, that we could stay tranquil. I had the minutes registered in the meeting minutes book of them saying that. But of course, right, our documents went away [were buried by the mud]. The minutes were a document. And they said no, that there was no risk.²³

The case reflects a transnational tendency to privilege, within the framework of governance policies, risk management based on technical knowledge to the detriment of human rights (Perelman 2019) and to overlay compensation mechanisms over participatory discussions (Castro 2015). EIAs end up being a mere formality, with no factual possibility of preventing project implementation. The authorities in charge of licensing hardly consider the possibility of disapproving projects, as the argument of regional economic development and the presence of mining companies in the councils weigh heavily. In general, it is assumed that potential problems can be eventually avoided, mitigated, or compensated (Zhouri 2008).

If the Fundão dam breach was the first to attract international attention due to its dimensions, wrongdoings committed by Samarco have been occurring for much longer. Between 1996 and 2015, the company was fined 18 times for environmental violations. The company’s first strategy to respond consists in pleading not guilty in all cases, using legal and technical arguments, which undermines legal authorities’ legitimacy and State regulatory bodies’ competence. If this strategy fails, Vale tries to reduce the fines applied or to delay the process as much as possible, aiming at its statute of limitations (Milanez et al. 2018, 22). The signing of Terms of Conduct Adjustment (TAC), an extrajudicial settlement procedure, can also be considered an institutional strategy: it displaces the State from decisions on violations committed by the company, making it a mere mediator. Such agreements are often used to avoid sanctions for labor analogous to slavery, violation of Indigenous rights, and environmental damage, as discussed in the next section (Milanez et al. 2018, 22). TACs were also the tool used to repair the damage caused by the Fundão failure.

4. Remediation governance

After the dam collapsed, State and companies proved unable to deal with it. In part, this is due to the fact that a disaster of this magnitude was unprecedented and, therefore, they had no experience on how to respond to it; in part, to the fact that the risks to which these populations were subject were always veiled; and, in part, to the companies’ attempt to minimize the understanding of the damages, limiting as much as possible their responsibilities and, consequently, their expenses.

There are numerous individual complaints, two public civil lawsuits, and some criminal investigations in national and international courts related to the Fundão dam failure. For this paper, I focus on summarizing the negotiations resulting from the two public civil lawsuits (ACPs – Ações Civis Públicas), which led to the signing of Conduct Adjustment Agreements (TAC – Termos de Ajustamento de Conduta); an extrajudicial settlement that, by definition, ought to be negotiated between the two parties, the judicial State body being the mediator. This choice is due to the fact that the terms signed within the scope of the two ACPs are those having the greatest impact on affected people’s lives, collectively. But why an out-of-court settlement? In Brazil, there is no unified legal body dealing with diffuse and collective rights. Conduct adjustment agreements are designed to resolve conflicts caused by the violation of this type of right,²⁴ while the State carries the obligation to guarantee and supervise their execution (Roland 2018). There are no regulatory frameworks that indicate guidelines for involuntary resettlements either (Xavier and Carneiro 2020). The option was then justified by the Executive power as the fastest way to respond, since litigation would require a long and complex procedure that could take years, while the situation required urgent action (Roland 2018, 22). On his part, Mariana’s mayor was particularly concerned about restarting Samarco’s activities, given the economic dependence on the companies,²⁵ and the Minas Gerais governor held his first press conference at Samarco’s headquarters (Borges and Maso 2017, 72).

Before analyzing the negotiated solution issues, it would be helpful to try to define what would be an ideal remediation process. Primarily, affected people would have the right to inform themselves about the harm suffered and to be counseled on their other rights, including where to seek help and through what procedure. Another important legal principle is the centrality of the suffering of the affected people, which would place them as primordial for the reasoning about practical strategies that would guide the process. Wrongs caused by third parties should also give the right to full reparation, that is, where the affected people achieve a status at least close to the previous one (Roland 2018, 6).

The first public civil lawsuit,²⁶ which gave place to the first agreement, was brought by the Union, the two federations involved, and public administration agencies against the company Samarco and its shareholders, Vale S.A. and BHP Billiton. The ACP’s main intention was to adequately account for the damage so that remediation could be established. The resulting agreement (TTAC – Termo de Transação e Ajustamento de Conduta) was signed in march 2016, excluding affected people’s participation in all its phases. One of the results of this agreement was Vale's creation of Renova, a private foundation meant to manage resources and to carry out the reparation, mitigation, and compensation programs for socio-economic and socio-environmental damages (Milanez et al. 2018, 22). The agreement forced the affected people to negotiate directly with the foundation, which spawned several problems. The registry form that the companies tried to impose since the first days after the crime in order to catalog the affected people had the power to attest whether they had been or not “impacted,” thus determining the eligibility criteria for the right to participate in compensation programs. The adopted definition of “impacted” people, however, was extremely restrictive to limit as much as possible the number of people to be compensated.²⁷

As a result, many of them were denied their rights to emergency compensation and, among those who had access to it, many were denied recognition of part of the damages. In several cases, taking advantage of affected people’s vulnerability, Renova offered smaller compensation in comparison to what they would be entitled (Borges and Maso 2017). Besides, outsourced companies were hired for the form’s application service, and most of their technicians were not prepared to work with the communities, which engendered more psychological violence.²⁸ Many people could not access their rights to remediation because they did not have access to information on that same right, the magnitude of the crime they had been subjected to, or the justice system in general. These problems and the discontent they generated gave rise to an important mobilization. The social movements that collaborated with the affected people since the beginning helped organize protests, including a march that was violently repressed, and the TTAC was suspended. The Public Prosecutor’s Office was then involved in a new ACP,²⁹ which sought to provide a more complete account of the damages, considering, for example, the affected people’s social suffering. The first result of this ACP was a Preliminary Adjustment Term (TAP),³⁰ signed in January 2017, intending to establish the hiring of a team of experts to assist the Public Prosecutor’s Office in carrying out the diagnosis of the socio-environmental and socio-economic damages resulting from the dam failure (Roland 2018). However, this TAP foresaw engaging private firms which were service providers hired by Vale and Samarco, and was therefore again contested (Milanez and Santos 2017, 139). In November 2017, an addendum to the TAP³¹ was signed. The document defined the hiring of other companies to carry out the damage assessment, in addition to the use of accessible language, the principle of transparency and participation, respect for collective logic, and the hiring of technical consultants.

The second ACP finally resulted in a new agreement – the Governance Conduct Adjustment Agreement (TAC Governança), signed in June 2018. The agreement brings in its name the ambiguity summarizing the problem of negotiated settlements. On the one hand, it aimed at correcting the main mistake of the previous agreement: the exclusion of affected people, that is, lack of participation. On the other hand, the conditions of extrajudicial negotiation favored companies. That is precisely the structure at the origin of the crime. Indeed, in addition to swapping the term “impacted” for atingidos,³² mechanisms were established to allow affected people to effectively participate in all stages of decision-making and to expand the structures for monitoring and controlling Renova activities. But that was not enough to guarantee a fair remediation process. The agreement was again elaborated exclusively by the public authorities in the role of affected people’s representatives, and by the companies, in a situation of extreme inequality. The companies could mobilize the best and greatest legal body, while the Public Prosecutor’s Office and the Public Defender’s Office had limited resources. Although it ensured a broad participation in the structures created for remediation, the agreement did not resolve the disparity in representation between the companies involved and the people affected. Before the sentence that ratified the agreement, a letter was written by affected people and partner organizations asking the responsible judge to carry out due diligence in the affected territory to better understand the case. The request was ignored, and, in addition, the magistrate amended one of the texts of an addendum agreement by prohibiting the hiring of technical consultants that were linked to social movements or religious entities, precisely those that assisted affected people from the first moment. From the R$155 billion in damages assessed by the second ACP, only R$2.2 billion were taken from the companies as collateral. Finally, the companies played, through Renova Foundation, a leading role in the remediation for the harm that they themselves are responsible for (Roland 2018).

The affected people, with the support of social movements, organized themselves into independent committees and met regularly to discuss the evolution of the remediation process, the reconstruction of devastated communities, and their own participation in decisions. It was also through these dialogues that these groups spread the idea that access to common goods, such as water, is a non-negotiable right. The idea of the river as a subject of rights, which already existed in the indigenous struggle, gained strength in mobilizations, as it matches the perceptions of peasants and fishermen in the region. This was extremely important to maintain a certain unity in situations of exhaustion and tension, especially considering the long wait for a response and the considerable experience of the companies in demobilization. Affected people, social movements, NGOs, and local associations were the main actors in the struggle for just reparation:³³ their diverse and uninterrupted actions, constantly criminalized, changed the course of negotiations, even if the results are still beneath ideal. The TAC Governança was not complied with. After the judge set aside the affair as a result of allegations of bias and the Public Prosecutor Office’s continued demonstrations of discontent with the remediation process, in September 2022, almost seven years after the disaster, the public entities made official the decision not to renegotiate the agreement with the companies implicated. Considering that the companies’ financial disbursement proposal does not correspond to the “need for full, swift and definitive reparation for the Doce River and the affected populations,” these entities must now resort to litigation.³⁴ The resettlement of buried communities has not yet been accomplished.

I interviewed people from two buried communities – Bento Rodrigues and Paracatu de Baixo – in August 2018, when the second agreement had been recently signed. The two communities were mainly dedicated to family farming and animal raising. According to newspaper reports at the time, they had approximately 600 and 300 inhabitants, respectively.³⁵ The percentages of black and brown population are 84.3% for Bento Rodrigues and 80% for Paracatu de Baixo (Wanderley 2015, 4). The bond of those populations among themselves and with the place they inhabited was profound and celebrated through popular Catholic traditions. At that moment, the master plan for Bento Rodrigues’ reconstruction had already been approved and was in progress. Master plans are market tools meant to propose urban solutions, which often result in pasteurized recipes that ignore local socio-spatial practices (Xavier and Carneiro 2020, 21). In a meeting for “Survey expectations” on Bento Rodrigues' reconstruction, the community formulated a single demand: a place that made it possible to maintain the quality of life and the socio-affective, economic, and territorial relationships that it previously had (Xavier and Carneiro 2020, 28). The first versions of the master plan were elaborated without any community participation and “resembled a generic condominium that, in almost everything, differed from the old Bento Rodrigues spatialities and territorialities” (Xavier and Carneiro 2020, 29). Santos, Bento Rodrigues’ school principal, describes some difficulties in approving the project for the school reconstruction:

I told them that it is no use having a huge building, with several departments. We don’t even have employees for that. I saw there a reception as if we were going to have a doorman who directs the kids somewhere  …  We are a small team, right, it won’t happen. […] It looks nice, but then it does not make sense.³⁶

The master plan for Paracatu de Baixo, in turn, was in the final stage of discussion. When I interviewed Luzia Queirós, the Commission of Affected People representative, she counted already 25 versions of the master plan. She vented:

I do not know why they made it take so long, so late. If they had listened to us and done it, understood it in our own way, that would have already ended. We would not be in this hell. So it was because of bureaucracy, of things they invented.

During this process, the community had to learn how to deal with the bureaucratic world: “After we got the technical consultancy, we had to learn what each of the difficult words meant and had to interpret them.” Xavier and Carneiro (2020, 30) confirm that “the master plans were presented to the community using graphic and oral technical language that was not very accessible and understandable. Thus, as in a closed cycle, projects were developed, presented, rejected, and redone.” After fighting so hard for participation in the community reconstruction, Luzia Queirós concludes:

wherever technology failed, the capiau³⁷ are outdoing themselves. Because we are ignorant, but in our ignorance, we showed that we know much more than books and encyclopedias, because we know what we are feeling in our skin. So, you can look in all the books, but the books will never teach. Because we have to show where is the wrong and where it has to be remediated. It’s us. And we still believe in justice, we still believe in law.

Other narratives along these lines were shared with me. Residents of Santana do Deserto, a community more than a hundred kilometers away from Mariana, had their lives disturbed by the works to contain the dam tailings in a nearby hydroelectric plant also belonging to Vale S.A. The siblings José Ribeiro Neto and Raimundo Ribeiro Filho explained³⁸ in great detail how the land behaves in that area and why they believe that the containment will break in a few years, but they were never consulted. Maria dos Reis,³⁹ member of the Landless Workers’ Movement (MST – Movimento dos Trabalhadores Rurais Sem Terra), affirms that a group of women from her community offered to participate in the region’s spring water recovery projects, but they were not taken into account. Daniel Krenak, deputy leader in the Atorã indigenous village, when talking about what he expects for the future, said:

We want the Sete Salões area demarcated and our kids in college. Because today you have to have knowledge to be able to debate it out there. Today, if I sit at a round table to debate and I don’t have a college degree, my voice doesn’t say anything, my voice is nothing.⁴⁰

5. Conclusion: rethinking governance

Among the most notable problems of governance systems is the fact that they give rise to a transfer of public roles and responsibilities to the private sector (Bebbington 2007; Zhouri 2008; Sousa Santos 2005; Castro 2015; Svampa 2011). Environmental and social equity dilemmas are therefore approached from within the capitalist structure (Castro 2015; Svampa 2015), with no panorama of changing it. Declaring their values as technical, therefore neutral and universal, is how transnational corporations and multilateral organizations could justify their independence from national States (Spurk 2016). The emphasis on transparency, risk management, sustainability, and compensation discourses replaces and thus prevents a previous political discussion with the groups potentially affected about the very existence of extractive projects (Leclerc 2017). The primary political antagonism between corporations and local communities is therefore structurally erased. Negotiated settlements, for instance, replace and prevent litigation in which the two parties would position themselves as antagonists and the affected people would be heard as such. The dynamics of participation restricted to actors with technical capacity are finally dissimulated as democracy (Zhouri 2008, 98). The technification and consequent depolitization of politics are also related to the society-nature dichotomy and the progressive incorporation of the former within the latter. The environment, that is, “nature,” is perceived as an autonomous reality, external to social relations, seized only by illustrated knowledge and part of a single conscience that must lead to universal consensus (Zhouri 2008, 98). In governance systems the field of what was once “nature” is expanded over society: it becomes an abstract entity, controllable through prediction, with no capacity for agency.

What I would like to suggest, following the foregoing analysis, is that the hegemony of techno-scientific solutions precedes the problem of public power transfer to the private sector in the discussion on governance. This suggestion can be supported by at least two arguments. First, throughout Latin America, the centralization of power in the State has never been a promising prospect for rural communities, which is why focusing the debate on the lack of State power may not be fruitful. In the region, mining was established in colonial times through the exploitation and genocide of native and African peoples. Soon it became a source of income for nationalist development and today it provides profits to shareholders around the world without ever having improved local equality indices. This process of internal colonization (Gonzáles Casanova, 2007) has continuous recourse to authoritarianism: first colonial, then dictatorial, and now neoliberal. In the context of the global expansion of extractive frontiers (Moore 2000), State coercive force is central to accumulation by dispossession, especially the agrarian one (Harvey 2003; Levien 2015). Sousa Santos (1999) refers to States as the main actors of globalization. In Latin America, this modus operandi is especially marked by different types of violence: semiotic, economic, legal-political, and military (Machado Araóz 2011).

Second, the predominance of technical-scientific-bureaucratic responses sustains the transfer of public functions to the private sector. As we saw in the case studied, when the Bento Rodrigues’ association finally managed to arrange a meeting with the company before the dam collapsed, the technicians in charge ensured that everything was under control based on their risk-management practices. The affected people’s demands for participation in decisions were not considered effectively legitimate precisely because they are rural populations, without a high degree of formal education. After having their past devastated by the irresponsibility of transnational companies, these people were denied the possibility of building their own futures, essentially linked to the community and the territory they inhabit, because their knowledge is not recognized as such. In addition, the urban middle class and the rural lower classes are ontologically alien. The judge and the corporation’s understanding and interest in the reality and needs of the affected people were extremely superficial and stereotyped, which made the reparation process even more harmful.

The critique elaborated by the STS field proves, then, to be an interesting tool in analyzing governance policies. For John Law (2004, 13) “the basic intuition [of STS] is simple: it is that scientific knowledge and technologies do not evolve in a vacuum. Rather they participate in the social world, being shaped by it, and simultaneously shaping it.” Drawing from this definition, we can argue that all the knowledge that sustains governance practices – the paradigm that allows the implementation of extractive projects in foreign territories (EIA and inspections methods, the means of participation) and their conflict resolution mechanisms (out-of-court settlements, indemnities, master plans) – is produced in a context of specific practices and, at the same time, produces new practices in the territories where the projects are implemented.

In order to contemplate the epistemic maneuver that founds the exclusion triggered by governance policies, we have to remember two things: first, as Latour (2006) taught us, the separation between nature and society (as politics and economy) is an invention of modernity, which is often not completely shared by the communities affected by extractive projects. This same division also relegates nature to science, which in the modern world can only be known through absolute and irreconcilable detachment from the object. Viveiros de Castro (2004, 468) writes that, in modern episteme,

to know is to objectify – that is, to be able to distinguish what is inherent to the object from what belongs to the knowing subject and has been unduly (or inevitably) projected into the object. To know, then, is to desubjectify, to make explicit the subject’s partial presence in the object so as to reduce it to an ideal minimum.

Furthermore, methods of knowledge based on subjectification are a matter of epistemic injustice: a subtle way in which some groups are excluded from participation in shaping law and public policy (Tsosie 2012).

Second, modern (neoliberal) politics/economy and science are inventions based on this very division. Jasanoff (2004, 14) treats science and society as a co-production: the same type of institutionalized knowledge that has the power to produce and reconfigure “nature” through science, produces and reconfigures society through laws, regulations, bureaucracies, and financial instruments. Institutionalized ways of knowing are continually enacted in new contexts; either because it would be too disturbing to examine them openly, because they are socialized into actors, or because they have their own mechanisms for managing dissent. Solidified in the form of administrative routines, these repertoires of knowledge and power offer constant backup positions from which answers to new problems can be built. When the produced knowledge is institutionalized in governance policies, its statements acquire a “truth-value,” a “fact-value,” that are stabilized and begin to be reproduced without any reference to their construction process (Latour and Woolgar 1988). This is how governance policies that are designed based on northern countries' liberal mentality can be deterritorialized and reterritorialized (Santos 2001) in places where the socio-natural formations⁴¹ are other. This is how, for the enlightened urban middle class, a pasteurized condominium project seems to be what anyone should wish.

But what is the issue, then? Well, this process has nothing to do with democracy, participation, or transparency. The small rural communities in the interior of Brazil and – I dare to state – in the interiors of colonized countries, “Southern” countries or however we want to call them, do not necessarily share either the nature-culture divide, nor the scientific conception of a distant and objectified nature. Often, their relationship with the space and with the human and non-human beings who share it with them can be social, spiritual, or even of kinship. Relating might be how they know. And, of course, their ambitions as a community of humans may have nothing to do with the ambitions of the urban middle classes. Governance is a significant contemporary resource to impose the nature-culture divide and the technicized politics where they erstwhile did not reign supreme. The meanings of ethics, justice, or the common good cannot be pasteurized through formulas conceived in the liberal tradition of hegemonic countries. There is no democratic practice possible in the worldwide countrysides that does not require an opening to other socio-natural formations, other social ambitions not mediated by science, and other knowledges based on relationships, care, attachment, and closeness. These knowledges generate other skills and other techniques. Democratic practices must treat science and collective experiences and subjectivities at least symmetrically in worlds where capitalist modernity does not prevail.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Luciana Landgraf Castelo Branco

Luciana Landgraf Castelo Branco is a contractual doctoral student in the “Knowledge, Science, Education” program at the University of Paris Cité, attached to the Population and Development Center (Ceped-IRD). She has carried out fieldwork in rural and indigenous communities in Peru and Brazil. Based on a transdisciplinary critique of development, her research currently focuses on the conflicts generated by the mining sector.

Notes

1 For a review on the sociology of governance, see the same article.

2 I would like to thank Bruno Milanez, Michele Leclerc, Angela Camana and the three anonymous referees for their valuable feedback. Any shortcomings remain my responsibility.

3 This information is in the Public Civil Lawsuit no. 0023863-07.2016.4.01.3800, available at: http://www.mpf.mp.br/mg/sala-de-imprensa/docs/acp-samarco.

4 This should not be interpretated as an argument against holding the other company accountable.

5 See: https://www.brasildefato.com.br/2017/05/07/venda-da-vale-completa-20-anos-e-foi-um-dos-maiores-crimes-cometidos-contra-o-brasil.

6 See: https://www.mining.com/top-50-biggest-mining-companies/.

7 See: https://www.statista.com/statistics/272707/ranking-of-top-10-mining-companies-based-on-revenue/.

8 See: http://www.vale.com/en/aboutvale/across-world/pages/default.aspx.

9 See: https://observatoriodamineracao.com.br/frente-parlamentar-da-mineracao-defende-os-interesses-do-garimpo-e-pressiona-por-mudancas-favoraveis-ao-setor-no-congresso/.

10 See: https://www.brasildefato.com.br/2019/01/28/como-o-lobby-de-mineradoras-trava-projetos-de-interesse-publico-sobre-barragens/.

11 See, for example: https://www.bbc.com/portuguese/brasil-49453037.

12 See: http://www.vale.com/esg/en/Pages/PoliciesAndCorporateDocuments.aspx.

13 All translations are my own.

14 See: https://exame.com/negocios/exame-premia-as-melhores-empresas-de-2014/.

15 See: https://www.samarco.com/wp-content/uploads/2020/12/2014-Anual-Sustainability-Report_compressed.pdf.

16 See: https://www.ihrb.org/focus-areas/benchmarking/2018-corporate-human-rights-benchmark.

17 See, for example: Kirsch (2014), or Echave et al. (2013).

18 See: http://www.meioambiente.mg.gov.br/instituicao.

19 I consider the separation between State and civil society problematic, which is why the expression appears in quotation marks. For a more elaborate critique, see: Zhouri (2008) and Leclerc (2017).

20 And two other entities: State Forestry Institute (IEF) and Minas Gerais Institute for Water Management (IGAM).

21 EIAs are an initial and fundamental part of the licensing process for project implementation, as they condition its approval.

22 All translations are my own.

23 Interview held on August 28, 2018.

24 Law No. 7347/1985, Art. 5, paragraph 6.

25 See, for example: https://www.youtube.com/watch?v=DF126Qfpc1w.

26 ACP no. 0069758-61.2015.4.01.3400.

27 Interview with Tchenna Maso, activist and lawyer for the MAB, held on October 2.

28 Information obtained in a semi-structured interview with Ellen Barros, member of Caritas, held on August 18, 2018.

29 ACP no. 0023863-07.2016.4.01.3800.

30 See: https://www.mpf.mp.br/mg/sala-de-imprensa/docs/termo-de-acordo-preliminar-caso-samarco.

31 See: https://www.mpf.mp.br/grandes-casos/caso-samarco/documentos/aditivo-tap.

32 Atingido could be translated as “hit” or “affected,” but what matters here is that the term relates to a political construction headed by the social movement of people affected by dams (MAB – Movimento das pessoas atingidas por barragens), originated in the 1980s.

33 It is important to emphasize that the impact of social mobilizations was reinforced by international pressure, including UN agencies, demanding, for example, companies’ accountability, greater consideration for affected people, and expense transparency (Borges and Maso 2017, 83). It is not less important, however, to point out that international organizations have at least an ambivalent role. The UN encourages CSR practices and works in partnership with private institutions that develop them. The leading document used by the Corporate Human Rights Benchmark, which gave an award to BHP Billiton, is the UN Guiding Principles on Business and Human Rights. The fact that such awards were offered to companies involved in crimes proves that their evaluation methods are not only not reliable, but can also be a disservice by creating a false image of such companies as socially and environmentally responsible, finally supporting them in perpetuating that modus operandi.

34 See: https://www.em.com.br/app/noticia/gerais/2022/09/09/interna_gerais,1392455/negociacao-de-acordo-de-mariana-e-encerrada-oficialmente.shtml.

35 See https://g1.globo.com/minas-gerais/noticia/2015/11/barragem-de-rejeitos-se-rompe-em-distrito-de-mariana.html. and https://www.em.com.br/app/noticia/gerais/2015/11/17/interna_gerais,708818/populacao-de-paracatu-de-baixo-pede-reconstrucao-de-povoado.shtml.

36 Interview held on August 23, 2018.

37 Adjective to pejoratively qualify those who live in the countryside and have a simple life, appropriated and inverted by the interviewee.

38 Interview held on August 29, 2018.

39 Interview held on September 26, 2018.

40 Interview held on September 28, 2018.

41 I borrow this expression from Marisol de la Cadena (2010).