The Enduring State: An analysis of governance-making in three mining conflicts

This article investigates the profound ambiguity of the state in the organization of contemporary business– society relations. On the one hand, there has been a decisive shift from government to governance, encouraging private actors, such as corporations, communities and NGOs, to address social and environmental concerns themselves, i.e. without the state’s involvement. On the other hand, however, the continued importance and relevance of the organized state is difficult to ignore. In this article we examine the role of the state in three cases of mining conflicts in Chile, one of the world’s most important mining countries. Through longitudinal, qualitative research of conflictive mining governance relations between state organizations, large corporations and local communities, we show that the modes of influence conducted by the Chilean state oscillate between direct, central steering (‘cathedral’) and indirect, dispersed vouching (‘bazaar’). Elaborating on Foucault’s concept of governmentality, we offer a hybrid theory of state organization, where the dematerialization of the state’s responsibility is seen not as the norm but rather as a particular mode of governance that sits alongside the underestimated, yet enduring, material involvement of the state. We were bored and tired of always being in conflict with the community over many and the community was also bored and tired of this. we had to change the way we related to one another. Our relationship played out in the courts of law and that’s not sustainable for anybody. It’s exhausting for everyone! (interviewee, Pelambres management)


Introduction
The role of the state in contemporary organized society is profoundly ambiguous. On the one hand, it is very much visible and present, acting continuously, shaping and steering societies legally, new governance regime, producing an ambiguous mishmash of materialized and dematerialized presence of the state. Borrowing the concept of 'hybrid cathedral-bazaar' governance from Boelens, Hoogesteger and Baud (2015), we show how the five modes of influence enact an intricate relation between the direct and material nature of state responsibility ('cathedral' mode) and indirect and dematerialized forms of state responsibility ('bazaar' mode). We argue that these different modes of influence assemble into a general, multifaceted 'governmentality' where the state organizes its own dispersion to ensure its enduring governing capacity.

(Re)encountering the Organized State: From Governance to Governmentality
In OS and the wider social sciences, a decisive shift from the notion of government to that of governance has taken place over the past half century. It is one that has emerged in response to numerous political, economic and cultural developments and crises that have played out on a global stage (Barry, Osborne, & Rose, 1993;Miller & Rose, 1990), signalling a diminishing role of the state and a shift of power towards private actors in contemporary political economies (Rhodes, 2007). Yet, following Arellano-Gault, Demortain, Rouillard and Thoenig (2013), perhaps this move towards governance and the, often uncritical, assumption of the reduced power of the state has been premature and misguided; for them, the state has never really gone away. Indeed, since the 1950s there has been a longstanding concern in OS about the rationality behind the control-based, rule-bound, hierarchical functioning of the state (Bendix, 1947;Blau, 1968;Hall, 1962;Warwick, Meade, & Reed, 1975;Weber, 2009) along with its theoretical and moral setup (Heady, 1959;Udy, 1959; see also du Gay, 2000). Even during the 'linguistic' and 'systemic' post-Weberian turns in the 1970s and 1980s, the notion of the organized state managed to retain its protagonism. While its shapes, roles and practices have undoubtedly changed, the state should be understood as persisting, enduring and powerful actor. There is hence a need for a renewed focus of OS scholars on understanding and conceptualizing the organization of the state (Djelic & Etchanchu, 2015;Schrempf-Stirling et al., 2016;Whelan, 2012).
In recent decades, interest in the organization of the state has endured, yet its focus has veered dramatically towards post-bureaucratic reforms, aiming at enabling and expanding governance networks. This is where stakeholders -involving both state and private, non-state actors -engage reciprocally, exchanging resources and political views, so that effective governance mechanisms can be set up (Bernstein & Cashore, 2007;Rhodes, 2007). Such governance efforts are mobilized as enterprising initiatives both within and beyond the state, calling for the empowering of free networking individuals, the flexibilization of hierarchical power structures and the conduction of a constant process of identity work by 'post-bureaucrats', beyond the auspices of traditional statecraft and governmental authority (Heydebrand, 1989;Johnson, Wood, Brewster, & Brookes, 2009; see also Deacon, 1998;Rose & Miller, 1992;Stokes & Clegg, 2002;Thomas & Davies, 2005).
The abovementioned inquiries coincide with efforts to introduce the normative (mostly Habermasian) notion of 'political CSR' (PCSR), which can be said to represent the ideal culmination of the governance era (Djelic & Sahlin-Andersson, 2006;Moon et al., 2011;Rasche et al., 2013). From a normative standpoint, advocates of PCSR have made the case that governance can foster a fairer, democratic, self-regulatory interaction rationale between (self-organizing) stakeholders, in the face of social inequality, globalization and environmental degradation (Mason & Simmons, 2014;Scherer & Palazzo, 2011;Wettstein, 2009). In this way, the idea of a 'governance without government' (Rosenau & Czempiel, 1992) has been furnished with narratives of the ethical and political. Private actors are portrayed as having an increased sensitivity and effectiveness in guiding problem-solving processes in the socio-political sphere towards win-win outcomes -for instance, by offering benefits to communities in exchange for a social licence to operate (e.g. O'Faircheallaigh, 2013), surpassing a state apparatus that has been deemed passive and rather obsolete (Bernstein & Cashore, 2007;Gemmill & Bamidele-Izu, 2002), especially in the so-called Global South (Morgan, Gomes, & Perez-Aleman, 2016).
While some calls have been made to expand the theoretical base of PCSR (Mäkinen & Kourula;, recognizing the state as a noteworthy actor (Frynas & Stephens, 2015;Whelan, 2012), emerging PCSR approaches often overlook crucial issues related to the state's sustained organized agency. One of these relates to the asymmetries between powerful corporations and those stakeholders directly affected by business conduct (Banerjee, 2017). Empirical research (e.g. Coleman, 2013;Himley, 2014;Li, 2016;Maher, 2018Maher, , 2019Moog, Spicer, & Böhm, 2015) has indicated that, despite their normative coherence, it seems only an ineffective and simulated governance -or, as Walters calls it, 'anti-politics' (2004, p. 33) -is generated when the dialogical and deliberative potential of those who defend public interests via enterprise strategies is undermined. This is because of their managerial dependency on more operationally capable, and thus, more powerful corporations (see Moog et al., 2015;Lemke, 2007). Another, even more significant, issue relates to the underestimation of the state's actual influence, because of PCSR's normative bias and alignment with an 'ideal future', in detriment of a historical assessment of the state's material organization and actual influence (Gond et al., 2011). What is at stake here is the 'demateralization of responsibility' that takes place when value-laden discourses distract the public from the inherently active link between the enduring state and corporations in the face of 'governance gaps' (Djelic & Etchanchu, 2015). How then to elucidate the operational and moral frontiers between governmental agency(ies) and governance networks, which have become ambiguous and blurry?
Our contention is that Foucault's notion of 'governmentality' (Foucault, 2008;Lemke, 2007;Vallentin & Murillo, 2012) allows conceiving the state as operating beyond predefined regulatory, disciplinary roles, producing the discursive and material conditions for self-regulatory governance. Here, it is important, first, to distinguish a practical, material dimension, where the state is involved in legally sanctioning and technologically enabling the activity of a milieu of private counterparts. Second, there is a discursive dimension where the state serves as placeholder for a variety of narratives and values dwelling on the bureaucratic virtues of legitimacy, control and authority, which, if absent, would render the idea of governance meaningless. As Lemke points out (2007, p. 54), Foucault's reading of the state strives to bring back 'an analytics of government' that 'does not start from the assumption that there is an external relationship between government and its objects'. Instead, it recognizes the state to be 'actively involved in constituting agents, identities and interests'; the state's power is both productive and historical, as well as a political process in its own right, rather than merely produced and adhered to specific institutions or polities (Enroth, 2014;Foucault, 2008;Jessop, 2007). In this sense, the state can be conceived as active, yet not always in a direct way (by steering socio-political affairs from the top). It can also act indirectly by organizing and dispensing what Foucault called the 'conduct of (business) conduct' (Vallentin & Murillo, 2012). From this point of view, despite its multiple, diffused embodiments, the state is considered to be constructive of even the most private-driven, 'post-government' governance arrangements (Lemke, 2007).
It is precisely such recognition of 'diffusion' that can help us understand and tackle the problem of the 'dematerialization of responsibility' (Djelic & Etchanchu, 2015, p. 657) in a constructive way. With Foucault, the organization of the state vis-à-vis governance can be analysed in terms of 'govern-mentality', rather than government; as a diffused, post-disciplinary power-making dynamic, productive of the interests or 'mentalities' of all agencies/agents and capacities to influence, whether state, community or corporate bound. This notion of governmentality, which is articulated through symbolic discourses and material practices (Jessop, 2007;Joutsenvirta & Vaara, 2015;Sum, 2009), such as policy, law, enterprise and everyday life (Jones, 2012;Vallentin & Murillo, 2012), can help us account for the emergence of corporate-led self-regulation, which is inextricably bound to the elusiveness of a stable place for statecraft and the diffusion of the state's agency (Jessop, 2007). One should be aware, however, of those voices who 'invoke a nostalgic image of the nation state as an actor defending public interests against powerful economic actors', taking 'for granted the separation of state and market' (Lemke, 2007, p. 56). We can see this 'nostalgia' towards a pre-globalization imaginary of an autonomous, 'muscular' state particularly in the context of the Global South, which has been colonized by this neoliberal logic of individualization and dispersion further and for longer than most other regions in the world (Connell & Dados, 2014).
In the current context of global corporate capitalism, the state enacts its influence mostly through its capacity to trigger, arrange and conduct meanings assigned to diverse technologies of power, objects of governance, governmental projects and modes of political calculation (Jessop, 2007). The state can thus be seen as dispersedly grounded in a multitude of material and semiotic elements and especially in the subjectivity of mediating stakeholders, experts and citizens who are proud to engage freely and committedly in dealing with 'governance gaps'. As Foucault puts it, 'the state is nothing more than the mobile effect of a regime of multiple [liberal] governmentalities ' (2008, p. 77). Based on these premises, it would seem clear that the state would play an active role in the present, not just through the promotion of particular governance and self-regulatory practices, but also through the dissemination of diverse discourses that assign positive meanings to the self-organized overcoming of (potential) socio-environmental harms, encouraging the commitment of stakeholders towards governance (Boelens et al., 2015). Elaborating critically on the work of those who have prophesied the 'return of the state' in a 'post-neoliberal' Latin America (and the Global South) (e.g. Grugel & Riggirozzi, 2012), we propose that a Foucauldian reading can reveal how contemporary governance crises within a deregulated, enterprise-driven society turn most severe precisely when the state actively organizes the passivity and diffusion of its capacity to intervene over the 'public interest' (Ferguson & Gupta, 2002;Swyngedouw, 2005).
We will now explore and analyse such dynamics of governance-making by discussing three cases of mining conflicts in Chile. By drawing on Foucault (2008), we will seek to demonstrate how the state is always present, intervening through diffusely conducting the entrepreneurial governmentality of both state and non-state actors, even if (and precisely because) it appears not to.

Methodology
Considering the exploratory nature of the research focus and the discursive and material orientation of our theoretical framework, our empirical study investigates three contested cases of governance-making in Chile, involving communities, corporations and state actors. The empirical field research took place in Río Jorquera between , in Huasco Valley between 2011 in Caimanes between 2012 and 2017. Our main methodological approach was guided by the general conventions of researching and writing case studies, as proposed by Hays (2004) and Stake (2013).
The focus on governmentality emerged after visiting these three communities several times and triangulating the data with a variety of different documentary and other secondary sources (Stake, 2013) over the course of eight years (2009 to 2017). In line with a Foucauldian approach, triangulation here should be understood as studying social relations of power and influence as a distributed and dispersed set of discourses and practices (Lemke, 2007;Vallentin & Murillo, 2012). This entails studying influences from the bottom up; that is, examining how they permeate throughout society and are co-produced through complex and changing relationships. Thus, studying the historical context is also an important aspect for us to consider, taking into account archival documents and videos in order to get a wider picture of historical relationships (Jessop, 2007). We believe this is precisely what research on governmentality should primarily endorse: a grounded accounting of emerging patterns of reality construction, at both semiotic/symbolic and material/ practical levels.
The field research was undertaken in three mining-affected communities. Table 1 highlights the enduring state by demonstrating the critical incidents when the Chilean state granted environmental licences to the respective companies as well as when it revoked these same licences. Table 2 (see appendix) provides an overview of the three cases in terms of the respective conflicts, mining projects, roles of different actors and CSR arrangements. The data analysis also required substantial amounts of archival data analysis of mostly publicly available documents and video reports. In total we examined 44 documents and 18 videos, totaling 430 minutes. More specifically, for Huasco Valley we analysed 18 documents and four videos; for Caimanes 16 documents and 12 videos; and for Río Jorquera 10 documents and one video. A list of secondary sources (videos, reports and articles) can be obtained from the first author on request. Table 3 (see appendix) provides details pertaining to the field research and interviewees. Table 4 (see appendix) provides background contextual information of the three cases.
The data for this study was collected mostly in direct, opportunistic fashion, as the lead author had unobstructed access to community leaders, elected officials, company officials and state authorities. Notwithstanding, we also made use of publicly available sources such as video documentaries and media reports. As Table 2 (see appendix) shows, the lead author visited Río Jorquera and Caimanes four times and the Huasco Valley on two occasions. We moreover, conducted desktop research periodically on the three cases between 2009 and 2017, focusing on perceptions of governance, role of the state and any significant instances of trying to explain community positions of conflict or good relations with the nearby mining projects.
In our collection of data, we combined case study fieldwork with desktop/secondary research. The fieldwork included semi-structured and informal opportunistic interviews with community residents, mining company executives, local government representatives, state prosecutors, civil society actors and independent experts. Background research around the mining company's community relations was conducted using diverse documentary and press sources, in order to prepare and contextualize the interview process. Most of the interviews were audio recorded and later transcribed in their original language. For the interviews we were not able to audio record, upon the preference of interviewees the lead author took notes throughout the conversation, which were then typed up within 24 hours. The interviewees' names have all been anonymized for confidentiality purposes.
Once transcribed, the lead author began conducting initial thematic codification, which he shared with key interviewees from the different case studies so as to validate his interpretations. Later on, co-authors coded a sample of the transcripts, reaching a level of 80% intercoder reliability on the first-order concepts. The first level of coding of the interviews was done separately for each case, generally following the critical incident analysis technique as set out in Hays (2004). This exercise generated a total of 138 first-order codes of a more descriptive nature. Further analysis led to aggregating categories into themes based on similarities and differences (see Figure 1). Table 5 (see appendix) provides a list of first-order concepts categorized by the five second-order themes, all supported by examples of raw data. We were then able to construct narratives that reflected interviewees' experiences of each theme.
We justify the use of the three mining cases as each one has experienced a different outcome with regard to the deployment of CSR and the role of the state. Similar approaches of using more than single case studies for analysing mining-community conflicts have been employed by Bebbington et al. (2008) who contrast two cases from Peru and Ecuador to demonstrate how in Ecuador community resistance can prevail over mining projects. In a similar vein, Özen and Özen (2017) compare and contrast two gold mining community cases in Turkey to explain why resistance triumphed in one case and failed in the other. We argue that our three cases also contain different narratives, appropriate for subsequent theorization.

Three Mining Conflicts in Chile
In this section, we describe governance engagements held between private actors, governmental institutions and local communities in the context of the mining industry in northern Chile. In particular, we account for the interaction between forms of business conduct, administrative-bureaucratic sanctioning and activism, invoked and deployed by these three actors across three different cases of mining operations, which we have labelled 'Huasco Valley', 'Caimanes' and 'Río Jorquera' (see Table 4 in the appendix for contextual background of the three cases). Following the conceptualization of Lukes (2005), the governance engagements we are interested in can be described as complex processes of influence, in which actors invest in, or become the object of, discursive (i.e. symbolic-rhetoric and practical) strategies, aiming at generating and/or consolidating a position of power in relation to other actors. We identify five different modes of influence, running across all three cases, as shown in Figure 1. First, 'deployment of CSR initiatives' encompasses communications and practices that seek to solve mining conflicts by creating a consensus between actors. Second, 'mobilization of resistance' points to how stakeholders and the general public are alerted about the dangers and unfairness associated with the advancement of mining operations. Third, 'division of resistance' refers to the effort of diminishing the coherence of discourses against the advancement of mining operations, while, at the same time, promoting a consensus of the value of mining. Fourth, 'prescription by the state' alludes to the way in which the state's prescriptive capacity is invoked, demanded and/or carried out by diverse actors, including the state's own organizations. Finally, 'nexus between state and enterprise' refers to the communications and practices, sometimes carried out in secretive fashion away from the view of the public, that aim at situating government as supportive of, or closely aligned with, enterprise initiatives. In what follows we will further characterize these modes of influence, by providing details of key events, testimonies and observations, including some quotes for the purpose of illustration.

Mode of influence 1: Deployment of CSR initiatives
CSR initiatives were deployed by the mining companies in all three cases from the beginning of the conflicts to the present day. CSR was widely used as a key communication tool, aiming to build good community relations and highlighting the benefits of mining. In 2014, 15 out of 22 neighbourhood associations in Alto del Carmen in the Huasco Valley voted in favour of dialogue with Barrick Gold, which led to selected community leaders engaging regularly in meetings with company executives to discuss impacts of the mine and to exchange technical data. Barrick's community relations employees have participated in more than 1,000 community meetings during that time, including a comprehensive door-to-door education campaign (Barrickbeyondborders, 2017). In the other cases, too, similar CSR communication techniques were used, including distributing pamphlets among community members in Caimanes, which highlighted the benefits and safety features of the dam (OLCA, 2004), and a sustainability workshop conducted by the mining company in Río Jorquera in 2009.
Yet, CSR communications also extended to sharing cultural identifications. Barrick, for example, invested in the publication of a book called Diaguita, whose main message was that the Diaguita people are historically a mining people in the Huasco Valley, delegitimizing detractors' claims about mining not being compatible with the rural, agricultural identity of the valley. Other similar instances were found in the case of Caimanes (a non-indigenous community), where Pelambres funded the construction of a 'medialuna' (a crescent-shaped corral used for rodeos), and in the case of Río Jorquera, where Kinross Gold invested in the devising of an 'ethno-mapping study' with the assistance of the community to identify areas of special interest (Kinross, 2017).
While appreciated by some actors, these efforts to appropriate cultural representations were bitterly criticized by members of the community. An interviewee at Caimanes, for instance, affirmed that the medialuna did not represent their culture or identity: 'They built a medialuna for us! They think we are "huasos" but huasos are from south of Santiago; we are in the north! That is offensive and an insult! We don't do rodeos here' (interviewee, Caimanes community). Meanwhile, in the case of Huasco Valley, local organizations raised their voice against the inadequacy of initiatives aiming at cultural promotion: At Río Jorquera, Kinross has implemented a range of sustainability and ethno/eco-tourism projects in dialogue with the local Colla people, also reaching an agreement over access to their land. The company donated fences and paid for health insurance subsidies and modest college scholarships to some Colla residents, as well as offering competency-building courses, such as cake baking. Some of these efforts were received sceptically by the community, who felt that their main grievances concerning the dust, spills and accidents caused by mining traffic on the road had not been addressed. When asked about his views of the CSR initiatives, the Colla leader said: 'They take with one hand and give with another' (interviewee, Community leader of Río Jorquera).
The data indicates that CSR initiatives across all three cases also revolved around the theme of 'conflict resolution', aiming to reach and legitimize settlements between companies and communities that would keep business going within a more welcoming, friendlier environment: We were bored and tired of always being in conflict with the community over so many years, and the community was also bored and tired of this. We realized we had to change the way we related to one another. Our relationship played out in the courts of law and that's not sustainable for anybody. It's exhausting for everyone! (interviewee, Pelambres management) A similar strategy was adopted by Kinross Maricunga in Río Jorquera between 2005 and 2008. In this case, the company reached an agreement with the local community, which helped structure their mutual governance relation (interviewee, CSR official at Kinross). This consisted mainly of a Dialogue Table that was to serve as an instance for the mutual cooperation between Kinross and the Colla community, guiding assistance in areas of education, health and entrepreneurship.

Mode of influence 2: Mobilization of resistance
Despite the deployment of various CSR initiatives, resistance against the mining companies has been mobilized at various levels. Community leaders from the Huasco Valley told us similar stories about the beginnings of the local resistance movement, which was instigated in 2000 by the parish church, a nun and a national environmental justice NGO -OLCA. After charismatic community leader 'Lucho' read the Environmental Impact Assessment report to the community, a local nun took charge of assembling the people to devise a resistance effort. Afterwards 'Lucho' went to the Chilean capital of Santiago to visit OLCA, which agreed to second one of their staff members to the community to raise awareness around environmental impacts by sharing the testimonies of invited Canadian speakers and showing relevant films (interviewees, local residents, Huasco Valley). Also, as seen in many documents by filmmakers, journalists and researchers, church representatives began to hold Sunday Mass by the local river, as part of an effort to assign symbolic value to the water sources threatened by the mining operations of Barrick.
In Caimanes resistance began in 2008 when elected leaders of the defence committee decided to take a much more aggressive stance against Pelambres. The campaign included filming video documentaries, holding street protests, roadblocks and the hanging of anti-Pelambres street signs and black flags at homes. In late 2010, twelve community members barricaded themselves into a local school and went on hunger strike for 81 days demanding the relocation of the whole Caimanes village or the demolition of the tailings dam. Five years later, further resistance emerged when a large group, led by the defence committee, blocked the access road to the mine. This protest, which went on for 76 days, was an expression of grievance, pointing to the fact that nothing had happened two months after a supreme court ruling, ordering Pelambres to demolish the tailings dam (interviewees, various community members and video documentaries).
In Río Jorquera, too, resistance was very active, especially in the early years. Roadblocks, and later on legal channels, had been used by the Colla community since 2004 to protest against the spills, accidents and dust caused by mining-related traffic. When we asked Colla leaders if they could choose between a life with the mining companies and their CSR initiatives and one without any mining operations and the significant CSR investments, the response was an unequivocal and immediate preference for life without mining and CSR (interviewees, Colla Community leaders).
The data shows that resistance was also mobilized at a more institutional level. At Río Jorquera, the Colla complaint of dust plumes caused by mining-related vehicles was investigated by COREMA, a regional environmental authority, which found insufficient evidence to support the claims (various interviews with community, authorities and Kinross). It is important to note that in all cases the resistance actions became more institutionalized as lawyers became involved. In Río Jorquera, the community were joined by a woman lawyer who, in 2009, managed to take Kinross to court over the road issue. At Huasco Valley, between 2012 and 2017, there were multiple legal challenges by community groups backed by NGOs and lawyers with the aim of stopping the mining operations by Barrick. The defence committee at Caimanes also decided upon a legal approach in 2008 when hiring Ossa and company lawyers on a 'no win no fee' basis. However, Pelambres reacted institutionally in 2012 when the company decided to countersue defence committee leader Cristián Flores and his lawyers, a case the company lost in 2013 (press reports).
Resistance also expanded on to the international scene in various ways. Demonstrations against the Pascua Lama mine were held in London, Cambridge and Barcelona in 2006. In the case of Huasco Valley, a local indigenous leader travelled to Toronto, invited by Canadian NGO 'Protest Barrick', to give talks and presentations at universities, rallies and at Barrick Gold's annual general meeting (AGM) in Toronto. A small community from Huascoaltino also had its case against Barrick accepted at the Interamerican Commission for Human Rights (ProtestBarrick, 2010). After Barrick's Memorandum of Understanding (MoU) agreement with the Huasco Valley community, Canada's largest anti-mining NGO published a report delegitimizing the process of arriving at the agreement (Wiebe, 2015). In similar fashion, two of the hunger strikers at Caimanes were invited to France by the French government and the Mitterand Foundation to receive honorary medals for their bravery in defending their territory and water (interviewees and press reports). Since 2013, London Mining Network has been engaging in civil disobedience via protests outside the AGM in London of Antofagasta Minerals (the owner of Pelambres). Finally, the Colla community of Río Jorquera were recognized by UNESCO in 2010 and have consequently drawn on this fact to defend their position against large mining companies.

Mode of influence 3: Division of resistance
In contrast to the mobilization of resistance we observed, the data across the cases indicates widespread efforts by mining companies to divide resisting communities, ultimately leading to the latter's splits. It is important to note that the data is here showing something different than CSR, as the mode of influence at play does not rely on the capacity to benefit or repair but rather on the capacity to instigate an internal conflict among resisting factions. Specifically, the data indicates two distinct ways through which the division of resistance has been enacted.
First, efforts to divide resistance revolved around targeted incentives that were offered to resisting actors. Barrick Gold, for example, decided in 2008 to invest in one of their then main critical opponents, the Water Vigilance Board, which has been considered as one of the critical events that weakened local resistance to the mining project. The Board, made up of mainly large, wealthy farmers who had been opposed to Pascua Lama due to the impacts on their grape crops, agreed not to oppose the project upon receiving US$65 million. Similar efforts were observed in the case of Caimanes. In 2003, Pelambres hosted those community residents who were in favour of the company, including those who had received CSR benefits, in order to show that the company enjoyed local support, while, in fact, a large proportion of the community opposed the mining project (OLCA, 2004).
Second, offerings of targeted incentives led resistors to face the dilemma of having to decide whether to persist in their disruptive activities or to try to settle with the companies. An example stems from 2014 when a video was published, showing a local female leader in the village of Perales at Huasco Valley shouting at other community members on the day of the vote on the matter of a MoU with Barrick. The woman, who is in favour of signing the MoU, irately accuses a young man of being an activist and terrorist who does not allow the local 'humble' people to sign the MoU. She repeatedly asks who funds him. The video also shows a male local resident who seems to be denied entrance to the meeting because he asks too many questions. This man, together with a local councillor, claims that the woman is there to coerce local residents to sign the MoU with Barrick and that she does not allow the residents to think freely (Convenio entre Diaguitas y Barrick Gold, 2014).
At Caimanes, diverse efforts to divide resistance were observed led by the mining company, aimed at taking advantage of deepening rifts in the allegiances between neighbours and family members. This was evident in the result of a referendum devised by the company in 2015, in which the community had to decide whether or not to seek a settlement. We found instances of families being torn apart because one member voted 'yes' while others voted 'no', sometimes even leading to evictions from family homes. This sense of inner division kept growing following the referendum, as the lawyers of the Caimanes defence committee, who had defeated the company in 2014 at the supreme court, decided to engage in dialogue with Pelambres and ended up being paid approximately US$4.6m (after the community voted in favour of a settlement with the company). As one community member explained emotionally: Now many siblings no longer speak to one another, those who voted for the money do not look at us in the eye anymore, I don't know why, I haven't done anything to them …On top of that my kids and I were evicted from the house I was renting because the landlord said he can't rent out to a traitor like me … Now I'm here in this wooden shack without windows. (interviewee, member of Caimanes defence committee)

Mode of influence 4: Prescription by the state
The data indicates distinct efforts by actors across the three cases to invoke and/or implement the capacity of state organizations to intervene directly -that is, prescriptively -over conflicts about mining operations and matters of governance around them. Here, the data reveals two levels at which the state's prescriptive influence is deployed, namely, the executive, governmental level, and the judicial level.
Interventions by the executive branch of the state were identified in all three cases. These revolved, in the first instance, around issues of legal compliance and infringement of policy instruments, such as the Environmental Impact Assessment report.
For example, At Río Jorquera in September of 2013, the SMA (environmental regulatory government agency) issued a 'notice of violation' to Kinross, related to deviations from the permitted configurations of the Maricunga mine camp, fining the company $4.6 million. Kinross appealed the sanction to no avail (Kinross, 2017). Later, in March 2016, after conducting another site visit, the SMA ordered the closure of Kinross' mine due to the drying up of at least 70 hectares of wetlands between the Negro Francisco and Santa Rosa lakes, which they blamed on the Maricunga gold mining project (SMA, 2016). Then a resolution was implemented, seeking the closure of Maricunga's water pumping wells based on allegations of irreparable harm and imminent risk to the Valle Ancho wetland, located approximately 7 kilometres from Kinross' groundwater wells (SMA website, 2017). Meanwhile, in the case of Caimanes the Chilean Water Authority, an administrative agency at central government, ruled to authorize the necessary water permits needed for the El Mauro tailings dam. The dam would contain 1.7 billion tonnes of toxic mining waste and destroy 140 archeological sites (OLCA, 2004). Related to this, it is worth noting that in 2012 the Minister of Health brought a group of journalists to Caimanes so they could report his drinking of a glass full of water sourced locally. The Minister began to be heckled right after this gesture was captured and left the scene swiftly (Chile se moviliza, video documentary, 2013).
Yet, efforts to implement the intervention of the executive branch of the state also involved the use of force via police action. For example, at Caimanes the police broke up a roadblock with the use of riot control gear, and then demolished the campsite in March 2015, on day 76 of protest. The roadblock was organized in protest at the absence of any action after the supreme court ruling in October in 2014 requesting Pelambres to disassemble the El Mauro tailings dam. Members of the community told us that such intervention was really a form of protecting the interests of the company instead of complying with the court order to dismantle the El Mauro dam (interviewees, Caimanes community).
Interventions by the judicial branch of the state were also identified. In the case of Caimanes, judicial influence over the years can be seen as active, yet full of oscillations and contradictions. At first, in 2006, the Regional Court of Appeals prohibited the construction of the tailings dam, following the submission of several legal challenges by the community. However, in 2008 the courts reapproved the permits for the construction of the El Mauro tailings dam (Chile se moviliza, video documentary, 2013). In 2013, again the state acted judicially to curtail the reputation and progress of mining operations, as the Tribunal of Ovalle court found the leader of the Caimanes defence committee (Cristián Flores) and the lawyers who counselled him (the Ossa firm) not guilty of illicit collusion, following an accusation by Pelambres. In October 2014, the Supreme Court of Los Vilos ordered Pelambres to demolish its El Mauro tailings dam, deeming it a danger for humankind (El Mercurio, 2013). However, the state went back to acting judicially in support of mining operations, as in August of 2016 the La Serena Court of Appeals reversed the decision after seeing evidence of CSR and an agreement with the community between Pelambres and Caimanes (two interviewees, management officials Pelambres and Court of Appeals, La Serena ruling document).

Mode of influence 5: Nexus between state and enterprise
This mode of influence emerges when the actual or potential interventions of state and enterprise actors are represented as compatible. Here, again, actors can be observed invoking or demanding the active involvement of the state. Previously, we have noted how corporations have relied legally and politically on statecraft to license mining operations; here, however, the focus is placed on indirect relations between the state and other actors, particularly in those through which the support for the actions is generated. In this sense, we contend influence can be observed as reliant on a verifiable nexus between state and enterprise. The data indicates two levels at which this nexus is established, namely, communications/public relations and strategy/policy making.
A background nexus between state and enterprise, in terms of communications/public relations, can be appreciated as early as 1996, when the Chilean and Argentine states agreed to declare the borderline territory at the top of the Andes as 'bi-nominal'. This was done in anticipation of future mega mining activity by Barrick Gold (La Nación, 1996). In 2009, the regional director of the Chilean Water Authority, a central government agency, declared in an interview with the authors that 'out of all the many mining companies here, I think Kinross' self-monitoring, reporting and conduct on water usage is exemplary' (Region III, DGA regional director). Similarly, in the context of mining at Huasco Valley, Salinas (2007) accounts for the public interventions of Member of Parliament Jaime Mulet, who, in 2004, declared his delight at seeing Barrick's decision to go ahead with the Pascua Lama mining project. Equally, in 2006, the Chilean government's official spokesman, Osvaldo Puccio, declared the country was the 'overall winner' through Barrick's actions.
Meetings were also held at the Ministry of Public Works, between government officials, Kinross Gold executives and community leaders, to discuss conflicts over road-related grievances, such as dust and accidents to Colla livestock (observant participation in meetings by authors, 2009). In the case of Caimanes, Adriana Hoffman, the then head of environmental authorities, organized an unprecedented visit to the community along with a group of politicians in 2001, in which she publicly addressed the conflict, calling for more dialogue between mining company and community (OLCA, 2004). According to Kinross's head of CSR in Chile, the relationship with the Colla community constitutes a 'best practice' example, due to the efficacy of the governance arrangement that involved the Ministry of Public Works in generating benefits for the parties involved (interviewee, Kinross management official).
Communicational legitimation as a means to portray state and enterprise interests as compatible can also be appreciated in efforts to censor. This was identified in the case of Caimanes where the state-owned television network edited the content of a documentary in which critical remarks were made about the El Mauro dam and Pelambres. In the televised version multiple references to 'the situation' and 'drought' made by well-known rap star DJ Mendez during his visit to Caimanes were shown, while critical views, which linked water shortages to the company, were removed. Community residents confirmed that their consistent mentioning of Pelambres to DJ Mendez during recording of the documentary had been edited out, including the helicopter visit he made over the tailings dam (TVN documentary, 2012 and various community interviewees).
Finally, in the context of Caimanes, the articulation of nexus between state and enterprise in terms of policy/strategy making can be appreciated in all its subtlety, as an indirect and rather intricate effort. Here, the local community resisting the mining project appeared to be perplexed at how the judicial decision of 2014 to remove the tailings dam had been followed 14 months later by a business-as-usual stance by the company, which continued to operate as if nothing had happened. At this point, crucially, the community was offered a referendum to decide the future of the tailings dam in the area, which, despite having been organized by the mining company, was staged as an official democratic electionary event, as if organized by the state. Furthermore, the company then used the pro-mining result of this pseudo-electoral process as evidence within judicial dealings with the state, in order to overturn decisions against the tailings dam. This series of events indicates the state's complicity in enterprising this scheme. Not only was the state keen to admit the results of a corporate-run referendum into the judicial process, and remained passive while the pseudoelectoral process in the name of government was taking place; it was also keen to re-engage executively with the company, after the pathway for settlement had been legitimized as a result of the referendum. As a former resistance leader at Caimanes explains: "This is all the capitalist state's fault, they put that piece of trash there (referring to the dam) and then left us alone, to fight the company alone; the state loves private companies coming in. We…fought the company alone and got the agreement for ourselves, and now the state wants to come in and be in the first row ready to cut the ribbon for the new projects we get!"

Discussion
The above findings have led us to appreciate how the making of governance between conflicting actors -corporations, communities and the state -unfolds densely yet dispersedly, through the establishment of 'new polycentric forms of social organization and mobilization' (Boelens et al., 2015, p. 281). Rather than relying on a singular logic of resolution, the exchange between stakeholders -both large and small -appears to hinge on diverse modes of influence, which are deployed during practical interplay in the field. In light of our data analysis, influence can be understood as the capacity to exert power in interactive fashion, by invoking and enacting discourses on private interests, regulatory authority and technical knowledge.
Yet, what seems to matter most with regard to the making of governance in our three cases are the differing ways in which influence is enabled; or in Foucauldian terms, the ways in which conduct is conducted. What is, analytically speaking, at stake here are the overarching logics that grant coherence to the autonomous efforts that hold sway over governance processes, taking place at heterogenous centers of conflict (Lemke, 2007). Across the cases -Huasco Valley, Caimanes and Río Jorquera -we have found five such modes of enabling influence: (1) corporate social responsibility (CSR); (2) mobilization of resistance; (3) division of resistance; (4) prescription by the state; and (5) nexus between state and enterprise. These modes represent distinct and complementary approaches to the making of governance, portraying, as a whole, the multifaceted character of stakeholder engagement in the resolution of problems and the crafting of agreements. Overall, following Foucault (2008), they show how the power to govern is to be read 'productively', beyond the logic of authority/compliance: governance is actively 'produced' only when key actors manage to legitimize their actions, practically and symbolically, as influential, in relation to the interests, expectations and actions of others.
By foregrounding the crucial role that the enabling of influence plays in governance-making, we are able to witness the diffusion of the state's agency and its ambiguous presence as both a mediator and regulator amid corporations, communities and the general public. Confirming the views of several OS scholars (Bernstein & Cashore, 2007;Rhodes, 2007), our study shows how the organization of the state seems to have veered committedly towards a post-bureaucratic, post-disciplinary logic for (reciprocal) co-governing between stakeholders. Our data often seems to corroborate that the state appears to have withdrawn almost completely from the scene of governance so that a looser, less direct, approach to the policing and resolution of conflicts can be enacted. In the literature, such shift has been associated with the emergence of a 'bazaar' type of governance, meaning a situated interaction where courses of action are negotiated, not through the coercive force of hierarchic authority -what has been called 'cathedral' governance -but through the reciprocal, rather diffused definition of value(s), meaning(s) and networked allegiances (Boelens et al., 2015;Demil & Lecocq, 2006;Depledge & Dodds, 2017).
Notwithstanding, what we see in our cases of mining conflicts in Chile goes well beyond the characteristics of 'bazaar' governance, which has also been identified as 'private regulation' (Djelic & Sahlin-Andersson, 2006;Rhodes, 2007). Our data clearly shows that the Chilean state is present. That is, we identify different levels of state organization and involvement in the complex web through which state, enterprise and community actors attempt to (co-)establish grounds for influencing their counterparts. So, the state seems to have retained some of its 'cathedral', capacity to influence directly, unilaterally and visibly by relying on legal and administrative mechanisms available; for instance, through judicial ruling, contractual oversight and the command of security forces at the mining sites. At the same time, however, the small influence that the state has been able to preserve seems, ultimately, to support the consolidation of a mining governance 'bazaar', where a multitude of micro and macro exchanges between corporations and community actors are deployed, on the grounds of highly deregulated land/transport rights, social services and cultural heritage. Although noticeable at times, the state's influence proves timid and dispersed, as control over these crucial socio-economic matters is surrendered to the reciprocal influencing between private interests either partially (i.e. through enabling state-enterprise 'nexus') or completely (i.e. through deploying, and resisting, CSR initiatives).
While several scholars would consider these findings to indicate the unequivocal irrelevance and even the obsolescence of the state as a relevant player in the multi-stakeholder scene of governance (Rosenau & Czempiel, 1992;Scherer & Palazzo, 2011), a careful reading of the data allows us to appreciate how, despite the diffusing of its visible agency, the state is, in fact, never absent. Delving deeper into Foucault's definition of the contemporary state as 'nothing more than the mobile effect of a regime of multiple governmentalities' (Foucault, 2008, p. 77), our analysis has led us to interpret two crucial features of state organization. First, we see statecraft as a contingently enacted, multi-layered technology of administration, which can either be invoked and mobilized or, instead, be disregarded by actors depending on their stakes or 'mentalities' regarding governance affairs and the circumstances of particular conflicts. Second, we appreciate how, rather than vanishing from the reality of governance-making, the institutionality of the state emerges in dynamic, dispersed and indirect fashion. That is, the state, understood as 'governmentality(ies)', is about the configuring of a particular mode of agency, which oscillates between active, direct involvement and passive, indirect vouching of the agency of others, often in implicit, covert ways.
Accordingly, elaborating on Djelic and Etchanchu's (2015) recent insights on governance and the history of neoliberal governmentality, we interpret our data in terms of the 'materialization' and 'dematerialization' of the state's responsibility over discourses and practices of governance. In line with Foucault's project, this represents an attempt to transcend dichotomies, such as regulationderegulation or government-governance, in search of a deeper account of how the state is organized (and organizes itself) not only to govern but also to produce, historically, the conditions for governance to take place. By looking at diverse modes through which influence is enabled -what Foucault called 'governmentality' or the 'conduct of conduct' -we can discern the way in which 'self-government and the will to self-determination is always already prefigured by "regimes of truth" which shape and guide such behaviour' (Garland, 1999, p. 29). In this case, these 'regimes' not only concern the truth of an active, direct invoking of state intervention in social life by the governed, but also the truth of 'freedom' during highly deregulated 'bazaar' governance-making engagements: the indirect, implicit appeal to legal-administrative guarantees and infrastructures that make stakeholder subjectivity and private interests possible at the scene of governance (Jessop, 2007). We realize it is with and not without the state's remaining 'cathedral' rooting that the freedom to govern is realized. This means that governance exchanges between private interests do not surpass but rather reproduce and disperse the state's historical discourses on the forceful need to govern (Vallentin & Murillo, 2012). What is really at stake, we then infer, are the specific configurations of what Boelens et al. (2015) have called a 'hybrid cathedral-bazaar' governance; Figure 2 depicts such a continuum.
When seen as part of a relational system, the modes of influence are particular arrays of 'governmentalities', that is, shared representations of actors around the conditions for the accomplishment of governance-making, registered at both discursive-symbolic and material-practical levels. Crucially, these 'governmentalities' oscillate between 'materialized' and 'dematerialized' forms of state responsibility (see Figure 2). Different modes of securing and deploying influence can be perceived across the three cases observed, ranging from the rigorous implementation of regulatory and policing statecraft, on behalf of traditional 'cathedral' government authority, which we have equated to state prescription (mode of influence 4), to the devising of autonomist, private initiatives within a 'bazaar' framework, which we have equated to CSR (mode of influence 1). These poles represent stereotypical cases of governmentalities, in which the construal of the state's due involvement, or lack thereof, is depicted most purely and schematically.
On the one hand, the state comes to be construed as committedly 'material' in its prescriptive involvement over public affairs (mode of influence 4). The data shows how, in the face of conflict between corporations and communities, the state is organized locally and nationally to carry out a responsive and rather paternalistic governmental oversight, capable of informing not only direct regulation in the form of sanctions but also concrete actions at ground level, such as the mobilization of police forces. Notably, the state is here organized as an accountable presence, actively involved in the steering of private initiative(s) and perceived as causally and morally responsible for the effects of such intervention.
On the other hand, in the case of CSR (mode of influence 1), the state is construed as completely 'dematerialized', with its discursive and practical presence reduced to mere 'virtual' concerns and objects, and its authority reduced to an elusive, if not outright absent, locus of responsibility(ies) (Djelic & Etchanchu, 2015). Notable images of such 'dematerialization' are found in CSR initiatives aimed at 'facilitating' the reconstruction of local, indigenous cultures by communities, and at 'supplementing' the development of social services and essential infrastructure. The most remarkable in these instances, the data shows, is the naturalized fashion in which issues that have been historically paramount for state administrations, such as the constitutional recognition of indigenous identity and the continued provision of welfare, have been 'dissolved' into events or artifacts of vague significance. Books on indigenous heritage and fences and other forms of small-scale infrastructure, for instance, have supposedly come to accomplish what the nation-state once managed to organize. Yet, they can only do so in a blatantly nonspecific, indeterminate manner. A sustained addressing of indigenous culture is replaced by brief publications, while a commitment to dealing with social needs is replaced by the offering of 'solutions' whose effective implementation and evaluation no particular corporate actor or 'shareholder' will assume full responsibility for. What these CSR images show, in the context of conflicts around mining projects, is the hollowing out of the state's agency, up to the point of 'dematerializing' its relevance in the governance of public affairs.
Our data analysis has led us to conclude that the remaining arrays of 'governmentalities' (modes of influence 2, 3 and 5) can be situated along the continuum between poles of 'materialization/ dematerialization' of the state's responsibility, understood as expressing varying degrees of active state involvement in governance making.
Overall, we come to realize the enduring role the state plays in governance-making. Even in the case of CSR, the data indicates, references to public interests under the tutelage of state administration, such as culture and commons, cannot help but be articulated, revealing the 'background function' the state provides for the interaction between stakeholders. Hence, the state's agency is more or less materialized in all modes of influence, demonstrating the polycentric, multifaceted character of its organization. In some cases, statecraft is directly invoked and wielded as a concrete tool, while in others the concerns and attributions of the state remain a 'virtual' reference at a discursive level. Yet, the state's agency proves essential in the practical and symbolic enabling of influence around mining conflicts and a steady, enduring staple in the conduct of governance conduct.
We believe this not only debunks simplistic claims about 'statelessness' in the long history of business-society relations, but also illustrates exaggerated assumptions about the waning of state authority in our current landscape of neoliberal governance (Djelic & Etchanchu, 2015). Our study of the varied 'governmentalities' in which the Chilean neoliberal state is implicated during the making of governance shows how 'virtual' and 'dematerialized' forms of state responsibility do not exclude but actually coalesce with the 'materialized' involvements that the state has traditionally embodied. It can thus be inferred that contemporary forms of state organization are not so much about dismissing and discarding authority as they are about diffusing it amid a number of actors willing to subjectively interpret and act upon the need to govern, | all the while retaining a minimal 'background capacity' to guarantee the conditions for such free exchange to occur.
Boelens et al.'s concept of a 'hybrid cathedral-bazaar' governance (2015) signals this new generalized yet customized 'governmentality', accounting for the irresolvable tension between market-based governance and what could be called the 'governance of governance', which state organization accomplishes, practically and symbolically, in dispersed fashion. In Figure 3, we have used Boelens et al.'s concept to finally locate the specific relations between business, community and state actors which we have interpreted in our analysis of data across three cases, showing the specific ways in which modes of influence enact 'cathedral' and 'bazaar' logics or 'mentalities' of governance. Crucially, such schematic rendition allows appreciating at a glance how different modes of influence come together as a whole to establish a general, multifaceted 'governmentality' where the state organizes its own dispersion in order to endure. Figure 3 can thus be read as a map, showing the specific junctures at which the 'materialization' or 'dematerialization' of the state's responsibility take form (see Figure 2). Consequently, it can also be read as a guide for agents such as community leaders and state officials, who are often severely disempowered in relation to corporate actors, to identify the ambiguous boundaries between 'bazaar'-compatible (i.e. dematerialized) and 'cathedral'-compatible (i.e. materialized) modes of influence. Hopefully, this could lead to a more effective seizing of both the resources and gaps in state organization, leading to a fairer and more inclusive conduct of governance.

Conclusion
In this article we address the profound ambiguity of the Chilean state in its influence over the governance of mining industry. We have shown that the governance of the Chilean mining industry is dispersed, involving a range of state, corporate and community actors. In particular, Foucault's notion of governmentality (Foucault, 2008;Lemke, 2007;Vallentin & Murillo, 2012) has led us to conceptualize the role of the organized state as being neither the steerer of socio-economic affairs, nor a diminished figure in a 'post-national constellation', but the vehicle for the 'conduct of business conduct', within the era of governance (Enroth, 2014). Moving beyond normative and stateless PCSR assumptions about communicative rationality(ies) at the political level (Scherer & Palazzo, 2011), we call for the material embodiments of governmentality to be studied further. In particular, following Boelens et al.'s lead (2015), research should explore: first, how the 'polycentric' organization of the state blurs the practical and discursive dichotomy between 'cathedral' institutions of state regulation and fragmented 'bazaar' exchanges between private enterprising actors; and second, how PCSR strategies are less about filling governance gaps vis-à-vis a waning state (Scherer & Palazzo, 2011) and more about private actors establishing a sufficiently stable mode of influencing, in direct response to the influence they attribute to state actors, often operating 'behind the scenes' (Moog et al., 2015;Coleman, 2013;Himley, 2014;Li, 2016;Maher, 2018).
In relation to the OS field, our analysis has led us to agree, first and foremost, with calls to empirically explore the role of the state in private governance regimes (e.g. Gond et al., 2011;Schrempf-Stirling et al., 2016) Particularly, the discursive and material realities of governance relations, here exemplified in the economic, cultural and environmental impacts of mining endeavours, must be addressed. Second, we conclude that it is of utmost importance to assume the state not as a conceptual 'add-on' to governance-making but as a complex social relation or 'governmentality', which unfolds historically and territorially within a widespread neoliberal political economy, involving not one but multiple actors with diverse stakes (du Gay, 2000;Jessop, 2007;Lemke, 2007;Vallentin & Murillo, 2012). Third, our analysis of three cases in Chile reveals that the state has not disappeared but turned ambiguous and dispersed, becoming a 'polycentric hybrid' between 'cathedral' and 'bazaar'. While demanding the withdrawal of governmental control and regulation (Nem Singh, 2012;Undurraga, 2015), the state demonstrates an economic, political and legal-bureaucratic involvement in the regulation of mining vis-à-vis the people and the environment, through both the materialization and the dematerialization of its responsibility (Djelic & Etchanchu, 2015). Overall, we have shown, both empirically and theoretically, that the state endures, making its presence felt in a variety of ways. As Nettl wisely affirmed of the state, as early as 1968: 'It retains a skeletal and ghostly existence, largely because, for all the changes in emphasis and interests of research, the thing exists and no amount of conceptual restructuring can dissolve it' (Nettl, 1968, p. 559). OS scholars and governance-makers would do well to recognize its importance and further analyse its contemporary organization.

Funding
This article was partly made possible by the European Union Marie Curie Individual Fellowship, project number 707485.

Supplemental material
Supplemental material for this article is available online.