A Road to Conflict: Stakeholder’s and Social Network Analysis of the Media Portrayals of a Social-Environmental Conflict in Bolivia

Abstract Society’s understanding of a conflict is mediated by information provided in mass media, for which researchers stress the importance of analyzing media portrays of stakeholders in a conflict. We analyze information from the Bolivian press regarding the construction of a road crossing the Isiboro-Sécure Indigenous Territory and National Park (TIPNIS). Using stakeholder’s and social network analyses, we explore stakeholder’s positions and alliances as represented in the media and contrast it with previous scholarly work. We found that some actors cited as central in scholar analyses of the conflict are largely absent in the media (e.g., private investors, conservationist sector) and that the media tend to present stakeholders as having more homogeneous positions than the academic literature does while also neglecting some important alliances in their account. The media also suggests that Indigenous communities are forging stronger alliances with urban sectors and civil society, alliances not stressed by researchers.


Introduction
Areas targeted for road expansion face important governance challenges and constitute a prominent breeding ground for social-environmental conflicts, particularly in the Global South (Martinez-Alier et al. 2016;Raftopoulos 2017). Indeed, infrastructure expansion often involves a multiplicity of actors with competing interests, operating at different scales, and with different opportunities to influence decision-making, which generates an intricate network of stakeholders with opposed views, power differences, and conflictive social interactions (Del Bene, Scheidel, and Temper 2018; P erez-Rinc on, Vargas-Morales, and Crespo-Mar ın 2018; Scheidel et al. 2018;Temper et al. 2018aTemper et al. , 2018b. Understanding how these stakeholders relate to each other is critical for conflict resolution. Because society's understanding of the network of stakeholders involved in conflicts is largely mediated by information provided in mass media (Slone 2000;Evans 2010), conflict research has highlighted the importance of analyzing the role of media in shaping inclusive governance (Boyle and Schmierbach 2009;Hamelink 2011).
Conflict research suggests that the media should not be understood as a medium for the transmission of neutral information but rather as a tool that decision-makers employ in order to promote their own agendas (e.g., Sampedro Blanco 1997;Stanig 2015; Van Aelst and Walgrave 2017), or as an independent actor that creates pressures for action on issues it deems newsworthy (Patterson 1997;Eberwein, Porlezza, and Splendore 2016). On the one side, the control of media by actors with a stake in political affairs often results in selection, coverage, or statement biases in the framing of conflicts, contravening the highest standards of journalistic ethics (D'Alessio & Allen 2000). Such control might explain the larger visibility of those stakeholders with an agenda favorable to media, corporate, or political interests around a particular issue (e.g., Groeling 2013;Eberl, Boomgaarden, and Wagner 2015). On the other side, the media might also take a vested interest on particular conflicts (Carroll and Hackett 2006;Malinick, Tindall, and Diani 2013) and be more likely to report extreme positions with a high marketing potential (Klijn 2016). This overall issue of information impartiality has profound implications for political power, decision-making, and democracy in general, as it determines how conflicts enter the public sphere (Oliver and Myers 1999;Entman 2007).
However, while the role of the media is gaining increasing attention in research on armed conflicts and wars (e.g., Puddephatt 2006;Weidmann 2015), its role in the context of social-ecological conflicts is still meagre at best. To contribute to fill that gap, in this work, we analyze local media portrayal of stakeholders and their characteristics and alliances around a conflict arising from an infrastructure development project (i.e., the construction of a paved road) planned on an Indigenous territory and national park in South America.
In the last two decades, South American countries have experienced a high expansion of their road networks, mostly to promote regional economic integration (Laurance and Balmford 2013;Fraser 2014;Harvey and Knox 2015;Gallice, Larrea-Gallegos, and V azquez-Rowe 2017;Fern andez-Llamazares et al. 2018). The construction of such large transportation infrastructures is often contentious. Advocates of roads expansion argue that roads provide access to natural resources, facilitate market access to rural producers, and enable economic integration (Clements 2013;Saguier and Brent 2014;Riggirozzi 2015). In their view, roads reduce the costs of spatial mobility, bringing services and economic development to less-economically developed areas (De Lancie 2008;Perz et al. 2010;Pellegrina 2014). Opponents claim that road construction causes severe environmental impacts through increased deforestation, habitat loss, pollution, illegal hunting, and wildlife trade (Coffin 2007;Ahmed et al. 2014;Laurance et al. 2014;Baraloto et al. 2015). Furthermore, opponents also argue that roads have high social costs, as they increase land insecurity, especially when road construction violates local communities' customary land rights and creates conflicts over natural resources (Perz et al. 2008;Van der Ree, Smith, and Grilo 2015).
Within this context, we focus on the planned construction of a deeply-contested road crossing the Isiboro-S ecure Indigenous Territory and National Park (TIPNIS) in the Bolivian Amazon, which has deemed to be one of the highest-profile socialenvironmental conflicts in the history of South America (Hope 2016;Fern andez-Llamazares et al. 2018). Specifically, we explore (1) which stakeholders of the TIPNIS conflict are portrayed by the media, (2) how their position and involvement in the conflict is represented, and (3) how relations amongst stakeholders are represented. We then contrast this information with previous scholarly work in the area to discern differences between the narratives portrayed by the media and the findings in the academic literature. We conclude discussing the main strategies and discourses employed by the lowland Indigenous communities of TIPNIS in their efforts to attract media visibility and influence decision-making against the road construction.

TIPNIS Social-Environmental Context
The Isiboro S ecure Indigenous Territory and National Park (TIPNIS, for its acronym in Spanish) is located on the eastern slope of the Andes, between the departments of Beni (Moxos province) and Cochabamba (Chapare province). This unique biogeographic region, extremely rich in wildlife, is known as one of the most biodiverse regions on earth (Fern andez and Altamirano 2004;Hoorn et al. 2010). The TIPNIS covers 1,236,296 ha of which 88% have the dual protection status of National Park (granted in 1965 to the full area) and Native Community Land (TCO in Spanish, granted in 1990). The 12% of the National Park which does not overlap with the TCO is referred to as Polygon 7 (Figure 1). Despite such dual protection status, the area suffers important environmental threats, including the expansion of coca farming (Lilienfeld and Pauquet 2005;Siles 2009;Fern andez-Llamazares et al. 2018) and oil and gas exploration (Fernandez-Llamazares and Rocha 2015), activities that would largely benefit from the construction of the proposed road (McNeish 2013;. The national park is mostly inhabited by lowland groups distributed along the Isiboro and S ecure rivers and by some highland groups in the southern piedmont region. About 12,000 lowland Indigenous Peoples from the Mojeño, Yuracar e, and Tsimane' groups live in 64 communities with legally recognized communal rights over land overlapping with the National Park. Despite their sociocultural particularities, lowland Indigenous groups mostly rely on subsistence activities, including small-scale agriculture, hunting, fishing, and gathering. The political organization of lowland Indigenous Peoples is articulated around two representative bodies: the Subcentral S ecure (representing lowland Indigenous communities settled around the S ecure River) and the Subcentral TIPNIS (representing all lowland Indigenous communities within TIPNIS), both of which are part of the Confederaci on de Pueblos Ind ıgenas de Bolivia (CIDOB), the main umbrella organization for Bolivian lowland Indigenous Peoples.
In the southernmost part, within the Cochabamba Department and close to the coca-producing area of Chapare, the National Park is inhabited by Aymara and Quechua colonizers from the Bolivian highlands and the Chapare region. These communities, locally known as "colonists" and more recently as "interculturals," migrated to the area from the 1970s onwards and began to produce cacao, bananas, cassava, maize, citrus fruits and, most notably, coca. With time, their numbers increased and they started encroaching into the Park, a situation that led to the drawing of a "red line" marking the boundary between the National Park and the lands of lowland Indigenous groups on one side and the area occupied by the colonizers on the other. In 2009, Polygon 7 was separated from the demarcated territory granted to lowland Indigenous groups. In Polygon 7 individual land titles were granted to farmers and the area was officially recognized as distinct from the TIPNIS (Webber 2012). The area is now mainly populated by about 100,000 people who live in permanent settlements and rely on market-oriented economic activities, notably coca cultivation (McNeish 2013;Lalander 2017). Colonizers are organized in 52 agrarian trade unions and participate in one of the six federations of Chapare coca producers, Federaciones del Tr opico, for which the current president of Bolivia, Evo Morales, is a leader (Grisaffi 2010). They also participate in the Indigenous Council of the South (CONISUR for its Spanish acronym), an organization representing 21 communities who live close or inside Polygon 7 (Von Stosch 2014).
Villa Tunari-San Ignacio de Moxos Road Project: The TIPNIS Conflict (2009 In 2000, the Union of South American nations launched a trans-regional development strategy to link South America's economies through new transportation, energy, and telecommunications projects (Zibechi 2006;van Dijck 2008). The strategy, known as Initiative for the Integration of the Regional Infrastructure of South America (IIRSA), is expected to integrate the region's road networks, river ways, hydroelectric dams, and telecommunications to allow greater trade across the region (Veltmeyer and Petras 2014;Lalander 2017). The TIPNIS road was part of this initiative.
Although the plan for the TIPNIS road dates back to the early 2000s (Lilienfeld and Pauquet 2005), the conflict around its construction only sparked around 2009, when the Bolivian government signed a contract with a Brazilian company for the construction of a highway connecting Villa Tunari (Cochabamba) and San Ignacio de Moxos (Beni) (Figure 1). The road was planned in three sections, with Section II traversing the TIPNIS core area. Section II was contentious from the onset for two reasons established by the Bolivian Constitution. First, given its legal status as a National Park, infrastructure development was not allowed in the area. Second, given its legal status as an Indigenous territory, the project required the Free, Prior and Informed Consent (FPIC) of the local Indigenous communities (Achtenberg 2013), but this right was violated by the Bolivian Government (McNeish 2013).
After several months of frustrated negotiation in which conservationists and lowland Indigenous Peoples' representatives attempted to persuade the government to abort the project, the opposition to the road construction reached its peak in August 2011, when about 1000 lowland Indigenous Peoples marched from Trinidad to La Paz, covering 600 km, in their attempt to stop the road construction (VIII Indigenous March; see Supplementary Materials, S1). Simultaneously, intercultural federations and individuals intensified activities supporting the road. On September 25th 2011, road supporters blocked the VIII Indigenous March and the police exerted a brutal repression against marchers. The incident awakened the interests of international non-governmental organizations (NGOs) and human rights observers and increased the social support to the March (Petherick 2011). When the VIII March reached La Paz in October 19th, road opponents were strong enough to push the government to promulgate a law attributing TIPNIS the status of strictly protected, or untouchable ("intangible" in Spanish), making the area out of bounds for all future forms of state or development projects (Law 180 of TIPNIS, Delgado-Pugley 2013).
Soon after, CONISUR organized a "countermarch" with the participation of around 2,000 people in favor of the road. When, after two months of walking, this "countermarch" arrived at La Paz in February 2012, the government promulgated a new law, the 222 Consultation Law, according to which the government would consult the inhabitants of TIPNIS about the planned road construction (Gaceta Oficial del Estado Plurinacional de Bolivia, 2012). In response, lowland Indigenous Peoples organized another march (the IX Indigenous March) to demand the abolishment of the Consultation Law although this time negotiations were unsuccessful and the marchers faced a new wave of repression (Fabricant and Postero 2015).
Consultation was undertaken under an exceptionally high pressure (Fontana and Grugel 2016) and the results obtained (i.e., 80% of the TIPNIS communities supported the road construction) have been highly contested, with reports showing that many communities had never been consulted or had been coerced or manipulated (FIDH and APDHB 2013;Lalander 2017). After the consultation, the conflict entered a dormant phase that recently ended following the promulgation of a law annulling the TIPNIS untouchable status in August 2017 (Amazon Watch 2017).

Methodology
We collected information from the Bolivian press about the TIPNIS conflict focusing in the years 2011 and 2012, the period of major visibility of the conflict in the media. We analyzed this information combining stakeholder analysis (hereinafter SA) and social network analysis (hereinafter SNA) (Lienert, Schnetzer, and Ingold 2013). We defined stakeholders as actors who will be affected by or can influence decision-making processes toward a concrete goal. Stakeholders can be individuals but more commonly are public or private organizations, such as international organizations, governments, NGOs, businesses, and local communities (Mayers 2005). Previous work shows that stakeholders differ in their capacities and resources, including political, financial, cognitive, and moral (Dingwerth and Pattberg 2009). We use SA to identify (1) which of the stakeholders affected or being affected by the TIPNIS road are featured by the media and (2) how their participation and position in TIPNIS-related events is presented. We then use SNA to identify the relations amongst different stakeholders portrayed by the media through the featuring of stakeholder participation in different conflict-related events (e.g., marches, protests, meetings).

Sampling and Data Collection
We followed a two steps sampling procedure. In the first step, we identified stakeholders in the conflict; in the second step, we collected information on media portrayal of stakeholders' position, visibility, and relations. To identify stakeholders in the TIPNIS conflict, we selected four standard, broadsheet, general-interest, paid-for newspapers in Bolivia: two at national and two at regional levels. Following Quandt (2008), our selection of newspapers was based on (1) orientation towards professionalized "high-quality" journalism; (2) market leadership and high circulation rates; and (3) available in electronic format through online repositories. We deliberately excluded special monothematic newspapers (e.g., financial newspapers), tabloids, freesheets, and regional newspapers in departments other than Cochabamba or Beni. The selected journals were La Raz on and P agina Siete (national level) and La Palabra del Beni (department of Beni) and Los Tiempos (department of Cochabamba). For each journal, we screened all the titles of press articles in the policy section appearing between 1 January 2011 and 31 December 2012 and read all the articles related to the TIPNIS conflict. Following our definition of stakeholder, we noted all the individuals, groups, and organizations that could affect, be affected, or had publicly expressed a position about the construction of the TIPNIS road.
In the second step of the sampling procedure, we chose one newspaper to gather information on how stakeholder's participation and position in TIPNIS-related events are portrayed. We selected La Palabra del Beni, a departmental newspaper published in Trinidad, renowned for its clear opposition to the government and its support for increasing departmental autonomy. La Palabra del Beni was selected for three reasons.
First, it constitutes the highest-circulation regional newspaper in the study area, thus having greater potential to exert social influence at both local and regional levels. Second, it is the only regional newspaper in the Beni Department with a complete online repository. And third, it is the only regional newspaper that is not owned by or directly connected to, any political figure or party. We retrieved all the articles mentioning the TIPNIS conflict and stored them in Nvivo 11. We read all the retrieved articles and noted the following information: (1) type of event described (e.g., demonstration, meeting) and (2) stakeholder's position in the event regarding the road construction (i.e., supportive, neutral, or adverse). We entered information in a database where each row corresponded to an actor/article (where multiple actors mentioned in an article are listed in different rows).

Data Analysis
We used contextual information provided in the newspapers to classify stakeholders according to two criteria: sector and geographical scale. Thus, stakeholders were classified as representing the interest of (1) the administration (e.g., Executive, local government), (2) Indigenous Peoples (e.g., Subcentral TIPNIS, CIDOB), (3) the civil society (e.g., NGOs, social movements), or (4) other sectors (e.g., researchers, international institutions). Stakeholders were also classified according to the geographical scale at which they primarily operate: (1) local (e.g., Mayor of Villa Tunari), (2) regional (e.g., regional government of Cochabamba), (3) national (e.g., Bolivian Senate), or (4) international (e.g., foreign investors). Such categories have a purely analytical purpose, as in real life stakeholders might represent more than a single sector and/or operate at different scales.
Events featured in the news were classified as dialogue (e.g., meetings) or unilateral events (e.g., claims, protests, rallies, demonstrations, or marches). We define dialogue events as instances in which multiple actors converge to exchange opinions and unilateral events as actions organized to assert a particular position in the conflict (Routledge 2003). We differentiate between these categories because unilateral events might represent a higher degree of alliance and collaboration between stakeholders than meetings.
We also aggregated information to create three variables capturing stakeholder's position, visibility, and participation in the TIPNIS conflict (Schmeer 1999;Varvasovszky and Brugha 2000). Position refers to whether a particular stakeholder is featured as supporting or rejecting the road construction and was calculated by aggregating information from all the press articles reviewed in which the actor was mentioned into four categories: supporters, or stakeholders whose position was featured as supportive in at least 30% of the mentions; opponents, or stakeholders whose position was featured as adverse in at least 30% of the mentions; neutral, or stakeholders whose position was featured as neutral in at least 30% of the mentions; and ambiguous, or stakeholders who simultaneously had at least 30% of the mentions as supportive and 30% as adverse.
Visibility refers to the stakeholder's presence on the media in relation to the TIPNIS conflict and was measured as the share of days the stakeholder was mentioned from the total number of days when La Palabra del Beni had an article on the conflict. Based on the distribution of the values in this variable, we defined five values: major (when the stakeholder was mentioned in >5% of events), high (between 5% and 2.5%), regular (2.5-1.5%), low (1.5-0.5%), and minor visibility (less than 0.5%).
Lastly, participation refers to the percentage of events-from the total list of TIPNISrelated events-in which a stakeholder had reportedly participated. To estimate the participation of a stakeholder, we constructed a two-mode network (or stakeholder_ by_event affiliation network) in which the rows are the stakeholders and the columns the events. A value of 1 in a cell of this binary matrix means that that actor X participated in event Z (Diani and Kousis 2014;Halgin, Borgatti, and Brass 2014). We imputed data from the stakeholder_by_event affiliation network into UCINET 6 to generate two visualizations of how the regional press has presented stakeholders' participation in TIPNIS related unilateral and dialogue events. We also calculated the share of events in which each stakeholder had participated. Based on the distribution of this variable, we defined four levels of participation: very high (stakeholder who participated in >30% of events), high (10% to 30%), medium (3% to 10%), and low participation (3%).
Drawing on the affiliation network stakeholder-by-event, we created the stakeholderby-stakeholder squared matrix of overlaps through the cross-product method (Hanneman and Riddle 2005). This matrix depicts how many stakeholders were featured as present in the same events and is our measure of network relation, as co-participation in events is considered a network tie in the analysis of policy networks (Leifeld and Schneider 2012;Fischer and Sciarini 2016). We then conducted a one-mode coreperiphery analysis to the stakeholder-by-stakeholder matrix (Borgatti and Everett 1999) to discern between the set of actors who are closely related (i.e., have high density of ties amongst themselves by sharing many events in common, the core) from the actors who are less related because they have participated in few events in common (i.e., the periphery). The analysis allows to picture differences in stakeholders' capacities to coordinate actions: actors in the core have a structural advantage compared with actors in the periphery because they are better able to coordinate their actions (Hanneman and Riddle 2005).

Stakeholders Analysis
In the systematic reading of the 2011-2012 numbers from the four newspapers, we identified 34 different stakeholders (Supplementary Materials, S2). Out of the 34 stakeholders identified in the press, 14 represent different levels of the public administration (i.e., three stakeholders represent the local, two the regional, and nine the national level), 10 represent the civil society (four regional, five national, and one international), six stakeholders represent Indigenous Peoples and local community organizations (three local and three national), and four stakeholders represent other sectors at national (two) and international (two) levels. Thus, overall most (n ¼ 19) stakeholders represent the national scale, whereas only six represent the local and other six the regional levels. Altogether, we only found three stakeholders at the international level.
The systematic reading of La Palabra del Beni resulted in a total of 4059 TIPNISrelated press articles: 1701 articles published in 2011 and 2358 in 2012. In these articles, there were a total of 336 citations regarding stakeholder's participation in 103 TIPNISrelated events. Out of the 103 TIPNIS-related events identified, 50 correspond to dialogue events (i.e., meetings) and the other 53 correspond to unilateral events (i.e., claims, protests, activities, and the three aforementioned marches).
Overall, La Palabra del Beni portrayed stakeholder's positions regarding the road construction as largely divided: 12 stakeholders supported the construction of the road, 12 opposed it, six were neutral, and four were ambiguous (Figure 2, see also Supplementary Materials, S2). Stakeholders portrayed as supporting the construction of the road include the public administration (i.e., the national government, the Cochabamba government, the mayor of Villa Tunari), representatives of highland groups (i.e., CONISUR, Federaciones del Tr opico, intercultural organizations), and the private sector (i.e., national enterprises and foreign investors). Stakeholders portrayed as opposing the construction of the road include lowland Indigenous Peoples organizations (i.e., Subcentral TIPNIS and S ecure, part of CIDOB), social movements, NGOs, and some members of the public administration (i.e., opposition deputies and senators, major of Trinidad). Stakeholders portrayed as neutral include four representing the national administration (i.e., executive-ministers, Senate from the Movimiento al Socialismo (MAS), judicial authority, and ombudsman), national researchers, and international organizations. Stakeholders portrayed as ambiguous are from the public administration (i.e., deputies MAS, Regional government Beni, and National Service of Protected Areas) and Bolivian Workers' Center (i.e., COB for the Spanish acronym, Bolivia's largest labor union). La Palabra del Beni gave major visibility to six stakeholders (i.e., four Indigenous organizations and two stakeholders of the public administration) (Figure 2), high visibility to nine stakeholders: deputies from MAS and from the opposition, the judicial authority (who played a main role in the development and interpretation of the laws regarding the road through TIPNIS), and intercultural organizations. Other actors given high visibility were the local government of Beni, national NGOs, and the COB. Seven stakeholders received regular visibility: two stakeholders from the public administration at the national level (i.e., senate-MAS and the Ombudsman), two stakeholders from the civil society (i.e., Civic Committee of Beni, analysts, and national researchers), and international institutions, who mainly engaged as international observers during the marches and the consultation process. The newspaper gave low visibility to six stakeholders, including three stakeholders belonging to the administration sector (i.e., the mayor of San Ignacio de Moxos, the local government of Cochabamba, and the protected areas service, SERNAP) and three stakeholders from the civil society (i.e., Federaciones del Tr opico, hauliers and social movements). Finally, the newspaper gave minor visibility to the mayors of Villa Tunari and Trinidad, the senate-opposition, the Civic Committee of Cochabamba, the national enterprises, and the Federaci on de Ganaderos del Beni y Pando (FEGABENI), i.e., a regional association of cattle ranchers.
Overall, Figure 2 shows that La Palabra del Beni portrays supporters as stakeholders with medium to high participation and operating from local to national scales. The newspapers give supporters minor to high levels of visibility. Minor visibility was given to supporters operating at local and regional scales. In contrast, opponents appear as more scattered, with one group of stakeholders operating from local to national scale with a very high participation, and another group operating at the national scale and mostly with an urban profile having lower participation. Opponents are generally given some of the highest levels of visibility. Neutral and ambiguous stakeholders, only operating at national or international scales, appear as having medium levels of participation, but are given major to high visibility.
The 34 stakeholders had a total of 336 references in the 103 TIPNIS-related events identified, but participation levels differed widely between actors. For example, the Subcentral TIPNIS was featured as participating in 41 of the events identified, whereas FEGABENI did not appear as participating in any. Our data suggest that road opponents were featured as participating in more events than road advocates. Thus, the three actors within the category of very high participation represent lowland Indigenous groups: Subcentral TIPNIS (which appeared as participating in 41 of the 103 identified events), CIDOB (39 events), and Subcentral S ecure (35 events). These three stakeholders were present in more than 30% of the events and were also portrayed as having high visibility. Within the high participation category, we found eight stakeholders belonging to the Indigenous, administration, and civil society sectors and featured as supporters, opponents, neutral, and ambiguous. All stakeholders with high participation were also portrayed as having major visibility except social movements, who had a low visibility despite their high participation. Note also that executive-President was the only stakeholder identified as major visibility but low participation (11.76%). A heterogeneous group of 14 stakeholders from multiple sectors and positions were featured as having medium participation. Similarly, the category of low participation included nine stakeholders from all positions and different levels of visibility (except major visibility). Finally, La Palabra del Beni featured foreign investors as having high visibility but participating in less than 1% of the events.

The Stakeholder-by-Event Social Network
We used information from the stakeholder-by-event social network to compare the ranking of the stakeholders in terms of degree among the network of dialogue events, the network of unilateral events, and the complete network. We found that the top 10 stakeholders of the complete network of events (n ¼ 103) are the same than those represented in the network of dialogue events (N ¼ 50), with slight changes in the ranking. Moreover, eight of the top 10 stakeholders of the complete network are also represented in the network of unilateral events (N ¼ 53) (Supplementary Materials, S3). The two new stakeholders in the unilateral network are the Civic Committee of Beni (opponent), and national researchers (neutral), with ranks 7 and 8, respectively.
Given the similarity of the stakeholders across the three networks, we created a visualization of how La Palabra del Beni presented stakeholders' participation in TIPNIS events (both dialogue and unilateral) (Figure 3). In that figure, node size shows degree centrality (i.e., the number of direct links between stakeholders and events), so the bigger a stakeholder node (colored nodes), the more events had the stakeholder attended, and the bigger the event node (grey nodes), the more stakeholders have participated in the event. The results presented in Figure 3 suggest that the three lowland Indigenous organizations that oppose the road (i.e., Subcentral S ecure, Subcentral TIPNIS, and a sector of CIDOB) participated in more events than the other stakeholders. From those supporting the road construction, the two stakeholders participating in more events are the Executive (President) and CONISUR. It is worth noticing Figure 3. Bipartite graph of the two-mode network stakeholder-by-event of the TIPNIS conflict.Note: Coloured nodes are the stakeholders (red: opponents; green: supporters; blue: neutral; and brown: ambiguous) and grey nodes are the events (squares: dialogues; triangles: unilateral events). Node size shows degree centrality (i.e., the number of direct links between stakeholders and events). A colour version of the figure can be found online.
that, while the opponents seem to participate together in many events, there are also a substantial number of small events in which only few actors participate. Interestingly, none of the events seem to have a degree centrality much larger than the others, indicating that through the period of analysis, the conflict was characterized by a high number of events attended by specific stakeholders, rather than by central events highly attended.
The results of the stakeholder-by-stakeholder core-periphery analysis confirm previous results. Figure 4 shows a core composed by a group of opponents (i.e., Subcentral TIPNIS, Subcentral S ecure, and CIDOB) (density of connections ¼ 26.67; fitness ¼ 0.833), who are all very likely to co-participate in a considerable number of events (either dialogue or unilateral). Although to a lesser extent, opponents were also connected to the core group of road supporters (i.e., CONISUR and Executive president) as the two groups participated in dialogue events. The remaining 29 stakeholders had a lower number of ties between them (density of connections ¼ 0.5) indicating that they are featured as co-participating in fewer events. These actors conforms the network periphery (Figure 4).

Discussion
Bolivian newspapers portray a high plurality of stakeholders with different positions in the TIPNIS conflict, but-through the selection of the events reported and the actors mentioned-it also grants more visibility to some stakeholders than to others and provides a particular view of the relations between different actors. To understand the the thicker the tie, the more events two actors have in common. The node color represents stakeholder's position (red: opponents; green: supporters; blue: neutral; brown: ambiguous) and the size stakeholder's visibility (the bigger the node, the more visibility had the stakeholder). A colour version of the figure can be found online.
role of media in shaping public opinion, in this section, we compare media portrayals of stakeholders in the TIPNIS conflict with information from previous scholarly work on the topic. Being such a high-profile social-environmental conflict, the topic has received considerable attention in scholarly work. Some authors have analyzed the conflict using a political economy approach that highlights the social tensions between extractivism, Indigenous rights, and nature conservation (Fabricant and Postero 2015;Hope 2016), others have addressed the contradictions in Bolivia's politics of indigeneity (McNeish 2013;Canessa 2014;, and others have taken a conservation angle (e.g., Fern andez-Llamazares et al. 2018). Overall, this literature clearly shows that there is a complexity of local, regional, national, and international actors as well as of social, material, and territorial interest around the planned construction of the road (see McNeish 2013). The question is whether the media portrays this complexity in the same terms and scope.

Stakeholders' Diversity
Much in line with previous scholarly work (e.g., McNeish 2013), the media portrays a high diversity of actors with a stake on the TIPNIS conflict. Thus, although the media presents lowland Indigenous Peoples and the Bolivian central government as the central actors in the conflict, it also features a large number of other stakeholders with some level of participation in it. The diversity of stakeholders portrayed in the press, together with the fact that they represent multiple social sectors (e.g., administration, Indigenous Peoples, civil society, and other sectors) and geographical scales (i.e., local, regional, national, and international), confirms the strategic importance of the planned road beyond local and regional levels (Hope 2016;Fern andez-Llamazares et al. 2018). There are, however, two important differences between the media representation of stakeholder's diversity and results from previous scholarly work. First, some actors appearing as central in the academic literature of the conflict are largely absent in the media. And second, the media tends to present more homogeneous positions towards the road than the academic literature on the conflict. We elaborate on each of these two points.
Some of the stakeholders that have been singled out as having a central role in scholarly analysis of the TIPNIS conflict have a peripheral role in media portrayals. For example, researchers have discussed the critical role of foreign investors in the road construction Sanchez-Lopez 2015). However, in our analysis of press articles, foreign investors have low visibility and low participation in public events. The discrepancy might be explained by the fact that the media tends to focus on vocal actors (Carroll and Hackett 2006;Malinick, Tindall, and Diani 2013) and extreme positions (Klijn 2016), and foreign investors probably do not fit these categories. Moreover, considering that most media reports largely focus on events happening in the public sphere, it is not surprising that some actors operating beyond the media spotlight might have remained invisible.
Similarly, while the literature on the topic has often portrayed the conflict in terms of the social tensions between extractivism, Indigenous rights, and nature conservation (Fabricant and Postero 2015;Hope 2016), stakeholders in the conservation sector (i.e., conservation organizations and SERNAP, the public institution for the management of Bolivia's protected areas) were not portrayed as central by the media. Given that local and international environmental NGOs have often been the strategic allies of Indigenous People's fights for their land rights across the Amazon (e.g., Schwartzman and Zimmerman 2005;Postero 2007;Earle and Pratt 2009), this finding is indeed surprising. In fact, this mismatch might be a result of the contradictory actions of the Bolivian government regarding environmental issues. Thus, while Bolivia recognized Pachamama's (i.e., Mother Earth) rights in its 2009 Constitution, the Government has recurrently attacked and threatened several environmental organizations. For example, President Evo Morales threatened to expel any NGO or foundation attempting to obstruct the exploitation of the country's natural resources (Fern andez-Llamazares and Rocha 2015); and in 2015, Vice-president Alvaro Garc ıa Linera attacked four wellrespected Bolivian research organizations (i.e., Centro de Estudios para el Desarrollo Rural y Agrario (CEDLA), Centro de Informaci on y Documentaci on Bolivia (CEDIB), Fundaci on Tierra, and Fundaci on Milenio), accusing them of political meddling to advance the interests of foreign governments and corporations using foreign funds to promote colonial interests around environmental protection. While many academics have been bewildered by these developments and extensively written about them in both academic publications and the international press (e.g., Fern andez-Llamazares and Rocha 2015), these tensions have remained relatively unspoken in the Bolivian press. We also found that the press tends to present stakeholder's position in a more homogeneous way than scholarly work does. For example, in contrast with information presented in scholarly accounts of the conflict (e.g., Hope 2016), the media largely features lowland Indigenous Peoples as a homogeneous, unified block in their opposition to the road. Previous work based on ethnographic data suggests that the situation is much more complex, with certain representatives of the lowland Indigenous Peoples under CIDOB, the sub-centrals of the TIPNIS, and the Central de Pueblos Etnicos Mojeños del Beni (CPEM-B, a stakeholder which was not even mentioned in the media) having decided to support the road, often with explicit political aspirations (Hope 2016). Moreover, in addition to portraying their position as static and homogeneous, the media tends to assume that the opinion of all lowland Indigenous Peoples is legitimately represented by the voices of a limited number of Indigenous leaders and representatives. While this could certainly be the case, researchers working in other contexts have noted that there are often discrepancies between the attitudes of local leaders and local communities towards infrastructure development (i.e., Denton 2002;Ogra 2008;Bhattarai, Beilin, and Ford 2015;Jost et al. 2016), as well as disconnection between the leadership and base when it comes to activism (Lucero 2008;Shah 2010). These findings suggest that the media could be providing a biased portrayal of stakeholders' positions. Overall, the voices of lay people from the Indigenous communities within TIPNIS remained largely unheard in most of the press articles surveyed.

Stakeholders' Network
The complexity of relations established among stakeholders as portrayed by the media sits well within the literature reporting that Indigenous groups in remote parts of the Amazon are able to martial international support for the defense of their rights (e.g., Ramos 1994;Conklin and Graham 1995;Turner 2008). Thus, the media portrays the three lowland Indigenous organizations opposing the project (i.e., Subcentral TIPNIS, Subcentral S ecure, and CIDOB) as the core of the network of stakeholders in the TIPNIS conflict. In other words, the media portrayals of actors' networks suggest that stakeholders localized far from political and urban centres and having limited electoral relevance can gain high visibility in a conflict arguably through their connectivity with other stakeholders. Thus, using the structural advantage that derives from being at the core of the stakeholders' network (Hanneman and Riddle 2005), lowland Indigenous organizations might have sparked high levels of social mobilization around them, i.e., the two marches in defence of TIPNIS (Lorenzo 2011 Media representation of stakeholders' alliances also points at interesting associations that, to our best knowledge, have not been discussed in the academic literature. Thus, our results show that lowland Indigenous organizations at the core of the TIPNIS conflict have been able to amass support from urban sectors and the broader civil society. These alliances present an interesting venue in which to focus further research, as they suggest that Indigenous Peoples are expanding the type of stakeholders who had previously supported Indigenous struggles in the Amazon, i.e., mostly peasant organizations and local and international environmental NGOs (e.g., Tanaka 2012; Villanueva 2013). Indeed, such new alliances might indicate that Indigenous organizations involved in this conflict are articulating their fights around issues of social equity (Robinson 2003; Orta-Mart ınez and Finer 2010; Veltmeyer and Bowles 2014; de la Cuadra 2015; Temper, Del Bene, and Mart ınez-Alier 2015) and therefore getting closer to the "environmental justice" discourse (Martinez-Alier et al. 2016) than to the "conservationist" discourse. Such discourse shift certainly deserves more scholarly attention.
As with the role assigned to stakeholders, media portrays of stakeholders relations do not make visible some of the important alliances driving the conflict. For example, President Evo Morales, one of the most vocal defenders of the road, has long been a leader of Federaciones del Tr opico, Chapare's organization of coca producers (Grisaffi 2010). However, this association does not appear in the results from our SNA. On the contrary, the close links between the President and the federations of coca producers has been highlighted in the academic literature (e.g., Fern andez-Llamazares et al. 2018), with several authors suggesting that, given the interest of coca producers in the TIPNIS area, such association has played a significant role in the way in which the conflict has been officially addressed (Wickstrom 2013;.

Conclusion
In this article, we have analyzed information released in the Bolivian press between 2011 and 2012 in relation to the conflict over the construction of a road through TIPNIS and discussed how results of this analysis compare with scholarly information of the same conflict. We found two important differences in the information on the TIPNIS conflict reported in the academic literature and the information portrayed in the Bolivian press. First, some actors cited as central in scholarly analyses of the conflict are largely absent in the media (e.g., private investors, conservationist sector). And second, the media tend to present stakeholders as having more homogeneous positions towards the road than the academic literature does, while also neglecting some important alliances in their account. Another finding that is distilled from our analysis of the TIPNIS conflict is that Indigenous communities in the Bolivian Amazon are expanding the scope of their claims from the local to the global, increasingly engaging with the discourse of environmental justice, and forging new alliances with urban sectors and civil society. Place-based analyses such as the one presented in this paper hold the potential to uncover the potentialities and prefigurative politics of an emerging global movement for environmental justice (Martinez-Alier et al. 2016;Temper et al. 2018a). They can help us to better understand, contest, and transform different forms of hegemonic power, and to ground in socio-political realities the epistemologies and ontologies of the numerous Indigenous communities impacted by the most explosive era of road expansion in human history (Laurance et al. 2014;Alamgir et al. 2017).