Navigating the Map of Environmental-policy Narratives and its Influence within the Conservation Network of the Colombian-Peruvian Amazon Authors: Carlos Eduardo González-Rodríguez, Oscar Buitrago-Bermúdez, Carolina Gonzalez, Johana Marcela Castillo-Rivera, Jean-Francois Le Coq Published: 2025 ABSTRACT The Colombian-Peruvian Amazon’s global socio-ecological relevance attracts multiple organizations and technologies related to cooperation and environmental governance, configuring a complex network of interactions, relationships, and relational structures. This case study seeks to provide clarity on the complex functioning of the cooperation network in the Amazon biome, composed of institutional and technological nodes. It identified 17 thematic environmental narratives, their patterns of concordance by node (entities/institutions operating within the network), and how they influence inter-network links. Studying cooperating networks in dynamic contexts is made complex by progressive and diverse environmental threats, economic challenges, and barriers to accessing information. The article proposes a methodological approach to untangling these complexities, leveraging (a) social network analysis; (b) web scraping for data collection; (c) text mining to categorize narrative themes by node; and (d) network modeling using Exponential Random Graph Models. We propose and evaluate hypotheses on the influence of narrative concordance, contribution, homophily, and environmental-governance internal structural patterns that are important in creating network links. Study results indicate three dominating narratives: Research & Education, Communities, and Policy, revealing a higher distribution, average contribution, and significance for these discourses. Additionally, international nodes contribute predominantly to Research and Technology topics, surpassing Colombian-Peruvian nodes, suggesting a pattern of influence over these network themes. Lastly, we identify opportunities for improving system interventions for slowing the Amazon biome’s degradation, composed of a diversified, well-integrated cooperation network, and underscoring the need for environmental-policy frameworks that actively integrate local perspectives and capacities, while leveraging international networks to bridge regional limitations. Introduction The Amazon biome is an ecosystem of global importance, due to its exceptional biodiversity, the presence of Indigenous cultures, and its critical role in carbon sequestration (WWF 2020). Yet over-exploitation―driven by an anthropocentric vision and the development model―has accelerated its degradation (Barona et al. 2010; Orjuela and Anacona 2010) bringing it dangerously close to a tipping point that could transform it into a savannah (Lovejoy and Nobre 2018). Faced with this scenario, a complex network of institutional interactions has emerged, aimed at reversing the loss of the biome and its vast biodiversity. This relational configuration, in the form of an environmental governance network (Armitage et al. 2012) with diverse cooperation schemes, allows us to observe the characteristics of interactions―including structures, interdependence, leadership, and discourses―among actors involved in the management, administration, and development of technologies in the region. In this sense, the concept of a network is approached from a methodological Social network analysis (SNA) perspective, in which agents and organizations (nodes) build relationships based on various forms of cooperation (links) through their efforts in the Amazon (Contractor et al. 2011), mobilized in pairs, triads, or groups with other partners, guided by the goal of conserving the biome. The interactions between the nodes represent efforts (economic, technological, and institutional) to mobilize in a coordinated manner toward a common purpose or set of values. This Amazon network, with its multidisciplinary, open, and cross-border structure, interacts with a multidimensional network that encompasses various types of actors, including technological nodes; grassroots communities; research centers; government entities, and international organizations. In this context, network analysis offers a wide range of tools to study these interactions, providing valuable information to (1) guide policies, resources, and decision making, both in the public and private spheres, and (2) promote the active participation of the actors involved and their various forms of cooperation, in identifying problems and designing and implementing interventions that are tailored to local and social contexts and needs (Biermann et al. 2016). Environmental networks, especially, play a key role in the governance of natural and ecological resources (Barlow et al. 2011; Bodin and Crona 2009a), highlighting the importance of constructing consensus, defining power relations, seeking cohesion among those involved, and the producing and transferring knowledge and technologies. Following Denton (2017a), these networks function as stable groupings of actors engaged in interdependent actions characterized by the exchange of material and immaterial resources. This interdependence reflects the functional connections among stakeholders and governance rules across topics, configuring a system of social organization that is increasingly relevant for resource management and environmental policy development (Prell 2016). Therefore, interpreting environmental reality with an organizational network theory perspective is the most appropriate analytical mechanism to investigate the complex spheres of authority within transnational environmental governance (Andonova et al. 2009; Hadden 2015). In operational terms, the association and coordination of actions among national or transnational organizations are mediated by their communication channels and primarily by each actor’s unique thematic narratives. The communications, languages, and dominant topics within environmental networks provide evidence of patterns in the exchange of discourses, scientific positions, viewpoints, and work themes through their connections with other partners in the Amazon. Environmental network narratives contribute to determining achievements and identifying tools to address them, to reach decisions regarding the distribution of benefits and responsibilities in policy implementation and management, to creating rules for inclusion and exclusion, and to generating rational and legitimate choices (Jones et al. 2014). These narratives strengthen communication and social cohesion (Dunbar 2014; Wiessner, 2014), particularly concerning complex environmental information, thus generating knowledge that is useful to society (R. Lejano et al. 2013). Additionally, they serve as vehicles for rationalization, content creation, and the exchange of various critical dialogues on environmental issues (Koch et al. 2021a; Ryan 2007). Notwithstanding, civil society, public policies, and the private sector have persistently called for stimulating coalitions, networks, and collaborative efforts at the organizational level. This mobilization is crucial for conserving and protecting biodiversity in a biome as challenging as that of the Amazon. Conversely, few studies address the functioning of organizational networks and offer patterns, trends, and relationships among the actors involved (Reyes Sarmiento 2016). The flow, type, and narrative contributions disseminated among the actors and exchanged in network relationships remain unknown. Therefore, there is little information available to map disagreements, conflicts, and mismatches in the commitments of the actors involved, as well as potential misalignments in cooperative efforts at all levels. This absence of clear communication signals related to environmental narratives obscures the roles of some actors, establishes diffuse and uncoordinated commitments, and fosters other types of conflicts (Plummer et al. 2020). Thus, current knowledge about the map of actors and the exchange of available discourses hampers the development of robust and coordinated actions for implementing appropriate and effective public environmental policies. There seems to be no baseline information regarding the narrative trends underlying the cooperation declared by organizations working in the Peruvian and Colombian Amazon. The analysis of environmental governance in the Amazon biome lacks a robust body of literature, particularly on topics related to networks and narratives. A review of scientific publications in the Web of Science database for the 2016–2025 period reveals a marked disparity with Brazil, which accounts for 89% of all scientific literature focusing on the Amazon. In contrast, Peru and Colombia account for only 9% and 6%, respectively, reflecting existing bias and prioritization in scientific, technological, and public investment capacities within the Colombian-Peruvian Amazon region. The full search code is provided in the supplementary material (Table S1). It is crucial that governance and environmental co-management spaces provide the necessary resources to overcome and reconcile differing viewpoints, and understand the issues and potential solutions (Koch et al. 2023a). In this context, there is an opportunity for proposing an initial approach that analyzes and generates the map of narratives within policies and actions in the Amazon region. This article aims to provide analytical elements to enhance the understanding of the narratives of institutional, organizational, and technological actors in the Colombian and Peruvian Amazon. It is presented as a case study composed of neighboring countries that share the political administration of the Amazon biome, intending to contribute to academic studies and development work, as is the case in the Brazilian Amazon. We combine qualitative and quantitative methods to address this issue. We have: (a) developed a Social Network Analysis (SNA) to visualize and map the relationships among actors in the Colombian-Peruvian Amazon; (b) conducted web scraping and text mining to extract information from discourses and to categorize narratives; and (c) employed Exponential Random Graph Modeling (ERGM) to represent the network and evaluate the influence of endogenous patterns (internal structures of the network) and exogenous variables (node attributes, type of organization, weight of narrative proportions, etc.), obtaining statistics that measure the propensity for forming links within the network and for accepting or rejecting hypotheses about the structure formed among the nodes. SNA conceives “nodes” to be each of the entities or organizations operating within a network, which also facilitates a better visualization of how they interact and cooperate in this system of relationships. This methodological approach seeks to: (a) identify the dominant narratives in the Colombian-Peruvian Amazon cooperation and technology network; (b) explore the differences in discursive topics between local and international nodes, and (c) determine the concordance or divergence levels regarding a set of themes and the narrative weight among the nodes and the mapped network (homophily). To achieve this, we collected data through interviews with Amazon experts selected according to their affinity with the topic, followed by a review of secondary sources to complete and map the network. Additionally, we developed an algorithm using the R programming language to extract and analyze texts from the official websites of various environmental organizations operating in the Amazon, focusing on their stated purpose and scope, and other resources associated with each organization viewed as a node. This process generated a text corpus to determine the composition of the various narratives. Simultaneously, the extracted texts were standardized and assigned to predefined themes and narratives by matching processed words with key terms associated with each theme. This approach facilitated the creation of thematic word sets, enabling narrative profiling for each node. Environmental-Governance Narratives Narratives are verbal and written arguments that, when integrated, shape the conceptual frameworks or approaches of individuals or organizations. They consist of thematic descriptions and statements that establish trends, positions, and purposes (Polletta et al. 2011), and serve as a key instrument guiding the governance of environmental networks (Koch et al. 2021c). Actors may encounter difficulties finding common meaning and interpretative frameworks for addressing problems and co-managing environmental issues, which are essential for fostering shared positions, common perceptions, collaborative management, and effective action guidelines (Berkes 2009). Narratives play a crucial role as phenomena to be empirically observed and studied, and as a fundamental tool for understanding the emergence and dynamics of collaborations at both local and regional levels (Koch et al. 2021a). In this sense, narratives are mechanisms for creating meaning, articulating shared arguments, and fostering social cohesion within environmental networks. On the one hand, they convey information, and on the other, they shape power relations and cooperation strategies (R. Lejano et al. 2013). Narratives are a key factor in consolidating social ties and attracting other agents, and in stimulating mobilization and collective constructions of knowledge and policy influence (Dunbar 2014; Wiessner 2014). Narratives also play a significant role in shaping policy instruments and the public’s perception of environmental issues by simplifying and structuring complex matters, thus facilitating decision-making and coalition-building (Jones and McBeth 2010a). They are essential for understanding power and conflict dynamics in environmental governance, as they influence how problems and solutions are conceptualized, either facilitating or hindering cooperation and policy implementation (Dryzek 2005). Socio-ecological literature suggests that establishing a collaborative network does not guarantee successful environmental governance. Effectiveness depends on the economic structures of environmental governance and also on the actors and their collaborative links (Folke et al. 2007; Galaz et al. 2008). Agreements can fail to adequately align and match institutional structures with biophysical issues. This misalignment often leads to undesirable outcomes (Epstein et al. 2015). Tilly (2004) emphasizes the use of formal representations, such as network analysis, to bridge the gap between qualitative data and rigorous explanations. Similarly, Emirbayer (1997) highlights the importance of focusing on dynamic, relational processes rather than static views of social entities, which is essential to understanding the evolving nature of collaboration networks in environmental governance. In the Colombian-Peruvian Amazon case, the analysis and use of narratives in environmental governance is expanding. While the dynamics of power and cooperation have been studied, there remains a need to understand how these dynamics influence the formulation of public policies and resource management. Analytical efforts within a network of organizations can help map disagreements, identify conflicts among actors, and align cooperation efforts according to their respective interests. At the same time, such analyses provide a foundation for more inclusive and effective policies that foster shared understanding among diverse actors in conserving the Amazon’s biodiversity. Navigating the Narratives in Environmental Cooperation Networks Technology development within the governance network refers to the creation and application of tools, methods, and innovations that facilitate collaboration, information exchange, and decision-making among actors in an environmental governance context. By responding to a complex structure, the network of cooperation and technology development provides an opportunity to explore the relationships among nodes and describe patterns that strengthen the construction of narratives, based on the exchanges, shared work, and common achievements within the networks (Denton 2017a). These analyses of environmental governance networks facilitate identifying cooperation approaches (Ryan 2007) and allow an understanding their functioning, what motivates actors to act, and how they create diverse connections within networks, overcoming cultural, organizational, and disciplinary barriers. Furthermore, such analyses help identify formal and specific expressions of ideas that either support or hinder potential changes or limit transformations within the network itself (Koch et al. 2021a). Significant areas are yet to be explored regarding how environmental networks respond to and align with external, global, or political narratives, and how organizations manage the exchange of discourse among their peers. While these networks’ structures have broad meanings that facilitate cohesion and shared purpose, discourses represent specific expressions of ideas exchanged by organizations to articulate particular positions. Similar studies have addressed metrics that help in analyzing challenging environmental contexts, incorporating concordance and congruence analyses (Koch et al. 2021a). Concordance refers to the similarity of narratives between two organizations or individuals, while congruence is the integration of discourses, achieving a logical and comprehensible thematic and communicative flow. Identifying the relational patterns of functional cooperation and the exchange or comparison of narratives between organizations or nodes (at the dyadic links level) requires mapping and describing the multiple types of interactions within the environmental network. These patterns may reflect similarities in the attributes of a pair of nodes or their narratives (homophily) or, conversely, emerge as a greater tendency toward more diverse and complementary relationships (heterophily), both in organizations characteristics and discourses. Each pattern carries different implications for generating debates and exchanging opinions and viewpoints (Bodin et al. 2006; Prell 2016), the network’s governance and, consequently, in its operational plans. There are two scales of narratives among actors: the first is linked to social power, associated with local community nodes; the second is tied to the global technical discourse, reinforcing the hegemony and presence of international or transboundary actors. Both scales flow within the network, coexisting in the social interactions of the nodes. This dynamic leads to the construction of narratives with implications for representativeness in the decisions made within the network (Denton 2017a). Thematic categories within a network can reinforce authority through strategic discourse, leveraging an advantage in global scientific knowledge, while addressing the necessity for local input. However, organizations intervening in environmental matters bring their interests, experiences, and priorities, complicating the operational context within a multiscale system (international, regional, national, subregional). Participants must navigate dominant narratives, social trends, and policy changes, and simultaneously respond from their capacities and individualities to these dynamics. Narratives in scientific matters must be legitimized by communities to ensure they are understood, thereby avoiding misinterpretation or gaps between environmental and climate scientific discourses and public engagement (Bulkeley 2000; Leshner 2012; Reynolds et al. 2010). In networks, the common use of discourse by actors constructs a narrative network facilitated by the presence of a community that promotes it. Studies of social networks across various cases and contexts show that actors tend to forge links with others who are similar or who may align with their narratives, either directly or indirectly over time (McPherson et al. 2001). This dynamic contributes to the verbal interpretation of topics on environmental issues, that either foster ‘attraction’ among actors who frequently exchange opinions or ‘repulsion’ amongst those who do not. However, there are insufficient analytical and methodological tools to fully understand discourse exchanges and social interactions (Prell 2016) within environmental networks, including dominant narratives, alternate narratives, and other aspects—particularly when considering complex networks and interactions among multiple layers of organizations. Data and Methods Similar studies on climate change, water resource management, and environmental network governance, among others, have integrated methodologies using SNA, stakeholder analysis, Actor-Network Theory (ANT), and narrative analysis. While they employ different methods, they often converge on specific techniques (Koch et al. 2021c; Barone 2018; Denton 2017a). Exponential Random Graph Models (ERGMs) have also been used to study environmental governance, where connections between actors and shared narratives are intrinsically interdependent. These models allow analyzing and understanding how relationships in a network depend on one another, capturing patterns such as reciprocity, triad formation, and homophily (Bodin 2017; Koch et al. 2023a). In this study, an inferential network analysis is employed to explain the relational structure of the network, where the interactions of one node depend on the interactions of others, as well as the covariance of narratives alongside other relational variables. This assumption of interdependence violates the assumptions of standard statistical analyses, which is why ERGMs are used. These models are specifically designed to handle and explain such interdependent network data (Koch et al. 2021a, 2023a; Wang et al. 2016). The methodological flow explained in Figure S1 (Supplementary Material) covers: (a) identifying and using a network structure of organizations interested in the Amazon in Colombia and Peru, employing SNA techniques; (b) extracting and analyzing information published by each node’s website, virtual media, reports, and projects, through web scraping and subsequent text analysis; (c) constructing narrative profiles for each node, using text mining; (d) analyzing concordance among narratives within the network; and (e) using network modeling to evaluate the influence of narratives, their contributions, exchanges, and themes in actor relationships using ERGMs. These steps are presented in the following sections: Network and Social Structure Data collection was conducted using primary and secondary sources. Primary data was obtained through interviews with influential actors from organizations (nodes) engaged in activities in the Colombian-Peruvian Amazon region. Secondary data was gathered from information reported by respondents, including official websites, blogs, and other open resources. This data was collected in two phases: initially between 2021 and 2022, and subsequently in a second phase between 2022 and 2023. This approach allowed expanding and enriching the information base with appropriate resources for comprehensive data collection. Digital surveys were designed and implemented to record attributes and connections among actors. The process began by identifying organizations with expertise and key relationships rooted in the Amazon region. Respondents were selected from a database of organizations registered in two projects: (1) SERVIR-Amazonia, which promotes collaboration among governments and seeks to enhance local capacity by leveraging geospatial data and information across the Amazon (SERVIR Amazonia 2020); and (2) the Climate Services for Resilient Development (CSRD) Partnership, a public-private collaboration that develops tools, services, and approaches to strengthen the climate resilience of developing countries (CIAT 2020). The surveys were conducted through video calls and field visits in Tarapoto (Peru), and Mocoa and Leticia (Colombia), where information was collected from organizational partners operating in the Amazon. The data focused on cooperation within the Amazon, technologies used, and partners involved in using, developing, and transferring technologies. The entire process followed a convenience sampling method (Simkus 2022) (see Appendix S2, Supplementary Material). After identifying the associated actors mentioned in the surveys (see Appendix S1, Supplementary Material) and the technologies and innovations, the next step was to list the organizations cooperating or interacting with others on Amazon-related topics using available digital information. This exercise was conducted in two phases―in 2021 and in 2023―allowing for the update and expansion of the data required for analysis. The collected information included collective work relationships, mentions of support, funding, transfer, co-development among different nodes, technologies, and general data. A total of 66 interviews were conducted across two groups of actors: the first comprised scientific/administrative leaders (n = 42) and the second, technology developers (n = 24). Respondent profiles across both groups were distributed as follows: scientists (25.5%), managers/directors (24.5%), education (12.8%), agriculture (8.2%) and extension (7.7%), community/social work (4.6%), administrative (3.6%), communications (2.0%), and other (11.2%). Details on items, frequencies, and organizations are provided in Appendix S1 (See Supplementary Material). Regarding the secondary sources reviewed, 924 web sources (virtual information, official and open source) were examined and organized into categories. Many derive from official institutional websites, with particular emphasis on those detailing organizations’ key functions and activities (n = 265 sources) and general institutional information (n = 236 sources), which together illuminate projects, strategies, and areas of intervention. The remaining sources span institutional portfolios, areas of intervention, data and repositories, environmental research, impact, and funding—providing a comprehensive, up-to-date view of ongoing initiatives. For more details, see Appendix S2 in Supplementary Material. The mapped network identified 382 nodes connected to cooperation and the protection of the Colombian-Peruvian Amazon. Social network analysis was applied with a range of node (multipartite) types. This data structure also allowed gathering additional attributes at both the link and node levels, enhancing and strengthening the analysis (Prell 2016). This approach facilitated a more comprehensive representation of the interactions among actors united by common environmental themes (Bodin et al. 2006) and collective actions. Narratives and Text Analysis Organizational narratives were analyzed using an approach based on dyadic relationships, or node pairs, without assigning directionality to the relationships. This was based on the consideration that social interactions and narrative exchanges occur bidirectionally. Extracting discourses and determining the narrative contributions of each organization or node within the Amazon environmental network required processing large volumes of information. This offered both new opportunities and posed challenges in analyzing massive datasets and exploring different perspectives on narrative construction from secondary sources (R. Lejano et al. 2013; Segev 2020). To identify discourses, algorithms were developed to facilitate an information extraction model using web scraping from the “http” addresses of each organization or node. The model included websites containing: (a) organizational information; (b) ongoing or completed projects; (c) the organization’s topics of interest related to the environment, conservation, and other associated areas; (d) available reports on interventions, projects, and technologies under development or completed; and (e) connections with the Amazon, among others. The information was processed in R Studio 4.3.3 using various programming packages, including network and igraph (Csárdi et al. 2024), ggraph (Pedersen 2017), rvest (Wickham 2014), xml2 (Wickham et al. 2015), tidytext (Silge and Robinson 2016), stopwords (Benoit et al. 2017), and stringr (Wickham 2009). The texts were analyzed to extract keywords associated with topics of interest, calculating the contribution of each narrative by node (Fig. 1).Fig. 1Flow of text and narrative acquisition, and analysis by network node Flow of text and narrative acquisition, and analysis by network node Categorizing the Narratives by Node To establish a comparable baseline for each organization or node, a group of 17 topics or narratives and associated keywords was created. This was based on expert suggestions and the topics identified from the analysis of the results obtained. Narrative concordance was calculated for each of the 17 narratives for each relationship between nodes (dyadic level). This means that, for each theme, two pairs of nodes are explored for proximity, distance, and absence of common themes in common narratives. These are: Environmental, Forestry; Challenges; Animal; Water; Climate; Food System; Agriculture & Livestock; Communities; Network; Region; Research & Education; Donor; Research; Technologies & Innovation; Policy, and Economy. Tables S1 and S2 (Supplementary Materials) present these categories with brief descriptions of their content and a sample list of terms and keywords associated with each narrative category. For this study, the construction of thematic sets was based on specialized dictionaries (Alliance Bioversity & CIAT 2010; HUMBOLDT 2024; IDEAM 2024; PNUD 2023; UNEP 2017; WWF 2024), which were defined through expert consultations and secondary information. Additionally, due to the presence of international actors in the network, the word-sets were developed in Spanish, English, and Portuguese. Through these categories, the texts analyzed by node allowed us to identify the number of words associated with each topic, facilitating the calculation of proportions, word counts, and overall trends in the discourses. Matching Analysis by Pairs of Nodes or Dyadic Relationships We conducted a higher-resolution analysis of narrative trends and their significance in shaping cooperative relationships. We evaluated the concordance of each theme between pairs of actors, by comparing the percentage of their contribution to the total. This study did not address narrative congruence, nor did it aim to simplify or synthesize all trends into a single variable, as developed in the study by Koch et al. (2023a). This decision was influenced by the number of nodes involved, the geographic factor, and the accessibility of all the nodes’ data, in order to gather information through surveys and secondary sources. Instead, our proposed methodological approach offers a lower-cost alternative, with broader coverage and greater replicability in analyses, enabling periodic monitoring. The approach describes the interactions between pairs of nodes, focusing on comparing their narratives, which may align in varying proportions across one or more themes. As an example, Fig. 2 illustrates the relationship between two nodes with different narrative proportions. Node B has a higher proportion of discourse in the agriculture narrative (50%), while Node A participates to a lesser extent (20%). Conversely, the Food System narrative accounts for 50% in Node A, compared to 30% in Node B. This comparative analysis is performed for each pair of nodes in the mapped network where a connection or relationship in the Amazon was identified.Fig. 2Dyadic relations and intensive narratives by node Dyadic relations and intensive narratives by node The concordance between narratives used in this article is based on the distance between proportions, which is the difference in narrative contributions for a category among the 17 topics between two nodes, A and B, with already identified cooperative links. Figure 3 proposes three types or scales of concordance: (a) “Closeness between narratives” where both nodes show some similarity in their contribution to the evaluated \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{topic}}_{1}$$\end{document}topic1 (e.g., Food System); (b) “Distance between narratives”, where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{topic}}_{2}$$\end{document}topic2 exhibits differences in contributions, identifying a level of dissimilarity between the nodes (e.g., Agriculture); and (c) “Uncommon narrative”, where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{topic}}_{3}$$\end{document}topic3 is present in one node, but absent in the other (e.g., Economy).Fig. 3Graphical representation of narrative concordance between dyadic node relations Graphical representation of narrative concordance between dyadic node relations The specification of the matching types is expressed as follows: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$i,{j}$$\end{document}i, j being two nodes and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${Na}$$\end{document}Na a category of narratives. The difference is calculated as \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${Dif}{f}_{N{a}_{i,j}}= \% N{a}_{i}\,- \% N{a}_{j}$$\end{document}DiffNai,j=%Nai−%Naj, then if \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${Dif}{f}_{N{a}_{i,j}\,}{is\; on\; the\; range}\pm \mu ,\sigma$$\end{document}DiffNai,jis on the range±μ,σ of all the differences in distribution for narrative A, concordance is defined as “closeness between narratives”. Otherwise, if \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${Dif}{f}_{N{a}_{i,j}\,}$$\end{document}DiffNai,j is out of this range, the concordance is “Distance between narratives”, while when there is the absence of the topic in one of the pairs of nodes with cooperative links \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${Dif}{f}_{N{a}_{i,j}},$$\end{document}DiffNai,j, it is defined as “Uncommon Narrative”, since it implies that it is not possible to compute it. Additionally, within the portfolio of available social-network metrics used to identify the dominance of themes within the network by node, the following centrality measures were applied: Closeness and Betweenness, which differ in their interpretative implications. The first measure, Closeness, is understood as an indicator of the efficiency with which information can be disseminated sequentially from one node to all others (Opsahl et al. 2010). The second, Betweenness, quantifies the number of times a node acts as a bridge along the shortest path between two other nodes (Newman 2005). Exponential Random Graph Models By combining qualitative and quantitative data in an ERGM, this study offers innovative perspectives on how and what social dynamics influence the emergence of a common narrative among heterogeneous actors in the context of environmental cooperation (Koch et al. 2023b). Beyond the metrics presented, it is essential to model the network by exploring the contribution levels of exogenous and endogenous variables within the mapped relational structure. To enhance network functionality, ERGMs were applied with hypothesis testing based on their exponential distribution. Individual network covariates (i.e., node attributes) and structural properties of the network (such as transitive triplets) are useful for predicting overall network properties. Network modeling techniques were employed to determine whether specific organizational attributes, such as narrative contribution and vertex typology, influence the likelihood of establishing connections among nodes within the Colombian-Peruvian Amazon cooperation and technology network. The specification of the ERGMs can be found in Appendix S3 and S4 (Supplementary Material). Model Specification and Hypothesis Evaluation Identifying the variables and relationships that contributed to understanding the network’s functionality, the programming language functions used for modeling, and the proposed model are listed. It is important to note that, given the size of the network and the possible number of statistics to be included, ERGMs can easily become overfitted and produce degeneration issues, a factor carefully considered during specification. On the other hand, given the complexity and coverage of the data, rather than inferring temporal precedence, the analysis focuses on cross-sectional associations, specifically on the relationship between shared narratives and the presence of social or collaborative ties. The evaluated variables were connected to guiding questions, along with a more detailed explanation of the associated hypotheses, as follows: Does the form of relations between the participants in the Colombian-Peruvian Amazon network influence the formation of cooperative links? Hypothesis 1: Centrality (H1): The tendency of connection between two nodes in the Colombian-Peruvian network influences the formation of cooperative ties. Nodes with high centrality―those with numerous social connections—tend to be more visible and well-informed about the network’s activities. These nodes are considered influential due to their access to valuable information and resources (Bodin and Crona 2009a). Therefore, the greater the number of social connections, the higher an actor’s capacity to co-create a shared narrative and form cohesive (sub)groups (Koch et al. 2021a). This network centralization trend, where multiple edges converge on a small number of popular nodes, is an important variable to be considered. In ERGM, this is primarily modeled using the Geometrically Weighted Degree (GWD) statistic (A Levy 2016; Krivitsky et al. 2022; Snijders et al. 2006) and is referred to as ‘gwdegree’.Hypothesis 2: Triadic closure (H2): The configuration of closed triangles, an expression of the principle of triadic closure in network analysis, influences the formation of relationships and cooperation structures in the Amazon. Triadic closure refers to the tendency of two nodes that share a common neighbor to form a direct connection, creating a closed triangle or triad. This dynamic micro-level process leads to clustering relationships into tightly connected groups. The concept of closed triangle configuration is used to explore how clustering relationships into this geometric form impacts cooperation and relational structures. These cliques or substructures reflect levels of trust among nodes and the circulation of information and actions. In ERGM modeling language, the term ‘gwesp’ (Geometrically Weighted Edgewise Shared Partners) is used. A positive resulting coefficient indicates a higher likelihood of trust in the observed network’s triangles, beyond what would be expected by chance. This supports the hypothesis that trust is more likely to occur within subgroups (i.e., cliques) in the network (Gorris and Koch 2024). Hypothesis 1: Centrality (H1): The tendency of connection between two nodes in the Colombian-Peruvian network influences the formation of cooperative ties. Nodes with high centrality―those with numerous social connections—tend to be more visible and well-informed about the network’s activities. These nodes are considered influential due to their access to valuable information and resources (Bodin and Crona 2009a). Therefore, the greater the number of social connections, the higher an actor’s capacity to co-create a shared narrative and form cohesive (sub)groups (Koch et al. 2021a). This network centralization trend, where multiple edges converge on a small number of popular nodes, is an important variable to be considered. In ERGM, this is primarily modeled using the Geometrically Weighted Degree (GWD) statistic (A Levy 2016; Krivitsky et al. 2022; Snijders et al. 2006) and is referred to as ‘gwdegree’. Hypothesis 1: Centrality (H1): The tendency of connection between two nodes in the Colombian-Peruvian network influences the formation of cooperative ties. Nodes with high centrality―those with numerous social connections—tend to be more visible and well-informed about the network’s activities. These nodes are considered influential due to their access to valuable information and resources (Bodin and Crona 2009a). Therefore, the greater the number of social connections, the higher an actor’s capacity to co-create a shared narrative and form cohesive (sub)groups (Koch et al. 2021a). This network centralization trend, where multiple edges converge on a small number of popular nodes, is an important variable to be considered. In ERGM, this is primarily modeled using the Geometrically Weighted Degree (GWD) statistic (A Levy 2016; Krivitsky et al. 2022; Snijders et al. 2006) and is referred to as ‘gwdegree’. Hypothesis 2: Triadic closure (H2): The configuration of closed triangles, an expression of the principle of triadic closure in network analysis, influences the formation of relationships and cooperation structures in the Amazon. Triadic closure refers to the tendency of two nodes that share a common neighbor to form a direct connection, creating a closed triangle or triad. This dynamic micro-level process leads to clustering relationships into tightly connected groups. The concept of closed triangle configuration is used to explore how clustering relationships into this geometric form impacts cooperation and relational structures. These cliques or substructures reflect levels of trust among nodes and the circulation of information and actions. In ERGM modeling language, the term ‘gwesp’ (Geometrically Weighted Edgewise Shared Partners) is used. A positive resulting coefficient indicates a higher likelihood of trust in the observed network’s triangles, beyond what would be expected by chance. This supports the hypothesis that trust is more likely to occur within subgroups (i.e., cliques) in the network (Gorris and Koch 2024). Hypothesis 2: Triadic closure (H2): The configuration of closed triangles, an expression of the principle of triadic closure in network analysis, influences the formation of relationships and cooperation structures in the Amazon. Triadic closure refers to the tendency of two nodes that share a common neighbor to form a direct connection, creating a closed triangle or triad. This dynamic micro-level process leads to clustering relationships into tightly connected groups. The concept of closed triangle configuration is used to explore how clustering relationships into this geometric form impacts cooperation and relational structures. These cliques or substructures reflect levels of trust among nodes and the circulation of information and actions. In ERGM modeling language, the term ‘gwesp’ (Geometrically Weighted Edgewise Shared Partners) is used. A positive resulting coefficient indicates a higher likelihood of trust in the observed network’s triangles, beyond what would be expected by chance. This supports the hypothesis that trust is more likely to occur within subgroups (i.e., cliques) in the network (Gorris and Koch 2024). Is the creation of cooperative relationships in the Colombian-Peruvian Amazon related to the concordance in the node narratives? Hypothesis 3: Narrative Concordance (H3): The similarity in narrative sets, or concordance between two nodes, is significant in forming dyadic cooperative relationships within the biome. The greater the thematic similarity in narrative weights between two nodes, the more likely they are to establish connections. This hypothesis is incorporated into the model as an explanatory variable or structural covariate, representing a network that only includes links between nodes with narrative similarity in more than seven topics, forming a subnetwork of narrative concordance. In ERGM, the term ‘edgecov’ is used, adding a statistic to the model that equals the sum of covariate values for each edge present in the network (Morris et al. 2008; Van der Hulst 2009). Evidence suggests that sharing narratives and common themes facilitates communication and fosters mutual understanding among actors, which are essential factors for effective cooperation (Leifeld 2013; Moore and Westley 2011). Moreover, in the context of natural resource governance and environmental sustainability, thematic similarity not only enhances coordination, but also strengthens the resilience of social networks by enabling the formation of effective coalitions (Moore and Westley 2011). Hypothesis 3: Narrative Concordance (H3): The similarity in narrative sets, or concordance between two nodes, is significant in forming dyadic cooperative relationships within the biome. The greater the thematic similarity in narrative weights between two nodes, the more likely they are to establish connections. This hypothesis is incorporated into the model as an explanatory variable or structural covariate, representing a network that only includes links between nodes with narrative similarity in more than seven topics, forming a subnetwork of narrative concordance. In ERGM, the term ‘edgecov’ is used, adding a statistic to the model that equals the sum of covariate values for each edge present in the network (Morris et al. 2008; Van der Hulst 2009). Evidence suggests that sharing narratives and common themes facilitates communication and fosters mutual understanding among actors, which are essential factors for effective cooperation (Leifeld 2013; Moore and Westley 2011). Moreover, in the context of natural resource governance and environmental sustainability, thematic similarity not only enhances coordination, but also strengthens the resilience of social networks by enabling the formation of effective coalitions (Moore and Westley 2011). Hypothesis 3: Narrative Concordance (H3): The similarity in narrative sets, or concordance between two nodes, is significant in forming dyadic cooperative relationships within the biome. The greater the thematic similarity in narrative weights between two nodes, the more likely they are to establish connections. This hypothesis is incorporated into the model as an explanatory variable or structural covariate, representing a network that only includes links between nodes with narrative similarity in more than seven topics, forming a subnetwork of narrative concordance. In ERGM, the term ‘edgecov’ is used, adding a statistic to the model that equals the sum of covariate values for each edge present in the network (Morris et al. 2008; Van der Hulst 2009). Evidence suggests that sharing narratives and common themes facilitates communication and fosters mutual understanding among actors, which are essential factors for effective cooperation (Leifeld 2013; Moore and Westley 2011). Moreover, in the context of natural resource governance and environmental sustainability, thematic similarity not only enhances coordination, but also strengthens the resilience of social networks by enabling the formation of effective coalitions (Moore and Westley 2011). Is the contribution of a node-level narrative important in shaping relationships in the Amazon network? Hypothesis 4, Intensity: Highest Contribution in a Narrative (H4): The highest percentage of contribution expressed by a node in a specific narrative (e.g., Agriculture & Livestock, Forestry, Animals, Climate, etc.) influences the formation of relationships within the network. The key to understanding the cooperation, collaboration, and environmental governance processes lies in the power within the environmental networks, and the key role that trust and constancy play in their discourses (Gorris and Koch 2024). Narratives within an environmental network are constantly under negotiation, providing signals about its functionality, technical pathways, collaboration, internal roles, and more. Narratives that dominate this system of cooperation at an organizational level are closely tied to expertise, authority, and experience in the topic (Denton 2017a). This suggests that an edge is more likely to form when the organization (node i) has a particularly high narrative contribution. In the model, this is represented in ERGM using the term ‘Nodecov.’ Hypothesis 4, Intensity: Highest Contribution in a Narrative (H4): The highest percentage of contribution expressed by a node in a specific narrative (e.g., Agriculture & Livestock, Forestry, Animals, Climate, etc.) influences the formation of relationships within the network. The key to understanding the cooperation, collaboration, and environmental governance processes lies in the power within the environmental networks, and the key role that trust and constancy play in their discourses (Gorris and Koch 2024). Narratives within an environmental network are constantly under negotiation, providing signals about its functionality, technical pathways, collaboration, internal roles, and more. Narratives that dominate this system of cooperation at an organizational level are closely tied to expertise, authority, and experience in the topic (Denton 2017a). This suggests that an edge is more likely to form when the organization (node i) has a particularly high narrative contribution. In the model, this is represented in ERGM using the term ‘Nodecov.’ Hypothesis 4, Intensity: Highest Contribution in a Narrative (H4): The highest percentage of contribution expressed by a node in a specific narrative (e.g., Agriculture & Livestock, Forestry, Animals, Climate, etc.) influences the formation of relationships within the network. The key to understanding the cooperation, collaboration, and environmental governance processes lies in the power within the environmental networks, and the key role that trust and constancy play in their discourses (Gorris and Koch 2024). Narratives within an environmental network are constantly under negotiation, providing signals about its functionality, technical pathways, collaboration, internal roles, and more. Narratives that dominate this system of cooperation at an organizational level are closely tied to expertise, authority, and experience in the topic (Denton 2017a). This suggests that an edge is more likely to form when the organization (node i) has a particularly high narrative contribution. In the model, this is represented in ERGM using the term ‘Nodecov.’ Is the formation of links between two nodes related to a similar contribution and theme in the Colombian-Peruvian Amazon? Hypothesis 5: Narrative Homophily (H5): Two nodes with similar contributions or thematic interests in the same narrative or topic exhibit characteristics that facilitate cooperation and collaboration. The formation of dyadic connections based on similar node attributes is referred to as ‘homophily’ (Howe et al. 2023; Koch et al. 2021c; Phillips et al. 2013). Therefore, similarity in narrative weighting between two nodes is crucial for creating cooperative ties within the network. Let \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${x}_{i}$$\end{document}xi and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${x}_{j}$$\end{document}xj represent the narrative contributions for nodes \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$i$$\end{document}i and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$j$$\end{document}j, respectively. Considering that each node has a percentage of distribution across various themes, this statistic evaluates the likelihood of forming a dyadic connection between nodes \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$i$$\end{document}i and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$j$$\end{document}j. In the ERGM used, this is represented by the term ‘absdiff.’ Hypothesis 5: Narrative Homophily (H5): Two nodes with similar contributions or thematic interests in the same narrative or topic exhibit characteristics that facilitate cooperation and collaboration. The formation of dyadic connections based on similar node attributes is referred to as ‘homophily’ (Howe et al. 2023; Koch et al. 2021c; Phillips et al. 2013). Therefore, similarity in narrative weighting between two nodes is crucial for creating cooperative ties within the network. Let \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${x}_{i}$$\end{document}xi and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${x}_{j}$$\end{document}xj represent the narrative contributions for nodes \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$i$$\end{document}i and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$j$$\end{document}j, respectively. Considering that each node has a percentage of distribution across various themes, this statistic evaluates the likelihood of forming a dyadic connection between nodes \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$i$$\end{document}i and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$j$$\end{document}j. In the ERGM used, this is represented by the term ‘absdiff.’ Hypothesis 5: Narrative Homophily (H5): Two nodes with similar contributions or thematic interests in the same narrative or topic exhibit characteristics that facilitate cooperation and collaboration. The formation of dyadic connections based on similar node attributes is referred to as ‘homophily’ (Howe et al. 2023; Koch et al. 2021c; Phillips et al. 2013). Therefore, similarity in narrative weighting between two nodes is crucial for creating cooperative ties within the network. Let \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${x}_{i}$$\end{document}xi and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${x}_{j}$$\end{document}xj represent the narrative contributions for nodes \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$i$$\end{document}i and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$j$$\end{document}j, respectively. Considering that each node has a percentage of distribution across various themes, this statistic evaluates the likelihood of forming a dyadic connection between nodes \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$i$$\end{document}i and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$j$$\end{document}j. In the ERGM used, this is represented by the term ‘absdiff.’ Is shaping relationships between two nodes related to a specific combination of narratives by node? Hypothesis 6: Homophily by Node Type (H6): Two nodes with similar organizational attributes or characteristics are more likely to establish cooperative connections within the Amazon network. Categorical attributes for nodes, such as national agricultural research systems (NARS), national government agencies, international non-governmental organizations (NGOs), grassroots organizations, private companies, universities, and others, indicate whether nodes in the network tend to form connections with those that share a specific attribute (Hunter and Goodreau 2019). In ERGMs, this propensity is captured using the statistic ‘Nodematch.’ Hypothesis 6: Homophily by Node Type (H6): Two nodes with similar organizational attributes or characteristics are more likely to establish cooperative connections within the Amazon network. Categorical attributes for nodes, such as national agricultural research systems (NARS), national government agencies, international non-governmental organizations (NGOs), grassroots organizations, private companies, universities, and others, indicate whether nodes in the network tend to form connections with those that share a specific attribute (Hunter and Goodreau 2019). In ERGMs, this propensity is captured using the statistic ‘Nodematch.’ Hypothesis 6: Homophily by Node Type (H6): Two nodes with similar organizational attributes or characteristics are more likely to establish cooperative connections within the Amazon network. Categorical attributes for nodes, such as national agricultural research systems (NARS), national government agencies, international non-governmental organizations (NGOs), grassroots organizations, private companies, universities, and others, indicate whether nodes in the network tend to form connections with those that share a specific attribute (Hunter and Goodreau 2019). In ERGMs, this propensity is captured using the statistic ‘Nodematch.’ Table 1 summarizes the types of analyses based on the formulated questions and hypotheses, highlighting the types of statistics as follows: node-based covariates (1), which are associated with attributes specific to each nodal entity; dyadic-level covariates (2), corresponding to variables evaluable at the level of relationships between pairs of nodes; and structural or triangular-level covariates (3), representing parameters associated with forms of relationships between nodes, such as triangles, stars, and other configurations. This classification facilitates the organization and interpretation of the analytical framework for examining the network.Table 1Summary of hypotheses and variables used to test them in the Exponential Random Graph Model (ERGM)HypothesisVariableGraphical representation (Nodes)edges\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\boldsymbol{\theta }}}_{{\boldsymbol{1}}}{\boldsymbol{edges}}$$\end{document}θ1edges This is the basic element of the networks that controls the density. Type: Structural statistic (3)Centrality (H1)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\boldsymbol{\theta }}}_{{\boldsymbol{2}}}{\boldsymbol{GWdegree}}$$\end{document}θ2GWdegree This is a star-like configuration; a nodal covariant representing the degree of centrality value in the cooperation network. Trust relationships tend to occur around facilitators.Type: Structural statistics (3)Triadic structures (H2)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\boldsymbol{\theta }}}_{{\boldsymbol{3}}}{\boldsymbol{GWesp}}$$\end{document}θ3GWesp This is the propensity of connections to form multiple shared cyclic pairs and close triangles.Type: Structural statistics (3)Narrative concordance (H3)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\boldsymbol{\theta }}}_{{\boldsymbol{4}}}{\boldsymbol{Concordance}}$$\end{document}θ4Concordance statistic that allows us to consider the concordance in a set of narratives as important in the formation of cooperative links in the network.Type: Covariate linkage statistic (1)Intensity: Highest contribution in a narrative (H4)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\boldsymbol{\beta }}}_{{\boldsymbol{1}}{\boldsymbol{ \% }}{\boldsymbol{Narrative}}}$$\end{document}β1%Narrative Degree effect in response to a specific attribute of the node. In the case of the Amazon network, 17 narrative weights (%) are evaluated.Type: Node covariate statistic (1)Narrative Homophily (H5)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\boldsymbol{\beta }}}_{{\boldsymbol{2}}{\boldsymbol{Homophil}}{{\boldsymbol{y}}}_{{\boldsymbol{Narrative}}}}$$\end{document}β2HomophilyNarrativeHomophily at the level of dyadic linkages between a pair of nodes, determined by assessing the coincidence and similarities per narrative.Type: Statistical dyadic structures (2)Homophily by node type (H6)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\boldsymbol{\beta }}}_{{\boldsymbol{3}}{\boldsymbol{Homophi}}{{\boldsymbol{ly}}}_{{\boldsymbol{type\_organization}}}}$$\end{document}β3Homophilytype_organization Homophily by organization type (NARS, foundation, Government agency, Technological nodes, etc.).Type: Statistical dyadic structures (2) Table 1Summary of hypotheses and variables used to test them in the Exponential Random Graph Model (ERGM)HypothesisVariableGraphical representation (Nodes)edges\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\boldsymbol{\theta }}}_{{\boldsymbol{1}}}{\boldsymbol{edges}}$$\end{document}θ1edges This is the basic element of the networks that controls the density. Type: Structural statistic (3)Centrality (H1)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\boldsymbol{\theta }}}_{{\boldsymbol{2}}}{\boldsymbol{GWdegree}}$$\end{document}θ2GWdegree This is a star-like configuration; a nodal covariant representing the degree of centrality value in the cooperation network. Trust relationships tend to occur around facilitators.Type: Structural statistics (3)Triadic structures (H2)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\boldsymbol{\theta }}}_{{\boldsymbol{3}}}{\boldsymbol{GWesp}}$$\end{document}θ3GWesp This is the propensity of connections to form multiple shared cyclic pairs and close triangles.Type: Structural statistics (3)Narrative concordance (H3)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\boldsymbol{\theta }}}_{{\boldsymbol{4}}}{\boldsymbol{Concordance}}$$\end{document}θ4Concordance statistic that allows us to consider the concordance in a set of narratives as important in the formation of cooperative links in the network.Type: Covariate linkage statistic (1)Intensity: Highest contribution in a narrative (H4)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\boldsymbol{\beta }}}_{{\boldsymbol{1}}{\boldsymbol{ \% }}{\boldsymbol{Narrative}}}$$\end{document}β1%Narrative Degree effect in response to a specific attribute of the node. In the case of the Amazon network, 17 narrative weights (%) are evaluated.Type: Node covariate statistic (1)Narrative Homophily (H5)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\boldsymbol{\beta }}}_{{\boldsymbol{2}}{\boldsymbol{Homophil}}{{\boldsymbol{y}}}_{{\boldsymbol{Narrative}}}}$$\end{document}β2HomophilyNarrativeHomophily at the level of dyadic linkages between a pair of nodes, determined by assessing the coincidence and similarities per narrative.Type: Statistical dyadic structures (2)Homophily by node type (H6)\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\boldsymbol{\beta }}}_{{\boldsymbol{3}}{\boldsymbol{Homophi}}{{\boldsymbol{ly}}}_{{\boldsymbol{type\_organization}}}}$$\end{document}β3Homophilytype_organization Homophily by organization type (NARS, foundation, Government agency, Technological nodes, etc.).Type: Statistical dyadic structures (2) Summary of hypotheses and variables used to test them in the Exponential Random Graph Model (ERGM) The model in its general expression of the Colombian-Peruvian Amazon network and the incorporation of endogenous and exogenous variables can be formulated as follows for the Amazon network (see Eq. 1):1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\begin{array}{lll}p({\boldsymbol{y}}| {\theta }_{1},{\theta }_{2},{\theta }_{3},{\theta }_{4},\beta )&=&\frac{1}{\kappa \left({\theta }_{1},{\theta }_{2},{\theta }_{3},\beta \right)}\exp \left\{{\theta }_{1}edges({\boldsymbol{y}})\right.\\ &&+\,{\theta }_{2}GWdegree\left({\boldsymbol{y}}\right)+\theta_{3}\mathrm{GWesp}\left({\boldsymbol{y}}\right)\\&&\left.+\,{\theta }_{4}Concordance\left({\boldsymbol{y}}\right)+{\beta }^{T}{\boldsymbol{g}}({\boldsymbol{y}},{\boldsymbol{x}})\right\}\end{array}$$\end{document}p(y∣θ1,θ2,θ3,θ4,β)=1κθ1,θ2,θ3,βexpθ1edges(y)+θ2GWdegreey+θ3GWespy+θ4Concordancey+βTg(y,x)Where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\theta }_{1}$$\end{document}θ1 is the coefficient associated with the (edges), \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\theta }_{2}$$\end{document}θ2 is the integrated coefficient that captures the structural patterns of the network associated with Centrality (H1), \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\theta }_{3}$$\end{document}θ3 corresponds to Triadic Structures (H2), and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\theta }_{4}$$\end{document}θ4 represents Narrative Concordance (H3). Regarding the exogenous variables of the network, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${\beta }_{n}$$\end{document}βn represents the nodal attributes used to evaluate the following: (a) Intensity, or the topic with the highest contribution at the narrative level (H4); (b) Narrative Homophily (H5); and (c) Homophily by Node Type (H6). Results Figure 4 illustrates the mapped network, consisting of 382 nodes, including Colombian and Peruvian organizations, international organizations, technological nodes, associations, and platforms, with a total of 783 interactions. From these interactions, we derived the narrative patterns of each node, along with their percentage contributions, represented as thematic weights based on the previously defined categories. Additionally, the most prominent types of organizations (see Table 2) are government entities, research centers, and technology centers, accounting for 24%, 16%, and 13% of the nodes, respectively.Fig. 4Mapping of nodes in the Colombian-Peruvian AmazonTable 2Distribution of nodes by typologyAttributes of node organization types# nodes%Government (National−Subnational)9024%Research organizations and universities (General−International)6116%Technology5113%NGOs―International (General)4712%Organizations (other than financial or research) (Regional−International)4111%NGOs―national, local, & regional (Farmers−General)3810%Research organizations and universities (National-Regional Universities)195%Foundations144%Private companies (other than financial)82%Financial Institution(International)72%Research organizations and universities,National (NARS)51%Research Organizations and Universities (International (CGIAR))10%Total382100% Mapping of nodes in the Colombian-Peruvian Amazon Table 2Distribution of nodes by typologyAttributes of node organization types# nodes%Government (National−Subnational)9024%Research organizations and universities (General−International)6116%Technology5113%NGOs―International (General)4712%Organizations (other than financial or research) (Regional−International)4111%NGOs―national, local, & regional (Farmers−General)3810%Research organizations and universities (National-Regional Universities)195%Foundations144%Private companies (other than financial)82%Financial Institution(International)72%Research organizations and universities,National (NARS)51%Research Organizations and Universities (International (CGIAR))10%Total382100% Distribution of nodes by typology Note: Centrality by degree indicates the positioning of a node based on the number of interactions and connections, with larger node sizes corresponding to higher centrality. Nodes are categorized as follows: ‘Others’, which includes countries from Oceania, South America, and Asia that generally exhibit low interaction; ‘Technology’ corresponds to technological nodes; ‘Network’ represents nodes such as associations, organizational platforms, and organizational networks; ‘International’ represents nodes or organizations that have a presence in different countries. The actor mapping reveals that at least 53% of the nodes belong to the categories of national and regional government agencies, ministries, and offices, as well as international research centers, universities, and technological nodes (see a detailed list in Table S5, Supplementary Material). Dominant Narratives in the Colombian-Peruvian Amazon Cooperation and Technology Network The most dominant narratives represented by the nodes in the conservation and technology network of the biome are primarily associated with terminology related to Research & Education, averaging 13.72% of the discourse proportions. This is followed by the Communities terminology (comprising gender, Indigenous and peasant communities, families, traditions) (12.47%), contributions related to Policy (covering a broad portfolio of terms and actions inherent to the political sphere, such as laws, legal, judicial expressions) (12.31%), and lastly narratives tied to Technologies & Innovation that include concepts related to spatial analysis, innovations, and software platforms (10.14%). These dominant narratives account for 50% of the content represented, transmitted, and shared by the operational nodes in the Amazon (see Fig. 5).Fig. 5Distribution of nodes by narrative category Distribution of nodes by narrative category Differences in Discursive Topics Across Node Categories When comparing the narratives of the Colombian and Peruvian nodes with those of international and technological nodes, we see that both countries share similar patterns in narrative composition and contribution, particularly for Research & Education (13%), Policy (14%), and Donor (6%). Whereas there is a higher contribution to the Technologies & Innovation narrative, with 14% and 10% of nodes in Brazil and the U.S., respectively, compared to local nodes in Peru and Colombia. Regarding the Communities narratives (15% in Colombia, 13% in Peru), local nodes in both countries exhibit a greater narrative weight than for other nodes (see Fig. 6).Fig. 6Contribution of narratives by node types (%). Red squares highlight the Research & Education narratives, and blue squares highlight the Communities narratives Contribution of narratives by node types (%). Red squares highlight the Research & Education narratives, and blue squares highlight the Communities narratives Combination of Narratives in Cooperative Relationships within the Network The results reveal that the relationships or links with the most concordant themes are Communities (497), Research & Education (485), Policy (402), Environmental (288), and Technologies & Innovation (251) (Fig. 7). There is a mutual relationship regarding discourse priorities among the nodes operating in the Amazon, particularly in narratives related to Research & Education, Communities, and Technologies & Innovation. This indicates that, in their relationships and interactions with their partners, organizations and technological nodes prioritize these topics or assign them greater weight in their communications and documents.Fig. 7Graph of the most relevant matching concordant narrative themes in the Amazon network Graph of the most relevant matching concordant narrative themes in the Amazon network An important finding regarding the narratives mediated by exchanges between nodes in the network is the role of Technologies & Innovation as the narrative category most closely linked to other themes through exchanges or cooperative interactions. The proximity of Technologies & Innovation and Research & Education emerges as the most dominant discursive components in the Colombian-Peruvian Amazon. These two categories influence debates and interactions; not only were they prioritized in the nodes’ discourses, but they were also found to hold a central position in the analysis of the overall themes. Conversely, the most distant themes were Animals, Water, and Energy. This assessment is based on the primary narrative priorities of each node, reflecting the most significant exchanges for each mapped node in the network (Table 3).Table 3Centrality metrics by narrativeNo.LabelClosenessBetweenness1Technologies &Innovation0.5000.5292Research &Education0.3330.5293Region0.2500.4824Policy0.2000.5295Network0.1670.5296Forestry0.1430.3917Environmental0.1110.5298Food System0.1110.2109Donor0.0910.52910Economy0.0910.09011Communities0.0830.52912Climate0.0770.52913Challenges0.0710.52914Agriculture & livestock0.0590.06715Animals0.0590.00016Water0.0000.00017Energy0.0000.000 Table 3Centrality metrics by narrativeNo.LabelClosenessBetweenness1Technologies &Innovation0.5000.5292Research &Education0.3330.5293Region0.2500.4824Policy0.2000.5295Network0.1670.5296Forestry0.1430.3917Environmental0.1110.5298Food System0.1110.2109Donor0.0910.52910Economy0.0910.09011Communities0.0830.52912Climate0.0770.52913Challenges0.0710.52914Agriculture & livestock0.0590.06715Animals0.0590.00016Water0.0000.00017Energy0.0000.000 Centrality metrics by narrative Concordance within Narratives The results show greater proximity in dyadic relationships for the narratives on Communities (76.6%), Environmental (76.8%), and Networks (87%). Conversely, the topics with the greatest distance in the narrative contribution are Animals (47.1%), Agriculture & Livestock (37.9%), Forestry (21.3%), and Energy (33%) (Fig. 8).Fig. 8Concordance by narrative in the Amazon network Concordance by narrative in the Amazon network The concordance levels allow adding attributes to the dyadic relationships between nodes in the mapped network structure. These graphs (Fig. 9) visualize how relational structures in Amazonia are not completely homogeneous in terms of thematic interests (red dots represent the nodes). The results indicate that the Colombian-Peruvian Amazon network is based on relationships between nodes with significant concordance in numerous narratives; collaborative relationships were also identified between nodes with differences in narrative weights and uncommon themes. These three structures (Fig. 9) are included in the network modeling as covariates of the edges in the comprehensive network, aiming to explore their impact on forming connections.Fig. 9Graphs of narrative concordance levels in the Colombian-Peruvian Amazon cooperation network Graphs of narrative concordance levels in the Colombian-Peruvian Amazon cooperation network Network Modeling This section presents the results of the model testing the hypotheses related to relational structures, contribution, homophily, and networks of narrative concordance (closeness, distance, and uncommon), as well as homophily by node type in the formation of interpersonal trust relationships. The results of the hypothesis significance tests are presented in Table 4. Evaluating the 17 narrative categories produces an extensive model response, of which only the most relevant and pertinent aspects are analyzed. The goodness-of-fit diagnostics indicate that the model is appropriate and adequately represents trends, edge proportions, and minimum geodesic distances (see Appendix S5, Supplementary Material). Meanwhile, the Markov Chain Monte Carlo (MCMC) diagnostics confirm the quality and accuracy of the parameters obtained in the modeling (see Appendix S6, Supplementary Material).Table 4Results summary of the ERGM modelHypothesesResultsEdges (control)–3.504***Centrality Hypothesis (H1)–0.0215Triadic structures (H2)0.8219***Narrative Concordance Hypothesis (H3)0.794**Highest Contribution in a Narrative Hypothesis (H4)Challenges(H4)–0.02218*Climate(H4)0.03153***Economy(H4)–0.0883***Environmental(H4)0.01579**Forestry(H4)0.04253***Network(H4)–0.01655*Narrative Homophily Hypothesis (H5)Agriculture & livestock(H5)–0.0574**Challenges(H5)0.0376**Communities(H5)–0.03367***Economy(H5)0.0244+Energy(H5)0.0662*Environmental(H5)–0.01539+Forestry(H5)–0.0253+Policy(H5)–0.03357***Research & Education(H5)–0.02360***Technologies & Innovation(H5)–0.01670*Water(H5)–0.0576**Homophily by node type Hypothesis (H6)–0.250*Num.Obs.73153AIC7749.7BIC8126.9MC Std. Err. = 0.4039Null Deviance: 101412 on 73153 degrees of freedomResidual Deviance: 7668 on 73112 degrees of freedom+ p < 0.1, * p < 0.05, ** p < 0.01, *** p < 0.001AIC akaike information criterion, BIC Bayesian Information Criterion, Num.Obs. number of observations.For Hypothesis H3, the network with narrative distance and uncommon themes was not significant. For Hypothesis H4, the narratives Agriculture & Livestock, Animals, Communities, Donor, Energy, Food System, Policy, Region, Research & Education, Technologies & Innovation, and Water was not significant. For Hypothesis H5, the narratives Animals, Donor, Network, and Region were not reported as significant. Table 4Results summary of the ERGM modelHypothesesResultsEdges (control)–3.504***Centrality Hypothesis (H1)–0.0215Triadic structures (H2)0.8219***Narrative Concordance Hypothesis (H3)0.794**Highest Contribution in a Narrative Hypothesis (H4)Challenges(H4)–0.02218*Climate(H4)0.03153***Economy(H4)–0.0883***Environmental(H4)0.01579**Forestry(H4)0.04253***Network(H4)–0.01655*Narrative Homophily Hypothesis (H5)Agriculture & livestock(H5)–0.0574**Challenges(H5)0.0376**Communities(H5)–0.03367***Economy(H5)0.0244+Energy(H5)0.0662*Environmental(H5)–0.01539+Forestry(H5)–0.0253+Policy(H5)–0.03357***Research & Education(H5)–0.02360***Technologies & Innovation(H5)–0.01670*Water(H5)–0.0576**Homophily by node type Hypothesis (H6)–0.250*Num.Obs.73153AIC7749.7BIC8126.9MC Std. Err. = 0.4039Null Deviance: 101412 on 73153 degrees of freedomResidual Deviance: 7668 on 73112 degrees of freedom+ p < 0.1, * p < 0.05, ** p < 0.01, *** p < 0.001AIC akaike information criterion, BIC Bayesian Information Criterion, Num.Obs. number of observations.For Hypothesis H3, the network with narrative distance and uncommon themes was not significant. For Hypothesis H4, the narratives Agriculture & Livestock, Animals, Communities, Donor, Energy, Food System, Policy, Region, Research & Education, Technologies & Innovation, and Water was not significant. For Hypothesis H5, the narratives Animals, Donor, Network, and Region were not reported as significant. Results summary of the ERGM model Null Deviance: 101412 on 73153 degrees of freedom Residual Deviance: 7668 on 73112 degrees of freedom AIC akaike information criterion, BIC Bayesian Information Criterion, Num.Obs. number of observations. For Hypothesis H3, the network with narrative distance and uncommon themes was not significant. For Hypothesis H4, the narratives Agriculture & Livestock, Animals, Communities, Donor, Energy, Food System, Policy, Region, Research & Education, Technologies & Innovation, and Water was not significant. For Hypothesis H5, the narratives Animals, Donor, Network, and Region were not reported as significant. Centrality (H1) and Triadic(H2) Centrality was not significant, so the hypothesis that degree centrality is relevant for forming links in the network cannot be accepted. Conversely, the hypothesis regarding triangular structures or Triadic (H2) is significant and positive. This indicates that the Colombian-Peruvian Amazon cooperation network exhibits closed triangular structures and subgroups for conservation and technology, revealing consolidation patterns and association preferences that favor the formation of cooperative connections within the network. Narrative Concordance (H3) The significance of the statistic and a positive result (Table 4) indicate that, when nodes share more than seven narratives over short distances, they are more likely to form cooperative connections or links. This means that closer contributions on topics that are more aligned with the partner’s narrative will increase the likelihood of cooperation. This concordance pattern highlights the importance of consolidating consistent and intense discourses in communications, which can enhance the probability of creating cooperative links in the Colombian-Peruvian Amazon. Conversely, networks with connections that have narrative contribution gaps or low levels of concordance do not show statistical significance for relationship formation. Intensity: Topic with the Highest Contribution in a Narrative (H4) Higher narrative contributions in themes such as Climate, Forestry, and Environmental are associated with greater probabilities of forming dyadic cooperative relationships with other nodes. Conversely, the significant but negative coefficients corresponding to high contributions in the narratives of Challenges, Economy, and Network indicate a lower likelihood of forming cooperative links with other partners. Narrative Homophily (H5) The model results reveal that Energy, Economy, Challenges, and Region are narratives where greater differences increase the propensity for forming a cooperative relationship or link with another partner, as indicated by their statistical significance and positive result. This means that when these themes dominate, it is less likely that a relationship will form. Conversely, lower differences in these narratives indicate a higher probability or propensity to establish cooperative connections in the Colombian-Peruvian Amazon. Narratives such as Agriculture & Livestock; Communities; Environment; Forestry; Policy; Research & Education; Technologies & Innovation, and Water tend to foster links when narrative differences between two nodes are minimal. In other words, these themes exhibit patterns or tendencies to generate connections when there are narrative homophily nodes that are more likely to form links when they share close thematic alignment. Homophily by Node Type (H6) These statistics are significant and negative, indicating that there is a lower probability of forming links when the partner is of the same type. Discussion General Composition and Dominance of Narratives in the Network The social system’s response to the degradation of the Colombian−Peruvian Amazon is configured as a cooperation network composed of 382 nodes that interact and form a complex system of narratives. However, this network operates in a context where the biome has received less scientific attention than other areas in the Amazon region, limiting the availability of key information to strengthen its environmental governance. In this regard, this article presents an integration of methods that allows exploring the composition of narratives and trends in discourses resulting from the presence of an Amazonian cooperation network. The results reveal the dominance of narratives related to Research & Education (averaging 13.73%), within the context of greater participation by international actors (42% universities, international organizations, NGOs, etc.), suggesting that the biome is perceived as a type of laboratory, classroom, and library for generating international knowledge. Other dominant narratives are linked to the Communities category, emphasizing concepts such as women, family, Indigenous peoples, customs, ancestors, and traditions (13%). This is followed by governance and policy narratives, focusing on legislative, regulatory, and administrative aspects (12%). There is, however, a lower narrative contribution in key areas such as food systems, climate, agriculture and livestock, water, and fauna. Both Peruvian and Colombian nodes exhibit similar patterns in narratives related to communities, policy, and research, but with lower contributions compared to their counterparts in the U.S., Brazil, and Europe. This pattern contradicts the debates surrounding the potential imposition or hegemony of environmental narratives from the Global North and their misalignment with local and regional realities, as well as the potential marginalization of local knowledge and narratives in favor of external actors (Fairhead & Leach 2003; Forsyth 2003) thereby oversimplifying the social and cultural complexities of grassroots communities (Brosius 1999). Data and Trends for the Network The narrative distances between actors can either support or hinder incremental changes and radical transformations, as these narratives hold the power to legitimize or impede behavioral and policy changes in environmental governance (Koch et al. 2021a). By integrating the information on discourse contributions at the node level and when mediated by relationships with peers in the network, a narrative concordance (proximity between narratives) of 61% was observed, while 18% corresponded to greater discursive distance, and 16% of the edges were associated with uncommon themes. This indicates that the cooperation network for these two countries is characterized by a predominant narrative concordance, mainly concentrated in the categories Communities, Research & Education, Policy, Environmental, and Technologies & Innovation among the 17 mapped narratives. Similarly, when incorporating subnetworks as covariates for concordance, distance, and uncommon themes into the model, the results revealed greater significance for the statistics corresponding to the hypothesis that higher narrative concordance implies a greater propensity to create cooperative relationships in the Colombian-Peruvian Amazon. According to R. Lejano et al. (2013), this occurs when there is narrative concordance in environmental governance networks, facilitating cooperation, as these actors share a common vision and can act in a coordinated manner. However, there is a risk that complementary or divergent perspectives may be marginalized, limiting the network’s ability to adapt to new challenges or address complex problems comprehensively. Thematic Leadership Centrality metrics in the relationship between narratives reveal that the categories Technologies & Innovation, Research & Education, and Region hold influential positions, given their high closeness-centrality levels, with coefficients of 0.5, 0.3, and 0.25, respectively. Conversely, network modeling indicates that when a node is characterized by narrative biases (high proportion) in themes such as (1) Economy, (2) Challenges, or (3) Network, it has a lower probability of forming relationships with peers in interventions. Regarding the Economy category, this may reflect a disconnect between theories of growth, development, industry, and production and conservation-focused approaches. These narratives have limited resonance with actors prioritizing the protection of the Amazon biome. Aligning economic narratives with conservation objectives proves challenging, as traditional economic approaches often do not fully align with long-term sustainability and conservation priorities (Horan 2022). This misalignment generates confusion and inconsistency in the implementation of environmental policies related to climate change and sustainability (Liverman 2009a). The Challenges, and Network categories can be interpreted as declarative or symbolic discourses that require integration with more operational or impactful narratives (Lezama 2018). A node with a high narrative contribution on issues related to biome risks and degradation, or with a discourse centered on cooperation, collaboration, and centrality, may have a significant probability of not forming links with other partners in the Colombian-Peruvian Amazon network. For instance, organizations such as the Amazon Cooperation Treaty Organization face difficulties in coordinating and scaling down their cross-border operations, which limits their ability to establish effective collaborations with actors in Colombia and Peru (Mostajo 2024; Vania & Castilla 2024). The Most Attractive Narratives for Creating Connections The greater the difference in themes, such as Energy, Challenges, and Economy, between two nodes, the more likely they are to form a link in the network (heterophily). Conversely, narratives such as Agriculture & Livestock, Communities, Environmental, Forestry, Policy, Research & Education, Technologies & Innovation, and Water exhibit homophilic or similarity patterns, making them more prone to creating relationships in the Amazon. In this context, high narrative contributions at the node level influence relationship formation, specifically for Climate, Forestry, and Environmental. These themes may reflect consistent positions in discourses and levels of specialization, yet they do not emerge as dominant narratives in the overall discourse analysis of the network. Role of Subgroups The network’s structural form is supported by multiple relationships between nodes that configure closed triangles, identifying many subgroups interacting and cooperating in the conservation of the Amazon. This endogenous covariance in network modeling demonstrates preferences in associations, which significantly influence the formation of links between nodes. The result is relational clusters that reinforce the exchange of information, discourses, and narratives, thus addressing thematic concordance trends and thereby influencing interventions, research, and policy actions. The narratives within these clusters or niches allow actors to legitimize or challenge certain environmental policies, leading to an alignment of shared discourses and actions within the network (Van Der Leeuw 2020). Opportunities for Strengthening Environmental Governance in the Amazon using Network Results and Narrative Composition for Colombia and Peru The Food System narrative has one of the lowest levels of participation in the network’s discourses, despite its importance in the literature for food security and the health benefits of biodiversity-rich traditional food practices (Johns & Eyzaguirre 2006). The Amazon region has faced food security challenges due to biodiversity loss, agricultural landscape transformation, deforestation, and unsustainable farming practices, which have impacted traditional food practices and communities’ diets (Padoch & Sunderland, 2013). This disconnect in exchanges on this theme among the mapped network nodes reveals an opportunity for environmental governance and policies to stimulate interventions and address gaps. This includes enhancing interactions with dominant themes and emphasizing key topics such as gender, food systems, and economy.While scientific leaders consider water, territorial analysis, cultural inclusion, and environmental governance to be relevant, conversely, current narrative trends suggest they are not relevant. Therefore, this widens collaboration gaps across the network between future needs/topics of interest and the current thematic corpus. There is an opportunity to translate scientific evidence into grounded policy guidelines and instruments, connecting research with decision-making and realigning prevailing narratives with long-term priorities. This is explained in Fig. S1 in Appendix 1, Supplementary Material.Negotiation and credibility: the framing or strategic “telling” of narratives is crucial for policy information to achieve the highest credibility among the different stakeholders (Paschen & Ison 2014). Proper use of narrative becomes a key tool for consolidating and strengthening collaborative relationships, not just as a communication mechanism, but also as a means of establishing the authority of actor networks and influencing policy decisions. Narratives can promote or legitimize certain practices over others and are also seen as a mechanism for social cohesion within environmental governance. According to Denton (2017b), these narratives function as multiscale negotiation strategies between local and global actors, and are institutionalized in dominant discourses that shape environmental policies and practices. Organizational networks use narratives to strengthen their legitimacy and achieve their goals while adapting and transforming based on the interests of the actors involved (Warner 2019). This narrative negotiation process enhances the networks’ power and their capacity to influence environmental governance, opening new research lines to understand the underlying dynamics of cooperation networks and their impacts.Misalignment between layers: local narratives and the lack of mutual understanding among different actors (such as local communities, governments, and international stakeholders) often lead to stagnant debates, affecting the effectiveness of environmental policies (Ostrom, 1990). Additionally, a lack of understanding or recognition of other actors’ narratives can result in uncoordinated efforts and weak policy implementation (Fløttum & Gjerstad 2017). In this sense, narrative concordance in this complex (socio-environmental) context has been a criterion not covered by science and Amazonian governance itself, thus amplifying collaboration gaps (Bergsten et al. 2019) and favoring the forces that degrade and subjugate the biome and its communities. Therefore, interventions between nodes mediated by concordant or complementary narratives can improve their effectiveness and impact, which presents an opportunity to address them in technical roundtables, transnational agreements, or other spaces for political construction.Perspectives and studies on environmental narratives: Over recent decades, various studies have delved into the analysis of narratives related to climate change and environmental governance. Some approaches focus on predominant historical narratives, exploring how they influence the formulation of local climate policies (Ingram et al. 2019). Others examine the immediate effects of language in political debates through narrative analysis (Donald et al. 2022), while discourse analysis seeks to uncover the deeper meanings behind these narratives. Power narratives, on the other hand, explore differences in how intergovernmental organizations and local NGOs construct their discourses on climate change, often with divergent objectives (Denton 2017c). An emerging approach in this field involves narrative mechanisms in environmental governance. This framework analyzes the interaction between collaborative relationships, narrative congruence among actors, and the influence of social structures in constructing effective discourses (Koch et al. 2021b). Collectively, these perspectives provide valuable tools for understanding how narratives structure political and social debates, and shape actions and policies in environmental contexts such as the Amazon. This underscores the need to continue advancing the understanding of environmental narratives and their impact on future policies and practices.Monitoring, evaluation, and impact of narratives in the Amazon: it is crucial to promote exercises for monitoring and evaluating narrative trends in the Amazon biome, especially where trends are conflicting or fail to achieve consensus (Lawton & Rudd 2014). These efforts should be integrated into decision-making processes, mobilizing scientific evidence into political agendas. Mechanisms and projects should also be established to explore the impact of international narratives on local dynamics, aiming to integrate their perspectives into environmental governance.Enhancing cross-border cooperation in Amazon conservation: The high participation of international and external nodes (Fairhead & Leach, 2003) presents an opportunity for the national and local governments of Peru and Colombia to promote cross-border cooperation agreements (Transnational, regional, and local). By developing joint projects and creating strategic incentives, both countries can strengthen the involvement of the public sector and improve access to key resources, technologies, and political spaces for effective management of the Amazon biome. This approach would better integrate local and regional knowledge into environmental agendas, counteracting the potential hegemony of external narratives and ensuring a more inclusive management that is adapted to local realities.Environmental governance in the Amazon can shape and connect narratives: The difficulty of harmonizing economic narratives with conservation goals in the Amazon represents a strategic opportunity to strengthen environmental governance (Horan, 2022). Efficient governance is not possible without understanding the economic background of the region, making it essential to engage organizations with experience and knowledge to integrate and connect these narratives in specific projects, thereby reducing current collaboration gaps. In these areas, Agroclimatic Technical Committees (MTAs) are key spaces for collaboration to align economic narratives with environmental and sustainability issues, articulating them with national and Amazon region development plans, as they promote work packages that reconcile common values and objectives. The participation of international donors in activities where these narratives intersect promotes more coherent, comprehensive, and effective governance, capable of addressing the complexity of the territory and its actors. The Food System narrative has one of the lowest levels of participation in the network’s discourses, despite its importance in the literature for food security and the health benefits of biodiversity-rich traditional food practices (Johns & Eyzaguirre 2006). The Amazon region has faced food security challenges due to biodiversity loss, agricultural landscape transformation, deforestation, and unsustainable farming practices, which have impacted traditional food practices and communities’ diets (Padoch & Sunderland, 2013). This disconnect in exchanges on this theme among the mapped network nodes reveals an opportunity for environmental governance and policies to stimulate interventions and address gaps. This includes enhancing interactions with dominant themes and emphasizing key topics such as gender, food systems, and economy. The Food System narrative has one of the lowest levels of participation in the network’s discourses, despite its importance in the literature for food security and the health benefits of biodiversity-rich traditional food practices (Johns & Eyzaguirre 2006). The Amazon region has faced food security challenges due to biodiversity loss, agricultural landscape transformation, deforestation, and unsustainable farming practices, which have impacted traditional food practices and communities’ diets (Padoch & Sunderland, 2013). This disconnect in exchanges on this theme among the mapped network nodes reveals an opportunity for environmental governance and policies to stimulate interventions and address gaps. This includes enhancing interactions with dominant themes and emphasizing key topics such as gender, food systems, and economy. While scientific leaders consider water, territorial analysis, cultural inclusion, and environmental governance to be relevant, conversely, current narrative trends suggest they are not relevant. Therefore, this widens collaboration gaps across the network between future needs/topics of interest and the current thematic corpus. There is an opportunity to translate scientific evidence into grounded policy guidelines and instruments, connecting research with decision-making and realigning prevailing narratives with long-term priorities. This is explained in Fig. S1 in Appendix 1, Supplementary Material. While scientific leaders consider water, territorial analysis, cultural inclusion, and environmental governance to be relevant, conversely, current narrative trends suggest they are not relevant. Therefore, this widens collaboration gaps across the network between future needs/topics of interest and the current thematic corpus. There is an opportunity to translate scientific evidence into grounded policy guidelines and instruments, connecting research with decision-making and realigning prevailing narratives with long-term priorities. This is explained in Fig. S1 in Appendix 1, Supplementary Material. Negotiation and credibility: the framing or strategic “telling” of narratives is crucial for policy information to achieve the highest credibility among the different stakeholders (Paschen & Ison 2014). Proper use of narrative becomes a key tool for consolidating and strengthening collaborative relationships, not just as a communication mechanism, but also as a means of establishing the authority of actor networks and influencing policy decisions. Narratives can promote or legitimize certain practices over others and are also seen as a mechanism for social cohesion within environmental governance. According to Denton (2017b), these narratives function as multiscale negotiation strategies between local and global actors, and are institutionalized in dominant discourses that shape environmental policies and practices. Organizational networks use narratives to strengthen their legitimacy and achieve their goals while adapting and transforming based on the interests of the actors involved (Warner 2019). This narrative negotiation process enhances the networks’ power and their capacity to influence environmental governance, opening new research lines to understand the underlying dynamics of cooperation networks and their impacts. Negotiation and credibility: the framing or strategic “telling” of narratives is crucial for policy information to achieve the highest credibility among the different stakeholders (Paschen & Ison 2014). Proper use of narrative becomes a key tool for consolidating and strengthening collaborative relationships, not just as a communication mechanism, but also as a means of establishing the authority of actor networks and influencing policy decisions. Narratives can promote or legitimize certain practices over others and are also seen as a mechanism for social cohesion within environmental governance. According to Denton (2017b), these narratives function as multiscale negotiation strategies between local and global actors, and are institutionalized in dominant discourses that shape environmental policies and practices. Organizational networks use narratives to strengthen their legitimacy and achieve their goals while adapting and transforming based on the interests of the actors involved (Warner 2019). This narrative negotiation process enhances the networks’ power and their capacity to influence environmental governance, opening new research lines to understand the underlying dynamics of cooperation networks and their impacts. Misalignment between layers: local narratives and the lack of mutual understanding among different actors (such as local communities, governments, and international stakeholders) often lead to stagnant debates, affecting the effectiveness of environmental policies (Ostrom, 1990). Additionally, a lack of understanding or recognition of other actors’ narratives can result in uncoordinated efforts and weak policy implementation (Fløttum & Gjerstad 2017). In this sense, narrative concordance in this complex (socio-environmental) context has been a criterion not covered by science and Amazonian governance itself, thus amplifying collaboration gaps (Bergsten et al. 2019) and favoring the forces that degrade and subjugate the biome and its communities. Therefore, interventions between nodes mediated by concordant or complementary narratives can improve their effectiveness and impact, which presents an opportunity to address them in technical roundtables, transnational agreements, or other spaces for political construction. Misalignment between layers: local narratives and the lack of mutual understanding among different actors (such as local communities, governments, and international stakeholders) often lead to stagnant debates, affecting the effectiveness of environmental policies (Ostrom, 1990). Additionally, a lack of understanding or recognition of other actors’ narratives can result in uncoordinated efforts and weak policy implementation (Fløttum & Gjerstad 2017). In this sense, narrative concordance in this complex (socio-environmental) context has been a criterion not covered by science and Amazonian governance itself, thus amplifying collaboration gaps (Bergsten et al. 2019) and favoring the forces that degrade and subjugate the biome and its communities. Therefore, interventions between nodes mediated by concordant or complementary narratives can improve their effectiveness and impact, which presents an opportunity to address them in technical roundtables, transnational agreements, or other spaces for political construction. Perspectives and studies on environmental narratives: Over recent decades, various studies have delved into the analysis of narratives related to climate change and environmental governance. Some approaches focus on predominant historical narratives, exploring how they influence the formulation of local climate policies (Ingram et al. 2019). Others examine the immediate effects of language in political debates through narrative analysis (Donald et al. 2022), while discourse analysis seeks to uncover the deeper meanings behind these narratives. Power narratives, on the other hand, explore differences in how intergovernmental organizations and local NGOs construct their discourses on climate change, often with divergent objectives (Denton 2017c). An emerging approach in this field involves narrative mechanisms in environmental governance. This framework analyzes the interaction between collaborative relationships, narrative congruence among actors, and the influence of social structures in constructing effective discourses (Koch et al. 2021b). Collectively, these perspectives provide valuable tools for understanding how narratives structure political and social debates, and shape actions and policies in environmental contexts such as the Amazon. This underscores the need to continue advancing the understanding of environmental narratives and their impact on future policies and practices. Perspectives and studies on environmental narratives: Over recent decades, various studies have delved into the analysis of narratives related to climate change and environmental governance. Some approaches focus on predominant historical narratives, exploring how they influence the formulation of local climate policies (Ingram et al. 2019). Others examine the immediate effects of language in political debates through narrative analysis (Donald et al. 2022), while discourse analysis seeks to uncover the deeper meanings behind these narratives. Power narratives, on the other hand, explore differences in how intergovernmental organizations and local NGOs construct their discourses on climate change, often with divergent objectives (Denton 2017c). An emerging approach in this field involves narrative mechanisms in environmental governance. This framework analyzes the interaction between collaborative relationships, narrative congruence among actors, and the influence of social structures in constructing effective discourses (Koch et al. 2021b). Collectively, these perspectives provide valuable tools for understanding how narratives structure political and social debates, and shape actions and policies in environmental contexts such as the Amazon. This underscores the need to continue advancing the understanding of environmental narratives and their impact on future policies and practices. Monitoring, evaluation, and impact of narratives in the Amazon: it is crucial to promote exercises for monitoring and evaluating narrative trends in the Amazon biome, especially where trends are conflicting or fail to achieve consensus (Lawton & Rudd 2014). These efforts should be integrated into decision-making processes, mobilizing scientific evidence into political agendas. Mechanisms and projects should also be established to explore the impact of international narratives on local dynamics, aiming to integrate their perspectives into environmental governance. Monitoring, evaluation, and impact of narratives in the Amazon: it is crucial to promote exercises for monitoring and evaluating narrative trends in the Amazon biome, especially where trends are conflicting or fail to achieve consensus (Lawton & Rudd 2014). These efforts should be integrated into decision-making processes, mobilizing scientific evidence into political agendas. Mechanisms and projects should also be established to explore the impact of international narratives on local dynamics, aiming to integrate their perspectives into environmental governance. Enhancing cross-border cooperation in Amazon conservation: The high participation of international and external nodes (Fairhead & Leach, 2003) presents an opportunity for the national and local governments of Peru and Colombia to promote cross-border cooperation agreements (Transnational, regional, and local). By developing joint projects and creating strategic incentives, both countries can strengthen the involvement of the public sector and improve access to key resources, technologies, and political spaces for effective management of the Amazon biome. This approach would better integrate local and regional knowledge into environmental agendas, counteracting the potential hegemony of external narratives and ensuring a more inclusive management that is adapted to local realities. Enhancing cross-border cooperation in Amazon conservation: The high participation of international and external nodes (Fairhead & Leach, 2003) presents an opportunity for the national and local governments of Peru and Colombia to promote cross-border cooperation agreements (Transnational, regional, and local). By developing joint projects and creating strategic incentives, both countries can strengthen the involvement of the public sector and improve access to key resources, technologies, and political spaces for effective management of the Amazon biome. This approach would better integrate local and regional knowledge into environmental agendas, counteracting the potential hegemony of external narratives and ensuring a more inclusive management that is adapted to local realities. Environmental governance in the Amazon can shape and connect narratives: The difficulty of harmonizing economic narratives with conservation goals in the Amazon represents a strategic opportunity to strengthen environmental governance (Horan, 2022). Efficient governance is not possible without understanding the economic background of the region, making it essential to engage organizations with experience and knowledge to integrate and connect these narratives in specific projects, thereby reducing current collaboration gaps. In these areas, Agroclimatic Technical Committees (MTAs) are key spaces for collaboration to align economic narratives with environmental and sustainability issues, articulating them with national and Amazon region development plans, as they promote work packages that reconcile common values and objectives. The participation of international donors in activities where these narratives intersect promotes more coherent, comprehensive, and effective governance, capable of addressing the complexity of the territory and its actors. Environmental governance in the Amazon can shape and connect narratives: The difficulty of harmonizing economic narratives with conservation goals in the Amazon represents a strategic opportunity to strengthen environmental governance (Horan, 2022). Efficient governance is not possible without understanding the economic background of the region, making it essential to engage organizations with experience and knowledge to integrate and connect these narratives in specific projects, thereby reducing current collaboration gaps. In these areas, Agroclimatic Technical Committees (MTAs) are key spaces for collaboration to align economic narratives with environmental and sustainability issues, articulating them with national and Amazon region development plans, as they promote work packages that reconcile common values and objectives. The participation of international donors in activities where these narratives intersect promotes more coherent, comprehensive, and effective governance, capable of addressing the complexity of the territory and its actors. This study highlights the need for a diversified and well-integrated cooperation network for the conservation of the Colombian−Peruvian Amazon. It reveals that organizations intervening in environmental matters bring interests, experiences, and priorities that complicate operations at multiple levels: international, regional, national, and subregional. Participants face the challenge of navigating dominant narratives, social trends, and political changes, while adapting with their own resources and perspectives (Jones & McBeth 2010b). The inclusion of diverse perspectives and the alignment of policies with local, often transnational, realities are crucial for strengthening the environmental governance of the biome. By understanding and leveraging the dynamics of these networks, more effective interventions can be developed that highlight the unique characteristics of the Amazon, promoting a conservation approach that is both inclusive and adaptable (Guio & Rojas, 2019). This approach will foster collaboration among actors who, while not always fully aligned in their conservation strategies, priorities, intervention methods, or land management models, can find common ground to work together for the benefit of the Amazon biome (Ostrom 2009; Shanahan et al. 2011). Conclusions The Colombian−Peruvian Amazon presents a complex configuration of institutional and technological relationships and narratives layered within a cooperation framework. This intricate network has not been thoroughly studied, leaving aspects that could be instrumental for environmental governance and the protection of the biome unexamined. This article contributes to the analysis of narratives by employing a methodological integration of Social Network Analysis, web scraping, text mining, and Exponential Random Graph Models. By mapping narratives and analyzing their concordance patterns, key discursive dynamics that either facilitate or hinder cooperation among actors are identified, providing a more detailed understanding of the structure of interaction in the region. This analysis highlights how dominant narratives shape the network while simultaneously revealing that centralized environmental policies hinder the integration of local complexities and the decentralized interactions necessary for effective governance. It underscores the need to construct environmental governance frameworks that actively integrate local perspectives and capacities while leveraging international networks as a bridge to overcome limitations. In this context, the study evaluated several hypotheses regarding the importance of centrality, the role of subnetworks, contributions, and homophily across 17 thematic narratives (e.g., Communities, Research & Education, Environmental, Technologies & Innovation, Networks and Cooperation, Donors and Funding, among others), derived from the interaction of 382 nodes identified in the Amazon. The results indicate the dominance of three narratives: Research & Education, Communities, and Policy, revealing a higher distribution, average contribution, and significance for these thematic narratives. Additionally, international nodes contribute predominantly to research and technology topics, surpassing Colombian and Peruvian nodes and suggesting a pattern of influence over these themes within the network. This aligns with other analyzes of narrative closeness centrality, such as those by Bodin & Crona (2009b) and Koch et al. (2021a), which emphasize that highly central narratives dominate discursively and connect strategically with other topics in the network. Moreover, studies such as those by R. P. Lejano et al. (2013) and Prell et al. (2009) support this perspective, highlighting the role of narratives as bridges between different themes within social and environmental networks. The Amazon actor-network is significantly clustered, with strong concordance across multiple narratives, and exhibits homophily in key themes such as Forestry, Agriculture, Climate, Water, Policy, and Research & Education. This significantly explains the formation of relationships among nodes with a strong influence of technical, educational, and research narratives and substantial contributions from international nodes. Supplementary information Supplementary information Supplementary information