Discursive construction of environmental perceptions in amazonian digital Media: a comparative textometric analysis

Abstract

The socio-environmental crisis in the Ecuadorian Amazon—marked by extractive expansion, deforestation, and human rights violations—requires an understanding of how regional public discourses on conservation build representations of environmental perception. This study draws on textometric discourse analysis to examine environmental representations as they are constructed through public discourse in digital spaces. A corpus of 281 texts was compiled using BootCat and divided into two subcorpora: academic (149 texts) and communicational (132 texts). Sentiment analysis using Orange Data Mining revealed that the academic subcorpus exhibits affective neutrality, stemming from scientific legitimacy based on objectivity, while the communicational subcorpus displays a bimodal structure that oscillates between denunciation and the assertion of rights, characteristic of environmental activism. Topic modeling using Latent Dirichlet Allocation identified territorial planning, indigenous rights, and environmental education as the core themes of the academic discourse; the communicational discourse addressed the issues of illegal mining, prior consultation, and community communication. When these results are combined with the area of discursive overlap identified in the corpus, hybrid patterns of environmental communication emerge that provide insights for designing strategies based on ecological justice and recognition of the more-than-human world.