Constructing the Truth: Text Mining and Linguistic Networks in Public Hearings of Case 03 of the Special Jurisdiction for Peace (JEP)

TL;DR

The study addresses how to reconstruct contested truths in Case 03 of Colombia's Jurisdicción Especial para la Paz (JEP) using natural language processing and semantic networks applied to public hearings. It introduces a pipeline that combines transcription-based text processing, sentiment analysis, and network-based topic modeling with bigrams and skipgrams to detect narrative patterns across victims and comparecientes and to compare regional subcases. Key findings show a pronounced negative sentiment intensity alongside regionally differentiated thematic clusters, with higher modularity when analyzing subcasos, particularly in skipgram and bigram networks, supporting a nuanced understanding of truth-telling in transitional justice. The work provides a replicable methodological framework and open data/tools that can be extended to other transitional justice contexts to support truth, justice, repair, and non-repetition efforts.

Abstract

Case 03 of the Special Jurisdiction for Peace (JEP), focused on the so-called false positives in Colombia, represents one of the most harrowing episodes of the Colombian armed conflict. This article proposes an innovative methodology based on natural language analysis and semantic co-occurrence models to explore, systematize, and visualize narrative patterns present in the public hearings of victims and appearing parties. By constructing skipgram networks and analyzing their modularity, the study identifies thematic clusters that reveal regional and procedural status differences, providing empirical evidence on dynamics of victimization, responsibility, and acknowledgment in this case. This computational approach contributes to the collective construction of both judicial and extrajudicial truth, offering replicable tools for other transitional justice cases. The work is grounded in the pillars of truth, justice, reparation, and non-repetition, proposing a critical and in-depth reading of contested memories.

Figures (24)

• Figure 1: Flujograma del preprocesamiento.
• Figure 2: Análisis general: palabras más frecuentes.
• Figure 3: Análisis general: nube de Palabras.
• Figure 4: Análisis general: sustantivos más frecuentes.
• Figure 5: Análisis general: verbos más frecuentes.