DOI: 10.18485/judig.2026.2.ch8 ISSN:

Lexicon-Based Exploration of Emotion Dynamics in TEI Novels: Case Study on SrpELTeC Corpus

Ranka Stanković, Milena Šošić
Show PDF Cite
This paper presents a hierarchical analysis of emotional dynamics in the SrpELTeC corpus of Serbian novels from the period 1850–1920s. The study combines TEI-encoded texts with a Serbian-tailored emotion lexicon EmoLex.SR , based on Plutchik’s model of basic emotions. Emotional signals from the corpus were extracted through lexical matching and analyzed across multiple textual hierarchically organized segments — words, sentences, paragraphs, chapters, literary works and the full corpus, which enabled the exploration of both local variation and global patterns of emotional signals. The approach combines emotional words density and intensity measures with hierarchical and chronological grouping, to examine short- and long-term changes in emotional structure across the corpus. First empirical results showed a stable baseline dominated by
trust
and
anticipation
emotions, alongside with the significantly increased emotional variation in the early twentieth century. These findings indicate that emotional organization in Serbian prose evolves in both magnitude and structural complexity over time and that proposed methodology enables consistent comparison and clear computational visibility across literary work segments and time periods.

More from our Archive