Multitextuality in RPGs: A Ludonarrative Synergy Model for Video Game Text Analysis

Solska's paper proposes a ludonarrative synergy model for analyzing the textual material of role-playing video games, extending Galloway's distinction between diegetic and non-diegetic elements into a categorization of game text types. The central thesis holds that RPGs such as Divinity: Original Sin 2 contain a greater volume and variety of textual material than other genres, and that this multitextuality can be systematically classified by whether a text belongs to the fictional world (narrative and world-building) or to the mechanical framework of play (ludic and interface content). The model draws on text and discourse linguistics, natural language processing, and computational linguistics, and is implemented through a two-level Python-based method combining micro-level lexical and syntactic analysis with macro-level clustering.

The empirical test applies TF-IDF lexical analysis, syntactic n-gram extraction, and K-means clustering with k=2 to a corpus of Divinity: Original Sin 2 texts. The results show that the two categories exhibit distinct patterns. Diegetic texts are characterized by lore vocabulary such as "source," "divine," and "godwoken," along with narrative n-grams such as the banishment to the Void. Non-diegetic texts are marked by mechanics terminology including "damage," "level," "potion," and "turns," alongside ludic n-grams relating to statistics and abilities. The author concludes that these computational findings support the validity of the diegetic/non-diegetic categorization within the proposed model.

The scope of the study is focused on a single game, Divinity: Original Sin 2, and on the RPG genre as a representative case, with the analysis confined to its English-language in-game text. The work is situated within game studies and localization research, and its contribution lies in demonstrating how computational methods can operationalize theoretical categories for video game text analysis.