DOI: 10.65266/WFUT5670
Full-text access available through the Modern Language Association (MLA) International Bibliography: https://about.ebsco.com/products/research-databases/mla-international-bibliography-full-text
Bibliographical Information: Marca, Pablo a. “A New Classification of Fairy Tales Using Topic Modeling.” Annali d’italianistica 43 (2025): 37-64.
Abstract: Fairy tales have traditionally been classified based on diverse criteria such as plot, the Aarne-Thompson-Uther (ATU) index, functions (Propp), authorship, time period, or thematic interpretations devised by scholars. These manual categorizations, while useful, can be subject to criticism. However, classifying fairy tales can be a way of simplifying the study of this vast corpus. The article takes advantage of the emergence of AI and machine learning to present a new categorization of fairy tales by topic modeling a corpus of European classic fairy tales. Topic modeling, a statistical modeling technique that uses machine learning, identifies word co-occurrences within a corpus, organizing them into clusters or topics. The results confirm the validity of plot-based categories (the ATU index), but also the possibility to find new connections between the tales. Thus, this method offers a new classification approach, opening new research avenues.
Key Words: fairy tales, Aarne-Thompson-Uther (ATU), tale-type index, topic modeling, machine learning.
