Revealing gendered patterns in Italy’s street names with large language models

Abstract This study investigates the Italian odonymic landscape through the lens of critical toponymy, employing computational methods for semantic classification to examine gendered patterns of commemoration in street naming. A national-scale database is constructed by integrating official toponymic registers with rule-based heuristic procedures and large language model-based classification. This hybrid infrastructure enables the systematic identification and categorization of odonyms dedicated to individual and collective persons, with particular attention to gender representation. The combined classification approach reliably detects personal commemorations and infers gender attributes, yielding results that closely align with manually curated datasets. The study also advances critical toponymy by showing how computational methods can operationalize analyses of collective memory and symbolic power in urban space. The results reveal the persistent gender asymmetries in Italian commemorative odonymy.