Formulas and Semantics in Digital Diplomatics: Reflections from Évora
The IV International Conference on Palaeography and Diplomatics, held on 7–8 October 2025, provided a vibrant environment for discussing the intersections between traditional diplomatics and digital humanities. Set in the historic university city of Évora and at the public library, the event brought together scholars from diverse backgrounds, including paleography and philology, as well as computer science, to reflect on how digital tools are reshaping our understanding of written culture.


Our contribution, titled ‘Formulas and Semantics in Digital Diplomatics’, explored how the integration of semantic and structural approaches can open new perspectives in the study of notarial documents. We focused on a small corpus of fifteenth-century sales charters from the ‘Sabino Loffredo’ Municipal Library of Barletta, produced by notaries operating in the Kingdom of Naples. The goal was to examine whether certain formulaic expressions, such as tenere et possidere precario nomine, could serve as sociological markers, reflecting the nature of the actors involved rather than being mere neutral legal conventions. The analysis revealed that, while clauses like the defensio occur in every transaction regardless of the social status of the parties, the precarium clause appears exclusively when at least one of the actors is a religious institution or a part of such. According to the Vocabulaire international de la diplomatique, the precarium expresses a voluntary but revocable concession, a formula often used by ecclesiastical institutions to secure the temporary possession of a good or property. Its selective occurrence, therefore, suggests that some formulaic expressions may encode not only legal meaning but also the social identity of the participants, functioning as subtle markers of institutional belonging.
To explore these patterns, we adopted a combined digital methodology developed within the project DiDip – From Digital to Distant Diplomatics. Using two complementary tools, ANNIE (Annotation Interface for Named-Entity & Information Extraction) and FLAME (Formulaic Language Analysis in Medieval Expressions), both developed by Tamas Kovacs, we were able to move from a close reading of single documents to a scalable, data-driven interpretation. ANNIE enables the manual semantic annotation of entities, such as persons, roles, locations, and formulaic clauses, whereas FLAME clusters texts based on their text similarity. By merging the outputs of both tools in a Python notebook, we generated a heatmap that visualized the relationship between social actors and the recurrence of specific clauses. The precarium appeared as a cluster confined to transactions involving ecclesiastical institutions, whereas neutral formulas extended across all quadrants.

This visual correlation confirmed that the formulaic language of medieval notaries was not a rigid framework but a vivid expression of context, shaped by the legal and social environment in which it was produced.
The presentation was very well received. Several participants were curious about how this integrated workflow could be scaled up, which will be the next step in the work’s progress. The discussion highlighted how this kind of research can bridge the gap between computational analysis and historical interpretation, without losing the interpretive depth that characterizes diplomatics as a discipline.



Beyond our presentation, the conference itself offered an inspiring overview of current digital approaches to manuscript and documentary studies. Dominique Stutzmann (CNRS–IRHT) opened the event with a talk on artificial intelligence and medieval manuscripts, discussing projects such as HIMANIS and Horae_Minit. His reflections on bias, opacity, and the epistemological implications of machine learning in palaeography resonated deeply with our own methodological challenges. Later, Antonio J. Díaz-Rodríguez (University of Córdoba) presented the ambitious Proyecto DARCO – Digitalización del Reino de Córdoba, which aims to digitize millions of pages of early modern notarial protocols and to develop EscribanIA, an AI-based model for handwritten text recognition. Even though our discussion time was short, the project’s scale and ambition illustrated how digital paleography and diplomatics are rapidly converging towards a new phase of data-rich historical research.
The two days in Évora were a reminder that digital diplomatics have become a meeting point for disciplines that once seemed far apart: computer science, linguistics, archival studies, and social history. The atmosphere was collaborative and intellectually stimulating, with discussions continuing well beyond the formal sessions. For our part, presenting the first outcomes of the ANNIE–FLAME integration felt like a first, concrete step in enabling a more nuanced and interconnected understanding of how medieval documentary language functioned as both a legal instrument and a reflection of social identity. Looking back, what struck most was the shared conviction among participants that digital diplomatics is not a replacement for traditional methods but an evolution of them. The diplomatic analysis of documents, including their language, structure, and material form, remains a core task. What changes is the scale: the ability to visualize and compare hundreds or even thousands of texts at once, while still remaining faithful to the interpretive rigor of the field. In this sense, the Évora conference was not only an academic gathering but also a collective reflection on how to integrate technological innovation into the critical humanistic tradition. Leaving Évora, with its quiet cloisters and sunlit arcades, I felt that our work within DiDip stands at the intersection of two traditions: the analytical precision of diplomatics and the exploratory power of digital methods.
OpenEdition suggests that you cite this post as follows:
Giuseppe Consolo (October 30, 2025). Formulas and Semantics in Digital Diplomatics: Reflections from Évora. DiDip. Retrieved November 6, 2025 from https://didip.hypotheses.org/3891

