Visualizzare l'Invisibile: Dal Knowledge Graph alla Piattaforma Digitale

Every textual work may be seen as the product of regulated combinatorial procedures operating on discrete formal and semantic units. Whether these rules are adopted intentionally, as in the case of constrained writing, or act as implicit conventions, they shape the formal and semantic architecture of the text. Yet current bibliographic cataloguing systems focus almost entirely on the bibliographic product, leaving the generative mechanisms of texts implicit and semantically underspecified. This research addresses this descriptive gap by introducing DeSMòS (Descriptive Semantic Model for Structured Texts), an OWL ontology developed to formally represent literary constraints as generative devices and make them computationally accessible.

The research is driven by two Research Questions (RQs): 

RQ1: What is a literary constraint and which entities are necessary to describe it formally?

RQ2: Which strategies for querying and visualising a Knowledge Graph (KG) can make explicit the latent compositional mechanisms structuring literary works, ensuring non-expert accessibility in compliance with copyright restrictions?

To address RQ1, the methodology integrates a domain analysis of potential literature, specifically the production of the Oplepo (Opificio di Letteratura Potenziale), treated as a paradigmatic case study of regulated textual forms, with Semantic Web technologies. DeSMòS is aligned with international standards, including CIDOC CRM, LRMoo, PROV-O, and SKOS, ensuring full interoperability and adherence to FAIR principles. The model formalises three core areas: (i) compositional rules; (ii) creation events, including texts and agents involved; (iii) textual evidence, distinguishing between intratextual traces and authorial declarations.

Building on this formalisation, the research aims to transition from the KG to the design of a digital platform supporting data exploration and graph querying through faceted navigation, visual metaphors, and SPARQL-based interfaces. A central challenge lies in balancing the semantic granularity of structured data with the legal restrictions imposed by copyright. Taken together, the two phases enact a shift in perspective from the description of what a text is to the formal representation of how it is made, opening new analytical possibilities for the study of complex textual forms, including combinatorial structures, linguistic interplay, and hybrid text-image architectures.