Daejeon, July 27–31
Umberto Eco is one of the leading Italian intellectuals, semioticians, and writers of the late twentieth century For a bibliography updated to the early 2000s, see Contursi (2005). For a recent bibliography, see the bibliography in Palazzolo (2019) and D'Agata (2025). The first edition was published by Bompiani, Milan, 1980 (Eco 1980/1983); while the revised and corrected edition was published by Bompiani, Milan, 2012 (Eco 2012). The edition currently on sale is based on the latter, published by La Nave di Teseo with illustrations by Eco himself (La Nave di Teseo, Milan, 2020). www.variantidellarosa.it
The ‘IDEA’ paradigm emerges from the need to overcome the traditional compartmentalization of digital philology (D’Agata 2022). Interpretation refers to the set of exegetical and hermeneutic practices (Sichera 2017) accompanying a text and which, within the hyperedition, are no longer confined to footnotes. Instead, they unfold through multiple layers of annotation and visualization, involving an active role of both encoder and reader, who select what to mark and display according to their specific textual focus. Didactics concerns the potential of transforming the portal into a tool for schools and universities, through interactive exercises, guided pathways, and multimedia materials that turn textual complexity into an opportunity for continuous discovery. Edition identifies the core processes of textual representation based on contemporary digital philology, while Annotation involves the use of formal languages (XML-TEI and DSL) that make the text (and its commentary) computable, queryable, and reusable (Wilkinson, et al 2016) – a process that, in turn, cyclically depends on the critical interpretation of the editor.
The shift from IDEA to IDEA(S) introduces Storytelling as a fifth, integrative dimension (Fig. 1). In this model, storytelling is not conceived as a simplification or popularization of scholarly content, but as a strategy for structuring and communicating philological knowledge. Variants are presented as narrative moments in the life of the text, allowing readers to follow the evolution of authorial choices, stylistic adjustments, and semantic shifts over time.
More specifically, the digital edition of the variants was developed through the collation of the 1980 and 2012 texts (acquired via OCR) using automatic collation algorithms (CollateX https://collatex.net Cfr. https://journalofdigitalhumanities.org/3-1/juxta-commons/. As of today (05/2026) the Juxta commons website is unavailable. https://github.com/cs6-uniwue/Variance-Viewer Cfr. https://tei-c.org http://evt.labcd.unipi.it
The variants were then annotated with a Domain Specific Language (Parr 2007; Parr 2014; Zenzaro, et al 2025) through the Euporia software (Boschetti, et al. 2023; Boschetti / Del Grosso 2020; Mugelli et al. 2016), developed by CNR-ILC. The DSL provided generic annotations concerning the communicative context of each variant (the characters involved, the thematic and narrative elements of the paragraph) as well as linguistic annotations (Fig. 3), particularly based on Tullio De Mauro’s (1999-2000) usage labels (“fundamental vocabulary,” “high usage,” “low usage,” “literary,” “specialized”). This framework enabled a range of statistical analyses, which helped define Eco’s corrective interventions not as simplification, but as a rhapsodic exercise in labor limae.
From the standpoint of engagement, the crucial elements lie not only in the richness of philological data, but also in their transformation into narrative experiences. The portal includes a section dedicated to the storytelling of variants, where significant transformations are visualized using TRAViz http://www.traviz.vizcovery.org
Another example of storytelling is with StoymapJS and Timeline JS (Fig. 5). The aim is to make the variant not a technical artifact accessible only to specialists, but an entry point into understanding the novel’s construction of meaning A new release planned for 2026 consolidates the IDEA(S) paradigm by integrating research infrastructure and storytelling-driven engagement: 1)A variant search engine, developed as a web application based on eXist-db, enabling complex queries on words, usage labels, characters, and themes, as represented in the Euporia-generated annotations; 2) An experimental linguistic and interpretive commentary system, generated in natural language and supported by large language models (GPT, Gemini, Claude, DeepSeek), combining philological accuracy with narrative-driven strategies to present variants as evolving stories of textual change, making scholarly interpretation accessible to non-specialist audiences without sacrificing critical depth.
To illustrate the IDEA(S) model in practice, we conclude with a single variant treated across all the encoding layers it traverses in the portal. The case is variant XIII of The Name of the Rose, located at the opening of the description of Guglielmo di Baskerville. In NR80/83, William’s portrait begins with a sentence overtly modeled on the incipit of A. Conan Doyle’s A Study in Scarlet: «Era dunque l’apparenza fisica di frate Guglielmo tale da attirare l’attenzione dell'osservatore più distratto» – a near-translation of «His very person and appearance were such as to strike the attention of the most casual observer». In NR12, Eco silently removes the Holmesian opening and rewrites the paragraph so that the description now starts with the more neutral «La statura di frate Guglielmo».
At the philological-data layer, the variant is encoded according to the TEI parallel-segmentation model:
This level guarantees machine-readability and visualization through eXist-db (Fig. 6); it makes the variant queryable, FAIR-compliant, and reusable for subsequent computational analysis.
At the interpretive-annotation layer, the same variant is encoded in the Euporia DSL, which captures dimensions that TEI cannot easily express – speaker, topic, hermeneutic cause, possible-world status:
& XIII
*[0] @narratore §Guglielmo/descrizione
*[1] {Era dunque l'apparenza fisica di frate Guglielmo tale da attirare l'attenzione dell'osservatore più distratto. } = §alleggerimento/maior §riferimentoSherlock §variazionePersonaggio §mondoPossibile/globale
*[2] {La sua statura} : <La statura di frate Guglielmo> = §esplicitazione
At the public-facing layer, the same variant becomes a short narrative addressed to a non-specialist reader:
The Holmesian opening. NR80 begins its portrait of Guglielmo with a phrase clearly derived from Conan Doyle: “It was, therefore, the physical appearance of Brother Guglielmo that was such as to attract the attention of even the most casual observer.” The echo is that of the first description of Sherlock Holmes in A Study in Scarlet (“His very person and appearance were such as to strike the attention of the most casual observer”). NR12 erases everything. Thus begins the systematic “de-Holmesification” of the character, which continues in & XIV and XV...
This third layer, drafted by the editor and refined through LLM-assisted commentary (with prompts conditioned on the DSL annotation as structured input) A related reflection on the interaction between DSLs and LLMs has been accepted for presentation at the AIUCD 2026 conference: D'Agata, Christian / Del Grosso, Angelo M. / Boschetti, Federico, "Per una formazione del filologo computazionale del futuro: Domain Specific Language e Large Language Model" (forthcoming).
Variant XIII demonstrates the architecture of the hyperedition we propose. The same unit of textual change is rendered as TEI-data, as DSL-interpretation, and as narrative-experience, with traceable links between the three. The Storytelling layer does not replace the philological apparatus: it is generated from it. In this sense, IDEA(S) is not an additive paradigm: Storytelling is what holds together Interpretation, Didactics, Edition, and Annotation when the digital scholarly edition steps outside the academy. Le varianti della rosa offers a working prototype of this replicable workflow – collation, TEI encoding, DSL annotation, narrative rendering with LLMs – for digital editions that aim simultaneously at scholarly rigor and public engagement.