Digital Storytelling and the Art of Community: Reflections on DH Benelux 2026

Maastricht, 2–5 June 2026

The city of Maastricht was, in many ways, an ideal setting for the 13th edition of the DH Benelux Conference, which gathered some 250 scholars from across Belgium, the Netherlands, Luxembourg, and beyond at Maastricht University’s Faculty of Arts and Social Sciences (FASoS). We were among them, primarily to present our workshop on HistText.

The conference theme, “Digital Storytelling”, proved capacious enough to hold a diverse range of contributions without losing coherence. From computational analyses of large textual corpora to immersive 3D reconstructions, from participatory oral history archives to GIS-based spatial narratives, the programme reflected a field that has matured well beyond its early preoccupation with digitization and is now grappling seriously with the latest developments in computational methods. Two keynote lectures gave the gathering its intellectual spine.

 

Shawn Graham (Carleton University) opened Day 1 with a provocative talk — “Once Upon a Time: The Behaviour Space(s) of Stories” — that framed the current moment of generative AI through the lens of Terry Pratchett’s satirical observation that stories are parasitical life forms: once a story has you, you cannot escape until it is told. Graham mapped the geometry of what he called “story spaces” and asked, with a certain urgency, what digital humanities scholars can and should do now that AI has transformed storytelling into something closer to a hyperobject: everywhere, ungoverned, and accelerating. It was a very stimulating opening.

Sandra Camarda (University of Luxembourg) offered a more focused reflection on Day 2: “The Shape of the Playable Past: Narrative Systems and the Practice of History.” Her argument — that interactive historical environments can model the epistemic conditions of historical inquiry rather than merely representing its conclusions — resonated strongly with anyone who has wrestled with the gap between the messiness of archives and the tidiness of finished narratives. Both keynotes, very different in register, converged on the same uncomfortable question: who controls the story, and what do we lose when we delegate that control?

The parallel session programme was dense, and no single participant could do it justice. From our vantage point, we chose to focus on several clusters: “Session 1A (Texts & Communities)” set a strong methodological tone on the first morning, with presentations ranging from Ruben Ros’s large-scale text mining of “crisification” in Western media (1800–2025) to Yann Ryan and Anne Heyer’s use of LLMs to excavate popular sentiment in nineteenth-century Dutch newspapers. The methodological self-consciousness was notable: presenters were careful to articulate both what their computational approaches could reveal and where they hit interpretive walls. “Session 2B (3D & Multimodal Storytelling)” was a reminder that “digital storytelling” is not only a metaphor. The reconstruction of the Chest of Distress of Maastricht by Charles van Den Heuvel and colleagues, and Tim van der Heijden’s theorization of “thinkering in 3D” as a form of digital experimental media archaeology, both demonstrated the epistemic seriousness with which the field now approaches three-dimensional and multimodal forms. “Session 4A (Large Language Models)” brought the inevitable reckoning with LLMs into focus, though with more nuance than the subject sometimes receives. Alongside technically sharp papers on RAG in historical research and transformer models for orthographic variation in Early Modern Dutch, Louis Escouflaire’s analysis of how LLMs interpret the Smurfs’ distributional language offered a mix of fun and seriousness that overwhelmed the room. Overall, however, LLMs were nearly omnipresent across all the papers presented at the conference. The Lightning Talks on the afternoon  of Day 1 were a highlight in a different register — fifteen-minute windows that managed to pack genuine intellectual substance, from Marie Philippe’s computational analysis of hagiographical topoi on a global scale to Susan Hogervorst’s frank reflection on the archival dilemmas of participatory oral history.

Our own contribution to the programme was Workshop 6, held on the afternoon of the opening workshop day: “Introduction to HistText: An Application for Exploring Multilingual Text Corpora,” presented by Cécile Armand and Christian Henriot. HistText is an NLP platform designed specifically for the exploration and analysis of multilingual historical corpora. It allows researchers to interrogate large documentary collections through text mining, named entity recognition, and corpus comparison across languages. These tasks that have historically demanded either specialist programming skills or significant compromise on analytical precision. The workshop brought together participants working on a wide range of textual traditions and languages, and the discussions that followed the hands-on sessions confirmed what we had hoped: HistText addresses a genuine need in the community, particularly among scholars who command deep archival expertise but have been effectively excluded from the corpus-scale turn in digital humanities by the technical barriers involved.

DH Benelux is known within the community for the quality of its social programme, and the 2026 edition was a clear success. The opening reception at the Centre Céramique — Maastricht’s modern cultural centre on the banks of the Meuse, designed by Jo Coenen — combined drinks and conversation with a visit to the Maastricht Museum, where the city’s layered history is displayed through archaeology, historical artefacts (including a 1794 genuine guillotine), and interactive installation. The juxtaposition of medieval artefacts and digital displays felt, in context, offered a nice echo to the debates at the conference.

 

Beyond the formal events, much of what makes DH Benelux valuable happened in the FASoS garden during coffee breaks, on the group walks between venues, and in the corridor exchanges that no programme can schedule. The organization by The Plant (FASoS’s Playground and Laboratory for New Technologies) and colleagues at Maastricht University was seamless: attentive to detail without being intrusive, and clearly shaped by a genuine effort in making participants feel welcome. For a conference of 250 people spanning three parallel tracks across multiple buildings, that is no small achievement. The conversations initiated during and after the workshop will, we hope, open into longer collaborations.

---

---