Digital Humanities Conference “Inclusivity and Citizenship“ 2025 – our highlights

Lisbon, five days and lots of workshops, talks and panels addressing ideas, methods, tools, data, and software for digital Humanists during the Digital Humanities Conference “Inclusivity and Citizenship” (DH, 13.-18th of July 2025, https://dh2025.adho.org/). We are delighted that four IC members, Paul Bayer, Torsten Hiltmann, Till Grallert and Carolin Odebrecht, had the chance to attend this important conference and to present some of the key aspects of our centre with our joined poster contribution “Centering Digitality”!

We would like to highlight some important contributions below, knowing that this short overview would just represent a small excerpt of what happened at the conference.¹ However, we hope that some of our highlight might enlighten you and that we can inspire you for topics of the conference surrounding digital humanities, cultural and social science!

Our Contribution: Centering Digitality

The IC D2MCM's contribution © IZ D2MCM 2025

Our poster contribution presents the concept and rationale of the Interdisciplinary Centre for Digitality and Digital Methods at Campus Mitte of Humboldt Universität zu Berlin (IC D2MCM). The IC focuses both on digitality and digital methods as an epistemical construct and bridges diverse fields such as Digital Humanities (DH), social sciences, media and cultural studies, law, economics, and religious studies. We discussed the goals and the governance of our centre, where especially the member-driven working groups were widely acknowledged as being more important than digital services usually associated with DH centres. This is underlined by the section part of the contribution: in the working group Digitality² we asked ourselves what constitutes the phenomenon of digitality? With many insights from the discussion at the conference, we developed a good starting point for our future work: we are building a theory-driven centre!

Our Highlights

or what we managed to attend during this immense interesting conference

Godot in the Digital Humanities

For Paul Bayer (one of our student assisants), a highlight of the conference occurred already on its first day. At the workshop “Creating Interactive 3D Applications with the Open-Source Game Engine ‘Godot’ – A DH Hackathon/Game Jam”³, the speakers and participants discussed gamification and the use of the Godot game engine in the context of the Digital Humanities. Paul Bayer had independently conceived and organized the workshop “Interaktive Visualisierung von Wissen mit der Godot-Engine”⁴ in July 2024. It was therefore encouraging to see how Godot is being applied elsewhere in the DH world. While the speakers focused on applications in museum studies, archaeology, and cultural heritage⁵, Bayer’s own approach involves using games as mnemonic devices, similar to a memory palace. Fascinatingly, a Dutch conference participant mentioned knowing someone who had written a PhD on medieval memnonic techniques and who had developed a digital app for making memory palaces.⁶ This is exactly the kind of serendipity that can happen at conferences like this!

Text, Data and Responsibility

In the context of creating digital edition based on historical manuscripts, using HTR and TEI, researchers discussed whether „manuscript encoding still have a place in Digital Humanities”?⁷ The answer depends on the re-usability of guidelines, infrastructure, software and code. This holds also for modern manuscripts⁸ where transcription guidelines seems to rescue HTR project for encoding Doctors’ manuscripts, which are indeed are illegible and highly variant. Next to manuscripts, the encoding and analysing of periodicals⁹ addresses discourse analytics as a proxi for the interplay between language, culture, and identity.

An important impulse was presented with antiracist, decolonial, and inclusive TEI guidelines.¹⁰ It proposes to identify words/phrases that would be nowadays, e.g. culturally discriminating, contextualise them, and explain where the terms might possibly used differently in terms of ethics. Documentation is also discussed as a method serving as a powerful tool for positive change by making our research and datasets more transparent with regard to data contexts and biases.¹¹ A comparison of FAIR testing tools presented insights in the different implementation of what FAIR could mean for and to researchers.¹²

genAI-based approaches

With AI Emily, a pilot parallel corpus of 40 original and 360 AI-generated poems by, and in the style of, Emily Dickinson, was tested, whether stylometrics can classify original poems by Emily Dickinson from AI-generated poems.¹³ Preliminary results seems to show that AI systems have “their own style”. Researchers also tested personal LLM by creating a performative identity¹⁴ based on personal writings and research outputs by using the HALO ontology for classifying hallucination errors. ¹⁵

Vision models were also trained, based on manual annotated ground truth digitalisations and tested in order to identify gendered poses in Renaissance Art.¹⁶ Outcome: the model fails to distinguish culturally distinct gestures, which highlights a key focal point in the discussions of limitation of AI models related to domain-specific and cultural influenced concepts to during the entire conference.

Handwritten Text Recognition

There were several talks on HTR. At several points, ways in which national and institutional divides shape the conditions in which HTR can be practiced were addressed.

One example of this is how training general-purpose HTR models for the cultural heritage of some Nations gets helped or hindered by how many HTR projects there have been in the past for these specific national languages and scripts, and how big these projects were. For nations like France with a dozen different, smaller projects, this meant that there existed many distinct, project-specific standards, heightening the difficulty of generalizing them. Meanwhile the fewer, bigger projects in the Netherlands meant that a lot of the data was very uniform, which aided the work. ¹⁷

On the automatic transcription of Spanish calligraphic treasure fleet records, Rodrigo Vega-Sánches detailed how many hitherto developed HTR models for Early Modern Texts were trained on Northen European styles, which caused problems when applying them to Southern European historical materials. This makes his and his team’s efforts to train models for four distinct Spanish calligraphic styles a very important task.¹⁸

Panel Digital History

Another highlight of the conference was the panel on digital history. The panel discussion addressed the need for a new data culture in history and beyond, with the aims to understand the fundamental epistemological affordances of the post-digital moment; to develop the necessary quotidian practices and disciplinary protocols; and to negotiate new understandings of history as a discipline of societal relevance.¹⁹

Torsten Hiltmann’s input was about the process of integrating the Digital into the academic field of history. He argued that now is the the time that history changes through datafication, which puts an end to the “eternal promise of Digital History”. After beginnings in cost-intensive datafication projects of the past, and an intermediate stage of increasing accessiblity, he now sees the field at a point where the digital is the new normal. The foundations of history are changing. History is now part of the digital world. This brings about a new challenge, however: Especially with the advent of generative AI, more and more complex systems are becoming easier and easier to use. Wrong outputs still look good when they are presented as smooth natural language. Hiltmann argues that the step towards using LLM can be called post-digital. As a possible way forward in this new environment he identifies three priorities for the field of history: 1. Documentation, 2. Guidelines, 3. Evaluation. The most consequential decisions are made the first two of these steps. Digital historians also need to engage with the way that digitality and AI change “doing history”. Beyond “the eternal promise of Digital History” they should aim to “explain what is [now already] happening and how it happens”.

Min-Woo Lee detailed challenges for digital history when it comes to mapping historical places and names across time in Korea. For the different epochs since the Middle Ages, documents exist in four different languages: Classical Chinese during the Joseon dynasty, Japanese during the occupation, English during the Korean War, and finally Korean in Hangul script. Beyond linguistic and paleographical differences, which already pose challenges to digital standartisation, each change of power brought with it a different paradigm of record-keeping: Lee showed slides land that showed the difference between Joseon Yangan and land registers from the Japanese occupation. While the former took rather the form of a handwritten table, the latter was a topographical map. The way people thought about land itself was very different and that was reflected in the records. Mapping terms between the two formats is hard. Names changed and land was restructured. Lee argues that we need a digital history that embraces the kind of ambiguity that arises from changing types of sources like these.

Ian Kisil Marino’s talk revolved around the topic of informal digital history. For defining Informality, he referenced the concept of Informal Archives (Auerbach 2018), Digital Informal Archives (Mariono, Silverira, Nicodem 2022), and the Laboratory Turn (Pawlicka-Deger 2020). In his research, Marino investigates institutions, archives and practices of digital history in Latin America. He brought attention to the fact that the Digital Humanities are often seen as something ubiquitous but that the international situation differs wildly, especially in the Global South. In the Brasilian speakers case at his university, there is a Digital History website, but no dedicated money and or staff for Digital History. How, then, are Digital Humanities done in Latin America? The answer is complicated. It’s hard to monitor how many DH projects, associations, labs, journals, and courses there are. In terms of institutionalization, Marino’s findings indicate that history is the main discipline within Latin American Digital Humanities. In terms of activities, DH people do research, provide services and are involved in teaching. Building a data culture is a common goal between the different locations. Finally, much of Digital Humanities work in Latin America is informal. Its study is in large part requires methods from ethnography.

Julianne Nyhan presented on the state of Digital History regarding Oral History. While in the past, the digital stood for dissemination and accessibility of audio data, now with advanced AI methods, more ways of analysis become possible. In her work analyzing interviews, she Whisper AI to access parts of interviews that get omitted by human transcriptors. These are paralinguistc phenomena such as laughter, hesitation, and omission. Using this example, she warned of the “paradox of apparent well-formedness”: Outoputs that are formally coherent and computationally valid but fraught in ways that are only apparent to experienced humanists. In the discussion between the panel speakers, issues such as the lack of a good way to detect hallucination were discussed, as well as the artificial split between regular historians and digital historians. Historical questions and digital methodologies need to be reconciled, and we need a strong human skill to critique the ouput of machines, the speakers argued.

Conference program https://dh2025.adho.org/browse-the-program-agenda Accessed 30/07/2025. ↩
Working group Digitality https://izd2m.hu-berlin.de/content_en/wg/digitality/digitality_about_en.html accessed 30/07/2025. ↩
Peter Mühleder, Franziska Naether, Dirk Goldhahn, Patrice Bleckmann: “Creating Interactive 3D Applications with the Open-Source Game Engine “Godot” – A DH Hackathon/Game Jam”, URL: https://www.conftool.pro/dh2025/index.php?page=browseSessions&form_session=272&presentations=shows accessed 30/07/2025. ↩
Paul Bayer: Blog post “Interaktive Visualisierung von Wissen mit der Godot-Engine”, 13.02.2025, URL: https://izd2m.hu-berlin.de/blog-posts/2025/02/13/bp-workshop-godot.html accessed 30/07/2025. ↩
Works cited by the speakers: Virtual Reality Oracle project (https://vroracle.co.uk/), Giza Project: (http://giza.fas.harvard.edu/giza3d/), Virtual Viking Longships Project (https://virtualvikings.sites.grinnell.edu/). Another work mentioned by a participant during discussion: Walden, A Game, (https://www.waldengame.com/). All accessed 30/07/2025. ↩
Munx VR, Linguisticator (https://linguisticator.com/p/munxvr). For another thematically adjacent project, compare: Museum of all Things (https://godotengine.org/article/museum-of-all-things/). All accessed 30/07/2025. ↩
“Vital Signs Between the Lines? Reconsidering Textual Genesis Encoding in a Digital Future”: Brett Barney, Katrin Henzel, Joshua Schäuble, Nooshin Shahidzadeh Asadi, Ashlyn Stewart ↩
“Transcribing Western modern manuscripts (1500-2020): an economical, ecological and secured approach”: Simon Gabay, Tobias Hodel, Ronald Sluijter, Élodie Paupe, Jean-Claude Rebetez, David Rabouin, Vincent Giovannangeli, Walter Boente, Elodie Bascoul, Marion Philip, Marie-Laure Massot, Vincent Ventresque, Serena Crespi, Pauline Jacsont, Yvan Jauregui, Loraine Chappuis, Esther Solé, Elias Zimmermann, Maxime Humeau, Myriam Lamrayah, Justine Falciola, Alix Chagué ↩
“Accessing Historical Periodicals: Newspaper Discourse on Slovene Language” by Vojko Gorjanc, Ajda Pretnar Žagar, Filip Dobranić, Darja Fišer ↩
“Reading Spanish NovEllas through an Antiracist, Inclusive, and Feminist Text Encoding Framework” by Sarah Revilla-Sanchez, Elizabeth Lagresa-González ↩
“Documenting datasets as a tool for Change”: Sarah Lang ↩
“How equal are tests of FAIRness? - A comparative evaluation from a domain-specific perspective” by Steffen Pielström, Kerstin Jung, Patrick Helling ↩
“Understanding AI Emily: Designing an AI-generated lyric poetry dataset for evaluation Experiments” by Judith Bishop, Ruby Mineur ↩
“Talking to Myself: Examining Narrative Identity with Personalized Large Language Models” by Sarah Grace Immel, Beatrice Alex, Susan Lechelt ↩
See URL: https://arxiv.org/abs/2312.05209, accessed 30/07/2025. ↩
“Exploring Gendered Poses in Renaissance Art: A Computational Analysis of Activity and Passivity” by Brianah N. T. Lee, Giulia Speca, Celis Tittse, Lisandra Costiner ↩
Gabay et al. 2025. ↩
“Progress of The New Spain Fleets Project: accurate Handwritten Text Recognition models for 16th-17th century Spanish calligraphies” by Rodrigo Vega-Sánchez, Edna Brito-Ramos, Francisco Cruz-Ríos, Fryda Montiel-Alejos, Andrea González-Aceves, Abril Hernández-Ronquillo, Martín Díaz-Vázquez, Ricardo Valadez-Vázquez, Lidia Camacho-Gamez, Guillaume Candela, Mariana Favila-Vázquez, Flor Trejo-Rivera, Alexander Sánchez-Díaz, Patricia Murrieta-Flores ↩
See URL: https://www.conftool.pro/dh2025/index.php?page=browseSessions&form_session=152 (Accessed 17/09/2025) ↩

Cite as: Carolin Odebrecht, Paul Bayer: Our highlights from the DH conference in Lisbon (2025). In: IZ D2MCM Blog [Weblog], 17.09.2025. URL: https://izd2m.hu-berlin.de/blog-posts/2025/09/17/bp-dh-lissabon-2025.html. Zuletzt aktualisiert am: 17.09.2025.