With the start of my PhD project a few months ago, I entered the diverse field of Digital Humanities (DH). I started asking myself new questions and thinking about the humanities as a field in a different way: How can we use computational methods to identify patterns across large textual corpora that might not be visible through close reading? How can we balance quantitative text analysis with traditional qualitative interpretive methods in humanities research? It was really rewarding to attend my first conference in DH and to see that these questions—and many others—are shared by a whole community. The conference in question was Digital Humanities in deutschsprachige Raum 2025—thanks to the generosity of the DHd Verband and their Early Career Scholarships.
I arrived on Monday to attend the first two workshop-days. I spent all Thursday on Fortgeschrittenes Prompt und AI Agent Engineering für wissenschaftliche Textproduktion by Christopher Pollin et al., on how to effectively use GenAI for this purpose. In contrast, later that day we attended the keynote speaker Mark Dingemanse, about what makes LLMs so irresistible. With a human-like interface, LLMs trick us into trusting them by “exploiting human methods of sense-making and meaning ascription” (Dingemanse, 2025). This critical conference was a good reminder to engage with problematic scientific and ethical issues in the use of AI in academic practice, and not to be swayed by AI hype (Andrews et al., 2024).
One of the first presentations I attended was on the relationship between canonicity and Wikipedia: Wikipedia als Hallraum der Kanonizität: »1001 Books You Must Read Before You Die«. As part of the MSCA Doctoral Network MECANO (the Mechanics of Canon Formation and the Transmission of Knowledge from Greco-Roman Antiquity), I am becoming increasingly familiar with canonisation processes through my project on the reception of ancient authors in contemporary academic discourse. The approach of this contribution was a classical evaluation of a list of works considered to be canonical, in order to assess the suitability of two Wikipedia metrics as markers of canonicity: sitelinks (number of languages the article is translated to) and QRank (representing the total links to all wikipedia sites and projects) (Rohe et al., 2025).
In the context of my project, I was also very interested in some of the presentations in the Digital History contributions, such as Named Entity Uncertainty Mining: Von der intellektuellen zur computergestützten Untersuchung unsicherer Annotationen by Sophie Schneider et al., as NER in historical documents is a highly domain-specific task with many difficulties, and the bibliographic review presented in the presentation was comprehensive.
I really enjoyed Kontext, Unsicherheit und Geschlecht im Fokus der Modellierung: Datenprinzipien für die feministische Filmgeschichte by Pauline Junginger, in which attention to the contextualisation of data, to the ambiguities present in the data and to “gender” as a (provisional) category of analysis (Junginger, 2025) allows for a more precise analysis of fragmentary, uncertain data about women film pioneers.
I attended most of the panels about Computational Literary Studies, more result driven and also more methodological presentations, as my specialisation is NLP. For example, I found interesting the work Qualitative Genre-Profile und distinktive Wörter: Eine Studie zu Keyness in Subgenres des französischen Romans, which compared different statistical measures to extract the genre-specific vocabulary of a corpus (Röttgermann et al., 2025).
The Digital Humanities scene in Saxony, particularly around Leipzig, made a strong showing at this year’s DHd conference. As someone familiar with this regional research hub, I was pleased but not surprised by their significant presence. The Leipzig University’s Computational Humanities group showcased several impressive projects. Two posters were somehow related to the group: Under Construction: Ein DH-Lab im Aufbau and Explorative interface for accessing a biobibliographical data set of GDR authors. From the same institution, Wiebke Helm, Janos Borst and colleagues presented Buchkindheiten digital, sharing their progress on computer-assisted analysis of illustrations in historical children’s and youth literature. In the workshop days, the group was present in Quellcodekritik aus der Ferne: Distant Viewing und kritische Analyse von Quellcode. The Sächsische Akademie der Wissenschaften contributed several noteworthy presentations as well. Philipp Sauer’s comprehensive talk Work(s) in progress: Datenmodellierung zum Kulturerbe Tanz in der DDR als Prozess stood out, detailing the ongoing data modeling process for dance heritage from the German Democratic Republic era. I was also intrigued by the poster Vernetzung von Kulturdaten in Sachsen: Auf dem Weg zum DIKUSA Forschungsdatenregister als Schlüssel zur Datenintegration, which outlined efforts to create better integration of cultural heritage data across the region. And finally, from Universitätsbibliothek Leipzig, Annika Schröer’s poster Anticipating digital research: Der Codex Sinaiticus gibt seine Daten frei highlighted how one of the world’s oldest biblical manuscripts is being prepared and updated for computational research.
The high-level meta-reflection present throughout the conference was striking, manifesting both explicitly (in the closing keynote on Unfertigkeit als epistemischer Wert in den Digitalen Geisteswissenschaften by Mareike König, or the Metareflection space, where Manuel Burghardt advocated for Digital Environmental Humanities and Evelyn Gius discussed the operationalisation of DH operationalisation processes) and implicitly (as in the Modellierung contributions, where presentations on specific topics simultaneously examined what constitutes good modeling, confronting FAIR principles—expanding them, debating their application in all the projects and criticising them in their limitations without hesitation).
These thoughtful reflections were further enhanced by the panel format itself, where extended time frames allowed for deeper elaboration on crucial topics—from feminist perspectives inside DH to the objectives of computational literary studies and the integration of generative AI workflows in humanities research.
What emerges clearly from these discussions is that Digital Humanities as a discipline has moved well beyond any notion of being a „zero theory“ field (Anderson, 2008; Kleymann et al., 2022). The humanities continue to exert significant influence and intellectual weight in this space, effectively rendering that particular debate settled.
As Goethe wrote in his novel Wilhelm Meister: „Tätig zu sein (…) ist des Menschen erste Bestimmung, und alle Zwischenzeiten, in denen er auszuruhen genötiget ist, sollte er anwenden, eine deutliche Erkenntnis der äußerlichen Dinge zu erlangen, die ihm in der Folge abermals seine Tätigkeit erleichtert.“ („To be active is man’s first destiny, and he should use all the times in between when he is forced to rest to gain a clear knowledge of external things, which will subsequently make his activity easier again.“) Even while DH are eminently applied, they need self-reflection to evolve beyond its current boundaries.
This quote also resonates with my understanding of conferences as valuable „times to rest“—moments I particularly cherished during lunch breaks, connecting with my team and enjoying our time together. Yet these are not periods of true otium (complete leisure or idleness), but rather essential mental reprieves—breaks from paperwork and ordinary errands—that create space for deeper thinking. Without such reflective intervals, scientific research would indeed remain stagnant.
References
Anderson, Chris. “The End of Theory: The Data Deluge Makes the Scientific Method Obsolete.” Wired, June 23, 2008. http://www.wired.com/2008/06/pb-theory/.
Andrews, Mel, Andrew Smart, and Abeba Birhane. “The Reanimation of Pseudoscience in Machine Learning and Its Ethical Repercussions.” Patterns 5, no. 9 (2024): 101027. https://doi.org/10.1016/j.patter.2024.101027.
Dingemanse, Mark. Why Are LLMs So Irresistible? Oral presentation at DHd 2025, Bielefeld, March 2025.
Junginger, Pauline. “Kontext, Unsicherheit und Geschlecht im Fokus der Modellierung: Datenprinzipien für die feministische Filmgeschichte.” In DHd 2025: Book of Abstracts, 334–338. Bielefeld, March 2025.
Kleymann, Rabea, Andreas Niekler, and Manuel Burghardt. “Conceptual Forays: A Corpus-based Study of ‘Theory’ in Digital Humanities Journals.” Journal of Cultural Analytics, December 2022. https://culturalanalytics.org/article/55507.
Rohe, Jonas, Viktor J. Illmer, Lisa Poggel, and Frank Fischer. “Wikipedia als Hallraum der Kanonizität: »1001 Books You Must Read Before You Die«.” In DHd 2025: Book of Abstracts, 202–206. Bielefeld, March 2025.
Röttgermann, Julia, Keli Du, and Christof Schöch. “Qualitative Genre-Profile und distinktive Wörter: Eine Studie zu Keyness in Subgenres des französischen Romans.” In DHd 2025: Book of Abstracts, 181–186. Bielefeld, March 2025.