Tuesday June 3rd – Morning Sessions
Christoph Hafner (City University of Hong Kong)
Lecture: Approaching ‘Digital Discourses’ from a Digital Literacies Perspective
Digital Humanities scholarship shares an interest in digital texts with various branches of applied linguistics and literacy studies, including those concerned with digital genres, computer-mediated communication, and critical digital literacies. In particular, one of the aims of Digital Humanities scholarship is to critically interrogate digital technologies and the social practices that surround them. One way to achieve this aim is through the analysis of ‘digital discourses’, i.e. ‘digitally mediated texts and interactions and their associated ideologies’ (Hafner, 2018). In this presentation, I describe and explain a digital literacies perspective on digital discourses and practices. This perspective is grounded in a social practice approach to language and literacy that treats digital literacy practices as socially situated events blending online and offline contexts. Drawing on a range of examples, I argue that such a social practice approach to digital discourses provides a contextually rich understanding of sociotechnical structures and associated power dynamics that may inform Digital Humanities research.
Workshop: Doing Digital Ethnography: Multiple Perspectives on Digital Cultures and Practices
In order to gain an in-depth understanding of digital cultures and practices it is helpful to extend the investigation beyond texts alone to incorporate the way that such texts are experienced by their producers and consumers. Digital ethnography, a methodology that is grounded in the field of anthropology, provides an approach to research on digital practices that seeks to capture the emic perspectives of participants in such practices. In this workshop, I aim to provide an accessible introduction to this approach. I will provide definitions of ethnography and digital ethnography and describe underlying assumptions. Then, drawing on some example studies, I will demonstrate some common strategies and techniques, as well as address potential issues such as ethical concerns.
Tuesday June 3rd – Afternoon Sessions
Josef Schmied (TU Chemnitz)
Lecture: Understanding AI tools for revising and editing research writing
This presentation explains how non-native PhD students can use current Artificial Intelligence (AI) tools to improve their own texts for publication (journal articles) or submission (PhD theses). First, we discuss linguistic perspectives to understand how AI tools work, especially Large Language Models (LLMs) like ChatGPT, Gemini, Claude, Perplexity, Mistral, DeepSeek, etc. Then, we look at AI detection and correction tools. We focus on individual preferences and how authors can use all tools to develop their personal style within the genre and publishing frames. Finally, we debate basic principles of academic integrity that have to be maintained when using AI tools.
Workshop: Practising revising and editing research writing with Large Language Models
Based on their own experience, participants chose their preferred LLM to improve standard academic texts, including their own, and discuss their preferences in small groups to try a SWAT analysis of their individual LLM experience.
Wednesday June 4th – Morning Sessions
Marcus Müller (TU Darmstadt)
Lecture: Chatting with Large Language Models in corpus linguistics
In my lecture, I will explore the application of Large Language Models (LLMs) in corpus linguistics projects conducted at the Discourse Lab. We mainly employ LLMs for two types of tasks: corpus preparation and classification of text segments. I will illustrate this with two research projects.
The first example is drawn from the interdisciplinary project Terminological Innovation in International Relations, which traces the trajectories of key political science terms—such as regime, governance, and international anarchy—as they circulate between academic discourse, policy consulting, and practical politics (1976–2000). This requires constructing a complex, multi-part corpus comprising texts that are in part available only as image scans or even in paper form. As a result, corpus creation demands considerable effort. In this context, LLMs have proven to be transformative in addressing OCR-related challenges, although they also bring their own complications. A second application of LLMs within the project involves the disambiguation of key term senses across different contexts.
The second use case is situated within the context of the DFG research group Controversial Discourses, where we investigate political discourse in Germany since 1990. Among our methodological goals is the identification of argumentation topoi to detect implicit knowledge structures in argumentative texts. Recognising these topoi is often challenging, as even experts find it difficult to agree on where exactly a topos is located, its boundaries in a text, and the contextual information necessary for its recognition. I will present classification experiments using LLMs to identify the utility topos in a dataset of bioethics discourse.
Workshop: Chatting with Large Language Models in corpus linguistics
In the follow-up workshop, I would like to revisit the topic of LLMs in applied corpus linguistics and discuss with you the opportunities and risks of using LLMs in the humanities and social sciences. We’ll do this through a hands-on classification exercise on word sense disambiguation, using the Hugging Face chat platform: https://huggingface.co/chat/
Wednesday June 4th – Afternoon Sessions
John Bateman (University of Bremen)
Lecture: Scaling-up qualitative multimodal analyses: theoretical principles for practical variable-scale projects
As the kinds of communication that are considered legitimate targets of analysis grow in their medial complexity, one needs ever stronger theoretical and practical tools for engaging with them effectively.
In this lecture, I set out the current views we have developed for multimodal analysis, showing the tight interconnection of appropriate theoretical constructs, such as semiotic mode, medium, genre, material, and canvas, with practical procedures for analysis, both manual and automatic. Examples will be given of how these constructs combine to allow research on complex multimodal communication at all scales ranging from individual close readings to larger-scale corpus work with annotated data sets.
Workshop: Scaling-up qualitative multimodal analyses: mixing manual and automatic analyses of complex multimodal artefacts
In the workshop component, we will explore how the constructs introduced in the lecture are actually applied to data of various kinds. Examples will be worked through in order to practise the application of the constructs, with a specific emphasis on how research can be guided and organised in order to cope with complexity. The relationship between theoretical description and coding and annotation for the purpose of finding reliable patterns will receive particular emphasis. Participants are encouraged to ‘bring their own’ examples of challenging materials to be subjected to multimodal analysis so that the constructs introduced can be worked through on the basis of concrete examples.
Thursday June 5th – Afternoon Sessions
Astrid Ensslin (University of Regensburg)
Lecture and Workshop: Videogames, Digital-born Fiction and DH: Distant Reading/Play and Ludolinguistic Experiments
Abstract: In this interactive lecture and hands-on workshop, we‘ll be looking at the challenges and opportunities of studying videogames and other forms of digital-born narrative within a Digital Humanities framework. We explore ludolinguistic discourses, theories and analysis and examine games as a form of digital-born fiction. Through critical engagement with a blend of theoretical material, mixed media demos and analytical/creative practice, students will gain insights into core aspects of medium-specific multimodal discourse analysis, Critical Making, distant reading and play, as well as accessible ludolinguistic corpus analysis.
Friday June 6th – Morning Sessions
Lynne Bowker (Université Laval)
Lecture: The evolving automation of translation: Applications and implications of less and more automated translation tools
As translation technologies have evolved (e.g. concordancers, translation memories, machine translation, generative AI), the degree of automation has increased. At the same time, user interfaces have become simplified, and tools have become more easily accessible. But the ease of access and use make it possible to engage with translation in a superficial way and to operate the tools in an uninformed or non-critical manner. In the race to discover what AI can do, we risk employing it in cases where it does not offer a clear advantage and where it may even have negative effects. Less automated tools require more interpretation on the part of users, but this can allow for increased nuance and improved quality. This lecture will explore the strengths and limits of less and more automated translation tools.
Workshop: Concordancers, NMT and LLMs: Comparative analysis and practical applications for Translation
This practical session will include hands-on comparisons of online bilingual concordancers, Neural Machine Translation and translation tools based on LLMs. It will cover activities such as interpreting concordance lines; generating effective translation prompts; pre-editing and translation-friendly writing; and creating glossaries.
Friday June 6th – Afternoon Sessions
Wu Ping (Beijing Language and Culture University)
Lecture: Using a Large Language Model in Emotional Analysis: A Case Study on Yu Hua’s Novel To Live
Based on linguistic appraisal theory, this study offers a sentiment analysis determining the emotional tone of Yu Hua’s novel “To Live”, its translation by Michael Berry, and two machine translations by GPT-4 and ERNIE-4. The sentiments expressed in these four versions are quantified with a large language model (LLM). Using statistical and visualization tools, the LLM is shown to reveal accuracy differences between the sentiments expressed in the original and the human and machine translations. Results improve our understanding of Yu Hua’s work and evidence the potential of integrating linguistic theory with computational linguistics to provide a new interdisciplinary research paradigm.
Workshop: Practice of Using a Large Model in Emotional Analysis: Segments from Il Colibrì
This demonstration experiment is a complete Python demo showcasing an emotion analysis practice using large language models (LLMs). It explores a new research paradigm integrating computer science and linguistics under the perspective of digital humanities in the AI era. The demo employs selected excerpts from the Italian novel Colibrì, with English translations generated by GPT-4 and ERNIE-4. A multi-stage, LLM-driven computational approach is applied to run several emotional tasks in both translated texts. Finally, the results are visualized using Python for comparative assessment.