When digital humanities meet East Asian studies: the state of play and the way forward

By Kathleen Schlütter

In the second week of July 2024, about 130 researchers in East Asian studies from Europe and beyond came together in Staatsbibliothek zu Berlin (Stabi) to discuss the digital humanities (DH) approaches and projects in their field at “Charting the European D-SEA”. It was the first meeting of its kind in Europe, and I would like to share some observations that are relevant to the broader discussion about the future of area studies and DH.

I am a cultural historian working with digital humanities methods, without any particular relation to East Asia. However, I am working on how the humanities and social sciences, and area studies in particular, are changing. I attended this event because it is directly related to my research project on the current developments in area studies in Germany and my engagement in area histories within the historical sciences consortium for the National Research Data Infrastructure (NFDI), 4Memory.

The power of cooperation

The D-SEA conference was a combination of hands-on workshops and presentations. It was a joint event of a non-university research institute and a library, the Max Planck Institute for the History of Science (MPIWG) and the Stabi. These two partners have known each other for  several years and cooperate on a project level in the context of the CrossAsia portal. To me, this was a great example of what can be achieved when there is a willingness to cooperate and a common goal. It also shows the potential of libraries as research organizations and partners for DH projects (see Bonte 2023).

The density of innovative approaches and thinking ahead of their fields left me with positive impressions, many learnings and new ideas. From the current discussions about the state of area studies in Germany (or Great Britain, or the US, or France), one would expect conversations about underfunding, a lack of institutional support, and poor career prospects. In addition, recent developments in large language modeling (LLM), natural language processing (NLP), and others are rightly criticized for being too centered on the English language and/or the Global North, making it difficult to work with languages with non-Latin characters (Viola and Spence 2023; Wagner 2018). And let’s not forget the possible ethical and legal challenges and different perceptions of “FAIR” data use and publication principles when working with sources and data from contexts outside of Europe (Wilkinson et al. 2016; Arnold et al. 2022). While all of this is true and needs to be addressed (Ommert and Schlütter 2024), it was not the focus of this conference. Instead, I attended presentations by researchers who have successfully found ways to address their domain-specific DH challenges.

Not all non-Latin script languages are underserved by digital tools

This seems to be largely related to two facts: (1) Governments in the People’s Republic of China (PRC) and South Korea, and to a lesser extent in Taiwan and Japan, specifically fund digital humanities projects. (2)  Most of the attending researchers work on Greater China, and obtaining source material from this region has worked rather well in recent decades. Also, Chinese (in its various historical and contemporary forms) and Korean are rather well-implemented languages in terms of optical character recognition (OCR). Japanese institutions have been more reluctant to share their digitized historical sources and allow access from outside of Japan. This is part of the answer as to why scholars working on Japan still “wrestle with fundamental challenges like character recognition,” as Japanologist Alíz Horváth (Eötvös Loránd University) said in her presentation on OCR in Japanese Studies. The size and economic power of a research area are important, but so are cultural issues.  

Area studies and DH: an additional layer of complexity

The use of digital tools in the research context is still a matter of personal interest. Some of my coffee break conversation partners told me that, in their experience, East Asian studies are still largely non-digital. They share this with most other humanities and social science disciplines. As sinologist Peter K. Bol (Harvard University) pointed out in his presentation on historical developments in DH and Chinese studies, none of these developments came from the humanities themselves, and this is not what we were originally trained for. However, digital methods do open up new possibilities for analysis and allow new questions to well-known sources. Since many fields of research are currently moving in this direction, they should, in my opinion, not be ignored by area studies. If area studies fails to recognize current technological trends, it once again risks being seen as insufficiently innovative after the severe structural crisis of area studies in the aftermath of the Cold War (see Middell (ed.) 2013; Mielke and Hornidge (eds.) 2017). More importantly, infrastructural developments in research technologies might not take into account sufficiently the needs of area-related research fields. Finally, research in the “Global North” needs to take more account of the context in which theories emerge and research results are produced. The D-SEA conference illustrated the potential of such context-sensitive approaches in DH, and I think academia needs more of this.

The question of how to integrate digital methods into a curriculum remains unanswered. Where to make room in already packed curricula? What level of IT expertise is required: just being able to use a tool for a certain kind of analysis, or being able to code in Python or R? This may not be possible with a one-semester course in digital methods. It becomes even harder if you are studying a specific world region and need to learn one or more languages in order to be able to do valuable research there. While the introduction of statistical methods and more quantitative tools into political science and sociology has already been a serious challenge for area studies (Stevens, Miller-Idriss et al. 2018), coding adds a significant layer. The good news is: While historical, social, geographic, and linguistic contexts under study are unique in their own right, the challenges when it comes to digital methods may be more similar than one might think: dealing with multilingual corpora, analyzing and visualizing networks, geographic spaces, mobilities, topics… So, there is a lot to learn from each other in terms of problem solving, data storage and platform building.

Technological progress is very fast

As digital humanities specialist Shih-Pei Chen, one of the co-organizers, put it: “Change is fast, ChatGPT is only two years old and has changed everything.” East Asian researcher Matthias Arnold (University of Heidelberg) presented the open access platform “Early Chinese Periodicals Online (ECPO),” which has brought together collections of newspapers from the Chinese Republican era and provides scans and metadata for articles from some 300 periodicals, as well as some full-text passages. Arnold reported on the various technical challenges his team had to surmount to achieve OCR obstacles in the years until 2022, concluding that this research is “two years old, it is stone-aged now,” referring to the fact that OCR technology has advanced at a monthly pace. This was also illustrated in a workshop on ChatGPT and humanities research by IT developer Calvin Yeh (MPIWG). He pointed a variety of ways in which researchers can incorporate LLM tools into their research process to make it more efficient, such as summarizing texts, analyzing data, or prompting a syllabus for a seminar, or which alternatives existed to ChatGPT, such as Claude (more suitable for longer texts and better data protection). On a more anecdotal level, this was illustrated by the conference poster with an image generated by ChatGPT – so without copyright, but credited to D-SEA co-organizer Jing Hu (PhD student in East Asian studies, Stabi/Leiden University) for the prompts used for image generation.

A concrete bottom-up approach leads to technological solutions that are beneficial to others. The question is, how to know about them

A tool or platform emerging from a concrete research question, can be highly useful for others. A striking example for me was HistText, an open source platform conceived by a team around the sinologist Christian Henriot at Aix-Marseille University, working on a project about elites in China between 1830 and 1949. The system allows for the analysis of large corpora of full-text sources, including dealing with language complexities such as having sources in both Chinese and English, and searching, querying, and extracting data from them. HistText can by applied to any kind of language and corpora. While it was already known to many researchers in Sinology, it was new even to researchers in other East Asian research fields, such as Japan. Other examples include MARKUS and its subsystems, COMARKUS, ENTMARKUS, and IMMARKUS, tools designed for Chinese and, more recently, Korean texts, which allow the construction of entity relationship schemes and the automatic annotation of texts and images. They can be used for other languages as long as they are encoded in Unicode, and a set of keywords is provided (De Weerdt 2020). MARKUS has been developed since 2016 by the sinologist Hilde de Weerdt and her team at KU Leuven. The experience of the researchers behind databases like LoGaRT, where 8,000 digitized Chinese local gazetteers are made available for historical research (Shih-Pei Chen and Calvin Yeh, MPIWG), or the China Biographical Database (CBDB) with biographical information from over 530,000 individuals from the 7th to the 19th century (initiated by social historian Robert M. Hartwell in the 1990s and further developed by the sinologist Peter A. Fuller, now supervised by Peter K. Bol in a cooperation of Harvard University, Academica Sinica, and Peking University) would surely be helpful to scholars in other area studies fields who are trying to create something similar on the basis of their sources. And this is only what I learned from some of the presentations I attended. Other talks were about social media analysis, network analysis or spatial analysis. On a side note: medium-sized or larger research projects seem to be particularly suitable to develop research infrastructures that can be used beyond the concrete project. The MARKUS tools as well as HistText received, among other presented projects, multi-annual funding from the European Research Council (ERC).

As a kind of conclusion, one question remains for me: How does one come to know about such tools, especially if they come from an area-related research field that one is not connected to? This is related to the fact that thinking beyond one’s own discipline, research environment, and personal research habits is a challenge in general. While it has always been the case that a research question can only be answered with the methods and tools available at that moment, the speed of progress in the field of DH is remarkable and challenging. I have no answer to this, but I think it is crucial to invest time in getting to know new approaches and projects from other disciplines and areas than one’s own. In terms of DH and area studies, this might happen at future events such as D-SEA or in summer schools that address PhD or MA students and include projects and approaches on and from other regions of the world (South Asia, Sub-Saharan Africa, Maghreb, Latin America, Eastern Europe…). This could be a way to provide a basic understanding of the different technologies at an early stage of a researcher’s career, in order to be enable future generations of area specialists to consider if and which DH specialization might be useful for their research questions.

I am grateful to Thorben Pelzer and Eva Ommert for their comments on earlier versions of this text.

---

About the Author

Kathleen Schlütter is a research associate at the Research Centre Global Dynamics and conducts research on the changes in knowledge production in the global condition, currently using the example of ‘area studies’ in Germany. She wrote her doctorate at the University of Leipzig in Global Studies on French higher education and research policy under Jacques Chirac and Nicolas Sarkozy. Her research interests include in particular higher education and research policy in the knowledge-based economy, digital humanities, transnationalisation, cultural transfer, history of science and France in the 20th and 21st centuries.

---

References

Arnold, Matthias, Duncan Paterson, und Jia Xie. „Procedural Challenges: The FAIR Principles and PRC Electronic Resources – a Case Study of Chinese Republican Newspapers“. International Journal of Digital Humanities 4, Nr. 1–3 (30. November 2022): 147–70. https://doi.org/10.1007/s42803-022-00055-6.

Bonte, Achim. „Forschungsnahe Bibliotheksdienste. Eine Standortbestimmung“. Bibliothek Forschung Und Praxis 47, Nr. 3 (1. Dezember 2023): 453–60. https://doi.org/10.1515/bfp-2023-0035.

De Weerdt, Hilde. „Creating, Linking, and Analyzing Chinese and Korean Datasets: Digital Text Annotation in MARKUS and COMPARATIVUS“. Journal of Chinese History 4, Nr. 2 (Juli 2020): 519–27. https://doi.org/10.1017/jch.2020.23.

Middell, Matthias, ed. Self-Reflexive Area Studies. Global History and International Studies 5. Leipzig: Leipziger Universitätsverlag, 2013.

Mielke, Katja, and Anna-Katharina Hornidge, eds. Area Studies at the Crossroads. New York: Palgrave Macmillan US, 2017.

Ommert, Eva, und Kathleen Schlütter. „Digital Monolingualism, Archives at Risk, and Global Views on Open Access“. Text/html. Forum Transregionale Studien, 6. März 2024. https://doi.org/10.58079/VYM1.

Stevens, Mitchell L., Cynthia Miller-Idriss, and Seteney Khalid Shami. Seeing the World: How US Universities Make Knowledge in a Global Era. Princeton studies in cultural sociology. Princeton, Oxford: Princeton University Press, 2018.

Wagner, Cosima. „Workshop-Bericht ‚Nicht-lateinische Schriften in multilingualen Umgebungen: Forschungsdaten und Digital Humanities in den Regionalstudien‘“, 2018. https://dhd-blog.org/?p=10669.Viola, Lorella, und Paul Spence, Hrsg. Multilingual Digital Humanities. London: Routledge, 2023. https://doi.org/10.4324/9781003393696.

Wilkinson, Mark D., Michel Dumontier, IJsbrand Jan Aalbersberg, Gabrielle Appleton, Myles Axton, Arie Baak, Niklas Blomberg, u. a. „The FAIR Guiding Principles for Scientific Data Management and Stewardship“. Scientific Data 3, Nr. 1 (15. März 2016): 1–9. https://doi.org/10.1038/sdata.2016.18.

---

---