Digital Humanities in the Max Weber Foundation

【概要および講演者】

Harald Kümmerle

Datafication as observed in speeches in the Japanese Diet – an analysis using topic modelling

Topic modelling is a method for mining large text corpora that has become popular over the last decade. With topic models, it is possible to carry out a “distant reading” of a corpus and identify texts that are most suitable for a classical “close reading.” However, epistemological issues still limit the potential for operationalization.

This talk shows how to make use of topic models in a research project on the discourse regarding data in Japan. The corpus consists of relevant speeches in the Japanese Diet, a source that is highly rich in structure and metadata. Harmonizing conceptual history with a praxeological perspective, several aspects of the process of datafication since 2011 are traced through key texts. This use case brings out the strengh of topic models as a mixed-membership model for text analysis.

Harald Kümmerle is senior research fellow at the German Institute for Japanese Studies (DIJ) in Tokyo. He studied mathematics at the Technical University of Munich (TUM), Japanese Studies at the Martin Luther University Halle-Wittenberg (MLU), and Japanese as a Foreign Language at Keio University in Tokyo. His doctoral thesis (Japanese Studies; defended in 2019) concerned the institutionalization of mathematics as a science in Meiji- and Taishō-era Japan. During the time as a doctoral researcher, he was junior visiting research fellow at Keio University, visiting research fellow at the Centre for Science Studies at the German National Academy of Sciences Leopoldina and research fellow at the MLU Halle-Wittenberg. His fields of interest include the history of mathematics, digital humanities, new materialism, and critical data studies. Currently, he also is an associated researcher at the Graduate University for Advanced Studies, SOKENDAI.

Esther Meier and Sebastian Kindler

Soviet and German Prisoners of War and Internees. A German-Russian Documentation and Digitisation Project

During the Second World War, up to 5.7 million Soviet soldiers and officers were taken prisoner of war by the Wehrmacht. More than three million died from the conditions of captivity or were murdered. The German treatment of Soviet prisoners of war is one of the greatest war crimes of the Second World War.

On 22 June 2016, the project “Soviet and German Prisoners of War and Internees” was initiated by a joint declaration of the Foreign Ministers of the Russian Federation and Germany, Sergei Lavrov and Frank-Walter Steinmeier. Within this project, the GHIM is conducting the archival research on Soviet prisoners of war. The names and biographies of many Soviet prisoners of war are still unknown today. The GHIM searches and digitises personal documents on Soviet prisoners of war in Russian, German and international archives. The digitised material is indexed by a team at the GHIM and by external providers. The data and the digitised documents are included in the databases “Memorial Archives” in Germany and “OBD-Memorial” in Russia. Bringing together sources of different origins in databases opens up new possibilities for scientific research and clarifies numerous fates.

Esther Meier is the scientific coordinator of the project “Soviet and German Prisoners of War and Internees” at the German Historical Institute Moscow. She manages the archival research on Soviet prisoners of war in Russian, German and international archives. Esther Meier published on World War II, the Soviet War in Afghanistan and Soviet urbanisation. She worked at the University of the Federal Armed Forces in Hamburg and at the University of Hamburg.

Sebastian Kindler is research fellow in the project “Soviet and German Prisoners of War and Internees” at the German Historical Institute Moscow. He manages the collection, the maintenance and the indexing of the data and the digitised material. He worked on the database “Soviet War Gravesites in Germany”, which is run by the German-Russian Museum Berlin-Karlshorst. Sebastian Kindler is doing his doctorate on German propaganda photography in the Second World War.

Jörg Hörnschemeyer

Digital Humanities at the German Historical Institute in Rome (DHI Rome)

Digital Humanities have played an important role in the historical and musicological research at the DHI Rome for the last two decades. The DHI Rome provides a broad spectrum of research data on historical periods from the early Middle Ages to Contemporary History. The presentation aims to give some insights into different Digital Humanities projects.

Jörg Hörnschemeyer studied HKI (Computer Science for the Humanities), German Philology and History at the University of Cologne. In 2005, he finished his studies with a thesis on designing and implementing an SVG-based WebGIS application. From 2003 he worked as a software developer for several research projects including: the image database project Prometheus (prometheus-bildarchiv.de) in 2003; the Arachne project (https://arachne.uni-koeln.de/drupal/) at the Archeological Institute of Cologne between 2003 and 2005; the Critical online edition of the reports written by the nuncio Eugenio Pacelli 1917–1929 in 2008. Since 2010 he has worked as research associate at the German Historical Institute in Rome. He received his PhD in 2017 with a thesis in Genetic Editions.

Mareike König

Academic blogging in the humanities: the platform de.hypotheses

Hypotheses is a non-commercial blogging platform for the humanities and social sciences. The platform provides a free service that facilitates setting up academic blogs, brings them together under one roof and ensures greater visibility as well as content archiving. Founded in 2009, Hypotheses now hosts several thousand blogs, produced by a vast community of bloggers from countries all over the world. The platform receives over one million visits each month. This presentation will focus on the German branch of Hypotheses, provided by the Max Weber Foundation in cooperation with OpenEdition. It will deal with academic blogging in the humanities as a way of communicating research results with peers as well as larger audiences. We will present the platform and its services and then discuss writing practices and motivations based on the results of a large survey among German academic bloggers in 2019.

Mareike König is deputy director and head of the Digital Humanities department and the library of the German Historical Institute in Paris. She is also project manager of de.hypotheses.org, the German speaking blogging platform for the humanities and social sciences. She received her PhD in history from the university of Rostock. Among her research interests are Digital History, science communication in social media and French-German Relations in the 19th century. Twitter handle: @Mareike2405

Anne Klammt

Re-Purposing research data: Case study ‘Deutsch-Französische Kunstvermittlung 1871–1940 und 1945–1960’

The reuse of research data is an urgent issue not only for the Digital Humanities, but for the humanities as a whole. Databases, digital editions, and, increasingly, collections of born digital material form an essential part of the intellectual output of research. Yet the prospects for future use are precarious. While there is growing awareness of standards, procedures, and licensing in the area of FAIR data initiatives at the national and international level, the reuse of research data in the humanities, with the exception of linguistics, is rarely addressed in depth. In our view, re-purposing means reframing data from the perspective of future use. A use that may not follow the research interests of the data producers. More to the point, neither future users and producers will share the same insights and experiences, nor will they necessarily view the data in the same way. However, it is likely that future researchers will refer to various other data collections and even mix the data. The project “Data curation using the example of the data collection Deutsch-Französische Kunstvermittlung 1871–1940 und 1945–1960” primarily addresses the latter scenario, seeking ways to share the maximum amount of information about the scope and origins of individual items within a data collection through human- and machine-readable interfaces (GUI and API). A first attempt is to re-conceptualize the previous database with reference to the event-based CIDOC CRM (CIDOC Conceptual Reference Model). By associating an entry in the database with the type “information object” (E73), the events that led to the current state of the record can also be modeled. In the case of the data entries in the data collection, these events are the creation of the datasets in the 2000s and the semantic enrichment with controlled vocabularies in 2021. For the implementation of a prototype, the project used the Linked Art Data model. It is anchored in the domain of museum and heritage professionals and therefore not too far from the domain of art history. Accordingly, we hope to initiate real-world (re)use cases. The presentation will discuss the conceptual design and initial interim results.

Anne Klammt is an archaeologist by formation. After experiences in fieldwork in Austria, Egypt, Germany and Ukraine she specialised in landscape archaeology and early medieval Europe. Her research led her to the application of computational methods, which she taught for several years in transdisciplinary courses at the Universities of Göttingen and Regensburg. From 2015 to 2019, she was the managing director of the Mainz Centre for Digitality in the Humanities and Cultural Studies (mainzed) before taking up a position as a scientific coordinator of CLARIAH-DE. Since 2020, she has been Director of Research and Head of Digital Humanities at the DFK Paris. Her ongoing projects include the founding of the journal Construction Kit: a review journal for research tools and data services in the humanities (CKIT).

Jana Keck

Human in the Loop: Semi-Automated Genre Classification for Historical Newspapers

Human-in-the-loop (HITL) is a branch of artificial intelligence that leverages both human and machine intelligence to create machine learning models. How does such a cooperation between human and machine look like in computational periodical studies? This presentation will give insight into the workflow of these cooperative tasks to develop a genre classifier for digitized historical German-American newspaper texts by using unsupervised and supervised machine learning algorithms. This approach of semi-automatically classifying nineteenth-century newspaper texts into different genres represents an iterative human-in-the-loop process that allows both sides to interact continuously and it enables different modes of reading of both numeric and textual data. The objective of this presentation is to open up the black box of digital scholarship and to show the potential of machine learning algorithms to study literary conventions of fiction and non-fiction in historical newspapers.

Jana Keck is research fellow in Digital History at the German Historical Institute Washington (GHI), where she is involved in research activities that bring together hidden or neglected sources and Citizen Science to study the history of transatlantic migration. Before joining the GHI in 2020, she was working at the University of Stuttgart in the DFG-funded research project “Oceanic Exchanges: Tracing Global Information Networks in Historical Newspaper Repositories, 1840-1914” (DFG), which boosted a team of researchers in computational periodical studies from Europe and the Americas to investigate the transnational news circulation in nineteenth-century newspapers. In her PhD project, for which she was recently awarded the Peter Haber Price for Digital History as part of the 53rd Deutscher Historikertag, she examines reprinting practices and genre conventions in German-American newspapers to study popular newspaper literature and the migrant networks they created in the nineteenth century.

参加をご希望の方は、上記の「ウェブサイト」をクリックし、ドイツ日本研究所(DIJ）のウェブサイトから直接参加登録を行ってください。