Historian Prof. Torsten Hiltmann aims to make use of machine learning for medieval research
Centuries-old manuscripts, documents and heraldic images: at first glance, medieval research and artificial intelligence seem to be a contradiction in terms. After all, historical studies and the like were long seen as being subjects greatly removed from the world of IT. However, methods such as machine learning on the part of computer programmes, which learn new things and correct themselves, open up new opportunities for historians doing research. "By using artificial intelligence, we can make things visible which have so far been hidden", explains Dr. Torsten Hiltmann, who has been Professor of Digital Humanities in Historical and Cultural Studies at the Department of History at Münster University since March 2019. "There are lots of documents from the Middle Ages on which, due to their sheer quantity, hardly any research has been done. The reason is that special skills are needed to read them. Although digitalised images of these texts exist in many cases, their content has barely been explored", he adds.
The aim is to change this state of affairs by using artificial intelligence in order to enable full-text and keyword-based searches to be carried out on those inventories which have already been digitalised but not yet studied. Currently, in addition to textual analysis, a range of new digital methods is being developed – and these are being used not only in History, but also in Archaeology, Islamic Studies, Arabic Studies, Protestant and Catholic Theology and Literary Studies. All the activities being carried out in the Digital Humanities are bundled and monitored at the Center for Digital Humanities (CDH) at Münster University. Since the CDH’s inception in 2017, Torsten Hiltmann has been actively involved as a member of the Executive Committee of its Competence Centre. He is also the head of the Digital Humanities Service Center. In his dual role as professor and head of the Service Center, one of Hiltmann’s aims for the future is to offer seminars in which students and staff are taught methods used in the digital humanities, but which also enable participants to take a critical look at such methods.
"The aim of the CDH is to bring together research and IT infrastructure. For us, it’s important to show what’s already possible today", says Hiltmann. Any expertise currently not available at Münster University is brought in by the researchers, not least in the form of visiting guests – for example, Dr. Dominique Stutzmann from the Institut de recherche et d'histoire des textes (IRTHT) in Paris, who is presenting the "Himanis" project in the research colloquium entitled "400 to 1500. Middle Ages". Around 80,000 pages of documents in manuscript form relating to the French kings of the 14th and 15th centuries have been made accessible by means of an algorithm developed and used by those involved in the project. Now, for the first time, historians can search through this huge collection online and, for example, systematically research the background to the rise of nation states in medieval Europe.
"It means that, in future, recognizing heraldic images on illustrations and in manuscripts can produce entirely new research findings", adds Torsten Hiltmann, who is currently investigating how heraldic communication functioned in the High and the Late Middle Ages. With his project entitled "The Performance of Heraldic Images" – funded by the Volkswagen Foundation – his aim in the next four years is not only to devise special web technologies enabling researchers to study such images, but also to contribute to a digital reorientation of the auxiliary sciences of history.