BeCoRe

Medieval documents were more than just legal documents: they carried seals with pictures and graphic signs, which an illiterate population could see and understand. Some of them carried even decorations in colour. Many used special scripts. Others were just kept simple. Recent research has pointed out that these graphical features could convey authority: The claim to power as well as legal validity. So far, research has only investigated the phenomenon on selected holdings and especially in the early Middle Ages, simply because only from this period the quantity of documents is small enough that individual researchers could see them all. With the digitisation of archival documents, however, extensive databases came into existence. The allow comparing the large numbers of documents from the 12th-15th centuries. In particular, the world`s largest database of documents, "Monasterium.net", is a good place for such research. The project will compare the design of documents on a wide geographical scale and over a long period. France and Austria are predestined for an international cooperation on this scale, since document research has a very special tradition in both countries (École des Chartes, Paris; Institute for Austrian Historical Research, Vienna). The project therefore examines documents from important French and Austrian monasteries (Cluny, Fontenay, Göttweig, Lambach and Lilienfeld) which represent monastic traditions in close connection with an order (Cistercians in Fontenay and Lilienfeld) and less strict organisational form (Benedictines). The documents from the archives of these monasteries and documents, which were issued by the abbots of these monasteries, will be examined for their graphic properties and these properties will be placed in the context of simultaneous document production. The project will supplement the holdings in Monasterium.net with documents that have not yet been digitised and enrich the data in the database in a way that systematic comparisons can be made. Monasterium.net currently contains over 600,000 documents. Therefore, the project will use automatic methods of image analysis to study the layout and classify the documents according to visual properties. The project can link these semi-automated classifications with information about exhibitors, times and regions and come to a rich picture of medieval documents that takes into account their visual and symbolic functions. Thus, the role of documents as a sign of authority and cohesion can also become visible for a period when the sheer volume of this type of document has so far prevented a clear picture.

The project compared charter production in the context of French (Cluny, Fontenay) and Austrian monasteries (Lambach, Göttweig, Lilienfeld) in the late middle ages. It tested how established diplomatics methods can be enhanced with machine learning based procedures. It added substantial new datasets to the worlds largest database of medieval charters Monasterium.net. Throughout the project, we could create scholarly editions of unpublished documents held in France, (Fontenay Abbey, Cluny Abbey as work in progres) and the Lambach Abbey inn Upper Austria. For this, the project could integrate IIIF services by the Bibiliothèque nationale de France (Gallica) and transcriptions of a former archivist of Lambach Abbey. Part of the editions is enriched with references to authority files or provides new authority data on persons and places mentioned in the charterfs. The editions are now available on the Monasterium.net platform for a general public and for research. Technologically, we developed workflows for converting transcriptions created in the Transkribus environment and in TEI into Monasterium.net enabled TEI. Major contributions to the software behing Monasterium.net improved editing and display of charters in the portal. The project created insights in the selection of computer vision methods for visual similary of charter images. We crated reference data for identification of issuers in charters abstracts through a mix of automatic (combination of rule based and automatic named entity recognition) and manual processes. It made progress in automatic recognition and automated indexing of graphic elements in our corpus of digitized charters. Historically, we could make the concept of "Urkundenlandschaft" (Fichtenau) fruitful and learned how much charter production is defined by regional connections and local habits and less by organisational connections through monastic orders, although the monasteries were anchor points in this local and regional practices.