6 September 2022, via Zoom
__________________________
Standards to be considered in a Virtual Research Environment for Digital Editions of Charters
Standardisation work is necessary for Digital Diplomatics focused on structuring texts of medieval charters and modelling the metadata (e.g. in the Charters Encoding Initiative or the SKOS version of the Vocabulaire Internationale de la Diplomatique). In the ERC project From Digital to Distant Diplomatics we attempt to integrate new technologies to work with medieval charters and build them into a modernised version of Monasterium.net, Europe’s largest charter portal. Automatic transcription, visual analysis of layout, graphical elements, and objects on the charter image, named entity extraction are only the more obvious methods that could bring diplomatics forward under the conditions of the emerging large-scale data sets.
The presentation will introduce the architecture and pipelines of the virtual research environment (VRE) of the project. It will discuss existing standards, such as the TEI or iiif and call for further work (e.g. on the digital representation of the Vocabulaire Internationale de la Diplomatique). It will also try to present some recommendations on how to design a digital editing project at the current state of research, so that it can be re-used with the imagined VRE, profit from becoming part of a European scale research facility, and allow for the interchange with new tools developed in the project and the corpus made accessible in the VRE.
________________
Charter encoding in the TEI P5 with the CEI_TEI ODD
The Charters Encoding Initiative (https://www.cei.lmu.de/) is a fork of the TEI P4 describing a series of elements for the specialist description of charters metadata. Primarily used by the Monasterium.net charters archive, it fell out of step with the TEI as time progressed, resulting in the need for a new version, implemented as a TEI ODD extension, which was created in 2019 as part of a project funded by the Austrian Science Fund (FWF). This talk will describe how both observed use and theoretical concerns came together in building the current CEI_TEI extension and will describe its intended use cases, including how it can be used to increase the searchability and reusability of charters (meta-)data. The paper will additionally explore some of the ways that the diplomatics concepts might be extended and generalized to apply to other genres of textual materials covered by the TEI.
________________
Some limits and possibilities in the automatic transcription of charters
Recent advances in Machine Learning are already showing benefits for Medieval Studies, including Handwritten Text Recognition (HTR) or the automatic transcription of manuscript documents. However, applying HTR to single-sheet charters remains challenging due to the large variety of scribal hands and practices, layouts, abbreviations, sometimes mixing Latin and vernacular languages, and so on. One way to ameliorate this problem is to use fully open software and to share models and Ground Truth corpora for training, as is already being done for manuscript books, but this raises further questions such as what standards of transcription should be used. Furthermore, HTR gives a document-oriented transcription, and moving from this to a text-oriented model such as CEI is never trivial and cannot easily be automated. This presentation will therefore discuss limits and possibilities in the context of existing initiatives such as eScriptorium and HTR-United, and what this may mean for charter studies in the near future.
________________
Challenges of diplomatic analysis to digital editions
Electronic support offers enormous facilities for disseminating images and texts, and in recent years it has been widely used in numerous projects based on historical documentation. However, the possibilities of computational methods for analysis are probably still underdeveloped: platforms that give access to digital representations of charters offer limited searching functionalities, and frequently, users trained as historians do not have the technical capacity to exploit the possibilities of the computer. The purpose of this paper is to identify some critical aspects of the diplomatic analysis of medieval charters and to assess the requirements of a digital edition to respond to these needs, understanding that the definition of existing and future edition models depends essentially on the needs of research, to which they must offer appropriate technical solutions.
________________
Corpus analysis of historical documents: A study of spelling standardisation using the MELD corpus
The MELD corpus (Stenroos et al, 2017) provides detailed transcriptions of around 2000 English legal documents from 1406 to 1525, a period when spelling was beginning to be standardised. In contrast with many such corpora, original spelling is retained.
This paper examines the variation in spelling of the word said. Previous work has examined the move to the use of digraphs in the spelling of English vowels, but the MELD corpus permits detailed examination of the variant digraphs used over space and time. The results of the analysis show that the two parts of the digraph, <a> vs <e> and <i> vs <y> emerge at different rates, <i> becoming fixed earlier than <a>, and that Cambridgeshire and Middlesex are using the standardised <ai> form earlier than Suffolk and Norfolk. I will explain the process used and compare to other corpus approaches with similar documents.
Reference
Stenroos, Merja, Kjetil V. Thengs and Geir Bergstrøm (2017), A Corpus of Middle English Local Documents (MELD), version 2017.1. University of Stavanger. http://www.uis.no/meld (Accessed: 31 December 2020)
________________
Digitisation problems and copyright issues in the example of a large-scale Hungarian charter edition project
The charter edition series Angevin Archives (Anjou-kori oklevéltár) is one of Hungary's most significant historical source edition projects. The serial aims to edit every Hungarian-related medieval charter issued during the rule of the Angevin dynasty (1301-1387). The project started in 1990 with the first volume, and it has been continuously expanding with new volumes.
Similarly to other recent charter edition projects in Hungary, the editorial board decided to publish extracts in Hungarian instead of a full-text edition (Tibor, 2006). The series started before the digital world and has seen several waves of digitisation over the years. In parallel with the digitisation of archival material (György, 2010) – such as medieval charters, inventories and other documents – there were many attempts to publish the volumes of the Angevin Archives on CD, DVD and recently online (Rábai, 2017). We plan to summarise these steps and present the current situation while emphasising the advantages and disadvantages of the freely available pdf version. In connection with this, we aim to give an insight into our efforts to use latex-based processing and build a database with all the material included in the indices of the serial (places, people, objects). Furthermore, we intend to share our experiences concerning full-text edition with the help of speech recognition software (Mihajlik et al, 2017). Concluding our paper, we must mention financing problems and the uncertainty concerning individual and institutional copyright issues.
References
Almási, Tibor: The documentary heritage of Angevin Hungary comes to light. Kwartalnik historyczny 113: 4 (2006) 47-57.
Rácz, György: Collectio Diplomatica Hungarica. Medieval Hungary online. Archiv für Diplomatik 56 (2010) 423-444.
Rábai, Krisztina: Charters of the Angevin Period. Chronica 13 (2017) 133-140.
Anjou okmánytár. Hungaricana. <https://library.hungaricana.hu/hu/collection/kozepkori_magyar_okmanytarak_anjou>
Mihajlik, Péter; Szabó, Lili; Tarján, Balázs; Balog, András; Rábai, Krisztina. "First Results in Developing a Medieval Latin Language Charter Dictation System for the East-Central Europe Region". In: Lacerda, Francisco; House, David; Heldner, Mattias; Gustafson Joakim. Proc. Interspeech 2017. Stockholm, International Speech Communication Association (ISCA) , 2058-2062.