Authors' Libraries as Knowledge Spaces: Digitization of the Catalog of the Arnim Library
- Hosting organisations
- University of Graz
- Responsible persons
- Anke Jaspers and Georg Vogeler
- Start
- End
The catalogue of the library of Ludwig Achim and Bettina von Arnim at Wiepersdorf Castle documents around 4,000 titles from a historical authors’ library. In addition to bibliographic information, it records the original locations of the books in cupboards and on shelves. The catalogue thus makes a historical order of knowledge visible and provides insights into reading practices, the history of the collection and the organisation of knowledge.
The project makes this previously inaccessible material digitally available and prepares it for further research. It aims to transform the historical catalogue into a structured, machine-readable and reusable format, thereby creating new opportunities for research on authors’ libraries, historical library catalogues and knowledge spaces.
About the project
The catalogue of the Arnim Library was compiled by Werner Wilk in 1929 and expanded by Walther Encke in 1934. It comprises 177 leaves with typewritten rectos and later handwritten additions on the versos. The entries are organised by fields of knowledge, genres and publication types and are arranged alphabetically within these categories.
A distinctive feature of the catalogue is its double ordering structure: in addition to its thematic classification, it uses location marks to document the original spatial arrangement of the books at Wiepersdorf Castle. This makes it possible to reconstruct not only which works formed part of the library, but also how they were arranged and related to one another within the historical collection space.
As part of the project, the usage rights for the catalogue were clarified and 377 TIF files were provided by the Herzogin Anna Amalia Bibliothek in Weimar. The files were converted into PNG format and subsequently processed in Transkribus. Initial transcriptions of the typewritten pages were produced, corrected and annotated to create ground-truth data for training a dedicated model. This resulted in the Transkribus model “Print M1 Arnim-Bibliothek”. Further typewritten and handwritten pages were also transcribed in Transkribus. The work was accompanied by project team meetings and cooperation talks concerning the origins and transmission history of the catalogue.
The recorded data are to be converted into a simple machine-readable format containing the following information in particular:
- catalogue entry number
- bibliographic entry
- historical location mark
In addition, a mapping to library description standards such as BIBFRAME is planned. Once the data preparation has been completed, the structured data are to be published on Zenodo together with suitable metadata and an open licence. The project will also examine whether the OCR and HTR ground-truth data produced during the project can be made available.
The code developed for extracting and structuring the data is to be made accessible through a GitHub repository and archived on Zenodo after the completion of the project. In this way, the project supports the FAIR principles and creates a reusable foundation for further research on authors’ libraries, Romantic networks and historical orders of knowledge.