Beschrijving
The project ‘First-Hand Accounts of War: War letters (1935-1950) from NIOD digitized’ (2020-2023) is concerned with the digitization and so-called ‘datafication’ of historical – mostly handwritten – wartime letters. The original paper documents are part of the archival collection of the NIOD Institute for War, Holocaust, and Genocide Studies in Amsterdam, The Netherlands. The collection known as ‘Collection 247: Correspondence’ contains personal letters from the period before, during, and after World War II and the War of Independence in Indonesia (1945-1950). The letters give a first-hand account of experiences of a wide range of different people: soldiers at the front, Jewish children in hiding, or people in forced labour camps – to name but a few. ‘First-Hand Accounts of War’ preserves and scans the original documents, but also aims to take digitization a step further.All the scanned letters are transcribed to transform them into machine-readable texts (e.g., plain text (.txt) and Extensible Markup Language (.xml)). These transcriptions are partly made by a group of dedicated citizen scientists affiliated to the NIOD, and partly by the project team members. The resulting ‘Ground Truth’ transcriptions are then fed to the computer to train models for Handwritten Text Recognition (HTR) using READ-COOP’s Transkribus software environment. These models are already made available to the public in early 2023. They were also used within the project to digitize the bulk of the collections by automated transcription.
In addition, an enriched machine-readable digital dataset was created and will be made available online (e.g., through the web- sites Oorlogsbronnen.nl (general public), Archieven.nl (historical researchers, professionals), and DANS (digital humanists, researchers, developers, etc.)). This ensures that the collection is not only better preserved for future generations, but also generates new possibilities for digital humanities or digital history research. In addition to the more tangible deliverables, ‘First-Hand Accounts of War’ aimed from the start to bring together experiences, perspectives, and contributions of information specialists, archivists, and (digital) history scholars. The project is as much about the envisioned end products, as about learning from and re- flecting on the collaborative process along the way. This relates to the multidisciplinary project team itself, as to the collaborations with researchers, archivists, students, and colleagues involved in similar projects.
This poster at the DH 2023 showed how multidisciplinary conversations (both in
person and in the blogosphere) and collaborative endeavours in ‘First-Hand Accounts of War’ generated reflexive insights on processes of digitization and the transformations the archival collection underwent. This includes, amongst other things, the choices and decisions made in creating a digital representation of a physical historical archive. The poster invites, for example, to (re)consider how the generation of transcriptions and metadata shapes the knowledge one can retrieve from digitized historical data or the reception of such data. Take, for example, the descriptions in (digitized) archival collections: The concepts, labels, or categories used are not only functional, they also have an epistemological function and meaning. In addition, they ‘(…) [rely] on language that can change the meaning of historical materials; it is created to suit particular audiences and needs (…).’ (Rawson 2018). Terms used in describing sources are often the first and foremost framework of the historical collection with which researchers engage– whether in a ‘paper archive’, or engaging with the sources in a digital context. This is important because archives are (and have been) always about power, as they both reflect and constitute power relations (Cook / Schwartz 2002; Rawson 2018; Zaagsma 2022).
To conclude, the poster displayed how the divergent professional and academic backgrounds of the different team members in ‘First-Hand Accounts of War’ are reflected in the resulting project outcomes and deliverables by briefly addressing also the following topics: 1) legal issues (e.g., related to the EU’s GDPR), 2) the (re)use of Ground Truth data and HTR-models, and 3) the acknowledgement of contributions of partners, project members, and citizen scientists.
Periode | 13 jul. 2023 |
---|---|
Evenementstitel | Digital Humanities 2023: Collaboration as Opportunity |
Evenementstype | Conferentie |
Locatie | Graz, OostenrijkToon op kaart |
Mate van erkenning | Internationaal |