Ever since the liberation in 1945, the NIOD Institute for War, Holocaust, and Genocide Studies has collected many, often personal documents from the World War II era. A valuable, often underappreciated, part of the NIOD-collection consists of letters from the period shortly before, during and after the German occupation of the Netherlands (circa 1935-1950). The project ‘Oorlog uit Eerste Hand’ (War from First Hand) is concerned with the digitalisation of this collection of war letters.
The project is made possible by funding received from the Mondriaan Fonds as part of the ‘75 years of Liberty’ fund and by the NIOD Institute for War, Holocaust and Genocide studies.
The war letters collection at the NIOD contains letters from among others persecuted Jews, political prisoners, resistance members, displaced persons, refugees, volunteers at the Eastern Front or men in the Arbeitseinsatz (forced labour program of the Nazis). Though their situations differ widely, they all wrote letters far from home to their loved ones, their parents, sometimes to friends, acquaintances or family members on the move elsewhere. Likewise, the worrying people remaining at home wrote letters to their loved ones in faraway places. These letters tell the intense stories of interpersonal contact between people in times of war and occupation. They provide an insight in living in times of uncertainty, violence, occupation, war, oppression, persecution, collaboration and scarcity – a life in which freedom and democracy were often mere abstractions.
‘War from First Hand’ aims to digitise and preserve the NIOD’s collection of war letters, and to make this collection more readily accessible. Not only in the reading room at the NIOD, but also, where permitted by legislation, online through the NIOD and the Netwerk Oorlogsbronnen websites. Initially we digitise the collection by scanning, transcribing, and annotating the letters. ‘War from First Hand’ wants to take it a step further by also creating a structured, academic dataset that can be used in research for the application of for example quantitative text analysis. This ensures the collection is not only better preserved for future generations, but also made available in a way offering new research possibilities.
A digitally accessible letter collection offers new possibilities to explore and investigate questions regarding emotions, experiences, expectations, uncertainty, stress, and resilience. Moreover, it offers insights in the functioning of the occupation apparatus. The possibilities of digitised historical war letters as dataset in a more technical or methodological aspect further enhance this value. The integration of techniques for computational data and text analysis (‘text mining’) in historical scholarship allow for systematic and longitudinal studies of large quantities of historical texts. Such new digital techniques for analysis have resulted in the emergence of new fields of study such as Digital Humanities or Digital History. Creating and making available a machine-readable and structured digital dataset of historical egodocuments can provide an important impulse for both methodological and substantive innovation in the academic field of history – and with that also for research into World War II and the German occupation of the Netherlands.
Thanks to the developments in artificial intelligence it has become feasible to automatically transcribe digitised handwritten historical records. The European READ project has resulted in the development of Transkribus. This is a piece of software which, after a period of training, can transcribe digitised handwritten records, as well as infuse it with the relevant metadata and structure. We refer to this process as Handwritten Text Recognition (HTR). HTR allows handwritten texts to be converted into ‘machine-readable texts’. This is a basic requirement for full text searches for example, but also enables text mining. This requires a computer model to first be trained using a small portion of the letter collection, which has been transcribed manually. These transcripts are made by a group of volunteers. Subsequently an HTR-model is developed for handwritten Dutch texts from the first half of the twentieth century. This model will eventually also be made available for future use by academics and heritage institutions.
‘War from First Hand’ starts on 1 July 2020. The digitising is expected to be completed by the summer of 2023. The project team includes; Annelies van Nispen (information-analyst), Carlijn Keijzer (policy advisor), Milan van Lange (researcher) and Sergio Leatomu (conservation).