Sentiment analysis between gaining insight and distortion? A case study on personal accounts of the Holocaust and its aftermath

Personal accounts are important sources of information about historical events and provide insight into individual lives and experiences. Accounts by Holocaust survivors are essential for Holocaust research because they document events and perspectives that are not mentioned at all or are distorted in official documents or perpetrator reports. They are therefore a central component of remembrance culture and help to keep events alive for future generations. This is particularly important in view of the dwindling number of eyewitnesses and the much-discussed decline in the willingness to actively engage in remembrance culture. 

The significance of emotion-based and emotion-historical approaches is currently being discussed in historical education and research. Computer linguistic methods such as sentiment analysis are already established in some areas of digital history. Computer-assisted, primarily computer linguistic methods of natural language processing (NLP) promise opportunities for in-depth research into individual and collective experiences, interpretations and memories. Using the diaries of Theresienstadt survivor Martha Glass, the researchers involved in the project are testing and evaluating lexicon-based approaches as well as machine learning methods in order to open up new avenues for historical research. However, with regard to the evaluation of ego documents, the question of the influence of artificial intelligence is increasingly being raised, especially how the use of (generative) AI enables new insights, but can also contribute to distortions in results. Critical reflection on machine analysis methods with regard to the evaluation and interpretation of personal accounts can make an important contribution to digital tool and source criticism in the field of history, as well as to overarching discussions on the use of artificial intelligence in memory culture education.

The research project of the Institute for the History of German Jews and the inter-university Chair of Digital Humanities at the University of Erfurt and the University of Applied Sciences Erfurt is investigating how artificial intelligence (AI) is used in the analysis of personal accounts of the Holocaust and its aftermath, and where the opportunities and risks lie. The project also aims to develop a typology or ontology for emotional forms of expression in historical biographical sources. By subsequently preparing the project results for teaching and communication purposes, the project participants are providing important impetus for digital source criticism and the reflective use of methods and tools in working with personal accounts.

The collaborative project is being carried out by the inter-university Chair of Digital Humanities at the University of Erfurt and the University of Applied Sciences Erfurt (Professor Anna Neovesky with Nina Brolich), as well as the Institute for the History of German Jews in Hamburg (Dr Anna Menny and Helena Geibel).

The NFDI4Memory Incubator Funds are providing €24,500 in funding for the project at the Institute for the History of German Jews.

News post: Sentiment analysis between gaining insight and distortion?