The pilot study DWA Austria is a cooperative project of the Research Unit Linguistics at the ACDH-CH, the SFB German in Austria. Variation - Contact - Perception and the DSA | Research Center Deutscher Sprachatlas at the Philipps University of Marburg. As part of the pilot study, we are testing the possibilities of AI-supported recording of DWA questionnaires. On the basis of the pilot study, we aim to completely digitize all Austrian DWA questionnaires and carry out extensive analyses of the data.

With a total of about 50,000 survey locations the German Word Atlas / Deutsche Wortatlas (DWA) (Mitzka, Walther & Ludwig Erich Schmitt. 1951 – 1980. Deutscher Wortatlas. Gießen: Schmitz) is the most comprehensive survey of 20th century lexis in the German-speaking world. The survey was conducted indirectly by means of questionnaires between 1939 and 1942.

The questionnaires contained 188 words and 12 complete sentences, which were to be translated into the local dialect by the respective teachers of the school places written to. In addition, the school and birthplace of the teachers and pupils were asked for, which were used to answer the questions.

Completed DWA questionnaires from a total of about 3,700 place points are available for Austria. Previous research on the DWA data had to manually transliterate the questionnaires, which were mostly filled out by hand and in Kurrent script, in order to be able to use them for further scientific research.

With the help of Transkribus, a platform for the full-text digitisation of handwritten sources, selected Austrian DWA sheets are now to be recorded and made usable within the framework of a pilot study. In contrast to conventional text recognition products, Transkribus works with AI support. With the help of a training corpus of DWA questionnaires to be transcribed manually, a model is to be trained that can be used in a next step for the further indexing of Austrian DWA questionnaires.

The pilot study has resulted in the research project LexVAD20, funded by the FWF.