For all data types of a dictionary, as well as for theoretical purposes, corpora - a collection of electronically stored and digitally analyzable texts - provide a fundamentally new foundation. The corpora of the IDS, along with the research system COSMAS, are sufficient for answering scientific questions. Their potential uses are described in relation to the design of a German-Hungarian syntagmatic learner's dictionary. Until now, there has been a lack of empirical methods for obtaining and selecting the necessary data, as well as criteria for the typologization of the acquired word combinations and their systematic and user-friendly presentation for foreign language learners. In developing the dictionary concept, the focus is placed on the dictionary subject and the dictionary format. For determining the dictionary subject, a T.