Lexicogrammatical Patterns and Corpus Evidence in Schemann's Dictionary of Idioms

geyken, alexander;

This article presents a strategy for mapping entries in Schemann’s dictionary of German set phrases to lexicogrammatical patterns. With more than 10,000 headwords and 33,000 entries, Schemann’s dictionary is the most comprehensive printed dictionary of German set phrases. The result of the mapping yields more than 8,500 distinct pattern classes, which are used to generate effective corpus queries. A qualitative evaluation of these queries with respect to the 4 billion word corpus of the German DWDS project shows how the entries can be improved with corpus evidence and how the dictionary can be augmented with corpus examples.