With this poster, I outline an ongoing research project that uses historical sermons to study language change in German. While a long-term single-genre corpus like this has restrictions with regard to its representativeness of language use at a given time, it also has unique advantages over short-term multi-genre corpora. Because sermons are relatively uniform as a genre over long periods of time while at the same time being fairly close to oral speech, they are ideal for studying long-term developments of language in an oral setting. Besides outlining the design and structure of the corpus, I exemplify the benefits of this approach by reporting from a case study that shows how sermons can further our understanding of how German grammar changed over time.