The two academic disciplines linguistics and literary studies are often part of one common study program, but they differ in many respects: Their object of study, the methods they use, the type of knowledge they aim to generate, and also the presentation of their work in academic writing. I explore these differences by examining a corpus of German PhD theses from the two disciplines. The focus of this talk will be twofold: First, I will discuss how we can identify differences between corpora in a data-driven way, i. e. with only few theoretical assumptions. While many data-driven approaches rely on surface-based frequencies of words and sequences of words, I argue for the additional use of syntactic annotations for this purpose. Second, I will present and contextualize the differences between academic texts in linguistics and literary studies that can be detected in this way. I conclude by reflecting more generally on how the results of a data-driven analysis can be integrated into existing theories.