The importance of semantic descriptions of concepts by means of defining statements is a commonplace tenet of scientific and practical approaches to terminology. While the current understanding of defining statements remains bound to classical concepts of defining, there is limited knowledge about the types of conceptual information that may ease the transfer of knowledge. Furthermore, there is little insight into how defining statements differ epistemologically from non-defining (generic) statements; on the linguistic side, the same can be said about linguistic differences between defining and generic statements. Last but not least, it remains unclear how practical terminology work can benefit from corpus-based research on the description of defining statements. This paper aims to shed light on some of these questions by describing a corpus-linguistic study of knowledge-rich contexts in German and Russian web corpora. Hypotheses about linguistic features of knowledge-rich contexts are derived in a theory-driven manner and researched by means of corpus-linguistic methods. Significant features are then investigated further for the German data, using a multivariate method.