Methods of applied linguistics today are largely determined by text processing tools available not only to researchers, but also to a wide range of users. The article demonstrates that text corpora and corpus tools may be well applied to both research and teaching of the structure and composition of such complicated units as multi-component noun phrases (NPs). Based on original text corpora, the research focuses on syntactic and structural characteristics and transformations of multi-component NPs, namely noun+noun constructions, within a text. The article also proves their role as markers of text complexity and scientific style. As the complexity of NPs requires special skills facilitating both reading and writing scientific texts, the article suggests a procedure of treating NPs in special text corpora, compiled by the learner or novice author of their own discipline texts, that helps to identify, “unpack” and compose NPs. The method proposed is based on some published practice and the authors’ personal experience in teaching students to deal with NPs in scientific and technical texts.