Using AI and Corpus-Based Tools for Teaching Terminological Collocation in the Field of Aerospace Engineering

This study explores an integrated methodology for teaching terminological collocations in English for Specific Purposes (ESP) in the domain of aerospace engineering. A 1-million-word Aerospace Corpus, compiled from the MDPI Aerospace journal, was examined using Sketch Engine to identify high-value adjective + orbit collocations (e.g., Low Earth orbit, circular orbit, prograde release orbit, off-nominal orbit). Collocational strength was quantified using $\Delta P$ to distinguish conceptually bound terms (e.g., parking orbit, halo orbit) from operationally contrastive descriptions (e.g., circular orbit, elliptical orbit). These data were then used to select and prioritise targets for instruction, supporting claims that terminological collocations are central to conceptual understanding in specialised discourse and should be taught as units rather than as isolated words. Instructional tasks were sequenced according to Bloom's revised taxonomy so that learners first identify and interpret collocations, then apply them in controlled production, and ultimately justify orbital design choices and produce mission-style discourse using these terms. AI (ChatGPT) was incorporated as a complementary tool to generate draft exercises (e.g., verbcollocation gap-fill items, definition matching tasks, mission briefing prompts) and B2-level explanations of technical orbit types, while instructors curated output for technical accuracy, register, and cognitive level. The result is a replicable workflow in which (i) the corpus determines what to teach, (ii) Bloom's taxonomy structures how to teach it, and (iii) AI expands how much material can be produced efficiently without replacing expert judgment.