Linguistic research and foreign language teaching have been drifting apart from each other. At a time of declining social cohesion in polyglot Europe, scientifically sound and effective foreign language teaching is vital for mutual understanding, in particular, teaching of German, which will likely gain in importance as European lingua franca. CLOSER attempts to bring foreign language research and teaching closer together. Adopting quantitative research methods from usage-based cognitive linguistics, I investigate the acquisition of German two-way prepositions (2WYP), which poses significant challenges to foreign language learners. I use regression modeling and large data samples to pin down the factors which determine native choices and nonnative errors in authentic language use. Based on the findings, I develop a semi-artificial grammar learning experiment which provides insight into the interplay of input-driven implicit practice and metalinguistic explicit knowledge, which is crucial and unique to foreign language learning but widely unexplored in usage-based research. CLOSER is unique and original in supplying quantitative evidence to the ongoing linguistic debate about 2WYP. In unprecedented ways, CLOSER looks at the intricate interplay of forces at the implicit-explicit learning interface to bridge the gap to instructed foreign language teaching. Needless to say, however, research findings do not straightforwardly extend to instructed learning in classroom settings. Therefore, I develop online applications which generate teaching materials for optimized construction learning based on prototypical usage contexts and effective metalinguistic instructions. The CLOSER-inspired materials are then field-tested in a classroom study. As MSCA fellow at UiT under the supervision of Laura Janda, I will be able receive advanced training in quantitative methods and programming to successfully deliver CLOSER and build up networks and skills for its future exploitation.