We integrate insights from research in Second Language Acquisition (SLA) and Computational Linguistics (CL) to generate text-based questions. We discuss the generation of wh- questions as functionally-driven input enhancement facilitating the acquisition of particle verbs and report the results of two crowdsourcing studies. The first study shows that automatically generated questions are comparable to human-written ones. The second study investigates different types of questions, their perceived quality, and the responses they elicit. [For the complete volume, see ED578177.]