This study examined subject-auxiliary inversion errors in wh-questions produced by 88 Korean EFL elementary learners, focusing on whether types of wh-words and auxiliaries could affect inversion acquisition and whether explicit instruction on movement rules could facilitate this process. Guided writing tasks were used as a pretest and a posttest to analyze influence of wh-words and auxiliaries on learners’ inversion in the pretest and effects of instruction on movement rules in the posttest. Results showed that both whwords and auxiliaries significantly influenced learners’ inversion acquisition. Learners struggled more with why-questions than with what-questions, which were selected as representatives of adjunct and argument wh-questions, respectively. More inversion errors occurred in wh-questions requiring do-support than in those involving auxiliary be or modal will, although no significant difference was found between be and will. Experimental lessons with brief explicit instruction on auxiliary movement during regular classes significantly improved learners’ inversion accuracy, particularly in dosupport questions, which posed the greatest challenge in the pretest.

In conversation analysis (CA), wh‐questions are treated as invoking a claim that questioners have no knowledge about the information being solicited. This paper examines a particular form of wh‐questions that indexes an epistemic claim incongruent with such a claim of no knowledge. In particular, it examines wh‐questions that are marked with a committal suffix in Korean conversation, in which the committal suffix indicates that the speaker should have the information at hand. Using the method of CA, this paper shows that these wh‐questions indexing incongruent knowledge claims are used in contexts in which questioners know or should know about the information being solicited. First, they are commonly used to seek a particular piece of information questioners already know about, or should already know about, by reference to prior talk or shared knowledge with recipients. Second, wh‐questions marked with committal endings can be used as word searches. In these cases, they do not seek the other partyʹs active participation in finding solutions to the missing word(s) and thus are self‐directed. The analysis will suggest that wh‐questions with committal endings can serve to avoid a potential trouble or accountability in interaction.