How do Korean speakers receive internally headed relative clauses?

This study examines how Korean internally headed relative clauses (IHRCs) are received by native speakers. Although previous studies have greatly enhanced our understanding of IHRCs by proposing a theoretical framework with which to explain their constraints, no attempt has been made to provide an empirical study of how Korean speakers understand IHRCs with support from appropriate statistical methods. Accordingly, the current study conducted a Likert-scale task experiment and a question-and-answer survey whose results were thoroughly analyzed with multiple statistical tools. The findings are as follows. First, certain examples that have specific grammatical and aspectual features are judged as having higher acceptability. Based on this observation, we propose the possibility that the Korean IHRC is an example of a verb-class-specific construction in which innovative use spreads through semantic similarity. Second, Korean IHRCs are received in three different ways by native speakers: One group of speakers accepts the IHRCs with entity construal, another group with event construal, and a third group of speakers does not accept the IHRCs at all. It is argued that these split interpretations can be understood by observing constructions that are structurally similar to IHRCs. This study contributes to relevant areas by providing an empirical study of Korean IHRCs.