A Note on the Controversy over the Nature of Wh-island Effects of Wh-in-situ Questions in Korean

Wh-in-situ question in Korean and Japanese where the wh-phrase takes a non-local scope are generally assumed to be degraded, and this has often been explained in terms of a syntactic constraint such as the Subjacency or the MLC. Disagreements, however, exist. There has been a claim that such sentences are not degraded as long as they are read with the correct prosody and thus that the nature of the apparent wh-island effects in these languages is mainly prosodic (Deguchi and Kitagawa 2002, etc.). Another claim is that they exist but they are mainly processing in nature (Yoon 2009). In this paper I discuss what the result of an experimental study of Korean native speakers' grammaticality judgments on wh-in-situ questions with a non-local wh-Q association suggests to this controversy. In the test where such sentences in Korean were presented in the spoken as well as the written forms, it was found that the majority of the subjects judged the sentences to be highly acceptable but that almost two thirds of them misinterpreted them as Y/N-questions. I show that this result can be best explained in terms of processing.