This paper is concerned with split questions in Korean where the wh-question part and the tag are present within a sentence, such as John-i mwues-ul mek-ess-ni, sakwa(-lul)? 'What did John eat, an apple?' Recently, Arregi (2010) has made a proposal of how the tag is derived; the tag undergoes focus-movement to SpecCP, and PF-deletion takes place on the non-tag construction. Although his analysis seems to fit in the Spanish and English data, we argue that it cannot extend to Korean split questions. We have found evidence that the PF-deletion account cannot hold, in that it does not account for a wide range of the data found in Korean. In particular, we have noticed that the embedded verb complex may be overtly realized along with the tag in the tag-clause. This is unexpected under the PF-deletion account since everything but the tag is subject to be deleted. We propose that an LF-copying analysis straightforwardly explains those problematic cases for the PF-deletion account, and that it is a better analysis to account for split questions in Korean.
1. Split Questions
2. Biclausal vs. Monoclausal
3. PF-deletion in Korean SQs Revisited
4. LF-Copying in SQs
5. Conclusion
(0)
(0)