The island effect of a dative fragment results from an extraction out of an island when multiple fragment answers (hereafter, MFAs) occur. Kang (2014) suggests that the dative argument is attached to the embedded ditransitive verb, by reporting the re...
The island effect of a dative fragment results from an extraction out of an island when multiple fragment answers (hereafter, MFAs) occur. Kang (2014) suggests that the dative argument is attached to the embedded ditransitive verb, by reporting the results from a sentence completion experiment investigating the attachment preference of the dative argument in the sequence NP-NOM NP-DAT NP-ACC V-REL NP-ACC with the V-REL being a ditransitive verb in Korean. This analysis induces the dative fragment to be extracted from an island when MFAs occur. Although processing cost is not necessarily consistent with acceptability, the more costly processing tends to induce the sentence to be more marginal or unacceptable (cf. Han 2015). In this vein, the less processing is preferred. The alternative way to derive the available dative fragment in the MFAs analysis is to extract the dative argument only from the matrix clause. The dative argument in the sequence of NP-NOM NP-DAT NP-ACC V-REL may occupy either the embedded or matrix clause. At this point, what is at stake is that the processing cost of the ambiguous dative arguments is the same. Under the standard assumption that island effects result from a movement operation, the island sensitivity shown in the MFAs is also involved in the adjunct island.