This paper attempts to give a new analysis of (preposition) stranding in terms of phase extension which has been originally proposed by den Dikken in his series of recent papers (2006, 2007, and 2008). In doing so, this paper tries to explore a hidden ...
This paper attempts to give a new analysis of (preposition) stranding in terms of phase extension which has been originally proposed by den Dikken in his series of recent papers (2006, 2007, and 2008). In doing so, this paper tries to explore a hidden link between P(repostion) and C(omplementizer) with respect to stranding. In particular, this paper is focused on the asymmetry between leftward movement and rightward movement with respect to stranding. To be more specific, this paper attempts to answer the following questions. Why is P-stranding allowed with leftward movement but not with rightward movement in English Why is C-stranding not allowed in both leftward and rightward movement like Extraposition in English And why are P- and C-stranding not allowed both in rightward and leftward movement in Korean and Japanese
It is widely known that English allows P-stranding in general. In this regard, canonical sentences are out of question in allowing stranded prepositions in elliptical contexts. Discrepancy, however, arises with PPs in adjunct function. In those examples, P-stranding is salvaged by deleting the illicit part. Some English sluicing data in fact show that P-stranding is in fact forced: some examples show that pied-piping is not allowed in English ellipsis contexts. Regarding C-stranding, it looks like the complementizer that is different from prepositions with respect to stranding. That cannot be stranded by the movement of TP to either direction. In principle, the whole that-clause (CP) can move to the front by leftward movement such as passivization and topicalization. However, TP alone cannot move leaving that behind. The fact is that that cannot be stranded by rightward movement like extraposition, either.
P-stranding is disallowed in Korean at all. Ungrammatical strings, however, can be repaired by ellipsis if their ungrammaticality is due to stranded Ps. In ellipsis contexts, if stranded Ps are deleted along with the containing TP, the whole derivation (or representation) is repaired and becomes a licit one. Canonically, C-stranding is disallowed, whether the stranding is due to leftward movement or rightward movement. Even if stranded Cs are deleted, however, a deviant representation (or derivation) is not repaired.
For these findings, one and the same answer turns out to be possible on the basis of phase extension under the assumption that D to P incorporation or T to C raising occurs inside PP and CP, respectively. If D incorporates with P, then PP becomes a phase, the extraction of DP out of which would violate the PIC. In the same way, if T raises to C, then CP becomes a phase, then the extraction out TP over CP boundary would also result in the violation of the PIC. The present paper also deals with the reason why some instances of stranding are repaired but others cannot. In addition, a special requirement of the second wh in multiple sluicing in English that it must be a PP is explained in the same way as a result of D to P incorporation and subsequent phase extension to PP.