INALIENABLE POSSESSION IN LOCATIONAL CONSTRUCTIONS

In this article, we will argue that Dutch inalienable possession const ructions such as (la) and (Ib) have the same underlying structure in ( 2a): (1a) is derived by means of an obligatory movement of the predica te of the BIJ-phrase into the specifier of the locational PP, as in (2 b); (1b), in its turn, is derived from the structure in (2b) by incorp oration of the functor BIJ into the verb, as a result of which the NP is assigned dative Case. This analysis is consistent with Hoekstra's ( 1994) hypothesis, according to which inalienable possession is syntact ically encoded by means of a functor P (BIJ) that takes the possessor as its internal and the possessum as its external argument: [(SC)[(NP) Possessum][P [(NP) Possessor]]. (1) a. Ik zet het kind bij Jan/hem op de linkerknie, I put the child with Jan/him on the left knee b. Ik ze t Jan/hem her kind op de linkerknie. I put Jan/him the child on the le ft knee 'I put the child on John's left knee.' (2) a. ... V... [(pp) s pec P-loc [(SC) NP1 [BIJ NP2]]] (underlying structure) b. ... V... [(p p) [BIJ NP2](j) P-loc [(SC) NPl t(j)]] (=(1a)) c. ... BIJ(i)+V [(pp) [ t(i) NP2](j) F.. [(pp) t(j) P-loc [(SC) NPl t(j)]]] (=(1b)) Independen t evidence in favor of the proposal in (2) will be given, among which rather complex data involving island effects on movement from out of t he locational PP. Further, the discussion will be complicated by inter ference of apparently similar, but actually quite different constructi ons, which we will discuss as well.