Nouns used to pick out ontologically dependent entities such as holes and flaws, unlike those picking out “ordinary” entities, such as coats and tables, cannot felicitously stand as indefinite subjects of a locative copular sentence (#A hole is in the bucket), but appear freely in there-sentences (There is a hole in the bucket). This contrast is further evidence in favour of the idea that the two sentence types have different underlying predication structures (cf. Barwise and Cooper, 1981; Francez, 2007; Hazout, 2004; McNally 1998a; Williams, 1984, 1994), but also that the preposition in, occurring in both constructions, is ambiguous between a locative and relational meaning. That locative in is distinct from the in that relates a dependent entity to its host is confirmed by inferences between and among sentences containing these two forms. Locative in, whose meaning is roughly that of enclosure or containment, is licensed as a predicate in a locative copular sentence; this sentence type is used to state the location of an entity. However, because a dependent entity’s location is entirely contingent on its host, it cannot be felicitously introduced in this way. By contrast, it is possible to introduce a dependent entity by stating that its host has the dependent entity in it, which is what a there-sentence does. Following Hornstein, Uriagereka and Rosen (1994), the underlying representation of a there-sentence which realizes this relation does not contain a preposition; rather, relational in is derived via incorporation into the copula, like have (Kayne, 1993).