In this paper, I discuss a debate regarding the dative alternation in English. The debate consists of two opposing approaches to the dative alternation. On the one hand, alternative projection approaches claim that the double object variant and the prepositional dative variant of the dative alternation have distinct underlying structures; i.e. one is not derived from the other. On the other hand, transform approaches claim that the double object variant is syntactically derived from the prepositional dative variant. The main empirical focus is on (i) certain so-called non-alternating double object variants—whose existence is taken as support for the alternative projection approaches—that can in fact surface in the prepositional dative variant when the indirect object is heavy enough, and (ii) a certain class of verbs whose indirect object is always interpreted as a possessor/affected object regardless of whether the verb surfaces in the double object variant or the prepositional dative variant. These facts pose a serious problem for alternative projection approaches. I show how an analysis where the indirect object moves from the complement of a preposition to the specifier of an applicative phrase to derive the double object variant can handle these facts.