This dissertation deals with infinitival constructions in standard Dutch. There are three types: some infinitivals are obligatorily transparent, others are (apparently) optionally transparent, and still others are opaque. An in depth study of the fine structure of the obligatorily and optionally transparent complements shows that the possibility of transparency does not correlate with the presence of functional heads. In the class of obligatorily transparent complements, we find complements as small as VP, but also complements as large as TP. As both obligatorily transparent VPs and TPs show the same range of transparency phenomena, the possibility of transparency phenomena does not depend on the presence of particular functional heads. The same holds for the optionally transparent complements: within this class, the fine structure of the complement may range from vP to TP, but regardless of the fine structure, the same transparency effects arise. Transparency is instead ascribed to the absence of phase heads between the complement and the matrix clause. Phase heads being locality boundaries, a relation between the matrix and the complement cannot cross a phase head. This explains the observation that opaque infinitivals are CPs. The status of a head as a phase is argued to be variable. It is proposed that standard Dutch v is a phase head only if it checks accusative case. With this assumption, the observation that the various transparency phenomena correlate with the possibility of long object raising follows straightforwardly.