Recently, a school of thought claiming that ellipsis sites are determined by phasehood has emerged. More specifically, Gengel (2007, 2008) proposes that ellipsis sites correspond to Spell-out domains (i.e., complements of phase heads). However, Bošković (2014) and Harwood (2015) argue that not only Spell-out domains but also entire phases can be elided, by accounting for peculiar properties of VP ellipsis in multiple auxiliary verb constructions, i.e., sentences with the middle/aspectual field. In this paper, I claim that the connection between ellipsis sites and phasehood is not that tight. I will argue this by showing that Bošković (2014) and Harwood (2015) might be premature, in that they did not take into account the interplay between head movement and phasehood, which changes the phasehood of certain phrases, and thus, their conclusion that ellipsis is constrained by phase might be inconclusive. Furthermore, I argue that English gapping in DP serves as a counterargument to the idea that ellipsis is constrained by phasehood.
1. Introduction
2. Bošković (2014) and Chomsky’s (2015) Labeling Theory
3. Harwood (2015) and den Dikken’s (2006, 2007) Phase Extension
4. “Gapping” in DP as a Counterargument to Bošković’s Ellipsis Generalization
5. Concluding Remarks
References