Pseudogapping Puzzles
Howard Lasnik · 1999
Abstract In this chapter, I explore some of the properties of the so-called pseudogapping construction. This construction is in important respects reminiscent of VP ellipsis, except that it leaves behind an element of the VP as a remnant. I begin by summarizing the analysis of Lasnik (1995c), which is based on the important proposal of Jayaseelan (1990) that pseudogapping is simply VP ellipsis, with the remnant having moved out of the VP. I argue that Jayaseelan’s basic proposal is correct, except for the specific movement rule he invokes, heavy NP shift. I argue, instead that the movement rule involved is “object shift,” overt raising to Spec of Agr0• In the course of the presentation, I deal with a number of puzzles that arise, among them certain cases of apparent overgeneration. Finally, I consider the proposal of Bouton (1970) and Lappin (1992) that “antecedent-contained deletion” is derived via pseudogapping, and argue, along with Fiengo and May (1992), that the proposal is only half right, accounting for some ACD instances but not others. I conclude by discussing alternative hypotheses for the latter type of ACD.