Ellipsis. Guess where, when and how

SAAB, ANDRÉS

Workshop; Sluicing+@50; 2019

This talk focuses on ellipsis at the syntax-morphology interface. Mainstream, structural (i.e., non-atomistic) approaches to ellipsis assume that as far as the formal mechanism of ellipsis is concerned, the difference between phonological deletion and non-pronunciation cannot be empirically distinguished. This chapter argues that this is incorrect and that there is vast evidence that both mechanisms are not extensionally equivalent. The crucial evidence comes from two sources: (i) the (in)sensitivity of phonological and (late) morphological features to the identity condition on ellipsis and (ii) the timing of ellipsis. As for the first point, all the evidence suggests that phonological and late morphological features are fully insensitive to the identity condition on ellipsis, favoring thus a nonpronunciation approach over a PF-deletion one. As for the second source evidence, which I take as decisive, there many ways to show that phrasal ellipsis is exclusively resolved at narrow syntax, before any morphological operation and, consequently, before lexical insertion.