From syntax to postsyntax and back again

Martina Martinovic

University of Florida

Monday, February 11, 2019
10:30 a.m.–Noon

Humanities Center Conference Room D, Rush Rhees Library

Abstract

A fairly widely adopted view of the syntax-postsyntax(PF) interface is that narrow syntactic processes precede any PF processes (Spell-out), meaning that, once a particular domain (commonly called a phase) is spelled out, it is no longer accessible to syntax (Chomsky 2000, 2001, 2004, etc.). This talk presents ongoing research of the interaction between these two modules of the grammar, and proposes that the boundary between them is much more permeable than traditionally assumed. Specifically, I argue that syntax and PF (postsyntax) can be interleaved in such a way that a syntactic phase first undergoes Spell-out, and then participates in further narrow syntactic computation. I provide two pieces of evidence for this claim from the Niger-Congo language Wolof. The first one addresses a phenomenon in which elements that are in the final structure separated by intervening syntactic material undergo vowel harmony (Ultra Long-distance Vowel Harmony; Sy 2005). I show that at the moment of Spell-out the harmonizing elements are in a local configuration, only to be separated by syntactic movement in a later step in the derivation, resulting in a surface opacity effect. The second argument comes from the behavior of the past tense morpheme, which is in one configuration affixed onto the verb and carried along with it up the clausal spine, and in another stranded by the moving verb, exhibiting a Mirror Principle violation. I show that the past tense morpheme is affixed onto the verb in postsyntax (Marantz 1988, Embick & Noyer 2001), and that the syntax/postsyntax interleaving explains its variable position. The architecture of the grammar in which the syntax and postsyntax interact in a way proposed in this talk predicts precisely these types of surface opacity effects and removes the burden of accounting for them from narrow syntax. This spares us from positing idiosyncratic syntactic operations to account for anomalous phenomena that are in fact the domain of morphology or phonology, and allows us to maintain a view of syntax as cross-linguistically relatively uniform.