Phase-Constrained Obligatory Late Adjunction
Erik Zyman
June 2019
 

NOTE: This is the version that was originally submitted. The (substantially modified and improved) published version can be found here: https://onlinelibrary.wiley.com/doi/full/10.1111/synt.12226 • • • DOI: https://doi.org/10.1111/synt.12226 • • • It has sometimes been argued that adjuncts can merge late. This article provides a new argument for late adjunction and a new analysis of it. “Exactly” (or “precisely”) can adjoin to a wh phrase, and wh movement can affect either the entire adjunction structure (“What exactly is it?”) or only the host (“What is it exactly?”). Structures like “what exactly” can be generated VP internally, but surprisingly—and despite appearances—“exactly” cannot be stranded VP internally. The article argues that this can be understood if adjuncts merge late obligatorily, at the phase level: for H a phase head with complement XP, all adjunction within the HP phase must occur immediately before XP spellout. Further predictions of this analysis are shown to be correct. The larger picture that emerges is one in which the syntax prioritizes satisfying featural requirements: if so, it is not surprising that it waits until the last possible moment (within each phasal subderivation) to add in adjuncts.
Format: [ pdf ]
Reference: lingbuzz/006317
(please use that when you cite this article)
Published in: Syntax
keywords: derivational timing, “exactly”-stranding, late adjunction, late merger, phasal spellout, phases, syntax
Downloaded:349 times

 

[ edit this article | back to article list ]