Unprobed movement: three case studies

TitleUnprobed movement: three case studies
Publication TypeArticle de revue
Année de publicationSoumis
AuthorsCecchetto, Carlo, and Caterina Donati

This paper investigates the status of unmotivated movement in syntax, which, following current minimalist terminology, we call ‘unprobed movement’.  By elaborating on the theory of labeling developed by Cecchetto & Donati (2015a), in which the label is provided by the probe of the merging operation, we assume that unprobed movement is possible in principle but is severely restricted in practice, as it leads to the creation of an object without a label.

Given this framework, we re-consider three movement configurations: the intermediate step of successive movement, which has no obvious trigger and is intrinsically unstable; QR, which has no morphological reflex, is systematically covert and appears to be clause-bounded; and head movement, which again has peculiar locality restrictions and correlates with extended projections and the creation of complex lexical items. The basic idea is that these three movements are instances of unprobed operations creating unlabeled syntactic objects. In particular, unlabeled syntactic objects, being invisible to syntax, can occur only in those points of the derivation where the syntactic derivation can stop (or has already stopped): post-cyclically, namely after the syntactic derivation has ended and unlabeled nodes bring no harm (QR,); at the phase edge, where the derivation stops and ships the phase to the interface, thus destroying any unlabeled syntactic node and putting the moved item back into the working space where it can be probed again (successive cyclic movement); again at the phase edge, when an unprobed item labels the syntactic object it forms thanks to its being a word, hence endowed with an intrinsic labeling capacity (head movement).

By putting successive cyclicity, QR and head movement into a new natural class, that of unprobed movements, we provide a new explanation to old problems and we make a strong case in favor of a label-driven syntax in response to recent attempts (cf. Chomsky 2013) to expunge labels from syntax and relegate them to the interface.