Wh-Phrases and Pied-Piping
The starting point of Donati’s chapter is the assumption made in OWM that wh-movement operates on a phrase (see OWM, rule (1b): ‘‘Move wh phrase’’). Donati observes that this approach is incompatible with the minimalist feature-based approach to movement, whereby only features are visible to the computational procedure, and phrasal properties of wh movement need to be derived. Donati then argues that pure feature movement should be excluded as a syntactic option under the assumption that features cannot be merged. Thus, pied-piping is always involved. She proposes that head movement (i.e., pied-piping of X) is the unmarked pied-piping operation, since it is minimal (i.e., drags along the least mate rial). The question then arises how to derive phrasal movement (pied piping of XP—e.g., a wh-phrase) as an option of grammar. Donati proposes to relate the choice between head movement and phrasal movement to the phrase structure status of heads and phrases, in combination with the Condition on Uniformity of Chains (Chomsky 1995) and an economy condition on Merge Copy. An implication of this approach is that there is no principled reason for wh-movement to be restricted to phrasal constituents (e.g., a wh-phrase) in any context. To illustrate this point, Donati discusses two cases of wh-head movement: free relatives and comparatives.
Watanabe’s chapter explores the existence of an uninterpretable formal feature whose role is to mark a constituent that must be pied-piped. This so-called pied-piper feature is needed to solve a problem that arises under a movement theory in which movement consists of three (successive) sub operations: Agree, pied-piping, and Merge. More specifically, if Agree renders the goal inactive, then both the goal category and some category that contains the goal become inaccessible for further computation in narrow syntax. Obviously, pied-piping cannot be ordered before Agree, because Agree must first locate the goal. Watanabe argues that the introduction of the pied-piper feature solves this problem: the pied-piper feature marks the category to be copied by pied-piping. He presents different types of empirical evidence in support of this feature: (1) the change of (Old) Japanese, a language with overt phrasal wh-movement and concomitant marking of the wh-phrase by the particle ka, into a wh in-situ language without ka-marking on the wh-phrase; (2) the appearance of a focus particle on an overtly moved wh-phrase in languages with optional wh-movement (e.g., Kikuyu).
Agbayani’s chapter discusses the nature of specifiers formed by movement by looking in particular at the peculiar behavior of wh-subjects in English, which have been argued to not undergo movement (George 1980; Chomsky 1986a), and which seem to display paradoxical behavior with respect to the wh-diagnostics of OWM (e.g., they obey Subjacency but apparently do not leave a gap). Agbayani offers an alternative perspective on the requirement that a specifier be created by overt (wh-)movement (i.e., the Extended Projection Principle), by closely examining the behavior of wh-elements as specifiers. In line with Chomsky (1995), Agbayani proposes that if a feature is extracted out of a category in overt syntax for checking, then PF requirements force subsequent pied piping of the category, unless economy blocks the pied-piping, as he argues is the case with wh-subjects in English. He argues that such an analysis of specifier formation via movement is expressed naturally in a ‘‘two-chain’’ theory where bare feature movement (Move F) and subsequent category pied-piping take place in overt syntax. Evidence for such an approach to movement comes from the peculiar behavior of wh subjects in English under the Vacuous Movement Hypothesis.
