The Antisymmetry of Syntax proposes a restrictive theory of word order and phrase structure that denies this assumption.

The possibility of extraction across awh-island is usually assumed to be dependent on whether or not the constituent in question can undergo “long” (i.e., nonlocal) Ā-movement across the island. However, the question of how to make a principled distinction between those elements which can violate locality and those which cannot is still rather controversial. I will propose that there are no well-formed locality violations in these cases, and that the grammaticality patterns observed derive from a semantic filter on the escape (...) hatch used to bypass the island. In other words, if a phrase is extracted across awh-island, it must adjoin to the island in order to get by. This adjunction site is restricted by the interpretive component: only traces which are interpreted as variables of typee can occur in this position, while higher-order variables are not allowed. This restriction is shown to capture the known facts aboutwh-island violations, as well as some less known phenomena, such as the absence of functional (and pair-list) readings acrosswh-islands. (shrink)

A method now popular for fixing the scopes of arguments involves a covert movement operation, named QR (for Quantifier Rule) by Robert May. May envisioned QR as a kind of adjunction operation, attaching the arguments so affected to phrases dominating that argument. From the surface representation in (1a), for instance, QR can fashion the representations in (1b) and (1c) by adjoining the object and/or subject argument to IP. (1) a. [IP Someone [VP loves everyone ]]. b. [IP everyone1 [IP someone (...) [VP loves t1 ]]]. c. [IP someone2 [IP everyone1 [IP t2 [VP loves t1 ]]]] As the representations in (1a,b) suggest, QR has syntactic consequences rather like those displayed by Topicalization, the process that derives (2b) from (2a). (shrink)

Comparatives are among the most extensively investigated constructions in generative grammar, yet comparatives involving attributive adjectives have received a relatively small amount of attention. This paper investigates a complex array of facts in this domain that shows that attributive comparatives, unlike other comparatives, are well-formed only if some type of ellipsis operation applies within the comparative clause. Incorporating data from English, Polish, Czech, Greek, and Bulgarian, we argue that these facts support two important conclusions. First, violations of Ross’s Left Branch (...) Condition that involve attributive modifiers should not be accounted for in terms of constraints on LF representations (such as the Empty Category Principle), but rather in terms of the principle of Full Interpretation at the PF interface. Second, ellipsis must be analyzed as deletion of syntactic material from the phonological representation. In addition, we present new evidence from pseudogapping constructions that favors an articulated syntax of attributive modification in which certain types of attributive modifiers may occur outside DP. (shrink)