The form of Semitic noun phrases

I propose a phrasal-movement analysis of word order in Hebrew and Arabic noun phrases. I argue that the positioning of nominal modifiers with respect to the head of simple nouns cannot be adequately handled by an N-raising derivation. This is so because nouns fail to move as heads in the grammars of Hebrew and Arabic. The only heads that seem to be able to move are those which do not assign genitive case. When modifiers (adjectives, numerals, demonstratives, etc.) appear to the right of the noun, it is the noun phrase itself which has raised to the left of the modifier, moving from specifier to specifier and pied piping all the material on its right. In this manner, inverse or mirror image ordering of post-nominal material is accounted for. Some modifiers have the option of appearing pre-nominally as well as post-nominally. 1 argue that these elements can be merged either as heads governing their complement and assigning genitive case to their sister or to the specifier of their sister, or as specifiers of projections containing a phonetically null head. In the former case, the modifier's sister is frozen in position, being licensed by genitive case. In the latter case, XP-movement of the sister of the null head is triggered, and agreement or concord is manifested on the modifier. I develop a configurational theory of agreement, which accounts for these phenomena. Many of the signature properties of the Semitic Construct-state fall out from the phrasal movement analysis. Finally, action nominalizations are dealt with. I provide arguments for a syntactic derivation in which a VP is associated with a nominalizing head configured in the clausal hierarchy as a functional head. I show that the position of the nominalizing head is above certain aspectual projections and lower than others. This explains why some adverbials, but not others, may occur inside a derived nominal. (C) 2003 Elsevier B.V. All rights reserved.