Rajesh Bhatt, University of Pennsylvania and Elena Anagnostopoulou, Tilburg University
In the literature, at least two ways of marking specificity have been proposed - either structurally by moving out of the VP (Diesing (1992)) or morphologically as in Turkish (Enç (1991)). We discuss data from Hindi which marks specificity in a way very similar to Turkish, by the postposition ko. We claim that while the presence of morphological marking of specificity is necessary, it is not sufficient. To be interpreted as specific, an NP/DP has to satisfy the structural condition and be morphologically marked.
The crucial data supporting our claim come from double object constructions. In particular consider the obligatory scrambling: Ram-ne (Erg) Sita-ko (Dat) billi (cat) di-ii (give-Pfv) vs. Ram-ne (Erg) Sita-ko (cat-KO) Sita-ko (Dat) di-yaa (give-Pfv) (both `Ram gave the cat to Sita'). These data suggest a very restrictive shortest-move-obeying view of scrambling. We discuss the nature of the specificity-induced movement and the nature of the site in which specificity is licensed. Finally, we also discuss a connected alternation seen with the verb `send'. We extend our analysis of the specificity- ko to handle the dative- ko.