Focus, Adjacency, and Nonspecificity Miriam Butt and Tracy Holloway King University of Konstanz, Xerox PARC This paper investigates an intriguing interaction between preverbal structural focus and the nonspecific interpretation of preverbal objects in Urdu 1 and Turkish. 2 In both languages, accusative objects can scramble freely and are interpreted as specific. Nominative objects are semantically incorporating nonspecifics and can- not scramble, but they behave like accusative objects when they are coerced into a specific interpretation by the presence of preverbal focus. The apparently con- flicting functions of the immediately preverbal position in Urdu and Turkish argue for two independent constraints: a differentiation between (non-)incorporating ob- jects and the licensing of focus in a way similar to that of case, thus confirming the close association between focus and case licensing in two genetically unrelated but typologically similar languages. 1 Object interactions Objects in Urdu and Turkish interact with morphology, syntax, and semantics. This section first examines the interaction of nominative objects with preverbal focus. Then the interaction of specificity with case marking and position are discussed. 1.1 Interaction between Focus and Unmarked Objects Turkish and Urdu both allow objects to appear in the nominative case, i.e., with- out overt case marking or “bare”. These bare objects are licensed in immediately preverbal position, which also licenses focus. 1.1.1 Turkish Kornfilt (1995) describes an intriguing interaction between focus and unmarked ob- jects in Turkish. Nominative objects, which are nonspecific, are normally restricted 1 The South Asian language Urdu is closely related to Hindi, which is mostly spoken in India. In this paper, we primarily draw our data from the dialect of Urdu spoken in Lahore, Pakistan and the dialect of Hindi-speaking informants from New Delhi, India. 2 The data and claims for Turkish are taken mainly from Hoffman (1995), who bases her findings on a corpus gleaned from the childes database (Mac Whinney and Snow (1995)), transcribed col- loquial speech, and contemporary novels. Further data is taken from Kornfilt (1995) and additional fieldwork conducted independently.