Department of Linguistics, Yale University April ͳͺ, ʹͲͳ͵ The determiner restriction in nominalizations Gianina Iordăchioaia 1 University of Stuttgart 1. Introduction )t has been noted, at least as earlyas Lees ȋͳͻ͸ͲȌ, that some nominalizations in English are fully flexible with respect to determiner selection, while others are restricted ȋsee ȋͳȌȌ. ȋͳȌ a. John's/the/that/a performing/performance of the song b. John's/*the/*that/*a performing the song ȋZucchi ͳͻͻ͵Ȍ The question to answer is what properties of these two types of nominalizations create this contrast and how we can explain it theoretically. )n this talk ) contribute the following observations: ͳ. Full (1a) vs. defective (1b) nominalizations On the basis of a comparison to German, Romanian and Spanish ) show that the contrast in ȋͳȌ correlates with a distinction between nominalizations with full vs. defective nominal structure, where the former pattern with ȋͳaȌ and the latter with ȋͳbȌ. ʹ ʹ. Presence/absence of the nP layer and feature valuation between D and n )n a syntactic approach to nominalization ȋe.g., van (out & Roeper ͳͻͻͺ, Borsley & Kornfilt ʹͲͲͲ, Alexiadou ʹͲͲͳ, Borer ʹͲͲͷȌ, ) take the fully nominal structure to be introduced by the categorizing nP layer and argue that the restriction on determiners is due to unvalued gender and number features on D which can only be valued via Agree with the corresponding lexically valued features of n. Determiners that appear in defective nominalizations ȋwithout an nPȌ receive a 'default' value for the two features. The languages ) discuss here are particularly illuminating in this respect, as they mark gender morphologically and one can see that syntactic gender features ȋsee Picallo ʹͲͲ͸Ȍ are available in full nominals, but not in defective ones. ͵. 'Default' and 'expletive' determiners The distinction among determiners that appear with full or defective nominalizations seems to corroborate earlier conclusions concerning the expletive use of definite articles in English vs. Romance languages and German. Overview of the talk  Section ʹ: Full vs. defective nominalizations  Section ͵: Feature valuation between DP and nP  Section Ͷ: Default/Expletive determiners  Conclusions ͳ Thanks to Artemis Alexiadou, Carmen Dobrovie Sorin, Fabienne Martin, Friedericke Moltmann, the audience of NELS Ͷ͵ and of the Paris Workshop on Languages with and without Articles for useful comments and suggestions, and to Mariangeles Cano, Antonio Fábregas, Matías Jaque, Amanda Kahrsch, Susanne Lohrmann, Rafael Marin, Paula Menendez‐Benito, Sabine Mohr, Neil Myler, Christopher Piñón, Marcel Pitteroff, Florian Schäfer, Soledad Varela and Jim Wood for help on the data. This research was funded by the German Research Foundation through a grant to Project Bͳ, The Formation and Interpretation of Derived Nominals, within the Collaborative Research Center ͹͵ʹ, Incremental Specification in Context. ʹ A similar explanation is actually given in Chomsky ȋͳͻ͹ͲȌ.