In: Putnam, M. (Ed.), Exploring Crash-Proof Grammars. John Benjamins, Amsterdam, pp. 31-57. Implications of grammatical gender for the theory of uninterpretable features * Vicki Carstens University of Missouri carstensv@missouri.edu This paper argues that in languages with grammatical gender systems, the gender of nouns is uninterpretable yet intrinsically valued. Extensive evidence is then presented that nominal gender is infinitely reusable as an “active goal” feature in successive Agree relations, unlike Case; thus Agree relations apparently do not lead to uF deactivation /deletion marking for nominal gender. I point out that the approaches of Chomsky (2001), Chomsky (2007, 2008) and Richards (2008) lead to false predictions of crashes at the Conceptual-Intentional (C-I) interface for syntactic objects that contain nominal gender. In extending Minimalist theory to this under-explored variety of uF, I find strong support for the claims of Epstein, Kitahara, and Seely (this volume) that uFs do not require deletion; they are simply ignored at the C-I interface. I argue further that deactivation of DPs accompanies Case-valuation because PF can spell out only one value for a given formal feature. Hence a property of the Sensory-Motor interface gives rise to the phenomenon that Case- valued DPs are “frozen in place” (Chomsky 2001), not the C-I interface as is standardly assumed. No comparable “freezing” effect is connected with nominal gender because its value is not determined through Agree relations. 1. Introduction 1.1 Theoretical overview This paper begins with an exploration of grammatical gender and then outlines some consequences of its properties for Minimalist syntactic theory. The theoretical interest of the investigation arises from a general assumption, following Chomsky (2001), that features come in just two varieties: interpretable, valued features on the one hand, and uninterpretable, unvalued features (uF) on the other. The properties unvalued and uninterpretable are biconditionally related. 1 (1) uF biconditional: F is uninterpretable ßà F is unvalued (Chomsky 2001) Because they are meaningless, uFs cannot figure in interpretation. Chomsky argues that they must be eliminated before the syntactic object containing them is handed over to the * Thanks to Mike Putnam and to my fellow participants in the workshop Exploring Crash-Proof Grammars for their comments. For many lively and stimulating discussions on aspects of this project I am especially grateful to Michael Diercks, Sam Epstein, Hisa Kitahara, and Daniel Seely. 1 This section draws heavily on helpful discussion in Pesetsky & Torrego (2007). In particular, (1) and (2) are based largely on their formulation of how interpretability, valuation, and deletion interact in Chomsky (2001).