Abstract: Phi features are crucial for syntax and morphology, and number and person accordingly have been the focus of a significant amount of research. However, gender has received less theoretical attention, with the result that critical questions concerning its syntactic representation and morphological realization remain open (Is there a GenderP? Are gender features interpretable? How are the gender features found in the syntax spelled out?). In this talk, I present the results of an ongoing research project to develop a cross-linguistically viable proposal for the morphosyntax of gender. Assuming that lexical categories are decomposed into a category-defining head and a category-neutral root, I argue that gender is located on the nominalizing head ‘little n’ (cf. Ferrari 2005, Acquaviva 2009). I also argue that gender features can be either interpretable (interpreted as biological sex for many languages) or uninterpretable (e.g., the gender of an inanimate noun in French or Spanish). I initially motivate the analysis with an in-depth case study of Amharic (Ethiosemitic), and I argue both that the gender system of Amharic poses challenges for previous gender analyses, and that it provides clear support for gender on n. I continue by identifying and confirming a wide range of cross-linguistic predictions of the proposal, including the complex interplay of number and gender in Somali, the relationship between nominalizations and gender across languages, and the role of interpretable gender features generally in gender assignment. I conclude by showing how the proposal generates a typology of possible (and impossible) two-gender systems, and show how each type of predicted language is attested.