Resumen de Cardinal numerals and complex numerals as specifiers

Jacek Witkos, Dominika Dziubała Szrejbrowska

• The goal of this study is to argue for a more widespread application of a uniform representation of Numeral Noun Constructions (NNCs) which captures both patterns with higher numerals (≥5, NumH) agreeing in case with the modified noun (case matching pattern) or bearing a distinct case from a noun (case independence pattern) in Polish and in other languages. Our account draws on the analysis of cardinal numerals in Bailyn (2004) in which both agreeing and non-agreeing numerals are placed within the projection (QP) of the functional head FQ, but contrary to Bailyn (2004), it advocates a predominant cardinal-as-specifier representation of NNCc and complex NNCs in Polish, a language whose numeral/quantificational system is fairly challenging. We propose that many cases discussed in Bošković (2006), Ionin and Matushansky (2006), Kayne (2010), Danon (2012), Norris (2014) and Willim (2015) may have a uniform representation. In short, both the case matching and the case independence patterns are represented as equivalent to the cardinal-as-specifier representation. This serves to preserve derivational and structure building transparency and avoids the issue of look-ahead and the No Tampering ban (Chomsky 2000, 2001; Stepanov 2001, 2007). In the spirit of Burzio’s (1986) Generalization we submit that the case valuation capacity of FQ is conditioned directly by an independent case (feature) of NumP occupying FQ’s specifier position. Our proposal receives further support from recent work on the structure of complex numerals in Di Sciullo (2015, 2017) and Kayne (2016), where their component parts are combined in asymmetric structures involving silent functional projections.