The underlying unity of Reference and Quantification

The mapping of nominal arguments to semantic interpretation exhibits a certain amount ofintriguing empirical variation across languages; it has become increasingly clear, at least sinceLongobardi (1994), that a good deal of such polymorphy depends on a major parametricdivide, separating two types of languages: certain languages (e.g. Romance, but also Classicaland Modern Greek, Bulgarian, Arabic (Fassi-Fehri 2003)…; henceforth 'strong D' languages)exhibit overt association of nouns functioning as referential constants (proper names andreferential generics) with D (either by overt N-to-D raising, e.g. of proper names, or by anarguably expletive article), others do not (e.g. English, but also probably the rest of Germanic,Celtic…; 'weak D' languages).