Adnominal agreement can be defined as agreement in the domain of the noun phrase (NP), whereby the modifiers of the noun agree in categories like gender, number and case with the head noun. In many languages, including English, the modifiers agree with the head noun in number, cf. the English opposition between this book and these books. Likewise, many languages use the unmarked form of the noun (usually identical to the singular form) with numer-als higher than one in numeral constructions (cf. Turkish iki kitap ‘two books’, where kitap ‘book’ is the unmarked, singular form), while other languages, such as English, use the number-marked form of the noun in such numeral construc-tions (cf. English two books, with books in the plural). This paper uses a geneti-cally and areally balanced sample of 100 languages to show that languages like Turkish, in which nouns are unmarked for number in numeral constructions (with numerals above ‘one’), tend not to have adnominal number agreement. Although there are a few counter-examples to this universal, the correlation between the two features (having nouns unmarked for number in numeral constructions and lacking adnominal number agreement) is statistically highly relevant. We also show that this typological universal is not a consequence of different areal biases in the distribution of the features involved, i.e. that it is not a contact phenomenon. Lastly, we discuss the theoretical relevance of our findings and their implications for Jan Rijkhoff’s theory of ‘Seinsarten’ (Rijkhoff 2002, 2008).
© 2001-2024 Fundación Dialnet · Todos los derechos reservados