Le changement de construction des numéraux supérieurs à '5' en vieux slave
Résumé
In the new syntactic pattern of numerals higher than '5', the quantified noun agrees with the numeral, instead of being a complement in the Genitive plural. The new pattern is limited to oblique cases, and the majority of instances in OCS are found in prepositional phrases. This leads to the following hypothesis : in prepositional phrases with prepositions governing the Genitive case, the quantified noun in the Genitive, governed by the numeral, was reanalyzed as governed by the preposition, and promoted to the status of head of the numeral phrase, the numeral becoming a modifier, as is the case for smaller numerals from '1' to '4'. The extension to non prepositional oblique cases seems to be secondary. The immediate model, however, are not the smaller numerals, but the quantifiers of indefinite quantity such as koliko, which are characterized by a structural opposition between [Quant. NGen.] in the direct cases, the quantifier being the head, and [Quant.x Nx] in the oblique cases, where the quantified noun is the head and the quantifier an agreeing modifier. This structural opposition is old, and probably inherited for this type of non numeral quantifiers, it is found also in Latin (Nom.Acc. quantum argenti, but Abl. quanto argento, with determiner quantus agreeing with the head argento : the quantified noun argenti, which is a modifier in the direct cases, is promoted to the status of head of the noun phrase in the oblique cases). The coexistence of two syntactic patterns for the quantifiers of indefinite quantity, in Slavic or in Latin, is justified from the semantic point of view, it is used to differenciate the function of evaluation (" how big ? ") and the function of quantification (" how many / how much ? ") of the same morphemes quantus / quantum, kolikæ / koliko. The numerals higher than '5' borrowed this model through reanalysis in prepositional phrases : for numerals this syntactic duality is not meaningful, it merely reproduces that of the model. It is not necessary to assume with Žolobov 2006 that complex numerals such as '20', '30' played a role in the extension of the new syntactic pattern.