Gender mismatches in partitive constructions with superlatives in French

Petra Sleeman,Tabea Ihsane

doi:10.5334/gjgl.137

Abstract

This paper examines the syntax of partitive constructions with superlatives in French and the realization of gender in constructions with an animate noun, and focuses on contexts that allow gender mismatches (la plus intelligente de mes gentils professeurs ‘the most intelligent of my kind professors’). We show that such mismatches are allowed in French superlative partitives that contain a default masculine noun. To account for the mismatches, we propose that the nouns involved have an unvalued gender feature that can be assigned a sex specification in the course of the derivation. For superlative partitives we propose that this specification takes place in a Gender Phrase in the outer DP. Furthermore, we show that quantified partitives (e.g. three of the books) behave differently from superlative partitives, a fact we try to explain in terms of structural differences. We distinguish between partitives with an of-complement and non-canonical partitives with an ‘among’-PP or a preposed ‘of’-phrase. We also claim that our data can provide further insight into the role of locality in semantic agreement: we compare superlative partitives to quantified partitives in terms of agreement, and suggest that the two types of partitives should be placed in different positions on an Agreement Hierarchy.

Highlights

Partitive constructions that are discussed in the literature are often introduced by a quantifier (‘three of the books’), see, e.g., Hoeksema (1996), Zamparelli (1998), Cardinaletti & Giusti (2006; 2016)
On the basis of data from French, we have shown that, whereas superlative partitive constructions allow internal gender mismatches, canonical quantified partitives do not or do only marginally do so
We have argued that both canonical superlative partitive constructions and canonical quantified partitive constructions contain an unpronounced copy of the noun

Summary

Introduction

Partitive constructions that are discussed in the literature are often introduced by a quantifier (‘three of the books’), see, e.g., Hoeksema (1996), Zamparelli (1998), Cardinaletti & Giusti (2006; 2016). Gender mismatches in partitive constructions with superlatives may only occur with nouns of classes b) and c) The nouns of these classes have a masculine form to refer to males and a feminine form to refer to females. One of our informants does not accept a gender mismatch in a superlative partitive construction containing the class c) noun enfant ‘child’, judging the feminine use of this noun outdated, and only accepts the default masculine form in the inner and outer DP to refer to a female. To account for the data in (42)–(44), we propose that the empty noun in these examples differs from the one in the canonical ‘of’-partitives in (39)–(41) (recall the discussion of (33)–(35)), as it is not a copy of a lower N, but an independent noun taken from the Lexicon, which is, in this case, empty Support for this analysis comes from the fact that two different full lexical nouns (e.g. enfants ‘children’ and garçons ‘boys’) could be used in these constructions:.

Theoretical assumptions

Analysis of gender mismatches in partitive constructions

Conclusion