A statistical approach to rivalry in lexeme formation : French -iser and -ifier

Fiche du document

Discipline
Type de document
Périmètre
Langue
Identifiants
Relations

Ce document est lié à :
info:eu-repo/semantics/altIdentifier/doi/10.3366/word.2018.0130

Collection

Archives ouvertes




Citer ce document

Olivier Bonami et al., « A statistical approach to rivalry in lexeme formation : French -iser and -ifier », HAL-SHS : linguistique, ID : 10.3366/word.2018.0130


Métriques


Partage / Export

Résumé En

Rivalry in lexeme formation refers to a situation where multiple, rival lexeme formation processes may be used to fill a gap in a morphological family. In this paper we study one such situation, the rivalry between the suffixes -iser and -ifier in French to derive verbs from nouns and/or adjectives. We propose a statistical approach to the problem, and use multivariate logistic regression applied to a large dataset derived from existing ressources to establish that phonological, morphological, and semantic properties of the morphological family all contribute independently to predicting preference for one or the other suffix. One main result of this study is that rivalry can not be studied in terms of the relationship of a single base and a derived lexeme, as multiple members of the morphological family play a role in jointly predicting the choice of a suffix.

document thumbnail

Par les mêmes auteurs

Sur les mêmes sujets

Sur les mêmes disciplines

Exporter en