Text Article
French
ID: <
WXsTWiVUGtikcJODQJa5m>
Abstract
We study the order of post-word supplements in French on the basis of a data table composed of sentences extracted from three corpus: French Tree Bank, Est-Republicain and ESTER. More specifically, we observe the relative order of the SN and the SP, where these two components are subcategorised by the verb. We propose a model of order based on the logistics regression method (Agresti, 2007). In this article we focus on the role of the verb and its semantic class annotated according to the dictionary Les Verbes du Français (Dubois de Dubois Charlier, 1997). Based on data modelling, we show that the verbal lemma and its semantic class play an important role in choosing the order of supplements. The coefficients associated with the verbal lemmes can be interpreted, which makes it possible to estimate the preferences of each verb. The final model we propose is based on the length of the supplements, the fixed character of the V SP sequence, the verbal lemma associated with the semantic class and the corpus. The value of this work is also to show, following the work on English (Bresnan et al., 2007), that preference questions in syntax can be modelled using statistical tools widely used in the humanities (Howell, 1998).