Skip to Main content Skip to Navigation

Modélisation d'un discours étymologique. Prolégomènes à l'informatisation du Französisches Etymologisches Wörterbuch

Abstract : The Französisches Etymologisches Wörterbuch, the reference book in French and Romance linguistics, is currently underused because its lexicographic features make it hard to search and hard to read. The retroconversion of the 25 printed volumes of the FEW into a computerized dictionary, as desired by the scientific community, may address this problem. The density and the structural complexity of the dictionary may however make retroconversion appear as an unreasonable, even utopian, endeavor. Furthermore, computerization induces the risk of skewing the fewian discourse, thus opening the way to incorrect search practices. This doctoral dissertation proposes to study the feasibility of the retroconversion of the FEW, taking into account all of the constraints on the project. In the first part of the study, we model the fewian etymological discourse in order to address the searchability and readability issues of the dictionary in a way that preserves its structures. This modeling, formalized in XML, takes account of two complementary dimensions of the FEW, corresponding to two visions of the dictionary: a thesaurus of lexical units on one hand, a collection of monographs on the other hand. The second part of our study examines how to apply the model to the fewian text in an automated fashion. A retroconversion software was constructed to this end. The kernel of the software is comprised of about thirty algorithms that identify and tag, in an article of the FEW, the various relevant types of information. The application of the software to a corpus of 150 FEW articles results in a tagging that exhibits very few problematic errors. There is still room for improving the tagging, but it already demonstrates the feasibility of a retroconversion of the FEW that, assuming the availability of appropriate exploitation tools (search engine and search interface), should address most of the expectations of the scientific community, making the dictionary more accessible without eluding the complexity and depth of the fewian discourse.
Document type :
File URL :
Complete list of metadata
Contributor : Thèses Ul <>
Submitted on : Thursday, March 29, 2018 - 11:59:47 AM
Last modification on : Thursday, February 25, 2021 - 9:54:05 AM


  • HAL Id : tel-01749053, version 1



Pascale Renders. Modélisation d'un discours étymologique. Prolégomènes à l'informatisation du Französisches Etymologisches Wörterbuch. Linguistique. Université Nancy 2, 2011. Français. ⟨NNT : 2011NAN21011⟩. ⟨tel-01749053⟩



Record views