OFrLex: A Computational Morphological and Syntactic Lexicon for Old French

Proceedings of the 12th Conference on Language Resources and Evaluation (LREC 2020), pages 3217–3225 Marseille, 11–16 May 2020 c⃝ European Language Resources Association (ELRA), licensed under CC-BY-NC 3217 OFrLex: A Computational Morphological and Syntactic Lexicon for Old French Gaël Guibon, Benoît Sagot LLF (CNRS) – Université Paris Diderot & Almanach (Inria), Almanach (Inria) Paris, France {firstname.lastname}@inria.fr Abstract In this paper we describe our work on the development and enrichment of OFrLex, a freely available, large-coverage morphological and syntactic Old French lexicon. We rely on several heterogeneous language resources to extract structured and exploitable information. The extraction follows a semi-automatic procedure with substantial manual steps to respond to difficulties encountered while aligning lexical entries from distinct language resources. OFrLex aims at improving natural language processing tasks on Old French such as part-of-speech tagging and dependency parsing. We provide quantitative information on OFrLex and discuss its reliability. We also describe and evaluate a semi-automatic, word-embedding-based lexical enrichment process aimed at increasing the accuracy of the resource. Results of this extension technique will be manually validated in the near future, a step that will take advantage of OFrLex’s viewing, searching and editing interface, which is already accessible online. Keywords: Morphological lexicon, Syntactic lexicon, Lexicon Enrichment, Old French 1. Introduction Old French regroups romance languages qualified as Oïl languages used in the north of France, south of Belgium and in the Anglo-Norman islands spoken from 8th century to 14th century. They contrast with the Oc languages that come from the south of France. Contrary to Middle French, Old French possesses nominal declination. Both led to con- temporary French and possess relatively free word order: verbs are often in second position following a non subject constituent. Moreover, there is no spelling standardisation in Old French, even for proper nouns from the same author. The main textual databases with semi-automatic lemmas and part-of-speech tags (PoS) are the Base de Français Médiéval (BFM - Medieval French Base) (Guillot et al., 2017)1 with more than 4 million words and the Nouveau Corpus d’Amsterdam (NCA - New Amsterdam Corpus) (Stein and al., 2008)2 with more than 3 million words. The main treebanks for Old French are the Syntactic Reference Corpus of Medieval French (SRCMF) (Stein and Prévost, 2013) and the Old French subpart from the Modéliser le changement : les voies du français (MCVF) corpus (Mar- tineau, 2008). However, they do not share the same syntac- tic and POS tag sets, and only SRCMF is on open access with part of it in Universal Dependencies (UD) (McDonald et al., 2013) format3. In the available resources different kinds of text are gath- ered. Some vary in style (prose, verse), literary genre (re- ligious, historical, didactical, etc.), or even in time span (from 10th century to 13th century). Nevertheless, there is no available morphological lexicon,4 and a fortiori no 1http://bfm.ens-lyon.fr 2https://sites.google.com/site/ achimstein/research/resources/nca 3https://github.com/ UniversalDependencies/UD_Old_French-SRCMF/ 4A morphological lexicon is a collection of entries of the form syntactic lexicon5 for Old French. Most of the existing lex- icons and dictionaries are either not …

Content truncated to 500 words