French transdisciplinary lexis (V1)

The lexicon and phraseology of French academic writing

Agnès Tutin – LIDILEM – March 2013

This page contains some lists of French words and expressions particular to academic transdisciplinary lexicon – words, collocations and repeated segments. A project to expand this lexicon is in progress thanks to the French ANR Termith project.

  1. The Corpora

All of the statistical information concerning the lexicon (words, collocations, repeated segments) was extracted from the KIAP-LIDILEM corpus of French created from part of the KIAP corpus and enriched by the LIDILEM laboratory.

The corpus contains 2 million words of several kinds of academic writing in French (published articles, theses, reports, class projects) from three contrasting disciplines – Linguistics, Economics, and Medicine. See Table 1 for the number of words for each type of text. The corpus of scientific articles was extracted from the KIAP[1] corpus designed and built by Kjersti Fløttum and her team at the University of Bergen.

Linguistique Economie Médecine
Articles de revues (corpus KIAP) 285 881 mots 374 516 mots 164 315 mots
Thèses, rapports, cours 364 812 mots 286 653 mots 492 173 mots
Total 650 693 mots 661 169 mots 656 488 mots

Tableau 1 : Corpus des écrits scientifiques

The corpus was lemmatized using Cordial, a software developed by Synapse enterprises to identify single words in the transdisciplinary field. It was analyzed by Syntex for the collocations analysis, and finally, the AntConc toolkit was used to identify 3, 4, or 5-word repeated segments in the Cordial-lemmatized corpus.

  1. A transdisciplinary single-word lexicon

Word lists were created according to parts of speech (adjective, noun, verb, etc) that were common to the three disciplines using the Cordial-lemmatized corpus.

The list of adjectives contains adjectives that were identified at least 15 times in each discipline. We removed:

  • Adjectives that could be considered grammatical words, such as “autre”, “même”, etc.
  • Adjectives that were always employed in a fixed expression, such as “échéant” in “le cas échéant”.
  • Obvious tagging errors (such as past participles tagged as adjectives).
  • Polysemous words having clearly divergent meanings in the different disciplines.

The list of nouns also underwent manual revision. It contains nouns that were identified at least 15 times in each discipline. We removed:

  • Nouns always employed in fixed expressions.
  • Obvious tagging errors (such as adjectives tagged as nouns).
  • Polysemous words having clearly divergent meanings in the different disciplines.

The list of verbs contains verbs that were identified at least 15 times in each discipline. It has not been revised.

  1. Transdisciplinary Collocations

The term « collocation » refers here to recurrent lexical associations of two linguistic elements, which have a syntactic relationship (Tutin 2008). The syntactic relations were extracted using Syntex, a parser developed par Didier Bourigault at the ERSS.

The list of French Noun-Adjective collocations was compiled using Syntex. We retained the associations that were identified at least five times in at least two disciplines.  We removed the associations that appeared to be multi-word terms or that were the result of tagging errors, as well as most of the ordinal and cardinal numbers and the adjectives that could be considered grammatical words (such as “autre” or “même”).

The list of Noun-Prep.-Noun collocations underwent the same revision, as well as the Noun-Subject-Verb and the Verb-Noun-Complement collocations.

  1. Repeated Segments

The repeated segments here are strings of 3, 4 and 5 words that appear at least four times in the three disciplines. The status of these segments is variable: completely fixed expressions (ex: “c’est-à-dire”), collocations (ex: “mettre en relation”), routine formulae (ex: “en ce qui concerne”), etc. This list underwent manual revision.

Bibliography

Tutin, A. (2007). Traitement sémantique par analyse distributionnelle des noms transdisciplinaires des écrits scientifiques. Actes de TALN 2007. Poster presentation, 283-292.

Tutin, A. (2007). Modélisation linguistique et annotation des collocations : application au lexique transdisciplinaire des écrits scientifiques ». In S. Koeva, D. Maurel, M. Silberztein (Eds). Formaliser les langues avec l’ordinateur. Besançon, Presses universitaires de Franche-Comté.

Tutin A. (2007) (ed.). Lexique et écrits scientifiques. Revue Française de Linguistique Appliquée. December 2007.

[1] ) KIAP : Kulturell Identitet i Akademisk Prosa. Cf. http://kiap.uib.no/. We would like to thank Kjersti Fløttum and her team for giving us permission to use this corpus.