Lextutor has only one several usable mid-size corpus corpora.
For years the main French corpus on Lextutor, assembled by Thierry Selva from Le Monde articles in 1998 for a research project.
1 This is 2.7 million words of combined corpora draw from the following full corpora (described in detail below)... and used in the French Families project in 2022-2023 as a balanced (written, spoken; adult, children; academic, journalism, fiction; continental and Candian) corpus to indicate which family members to include in nuclear families in the FNFL (French Nuclear Families List). See the equvalent paper in English, and French one will appear here within 2023.
- WRITTEN ACAD academic 1 million+
- WRITTEN JOURNALISM 3 newspapers from Chambers 450,000 words (may yet get a Montreal paper into the mix) + part of LeMonde
- WRITTEN CHILDREN Petit Nicolas 130,000 words
- SPOKEN Paris 488,000
- SPOKEN Quebec 703,000
OCT 2022
This is the Chambers-Le Baron corpus, about 1.1 million words, divided into 11 academic subjects. A smaller partner for the BAWE on English side. More info online via reference:Chambers, A., & Le Baron, F. (2007). The Chambers-Le Baron Corpus of Research Articles in French/Le Corpus Chambers-Le Baron d'articles de recherche en français. Oxford: Oxford Text Archive. http://ota.ahds.ac.uk/headers/2527.xml
COMPOSITION
OCT 2022
This is the Chambers-Rostand corpus, about 450,000 words, from three sources: L'Humanité (Socialist viewpoint), Le Monde (centre-right), and La Dépeche du Midi (equal in article count though not especially equal in wordcount). More info online via reference:Chambers, Angela; Rostand, Séverine and University of Limerick, Ireland, 2003, The Chambers-Rostand Corpus of Journalistic French, Oxford Text Archive, http://hdl.handle.net/20.500.12024/2491.
This corpus is made up of 1723 articles taken from three daily French newspapers:
- Le Monde (576 articles / 355,046 words)
- L'Humanité (576 articles / 367,486 words)
- La Dépêche du Midi (570 articles / 257,299 words)
As stated on CorpusFinder
OCT 2022 - This is an adaptation of the CFPP2000, Corpus de Français Parlé Parisien des années 2000, assembled at Université Paris 3 Sorbonne nouvelle up to and including 2022. Size= 488,000 words in this rendition. XML format stripped and files joined into a single textfile for lexical purposes. Extensive details of original corpus plus original recordings and transcripts are available here - http://cfpp2000.univ-paris3.fr/
OCT 2022 - This is an adaptation of Gaétane Dostie and colleagues' CFPQ, Corpus du français parlé au Québec, assembled at the Université de Sherbrooke up to 2019, but eliminating the commentary and interaction aspects to focus on the lexis and run as a simple concordance output. The 31 sub-corpora,, 703,000 words in this rendition, carefully spelled out as to topic, participant background and age at https://applis.flsh.usherbrooke.ca/cfpq/, are joined here as a single textfile. Sentence boundaries and some spellings are nonstandard but can be interpreted. The motivation and exploitation of the original project are described here https://journals.openedition.org/corpus/2945
Summer 2022
130,000 words of highly natural kids talk from the 1960s without the passé simple verbs that have made French children's stories unnecessarily formal and difficult (for ex, Le Petit Prince).
Not particularly large, but includes all the main five collections:
- Le Petit Nicolas (original set)
- Le Petit Nicolas et copains
- Les récés du Petit Nicolas
- Les vacances du Petit Nicolas
- Le Petit Nicolas a des ennuis
The spoken corpus was assembled by Kate Beeching, University of West England (obtain the corpus on PDF and an account of its composition at http://www.llas.ac.uk/resourcedownloads/80/mb016corpus.pdf. The written member of the pair was a selection of comparable size from the the Le Monde corpus above meant to allow rudimentary spoken-written comparisons in French.OCT 2022 - brought together as 'Speech_Writing (x2)' for simpler comparison in output (via 'sort by sub-corpus'), while remaining available separately.
This is the Gutenburg download of de Maupassant's collected works, mainly used as a story corpus on the Boule de Suif suite of activities on Lextutor.
Sept 2014, this 37 million word corpus of mainly literary texts was assembled by Boris New and other French linguists associated with lexique.org to respond to the "serious lack of publicly available French corpora of any size" (described and available for download at lexique.org - read about it at http://www.lexique.org/public/lisezmoi.corpatext.htm.Because Lextutor can handle corpora up to just over 15 million words, Corptext was reduced to a mini-version (or 10% of the total corpus), answering the need for a useful 3 million word pedagogical corpus in French to, for example, furnish a sufficient number of examples of words for meaning inference).
Also, the full corpus was divided into three roughly equal parts (11, 12, and 13 million, roughly) called corpatext_a, corpatext_b, and corpatext_c, as a means of searching the whole corpus in pieces - or facilitating interesting range-type comparisons (like, is the frequency of a word or feature uniform across the three divisions).