Wikidata

Wikidata:Requests for comment/How to avoid to use male form as a generic form in property labels in French?

< Wikidata:Requests for comment
(Redirected from Wikidata:Requests for comment/How to avoid to use male form as a generic form in property labels in French ?)
An editor has requested the community to provide input on "How to avoid to use male form as a generic form in property labels in French?" via the Requests for comment (RFC) process. This is the discussion page regarding the issue.

If you have an opinion regarding this issue, feel free to comment below. Thank you!


THIS RFC IS CLOSED. Please do NOT vote nor add comments.

This request for comments focus on properties whose values are instances of human (Q5). For instance author (P50) is a property which often has a human being as a value. In French language there is often a difference between the male and the female form.

The following query show all the properties whose value is an instance of human (Q5) with their label on French, the list of aliases in French, the female form female form of label (P2521) and the male form male form of label (P3321).

The following query uses these:

  • Properties: female form of label (P2521)  View with Reasonator View with SQID, male form of label (P3321)  View with Reasonator View with SQID, property constraint (P2302)  View with Reasonator View with SQID, class (P2308)  View with Reasonator View with SQID, relation (P2309)  View with Reasonator View with SQID
    #title: List of properties which have instances or subclass of humans as possible value with male and female form in French
SELECT ?property ?propertyLabel ?maleform ?femaleform (GROUP_CONCAT(?alias; SEPARATOR = ", ") AS ?aliases) WHERE {
  ?property p:P2302 ?statement.
  ?statement ps:P2302 wd:Q21510865;
    pq:P2308 wd:Q5;
    pq:P2309 ?relation.
  ?property rdfs:label ?propertyLabel.
  OPTIONAL {
    ?property skos:altLabel ?alias.
    FILTER((LANG(?alias)) = "fr")
  }
  FILTER((LANG(?propertyLabel)) = "fr")
  OPTIONAL {
    ?property wdt:P2521 ?femaleform.
    FILTER((LANG(?femaleform)) = "fr")
  }
  OPTIONAL {
    ?property wdt:P3321 ?maleform.
    FILTER((LANG(?maleform)) = "fr")
  }
}
GROUP BY ?property ?propertyLabel ?maleform ?femaleform
List of properties which have instances or subclass of humans as possible value with male and female form in French

This show that the label is often similar to the male form and different from the female form.

Why is it a problem?

  • Using the male form as a generic form may reinforce gender stereotypes. By using the male form in Wikidata as a label we reinforce the idea that authors or filmmakers are male by default.
  • This may reinforce the idea that Wikidata is dominated by male contributors. Women coming to Wikidata may feel excluded from the community because there is a wide use of male form as a generic form.
  • Moreover this excludes also non binary people.

For all this reasons and many others, it would be good to use more gender inclusive language in Wikidata.

There are several ways to do it.

  • Avoid nouns and use verbal forms : for instance we could replace "auteur" by "écrit par"
  • Use a combination of female and male form. This could be "auteur/autrice" separated by the slash or "auteur ou autrice". This may exclude non binary people
  • Use an inclusive form such as "auteur•rice". This is the best solution to include men, women and non-binary people.

It would be good to have have a consensus on that topic.

See previous discussions here and here.

PAC2 (talk) 20:28, 21 November 2021 (UTC)[reply]

Total support thanks for opening the conversation Nattes à chat (talk) 21:30, 21 November 2021 (UTC)[reply]
German has apparently the same problem as French has. Problem is that—to the best of my knowledge—there is not even remotely a "standard gender inclusive German language" that finds reasonable agreement, even among the supporters of gender inclusive language. Which makes it pretty difficult to use gender inclusive language, unfortunately.
So, this is a problem of (some) languages in the first place, not of Wikidata. Please also note that for data users, the gender-neutral Q-identifier is much more important than labels. In case we want to help data users to use correctly gendered labels, we should increase coverage with both female form of label (P2521) and male form of label (P3321) whereever possible/necessary, so that they do at least have the choice to use correct labels if they want to. —MisterSynergy (talk) 21:50, 21 November 2021 (UTC)[reply]
Yes, female form of label (P2521) and male form of label (P3321) are also very useful. I've tried to complete the values in the request above. But we also need to have a consensus for gender neutral labels. PAC2 (talk) 06:17, 22 November 2021 (UTC)[reply]
Strongly support. Patafisik (talk) 17:14, 22 November 2021 (UTC)[reply]
I totally agree with MisterSynergy. The Q-ID is much more important than labels. Also, labels should be clear and concise to facilitate there use. In French, the name of a profession is always given in the masculine gender, because it is the gender that acts as the neutral (according to the Académie française, the neutral, in French, takes the forms of the unmarked gender, i.e. the masculine). So for me, the current solution of labeling an item with the masculine gender and aliasing it with the feminine gender is the best solution.
For me, the last two solutions are not desirable. Putting both genders in the label would reduce the readability and therefore the efficiency of the labels. Using the middle point is not a widespread use and is much debated in the French-speaking world. The French Academy is against it, as well as the French Ministry of Education.
Maybe the first proposition (avoid nouns and use verbal forms) can be a solution, but again, I think it's less efficient. Tubezlob (🙋) 10:31, 25 November 2021 (UTC)[reply]
@Tubezlob: I disagree with you. First, Q-id is useful for machines but humans think with words and therefore use labels. The Académie Française (Q161806) is a very conservative institution and academicians are not specialists of linguistics (Q8162) (see this book L'Académie contre la langue française https://www.editions-ixe.fr/catalogue/lacademie-contre-la-langue-francaise/). The French Ministry of Education recommends to use both feminine and male form ("Il est également demandé de recourir à des formulations telles que « le candidat ou la candidate » afin de ne pas marquer de préférence de genre, ou à des formules telles que « les inspecteurs et les inspectrices de l'éducation nationale » pour rappeler la place des femmes dans toutes les fonctions.", https://www.education.gouv.fr/bo/21/Hebdo18/MENB2114203C.htm). But moreover, knowledge equity is a central pillar of the Wikimedia movement's strategy (https://wikimediafoundation.org/our-work/open-the-knowledge/#a1-help-us-achieve-knowledge-equity). It's important to use a language which does not exclude women and non binary people. There is really no reason to say that the male form comes first as a label and the female form comes in the second place as an alias. PAC2 (talk) 06:23, 26 November 2021 (UTC)[reply]
No the Académie française has in the mean time agreed that feminization of functions and professions is something legit.
The masculine form is not neutral, and I agree that having a neutral form is optimal, but this would require a use of inclusive language. Nattes à chat (talk) 20:46, 29 December 2021 (UTC)[reply]
  • If there is a plurality of possible values and given that all labels should also be available as statements, it's probably preferable to create separate statements with name (P2561) or another property. This also allows to qualifiers and references. Otherwise, we can't really determine how they are used, where and by who(m). --- Jura 09:32, 27 November 2021 (UTC)[reply]
Jura1 maybe but we also need to find consensus for labels. PAC2 (talk) 06:12, 2 May 2022 (UTC)[reply]
  • There's no reason why the label for a property has to be a noun. If we take the example of author (P50) we could use verbs. "authored by"/"rédigé par" (translation from Google translate) would also do the job of gender neutrality and look nicer then "auteur ou autrice". Maybe we can make a list of problematic properties and migrate them from nouns to verbs? ChristianKl ❪✉❫ 23:45, 10 May 2022 (UTC)[reply]

Voting edit

After some discussions, there is no clear consensus. I propose to vote the following proposition :

Concerning properties whose target value is a human such author (P50), if the male form is different from the female form, it is recommended to use :
  • both male and female form if possible (ie "auteur ou autrice")
  • a verbal form if it's to complex to use female and male form.

Anyone can vote using {{Support}}, {{Neutral}}, {{Oppose}} or any other voting template with a comment.

Poke MisterSynergy, Jura1, Nattes à chat, Tubezlob, Patafisik.

  •   Support I think it's very important to have more gender inclusive labels in Wikidata, that using the male form reinforces gender stereotypes. PAC2 (talk) 06:35, 2 May 2022 (UTC)[reply]
  •   Support --Patafisik (talk) 07:24, 2 May 2022 (UTC)[reply]
  •   Oppose Pas de raison de déroger aux règles de bases de français. --Shev123 (talk) 07:42, 2 May 2022 (UTC)[reply]
    Shev123 What do you mean by "règles de bases". Time is changing and the usage now is not the same as in the past. PAC2 (talk) 19:51, 2 May 2022 (UTC)[reply]
    Les choses n'évoluent pas aussi vite que cela. En l'état, le Masculin ou lz féminin sont bien suffisants. Les autres usages sont minoritaires.--Shev123 (talk) 20:24, 2 May 2022 (UTC)[reply]
    oui, je propose justement d'utiliser le masculin et le féminin ("auteur ou autrice", "réalisateur ou réalisatrice"), exactement comme on fait pour sibling (P3373). PAC2 (talk) 07:12, 3 May 2022 (UTC)[reply]
  •   Oppose Lourd et sans intérêt. Ayack (talk) 08:59, 2 May 2022 (UTC)[reply]
  •   Neutral it is indeed important but I don't think that labels are the solution for this problem. Cheers, VIGNERON (talk) 17:33, 2 May 2022 (UTC)[reply]
  •   Support I strongly support any mean to (really) neutralize forms (for occupations or other stuffs). A complete blog article about it, using scientific references: https://www.bunkerd.fr/ecriture-inclusive/ (in french) --Teleosteen (talk) 08:06, 3 May 2022 (UTC)[reply]
  •   Support en l'absence d'une solution purement technique satisfaisante. Si cela peut la provoquer tant mieux ! / as there is no purely tech answer to that issue (if that creates the opportunity to find one, let's go!) Kvardek du (talk) 22:27, 3 May 2022 (UTC)[reply]
  •   Support J'ai des doutes sur un genre neutre en français. --Guerillero | Talk 16:45, 4 May 2022 (UTC)[reply]
  •   Support Seems very reasonable as it is already implemented in godparent (P1290), patronym or matronym for this name (P2976), sibling (P3373) and patronym or matronym for this person (P5056) --FreeCorp (talk) 06:50, 5 May 2022 (UTC)[reply]
  •   Support Karl Oblique (talk) 15:51, 6 May 2022 (UTC)[reply]
  •   Neutral meh. Can't say it will change anything for me either way, but I think simpler is better. --Jahl de Vautban (talk) 16:07, 6 May 2022 (UTC)[reply]
  •   Oppose Trop lourd. Nie l'existence de sexes intermédiaires entre mâle et femelle, ou mal déterminés. Le genre grammatical masculin est une forme de neutre en français. De plus, il est facile d'adapter ces libellés lors de réutilisation en dehors de Wikidata, notamment avec female form of label (P2521). Tubezlob (🙋) 12:09, 14 May 2022 (UTC)[reply]
    en quoi la mention du genre féminin nie plus l'existence des personnes non binaires que la seule mention du genre masculin ? Je ne comprends pas bien ton argument Tubezlob. PAC2 (talk) 20:29, 14 May 2022 (UTC)[reply]
    Idem. Je rajoute que les personnes intersexes peuvent avoir un genre déterminé (malgré leur intersexuation, veulent être genré·es au masculin ou au féminin exclusivement), contrairement aux personnes non-binaires ou fluides de genre. -- Teleosteen (talk) 07:48, 15 May 2022 (UTC)[reply]
    D'un côté, il y a une réalité biologique (mâle, femelle, entre les deux) ainsi que le ressenti de chaque personne sur son identité sexuelle. De l'autre, la volonté par de nombreux contributeurs de Wikidata ici d'obliger tout le monde à choisir entre mâle et femelle (ou genre masculin ou féminin, si je vous suis). Quand on voit écrit « auteur », ça ne sous-entend évidemment pas « auteur masculin ». Avec « auteur ou autrice », on dit qu'un auteur ne peut être qu'un homme ou une femme (ou, si je vous suis, une personne utilisant les genres masculin ou féminin). Je ne comprends pas l'intérêt de mettre en avant cette dichotomie, qui, me semble-t-il, cause beaucoup de mal à beaucoup de monde. Tubezlob (🙋) 13:39, 25 May 2022 (UTC)[reply]
    Il n'est pas question d'identité sexuelle, il s'agit bien d'une identité de genre. Désolé de me répéter, mais je dois insister : une personne intersexe peut (et c'est généralement le cas) avoir un genre bien déterminé. Merci, par respect pour ces personnes, de ne pas impliquer l'intersexuation dans le débat : si débat il doit y avoir sur le sujet évoqué, il porterait sur comment accorder les propriétés pour des personnes non-binaires (vu que la question porte exclusivement sur le genre et pas sur la réalité biologique).
    Le coup du « masculin neutre » ne fonctionne pas, car la représentation mentale de ce qui est accordé au masculin dit « neutre » est… masculine dans la réalité (si je dis un voisin non identifié m'a retransmis mon courrier dans la boîte aux lettres, la grande majorité des gens ne se figureront pas qu'il pourrait tout aussi bien s'agir d'une femme). Il y a de la vulgarisation scientifique à ce sujet (les sources sont fournies pour plus de détails) : https://www.bunkerd.fr/ecriture-inclusive/ et https://www.youtube.com/watch?v=url1TFdHlSI (vidéo de Scilabus).
    Maintenant, s'agissant de WikiData, dire que telle femme est podcasteur par exemple est très perturbant. Les ressources données ci-dessus expliquent très bien pourquoi.
    Enfin, je vois pas pourquoi, partant de la prémisse que le masculin neutre n'existe pas dans la pratique, la forme inclusive causerait du tord à beaucoup de monde ?
    Ça permet d'inclure toutes les personnes quel que soit leur genre. Si cette forme cause du tord pour les personnes non-binaires, je veux bien des sources et connaître les usages qui conviennent pour elles. Teleosteen (talk) 13:02, 27 May 2022 (UTC)[reply]
    Je ne vois pas tellement comment, sur l'élément concernant une femme, le fait qu'il soit écrit podcasteur (dans l'affichage brut de Wikidata) pour reprendre ton exemple, va tromper le lecteur qui ne saura plus si la personne est une femme ou un homme, car sa représentation mentale sera troublée… Cependant, je comprends tout à fait qu'il serait mieux si Wikidata pouvait afficher la forme féminine au lieu de la forme masculine (ce que font très bien les réutilisations de Wikidata, comme Wikipédia, car rappelons que Wikidata n'est pas vraiment fait pour être consultée telle quelle), et je m'oppose juste à des noms de propriété à rallonge…
    Je m'arrête là, car le débat semble clos vu la majorité pour. Je te remercie pour tes références et me renseignerai davantage sur la différence entre sexe et genre (et l'intérêt de tout genrer 😉). Bonne journée ! Tubezlob (🙋) 09:44, 28 May 2022 (UTC)[reply]
  •   Support. merci pour cette initiative. --Lewisiscrazy (talk) 21:18, 15 May 2022 (UTC)[reply]
  •   Support using only male form reinforces streotypes, and is a pace towards having more gender labels who are missing. Two is better than just one, and the masculine form is neither neutral nor universal. Nattes à chat (talk) 13:06, 16 May 2022 (UTC)[reply]
  •   Support Arpyia (talk) 15:47, 22 May 2022 (UTC)[reply]
  •   Oppose Labels shouldn’t contain “ou”. phab:T249447 is a better solution. --Thibaut (talk) 03:20, 3 September 2022 (UTC)[reply]
    @Thibaut120094: There is no rule saying that "ou" should not be used. sex or gender (P21) is a good example of the use of "ou".
    phab:T249447 could be a solution to avoid misgendering on item pages dedicated to women but there is no guarantee that it would be implemented at some point. Moreover phab:T249447 is not a solution for the label in item's pages dedicated to occupations and functions. PAC2 (talk) 17:33, 4 September 2022 (UTC)[reply]
  •   Oppose ibidem Thibaut120094 phab:T249447, proposals are bad tech design. Wikidata must indeed be thought through other languages than English, but this must be done in its development. We don't apply a bandage to treat an amputation, we do a graft. LD (talk) 04:03, 3 September 2022 (UTC)[reply]
  •   Support L'autre possibilité envisagée, imparfaite, a été formulée dans la requête Phabricator en 2020, elle n'est même pas encore assignée à ce jour. Autant dire que sa réalisation n'est pas pour demain, et de plus elle ne couvre pas le cas de l'affichage de l'item de wikidata lui-même. --Pa2chant.bis (talk) 07:50, 3 September 2022 (UTC)[reply]
  •   Oppose Labels shouldn’t contain “ou”. phab:T249447 is a better solution. - AvatarFR (talk) 09:55, 3 September 2022 (UTC)[reply]
  •   Support Generalized on properties since June without any major issue. phab:T249447 would be a bonus, but not sufficient by itself (cannot apply on property items themself) and we have no developper's feedback on this issue. The male form as generic form is ambiguous, especially for new contributors and should be avoided. — Metamorforme42 (talk) 12:01, 3 September 2022 (UTC)[reply]
  •   Oppose phab:T249447 is better. Şÿℵדαχ₮ɘɼɾ๏ʁ 14:01, 3 September 2022 (UTC)
  •   Oppose trop lourd, phab:T249447 est mieux. --Omnilaika02 (talk) 17:20, 3 September 2022 (UTC)[reply]
  •   Support Per Metamorforme42. Also, T249447 doesn't rely on lexicographical data, would be very complex to implement (not sure we want this kind of complexity in Mediawiki, maybe more in WikiFunctions), and would take years to happen. Cheers, — Envlh (talk) 10:32, 4 September 2022 (UTC)[reply]

Conclusion edit

First conclusion by PAC2 (challenged) edit

I count 10 votes in support of the proposition (  Support), 2   Neutral and 3   Oppose. I consider the proposition as adopted and I close the request for comments.

So for now in French :

Concerning properties whose target value is a human such author (P50), if the male form is different from the female form, it is recommended to use :
  • both male and female form if possible (ie "auteur ou autrice")
  • a verbal form if it's to complex to use female and male form.

Of course, the corollary is that the male form and the female form should be added as alias.

--PAC2 (talk) 04:49, 14 June 2022 (UTC)[reply]

WD:RFC: “Any uninvolved editor can close a request for comment”.
Please let someone else close the RfC.
btw, RFCs are not majority votes. Thibaut (talk) 02:51, 3 September 2022 (UTC)[reply]
Je veux bien m'occuper de faire la synthèse de cet appel à commentaires. Skimel (talk) 14:46, 5 September 2022 (UTC)[reply]

Conclusion (by Skimel) edit

Français :

Je compte 13 votes en faveur de la mesure (  Support), 8 opposés (  Oppose) et 2 votes neutres (  Neutral). Parmi les arguments avancés par les personnes en faveur de la mesure :

  • Rend Wikidata et les projets wikimédias plus inclusifs
  • Le masculin n'est pas universel ;
  • Cela permet de ne plus invisibiliser les femmes ;
  • C'est déjà mis en oeuvre dans plusieurs propriétés

Parmi les arguments avancés contre cette proposition :

  • "pas de raison de déroger aux règles de base du français" : il s'agit justement d'appliquer les règles du français en accordant correctement une profession selon son genre. Cet argument ne me semble pas opposé à la mesure.
  • Le masculin est neutre et englobe tous les genres : cet argument est contesté par plusieurs Wikidatistes, avec apport de sources scientifiques montrant que le masculin n'est pas véritablement neutre dans la perception commune du lectorat. En plus, Tubelzob semble avoir changé d'avis à la fin de la discussion. Cet argument ne me semble donc pas recevable dans la discussion.
  • Argument principalement mis en avant par plusieurs personnes (Thibaut, AvatarFR, SyntaxTerror, Omnilaika02) : c'est trop lourd, les labels ne devraient pas contenir le mot "ou" : il serait préférable de demander aux développeur.euse.s de Wikidata d'afficher un seul terme, en fonction du genre de la personne (notamment avec la propriété female form of label (P2521) si la personne est une femme). Le problème est posé depuis avril 2020 (plus de deux ans) sur Phabricator (T249447), sans qu'il n'y ait eu de réponse apportée.

Conclusion : il semble qu'un consensus se dégage en faveur de cette mesure. Son bien-fondé est assez largement accepté ; c'est plus sa mise en pratique technique qui soulève quelques oppositions. En attendant qu'une éventuelle solution technique jugée meilleure (comme celle proposée en T249447) soit mise en place, la proposition de PAC2 est acceptée. Skimel (talk) 15:12, 5 September 2022 (UTC)[reply]

English:

(Summary) There is quite a large consensus in favor of the propositon. Some opponents suggest that only the correct gendered form of the label be displayed, by using Sex or Gender (P21) and Female form of label (P2521). This has been asked on Phabricator (T249447) since April 2020 and has not been implemented since. Therefore, the proposition by PAC2 is accepted, until a better technical solution can be implemented. Skimel (talk) 15:20, 5 September 2022 (UTC)[reply]

Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata:Requests_for_comment/How_to_avoid_to_use_male_form_as_a_generic_form_in_property_labels_in_French%3F&oldid=1795753388"