Page MenuHomePhabricator

Add the neuter gender
Closed, DeclinedPublic

Description

Currently, we have gender support for

  • the male gender
  • the female gender
  • gender unknown (although caguely attributed "neutral" in the code sometimes)

Many germanic languages, including English, require the neuter gender at times, such as when having to make a distinction between:

  • (Missis Smith) "she"
  • (Mister Miller) "he"
  • (a bot program, a person nicknamed by a word of neuter gender) "it"

Gender agnostic wordings such as "user xyzzy" fits them all and is not outright
wrong in the neuter gender case, but having the correct grammatical gender at
hand even in these comaparatively rare cases will be much better language use.


Version: 1.18.x
Severity: enhancement
See Also:
T61643: Provide further options than only binary gender

Details

Reference
bz27744

Event Timeline

bzimport raised the priority of this task from to Normal.Nov 21 2014, 11:25 PM
bzimport set Reference to bz27744.
bzimport added a subscriber: Unknown Object (MLST).

related Bug 25834

Would this only be used for bots?

(In reply to comment #2)

Would this only be used for bots?

There are people that don't identify to a gender (or to the normal gender identities).

Which is what we try to keep our unspecified/undisclosed as (or at least should be doing)

(In reply to comment #2)

Would this only be used for bots?

No.
As said above, Nicknames of neuter grammatical gender may require it, too. This includes, for instance all diminiutives of names such as "Maria" (Mary, feminine) -> "Mariechen" (little Mary, neuter) for all local languages of the western part of the West German continuum from Swiss German over Swabian, Alsation, Palatinean, Ripuarian, (East) Limburgish to Low German (including Westphalian)

Usage is depending on language. It would not be used for bots having Names of non-neuter grammatical gender in Colognian, e.g.

Off-Topic:
As a side-note: it may be useful to make user "right" or group or status information available to message handling occasionally. We translate a user name to the string "the user named 'xyzzy'" in cases of unknown gender, since articles must be used with names standing alone, and there is no genderless article. It would of course be both clearer and more informative to use "the bot named 'xyzzybot'" instead. With some log messages, we would have take care to keep the status at the time of logging, then.

Another side-note: In order to get nicknaming with grammatically gendered words right, we would need to know both grammatical and natural gender. For example, a male boxer nicknamed "the Ape" ("Ape" is feminine as per grammatical gender) hit his father in law K.O. -> newspaper headline: "Die Aap hät singe Schwejervatter verkammesölt!" -> Grammatical analysis: "The Ape (feminine) is referred to with "his" (masculine) -> The nick's article goes the nicknames grammatical gender while references to the person otherwise go with the natural gender.

Yet another side note: It is not uncommon, that people use Nicknames that include articles, such as ":w:de:User:Der Hexer" or ":w:de:User:De Uli". For languages that require articles to be used with names, we have those articles in messages files of course. At least for the langues I know, adding Articles should be avoided, when names already contain ones.

(In reply to comment #3)

There are people that don't identify to a gender (or to the normal gender
identities).
Which is what we try to keep our unspecified/undisclosed as (or at least should
be doing)

Not to select a gender, whatever the cause, is possible.
To the best of my knowledge "unspecified" is not posing any grammatical or other
problems in the context of multiple language support (i18n).

Another context requiring the neuter gender surfaced in a radio interview with one of 15 members of the German Pirate Party, who were recently elected into the Berlin federal state parliament. The interviewed person is well known with his nick Plaetzchen (cookie) and confessed that he likes eating cookies, hence the name. Asked whether he wanted to be addressed "Mr. Cookie" or just "cookie" he said "das Plaetzchen" (the cookie) - "das" being the neuter article in German grammar.

Standard German does not put grammatical articles in front of names usually, but many regional and local varieties do. Having articles in front of names of processes, jobs, or bots, and some nick names, is at least established jargon, dispersing into standard language use through computer magazines, and other press. Grammatical articles always convey grammatical gender. German has three of them.

Analysis:

We have:

  • natural gender: "male", "female", "undisclosed", ("other" being discussed)

We need in addition:

  • grammatical gender: "masculine", "feminine", "neuter"
  • use article before name: "yes", "no"

so as to be able to completely map German language use in our German L10n.

Btw. the German Pirate Party is running a MediaWiki wiki. Das Plätzchen is of course using it.

Interesting story. Adding neuter gender (different from unspecified) is probably the next step we will do at some point.

undisclosed (what unspecified is now called) should already be a gender neutral for the languages that support it...

(In reply to comment #8)

undisclosed (what unspecified is now called) should already be a gender neutral
for the languages that support it...

Not really, many languages default to masculine form instead of grammatical neuter.

(In reply to comment #8)

undisclosed (what unspecified is now called) should already be a gender neutral
for the languages that support it...

That is not so. For most Germanic languages that I am aware of, the actual use is not compatible with the neuter Gender. You need to be aware that quite often, grammatical and natural gender deviate from each other, and while the natural gender governs the genderized use of very many languages in the world, this is not so for a large number of Indo-European languages, most notably Germanic ones.

In addition to what Niklas Laxström wrote, we also have at least one language that does not have a true default, and thus resorts to constructs like "the user named X..." which does not imply a gender. While this fits for neuter, too, it is definitely not the way you would talk about a neuter.

Without wanting to seem insensible to those of undetermined gender, wanted or unwanted, I am closing this issue as WONTFIX.

(In reply to comment #11)

Without wanting to seem insensible to those of undetermined gender, wanted or
unwanted, I am closing this issue as WONTFIX.

? Usually things get wontfixed because
a) It is infeasible to do. That does not appear to be the case.
b) It is a "bad idea". The comments so far don't really seem to indicate that that is the case

So why the wontfix?

  • Bug 59643 has been marked as a duplicate of this bug. ***

As said above, Nicknames of neuter grammatical gender may require it, too. This includes, for instance all diminiutives of names such as "Maria" (Mary, feminine) -> "Mariechen" (little Mary, neuter) for all local languages of the western part of the West German continuum from Swiss German over Swabian, Alsation, Palatinean, Ripuarian, (East) Limburgish to Low German (including Westphalian)

In Alemannic (gsw, which includes Swiss German, Swabian and Alsatian), the neuter gender is even more common than that. In many traditional varieties, the neuter gender is the normal way for referring to women. This does not depend on the presence of a diminutive ending. While this usage may still be found in rural areas, the situation may be more complex in more urbanized areas.

For example, in modern urban Bernese, women's names still take the neuter article when they are altered with the ending -i, but they take the feminine article in other cases, e.g. ds Natti (with the neuter article ds) vs. d Nathalie (with the feminine article d). While the ending -i may be considered a special kind of diminutive ending (restricted to names), it does not trigger the neuter article for men's names, e.g. der Köbi from der Jakob (with the masculine article der).

You will easily find sources for this in the dialect grammars, e.g. in: Werner Marti (1985): Berndeutsch-Grammatik. Bern: Francke, p. 81 §3.1.5.2.

Standard German does not put grammatical articles in front of names usually, but many regional and local varieties do. Having articles in front of names of processes, jobs, or bots, and some nick names, is at least established jargon, dispersing into standard language use through computer magazines, and other press. Grammatical articles always convey grammatical gender. German has three of them.

More precisely, it is only the Northern varieties of modern German that use no article in front of names, whereas the article is common in Central and Southern varieties, see Artikel+Vorname « atlas-alltagssprache and Artikel+Nachname « atlas-alltagssprache. The reason why German is commonly said to have no articles in front of names is only that Northern German has the most prestige (probably due to the raise of Prussia to the dominating state of Germany in the late 19th century).

It is therefore conceivable that the de-AT and de-CH localizations might introduce definite articles in front of names in the future. The Alemannic (gsw) localization certainly requires definite articles in front of names, so all messages that contain a username should check it for {{GENDER:}}.

But a problem arises: we only have one preference for users of any language. How to make an option such as neuter work, when it has such a wild variety of meanings across languages?

the neuter gender is the normal way for referring to women.

Then use neuter in the option for "female-addressed users", i.e. the second position? When the preference is offered we don't promise a specific grammatical category but an example, if you use neuter in the example then it's fine. Would some users remain uncovered?

But a problem arises: we only have one preference for users of any language. How to make an option such as neuter work, when it has such a wild variety of meanings across languages?

I imagine a solution where the number of options depends on the language:

  • A language with a three-way personal gender distinction (like Alemannic) would offer all three choices (plus the unspecified choice),
  • a language with a two-way personal gender distinction (like English) would offer both choices (plus the unspecified choice), and
  • a language without any personal gender distinctions (like Finnish) would drop the choice entirely.

I bet that there are languages with more than three possibilities.

The current choice (masculine – feminine – unspecified) is probably just another example of the bias towards English that is natural on the internet. With the English system in mind, you might wrongly assume that what is really a grammatical question (use of pronouns etc.) would be a question of natural gender (female or male).

What do we really want to know? As I understand it, we are not at all interested in the natural gender of the users (that is rather an interest pursued by commercial data miners). Instead, we want to know how to refer to the users in a grammatically correct way.

the neuter gender is the normal way for referring to women.

Then use neuter in the option for "female-addressed users", i.e. the second position? When the preference is offered we don't promise a specific grammatical category but an example, if you use neuter in the example then it's fine. Would some users remain uncovered?

Why, yes, the users who prefer to be referred to in the feminine gender would remain uncovered.

I imagine a solution where the number of options depends on the language

Our preferences system doesn't provide such a possibility.

Why, yes, the users who prefer to be referred to in the feminine gender would remain uncovered.

And who would be those?

I imagine a solution where the number of options depends on the language

Our preferences system doesn't provide such a possibility.

That's what I thought.

Why, yes, the users who prefer to be referred to in the feminine gender would remain uncovered.

And who would be those?

Most women. Some may prefer to be referred in the traditional way using the neuter grammatical gender. We cannot know in advance. The language's grammar provides two options (neuter grammatical gender or feminine grammatical gender). We only provide one of the options (feminine grammatical gender). Who are we to decide in advance for the users?

I see now that the thread T61643 is really more adequate to this issue. This does not concern exclusively the neuter grammatical gender, but may concern a wider array of grammatical questions. I will elaborate there.

I bet that there are languages with more than three possibilities.

In Colognian, we have 5 possibilities:

  • masculine
  • feminine with he standard addressing that resembles neuter
  • unknown
  • feminine with the uncommon feminine addressing
  • neuter addressing for neuter nicks, bots, etc.

All our translations are already taking care of these of 5 possibilities, even though at the moment only 3 of them are rendered.

I pledge for language specific GENDER handling. I know that there are other languages having more that 2 (or 3) grammatical genders.

Most women. Some may prefer to be referred in the traditional way using the neuter grammatical gender.

But you earlier said "the neuter gender is the normal way", I thought "normal" meant most common. Anyway, if it's a matter of style then this can be a further language variant, gsw-formal. That would be rather easy to do.

Most women. Some may prefer to be referred in the traditional way using the neuter grammatical gender.

But you earlier said "the neuter gender is the normal way", I thought "normal" meant most common.

That is clearly not what I said. That is only what you quoted. What I said was: "In many traditional varieties, the neuter gender is the normal way for referring to women." The qualifier "In many traditional varieties" is significant, as I went on to explain. In contemporary varieties of Alemannic, things are different and all three grammatical genders may be used when referring to people.

Sidenote: When discussing this issue with friends recently, they told me a case I had forgotten: There are other varieties of Alemannic where the neuter grammatical gender is often – but not always – used for men's nicknames.

Anyway, if it's a matter of style then this can be a further language variant, gsw-formal. That would be rather easy to do.

It is not a matter of style. It is a matter of preference that should be left to the user to decide upon. Exactly like the choice between masculine grammatical gender and feminine grammatical gender which is not a matter of style either.

Terminology is not helping us here. :( Practically speaking, why would a language variant not work as an option?

Practically speaking, why would a language variant not work as an option?

To have a more accessible example (it seems to me you are no expert in Alemannic), consider English: Could we achieve proper grammatical gender with a language variant? We would then have, for instance, the normal "en" version where users are referred to as "she", and a special "en-x-masculine" version where users are referred to as "he" (or the other way round). This would be very unsatisfactory for a number of practical reasons:

  • You would not have any control on how a message such as MediaWiki:Flow-notification-mention-email-batch-body/en appears to other users. The form in which it appears to another user – either: "Nemo_bis mentioned you in her post in '$2' on $3" or: "Nemo_bis mentioned you in his post in '$2' on $3" – would depend entirely on the other user's settings.
  • The setting you choose for yourself would invariably apply to all other users. The only choice would be between, for instance, a Wikipedia where all users invariably appear in the feminine grammatical gender or a Wikipedia where all users invariably appear in the masculine grammatical gender.
  • Making one of the grammatical genders the default choice (plain "en" instead of "en-{something}") is necessarily a controversial decision.

A language variant is neither necessary, nor useful in most cases.

If we want to improve on grammatical gender, we would have to make GENDER choices language depend. We already have completely genderless languages if I am not mistaken, and why not have in similar to PLURAL with varying choices for a bunch of dissimilar languages?