Grammatical case in Pages

[Ilya Hoilik]

Hello!

This problem has been actual in 3.4 too. I thought it will fix in 4.0 but it is not.

When I create or edit database I can choose how items will be called in my language. For example in Russian language "article" translates as "статья". And this 4 strings are used in all place in Pages. It correct for English, but not for Russian (and also for some other languages like as Belarusian, Ukrainian, maybe Hungarian, Latin and other)...

Russian language is a very complex language and with Grammatical cases. And every noun have 6 grammatical cases. For example:

статья (nominative)
статью (genitive)
статье (dative)
статью (accusative)
статьёй (instrumental)
статье (prepositional)

Based on this, many string in Pages are used wrong. I can provide some examples:

Add new {article} - Добавить {статья} (Correct: Добавить {статью})

Create New Article - Создать новую {Статья} (Correct: Создать новую {статью})

{Article} actions - Действия со {Статья} (Correct: Действия со {статьёй})

Report {Article} - Жалоба на {Статья} (Correct: Жалоба на {статью})

Go to {article} list - Перейти к списку {статья} (Correct: Перейти к списку {статей})

1 {article} in this category - 1 {статья} в этой категории (It's correct)

2/3/4 {articles} in this category - 2/3/4 {статьи} в этой категории (It's correct too, but...)

5 {articles} in this category - 5 {статьи} в этой категории (Correct: 5 {статей} в этой категории)

I want to suggest the opportunity to provide custom word for each string which are currently used wrong system. This is a very important for us and I think it's doable task.

[Meddysong]

I'm glad you mentioned this, @IPSRussia. I was going to start something similar today based on a bug I'd posted here.

The concept of case is, I imagine, quite daunting, as is the Wiki article, so it might be helpful to the IPS staff if I give a little example using English.

Consider the following sentence:

I love them

What's the reciprocal version?

They love me

It's not I love they and it's not they love I (or me love they etc), even though we're making reference to exactly the same people. What's happening here is that we're changing the form of the pronoun based on its grammatical role in the sentence. I and me are exactly the same people but we instinctively know that only one of the two forms can be used in the sentences above and we can pick them out correctly with no problem. You could do so in the following example without even pausing to think about it: "They gave [I/me] a book for Christmas because [I/me] asked for one especially."

What we're doing without even being aware of the rules or the terms is acting on our knowledge that the pronoun I is to be used in certain situations (when we want the subject of a clause) but then when we're referring to the object (or following a preposition - for me, not for I) we use the pronoun me. In grammatical terms, we're using the nominative case when we choose I, but the accusative case when me.

In modern English we only have this case distinction now in (some of)  the pronouns (I/me, he/him, she/her, we/us, they/them), whereas we used to change the form of our nouns too. (We still have an entirely different case called the genitive which we use quite naturally dozens of times every day - it's the thing that you would recognise by "apostrophe-s", where boy becomes boy's, children becomes children's etc.)

For many, many languages case-marking is an every-day feature and extremely prominent, just as much as it is for us to have I/me and boy/boy's. Several (as set out in Russian above) have more examples than we have in English. Even something as simple to us in English as "Hi, {username}!" is not strictly accurate in some languages. In Croatian, for example, you wouldn't greet me as "Hi, Tim!" but "Zdravo, Time!", my name changing because the grammatical function in this instance is rendered by something called the vocative case rather than the nominative (ie the basic form, the I part of I/me).

The problem in Pages is not that we're asked to provide just the singular and plural but that we're limited to the singular and plural of the nominative case because this works fine in English (assuming you're not looking to distinguish I from me, which we wouldn't be in this instance).

Borrowing from the original post, here are instances where we would need to make a distinction in Pages:

Add new {article} - accusative ("I add him" not "I add he")

Create New Article - accusative

Report {Article} - accusative ("Report him" not "report he")

And that original post also gave examples where other cases would be needed too. I know it's far beyond practicable to say "OK, enter every word in every case" (Hungarian has 18 of them!) and even if you offered that anyway you'd be getting complaints that "such and such a language string uses X in the genitive but in my language it should be the accusative". We can cope outside of Pages by writing the translations of specific language strings ourselves. In my case, for example, I render __indefart_blog_entry_comment as "a comment [<- in accusative form] on a blog entry" because I have freedom to do that rather than the approach in Pages of just inserting automatically the singular form provided for X.

Even outside of case-marking there are problems caused by the approach in Pages, which is what my original bug report was about. Take the following example:

You can usually get away with placing two nouns side by side in English to create a new one. Article + Comments/Reviews -> Article Comments/Reviews. But that doesn't work in most other languages.

In my language, I can't just take the word Artikolo and place it next to Komentoj. I would either need to change Artikolo into a plural adjective to accord with Komentoj (so, Artikolaj Komentoj) or, ideally, combine them to create a new noun, Artikolo-komentoj. I'm fairly sure that the Germans are constrained like this too (you couldn't just write Artikel Komentare, could you, @daniel-son, but maybe Artikelkomentare?) as are the speakers of Scandinavian languages (am I right, @TSP?). The speakers of Slavic languages would need to, I think, change Comments to the genitive case, and the speakers of Romance languages (French, Spanish, Italian ...) would need an entirely new structure altogether, akin to "Comments on/about Articles".

The point that I'm trying to convey, albeit it in a wordy way, is that the approach taken within Pages works perfectly for English and is in that respect a good, efficient system, but that it unfortunately doesn't work for other languages. I suspect the only solution if IPS wishes to maintain its excellent approach to internationalisation, is to automatically generate a series of language strings if another language is installed where we're currently asked to provide singular and plural and ask us to complete them. Not ideal, but I and I think others in the same position would much rather spend 10 minutes filling in strings than have totally ungrammatical text on our sites.

 

[Marcher Technologies]

This is indeed the same issue as is present in 3.x..... in essence, if you want to provide the ability to truly translate a given phrase to any language, you cannot break words out of the phrases to be translated separately, the phrase as a whole needs to be translatable separately from other phrases in entirety, lest it make no grammatical sense upon translation.

[TSP]

In my language, I can't just take the word Artikolo and place it next to Komentoj. I would either need to change Artikolo into a plural adjective to accord with Komentoj (so, Artikolaj Komentoj) or, ideally, combine them to create a new noun, Artikolo-komentoj. I'm fairly sure that the Germans are constrained like this too (you couldn't just write Artikel Komentare, could you, @daniel-son, but maybe Artikelkomentare?) as are the speakers of Scandinavian languages (am I right, @TSP?). The speakers of Slavic languages would need to, I think, change Comments to the genitive case, and the speakers of Romance languages (French, Spanish, Italian ...) would need an entirely new structure altogether, akin to "Comments on/about Articles".

​In norwegian I would translate with "Artikkelkommentar" (Yes, norwegian is very similar to german). I may also choose to do "Artikkel-kommentar", but I think the former is more likely.

But right now I can't really see an apparent problem with "%s Comments" etc. for my translation. I'll say that I support how Marcher says it though. 

Personally I've never been to worried about these kind of specialized language-needs. I guess it's a combination of laziness, that norwegian doesn't have the same kind of complex rules and that I'm more worried all these kind of in-built support for special language cases will in some way inhibit on performance on communities that only really need one language. I'll rather have performance than language perfection. (But again, this could just be because I don't feel I have the same needs in norwegian) 

[Marcher Technologies]

I realize the above post was critical without providing constructive feedback, so am coming back to this to provide a viable solution.

I understand the desire to have these items easily editable separately for the laymen... so what you could opt to do is add a yes/no for 'advanced language support' on the database configuration 'Language' tab, that when yes, the noun fields are hidden and instead all the relevant strings are listed to be customized as translatable fields and saved as custom strings for the given database's use. This would allow the easy editing you wish to provide for the general end user in languages like english with looser grammar rules in regards to nouns, while providing the flexibility for correct grammatical support for power users using languages with stricter grammar rules.

Personally I've never been to worried about these kind of specialized language-needs. I guess it's a combination of laziness, that norwegian doesn't have the same kind of complex rules and that I'm more worried all these kind of in-built support for special language cases will in some way inhibit on performance on communities that only really need one language. I'll rather have performance than language perfection. (But again, this could just be because I don't feel I have the same needs in norwegian) 

Considering that with the words seperated from the strings they are used in, the fully formed strings are being built at runtime, thus using such an advanced language option would actually be more resource efficient, to note.

[Ilya Hoilik]

IPS, are you try to solve this problem?

[Mark]

Providing full support for translation into other languages is an important feature for us. If you come across any cases where you are unable to translate something, please file a bug report. Please be as specific as possible - the first post in this topic is excellent, but sometimes people are a little more vague and since we don't speak the language in question, we're not always sure what needs to be done to fix it.

Pages is a particularly difficult case and an exception may need to be made at least for the first release of 4.0, but if you file a bug report with the details, we will certainly look into it and do our best.