[OT] Agreement and i18n (was RE: Language name questions)

From: Marco Cimarosti (marco.cimarosti@essetre.it)
Date: Tue May 28 2002 - 08:39:22 EDT


Misha Wolf wrote:
> Then comes the interesting question: What do we mean when we write
> "English" in a language selection menu on a Web page. The possible
> meanings include:
>
> - "English language" -- "slovenský jazyk" / "sloven?tina"
>
> - "in English" -- "slovensky"
>
> - "English Web page" - "slovenský", "slovenská" or "slovenské",
> depending on the gender of "Web page".
>
> I would welcome comments from Petra and others.

This reminds me of a very common i18n pitfall.

In some languages, grammatical agreement is much more common than in
English: e.g. in a number of languages adjectives must take up the gender,
number and case of the noun they refer to.

When a command or menu item is expressed with an adjective it comes natural
for English speakers to allow it to refer to a variety of objects, which are
indicated by a variety of nouns.

The problem is that, translating in a different language, the nouns which
indicates the various objects can have different genders and/or numbers and,
thus, call for different forms of the adjective.

A very well-known example of this situation is the menu "New" of Windows
Explorer (the program that is used to manipulate the file system under MS
Windows systems).

This menu has this form:

        &File
                &New
                        &Folder
                        &Shortcut
                        ...
                        Text document
                        Word document
                        ...

When translated in Italian, it becomes:

        &File
                &Nuovo
                        &Cartella
                        &Collegamento
                        ...
                        Documento di testo
                        Documento Word
                        ...

The adjective "nuovo" is in its masculine singular form, so it is OK for
masculine noun such as "collegamento" or "documento", but it is incorrect
when referred to the feminine noun "cartella": one should say "nuova
cartella", not "*nuovo cartella".

Another example is when i18n engineers assign a single slot to an indicator
such as "updated", which may be applied to a variety of items on a list:
when l10n engineers fill in the slot with the Italian translation, they have
to choose an arbitrary gender and number for the word "updated" (such as
"aggiornato", masculine singular), which will work OK for some items (such
as a "documento", masculine), but not for others (such as "immagine",
feminine singular, which would need "aggiornata"; or "impostazioni",
feminine plural, which would need "aggiornate").

Also defining user messages such as "%s is running" or "%s halted" is
calling for troubles: in Arabic, when the placeholder "%s" expands to a
noun, the verbal form should change according to the gender of the noun.

Even an apparently innocent message like "%d %s" (that should expand to a
number and a noun, such as in "120 files") is calling for troubles: in
Chinese, between numbers and nouns there must be a particle called "measure
word", and different nouns require different measure words. (Not to say
that, even in English, you cannot say "1 files").

Ciao.
Marco



This archive was generated by hypermail 2.1.2 : Tue May 28 2002 - 07:34:46 EDT