@Mujahids proposal is easily the best.
I am a Western learner; I come for the user group some of the answers are attempting to accommodate.
I am not fond of duplicating tags for the sole purpose of accomodating to my user group, some reasons are:
- duplicate content
- duplicate abstracts
- double the work to manage
- double the work to moderate
If this site gets out of commitment it will already have a number of moderation challenges to overcome. We don't need to add more by applying duplicate tag management to those moderators.
Instead when a user like myself enters an English tag it should actually link directly to the Arabic edition when posted. The abstract of that tag should then have both Arabic and English editions. Most likely the abstract displayed when hovering over being English.
This has the benefit that someone can put in Pronounciation and get back: الصوتيات
(the parent tag might be definite like that, not sure, I guess that is another dicussion for another time)
This means that they can also search for the English "Pronounciation" and Arabic "Phonics" (I assume that the two would be merged since they are inter-changeable) equally, for example: http://stackoverflow.com/tags/mongodb/synonymshttps://stackoverflow.com/tags/mongodb/synonyms one tag I frequent on SO. If you click its synonyms you will get the parent tag.
This also has the benefit that I do not need to try and use rtl Arabic script on a ltr UK keyboard and OS everytime I want to search for some Arabic questions. Instead I can just put in one of the English synonyms and get back the Arabic parent tag straight away.
This also provides the benefit that naive Arabic speakers who are willing to help learners and natives alike do not need to have two question lists open at the same time. They can get their feed all in one list, making it much easier for them to answer and keep ontop of their own activity, etc etc.
So yeah, @Mujahids has got my vote currently.
In reference to tag creation number: SO, whose tag count is stupidly high, is currently sat on 40,572. I would think that duplicating Arabic and English tags would most likely result in far less than this number, as such I think it is okay to assume we can go through with @F.Julian's idea.