[Ltru] Re: Scottish English (was: LANGUAGE SUBTAG REGISTRATION FORM)

[Addison Phillips]

[removing cross-post to ltru, which I think not yet appropriate]

> 
> Keep in mind: This is not abstract. I have a specific business need and 
> that semantic is best represented by a region. As previously noted, I'll 
> accept a classification of Standard Scottish English, but I don't think 
> this will be as accurate for what I'm trying to classify as "Scottish 
> English." A region tag would cover any Scottish dialect of English found 
> in the film and that's the reality of this situation.  

Except... it's not a "region" subtag. Instead, you intend to indicate a 
regional variation or collection of variations of a language, to wit, an 
accent and distinctive regional locution.

> 
> Films flagged with the Scottish English tag would be more likely 
> candidates for subtitles or redubbing than other English variants, and 
> the audience for these films would be different. 

I don't doubt this business case nor the need you've identified for it. 
The question is whether this requirement extends to other users. 
Sometimes private use tags are a better solution than trying to create a 
general purpose subtag.


> 
> To be honest, with so much documentation on the Web, a real-life use 
> case, and a set of ears, I can't believe there is so much controversy 
> over the legitimacy of my request. 

The question is whether what you seek to identify is a real language 
variation, in the sense that we can say it is distinct enough to warrant 
its own subtag. Again, I don't doubt the authenticity of your request 
and am inclined to support it. But it does open the question as to what 
level of detail we should register variants for. Given the profusion of 
subtags for the relatively limited (in terms of range and speaker 
population) language of Slovenian, one suspects that there is 
substantial latitude for registration for English variations.

OTOH, there was quite a long thread this year on this list about the 
practical limits of language tags. They cannot be expected to identify 
every distinct variation in language. In most cases we are all satisfied 
to say that "en" is English... and sometimes add a closer regional 
variation. But truly detailed linguistic variation begins to test the 
capabilities of tags. In the case of "en-scottish", I can see the 
justification, but am concerned about where the line gets drawn.

> I would think the controversy would 
> be over the best way to represent this linguistic entity as there's no 
> way other than a variant tag in RFC 4646 and it seems like there should 
> be. 

Why? Variants cover whatever the predefined subtags do not. What 
specific need is there that variants don't address?

> I will need to employ a private use tag if this request is not 
> granted and whatever the resulting tag is, I will need to encourage this 
> use throughout the entertainment industry. If spelling matters in 
> written contexts, surely accents matter in spoken contexts -- even if no 
> differences in word choice existed.

Yes. And private use might be a good solution in some cases (and not in 
others).

> 
> English is a language and Scotland is a well-defined region. There is no 
> controversy about these two statements -- why is there so much 
> controversy over putting the two together? Isn't this what the variant 
> tag is for?
> 

Scotland is a well-defined region. Alabama is a well-defined region. San 
Jose, California, is a well-defined region. The question is whether 
these describe useful language variations. I agree with Randy that the 
lack of a Scotland region is marginally surprising. But not much more 
than the lack of a "southern USA" region is. One may remark upon local 
dialects or accents (I have certainly abused Mark Twain's Puddinhead 
Wilson as an example in this regard) and still find that the most 
appropriate available tag is of a "higher order" (such as en-US).

This is not necessarily a weakness. It is just a measure of the overall 
granularity of language tags. I do think that 'scottish' is a reasonable 
subtag to register (we have one for Oxford spelling and Boontling, for 
heaven's sake). But there is also precedent on this list for not 
registering "practical" subtags that do not represent well-documented 
language variations (es-americas).

Addison

-- 
Addison Phillips
Globalization Architect -- Yahoo! Inc.
Chair -- W3C Internationalization Core WG

Internationalization is an architecture.
It is not a feature.