Scotland (was: Re: Add Likely Subtags first step)

Doug Ewell doug at
Sun Jan 25 23:18:09 CET 2015

Sorry, I should have read Philip's earlier post more carefully.

Language tags are always intended to denote a language or variety 
thereof. A "Scottish brand of English," whether distinguished by accent 
or vocabulary or spelling or whatever, is a variety, just as the writing 
system used to write that language is a variety. Saying that the speaker 
happens to reside in Scotland, or happens to be located in Scotland at 
the time of speaking, is not a language variety.

Region subtags have been used for a relatively long time for coarse 
distinctions like "American English" and "British English." But there 
has always been a very fuzzy line separating useful combinations from 
all the rest. In the case of "en-US" versus "en-GB", the tagger might be 
trying to specify the accent, but they're just as likely to be 
indicating differences in spelling (colour, realise, traveller) or 
grammar (the company are...) or some other aspect. There are a great 
many language-region combinations, like "ca-BT", where it's basically 
impossible to figure out what distinction the tagger had in mind.

To answer the question of how to tag "English as spoken in Scotland," it 
would probably help to ask oneself which distinction(s) one is trying to 

Doug Ewell | Thornton, CO, USA | ­

-----Original Message----- 
From: Doug Ewell
Sent: Sunday, January 25, 2015 13:11
To: Philip Newton ; Reece Dunn
Cc: ietf-languages at ; John Cowan
Subject: Re: Add Likely Subtags first step

We have a variant for exactly this purpose:

Type: variant
Subtag: scotland
Description: Scottish Standard English
Added: 2007-08-31
Prefix: en

Doug Ewell | Thornton, CO, USA | ­

-----Original Message----- 
From: Philip Newton
Sent: Sunday, January 25, 2015 12:52
To: Reece Dunn
Cc: ietflang IETF Languages Discussion ; John Cowan
Subject: Re: Add Likely Subtags first step

On 25 January 2015 at 20:22, Reece Dunn <msclrhd at> wrote:
> On 25 January 2015 at 18:58, Philip Newton <philip.newton at> 
> wrote:
> So how would you encode "English as spoken in Scotland"? Something
> like en-GB-x-scottish?

Something like that. There’s no registered subtag for that any more
than there is for ‘English as spoken in Wales’ or ‘English as spoken
in Didcot’.

Even better, though, would be ‘en-XS’, using a region subtag from the
range ‘XA–XZ’ reserved for private use, if sender and receive agree
beforehand that ‘XS’ will mean ‘Scotland’ for the purposes of their
communication; this subtag is unambiguously a region subtag, whereas
‘-x-scottish’ could be anything. See also section 4.6 ‘Considerations
for Private Use Subtags’ in RFC 5646.

(If you did want to use a private-use subtag, then ‘en-x-scotland’ or
‘en-x-scottish’ might be better, using it as a private region subtag,
rather than as a private variant subtag as in your example

>> Similarly with ‘de-(de-)1996’, which cannot mean ‘German as spoken in
>> 1996, or as written in that year in any orthography’, because that is
>> not what the variant subtag was registered as meaning. And
>> ‘uz-baku1926’ cannot mean ‘Uzbek as spoken in Baku in 1926’, nor can
>> ‘ja-Latn-hepburn’ mean ‘variety of Japanese written in the Latin
>> alphabet in Hepburn, Iowa’.
> What about en-US (American English dialect), en-GB (British English
> dialect), en-IE (Irish English accent) and others?

Here, ‘US, GB, IE’ are region subtags, not variant subtags, so you’re
comparing apples and oranges.

> I would imagine a better example would be the variety of English
> spoken in Oxfordshire.

That would also seem more likely to me than the speech of just one city.


More information about the Ietf-languages mailing list