Counting Heads

Addison Phillips [wM] aphillips at
Wed May 28 13:00:18 CEST 2003

I agree, but it violates the hypothesis I'm making about subsidiary tags
(which, of course, could be balderdash). IOW, you're treating zh-hakka *as*
a language (or near enough to a language), so the tag is really:

  lang = zh-hakka
  script = han(x)
  ortho = null

Rather than:

  lang = zh
  ortho = hakka
  script = han(x)

There are lots of tags that represent dialectical differences like this. The
problem is recognizing the "tipping point" when the dialect becomes more
important than the script, enough so that it warrants an exception to the
pattern. I mean, I'd guess you wouldn't have a problem with, say,
en-latn-boont? But if zh-hakka is on the other side of the divide we ought
to define the divide (and that may not be possible to do with any more
precision than a "duck test").


> -----Original Message-----
> From: John Cowan [mailto:jcowan at]
> Sent: Wednesday, May 28, 2003 11:47 AM
> To: Addison Phillips
> Cc: ietf-languages at
> Subject: Re: Counting Heads
> Addison Phillips scripsit:
> > I agree generally with Doug and John that script is far
> > more important than, say, spelling or dialect in most cases, but I also
> > suspect that at least a couple of counter examples can probably
> be cited
> > (zh-hakka has already been cited as possibly one such). Languages are
> > too messy for that not to happen!
> I don't think it is a counterexample, because "zh" really means "Sinitic
> languages" and "zh-hakka" is a particular Sinitic language.  When _Hakka_
> in particular is written (as opposed to writing Standard Mandarin in
> either hant or hans and then reading the characters with Hakka
> pronunciations)
> then script (latn vs. hans vs. hant) is surely subordinate to Hakka-ness.
> --
> John Cowan
> jcowan at
> All "isms" should be "wasms".   --Abbie

More information about the Ietf-languages mailing list