ID for language-invariant strings

John Cowan cowan at
Fri Mar 14 19:49:24 CET 2008

Peter Constable scripsit:

> > Except when they do:  the tag "yue" is simply the Mandarin name for
> > Cantonese.
> No, I very much disagree: language identifiers/tags are non-linguistic
> strings. In the particular case of "yue", it *incidentally* happens
> to take the same form as one of the linguistic names for the concept
> being referenced; but across the board the alpha-3 IDs in ISO 639 are
> non-linguistic, symbolic IDs.

Quite so.  But what is incidental or accidental, and what is essential
use of a language?  That is the distinction (at any rate, when it purports
to be a fixed one) that I am at pains to discredit.

> No, they are *not* completely arbitrary tokens. We would never, for
> instance, assign a reference name of "LANG QX13PB6". We want them
> to be linguistic in nature because we want humans referring to the
> code table to be able to cross-reference them to other documentation
> indicating the intended semantic if not be able to recognize directly
> the intended semantic.

How would this ability be impeded if they were arbitrary?  Indeed, the
tag alone serves that function.  The fact that "German" has been used
rather broadly in English is nothing but a distraction from its use as
the reference name for de, which denotes Standard German.

> Yes, I see the point you're making. And the Cobol example has a lot
> of similarity to the application scenarios I've described. But I'm not
> fully convinced.

"The man convinced against his will / Is of the same opinion still."
Take your time.  :-)

In politics, obedience and support      John Cowan <cowan at>
are the same thing.  --Hannah Arendt

More information about the Ietf-languages mailing list