Adding variant subtag 'erzgeb' for Erzgebirgisch (was: Requests that have been on hold)

[Randy Presuhn]

Hi -

> From: "Thomas Goldammer" <thogol at googlemail.com>
> To: "CE Whitehead" <cewcathar at hotmail.com>; <ietf-languages at iana.org>
> Sent: Sunday, August 09, 2009 10:58 PM
> Subject: Re: Adding variant subtag 'erzgeb' for Erzgebirgisch (was: Requests that have been on hold)
...
> well, I prefer vmf, and here is why:

Then that request should be our starting point.

> Ethnologue 16 has Erzgebirgisch as sxu dialect (
> http://www.ethnologue.com/show_language.asp?code=sxu ), but there are
> so many things screwed up in there (Hessian is now also listed as a
> dialect of sxu...).

German dialects tend to be regions in a multi-dimensional continuum,
with arguments for classifying something as A, B, or C depending on
just where one looks within that dialect's area and whether one looks
at lexicon, phonology, or grammar.

> I know there are not really decent publications on
> the classification of Erzgebirgisch, and there is no work going on, as
> far as I know, to change that - I do want to do that work by myself
> someday, but I'm bound to another project probably the next one or
> even two years.

This could be understood as an argument for considering the assignment of
a variant subtag to be premature at this time.

> But I wonder how difficult a process it is to change the tag then, if
> you decide now to hang it on sxu and some linguistic work perhaps
> eventually shows that it rather belongs to vmf (which I'm quite
> convinced of, not speaking as a speaker of that variety but as a
> linguist). So I do prefer vmf, although that's going against
> Ethnologue. But since Ethnologue does not give linguistically
> retraceable sources - just because there are none that really go
> beyond guessing - , I don't really feel bad with that, honestly. (And
> as you all know, acceptable sources for any kind of subgrouping within
> a group of related languages or dialects are such that use all the
> nice state-of-the-art methods of historical and comparative
> linguistics, which is basically gathering shared innovations all over
> the language system and stuff like that.)

There'd be two tags.  One is used to label the erzgeb variant of sxu, the
other is used to label the erzgeb variant of vmf.  It's conceivable that
both might be valid, referring to different subsets of what is collectively
called Erzgebirgisch.  (I'm *not* arguing whether this is or is not the
case with Erzgebirgisch; I'm just saying that this kind of situation is
entirely conceivable with German dialects, particularly when there are
overlays due to migration.)

> On the possibility to tag it for both: I guess the tags are supposed
> to have some relation to the actual language genealogy. And I'm not a
> person who thinks that a language or dialect can really derive
> genetically from two separate languages/dialects. Another
> language/dialect can influence it so that it eventually ends up having
> sort of more properties from that contact language/dialect, but still,
> the methods we have got to figure out subgroupings will still show the
> inheritance from somewhere else - like it is in the case of English
> which doesn't really look like Western Germanic, after all.
...

This is a good summary of why the registry is not structured in terms of
genealogy.  :-)   For those same reasons, I wouldn't obsess about
"correcting" a situation where X is tagged as a variant of Y, even after
research reveals that X is really a variant of Z.

There is not a *requirement* for the tags to reflect genealogy.
Though it *can* be helpful, I think the current state of a language
and what other language it most closely resembles are more useful
when we're constructing tags in terms of "X is a variant of Y".

Randy