Re: [geolang-comment] ISSUE 3: The "language" published subject

[Lars Marius Garshol <larsga@garshol.priv.no>]

* Lars Marius Garshol
| 
| Ouch! Do you have a reference on the problems with codes that appear
| to represent single languages but don't?

* John Cowan
| 
| Well, it also has to do with the purposes for which one defines
| "language".  For example, QUE (Quechua) is listed as a single
| language by ISO, and for bibliographic purposes it probably is.  But
| as far as the criterion of mutual intelligibility goes, there are
| about 30 Quechuan languages, closely related but by no means
| mutually intelligible.

I think the criterion must be bibliographic, then, because Bokmål
(Norwegian) and Nynorsk (the other Norwegian) are listed separately,
and they are most definitely mutually intelligible.

| ISO does not seem to have used a single standard at all.

Unless it's what's likely to be useful for bibliographic purposes.
 
* Lars Marius Garshol
|
| (1) attempt to create a workable definition of "language",
 
* John Cowan
|
| Hopeless.  The definition, as indicated above, depends on the
| purpose.

I pretty much agree.
 
* Lars Marius Garshol
|
| (2) define "language" extensionally by enumerating all members of the
|     class, 
 
* John Cowan
|
| This breaks down because "a language" is not a sharp notion, any
| more than "a cloud" is? 

Well, true, but I meant that we could enumerate the ISO 639 codes that
represented single languages and define "language" as the class of
those languages.

| Yet given a similar dialect continuum within Norway, and two
| distinct written standards, there is still an ISO code for
| "Norwegian" (as well as codes for "Bokmal" and "Nynorsk" as well).

Actually, those two written languages are definitely mutually
intelligible. The spoken dialects can be borderline at times, but
there's no question that the written languages are MI.
 
* Lars Marius Garshol
|
| (3) make a simple definition that is uncontroversial because of its
|     vagueness, like "the notion of 'language'; the class of all
|     'languages'", and
 
* John Cowan
|
| Probably the best we can do.

I am leaning in this direction, too.  Now that I have a copy of ISO
639 (and, thanks to an ISO screw-up, ISO 638) I can see that it
doesn't actually define "language" at all or do anything even remotely
like it. It does say something about what its criteria are, but they
are very very vague.
 
* Lars Marius Garshol
|
| (4) use somebody else's definition.
 
* John Cowan
|
| Whose?  There are too many candidates.

Well, this would essentially be a more devious form of (3), where we
take somebody else's definition (Encyclopedia Britannica's, or
WordNet's, or...) because it sounds better than the more naive (3)
approach, but it won't really achieve that much more.

-- 
Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
ISO SC34/WG3, OASIS GeoLang TC        <URL: http://www.garshol.priv.no >