On 09/13/2000 01:39:37 AM J%ORG KNAPPEN wrote:
>I once looked at the ethnologue and its subdivision of the german language
>is just ridiculous. Not small errors, a gross misconception. I don't trust
>the ethnologue in area where I don't know the fact well, since it fails in
one
>area where I know them.
I'm not a specialist on German, and so can't comment on those details. I
will say, though, that there will always be someone with this kind of
complaint on *any* enumeration of languages simply because different people
apply different operational definitions. That is one of the key issues Gary
and I discussed in our paper, and it is for that reason that we suggested
distinct namespaces that match different operational definitions. Even if
the Ethnologue fails in this spot under its own definitions (not a point
I'm conceding), it is a fallacy to conclude that the remainder is invalid.
Just because you don't trust it because you're not happy with the one piece
you have opinions on doesn't mean that it isn't useful for a lot of other
users that do want to use it for other areas. It is just as easy to point
to problems with ISO 639-x (actually easier, I think), but I'm not trying
to keep people from using that if it serves their purposes. The point isn't
to flog ISO 639-x but rather to say we need to get serious about moving
forward on providing identifiers for the thousands of other languages
people are interested in.
The 13th edn. of the Ethnologue lists a total of 15 lanuages under the
classification Indo-European/Germanic/West/Continental/High. I don't think
you'd say that *all* of these are wrong; and for the handful that may have
some problems associated, they don't eliminate the utility for people
interested in the thousands of other languages. As for the attributed
problems for that handful, please provide documented information to the
editor, and be willing to make language attitudes secondary to the
operational definitions that Ethnologue is applying. There is every desire
to make the catalog better wherever possible, and input is always accepted,
but there is need to ensure that the information can be corroborated, and
that the conclusions being claimed from the data conform to the definitions
being assumed. In the mean time, let's provide a solution to the language
identification needs of those interested in thousands of languages from
hundreds of language clusters other than High Continental West Germanic who
don't have *any* current way to tag data using identifiers that are even
close to the particular languages they are interested in.
If we want perfection, we'll never get there. We're all better off because
Unicode and ISO 10646 didn't insist on perfection. Let's not impose that
albatross on the domain of language identifiers.
- Peter
---------------------------------------------------------------------------
Peter Constable
Non-Roman Script Initiative, SIL International
7500 W. Camp Wisdom Rd., Dallas, TX 75236, USA
Tel: +1 972 708 7485
E-mail: <peter_constable@sil.org>
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:13 EDT