On 12/06/2002 06:19:13 AM Scripts2 wrote:
>I'm sending this to you personally: if you wish, I can also send it
>via the JAC list.
Well, I think it would be helpful to others on the list to hear us agreeing
on some things for a change. :-) I'm sending my response to the list, but
will delete some bits. I'll just forward your original.
>> This analysis is based on past research. The question, then, is whether
>> there is newer evidence that points to a different analysis.
>
>Yes: as I said earlier, in no way can Kalmyk and Oirat (much
>separated by space and time) have stayed so similar as is suggested
>in the analysis, when languages much closer in time or place have
>diverged as their normal pattern.
I wouldn't characterise that as new evidence, though. Rather, it's a
hypothesis along with a reason for why one might expect the hypothesis to
hold. The evidence would be linguistic or sociolinguistic information from
researching the language and language community in question.
>> I can't disagree that this situation looks like it would be a potential
>> candidate for language split, but I don't think we can simply conclude
from
>> the time and locations that this must be the case since there may be
other
>> factors involved. The time depths for the European diaspora (English,
>> French, Spanish, Portuguese) are comparable, and the distances are
>> generally greater, yet language splits have not occurred.
>
>The disapora in those cases is more recent.
>They are much bigger language groups, with more international
>contact. They are united by media influence in all the cases you
>mention.
Indeed, we can identify factors that account for why there has not been a
language split in English or Spanish, etc. I was not using that example as
an argument by analogy that no language split has occurred with Oirat /
Kalmyk, though. Rather, I was using that example to demonstrate that time
and distance are not sufficient factors for language split. Those are the
two factors you have given to support the hypothesis. If they are not
sufficient factors, then we can't consider the hypothesis to be confirmed.
>If those three factors were not involved, there would have been
>splits, like that I have described in previous emails for Kalmyk and
>Oirat.
But I haven't seen you actually describe a split for Kalmyk and Oirat; I've
only seen the claim that it has occured.
>I cannot ask the Ethnologue Editor to split
>> the "Kalmyk-Oirat" entry into two unless there is reliable evidence
>> indicating that these are distinct languages.
>
>Conversely, nobody, including the Ethnologue Editor can justify
>retaining the "Kalmyk-Oirat" entry as a single entity, unless there
>is reliable evidence indicating that these are not distinct
>languages.
That's the whole point. It's my understanding that the Ethnologue Editor
had already been given reliable evidence indicating that those are not
distinct languages. Barbara Grimes was very careful in her work, and is
careful in the information she supplies. When I asked her about this case a
couple of weeks ago, she didn't say, "We assumed one language because we
didn't have enough information to support adding a split, but that may well
be the actual situation." Rather, she confirmed that the evidence she
reviewed pointed to a single language.
>Normally, word lists, etc are used. I'm wondering how far that
>individual one was investigated in the first case.
I don't know the details on the sources Barbara Grimes had at her disposal.
>> Therefore, I won't be able to relate an ISO 639-2 entry for
>> "Kalmyk" to an entry in part 3, and there will be an interoperability
>> problem.
>
>There's a similar interoperability problem for Zapotec (and others)
>in ISO 639-2 and in the Ethnologue, so that's not a unique problem.
>You've provided details of ways to overcome this in your IUC papers
>etc.
>
>That issue already needs to be dealt with.
True, but that situation is different in a significant way: there are lots
of cases in which ISO has one code element where Ethnologue has many, and
we analyse those situations by saying that the ISO entries denote
collections of individual languages. Thus, we analyse it using a data
category that already exists in ISO 639. Here, though, it would be the
other way: Ethnologue has one entry corresponding to ISO's many. Since the
Ethnologue Editor isn't about to suggest that the things listed in the
Ethnologue are anything but individual languages (i.e. it uses a single
data category for all of its entries), then we have a problem of needing
new data categories for things smaller than an individual language in ISO
639-2, and I don't think we really want to go there. (There is, though, an
existing case of this problem already in ISO 639: Twi, Fanti and Akan are
listed separately in ISO 639-2, and Twi is also in ISO 639-1, but the
Ethnologue lists Twi and Fanti as dialects of Akan.) Even if we did, you're
not wanting to see Kalmyk refered to as a dialect. So, we'd have a problem
with two entities -- "Kalmyk" on the one hand and "Kalmyk / Oirat" on the
other -- for which we'd be claiming that they're the same kind of object
(they're both individual languages), but we're also claiming that one is a
component of the other. That just wouldn't work.
>> (And if I make a split in ISO 639-3
>> without having had evidence for the Ethnologue to justify a split, there
>> will still be an interoperability problem.) That is why I have been so
>> stubborn in saying that we should not proceed on the basis of a split
until
>> there is real linguistic or sociolinguistic evidence justifying a split.
>
>And why I have been so stubborn on this one is because the Ethnologue
>is usually right, there is a danger that any further questioning is
>not done in exceptional cases.
There's no question that there are places at which the Ethnologue's
analysis is incomplete. Linguistic or sociolinguistic evidence is needed
before improvements can be made, though. A hypothesis is not sufficient,
even one that has some reasonable motivation.
>> I'm not a voting members of the JAC, so I don't want to continue taking
so
>> much bandwidth in this discussion. I'll just say that if you add a code
>> element for "Kalmyk" now, it is going to create an issue that will
require
>> revisiting the matter in the very near future.
>
>It will, so let's do the analysis.
>
>I'm going to talk to further Mongolic experts that I have located.
Very good.
- Peter
---------------------------------------------------------------------------
Peter Constable
Non-Roman Script Initiative, SIL International
7500 W. Camp Wisdom Rd., Dallas, TX 75236, USA
Tel: +1 972 708 7485
|