Re: The Turkic Languages in a Nutshell



Peter T. Daniels wrote:
On Jun 22, 2:50 pm, Darkstar <darkstar...@xxxxxxxx> wrote:
Peter T. Daniels wrote:
On Jun 22, 12:04 pm, Darkstar <darkstar...@xxxxxxxx> wrote:
Peter T. Daniels wrote:
On Jun 22, 8:39 am, Darkstar <darkstar...@xxxxxxxx> wrote:
Peter T. Daniels wrote:
On Jun 21, 8:07 pm, "Brian M. Scott" <b.sc...@xxxxxxxxxxx> wrote:
On Sun, 21 Jun 2009 16:19:12 -0700 (PDT), "Peter T. Daniels"
<gramma...@xxxxxxxxxxx> wrote in
<news:d88cf03c-f16d-4bef-bc05-c76dd51703a8@xxxxxxxxxxxxxxxxxxxxxxxxxxx>
in sci.lang:

On Jun 21, 3:52 pm, "Brian M. Scott" <b.sc...@xxxxxxxxxxx> wrote:
On Sun, 21 Jun 2009 12:20:54 -0700 (PDT), "Peter T. Daniels"
<gramma...@xxxxxxxxxxx> wrote in
<news:a417b645-73e6-4e7c-867b-4ea1ae94c065@xxxxxxxxxxxxxxxxxxxxxxxxxxxx>
in sci.lang:
On Jun 21, 1:42 pm, Darkstar <darkstar...@xxxxxxxx> wrote:
I'm not sure if you're interested, but I should leave these links
anyway:
A general ethnological and historical description of the Turkic
languages and peoples with many illustrations:
http://turkic-languages.scienceontheweb.net/
The argumentation for the internal classification dendrograms and the
maps of the early hypothetical migrations of the Turkic peoples:
http://turkic-languages.scienceontheweb.net/migration_and_classificat...
Why is it unsigned?
It's Darkstar's own work, as may be seen by clicking on the
<Home> link at the bottom of either page: the home page (at
<http://indo-european-migrations.scienceontheweb.net/>) has
the title 'Historical Linguistics by Darkstar'.
I did go to the homepage and saw that it was his/her
website, but nothing Darkstar has ever posted here
suggests s/he has the competence to perform massive
lexical comparisons across the Turkic family (and if s/he
did do the work, why doesn't s/he provide the raw data).

Considering his record here, do you really think that the
existence of that page is very strong evidence of such
competence?

That's what I said.

Neither you have the competence to evaluate whether I have the
competence, so the link has been posted not for you.-

Your English really goes downhill when you get upset.
Why don't you just say whether you wrote all those pages about Turkic
glottochronology (and ethnolinguistics, which I don't care about and
didn't look at), and if you did, why don't you publish the data you
used in making the calculations? You probably already have an Excel
table or the equivalent containing the forms.

The "Swadeshes" are publically available through Wikistationary,
there's a link inside the article. I manually composed about a half of
them and tried to re-check the rest — at least partly and wherever it
was essential. I don't see the necessity to publish the cognate pairs
and the calculations — no one would be checking. Even if they would,
they'll probably get it mixed up, because everyone has his own
personal counting method. That's not just taking one word and checking
it. I don't even remember seeing anyone publish intermediate
calculations in lexicostatistical studies. The results largely match
Dyachok's and Dybo's outcome, that's probably all one should know.-

If you refuse to publish either your data or your methodology, your
findings cannot be replicated (let alone falsified), and you are not
doing science.

"Because I said so" is not science. Especially when "I" isn't
identified.

A superb model for publishing lexicostatistical work is Dyen, Kruskal,
and Black's study of modern IE languages, which is Transactions of the
American Philological Society 82/5 (1992).

But no one is going to replicate these calculations or any other piece
of work. No one is going to replicate Dyen's calculations. You can
only attempt to REPEAT ALL of his study, which has been done by
several people already (including myself).

Where did you publish the ensuing confirmation or refutation?

If someone wants to repeat the calculations alone, the lists are
readily available.-

You referred to "own personal counting method" (presumably including
yourself in your "everyone"). If you do not reveal your method, your
conclusions are worthless.

No, I argued in that paper that everyone has his own slightly
different method of determining cognation, that's why
lexicostatistical percentages would always differ for different
counters. But, if you choose an internal glottochronological reference
event with a known historical separation time and use it as an
internal standard, it would not be necessary to use a standardized
(Swadesh's) method of determining cognation because the systematic
error attributed to the personal style would be canceled out by
setting that internal standard. One would only need to consistently
use the same style throughout the whole study, but it would be
redundant to explain all the details of that style.

The error can only occur if you're biased, for instance, if you fail
to notice cognates in one pair of languages, but see to many of them
in another pair.

Despite all of the above, I did use the standard Swadesh-200 lists and
the standard word-by-word method, so my methodology is not very
different. It only has, what I call, a "personal style", some minor
internal differences, that is.
.

Relevant Pages