Re: Why do Pokorny's roots for water have an "a" in front?

From: Rick McCallister
Message: 70567
Date: 2012-12-11

I find this argument interesting for various reasons. 
First, I think that substrate words are fascinating but for some people, they are the tail that wags the dog.
Second, I'm working on an etymological dictionary for Nawat, a Uto-Aztecan language. It is obvious that at a micro-level, there is a tremendous gap among the rate of change of words. Some words in some families don't change at all, while others change profusely. Yet, the words that change and those that differ often change from family to family --if the differences between IE and UA are typical. So, at a macro-level, Swadesh is an approximation. I do question the order of words, since they seem to be more related to semantic pairs than a strict order of increasing change.
Maybe one of you can come up with a better Swadesh for the top 200 or so.


From: Bhrihskwobhloukstroy <bhrihstlobhrouzghdhroy@...>
To: cybalist@yahoogroups.com
Sent: Tuesday, December 11, 2012 4:47 AM
Subject: Re: [tied] Re: Why do Pokorny's roots for water have an "a" in front?

 
2012/12/11, dgkilday57 <dgkilday57@...>:
>
>
> --- In cybalist@yahoogroups.com, Bhrihskwobhloukstroy
> <bhrihstlobhrouzghdhroy@...> wrote:
>>
>> First peopling. Of course, there's no proof either - otherwise we
>> wouldn't discuss on this topic - and there are many other
>> possibilities, but our first task is to fix the extreme extensions of
>> what is reasonable. Such extremes are by definition scarcely
>> attractive, but they are nevertheless necessary. One of these extremes
>> is Drew's theory (Non-Anatolian PIE shortly before the earliest
>> Non-Anatolian evidence; I don't like Indo-Hittite theory, but this is
>> nevertheless a necessary extreme; S. Kalyanaraman puts himself beyond
>> this extreme); the other one (on the remote side) is (Europe's) First
>> Peopling. Why is it necessary? Because nothing can (still) exclude it.
>> Let's leave apart non linguistic consideration and let's concentrate
>> on linguistic arguments: the crucial point is how to judge anything
>> that isn't universally accepted as PIE heritage (Tavi doesn't accept
>> PIE reconstructions and his rejection is even more radical than what
>> he states, because he operates with different - if any - soundlaws, so
>> his theory is outside the scope of this discussion). As long as
>> alleged Pre-IE evidence can receive PIE etymologies according to
>> received soundlaws, the discussion must remain open. If and only if
>> all these regular PIE etymologies were right (beside being correct),
>> the equation PIE dispersal = Europe's First Peopling could gain
>> something in probability, although still not definitely proven. I
>> presume I can offer a regular PIE etymology for any alleged Pre-IE
>> relic and I think such etymology, although not always better, is
>> anyway never worse than its alternatives. This is the best linguistic
>> discussion we can deepen on this topic
>
> I have no doubt that you CAN offer a PIE etymology for any alleged pre-IE
> relic, just as you CAN offer a Celtic one for any alleged Ligurian relic,
> but we run into the same issues of falsifiability.
>
> As for your highly adjustable rate of linguistic change, biblical literalist
> Albert Cuny used it to argue that the vast variety of human languages could
> indeed have arisen in the few centuries since Noah's Flood supposedly
> occurred. Is this the sort of "science" we want to be doing?
>
> DGK
>
I think we all want a sort of linguistic science by which we can
recover - through indipendently tested linguistic method - some piece
of otherwise lost evidence of Prehistory. As long as linguistic
reconstruction is concerned, we all use the same method; only
exceptions are Tavi's comparisons and Stlatos' optional changes, but
they don't matter here. Note that the "PIE etymologies I can propose
for every alleged pre-IE relic" are entirely the same You all could
propose as well, since I don't add tghe slightest innovation to
received PIE diachronic phonology or morphology.
As to the external linguistic (historical) frame, we come into
another kind of research, no more strict linguistic reconstruction,
but rather linguistic-based historical reconstruction. This is the
field of research where we disagree. In order to proceed, we need
generalizations, but such generalizations on Linguistic Change and
other typical topics of General Linguistics would still need a much
larger data base than we at present have, so we can just make use of
educated impressions. I don't become a crackpot just because I have a
different opinion on a specific point of the discussion about the Rate
of Linguistic Change (i.e., that I don't believe in Glottochronology);
maybe I'm indeed a crackpot, but not for that reason... I can't
understand what the argument of Cuny's ideas specifically has to do
with our present discussion; we are dealing with long durations, not
short ones.
Please always remember that this discussion is about the oldest
thinkable extreme of PIE unity, so that opinions differ only about
that point, not any other, and that the most spread views about PIE in
the third and fourth millennia BC(E) are fully included in the model
of the oldest thinkable BEGINNING of PIE unity I'm proposing. Please
also avoid any reference to Mario Alinei or other Continuity Theories
like that, because they have even less to do with what I'm maintaing
than with Your ideas, so maybe they are more interesting than mine,
but pointless in our present case.
So, let's start from PIE in the third and fourth millennia BC(E).
We aren't concerned with any of the subsequent changes. At least in
its final stage, PIE could indeed be spoken on a large area (say from
Western Europe to Central Asia and India); this is our first question.
Maybe it didn't, but this hypothesis can be accepted as maximal
reasonable extreme. I think we all agree on that (except Tavi and S.
Kalyanaraman).
Second question: lexical diversity. We all know that specific
lexical units hugely diverge bewteen IE classes. Can we backproject
this divergence into PIE in the form of lexical richness? Common
answer: no, a unitary language cannot have so huge a lexicon. But our
question is: how huge can be the lexicon of a diasystem extended from
Western Europe to Central Asia and India? Does the backprojection (cum
grano salis!) of the lexicon of the different IE classes fit, at least
as an extreme possibility, the requirements of such a diasystem? I
think - although I'm not sure - that we can agree on this as well in
answering yes, at least for the sake of the argument, because the
point is a further one.
Third question (first crucial one): can such a lexical diasystem
have evolved, N.B. *at the slowest rate*, for several millennia? If we
don't believe in any Glottochronology, we must answer "maybe, we don't
know" and therefore take it into consideration as a possibility. If,
on the contrary, we believe in some kind of Glottochronology, our
answer can be: 1) (Classical Glottochronology) "yes, approximately for
six millennia" (Swadesh' Glottochronology puts PIE - IH - dispersal at
about ten millennia BC(E)) or 2) "no, the rate of *lexical* Linguistic
Change in Prehistory was faster than in historical times (this is Your
thought, isn't it?) or 3) "yes, the rate of lexical Linguistic Change
in Prehistory was slower than in historical times (and therefore we
must take that possibility into consideration)". So, do You (all)
believe in Glottochronology or not? If no, please take 'my' model into
consideration; if yes, please identifiy Your position between 1, 2, or
3, and consequently act. You appear to adhere to glottochronological
position n° 2 (= "Glottochronology is a valid science, but Swadesh was
wrong, because linguistic change in Prehistory was faster than
afterwards").
Fourth question (second crucial one): I've explicitly written
"reconstructable phonological diasystem", what I propose is barely
that the phonological features we reconstruct for PIE had been stable
for some millennia before the fourth. That a specific phonological
feature can be stable through millennia has nothing incredible, I
think; at least, it can be accepted as a theoretical extreme. Every
other phonological feature (points of articulation of laterals,
vibrants, and so on) can indeed have gone through multiple changes
(according to specific local varieties of PIE) starting from the first
PIE speaking community (of course a restricted one, not extending from
Western Europe to Central Asia and India).
In sum, You can see that the model I propose *just as an extreme
point of thinkable reconstruction* is made of:
- the expansion of a regional community to a wide area (from Western
Europe to Central Asia and India) at the time of First Peopling (this
is proven: First Peopling has indeed taken place in a relatively low
amount of millennia starting from very little human groups);
- the language of such expanding community did suffer phonological
changes, but these changes didn't affect the 100% of phonological
features; some features remained (by chance) unchanged for millennia;
- this language also experienced laxical change and variation, at a
rate whose extimation depends on the version of Glottochronology one
accepts; I don't believe in Glottochronology, so I take into
consideration the possibility that lexical change can be very slow or
very fast or anything else, but if I were compelled to choose anyway
some kind of Glottochronology I would choose a slower version than
Swadesh' one (since I think that linguistic dynamics were slower in
less complew societies) or at most precisely Swadesh' one, surely not
a faster version (as, on the contrary, You all seem to choose);
- during these millennia, the spread of Agriculture, Chalcolithic
innovations and migrations and so on produced the diffusion of
culturally important words, whose phonological features were mostly,
but non totally, different from local variety to local variety;
- I propose just two postulates: that the first community spoke what
we reconstruct as strictly unitary and spacially restricted Common PIE
and that the phonological features which in the following millennia
didn't got different from one local variety to another one were
precisely the ones we reconstruct as PIE in the fourth and third
millennia BC(E).
Everything, in this model, seems possible (to me). If there's
anything that a priori must be excluded, please have the patience and
courtesy to point it out (once again and more precisely than before)