From: Patrick Ryan
Message: 57335
Date: 2008-04-15
----- Original Message -----
From: "david_russell_watson" <liberty@...>
To: <cybalist@yahoogroups.com>
Sent: Tuesday, April 15, 2008 4:29 AM
Subject: Horse Sense (was: [tied] Re: Hachmann versus Kossack?)
--- In cybalist@yahoogroups.com, "Patrick Ryan" <proto-language@...>
wrote:
> --- In cybalist@yahoogroups.com, "david_russell_watson"
<liberty@...> wrote:
>
> the far greater number derive from a phoneme that was
> probably realized as [x]/[ç].
>
> Yes, it is rare. I resisted the correspondence when I
> first saw the comparative data pattern.
What correspondence? The various outcomes in the Indo-
European dialects of the sounds in question all lead to
the conclusion that they were _stops_, with only their
exact locations of articulation and types of phonation
ever being in any question.
***
Patrick:
Interlanguage-family comparison: PIE *g(^)W(h) and k(^)W correspond to velar
and palatal spirants.
I doubt this can be proved strictly within PIE.
***
> After trying every other alternative, I was forced to
> accept.
But nothing forces us to accept that. On the contrary
the majority of their earliest attested reflexes are
stops and affricates, with the affricates being easily
derivable from stops, and the fricative reflexes being
easily derived from the affricates.
***
Patrick:
See above.
***
> A back velar would have made much more sense.
>
> During the Pontic period of glides and neutral *a,
What's the "Pontic period of glides and neutral *a"? No
such stage is reconstructed for P.I.E.
***
I think you are again wrong.
We have discussed on this list a stage of early PIE (or pre-PIE) which had
integral *i, *a, and *o. Subsequent to that stage, the Pontic period saw
*i -> *YA; *a -> *A; *o -> *WA.
By *A, I mean the Ablaut vowel: *é/ó <- è/*°/*Ø.
Subsequently, the glides were discarded leaves traces like the
palatalization of dorsals that had preceded them.
***
> *kWa would have been the outcome of an earlier *ko. Perhaps,
> this sequence was phonemicized and generalized. When glides
> went the way of the wind, *kW and *gW were available for new
> employment.
When did glides go the way of the wind? Both *w and *j
are reconstructed for P.I.E., with reflexes of both of
them surviving as glides in some of the dialects.
***
Patrick:
I cannot date it exactly because a pre-Ablaut Pontic stage of PIE can only
be theorized only the basis of its antecedents.
***
More importantly, though, you give 'k' and 'g' for this
supposed earlier stage of P.I.E., not the 'x' for which
you're supposedly arguing.
***
Patrick:
Apples and oranges.
You have misunderstood what I wrote.
PIE *k(^)W has nothing to do with Pontic *kW(Y)A which derives from Pontic
*xW/Ø(Y)A; Pontic *kWA and *kA yields PIE *kA; Pontic *kYA yields PIE *k^A.
***
> Why?
>
> Because the earliest PIE speakers seem to have been voiceless
> spirant-shy: no *f, no *x/ç.
If your claim is that the velars of late P.I.E. ascend
to fricatives in the earliest P.I.E., then how does it
help your case to claim that the very speakers of that
earliest P.I.E. were "spirant-shy"?
***
Patrick:
Because it explains the necessity (for them) of converting [G] and [x] into
something they felt comfortable (or able) to reproduce.
I am open to other suggestions.
***
Moreover it doesn't really appear they had any problem
pronouncing voiceless fricatives. The main allophone
of P.I.E. *s is supposed to have been one, and some of
the laryngeals are well supposed to have been too. See
section 14.5. "Evidence for fricative articulation" and
section 14.7. "Evidence for voice" of Lehmann's 'Proto-
Indo-European Phonology', available online at
http://www.utexas.edu/cola/centers/lrc/books/piep14.html#txu-
oclc-3953445.xml-div-d0e19661 .
***
Patrick:
An allophone in certain positions is not a phoneme though, of course, it
_can_ become one.
We do not reconstruct [z] for PIE precisely because we can reconstruct no
contrasting pairs:
**zeC <-> *seC
Thank you for the link. Lehmann gave me a copy of his book.
***
> You will mention Hittite <h> as [x]; its employment for all
> notated 'laryngeals' suggests strongly Hittite <h> was [h]
> not [x].
Even if the Hittite sound were truly [h], Hittite isn't
P.I.E. and [h] could easily derive from an earlier velar
fricative. Though, again, what does it lend to a case
for earliest P.I.E. having velar fricatives to deny them
for later P.I.E.?
***
Patrick:
I think you are losing track of who is arguing what.
No, of course, Hittite is not PIE but do you know of any recorded language
closer to it than Hittite.
Hittite developments, though not necessarily definitive, should carry
appropriate weight.
Could? We are not talking about what could happen, Dr. Pangloss, but rather
what is likely to have happened - closer to the truth we may not come.
I do not have a problem with [s] but some do.
***
> I wondered about that, too. The better notation would have
> been *k(^)WH.
No, that's still impossible. There can be no 'k^W'. In
the standard system of notation '^' and 'W' represent
mutually exclusive modifications of a velar: palatality
and labiality respectively.
Maybe you can explain what you intend with this notation
in phonetic terms.
***
Patrick:
What I meant to notate, and perhaps I failed to do it effectively, is that
pre-PIE [x] and its allophone [ç] acquired glides when preceding *o (*W) and
*e (*Y), at which point all short vowel qualities, unless lengthened, were
conflated into *A.
Sometime after that, [x] and [ç], for whatever reason, they were replaced
with with *kW(W)A, which resolved into *kWA; and *ç^WA, for the latter of
which it does not take a North Caucasian to pronounce.
If you prefer, we could notate the latter as *çW.
*kW remains in PIE; *çW is notated in PIE as *kW also but I believe one can
detect slightly different reflexes for the two, and try to indicate this by
*^; perhaps a mistake.
***
-----------------------------------------------------------
So as best as I can guess, you're saying that, between one
early stage of P.I.E. and the next, its speakers lost the
ability to pronounce voiceless fricatives, and that thus is
explained the strange change of fricatives to stops that
you claim took place.
You can't prove that such a radical change in pronunciation
ever took place, however, and haven't, and what's missing
as well is the reason we should believe they were fricatives
in the first place.
***
Patrick:
Essentially correct.
I do not believe so much that PIE speakers lost voiceless fricatives (not
including [s]) but more probably originally non-PIE speakers could not
reproduce them, and substituted *kW when they adopted PIE.
You deny me the possibility of proof when you do not allow pre-Pontic forms
to be introduced. The surprising switch can only be followed if extra-PIE
data is allowed.
So, it is not that I cannot prove it in spades, but you refuse to let me
have the deck of cards.
I have dozens of examples to support this bizarre development but you will
never lookat the, will you?
***
> > That's correct. Please do provide examples to back your
> > claims here, and clarify what was intended with your non-
> > standard notations.
>
> I believe I have done that above.
With 'examples' I was actually requesting correspondences,
meaning lists of cognate pairs to prove that 'k(^)W', in
whatever sense you're using it, or any other of your other
reconstructions, is superior to the standard reconstruction
already widely accepted. I suppose you do understand that
P.I.E. as reconstructed by yourself is quite different from
that of Beekes, Lehmann, P. Gasiorowski, M. C. Vidal, etc.,
and even that of Pokorny, do you not?
***
Patrick:
Of course I understand it is different; when you wear a monocle, you see
only what one eye sees.
I do not fault them at all for the discrepancy.
They have been schooled to reject extra-PIE data; and without it, their
conclusions are perfectly reasonable.
As for *çW, the difference in responses between it and *kW can be explained
differently using the standard PIE approach. Or, perhaps I should simply
accept the idea that *çW simply lost its palatal quality and became
identical to the alveolar/velar *kW.
It is certainly the easier case to argue.
***
The idea that the P.I.E. velar stops derived from earlier
fricatives can't be made on the basis of cognates, however,
since P.I.E. has no known relatives to provide those. It's
a matter entirely of internal reconstruction, and I don't
believe anybody's been led to posit pre-P.I.E. fricatives
as the source of P.I.E. velar stops on that basis either.
David
***
Patrick:
As I have said, the case can be made from cognates if extra-PIE data is
permitted to be utilized; and, if you had ever read my essay in
http://geocities.com/proto-language/c-AFRASIAN-3.htm
you would see that I _have_ proved it.
Why would I have had to develop this theory if anyone had already recognized
the obvious?
***