Re: [tied] PIE dorsals revisited

From: Piotr Gasiorowski
Message: 12292
Date: 2002-02-06

I used to side with the two-series camp. I changed my opinion after a heated discussion on this list, where all the pros and cons were reviewed. I believe I started it with the following posting:
 
http://groups.yahoo.com/group/cybalist/message/4963
 
(where the history of the question is discussed)
 
The discussion continued for some time, and the name of the thread was cjanged in the process, so it may be difficult to follow now.
 
At present I regard the following interpretation of the three series (exemplified by their voiceless members) as very plausible:
 
*k^ = [k]  (plain velar, no doubt free to develop slightly palatalised allophones)
*kW = [kW] (labiovelar)
*k  = [q]  (uvular)
 
This possibility (mentioned in the EIEC) is consistent with the relative rarity of the *k series (by contrast to the strikingly frequent occurrence of *k^) and its occasional "colouring" influence on adjacent vowels (especially in Latin; see the postings referred to above).
 
Piotr
 
 
 
 
----- Original Message -----
From: David Russell Watson
To: cybalist@yahoogroups.com
Sent: Wednesday, February 06, 2002 9:38 PM
Subject: Re: [tied] Re: Why is PIE more centum than satem?
Here's my understanding of it:

     All of them were stops in P.I.E. and not affricates like
'ch' or 'ja' (or are you using 'j' as a glide?).
     *k^, *g^ and *g^h were palatal, at least as far to the
front as Eng. /k/ before a front vowel or glide (kin, keep,
cube) or maybe a bit further to the front than that, but not
too much more fronted, I think, for the reasons given below.
     *k, *g and *gh were velar, like Eng. /k/ before a central
vowel or a consonant (act, flex, caught, cob)*
     *kW, *gW and *gWh were either velar with additional
lip rounding, like Eng. /k/ and /w/ pronounced simultaneously
not in sequence, or uvular like Arabic /q/ but with lip rounding.

(*note that in my dialect of English the 'au' and 'o' here are
low central vowels, not rounded back ones.)

Here are my questions about it:
  
     Could *k^, *g^ and *g^h have been very much more
front than I've described, or have developed out of older
velars very much earlier than the break up of P.I.E. without
having quickly shifted to affricates once fronted?  Don't stops
in the post-alveolar, pre-palatal and palatal regions usually
quickly shift to affricate pronunciation when the sound system
has room for it?   The centum reflexes don't allow this series
to have already been affricate in P.I.E., do they?  Also, when
it's suggested that *kW, *gW and *gWh may have been
uvular, does this include the possibility that they were pure
uvulars originally without any labialization?
   
> So we've got 3 PIE sounds: *k, *kw and *k'?

Apparently so.  When I subscribed to this list I thought that
there were two velar series only.  Lehman and Burrow said
so, but Piotr has been convinced that there were three, so I
don't know what to think.

> I assume *k is a normal "k" sound (as imprecise as that is)
> what does *k' sound like using Eng, Ger, French or Skt as
> examples?

k' represents the same sound as k^, doesn't it?  And about that,
is there some standard system for representing the phonetic
symbols that this list uses?  Is it described somewhere online?

David