Re: [tied] What was *kW?

From: Piotr Gasiorowski
Message: 13869
Date: 2002-06-17

Since sequences like *k^w, *g^w also change into /p/, /b/ in "p-languages", it cannot be argued that a true labial-velar double articulation is necessary to make the change understandable. In fact, sound change need not be articulatorily gradual, and kW > p makes enough sense in acoustic/auditory terms. Note that in Italic and Celtic we also have *p..kW > kW..kW, as in Lat. quinque, coquo:, quercus. The change kw, gw > p, b has been discussed on Cybalist before -- similar changes have taken place in many known instances, e.g. in Romanian. The change xw
> f is probably even more common. The developments of the labiovelar series
in all the branches are easy to explain if we assume an original phonetic value like [kW] ([k] plus labialisation) -- a rather common sound, cross-linguistically.
In my dialect of Polish (and generally in the standard Polish pronunciation) there is a stable contrast between /t+S/, /d+Z/ (stop plus fricative clusters, spelt <trz drz>) and /tS/, /dZ/ (affricates, spelt <cz dz.>, the dot being placed over the zed in the latter digraph). Here are a few typical minimal pairs. They don't involve morpheme boundaries; in all the examples the contrast is realised morpheme-initially:
trzy '3' : czy 'whether'
trzysta '300' : czysta 'clean' (f.)
drzemy 'we tear' : dz.emy '(fruit) jams'
Some dialects have levelled out the contrast in favour of an affricate pronunciation across the board, just as many IE dialects lost the contrast between *k(^)w and *kW. In standard Polish, however, the absence of contrast is stigmatised, hence occasional hypercorrect forms like /d+Zem/ for <dz.em> in near-standard idiolects.
Greek maintains traces of *-k(^)w- reflected diferently from *-kW- between vowels. The former yields geminated -pp-/-kk-, the latter -p- (I ignore palatalisation effects).
----- Original Message -----
From: Wordingham, Richard
To: ''
Sent: Monday, June 17, 2002 4:21 PM
Subject: [tied] What was *kW?

Dear Fellow Members,
                    What was the phonetic realisation of PIE *kW etc?  I can
see various alternatives, of varying degrees of plausibility:

1) True labiovelars, i.e. co-articulated stops such as Igbo <kp>, <gb>.

2) Labialised velars, i.e. velars with lip-rounding.

3) Roughly [kw], but with some difference from /kw/.

Option 1 makes the sporadic change of PIE labiovelars to labials more
understandable.  ('Igbo' used to be written 'Ibo'.)  Curiously enough, Igbo
is one of the few (only?) non-Indic languages to have 'breathy' consonants
like Sanskrit 'bh'.

Option 2 is what I would normally take superscript notation to mean.

Option 3 needs some explanation, which I now give.

I don't know how robust the contrast of one phoneme /kw/ against the
sequence of two phonemes /k/ + /w/ would be in a language.  The nearest
English analogue is /c/ (affricate, not phoneme) versus /tS/.  In /c/ the
[t] is retracted and the [S] is shortened, yielding a phonetic difference.
Thus, in British English, 'courtship' is /kO:tSIp/ (I hope you can decode my
IPA to ASCII translation), not /kO:tSIp/.  However, I know of no minimal
pairs for /c/ and /tS/, and all the examples of /tS/ I can think of straddle
a morpheme boundary.  English 'ketchup' v. 'catsup' suggests that /tS/ would
rapidly become /c/ within a morpheme.

Does Polish or Hungarian offer stable contrasts between an affricate on one
hand and the plosive plus fricative on the other?

How much support would IE morphology give to the contrast of /kw/ and /kW/?
Morphemic alternation [kw]~[ku] in the declension of *kwon-, 'dog', would
preserve the contrast in that word.  *ekwo- 'horse' would only have support
from *oku- 'swift', but perhaps the word was coined late enough for the
satem dialects to readily distinguish /kw/ and /kW/.  (I think the centum
dialects all ultimately merged /kw/ and /kW/.)

So, what is the current understanding of the phonetic reality of *kW?