Re: k(w)->p

On 26 Jan 00, at 22:18, Raymond Brown wrote:

> At 11:13 am -0600 26/1/00, Matt Pearson wrote:
> >If /kp/ and /kw/ are consonant clusters, then the change from the
> >former to the latter may not be that plausible (although /p/ -> /w/
> >is found in some languages).  However, if /kp/ is a doubly-articulated
> >stop and /kw/ is a labialised velar stop - i.e., if both are treated as
> >single segments - then the change seems entirely plausible to me.
>
> And to me also.  I don't see the difficulty.
>
> Clearly V.Latin 'quattro' (Classical: quattuor) _did_ become 'patru' in
> Romanian, and 'lingua' did become 'limb@'.
>
> The theory of intermediate fricativation ( /kwattro/  -> */xwatro/ -->
> */Watru/ */p\atru/ --> /patru/ and, presumably, /liNgwa/ --> */liNGwa/ -->
> */linW@/ --> */limb\@/ --> /limb@/) is not merely unsupported by any
> external evidence but is IMHO just too fantastical for words.  The change
> /k_w/ --> /p/ and /g_w/ --> /b/ is what the evidence suggests and I
> personally see no problem.
Minor notational nitpick: (supposing for one moment that the theory holds
water in the first place, IMO it does so fairly well)  The stage you've got
above as */linW@/ should surely be */limw@/ to retain both voicing and
nasal-assimilation.

On intermediate frication, I'd have to agree with you and say that the
/x_w/ stage is neither attested nor neccessary, and I personally have
doubts about the /p\/ stage as well, though I would have to double-check
that one.

I doubt that either form was present as anything very much more than an
allophone of its 'chronological neighbours' rather than a phoneme in its
own right.

> What is conventially written as /kw/ is not here /k/ + /w/ but, as Matt
> says, a _labialized_ velar stop, a _single_ segment.  Likewise Latin -gu-
> (found only after -n-) was almost certainly a labialized voiced velar stop.
Absolutely, and in the chain of events I suggested...

On 26 Jan 00, at 1:02, Paul Bennett wrote:
>
> k_w -> x_w -> W -> p\ -> p
>
All the steps _were_ single segments.

Despite my defending my position on this, I made the above suggestion in
the context of (IIRC) Ed and Nik (my apologies if I've named the wrong
names) saying that they could conceive no plausible way in which /k_w/
could become /p/.  I felt that (regardless of its actual attestation and
provability, which I'm also ready to question [*]) this certainly was a
plausible path, and there could be others.

[*] 'ready to question' is not equal to 'currently questioning'