Re: [tied] IE numbers

On Sun, 14 Oct 2001 10:56:37 +0200, "Piotr Gasiorowski"
<gpiotr@...> wrote:

>Miguel,
>
>I'd be happier about your theory if you could show any convincing typological parallel for the curious development *pW > *kW. The reverse process, kW > p is easy to explain, since a bilabial element is already present in the input, but the spontaneous velarisation of *pW is hard to justify without some kind of special pleading.

I'm sure I've seen examples of *pw > k(w), but I can't come up with
any right now [if you don't count the doubtful Armenian k`alak` "city"
< p(W)lh1-, and all labialized consonants go to <k`> in Armenian
anyway]. There is of course /bW/ > /gW/ in Spanish (bueno, abuelo >
güeno, agüelo).

I don't see why /kW/ > /p/ would be so much easier to explain than
/pW/ > /kW/: the acoustic effect of both sounds is very similar, and
the development could go both ways (labialization is usually
accompanied by velarization).

>The cases in which Germanic has labials for expected labiovelars often involve potential assimilatory environments, e.g. *penkWe, *wlkWos with an initial labial, or *kWetwor- with a medial *-w-

Agreed about *penkwe > *pempe > fimf. There is also a labial -u in
*h2aukw-n- > "oven", but I don't see any assimilatory environment in
"liver", "twe-lve" or "sieve". And if wlkwos > wulfaz because of w-,
why does *wokw- give Gmc. *wahw-?