Re: IE *p and *b

--- In cybalist@yahoogroups.com, Piotr Gasiorowski <gpiotr@...> wrote:

>
> On 2007-12-10 01:51, stlatos wrote:
>
> > It's not exactly the same, but it's definitely due to very old
> > changes to dia. in the same area. As I said, there are many other
> > changes and voc. shared. For p specifically, it's almost impossible
> > that both p > h but p > v > w between vowels and before N in _both_
> > languages would occur. Also, the intermediate xW isn't immediately
> > obvious as a stage in both, but -pt- > -xt- and (likely) -sp- > -sxW-
> > > -skW- in Old Irish and tp- > txW- > tkW- > tHkH- in Arm. make it
> > likely. The xW is also shown by the rounding in *perut! i 'last year'
> > > *xWiruTi > urid and probably *pedyo+ 'step, walking, journey' >
> > *xWidyo+ > huide in OIr.
>
> Well, _if_ there was an intermediate stage like xW, then the remaining
> changes are pretty natural, especially *-xW- > *-GW- > -w-
> intervocalically.

Whatever the intermediate stages (and I'm saying f > v > w / V_V as
a part of this) why are Celtic and Arm. so similar in the dif.
outcomes in specific environments if not from a shared history?

Is there another (shared?) path at all likely to create -xt- and
tHkH- as I wrote above?

> But Celtic may have taken an alternative path: *p > *P
> > h > 0 initially

Why the rounding (as kW / gW)?

>and *p > *P > *B > w/b intervocalically and in V_R

It's not just p>b after V and before r/l/L; t>d and k()>g().

Before n a change like *p > *P > *B > *w > u could occur, but is it
likely in _both_ Celtic and Arm.? You also didn't mention that ep
does not > eu in this position in Celtic; it became e instead. The
change of ew > e before n (and prob. iw > i if there were any
examples) is likely to be old, during the IE dialect-stage. This is
because this new ew was changed before a C in many dif. IE languages
(like *bhervY+ 'boil', but *bhrevYgY+ 'roast' > 1 *bhriwYgY+ > G
phru:go:, L fri:go:, > 2 *bhrezYgY+ > Skt bhrajj-).

> environments (which is the "official" scenario, BTW).

What would cause P > 0 before a sooner than elsewhere? The dif.
between xW and a (assuming K still patterned with a and KW with o)
could cause something like xWa > xa to begin the differentiation.

Also, what about the timing of what would cause -pn- > -pp-? That
must be fairly old, so at least p > f before that.

If there had been a lasting stage of P, why no P > B or w in
*fkYroxW-po+ / *fkYor-xW-po+ > *korpo+ > *korxWo+ > OIr corra:n
'sickle'? Or *tul-x-po+ > *twilpo+ > *swilxWo+ > W chwil 'beetle'?