Re: IE *p and *b

On 2007-12-09 11:15, meska_jd wrote:

> One of characteristic changes in Celtic was *gW > b (*gWow- > bou-
> "cow) and falling with of aspirated and plain consonants after
> which consonant system becase balanced where every voiced consonant
> had a voiceless correlate. Development of *p has thrown the system
> in confusion again as *b was left without a pair. It is rather
> unusual that a balanced system would be taken out of balance without
> any good reason.

Hm. My experience is that even nicely balanced systems can be thrown out
of balance for no obviously good reason. Languages are vast systems and
we don't always understand their somewhat chaotic dynamics. Late Middle
English had a most satisfactory, perfectly symmetrical arrangement of
long vowels (i:, e:, E:, a:, O:, o:, u:), and then the Great Vowel Shift
came along and messed it up.

> I do strongly believe that *p behaviour means a bit
> more than many people usually suppose. Furthermore, later when
> lenition process started all medial -j- and -w- have been lost,
> meaning *p disappearance was very likely to happen due to similar
> mechanism.

I wonder if it's significant, but initial *p was lost earlier before *a
(also from vocalised laryngeals and resonants) than in other contexts.
We have various early spellings showing some kind of residual
consonantal reflex ([h], [B]) in other contexts, but never before /a/.
It seems that the lexical diffusion of *p > *P > h > zero was sensitive
to the openness of the following segment, and was spearheaded by words
in *pa-, like pre-Celt. *pate:r.

> Regarding *sp-: there are no native words with *sp- in Irish, all of
> them are either Latin or English borrowings.

What about the 'heel' word, OIr. seir (acc.du. di pherid, showing the
outcome of lenition), Wel. ffêr 'ankle'? The root is usually given as
*sperh(1?)- 'tread, kick away' (Hitt. ispa(:)ri, Skt. spHuráti, Lith.
spìrti, and even if one insists on a PIE reconstruction with aspirated
*pH (see LIV, p. 585), I would not expect different reflexes of *sp and
*spH in Celtic.

Piotr