On Wed, 02 Jun 2004 10:06:22 +0000, tgpedersen
<
tgpedersen@...> wrote:
>> >Aha. Maybe an early loan from PIE *p-g- "pole; catch"
>>
>> No. An early loan would have given *Vg-
>
>Yes, following the rules as given by M. Vidal for Pre-Pre-Basque. A
>loan into Pre-Basque (Roman times, and earlier?) would give *bVg-. An
>IE dialect that weakened stops in inlaut might have *pVG-. A loan
>from that into Basque might give *bVh-.
*p-, *t- and *k- disappeared in Basque before Roman times.
Loans from Latin have b-, d- and g- for p-, t- and k-, but
earlier loans (from Celtic) have t- > 0-, k- > 0- [no p-, of
course]. This suggests that the loss happened *just* before
the Roman period (Celtic loans are affected by the loss,
Roman loans aren't). So normally, your proposed "early loan
from PIE" would have had to take place no earlier than the
Roman period, which is a problem, because PIE didn't exist
anymore at that time, and the only loanwords to be adopted
by Basque as of the Roman period until teh modern era were
Latin (and later Romance) words.
Of course *p- might in principle be an exception. There is
no Celtic material to show what happened with it, and it's
quite likely that the loss of *p- in Basque was earlier than
the loss of *t- and *k- (in which case it would be an areal
feature shared with Celtic): when protected in the position
at the beginning of a verbal root, Basque maintains *t-
(e.g. e-torr-i) and *k- (e.g. e-karr-i) [and *d-], but there
are no verbal roots starting in *p- (i-pin-i is the only
exception, but the *p- is not original, coming from *bVh-).
The consonant(s) that give(s) hiatus medially in Basque
words do not include /G/ (Basque /g/ *is* [G] medially
between vowels, as in Spanish). I'm not sure *what* it was:
not p/b, not t/d, not k/g, not m/n, not s/z, not r, not l.
The most likely candidates would be /w/, /y/ and/or /h/.
=======================
Miguel Carrasquer Vidal
mcv@...