Re: [tied] Re: On the ordering of some PIE rules

From: Piotr Gasiorowski
Message: 48949
Date: 2007-06-11

On 2007-06-11 02:00, stlatos wrote:

> Wouldn't the tx>tHx>tH remove it from undergoing tt>ss? How do you
> know what features or sounds block the rule ahead of time?

I don't know of any independent evidence of *th2 behaving differently
from *t in Germanic.

> *gWhrendh+ti+ > *gWrinsti+ > OE gríst 'grinding'

See below.

> *prix+ > friþ 'peace', frist- / first 'truce, etc.'
> *bhondh-xY-to+ 'binding (place), stall' > Goth bansts 'barn', ON báss
> 'cow-stall"
> *drou+dhxY,t()+ 'putting trust/confidence in' > traust

*trausta- may well contain *-sth2-o- as the second element. The primary
meaning of <frist ~ first> in OE is not 'truce, etc.' but 'specified
period of time, deadline', and *pres-sth2-o/i- 'standing before'
accounts for the Germanic forms (cf. Skt. puroh-stHita). Goth. bansts
shows a /t/, but ON báss, OE bo:s(-ig), MLG banse, MDu. bo:s don't,
which throws suspicion on PGmc. *-st- rather than *-s- in *Bansaz ~
Banso:, whatever its etymology.

> That doesn't seem to work for *ksom+dhxY,to+ > *kondhto+ > *xanssa+
>> ho:s.

It does not apply here. When *-dH&1to- occurred as the second element of
the compound, the schwa was not vocalised, and possibly deleted already
in PIE. This loss in practically regular, and closely connected with the
rule that turns *newo-g^nh1-o- into *newog^no- (Gk. neognos).

> Of course it does: r was in the coda, when r>0 its mora remained:
> rur>ru_>ruu.
> Even if you disagree, what about OCS ryz^dI?

It's Slavic *rydjI < *ru:dja- (POl. rydz). There's some kind of Slavic
vr.ddhi here, though the details are not clear to me. I don't know if
and how it may be related to the length in OE ru:st.

> So, you now believe tt>st independent of PIE rules, but you're using
> this to argue against my version by placing it after a supposed *tst>ss?

I don't "believe" it. I consider such a possibility. If *[tst] is PIE
and *ss "Western IE", the loss of a vocalised laryngeal reflex in *-t&t-
may have produced a new *tt cluster handled differently (at least in
Gmc.) from inherited *tst but similarly to *pt, *kt > *ft, xt. This may
also apply to a few specifically Germanic formations with the full grade
of the root plus the *-ti- suffix, like *xaisti- (Eng. (be)hest) from
*xait- and *Grinsti- from *GrinD-.