Re: [tied] Albanian intervocalic s > h

On 6/19/05, alexandru_mg3 <alexandru_mg3@...> wrote:

> I.
> >[AK] wrote:
> >You are wrong. PIE *plok'-so > Alb. plah/plaf 'thick covering/rug',
> >suffixed form <pëlhurë> 'woven cloth, , textile', due to my
> metathesis
> >of liquids L - V > V - L / R -V > V- R, when vocal after liquid pass
> >in unstressed position. See also *(s)pok'-so > pah 'view,
> >perspective'.
>
> There is no *plok'-so, for Alb. plah, Abdullah.
> See Demiraj on the web at Leiden: Alb. plah has a regular ending: -
> ske/o
> url: http://www.indo-european.nl/cgi-bin/response.cgi?
> root=leiden&morpho=0&basename=\data\ie\alb&first=351&sort=albform
>
> (I don't want to say that Demiraj is always rigt but -ske/o is the
> regular ending here)
>

[AK]
To plait and to cover are two different words. So, my derivaltion is
not from *pel- 'to cover, hide' as is Demiraj's, but from *plek'- 'to
plait'. Similar formation you have in Eng. <flax> from o-grade form
*plok-so. Primary meaning, to my view, is to weave, to plait and,
secondary could be to cover, to hide.

>
> II.
> >[AK] wrote:
> >About PIE *dek's-to > Alb. i djathtë 'right' and *eg'hs-to > Alb.
> >jashtë 'out', if we accept Jens' rule that *e yields /ja/ after
> >secondary cluster, than *e > ja is regular. Also you must take into
> >consideration that in these examples we have syllabic structure
> CVCCCV
> >that was normally reduced to CVCCV, so as /k'/ as well as /g'h/ are
> >dropped.
>
> a)
> -të in djath-të is recent formation in Albanian.
> There is no -tht- formation in today Albanian, that could
> originate 'directly' from a PIE cluster
> See for a more older cluster *k'st the Albanian gjashtë > sek's-to
> => So at least sht but not tht
> (I strongly suspect that this sek's-to is not a PIE formation (as
> Piotr proposed us some time ago) but a more recent one,
> but at least sek's-to is much much older then djathtë...)
> url: http://www.indo-european.nl/cgi-bin/response.cgi?
> root=leiden&morpho=0&basename=\data\ie\alb&first=151&sort=albform
>
> b)
> See also Demiraj that talks abouth djathë too.
> Do you have also in Albanian the worb ngjath, as Demiraj
> presented, where there is no -të inside?
> url: http://www.indo-european.nl/cgi-bin/response.cgi?
> root=leiden&morpho=0&basename=\data\ie\alb&first=311&sort=albform

[AK]
If you agree that *sek's-to yields in Alb. <gjashtë> through droping
the <k'>, where do you see the obstacle to not be droped /s/, for
otherweise we will have two <th>: k' > PAlb.*st, and *st > sht, the
same sound in *dek's-to, as well as in *eg'hs-to > jashtë (*g'h as
/k'/ falls regulary). <ndjathë>, not <ngjathë> is an adver which in
juxtaposition <në të djathtë> 'on the right' should be heared
somewhere as <ndjathë>, but as far as I know regular form is
<ndjathtë>.

> So there is no CVCCCV here...it's simple *dek's-
>
>
> III.
> >[AK] wrote:
> > As I have said, Alb. <kohë> 'time' is in many contexts fully
> synonymic
> > with Alb. <moshë> 'age', the reason why we must also respect Latin
> > <cascus, -a, -um> 'old, ancient, primitive'. This is an argument
> more
> > that we must back up PIE recunstructions *keH1s-k'-aH2, for /*k'/ >
> > Sl. /s/.
>
> Sorry to ask you, but:
> What are the semantism of this k' here? What cognates do you have?
> Why is mandatory to suppose k'? (Seems that Derksen and Demiraj
> didn't see any k'...)

[AK]
Exactly the Slavic <c^as>.

> Best Regards,
> Marius
>
>
> P.S. Regarding the similar metathesis that you have proposed (none of
> them related to swe > wes), I'm not sure about your derivations.

[AK]
Are you sure about Latin <cervus>, Slavic <srna> or it must be said
from some imaginary authorities. Why you are not sure about <sorkadhe>
that is derived from *k'erw-k-ala > *k'wer-k-ala or you have a doubt
also about *k'we:rna > Alb. <sorrë>.

> To take one example:
> zog for example doesn't have solid cognates to be sure on it.
> To suppose in addition a metathesis regarding zog, will make less
> probable its derivation...that for sure is difficult from the
> begining.
> See also a different derivation for zog (with no metathesis):
> url: http://www.indo-european.nl/cgi-bin/response.cgi?
> root=leiden&morpho=0&basename=\data\ie\alb&first=491&sort=albform

[AK]
But, see: Certenity -

> Best Regards again,
> Marius

Konushevci