Re: colouring

On Tue, 12 Feb 2008 20:17:24 -0000, "Pavel A. da Mek"
<a.da_mek0@...> wrote:

>> Do you mean
>> "Szemerényi" length in the nom.sg.? It is not compensatory, strictly
>> speaking, since the vowel is lengthened also when no segments are lost,
>> as in *dje:us.
>
>It is said that what was lost here is "non-segmental phoneme" - syllabicity
>of the following syllable; but popularly speaking,
>I suppose that it could be described either as
>the compensation for the loss of the final reduced vowel,
>or maybe as metathesis and contraction:
>
>nom. dejéweze > djéwz& ( > djé&wz) > djé:wz > djé:ws
>acc. dejéweme > djéwm& > djéwm.
>gen. dejewèse > d&jwès > diwós
>
>nom. peh2etéreze > p&h2térz& ( > ph2.té&rz) > ph2.té:rz > ph2.té:r
>acc. peh2etéreme > p&h2térm& > ph2.térm.
>
>First the odd unstressed vowels were lost and the even ones reduced,
>then the shwas next to H or R gave them their syllabicity,
>whereas the remaining shwas transfered their mora to the preceding syllable.

I don't understand what you're proposing here. If the
lengthening is caused by a lost final mora, why do we have
it in Nom. *p&2té:r and not in Acc. *p&2térm., if they both
come from a thing like pVxVtVrV-CV?

The way I see it, Szemerényi's law works only where we have
a sequence -VCF# (vowel-consonant-fricative), and it doesn't
work where the sequence was -VCVF# _before_ zero grade. The
classical example would be Nom. *pó:ds vs. Gen. *péds, from
*pód-z vs. *péd-&s.

Where there *is* compensatory lengthening is in the Acc. sg.
of the "god"-word, mentioned above: *djé:m < *djéwm.

=======================
Miguel Carrasquer Vidal
miguelc@...