On 2007-06-19 03:02, stlatos wrote:
> --- Piotr Gasiorowski <gpiotr@...> wrote:
>
>> That would mean there is no
>> independent support for
>> such a change. Other sonorant clusters may develop
>> into geminates but
>> don't cause vowel lengthening.
>
> But you said:
>
> --- In cybalist@yahoogroups.com, Piotr Gasiorowski <gpiotr@...> wrote:
>
>> In our earlier discussion I
>> suspected a Proto-Germanic vr.ddhi formation, which is a possibility,
>> but perhaps the simplest solution is the best: the expected feminine
>> derivative of *kán-o:n would have been *kan-n-ih2. If at the time in
>> question the ancestor of Germanic did not tolerate geminated nasals,
>> degemination with compensatory lengthening would have produced
> *ka~ni: >
>> *ka:ni:, further transformations withing Germanic yielding secondarily
>> extended stems like *xo:n-jo:n- (ON hœna) etc.
True, but:
(1) That was a private hypothesis on my part, not communis opinio, even
if I fancy it looks elegant (probably more elegant than an unmotivated
vr.ddhi)
(2) If there's anything correct about my analysis, *kan-n-ih2 had an
_original_ morphological geminate, eliminated certainly in pre-PGmc.
times. The reduction of *-rl- to *-:r- would have had to be an inner
Germanic development, given the proposed origin of the sequence. I can't
think of anything comprable in Germanic. We do have secure examples of
*-ln- > -ll-, *-nw- > *-nn- and some decent if disputed evidence
evidence for *-sm- > *-zm- > *-mm- and *-tl- > *-Dl- > *-ll- in the
context of Verner's Law. None of them involves vowel lengthening. *z is
sporadically lost with a compensatory lengthening as in OE me:d ~ meord
'reward' < *mizdH-ah2 and I have myself proposed a few etymologies like
*xe:ra- 'hair' < *kes-r-ó- (Cf. ON vár 'spring' < *wes-r-ó-), but here
again it's (Vernerian) *z in _coda_ position that is vocalised.
Piotr