--- In cybalist@..., "Piotr Gasiorowski" <gpiotr@...> wrote:
> Some people claim that the alternation is due to the original
stress mobility of Germanic o-stems (as in Balto-Slavic), so that
e.g. a root-stressed nominative was accompanied by an end-stressed
dative. After the operation of Verner's Law just one root allomorph
was generalised and the other disappeared. The only traces that
survive are sporadic dialectal reflexes of variants like *xanxista-
/*xangista-. This explanation is somewhat circular but not
implausible. Levelling of this kind must have taken place in root
nouns at any rate, otherwise we would find alternations like *mu:s-
/*mu:z- (> NWG *mu:r-) for 'mouse' or *tanT-/*tund-
for 'tooth'. "Hare" (OE hara, ON heri, OHG hasô) is a consonantal
stem, and an original pattern like *xás-on-/*xaz-en-ó- is thinkable.
Every time the effects of Verner's Law were eliminated from some
Germanic paradigm, the cleaning process was carried out very
thoroughly. For example, of the numerous Old English Vernerian
alternations in the strong conjugation only was/were has survived.
Gothic speakers did the same thing much earlier. Had there been no
other but Gothic and Modern English conjugations to compare, Verner's
Law wouldn't have been discovered.
Yes, as I thought. Someone ought to do a systematic exposé of of
this. Actually I thought up something around "tooth". Something like:
I know this is horrible and with all kinds of details wrong. But look:
Du., Sca. tand-
So perhaps we have here a paradigm levelling from each of these cases?
As for the Ingvar/Ivar thing, if this is a levelled paradigm, there
is not much grounds for arguing which is right. In that case the rune
might have started out as standing for -ng-, and as in those dialects
where the nominative was generalized, the -ng- sound disappeared
before -x-, the rune was reinterpreted to mean i: .
By the way what did you think of my idea of paradigm levelling going
different ways in different dialects as being the starting point for
(phonetic) shibboleths and they then being "condensation points" for
what later appears to be "sound changes"?