--- In norse_course@egroups.com, Deep Stream <DeepStream@...> wrote:
> Hi All,
>
> I can follow the following discussion of -r
> endings, BUT, still can't completely follow the
> suggestion that the -r ending never creates its
> own syllable in words:
>
> --- Kurt Oertel <oertel@...> wrote:
> > Proto-Germanic:
> > armaz
> > -z is pronounced as a voiced s
> >
>
> Here it is clear that the (reconstructed) -az
> consitituted an own syllable.

It's just so hard to say really. Traditional Icelandic linguistics
seem to say that the -r was not a separate syllable, citing the
poetry as proof. Haukur also ascribes to this view. They say that
when the spelling changed to -ur, it was proof of the pronunciation
having changed.
I personally disagree with that. Historically, going farther back
to, say, Proto-Germanic, the reconstructed ending was its own
syllable, -az. And in other European languages, it definitely was
too, such as the (roughly) equivalent Latin endings -us, -is (but
sometimes just -s, as in 'rex') or Greek -os. I have trouble
believing that the ending had weakened so much as to not have any
vowel in it, while still retaining the r sound. And that it had later
on, in Icelandic, regained its vowel sound. IMHO, it was rather
pronounced something like [@r], i.e. with a weak vowel sound, which
Nordic scribes may have felt unnecessary to write, hence
orthographic -r, with no vowel between. Later, when ON u gained the
value it has in MnIce, which sounds a lot like [@], the scribes would
have reanalyzed it as -ur, changing the spelling. But this is just my
theory; I just want to emphasize that we shouldn't take an
orthographic change as sure sign of a phonetic change.
In any case, don't feel that pronouncing the -r ending with a weak
vowel between, i.e. as [@r], is any less authentic; no matter what
Haukur or any other Icelandic scholars say. They don't have concrete
proof any more than I do.

Óskar