--- In
cybalist@yahoogroups.com, "Richard Wordingham"
<richard.wordingham@...> wrote:
> > > Would you say Latin was well-nigh a 3-vowel (/a/, /e/, /o/)
> language?
>
> This was not planned as an ambush.
>
I didn't take it to be. It's a plain and friendly-enough civil
question of relevance to the matter.
> > I suppose I would. The Romans did fine in their own writing
using
> I and V
> > as both consonants and vowels. Vowel length would have to be
> handled by
> > doubling which would be unproblematic since there are no
hiatuses
> between
> > identical vowels.
>
> There are! See discussion of /VV/ below.
>
> What about _fi:o:_ 'I become' compared to _facio:_ 'I make'?
>
> I assumed you'd know the minimal pair (I assume) _qui:_ 'who
(rel)'
> and _cui:_ 'who (d.s.)'. It's one of the reasons I said 'well-
nigh'.
Right, and that's what I said yes to. If well-nigh still applies,
there is no error in this.
> Vowel length is difficult to handle as a geminate. /VV/ is
> ambiguous in the environment C_C. In the 4th declension we have,
in
> modern spelling -u:s (g.s., nom.pl.), but -uum (g.pl.), to which
we
> can add add 2nd declension forms _servus_ 'slave (nom.s.)', and
> _servum_ (acc.s.), both of which are two syllables.
> _arduus_ 'steep, high; troublesome' and _viduus_ 'bereft' add to
the
> troubles unless the modern spelling is wrong.
>
> /II/ has a two-fold ambiguity. In compounds of _iacio:_ throw,
> the 'adverb' is long by position; I was taught that <inicio> was
> actually /injicio:/.
Well, yes, of course! It doesn't work entirely for this version of
Latin which is the one we normally quote. I do not want to
compromise the valid one-vowel analysis of Sanskrit by insisting on
an invalid application of the same for Latin.
It comes closer to working if Latin is pushed back to the time right
before the classical texts, i.e. the layer that had equos, not
equus. The words you quote would be /facioo/, /cuii/
(or /cuei/?), /cuuii/ (/-ei/?), /-
uus/, /seruos/, /seruom/, /arduos/, /viduos/.
I get trouble with _fi:o:_ and _injicio:_ whose reality however is
less then secure. A sequence [i:o:] is at variance with the rule
shortening vowels in hiatus, and [inic-] is often metrically
demanded, but even more often [injic-] is. It looks like a
vacillating norm: the form with reduced -jic- > -ic- would be the
older one, while -jic- is a restoration. Sommer explains [fi:-]
before vowel as due to levelling with contracted forms like _fi:s_,
_fi:mus_. Thus, an analysis /fioo/, /inicioo/ would refer to a stage
of the language preceding these analogies where it would apparently
cause no difficulies. [fi:o:] would still be an unambiguous reading
of /fiioo/ if there is no /j/ in hiatus (_major_ is /maiior/), but
[injicio:] just cannot be accomodated without a phonemic /j/.
I understand there is no conflict in _consilium_, gen. _consilii:_,
because this genitive form is only a post-classical restoration; the
authentical form is _consili:_ /consilii/, as shown by metrics and
inscriptions.
But again, whatever fate awaits the idea of a three-vowel system for
Latin should have no bearing on the analysis of Sanskrit which is
its own matter. I just hope it will be observed that this is not a
silly whim out of a loony-bin.
Jens