On 05-01-10 08:41, petegray wrote:
> Pardon me if I'm being slow. Can we discount the influence of the
> nominatives onus and Venus here?
I don't discount it. When I said that *-eto- could be blended with
*-os/*-es- here, I meant the influence of the s-final form of the stem
as seen in the nom.sg.
> Why can it not be a formation within
> Latin, based on Latin elements: the noun onus, Venus, etc, plus the
> usual -tus < *-tos participle/adjective ending?
Greek has <onostos> 'blameworthy', suggesting a deeper date, and cases
like <honestus> show that the phenomenon is pretty old within Latin.
> This pattern is well known and widespread, but unexplained. It is worth
> noting that these nouns with nom -or, gen -o:ris, are not neuter but
> masculine (I think without exception). Neuters from an original *-os also
> exist: eg frigus (n) beside frigor (m) but the commonest pattern by far is
> : verb in -eo, masculine noun in -or-, adj in -idus. These masculine -or
> nouns are widespead in Latin though rare in other languages, and many are
> likely to be innovations within Latin, rather than inherited. Does that
> mean the complex of -eo verb, -or noun, -idus adj could also be a purely
> Latin innovation?
That's true. Latin doesn't show <-idus-> next to neuters in <-us,
-eris>, but it has them regularly (with only sporadic exceptions) in
*-es- stems that developed into <-or, -o:ris> abstract masculines. I
suppose that, as you suggest above, a new morphological "package" was
stabilised in prehistoric Latin, including a series of productive
derivatives: stative -eo:, inchoative -e:sco:, abstract noun -or/-o:ris
(plus -e:do:/-e:dinis in some cases), adj. -idus. Synchronically in
Latin, -eo: verbs seem to have been treated as the starting point of
further derivation, so we normally don't find -idus unless there is also
a stative verb in <-eo:> (or at least an <-o:r-> noun, as in the case of
<cru:dus>). <nu:dus> is an interesting exception. But, whatever the
inner Latin developments, the elements of the complex are inherited;
Latin-speakers only rearranged them into a new configuration. My
original question was not so much about the function of <-idus> in Latin
derivational morphology as about its curious phonological form.
> Before positing new phonemes, should we not explore further the possibility
> of patterning among existing phonemes, along the lines of Jens's theory of
> the *s/*t/*h1 morpho-phonological complex, which you mention? Does such a
> pattern have to have a phonological basis?
It does according to Jens, who points to the complementary distribution
of *-s-/*-t- and *-h1- in related forms. He would be the best spokesman
for himself, but I hope I don't distort his original idea if I sumarise
it like this: the alternation between *s and *t in some morphemes
suggests that they represent the same PIE morphophoneme (symbolised as
//c// by Jens), presumably by virtue of their historical origin as
positional variants one and the same pre-PIE phoneme. Now in stative
verbs related to *-es- nouns we find *-éh1-/*-h1- (pres. *-h1-jé/ó-)
instead, as if //-ec// had become *-eh1- there (as e.g in Gk.
krátos/kratéo:). An alternation between [s] and [h] in some contexts is
phonetically plausible, so Jens prefers to treat the whole complex as
the product of an ancient phonological development, and I quite agree.
> And if it's a Latin innovation,
> might not the existence of a mere two or three inherited forms have provided
> an origin for this widespread pattern?
I agree in principle, but remain interested in the details of the
process, like the origin of the "two or three" inherited forms that have
given rise to the pattern.
Piotr