On Thu, 20 May 2004 06:20:25 +0000, Rob
<
magwich78@...> wrote:
>--- In cybalist@yahoogroups.com, Miguel Carrasquer <mcv@...> wrote:
>
>> That doesn't explain why some roots are declined both
>> thematically and athematically (e.g. *weiks, *wik- ~
>> *woikos).
>
>Perhaps the latter is an adjectival form of the former.
Perhaps. But it shows that thematicity is not a function of
the root shape.
>> There is a difference between root nouns and composite
>> nouns. Root nouns are always accented on the final syllable
>> (perhaps the penultimate at an earlier stage). Composite
>> nouns come in two basic varieties: proterodynamic (PD), with
>> accent on the stem in the strong cases, on the suffix in the
>> weak cases; and hysterodynamic (HD), with accent on the
>> suffix in the strong cases, on the desinence in the weak
>> cases. A third type are the collectives, with accent and
>> lengthening of the vowel of the suffix.
>
>Aren't 'suffix' and 'desinence' the same thing?
No. In a compound stem like nom.sg. *h2ák^-mon-z (>
*h2ák^mo:n), the root is *h2ak^-, the suffix is *-men-, and
the desinence is *-z.
>> The "dog" word seems to be a "collective", where the long
>> suffix vowel also causes stress retraction:
>>
>> nom. *k^awá:n-z > *k^wó:n
>> acc. *k^awá:n-m > *k^wónm.
>> gen. *k^awa:n-ás > *k^wéns ~ *k^úns (regularized to *kúnos =
>> Ved. s'únas).
>
>I thought the long o: appeared from compensatory lengthening due to
>the loss of *-s.
No, it's not compensatory lengthening. The vowel is
lengthened even if the *-s remains (**pód-z > *pó:ds).
There are in fact two rules:
(1) V > V: /_C(C)z#
(2) z > 0 / R_#,
that do not necessarily go together.
Another question is the o-grade in the strong forms
(*k^won-), which I derive from an old long vowel **a: (>
*o). In my view, this is independent of the nominative
lengthening, which is why the /o/ (short, of course) also
appears in the accusative.
In this case, the long vowel is simply a consequence of the
morphology: compound nouns make their collective form by
lengthening the suffix vowel and stressing it [+ a rule
against two consecutive long vowels] (*wá:d-an > *wódr,
coll. *wad-á:n > *udór+h2 > *udó:r).
In the case of standard proterodynamic nouns, the o-grade is
caused by an automatic lengthening of the posttonic
(svarita) syllable, _if_ the preceding syllable is light:
*h2ák^-man-z > *h2ák^-ma:n-z > *h2ák^-mon-z > h2ák^-mo:n
*h2ák^-man-m > *h2ák^-ma:n-m > *h2ák^-mon-m > h2ák^-mon-m.
*h2ak^-mán-as > *h2ak^-mán-a:s > *h2k^-mén-os
Notice that this also applies to desinences of the shape
-VC, whence the /o/ in the gen.sg. and abl.sg. (but no
lengthening in the loc.sg. *h2ak^-mán-a > *h2ak^-mán(i)).
There is one case where /o/-grade *is* caused by the
nominative ending, and that is in proterodynamic nouns with
a heavy root. An example is *pu:nt-ah2- "path":
nom. *pú:nt-ah2-z
acc. *pú:nt-ah2-m
gen. *pu:nt-áh2-as
loc. *pu:nt-áh2-a
As the root is heavy (but the suffix light), the svarita
lengthening rule only affects the genitive ending:
nom. *pú:nt-ah2-z
acc. *pú:nt-ah2-m
gen. *pu:nt-áh2-a:s
loc. *pu:nt-áh2-a
This develops into:
nom. *pó:nt-&h2-z
acc. *pó:nt-&h2-m
gen. *p&nt-áh2-o:s
loc. *p&nt-áh2-&
At this point, the suffix, which is reduced (*-&h2-) in the
strong cases, loses its stress to the long vowel of the
desinence (the paradigm becomes amphidynamic):
nom. *pónt-&h2-z
acc. *pónt-&h2-m
gen. *p&nt-&h2-ós
loc. *p&nt-áh2-&
Szemerényi lengthening occurs at this point (it *has* to be
before zero grade, and also before the shortening of /o:/ in
closed syllables), and a schwa affected by it lengthens to
something that eventually turns up as /o:/:
nom. *pónt-o:h2-z
acc. *pónt-&h2-m
gen. *p&nt-&h2-ós
loc. *p&nt-áh2-&
After zero grade we have:
nom. *pónt-o:h2-z
acc. *pónt-h2-m
gen. *pnt-h2-ós
loc. *pnt-áh2,
leading to:
nom. *pónto:s
acc. *pónthm.
gen. *pn.thós
loc. *pn.tá:(i) (analogically: *pn.thí)
=======================
Miguel Carrasquer Vidal
mcv@...