On Wed, 15 Mar 2006 12:27:24 -0500, andrew jarrette
<
jarretteandrew@...> wrote:
>Thanks. But why do you suspect that most neuter u-stems were actually
>r/n heteroclitics? I fail to see the connection.
It's a conclusion drawn from two sources. The first follows
from my speculations about pre-PIE vocalism, where I suspect
that an earlier 3x2 vowel system (*/a/, */i/, */u/; */a:/,
*/i:/, */u:/) fell together as */e/; */o/, */e:/ (and zero).
The effect of stressed *ú(:) on a following *t is
interesting: it turns up as PIE *s. The standard example is
*méh1no:ts, *m(e)h1nésos "moon, month", which I would derive
from **mát-nu:t-s, *mat-nút-a:s (> *m@...:s > *m@...
> *meh1nésos). When the *u stands at the beginning of a
morpheme, it goes to *we/*u instead of *e, as in the pf.
ptc., say *wéid-wo:t-s, *wid-ús-os.
If *t is labialized by a preceding stressed *u (and *k, *g,
*gh are labialized to *kW, *gW, *ghW by _any_ neighboring
*u), perhaps other coronals show the same phenomenon. The
obvious place to look is *n, wich may have been labialized
to *nW > mW > m ~ w.
The second source is the curious Armenian phenomenon whereby
u-stem neuters, traditionally reconstructed as ending in *-u
in the NAsg. show up as having final -r. We have asr
(*pek^u/*pok^u, vocalism from oblique *pk- > a-), artawsr
(*drak^u), cunr (*g^onu) and several others. There has been
lots of speculation about the origin of this -r (e.g. that
it comes from *-s > *-z, which is impossible because it's
precisely the _neuters_ that show the -r).
Taking the two together, I arrive at the following
reconstruction:
NA *g^á:n-un "knee"
G *g^a:n-ún-a:s
DL *g^a:n-ún-a(i)
IAb *g^a:n-ún-a:t
pl. *g^á:n-un-(a)h2
Which by regular soundlaw would have developed into:
NA *g^ón-ur
G *g^én-w-os
DL *g^én-w-i
IAb *g^én-w-ot
pl. *g^ón-un-(a)h2
This distribution is found in the Armenian u-stem
adjectives, which are n-stems in the plural:
NA barjr *bh(é)rg^hur
G barju *bhr.g^h(u)wós
pl. barjunk` *bh(é)rg^hunesW (through *bhr.g^hónesW)
This also might explain the interchangeability of u-stem
(really (w)r-stem) and ro-stem adjectives.
A word like *dóru(r) "tree, oak", which happens not to be
attested as such in Armenian [I would expect *taLr, with
dissimilation as in meLr < *merur < *medhur "honey", or
perhaps *artur, with generalization of the oblique root],
may be derived as:
*dá:r-un > *dóru(r)
*da:r-ún-a:s > *dérwos
or as:
*dú:r-un > *dóru(r)
*du:r-ún-a:s > *drunós ~ *dr(u)wós (depending on the
relative timing of the accent shift versus the labialization
after stressed *ú). Cf. Ved. dá:ru, G. drún.as.
=======================
Miguel Carrasquer Vidal
mcv@...