On Wed, 27 Nov 2002 11:52:22 -0000, "Richard Wordingham"
<
richard.wordingham@...> wrote:
>Are you saying that the post zero-grade consonant stem forms included
>inflected forms that retained -ln-? If not, when could *sulniko have
>arisen?
>
>> >The suggestion of pre-Germanic -ln- > -nn- bothers me slightly; I
>> >though we had PIE *kl.ni- 'hill' > Germanic *hulli- (OE hyll, at
>> >any rate) 'do.', Latin collis 'do.'.
>>
>> What I said only applies if the *-ln- cluster was pre zero-grade.
>> The "hill" words are from **qVlH-Vn- or **qVlH-mVn- with some material
>> originally between the *l and the *n.
>
>Are you suggesting that *-ln- from before zero grading was different
>from *-ln- arising as a result of it? Or are you suggesting that we
>are seeing dialect differences that predate zero grading?
The latter. Or, to put it more carefully, we see "pre zero grade"
dialectal variation in this particular word, with Slavic opting for
*ln in <slUnIce> < *sh2uln-ik-om, Germanic and Avestan having *n(n) in
the oblique, and everywhere else only *l(l), presumably analogical
after the NA sg. (Skt. suàr, sú:ras and sú:rya- (yo-stem), Grk.
*ha:welios (yo-stem), Lat. so:l, so:lis, We. haul, OIr. su:il "eye"
(*suh2li-), Alb. diell "star" (< *s(h2)uel-), Goth. sauil, ON so:l
[besides n-stem Goth. sunno:, OHG sunno, OE sunna from the oblique],
Lith. sáule: (ya:/ye:-stem), Hitt. {d}UTU-liya (yo-stem).
To know whether this holds true in other words, and what the other
groups make of *ln medially, we need more evidence, and so far the
only other *ln-stem I've been able to find is the "eagle" word. The
circumstances are different in that word, because it contains a medial
*r, because it's not neuter but masculine, and it has a static
declension. I reconstruct:
N **xá:r-aln-z
A **xá:r-aln-m
G **xa:r-áln-as,
which should have given in the **-ln- > *-n(n) dialects:
N *h2óroN
A *h2órNm.
G *h2árN(o)s
This is consistent with what we find in Germanic (ON ari ~ o,rn, OE
earn, OHG aro ~ aru, Goth ara, from nom. *aran, ON from acc. *arnum),
Hittite ha:ras, haranas (< *h2órons, *h2árnos) and Greek orni(:)s G.
orni:thos/orni:khos [what's that suffix?], órneon (from acc. *orn-).
Slavic shows no *ln this time, but *l(l), as if from a paradigm:
N *h2óroL
A *h2órLm.
G *h2árL(o)s
This explains OPr. arelie (= arelis?), Lith. ere~lis (dial. are~lis),
Latv. è:rglis (< è:rdlis) [is that dl < ll?] and OCS orIlU. The *r in
the root is probably responsible for such forms as OIr. irar, ilar,
We. eryr, Bret. erer (*erur-) and Armenian urur, oror, which then also
would go back to the paradigm with *L.
We have no Indo-Iranian, Latin, Tocharian or Albanian attestations.
=======================
Miguel Carrasquer Vidal
mcv@...