[UTF-8 encoded, like Mark's original
message]
Walde & Pokorny ignore laryngeals as a
principle. If the Don's name belongs here at all (<speculation>
Tanais may be a "Cimmerian" name with the Thracian analogue of
Grimm's Law; in SE Poland we also have Tanew <
*ta:nu:, which might in turn be Gothic-influenced, as the
river runs close to a very early settlement area of the Goths </speculation>),
the short vowel might be the reflex of a vocalised laryngeal
(*da:-/*d@- in Brugmannian terms), but I think
a more convincing explanation is possible. The reduced vowel
*U which occurs in some Slavic river-names
(*dUnEprU 'Dnieper', *dUnEstrU
'Dniester') is the reflex of unstressed Iranian *a (also
shortened *a:) in a pretonic syllable. A similar substitution
is visible in the numeral '100', Slavic *sUtó < Iranian
*satá- (the regular Slavic reflex of the IE word would be
**sInto or **sIntU) and in the verb
*xUt-je-/*xot-je-'want' (most likely from
Iranian *xat-yá- < *snt-jé-, cf. Polish
chęć 'willingness' < *xant-i- <
*sont-i-). The variation
*U/*o in the verb is unmotivated
phonologically; *don-/*dUn- may be a similar
case of an unstressed back vowel being replaced by dialectal Slavic
equivalents.
The "*o/*u
ablaut" can't be taken seriously. Bringing in an isolated Latvian word that
doesn't quite fit semantically is a shot in the dark. In etymologically secure
cases like nox/nuks u
is characteristically Greek and ALWAYS occurs next to a reconstructed
labialised velar stop or *w (cf. such celebrated items as
hudo:r, gune:,
kuklos). The vowel quality must be due to the colouring
effect of the "dark" consonant (labialised stops lose their labial component and
have plain velar reflexes in such cases!) and definitely has nothing to do with
IE alternations. Professor Schmid should be aware of that, as the problem has
often been mentioned in the literature.
Greek dan-, on the other hand, is
easily relatable to da:n- via the normal ablaut pattern
{dex+n} : {dx+n}, though any analyst who wants
to connect Da:nuvius and the Danaoi
should at least risk a prediction specifying the forms in which
the nil grade has a right to occur. Also the second -a- in
Danaoi remains difficult to explain even if we admit
*d(a)x-n-o- as a variant of *dax-n-(e)u-
(like, say, Germanic *xurna- < *kr-n-o-
'horn' vs. Latin cornu:) or equate the Danaoi with Ramesses
III's dnjn ("Denyen"/"Danuna", one of the Sea Peoples; this
immediately brings to mind the story of Danaus and Aegyptos). We would probably
need something like *dx-n-ax-(j-) as well --
which means a constellation of hypothetical forms with very little tangible
attestation. These connections are extremely tempting, but we're on slippery
ground here. Suspecti sunt Danai et dona ferentes.
Piotr
----- Original Message
-----
Sent: Wednesday, August 23, 2000 1:10
AM
Subject: Re: [tied] Danaans [was
Poseidon]
Mark
writes:
The dictionary of Walde-Pokorney .. has only
the root *dā- and *dē- without explaining either the quantity or
the o-vocalism of "Don" (or the quantity of its o-vocalism). But since there is
an initial dā-na for Sanskrit, isn't it possible that the root had a
final laryngeal originally? The Greek word δημός ('fat', 'fodder',
'nourishment for beasts') could possibly be attached to "-dan ". Then,
being analyzed further as *dā-m-, it would remind of *dā-n-.
Nevertheless, I think that the most plausible explanation is given by W.P.
Schmid. He interprets the difference in quantity between dā-nu and
Δαναοι with quantitative ablaut relations, *danos being the outcome
of the zero grade (Schwundstufe). He also brings the forms with -o- with
those with -u- together, dan/dun- as in Latvian danava,
dunava /'pool, puddle'/. The vowel gradation has then several parallels,
e.g., nox - νύξ , also calix - κύλιξ, Lithuanian nagas /'nail'/ - óνυξ, etc. It
is clear, he writes, that the Δαναοι must have had their homeland in the region
which had the zero grade form dan-, not that of
dā-nu.