Celtic Germanic relationship

[Gabor Sandi]

>From: X99Lynx at aol.com
>Date: Fri, 1 Dec 2000 00:49:14 EST

[ moderator snip ]

<Gabor Sandi's reflections on this:>

I think that the evidence is fairly clear as to the direction of the
borrowing for these words.

In the case of the "iron" words, the Scandinavian words have no sign of the
intervocalic -s- (eg Swedish jarn). As intervocalic *-s- does not get lost
in Germanic languages (as opposed to turning into -z- > -r- under Verner's
Law under certain conditions), the best explanation is that the northern
prehistoric Germanic dialects/languages borrowed the word for 'iron' from a
language where the intervocalic *-s- had disappeared. Since a similar change
indeed affected the ancestors of the known insular Celtic languages (in
Irish, *-s- was "lenited" into -h-, then nothing, whereas in Welsh and
Breton all *s > *h, with eventual loss in many cases), it is reasonable to
posit an *-s- > *-h- > nothing for some continental Celtic languages as well
for the 1st millennium BC. Other Celtic dialects/languages would have kept
intervocalic -s- (I believe Gaulish did), and it is from such a source that
the ancestors of English iron and German Eisen would have been borrowed. It
is noteworthy that English has an /r/ where German has a /z/, suggesting
(using Verner's Law) that the reconstructed Celtic source had variable
stress: */'i:sarnos/ or */i:'sarnos/, with the further suggestion that
borrowing predated the action of Verner's Law, and therefore the PGermanic
shift of stress to the initial syllable.

On *ri:k-, the key to the direction of borrowing is not so much the
consonant /k/ but the vowel /i:/. If we assume an IE etymology for the word
in question, we are more likely to find a PIE */e:/ than a PIE */i:/
(evidence: Latin rjx, rjgnna; Sanskrit rbj-). Now, PIE */e:/ either remains
*/e:/ in the Germanic languages or opens up to a long low vowel /ae:/ or
even /a:/. The conditions for these alternative developments go well beyond
what we are discussing, but the point is that PIE */e:/ never becomes */i:/
in Germanic. This is, however, the normal development in the Celtic
languages, so that IF we posit an IE origin for the Germanic word family
*ri:k-, the best explanation is again borrowing from Celtic, with the
borrowing taking place before the Grimm's Law change *g > *k.

As a subsidiary question: if the borrowing went the other way, how could a
Germanic voiceless stop *k have been borrowed with a voiced stop *g into
Celtic? Whatever funny stuff Irish did to its consonants through its complex
history, it always kept *k and *g well apart, with voicing occurring (at a
later period) after nasals only (this is called convection, I believe - any
Celtic scholars out there to confirm this?). If the proposed borrowing
predates the action of Grimm's Law, the Germanic */i:/ would still remain
unexplained, of course...

With best regards,

Gabor Sandi