txakur/dzhagaru/cachorro....

[Eduard Selleslagh]

----- Original Message -----
From: "Douglas G Kilday" <acnasvers at hotmail.com>
Sent: Monday, January 15, 2001 12:19 PM

[snip]
> 3 of the 4 possible variations of /a:e/
> correspondence are found in presumed substratal words:

>   Lat. cerrus, It. cerro, Sp. carrasco 'holm-oak'
>   Lat. larix, It. larice, Sp. alerce 'larch'
>   Lat. betula, It. betulla, Sp. abedul 'birch'
>   It. cheppia, Sp. sa'balo, saboga, saboca 'shad'

> These are practically the only examples I have, so I don't know whether the
> vowel-alternation and prosthetic /a/ are strictly determined by phonetic
> environment.
[snip]
> DGK

[Ed]

In the case of Castilian you have to take into account the Arab influence:
sometimes words of Romance origin were adopted by the Spanish-Arab population
and then got back into Castilian in their arabicized form, often with the
definite article al- (and its alternate forms, depending on the assimilation of
the initial "sun" consonant of the following word) attached to it. <alerche> is
such a case: actually <al-lerce>, possibly via some Arabic form **al-lars vel
sim..  <abedul> is possibly a simplification of *al-bedul (no assimilation: b
is a "moon" letter). I don't think these a's are prosthetic/epenthetic like the
e of <especial>.

Note also that Arabic has not really a vowel /e/, except in some regional
speech, nor /o/. In the transition to Spanish, the Arabic vowels often undergo
surprising changes, or are added. Examples: <alcázar> from Lat. castra, via
al-kasr. Cast. Alicante from Lat. Lucentum, via al-(lu)kant, but in Catalan:
Alacant; it is pretty strange that Lat. c suvived as /k/, since the Arabs came
there (713 A.D.) after the palatalization of Lat. c (unless some regional
peculiarity intervened, like an early dialectal /e/ > /a/, or maybe the Iberian
name that served as a substrate to Latin had an /a/ - I really don't know).

<Carrasco> has all the characteristics of a somewhat complicated origin: it is
almost certainly a compound, with the suffix -(V)sco, which can be IE but just
as well Iberian or Basque, even though that wouldn't affect its meaning. I
would guess that the Latin form is derived from a substrate word with /a/. The
Spanish word cannot possibly be derived directly from the late-Latin form,
because the Latin c would have become /T/ (English th), not /k/ [In Sp. cerro
means 'small mountain, hill']. On the other hand, no such objection exists for
It. cerro.  Could <carrasco> and Lat. cerrus /kerrus/ be related to a pre-IE
root and/or Celtic, for a certain type of mountain landscape? In such case, the
suffix -sko would make a lot of sense. Just a thought.

Ed. Selleslagh