andera 'woman' Celtic ?

[Miguel Carrasquer Vidal]

Larry Trask <larryt at cogs.susx.ac.uk> wrote:

>It has been debated for generations whether Basque <hartz> `bear' might
>be borrowed from Celtic *<artos> or from a related IE form (this `bear'
>word is certainly IE).

>The Celtic source is phonologically awkward, though: we would have
>expected *<artos> to be borrowed as *<(h)artotz>, not as the observed
><hartz> -- assuming, of course, that it was the nominative that was
>borrowed, but the accusative would be even less suitable.

And assuming early Celtic borrowings were subject to the same
phonotactic constraints as Latin borrowings.

>One or two people have suggested an IE source other than Celtic, but no
>IE language with an even vaguely suitable form of the word is known to
>have been in contact with Basque.

>Just to complicate matters, it is very unusual for a Basque lexical item
>to end in a consonant cluster.  Save only for <beltz> `black', which we
>have good reason to suppose is a syncopated form of earlier *<beletz>,
>all seemingly native words which end in a cluster end in <-rtz>, like
><hartz> `bear' and <bortz> `five'.  There are some grounds for supposing
>that these clusters too are secondary and result from some kind of vowel
>loss, but the evidence is not sufficient to support this view securely.

The only logical explanation is indeed *(h)a'rtos > hartz, with
(pre-Roman period) loss of unstressed vowel.  But we should maybe
look for more Celtic borrowings with nom. -(V)s giving Basque
-tz.

=======================
Miguel Carrasquer Vidal
mcv at wxs.nl
Amsterdam