ara/ndano

Larry Trask larryt at cogs.susx.ac.uk
Fri Mar 19 15:50:02 UTC 1999


On Thu, 18 Mar 1999, Xavier Delamarre wrote:

> 5/ Basque _arhan_, _aran_ 'plum' is obviously from continental
> Celtic *agran(io)-, exactly as _andere, azkoin, bezu, izokin,
> gereta, mando_ etc. are from Gaul. _andera:, bessu, eso:ks, cle:ta:,
> mandu,_ .

Well, I'd be a bit cautious about asserting that these items are
"obviously" of Celtic origin.  In fact, only one of them appears to be
indisputably of Celtic origin.

The problem with deriving <aran> `plum' from *<agranio-> is that this
would violate the usual rules for adapting loan words to Basque
phonology: we would have expected something like *<agaran-> in Basque,
not the observed <aran>.

Basque <andere> `lady' has long been suspected of being borrowed from
Celtic, given the existence of a possible Celtic source represented by
Old Irish <ander> `young (woman)'.  But there is another possibility,
pointed out a few years ago by Joaquin Gorrochategui.

G notes that the Aquitanian ancestor of Basque exhibits both a female
name ANDERE and a male name ANDOSSUS, with ANDOS(S)- also appearing as a
first element in other male names.  He proposes that the second
represents a Pre-Basque word *<andots> with a Latin ending, and he
proposes to derive both <andere> and *<andots> from a stem *(and->, with
a female suffix *<-ere> in one case and the attested male suffix
<-(d)ots> in the other.  If <andere> is `lady', then this *<andots> must
be `lord'.

Support for this idea comes from the observation that, while `lady' is
<andere> down to the present day, the word for `lord' in the historical
period is <jaun>.  Now, this word has a very strange form for a noun in
Basque: normally we find initial /j/ only in non-finite forms of verbs,
and nowhere else.  As it happens, <jaun> looks for all the world like
the perfective participle of a verb, and so we may surmise that <jaun>
might derive from the participle *<jaun> of an otherwise lost verb,
meaning perhaps `exalted' or something similar, and having displaced the
earlier word for `lord'.

All this sounds more than plausible to me.

Basque <azkoin> `badger', with *many* regional variants, is a puzzle,
but few have suspected a Celtic origin.  Most of the variants point to
an original *<azkone>, and many commentators have been quick to suppose
a borrowing from Late Latin <taxone> `badger', itself reportedly of
Germanic origin.  Admittedly, though, the borrowing of <taxone> as
*<azkone> would be somewhat unusual phonologically.

A complication is the Zuberoan variant <harzku~>, which appears to point
to an earlier *<arzkone>.  This leads some observers to suggest a
derivative of native <(h)artz> `bear', with an unidentifiable second
element.  It is true that a badger somewhat resembles a bear cub.

As for a putative Celtic origin, I can find no suggestion beyond that of
Wagner, who wants the Basque word to derive from a Celtic *<brokkos>,
the source of regional English `brock', but just how this is supposed to
yield *<azkone> beats me.

The strictly Bizkaian Basque <bezu> `habit, custom' has sometimes been
derived from a supposed Celtic *<bessus>, but most specialists from
Schuchardt on have rejected this, not least because it appears that no
such Celtic word exists.  Schuchardt and Michelena prefer to see the
word as deriving from some Romance reflex of Latin <vitiu> `fault,
defect', which develops the sense of `habit, custom' rather widely in
western Romance.

Basque <izokin(a)> `salmon' is widely suspected of being *ultimately* of
Celtic origin, but not *directly*: instead, a Celtic loan into Late
Latin is favored by most commentators as the direct source of the Basque
word.

Basque <gereta> (and variants) `rustic gate' (and other senses) is
likewise suspected of being ultimately from Celtic, but probably
directly from a Late Latin *<cleta>, whose reflexes are prominent in
western Romance.  A puzzle here is the form of the Basque word, since,
by the usual rules, *<cleta> should yield a Basque *<leta>, and not the
observed <gereta>; it may be that a Romance development *<keleta> has
intervened, since this would yield the attested Basque form
straightforwardly.  If not, then the word has developed somewhat
irregularly, though not wildly so.

Finally, <mando> `mule' is the one word here which is more or less
universally believed to be diectly from Celtic.  The word is reportedly
found widely in IE languages to denote various equine animals, but the
sense of `mule' is said to be specifically Celtic.

The paucity of direct Celtic loans into Basque is a puzzle, since we
know that an ancestral form of Basque was in contact with Celtic for
centuries before and after the Roman conquest.  Michelena once hazarded
the suggestion that more Celtic loanwords into Basque must have existed
once, but that many were either displaced by Latino-Romance loans or
simply re-formed under Latino-Romance influence.

A Celtic origin has also been suggested for <arrau> ~ <arraun> `oar',
pointing to earlier *<arranu>, possibly to be identified with the word
represented by Old Irish <ram(e)> `oar'.  Don't know if there's anything
much behind this, especially since the nasal appears to be wrong.

Finally, <maite> `beloved' has been thought for a century to have been
borrowed from the Celtic word represented by Old Irish <maith> `good'.
The semantics requires some fancy footwork, but the Basque word has an
anomalous form for a native word, and is surely borrowed from somewhere.

I should add that Vasconists and Romanists have at times been rather
eager to see Celtic solutions for problematic words, but that a
Celticist colleague has pointed out to me that some of the convenient
Celtic solutions put forward do not appear to have existed in Celtic.

Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
UK

larryt at cogs.susx.ac.uk



More information about the Indo-european mailing list