Basque 'sei'

[Larry Trask]

[ moderator re-formatted ]

Ed Selleslagh writes:

>>> Iberian has two sibilants (and also rhotics) that might very well be the
>>> same as the Basque ones, including the affricated varieties.

[LT]

>> Iberian certainly had two contrasting sibilants (at least), but the phonetic
>> nature of the contrast is entirely unknown.  Aquitanian probably had at
>> least four, and perhaps six, of the things, but the Roman orthography was
>> defective, and the various sibilants were not written in any very consistent
>> manner.

> [ES]

> The nature of the contrast of the two Iberian sibilants is not entirely
> unknown, as some toponyms have survived: e.g. Saitabi(a) > Xa'tiva,

Very interesting.  I didn't know that this etymology had been established.  Has
it been?

> which indicates that the S must have been rather closer to Basque (apical) s
> than z.

Why?

True, at a much later stage, a shushy sibilant in Arabic sometimes came into
Castilian as [esh], developing later to jota.  But on what basis can we extend
this pattern centuries further back into the past?

> Some Iberian words are extremely similar to some Basque words

I'd be very careful here.

First, it is not trivial to identify Iberian words at all.  Many Iberian texts
use a mark which l0oks like a word-divider, but not all do so, and not all do
so consistently.  For example, one Iberian text contains the following
unsegmented sequence:

	IKBAIDESVISEBAR'TAS'ARTIDVRAGVNAN

Decisions about word-boundaries in such cases are difficult and debatable.

In practice, nothing much interesting can be said about Iberian words.
Instead, it is the seemingly recurrent morphs which attract attention.  These
are presumed to represent morphemes, and some of them do indeed resemble items
in Basque.  However, in almost no case do we know the meaning of the Iberian
item, and hence the comparisons must be strictly of forms -- not very
illuminating.  The few Iberian morphs to which meanings can be assigned with
some confidence do not, in general, look like anything in Basque.

In its phonological structure, Iberian seems to be rather similar to Basque,
and hence it is perhaps not very surprising that similar sequences exist in
both languages.  But, when we don't know the meanings of the Iberian items, we
have no right to claim that Iberian and Basque possess "extremely similar
words".

For example, Iberian has a recurrent element BIOS-, which looks like Basque
<bihotz> 'heart'.  But we have not the faintest idea what the Iberian item
means, and there is no more reason to assign it the meaning 'heart' than there
is to assign it any other meaning.  Why not 'life', for example, given Greek
<bios> 'life'?

Finally, recall that Basque is of no more assistance in reading Iberian than
is, say, Norwegian or Zulu.  This must count for something.

> (I know you don't accept this to be anything but coincidence),

There is no reason to see anything other than coincidence when all we have is
resemblances in form, with no meanings attached.

> and these similarities always point to a systematic correspondance of the two
> Iberian sibilants and Basque s and z.

Really?  This is news to me.  What evidence can you adduce to support this
claim?

> From toponyms and other words one can deduce that Aquitanian and Iberian also
> had the corresponding affricates (ts and tz). In Iberian script these are
> written exactly as the non-affricated ones. In other scripts and in
> Aquitanian in Latin script orthography is pretty confusing, but often
> resembles later usages.

The Iberian script distinguishes only two sibilants, and I know of no hard
evidence that the language actually had four.

The defective Roman script used for writing Aquitanian doesn't distinguish any
sibilants at all very clearly, except that X or XS seems to have represented
affricates, while S and SS appear to have represented fricatives.

> I don't know where you found that Aquitanian may have had two additional
> sibilants.

This derives from the conclusion that Aquitanian represents an ancestral form
of Basque, more or less identical to the Pre-Basque reconstructed from
Basque-internal evidence.

Modern Basque has the following sibilants: laminal <z tz>, apical <s ts>, and
palato-alveolar <x tx>.  Now, one feature of our reconstruction is that only
the first four could ever appear in the unmarked forms of lexical items, while
<x tx> were entirely confined to expressive variants -- mainly diminutives.  We
therefore surmise that Pre-Basque -- and hence Aquitanian -- *may* also have
had these last two, even though they don't show up clearly in the written
records.

Now, I know of one piece of evidence suggesting that <x tx> may indeed have
been present in Pre-Basque and in Aquitanian.  Basque has an ancient diminutive
suffix <-to>, long unproductive but still present in a handful of fossilized
forms, such as <neskato> 'little girl', from <neska> 'girl'.  As is normal with
diminutive suffixes, this one long ago underwent spontaneous palatalization to
<-txo>, the productive diminutive suffix today.  The palatalization of <t> to
<tx> is completely regular.

Now, Aquitanian appears to exhibit <-to>, for example in the female personal
name NESCATO, to be identified with modern <neskato>.  But it also exhibits a
frequent suffix written <-xo> or <-xso>, which presumably contains an
affricate.  For example, the Aquitanian female name ANDERE, identified with
modern Basque <andere> 'lady', is seemingly related to the Aquitanian female
names ANDEREXO and ANDEREXSO.  We surmise that this -XO ~ -XSO may represent
the modern palatalized diminutive suffix <-txo>.

>>> The Castilian s and z/c (theta) are the descendants of the old Basque-type
>>> distinction, I believe.

[LT]

>> I don't follow.  Castilian /s/ simply continues Latin /s/, except that it is
>> apical, whereas the Latin /s/, on the Basque evidence, was probably laminal.
>> But the Castilian theta derives ultimately, in most cases, from Latin /k/
>> before a front vowel; this is thought to have become some kind of affricate
>> before developing into theta (or into /s/, according to region).

[ES]

> In derivations from Latin this true, but in all other cases it is not. I
> didn't mean that the Castilian s/z distinction is descended directly from the
> Basque apical/laminal opposition, as a parallel evolution in particular
> words, but that its very existence is due to a pre-existing awareness of such
> a phonological distinction (it did in Iberian), something most European
> languages don't have (and ditto for the rhotics).

"A pre-existing awareness"?  Strange wording, almost mystical.

May I translate into terms more familiar to me?

I presume the suggestion is that the Castilian s/z contrast derives from areal
pressure, from the influence of neighboring languages that already had it.

But the modern s/z contrast dates only from the 16th century -- rather late for
areal influence from Iberian, I'd say, or even from Basque -- which in any case
has a distinction decidedly different from the Castilian one.

As for other European languages, I will remind you of Martin Joos's 1952 paper
in Language, in which he argued that an apical/laminal contrast in sibilants
was in the medieval period widespread in Europe.

>>> What about Arabic? It certainly has various sibilants.

[LT]

>> Yes, but it does not have an apical/laminal contrast, and I know of no
>> evidence that Arabic phonology had any effect on Castilian phonology, still
>> less on Basque phonology.

[ES]

> As far as Castilian is concerned, I am not so sure: does any one have any
> references or information on that subject?

No, but I no of no evidence or argument that Arabic ever had the slightest
effect upon Castilian phonology, apart from the forms of a few individual words
that entered Castilian after Arabic mediation, like <jabo'n> 'soap'.

> In Basque, it is rather the opposite: e.g. kuttun < kitab, i.e. Arabic words
> were adapted to Basque phonology.

Yes, and this is what normally happens in borrowing.

By the way, Basque <gutun> ~ <kutun> ~ <kuttun> (and other variants), which has
been applied to various kinds of written things, is thought to derive from the
Arabic plural <kutub> 'books'.  This appears to be one of the rare cases in
which Basque has taken over an Arabic word without Romance mediation.

Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
UK

larryt at cogs.susx.ac.uk