Pre-Basque phonology (part 2)

Larry Trask larryt at cogs.susx.ac.uk
Fri Sep 17 16:11:38 UTC 1999


On Wed, 15 Sep 1999, Roslyn M. Frank wrote:

> At another point in our earlier discussion, I asked whether the
> monosyllabic root-stems in southern Basque dialects (which have
> aspirated bisyllabic counterparts in northern dialects) were
> considered bisyllabic, i.e., whether the forms in the northern
> dialects were given preference in reconstructions.

Yes, and with good reason, as explained earlier.

> Specifically we were speaking of root-stems with an initial /b/.
> Larry's response indicated that the reconstruction took the northern
> variants as representing the earlier stage and, therefore, the
> southern ones would show a loss of aspiration and consequently a
> falling together of the two syllables into one.

Absolutely.  This is the only interpretation consistent with the
evidence.

> Yet it seems to me that there may be other ways to look at the
> problem, particularly since the status of aspirated consonant series
> in pre-Basque doesn't seem to be fully understood.

True, though in Pre-Basque it seems likely that aspiration was probably
entirely non-distinctive, and that the marginally or substantially
contrastive value which it now has in the north results from other
changes.

> Let me suggest two alternate scenarios and then I would like others
> to explain why these are excluded as explanations.

OK, but, looking ahead, note this.  It is usually easy to invent all
kinds of scenarios that are more or less consistent with the data.
But the issue is the evidence that can be adduced in favor of any given
proposal.

> For our examples we shall use the bisyllabic root-stem /behe/ (N)
> and its monosyllabic counterpart /be/ (S) whose meaning in both
> dialectal groups is `below'.

To be honest, I am not aware that <behe> `below' exists today as an
independent word in the north, though it is attested there in older
literature, and its former presence is assured by its numerous
derivatives, such as <behera> `downward' (<-ra> allative).  In the
south, the independent word still exists today; it is usually <bee> in
Bizkaian and <be> elsewhere.  As a suffix, it is everywhere <-be>, with
a variant <-pe>.

>  Alternate Reconstructions:

> 1) Let us assume that the parent form had a phoneme in */bh/ and
> that northern dialects retained it while the southern dialects lost
> it. This would mean that the parent form was monosyllabic in */bhe/.
> Next, over time the northern dialects lost track of the phonemic
> status of */bh/ and the aspiration was rearticulated as /behe/. Or
> even that there developed at some part an allophone of */bh/ in
> */beh-/.

> This would make the northern dialects more conservative in one sense
> for they would have retained a trace of the original aspirated
> phoneme, but also innovative in that, in the process, they converted
> the earlier monosyllabic item into a bisyllabic one. On the other
> hand, the southern dialects would be conservative in retaining the
> monosyllabic nature of the parent form, but innovative in terms of
> the loss of aspiration.

The problem here is lack of evidence, plus a degree of unnaturalness.

To begin with, aspirated voiced plosives like /bh/ are rare in the
world's languages, though certainly attested.

But there exists no trace of evidence in Basque for such sounds.  There
exists not a single Basque form recorded anywhere with an aspirated /b d
g/, and positing such unusual sounds requires something in the way of
evidence.

Moreover, as my last posting showed, it is not only words with initial
/b d g/ that are involved, but words with all sorts of initials.  For
example, to account for northern <nahi> `desire', southern <nai>, you
apparently need also to posit an aspirated nasal */nh/; for northern
<ziho> `tallow', southern <zio>, you need to posit an aspirated sibilant
*/zh/; and so on.  And what do you do about cases like northern <aho>
`mouth', <ahal> `ability', <erhi> `finger', and so on, all of which have
initial vowels?  An account that says nothing about all these other
cases cannot be preferred to the simple and obvious interpretation of
the mere loss of /h/ in the south.

> 2) Let us assume that the parent form had no aspirated consonants,
> i.e., that the earlier form of the item was */be/. Then let us
> further assume that at least some of the northern dialects developed
> aspirated consonants, a characteristic that might have also spread
> to some extent to the southern dialects but did not fully take hold.
> Over time the aspirated consonant (or allophone of */b-/) in */bh/
> which had developed in these dialects, perhaps under pressure from
> those dialects that lacked this characteristic, rearticulated all
> the monosyllabic CV root-stems in /bh/ into */bVhV/. In the case of
> the southern dialects the monosyllabic non-aspirated form */be/ was
> retained: there was no innovation.

But then how do you account for the widespread Bizkaian form <bee>, with
double /e/ but no aspiration?  Or for northern <zahar> `old', southern
<zaar> ~ <zar>?

And what about the loans from Latin, like northern <ahate> `duck',
southern <aate> ~ <agate> ~ <ate>?  What can this account say about
these, bearing in mind that the Latin source is <anate(m)>?

Anyway, a sequence *<be> > *<bhe> > <behe> is not something I've ever
seen before in any language.

> Canonical Reconstruction: 3) Let us assume that the parent form was
> */behe/, identical to that found in northern dialects today, as
> Larry has suggested. That reconstruction of events means the
> innovation/loss would have occurred in the southern dialects, while
> the northern ones would be viewed as more conservative (for this
> feature). This is the standard interpretation, if I understand
> Larry's remarks correctly.

It is.

> With respect to this interpretation I would like to ask what role,
> if any, was played in this scenario by the aspirated consonants of
> the northern dialects, e.g., /bh/.

No such segments exist, nor are they attested anywhere in Basque.
The only Basque consonants which can bear the aspiration are /p t k/
(word-initially, or at the beginning of a second syllable when not
immediately preceded by a sibilant), plus all liquids and /n/ (at the
beginning of a second syllable only).

> I would mention, as Larry has, that the alternation between /b/ and
> /p/ is a common feature in southern dialects as well as northern
> ones.

Yes, but the voicing alternations, in the vast majority of cases, have
well-understood origins.

A typical example:

	<gurdi> `cart' + *<bil> `round'

Rule 1: /i/ is lost at the end of a first element in word-formation,
except after /b/, when it remains:

	*<gurd+bil>

Rule 2: a plosive becomes /t/ before a morpheme boundary:

	*<gurtbil>

Rule 3: /t/ devoices a following plosive:

	*<gurtpil>

Rule 4: /t/ is lost before another plosive:

	<gurpil> `cartwheel' (the modern form)

All of these rules are entirely regular and abundantly documented in
Basque word-formation.  Not one is doubtful or *ad hoc*.

More examples:

	<begi> `eye' + <buru> `head' > <bepuru> `eyebrow'

	<begi> `eye' + <azal> `skin' > <betazal> `eyelid'

	<ogi> `bread' + <-gin> `maker' > <okin> `baker'

> Please excuse my simplistic descriptions of these hypothetical
> events. I am certain that others on the list can improve on them as
> well as point out whether (3) is, indeed, the best and/or only
> explanation that can be given to the data.

It is, certainly.

> Again, I insist that I have absolutely no interest in promoting one
> description over another, only in hearing from others on the list
> (and obviously from Larry) concerning their opinions as to which of
> the three hypothetical scenarios listed above best describes the
> data.  Also, I am interested in hearing any other reasons why one of
> the solutions *ought* to be preferred/adopted and the other two
> rejected.  I should mention that there are probably other
> solutions/scenarios that could be mapped in addition to the three
> sketched out here.

Yes, but only the standard view accounts satisfactorily for the data.
Since this also happens to be by far the simplest and most natural
interpretation, what possible ground can there be for querying it?

Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
UK

larryt at cogs.susx.ac.uk



More information about the Indo-european mailing list