comment-KabardianMonovocalism.htm






COMMENT

from John Colarusso, CIS, McMasters, CA



Subject: Re: Kabardian Vocalism and Vertical Vowel Systems


To: Patrick C. Ryan



"In "autosegmental" phonology, one has a "tier" for consonants, one for vowels, and then a "skeleton" or "melody" which is either a string of syllables or a string of CVCVCV, or CVCCV, etc.

If you use this technique (developed by Jonathan Goldsmith), then the issue of schwa (for which I use '6') is resolved as follows, with both warring camps in this matter being partly right.

Let us take the Kabardian word for 'alive' /psaw6/, which ends in an unstressed unpredictable vowel. Then a tier analysis looks like this:







Then one draws lines "licensing" the consonants to the Cs and the vowels to the Vs. The first two, /p/ and /s/ both go onto the first 'C', giving the sort of co-articulated "harmonic" cluster seen in Circassian (of which Kabardian is an eastern form) and in Ubykh, but not in Abkhaz or Abaza (which seem to have C(C)VC(C)V skeletons). The /w/ goes to the second 'C'.

The open vowel /a/ (some workers follow Circassian orthography and prefer 'e') goes to the first V and the schwa to the second. Now the trick here is to leave /6/ out of the V-tier and to stipulate that /a/ cannot be licensed progressively onto the second V.

Then to have some phonological form for the second V one must use a "default" vowel, namely /6/.

So, is /6/ predictable? Yes.

Is the shape of the word predictable? No.

Therefore, are all /6/s predictable? No, not if the word skeleton has some unusual shape. In fact all unpredictable /6/s are confined to core vocabulary, where word shape is less predictable than in peripheral vocabulary.

As to the issue of "long" or "open" /a/s, that is whether there is a third vowel or not, then one can repeat the argument, thus:







and perform the usual licensing with one added detail. The second /a/ is licensed retrogressively back onto the offset of the second syllable, that is, it is licensed onto the second C. This is because the onset of the second syllable colors the offset of the first unless the second syllable has an open vowel.

Then the first syllable will have one that is extra open and will fill its own offset (or coda, as it is sometimes called).

So long /a/ is predictable for the vast majority (all but one or two) words with the form CaCa. It reduces when the syllable is unstressed. Long /a/s that do not reduce when unstressed seem to arise from a [+low] segment in the C-tier, which, from morphological considerations, seems to be linked to the plural, which is /-ha/ in word final position. So such non-reducing /a:/s appear to be underlyingly /ha/ (with a predictable flip to /ah/).

Crucially for language evolution/ history, one should remember that the evidence from this family points to an earlier vocalic system with normal coloring.

The vowels then gave up their colors to the consonantal periphery, retaining only [+low] because this cannot go on to a consonant unless it is a laryngeal (and perhaps a pharyngeal or epigottal/adytal).

It seems that when one hears a syllable, the acoustics of its periphery and nucleus largely overlap and it is a matter of "cognitive choice" as to whether one assigns such coloring to vowels or consonants.

Clearly the unmarked choice is to assign most of these colorings ([round], [high],[advanced tongue root]), to the nucleus, but occasionally (for morphological reasons, I believe) a language finds it convenient to assign them to the periphery (either onset (glides) as in Chinese, or offset as in Northwest Caucasian and, I suspect, Proto-Indo-European, and perhaps the Ndu languages of Papua-New Guinea).

I suspect that such a choice confronted human speakers from the start and that human phonological evolution has been toward distinguishing more and more features (articulatory linked with acoustic targets in a sort of "phonetic cognitive space"), and that some kind of primordial lumping pattern was not the case, or more accurately, that it was irrelevant."



March 15, 1997







the latest revision of this document can be found at

HTTP://WWW.GEOCITIES.COM/Athens/Forum/2803

Patrick C. Ryan * 9115 West 34th Street - Little Rock, AR 72204-4441 * (501)227-9947

PROTO-LANGUAGE@email.msn.com