Re: [tied] All of creation in Six and Seven

From: Miguel Carrasquer
Message: 27255
Date: 2003-11-16

On Sun, 16 Nov 2003 02:29:38 +0100 (MET), Jens Elmegaard Rasmussen
<jer@...> wrote:

>This is most illuminating, but also intriguing in several points. Permit
>me a few follow-up questions and comments:
>
>1. Is there a distinctive accent in Arabic? I had the impression it was
>automatic and of no historical value.

The rules are automatic. For Classical Arabic, the accent is on
(1) the penultinate syllable if it is closed/long (e.g. kitá:bun)
(2) the antepenultimate when it is closed/long and the penultimate short,
or when the word has three short syllables (e.g. ká:tibun, kátaba)
(3) the closed/long syllable before the antepenultimate when the
penultimate and antepenultimate are both short; on the first syllable if
there is no such long syllable (e.g. muká:tabatun, kátabatuhuma:)
(from: A.S. Tritton, Arabic).

What the historical value of the Arabic accent is, I'm not sure.
Abstraction made of prefixes such as mu-, ma- etc, within the general
Semitic framework (triliteral CV-CV-CV), there are basically two choices:
stress on the first or on the second syllable. Both seem to occur
(including CCVC- and CVCC- with complete reduction of the unstressed
syllable [the pattern CCVC tends to get a prefix ?a-]).

Verbs have various prefixes and/or suffixes, some of them with a long
vowel, so there's a bit more room for variation there. I wouldn't want to
get into the details, because I'm not sure of them.

>2. What is the evidence to show that the old accent was not influenced by
>the mimation/nunation?

The modern languages have mostly lost not only -n/-m, but also the case
endings themselves. The final vowel (except in cases of contraction) was
never stressed, whether followed by nunation/mimation or not. In languages
like Akkadian, where in the construct state the final vowel was entirely
missing (C.S. s^ar vs. Nom. s^arru(m) / Acc. s^arra(m) "king"), that
presumably also had no effect on the accentuation of the word.

>3. I have a Geez grammar spelling 'seven' as /sa-be-(a-tu/, /sa-be-(u/.
>(I use /.../ here only to set off what belongs to the word from what does
>not). The part on the writing system notes the vowels of the signs here
>given as /sa/, /be/, /(a/ with a diacritic of shortness, while /tu/ and
>/(u/ are given without, but the system contains no specifically short u. I
>guess you are right that one could say that the -u is thus not
>phonemically long. However, old *short* /i/ and /u/ have developed into
>the Eth. central vowel /e/ (with diacritic of shortness). The writing is
>not with -Cu-u.

See my afterthought. Brockelmann et al. were probably using a historical
transcription with /u:/ for /u/, and, presumably, /i/ and /u/ for /&/.
I'm still puzzled about the stress.

>4. Isn't lack of reflex of the nunation in the Geez noun just a matter of
>grammatical reduction - a reduction not occurring in the numerals? If
>there is a real phonetic rule against accenting the nunation, could it not
>have been set aside in cases where the function of that element was a
>different one?
>5. Supposing - only for the sake of the argument - the added particle does
>cause final accent in the Geez numerals, as Brockelmann writes (on a basis
>unknown to me), could it not have done the same in the kind of Semitic
>from where pre-PIE took the loanword 'seven'?

I don't know why the Ge'ez numerals end in -u (alternative forms in -ä
[accusatives?] are also given in the web page I mentioned by Jouni Maho.
For ordinary nouns, the nominative in -u should have given -& in Ge'ez, and
was subsequently deleted (just like the genitive in -i > -& > -0), leaving
only an accusative in -ä.

I'll ask on sci.lang, where I've seen Jouni Maho hang out.


=======================
Miguel Carrasquer Vidal
mcv@...