Re: Arabic and BACK TO Self-segregating morphology
| From: | Wesley Parish <wes.parish@...> |
|---|
| Date: | Tuesday, December 20, 2005, 2:19 |
|---|
Quoting Jim Henry <jimhenry1973@...>:
> On 12/19/05, Gary Shannon <fiziwig@...> wrote:
> > There would be no inflections, and any change of vowel
> > would mean a change in the meaning, not in case,
> > tense, plurality, etc. Those would be marked by
> > particles.
<snip>
> You could avoid that by setting aside certain
> consonants for use only in suffixes,
> others only in prefixes, some only in roots;
> or by selecting certain vowels to occur in
> prefixes, others in suffixes, and others as
> the first or last vowel of a tri-vowel pattern.
> Either way, no prefix or suffix could ever
> look like the first or last syllable of any actual root with
> any possible vowel pattern. E.g., if "pe-"
> is a prefix or suffix, then you could:
> 1. make sure no roots begin or end with p-, or
> 2. make sure no vowel patterns begin or end with -e-.
>
> Either would ensure that a phrase like "kulaji pe-nalija"
> could only be parsed as KLJ + u-a-i ... pe- + NLJ + --a-i-a,
> and not ku- LJP + a-i-e ... NLJ + a-i-a = ku-lajipe nalija, or
> or KLJ + u-a-i ... PNL + -e-a-i + -ja = kulaji penali-ja.
> In this example you would also want to
> ensure either that K, J, and N never occur
> in prefixes or suffixes, or that u, i, and a never
> occur in prefixes or suffixes.
Which looks quite like the hithpoel, qatal, etc, conjugations currently in
Semitic languages. The major difference being that the Semitic conjugations
also include person and aspect - qatal, yiqtal, etc.
Wesley Parish
>
> Also, unless you want phonemic gemination, you need
> to exclude /n/ from the set of consonants that
> can form roots -- else the /n/ in a root
> and the /n/ in a derivational vowel pattern
> could end up adjacent.
>
> --
> Jim Henry
>
http://www.pobox.com/~jimhenry
>