Re: Self-Segregating Morphologies

At 8:55 pm -0400 12/5/02, Shreyas Sampat wrote:
>: Are there any systems out there greatly different from the four that
>: I have mentioned?
>
>I seem to recall a discussion on this;
Yes, it has been discussed more than before - and I shall be reading any
more mails in this thread with great interest as one BrSc's aims (both
versions) is to have self-segregating morphemes

>one person suggested a vowel-harmony
>system requiring that every syllable but the last of the segregated
>morpheme(s) are of one vowel set, and the last is of the other set, such
>that whenever a disharmonic vowel appears it signals the end of a segregated
>group.
Yes, I recall that.  See below:
--------------------------------------------------------

At 9:09 pm -0400 12/5/02, John Cowan wrote:
[snip]
>
>The Xuxuxi approach uses stress and vowel harmony.  Polysyllabic words
>are stressed on the first syllable and then go on up to and including
>the first syllable that breaks vowel harmony; any following unstressed
>syllables are monosyllabic words (particles).  The vowel harmony
>table is thus:
>
>1st syll        next sylls      final syll      harmony type
>a               a,e,o           i,u             low
>e,i             a,e,i           o,u             front
>o,u             a,o,u           e,i             back
Yep - remind me who's the author of XuXuXi?

BrScA has vowel harmony, but no vowel disharmony.  The rule is basically
very simple:
- all lexical morphemes have the written shape CVC
- all functional morphemes have the written shape C

The vowel in CVC is stress (and probably longish); all unstressed vowels
are unwritten and determined by vowel harmony. Also all spoken syllables
have the form CV, therefore you can work where the unstressed vowels will
appear in the (disyllabic) lexical morphemes and the functional morphemes;
the unstressed vowels attached to them are determined by the lexical
morpheme to which they are attached as suffixes/enclitics.

>> Is it important to self-segregate the morpheme level, or is word-
>> level self-segregation sufficient?
>
>I think the importance of two-level self-segregation appears only
It depends, I think, on the language and the desires and aims of the
language modeller.  Of course, if we have only word level self-segregation,
then we do not have proper morpheme self-segregation.  Word segregation is
normally achieved in print with white space; a consistent & regular system
of word stress should help in speech.

In a 'naturalistic' artlang, such considerations are probably not very
relevant; in an a_posteriori auxlang, one will obviously encounter
problems, but one should IMO try to avoid "sukero" type ambiguity as much
as possible.

When I was learning Speedwords long years ago in the 1950s, I found the
lack of clear morpheme boundaries in such an unusual language to be very
annoying.  Strangely, in fact, Dutton went to some length in his
complicated rules of pronunciation to have morpheme self-segregate.  But
this is certainly not the case with the written language; but it is most
important to spot the morpheme boundaries, not only to know what compounds
& derivates might mean, but also how to pronounce the words properly.  The
only way this can be done with certainty, is to learn _all_ the morphemes
of the language and be familiar with them!

it was because of this experience, that I wish have this feature in any
final version of BrSc.  as BrScB is under revision & development, I shall
be reading subsequent mails with interest  :)

Ray.

=======================================================
The median nature of language is an epistemological
commonplace.  So is the fact that every general
statement worth making about language invites a
counter-statement or antithesis.
                                 GEORGE STEINER.
=======================================================

Reply