Theiling Online    Sitemap    Conlang Mailing List HQ   

Plastic Surgery on a Conlang

From:Christian Thalmann <cinga@...>
Date:Friday, April 19, 2002, 20:33
Initially, my primary conlang set out to be aesthetic, efficient and
original.  It seems I've laid too much emphasis on the latter two goals,
as the aesthetics of the language accumulated flaws and defects with
time.

I have now decided to correct the most prominent and unnecessary sources
of visual and acoustic disharmony.  I'd appreciate any *FEEDBACK* on
whether you consider the changes to be more aesthetic than the original,
and on alternate suggestions.

If you shouldn't, for whatever reasons unfathomable, want to read
through the whole post, at least have a look at the Bottom Line.  /=P


Problems:
*********

(1) Stop+Nasal clusters, such as in |getne| "fearing".

(2) Stop+Stop+Fricative clusters, such as in |noggze|.

(3) Ugly typographical juxtapositions due to non-standard letter
    values, as in |kcare| or |kwaq|.


Reasons:
********

(1) In the native script, fricatives correspondent stops have similar
    representations, stops adding a descender to the fricative shape.
    I thought it would be cool and original if I extrapolated that
    scheme to nasals, giving them stop-like twins.

    I decided against nasal+stop combination because they would cross
    syllable boundaries, which would compromise the nice syllabic
    analysability of the script.

(2) My initial syllable structure concept used to be (T)(S)V(S), where
    T is a stop and S a non-stop consonant.  The final S could become
    a stop only if the same stop followed at the start of the next
    syllable, thus forming a geminate.

    Therefore, a word like |kotre| would be syllabized as |ko.tre|,
    giving us an open stressed syllable, pronounced /"ko:tr@/.  To
    allow the closed-syllable configuration /"kOtr@/ as well, I had
    to geminate the stop: |kottre|.

(3) Each letter of my Latin transcription corresponds to one letter
    in the native alphabet.  Unfortunately, those letters don't always
    behave like Latin letters do, so I had to assign some Latin letters
    to values that didn't suit them.

    From a grammatical point of view, this notation is very useful,
    since it allows sound-changing inflections like |cajze| /"hajZ=/
    --> |cejze| /"sejZ=/ to use the same consonant letter |c| for
    different allophones of the phoneme.


Solutions:
**********

(1) I abolished all stop-nasal combinations.  The active participles
    of consonantic verbs, which used to produce such combinations,
    now infix the nasal: |kot-| --> |konte| rather than |kotne|.

(2) I changed the syllabification rules to draw the syllable boundary
    between two consonants, regardless of stops.  This means all words
    like |kotre| are now pronounced /"kOtr@/, thus rendering the
    spelling |kottre| obsolete.  As a side effect, the pronunciation
    /"ko:tr@/ is now impossible.  Too bad, I kinda liked it.

(3) I am going to keep my transliteration scheme for linguistic
    purposes, but as per Aidan's suggestion (thanx, Aidan) I'm also
    developing an alternate spelling system for use with the "general
    public".

    The public transliteration will reflect pronunciation rather than
    phonemic structure.  For reasons of personal aesthetic preference,
    glides will be written as vowels except intervocally, and /S/, /Z/
    will be represented as s' and z', since I find sh, zh or j, x to
    be less pleasant.  I'll also use c for /k/.


Bottom line:
************

So, the formerly ugly words |getne|, |kcare|, |nokkce| and |kwaq| will
become =gente=, =chare=, =nocse= and =cuang=, or even =noxe= and
=quang=.  It's that nice?  =)


I'll apply these new spelling rules to my Babel text and see how it
comes out.

Thanx in advance for your -- as always -- prolific feedback...
::sigh::


-- Christian Thalmann


PS: The name of the conlang was deliberately never mentioned in this
    post since it seemed to act as a reply repellant in my previous
    posts.  =P

Replies

Christophe Grandsire <christophe.grandsire@...>
Raymond Brown <ray.brown@...>