Re: derivational morphology

In a message dated 9/30/01 11:50:48 PM, tb0pwd1@CORN.CSO.NIU.EDU writes:

<< Anyone want to talk about derivational morphology?  I'm trying to figure
out how other languages, nat and con, do it so I can find a way that's
aesthetically pleasing to me.  I love the sound of Hawai'ian; apparently
reduplication is part of their derivational morphology.  Anyone know how
that works?  Or have interesting methods to elucidate, like Hebrew? >>

    As I mentioned probably last week, it's a good idea, if you're creating a
language that's far removed from the date of its inception, to create
non-transparent derivational morphology.  For example, Mbasa, which I'm
working on right now.

    In Mbasa, you mark the dative with an initial [m], which assimilates to
the next segment ([n] before alveolars; [N] before velars, etc.).  However,
there's a related derivational marker with the same properties ([m], as
above) which derives words in a non-predictable manner.  Here are some
examples:

1.) pantsu=to swim; mbantsu=to splash ([p]>[b]/[m]_)  In this case the idea
is related to the dative
    market, in that splashing is a sort of indirect action that comes from
swimming.  It's as if the
    water is the direct object, and the splash the indirect affect.

2.) heSa=night; mheSa=to darken (/mh/=voiceless [m])  Here it's kind of
describing the action that
    night does when it descends upon the land.  Not so much an indirect
affect.

3.) hesa=to see; mhesa=to understand; Now, this isn't an affect at all, but
more of a metaphorical
    extension, via the metaphor "to see is to understand"; but, you could
argue that understanding is
    sort of a by-product of seeing.

4.) kesa=to want; Ngesa=to need (same rule as in 1 with the [k])  Here it's
not a by-product at all, but
    almost a strengthening of the original verb, since needing is a stronger
form of wanting.

5.) Gena=to take; NGena=to enslave ([G] voiced velar fric.)  Here it's also a
strengthening, but
    further than that, a very negative meaning is derived, which couldn't be
expected or predicted at
    all.

6.) kodZi=to cook; NgodZi=food; Here it didn't even derive a noun but a verb,
though the meaning
    is still close to some sort of a object status, even if it's probably a
direct rather than indirect
    relationship.

7.) nombi=to sleep; nnombi=to dream ([nn]=geminate [n])  And then, again, you
find a far more
    predictable relationship.


    This is the kind of things that happen to a language over time with
derivational morphology.  The [m] as a derivational marker probably was very
predictable in the beginning, but it's usage probably got reinterpreted and
extended so much over the time that a native speaker would be hard pressed to
figure out what the heck it means or what it would do to a new word.  It's my
impression that this is similar to reduplication in Hawaiian, in that it
can't be readily predicted.  Also, if you search the archives, I posted all
my derivational morphology on Gweydr, which I also intended to be
non-transparent.

-David