more thoughts on MLU

[Sheri Wells Jensen]

Folks,

This discussion touches on something I've been working on for a
while--trying to find a metric by which you can functionally compare
complexity of different languages or of sub-systems within different
languages.  I've been reading lots of good ideas here.  The MLU is a great
starting point, but here's the can of worms it has opened for me:
there's the trouble of equivalents.   Are all morphemes within and between
languages created equal or are some more equal than others?  Is retrieving
a content morpheme enough like adding an affix or applying a regular rule
that they all  really ought to be added together in an MLU count and each
counted as 'one'?  Is  marking agreement  as difficult as marking plurality
when the latter is arguably easier to understand?    Looking across
languages, is it less complicated to mark number on adjectives in a
language that already requires gender to be marked on adjectives?  Worse,
can you equate binary gender marking in Spanish to noun class marking in
Kiswahili?  Maybe you wouldn't have to worry about this if you were
reflecting patterns in language, but if you  want to deal with functional
cross-linguistic issues, it's unavoidable.

Sheri

* * * * * * * * * * * * * * * * * * * * * * *
Dr. Sheri Wells-Jensen
Bowling Green State University
MA TESL  Program
http://www.bgsu.edu/departments/tesl/

Office: 423 East Hall
(419) 372-8935
Homepage: http://personal.bgsu.edu/~swellsj/
* * * * * * * * * * * * * * * * * * * * * * *