upcase-word, &c. for non-latin alphabets

[Will Parsons]

I'd like to be able use the upcase/downcase/capitalize-word functions
(or equivalents) for non-Latin alphabets (utf-8 encoding).  Can anyone
give me any hints on how to go about it?  I assume that one step is to
define the correlation between uppercase and lowercase letters
somehow, but beyond that I don't have any idea.  (Since the standard
functions are built-in rather than elisp, I can't easily see how they
do it.)

-- 
Will