-------------------------------------------------------
Comments

 In my first modules page run I indicated that the Hebrew conjunction
 "w" is an affix. Since I also indicated that phrases can be
 coordinated in Hebrew I got a script error since an affix is
 incompatible with phrase coordination. I changed the designation of
 "w" from affix to word to avoid the error. The "w" is probably a
 clitic rather than an affix. Still, designating it as a word means
 that it is treated as an independent unit. At this point, all the
 sentences in the test suite that contain the coordination phenomenon
 fail to parse because in them the "w" is attached to the first word of
 the last coordinand. So I will need to find a solution for this
 somehow.

If it can attach to the first word no matter what that word is,
sounds like a clitic, and should therefore be segmented by a preprocessor.

 I think that the biggest bang for the buck will come from figuring out
 how to handle the "clitics". There are tons of words now which are not
 recognized because there is a clitic attached to them as a prefix. One
 option is to separate them from the words they prefix in the test
 suite and assume the preprocessor does that, which is not a far
 fetched assumption. Handling it within the lexicon may also be an
 option. 

Handling it within the lexicon sounds like it would be tedious,
since a) the clitics can probably attach to just about anything
and b) the syntactic effects they would have to have would depend
greatly on what they're attaching to.