Lab 3 grade
-----------------------------------------------------------------

Comments:


 In this case, the prefix "a" represents the 1SG future morpheme. Note
 that I am not marking the "a" for FUT again, even though the morpheme
 for 1SG in other tenses is different. I hope that is OK.

I think so.  As long as the future info is there in the gloss somewhere.
Someone who was particularly working on that morphological dependency
might put more info into the gloss.

 Hebrew has two genders: masculine and feminine. There are some clues
 in the orthography to indicate which gender the noun is likely to be,
 but there are enough exceptions to make the breakdown undesirable in
 the lexicon. Therefore, I am not following the "ragazz-a & ragazz-o"
 type notation and am treating each noun as having an inherent gender
 (a la gateau) . Therefore, the gender is presented as internal to the
 word and separated by a semicolon (I believe that the semicolon
 represents the idea that there is some inflection but that it is
 irrelevant to the example).

That's fine.  The thing about Italian is that the inflectional endings
for singular and plural are different for m & f.

 Initially, I thought I would represent the PL.M & PL.F information as
 hypthenated inflection. But that is not a good idea either. While the
 suffix "wt" is generally the suffix used for PL.F and the suffix "im"
 is the suffix used for PL.M, this is strictly true only in the case of
 adjectives. When it comes to nouns there are enough exceptions to make
 this generalization unprofitable. Therefore, I am treating the gender
 information as internal to the word even in the case of plural
 nouns. Each word in the lexicon will have to have information as to
 its gender and as to the type of plural suffix it combines with.

Sounds good.

Under coordination, consider some ungrammatical examples with
the coordinator in the wrong place, or lacking a coordinator.

 Unlike in English it is possible in Hebrew to relativize an NP that is
 a corrdinated structure, and say "The man, that he and his mother are
 doctors...". See below.

This, together with the GAP/SLASH analysis of long-distance dependencies
suggests that Hebrew relative clauses with resumptive pronouns
do not involve an LDD:  One way to handle resumptive pronouns is
to have them have non-empty SLASH values even though they are
overt --- then the relative clauses look similar in the gap cases
v. the resumptive pronoun cases.  This also allows you to require a pronoun
rather than a full NP in the 'gap' position.  But, that wouldn't seem to work
here.  What happens in subject relatives with verbal predicates
and coordinated subjects.  For example, "The man who wrote many books
and his mother wrote even more"?  (I'm interested because the subject
relatives ordinarily don't take resumptive pronouns, but the coordination
might force one.)

The relative marker "e" looks like a clitic --- since it 
can lean on just about anything.  So, I'd recommend assuming
that a preprocessor will segment it off.