---------------------------------------------------------
Notes

  * Demonstrative Adjectives - I have reassessed my interpretation of
    the -la and -ci suffixes, now considering them demonstrative
    adjectives.  Still need to add those particles to my grammar.
    TODO.  They are in my test suite, however:

On what basis?  If they attach to the right most word in the noun
phrase, then yes, they look like adjectives (clitics, not affixes).  If
they always go on the noun, then suffixes seems right...

Shared constraints here can be moved up to a supertype, e.g.,
avoir-aux-lex.

1SG_aux_avoir-aux-lex := inanimate_obj_pronoun-rule-dtr & negation_prefix-rule-dtr & arg-comp-aux-no-pred &
  [ SYNSEM.LOCAL.CAT.VAL [ COMPS.FIRST.LOCAL [ CONT.HOOK.INDEX.E.TENSE past,
                                             CAT.HEAD.FORM past_participle ] ] &
                         [ SUBJ.FIRST.LOCAL.CONT.HOOK.INDEX.PNG [ NUM singular,
                                                                  PER 1st ] ] ].

2SG_aux_avoir-aux-lex := inanimate_obj_pronoun-rule-dtr & negation_prefix-rule-dtr & arg-comp-aux-no-pred &
  [ SYNSEM.LOCAL.CAT.VAL [ COMPS.FIRST.LOCAL [ CAT.HEAD.FORM past_participle,
                                             CONT.HOOK.INDEX.E.TENSE past ] ] &
                         [ SUBJ.FIRST.LOCAL.CONT.HOOK.INDEX.PNG [ NUM singular,
                                                                  PER 2nd ] ] ].

  Unfortunately, I ran into trouble with the multiple prefixes, such
  as a dropped argument marker and the negative prefix.  I couldn't
  find how to specify which order these attach in.

Give neg-lex-rule a dtr value that is supertype to both the verbs and
the pronoun prefix lex rules.
 
  ** My sentences formed with est-ce-que have CP as their root node.
     Not sure if this is normal.

Yes, this is expected.

  The fact that se demander is a reflexive verb has necessitated that
  I add a new verb slot for reflexive verbs that not only adds the
  regular suffix, but also a prefix.  Still pending.  TODO.  As a work
  around, I have used the non-reflexive form of demander with si in my
  test suite.  TODO - FIX.

This is a good workaround, since it supports incremental development :)

  I'm also checking with my sources whether embedded declarative
  clauses require the subjunctive form of the verb.  Their initial
  thoughts were that the use of the subjunctive changes the sentence
  from a statement of fact to a statement of opinion.

My understanding is that some verbs require subjunctive clausal complements,
but that this is lexically specific, and not a general property of embedded
clauses.

Your semantics for the embedded clauses aren't quite right.  In particular,
the ARG1 and ARG2 of the matrix verbs aren't linked to the arguments.
The reason is that you're not connecting SUBJ and COMPS to the ARG-ST
in this type:

embedded-clause-verb-lex := verb-lex & clausal-second-arg-trans-lex-item & norm-sem-lex-item &  
  [ SYNSEM.LOCAL [ CAT.VAL.COMPS < [ LOCAL [ CAT.HEAD comp,
                                             CONT.HOOK.INDEX.SF prop-or-ques ] ] > ]]. 

#317 - Without the copula - NP then becomes an adjective 
Source: author
Vetted: f
Judgment: u
Phenom: cop
*Le garçon est enfant
le           garçon       est    enfant
the.3SG-masc boy.3SG-masc is.3SG child.3SG-masc 
'the boy is a child' 

This example still has the copula (est), it's missing the determiner
for the predicative NP.

Your predicative adjectives aren't quite working:

--- The semantics is wrong (the subject should be the ARG1 of the adjective's
KEYREL)
--- The agreement isn't functioning ... you're parsing "La fille est petit", as well.
--- There is no tense information, though "etre" should be contributing some.

The first two problems can be addressed in one go, and this is partially
an issue with the grammar matrix.  To fix this, change the definition
of intersective-mod-lex in matrix.tdl to:

intersective-mod-lex := norm-hook-lex-item & no-hcons-lex-item &
  [ SYNSEM [ LOCAL.CAT.HEAD.MOD < [ LOCAL intersective-mod ] >] ].

And the definition of adjective-lex in french.tdl to:

adjective-lex := basic-adjective-lex & intersective-mod-lex &
   [ INFLECTED -,  
     SYNSEM [ LKEYS.KEYREL.ARG1 #xarg,
              LOCAL [ CONT.HOOK.XARG #xarg,
                      CAT [ HEAD.MOD < [ LOCAL[ CONT.HOOK.INDEX #xarg,
                                                CAT [ HEAD noun, 
                                                      VAL.SPR cons ]]] >, 
                            VAL [ SPR < >, 
                                  SUBJ < >, 
                                  COMPS < >, 
                                  SPEC < > ] ] ] ] ]. 


  For some reason, the Parse chart shows that adposition somehow
  undergoes the inv-lr, implying that it is a verb, and it forms with
  its NP complement as a VP.  The copular version of être wants an
  adpositional phrase as its complement, and therefore, it won't
  unify.  I am rather stumped as to why my PPs are being thought of as
  VPs.  Nothing in my declaration of the adposition is verby, and
  nothing in the ivr-lr heirarchy is jumping out at me.

As noted today, the reason inv-lr was firing is that the HEAD value
on its DTR is not constrained.  However, that doesn't explain why you
weren't getting any parses.  The problem there was a path error.  This:

locational-adpositional-lex := basic-int-mod-adposition-lex & 
  [ SYNSEM.LOCAL.CAT [ VAL.COMPS < [ HEAD noun ] > ] ].

should be this:

locational-adpositional-lex := basic-int-mod-adposition-lex & 
  [ SYNSEM.LOCAL.CAT [ VAL.COMPS < [ LOCAL.CAT.HEAD noun ] > ] ].

That will get you parsing, but then you get too many parses, and will
also need to debug the semantics.


  Sentences with an adverb and no object look to be going through the
   basic-head-opt-comp as the verb and the vervb + adverb combo (nous
   mangerons lentement is an example of this)

If you make the mother of head-opt-comp [LIGHT -] it should help...

  le garçon et le enfant mange le gâteau - I don't know how to force a
  conjoined subject to declare itself plural

Say INDEX.PNG.NUM pl on the mother of the np1-top-coord-rule
and the n1-top-coord-rule.

  Je ai mangé lentement - I guess the [ POSTHEAD + ] setting on the
  adverbs is considering the participle as part of the HEAD.

If you look at the tree, you should see lentement combining with mange.
You can constrain the adverb to require [FORM finite] MOD value.


