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

The PRED value for pronouns should all be _pronoun_n_rel. (-1 total
between both).
The PNG information is expressed by features.

The ARG0 of the _can_v_rel should be the INDEX of the whole MRS.

  We have tried the complete sentences, selective edges from LKB
  charts, and parts of the sentence, all with no luck. Each attempt to
  generate requires killing emacs and reloading the lkb and our
  grammar. We are unsure of how to proceed, given the requirement that
  our grammar generate for week 9.

If you don't even get the maximum-number-of-edges error, then
the best strategy is to find the simplest sentence you can, comment out
all lexical rules not used in that sentence, and try generating.    If you
still can't generate, then those lexical rules contain a bug.  If you can generate,
put the lexical rules back in one by one until you can't again.  Then debug
the last rule you re-added.

Regarding your pipeline-refining... it's definitely good that you've brought
ambiguity down, and you can leave things as they are.  I wonder, though,
if you now have repeated constraints (e.g., negation lexical rules that are
substantially similar across all the types) that can be pulled out into supertypes.
These supertypes would be cross-classified with the types that control
the pipeline to give your actual lexical rules.

