A lexical entry with an affix is analyzed but doesn't seem to go through the lexical rule first

morphology
grammar_matrix

#1

I have a pseudogrammar with a lexical rule associated with an affix -SUBJ which is supposed to mark the verb for an abstract nonfinite form:

FRM-lex-rule-super := add-only-no-ccont-rule & infl-lex-rule &
[ DTR verb-lex ].

nonfin-lex-rule := FRM-lex-rule-super &
[ SYNSEM.LOCAL.CAT.HEAD.FORM subjunctive ].

nonfin-suffix :=
%suffix (* -SUBJ)
nonfin-lex-rule.

When I try to parse a pseudosentence that includes verb-SUBJ, I don’t get a parse and, to begin with, I do not see the verb going though the lexical rule, so in the parse chart, there is only one mention of verb-SUBJ, and it is the leaf level. No lexical rule.

The feature structure for tverb-SUBJ somehow has [ FORM finite ] on it, presumably because that’s what is on supertype of the verb. So I see why tverb-SUBJ won’t go through the rule, the values of FORM clash, but I don’t understand how tverb-SUBJ got to be analyzed without going thrugh the rule?

Something about the Morphology system that I don’t understand here.


#2

Something I am not seeing in this grammar produced by my current version of the customization system is this type of thing:

inflected :+ [ FRM-FLAG luk ].

infl-satisfied :+ [ FRM-FLAG na-or-+ ].

And then that would also show up in the lexical rule.

I wonder why? What should be generally be done on the Morphology page in order for the flags to be added by the customization system? That said, I am not sure at this point that that is the issue…


#3

tverb is showing up in the chart because the orthographic subrules can strip the -SUBJ affix and then the parser can locate the lex entry tverb. It’s showing up in the tverb-subj cell because the cells are labeled by the tokens, but that’s actually just tverb, and not tverb-subj.

The FLAG features come in from the morphotactics library when you make a position class obligatory, or if there is a requires/forbids constraint between it and some other position class.