[developers] Re: unary rules and packing
danf at csli.stanford.edu
Tue Mar 22 15:18:06 CET 2005
> for the current case, i now believe we are seeing a bug in the grammar:
> its `adj_intrans_compar_nale' is not supplying a COMPS values, and then
> the `hoptcomp' unary rule will cycle freely. dan, what do you think?
> after all, the generic lexical entries have hardly been used since the
> YY days, for all i know.
You're right that there is an error in this lexical type (and also in its
sister type 'adj_intrans_superl_nale') since the COMPS value should be
constrained to < > but isn't. I'll check in that fix with the new version
of the grammar that I'm about to check in.
More information about the developers