[lkb] Showing CTXT (context)

[Ann Copestake]

bond at cslab.kecl.ntt.co.jp said:
> G'day,
> related somewhat to the discussion of information structure on the [matrix]
> list, JongBok and I were wondering whether there was any way easy way of
> displaying the CTXT diff list in the lkb.  The Japanese and Korean Grammars
> use it for honorification/empathy, and I would also expect to use it for
> information structure.  Unfortunately, I only know how to see it within the
> feature structures.

> Is it possible to display it as we do the MRSs?  

There's a notion of what an MRS is independent of the FS encoding of it.  We 
could build MRSs from a non-FS grammar (in fact we build RMRSs from RASP). The 
code that manipulates MRSs actually extracts a new data structure from the FS 
and then deals with that.  There's no problem in principle in doing the same 
thing with CTXT but someone really has to define what's in CTXT so that 
whoever writes the extraction code knows what they are manipulating.  This 
could be quite simple - I actually had some code that did this for a student a 
few years ago but then removed it because nobody else was using it.  Obviously 
if you would like to do anything with CTXT other than simply admire it, there 
needs to be a definition of what the operations on this structure are.

Ann