[Prev][Next][Index][Thread]

Re: RM stuff

%%TER: Perhaps since we are no longer in parsing with its objective of exactly
%%TER: describing syntactic structure, the set of acceptable representations
%%TER: might be extended to:
%%TER: 
%%TER:      np_postmodifiers([pp( Prep, Phrase, PTs ), PP ... ]).
%%TER: 
%%TER: It is easy to envision 'the printer by the desk in the office at the end 
%%TER: of the hall'.

**SYL: One remark concerning your last point Terry: I would be inclined
**SYL: to stay away from modifs that could throw in some confusion about
**SYL: the parse structures. At least in the beginning. Let's keep
**SYL: verification as simple as possible.

==KEN: I'm not sure I get the last point.  Terry, are you suggesting
==KEN: that we do away with subtleties like 'conj_pp', etc.?  I'd be
==KEN: loath (or is that 'loathed'? ;-) to do it.  As an example, I
==KEN: call 'You can print on the laser printer OR on the plotter' as
==KEN: my first witness.  The disjunction would be lost.

No, just that in the 'the printer by ...' case, the PPs are on analysis 
not conj-ed in the and/or manner you illustrate.  Fact is, reattachment 
here is pushing the envelope of DIPETT's grammar--the user wants to create 
a (correct) structure that DIPETT will not produce on its own, always 
preferring to conj its PPs.  What to do? I ask.

==KEN: Also, I would completely nest your example sentence to
==KEN: something like:
==KEN: 
==KEN:    'the ((((hall end) office) desk) printer)'
==KEN: 
==KEN: which doesn't illustrate a conj_pp (or your simplified list of
==KEN: PPs).  Multiple PPs attached to a noun might be illustrated by
==KEN: something like:
==KEN: 
==KEN:    'the printer at school in the lab on the desk'.
==KEN:

You raise an interesting semantic point and one which might be expected
of someone looking deeply into noun semantics.  

Is the LOC Case a unitary one existing on the verb level alone, or are 
nouns themselves capable of having LOC aspects?  For simplicity and no 
other reason I was tending towards the former, which is suited to having 
all location specifications brought out to the verb (=clause/4) level 
in the flat list of PPs you saw me suggest earlier; your rendering of 
the representation suggests that each additional PP refines the location 
specification of those that have gone before in a hierarchical manner.
More discriminating.

The difference has a direct impact on how reattachment is implemented
and how someone would use it.  Hierarchical, the user should nest each 
PP below its superordinate, and this is what the screen might show at 
the end of the process:

	on(printer)
    		by(desk)
			in(office)
				at(hall end)

Treated at a purely verb phenomenon, interaction might produce:

	on(printer)
		by(desk)
		in(office)
		at(hall end)

To complicate matters in a clarifying way, let's add the PP 'at 9 a.m.'
to the sentence.  In both instances this marker of another Case would
appear on the top level:

	at(9 a.m.)
	on(printer)
		...

Stepping back a bit, I wonder if it is useful to reintroduce the verb
at this point; the example came out of the sentence "I printed the file
on the printer by the desk in the office at the hall end".  Is the LOC
that these PPs together specify that of the printing action, or that 
of the printer?  

Resolving these issues and settling on a format for the output rep is
important to me.  Let's keep on it, gentlemen, and by the time I have
finished gathering all my bills together for the revenuers we may have
decided the targets for design.

-- 
Terry Copeck (terry@csi.uottawa.ca)
References: