Sag&Wasow coordination question

[Stefan Mueller]

Hi,

NP coordination is a different story all together.

(1) a. [Kim and Sandy] sleep / *sleeps.
    b. Kim sleeps.

The coordination of two singular NPs usually gives a plural NP (for
exceptions see Mueller99a, p. 132 or Mueller99b, p. 76). So the CONT
value of the NP in (1a) is necessarily different from the CONT values of
the coordinated NPs. Since gender is assumed to be part of the INDEX
information which lives under CONT the coordination of two NPs which
different gender should not be a problem. (However, the determination of
the gender of the resulting NP is not trivial. See (Dalrymple and
Kaplan, To Appear) on this issue).

There is a way to do coordination in HPSG which was suggested by Pollard
and Sag (1994, 202). This is not the full solution but covers a lot of
cases:

The information that is shared in a coordination is only the syntactic
information, i.e. CAT and NONLOC. The semantic features may be
different.

(2) Mary is sleeping and the girls are playing in the yard.

Sentences like (2) are no problem for analyses like the one above, if
agreement is treated lexically. (This is an observation of Ivan Sag, see
also (Kathol, 1999))

The reason for this is that the agreement information of `is' is only
present in the valance features of `is' or `is sleeping'. Once you have
the full sentence you do not have the agreement properties anymore,
since the SUBJ and COMPS lists of sentences are empty. (If you project
ARG-S somewhere in CAT then of course you get problems).

The alternative is to have a head feature AGR and then some way of
neutralization of conflicting information. But this solution does not
work in general, since it may also neutralize conflicting valance
information in the case where two unsaturated heads are coordinated.
(This was pointed out by someone else before.)

Without these neutralization techniques coordination data really is
important for determining feature geometry. All features that would
clash in coordinations have to be either hidden in valance features or
be represented inside of CONT.

Since CONT usually travels on head pathes as well this is a possability
to avoid the conflict of head features.


Greetings

	Stefan

P.S. The subsumption based approach by Ingria doesn't work at all. For a
discussion see Mueller99b and the discussion at the HPSG-Mailing list we
had in 1997. Unfortunatly the e-mails got messed up in the archive of
the Mailing List but if someone is interested in this I can repost them.

Greetings

	Stefan

@article{DK2000,
  author    = {Mary Dalrymple and Ron Kaplan},
  title     = {Feature Indeterminancy and Feature Resolution in
Description"=Based Syntax},
  journal   = {Lanuage},
  year      = "To Appear"
}

@InProceedings{Ingria90,
  author        = "Robert J. P. Ingria",
  booktitle     = "Proceedings of the Twenty-Eight Annual Meeting of the
ACL",
  address       = "Pittsburgh, Pennsylvania",
  organization  = ACL,
  pages         = "194--204",
  title         = "The Limits of Unification",
  year          = "1990"
}

@InCollection{Kathol99,
  author    = "Andreas Kathol",
  title	    = "Agreement and the Syntax-Morphology Interface in {HPSG}",
  booktitle = "Readings in {HPSG}",
  editor    = "Robert Levine and Georgia Green",
  publisher = "Cambridge University Press",
  year	    = "1999"
}

@book{Mueller99a,
  author    = {Stefan M{\"u}ller},
  title     = {{Deutsche Syntax deklarativ.
	        Head-Driven Phrase Structure Grammar f\"ur das Deutsche}},
  publisher = {Max Niemeyer Verlag},
  address   = {T\"ubingen},
  series    = {Linguistische Arbeiten},
  number    = 394,
  note      = {http://www.dfki.de/~stefan/Pub/e_hpsg.html},
  year      = "1999"
}

@article{Mueller99b,
  author      = {Stefan M{\"u}ller},
  title       = {An {HPSG}-Analysis for Free Relative Clauses in
{German}},
  journal     = {Grammars},
  volume      = 2,
  number      = 1,
  pages       = {53--105},
  note        = {http://www.dfki.de/~stefan/Pub/e_freeRel.html},
  year        = 1999
}

@book{PS94,
  AUTHOR    = {Carl J. Pollard and Ivan A. Sag},
  title     = {Head-Driven Phrase Structure Grammar},
  Publisher = {University of Chicago Press},
  ADDRESS   = {Chicago, London},
  SERIES    = {Studies in Contemporary Linguistics},
  Year      = 1994
}



--
FB Sprachtechnologie
DFKI GmbH               Tel.: (+49 - 681) 302 - 5295
Stuhlsatzenhausweg 3    Fax:  (+49 - 681) 302 - 5338
D-66123 Saarbruecken    http://www.dfki.de/~stefan/

http://www.dfki.de/~stefan/Babel/Interaktiv/Babajava/