BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Agreement Resolution from an Optimality Theory Perspective - Profe
 ssor Bill Badecker (Johns Hopkins University)
DTSTART:20061207T160000Z
DTEND:20061207T173000Z
UID:TALK6033@talks.cam.ac.uk
CONTACT:Susan Rolfe
DESCRIPTION:There are two patterns of subject-predicate agreement that one
  observes when a subject noun phrase is comprised of conjoined noun phrase
 s:\n Partial Agreement\, where the predicate agrees with one of the conjun
 cts\, typically the conjunct that is linearly closest to the predicate\; a
 nd Agreement Resolution\, where agreement is based on inflectional values 
 that for person and number features reflect the derived semantic propertie
 s of the conjoined phrase and for gender (or noun class) features reflect 
 grammatically based principles of feature selection.  This talk focuses on
  the process of agreement resolution.\n\nPrevious accounts of resolution f
 or gender features exemplify one of two sorts of approaches:  Rule-based s
 tipulations to handle different combinations of conjunct genders (e.g.\, C
 orbett\, 1991) and Generalized\noperations over sets of features (e.g.\, f
 eature set union: Dalrymple & Kaplan\, 2000\; and feature set intersection
 : Wechsler & Zlatic\, 2003).\n\nThe former approach suffers largely from a
 n over abundance of expressive power (allowing for a host of unattested an
 d implausible resolution patterns)\, whereas the latter suffers from the o
 pposite flaw\n(an inability to capture a range of interactions one frequen
 tly observes\, for example\, between gender and number).  An approach usin
 g Optimality Theory constraint interaction\, in contrast\, can provide an\
 naccount of cross-linguistic variation in resolution patterns that is both
  explanatory and descriptively adequate. Resolution patterns derive cross-
 linguistically from the interaction of violable markedness constraints (e.
 g.\, *Fem >> *Masc >> *Neut) and faithfulness constraints (e.g.\, constrai
 nts that impose the condition that the phrase and all its\nconjuncts agree
  in gender\, or that the conjoined phrase agree on all dimensions with one
  of its conjuncts). The ranking of these constraints\nwill be shown to acc
 ount for the variation one observes in resolution patterns. The limited ci
 rcumstances under which resolution and partial\nagreement interact will al
 so be discussed.
LOCATION:GR-06/07\, English Faculty Building
END:VEVENT
END:VCALENDAR