Agreement: An investigation into the distribution of information
Project Overview
Project
Agreement: An investigation into the distribution of information
Project members:
Prof Greville G. Corbett
Dr Dunstan Brown
Dr Andrew Hippisley
Dr Julia Barron
Dr Carole Tiberius
Consultants:
Prof Nick Evans (University of Melbourne)
Prof Marianne Mithun (UC Santa Barbara)
Period of award
September 1999 - October 2002
Funder:
Economic and Social Research Council (ESRC) - R000238228
Agreement is a puzzling phenomenon, found widely in languages of different types. Basically, agreement is the 'displaced' expression of information. In the sentence: the system works intermittently, information about the number of systems is expressed redundantly on the verb (works versus work) which, were it not so familiar, would surprise us. The verb is marking the number of systems (not the number of working events). English agreement is rather straightforward; elsewhere agreement is much more complex. Twelve languages, taken from different families so as to maximise diversity, were investigated as part of this project; data was collected according to a consistent format across the languages, starting from the four main areas of variation (controllers, targets, domains and categories). This led to an initial typology, followed by a more detailed investigation of three languages, found at extreme points of the typological space (Russian, Mayali, and Yup'ik). Based on this complexity a general typology of agreement systems was developed. This was primarily a theoretical aim, but it has potential long-term applications in that agreement has implications for the design of parsers in natural language processing systems. The outputs of the project include a typological database and an annotated bibliography.