Agreement in Discourse

This project explores the function of agreement in natural texts. The concept of agreement has played a key role for various domains of linguistic theory (morphology, syntax, semantics), and there are a number of different approaches to modeling it. However, there is still no generally accepted explanation of its function: Why should languages so often develop agreement in their grammars? In his seminal work on agreement, Corbett (2006: 274-275; see also Lehmann 1988, Levin 2001: 21-27, Kibrik 2011) proposes four possible functions of agreement, among which the most important are:

  1. Agreement provides additional redundant (repeated) information to facilitate understanding for the hearer.
  2. Agreement helps the hearer to keep track of the different referents in a discourse.

Remarkably, the two central claims (agreement is redundant, and agreement is referential) continue to be repeated in the literature, despite the fact that, with very few exceptions, neither has ever been subjected to more rigorous testing (cf. Siewierska 1998, Bickel 2003), and both clearly admit counter examples.

Thus, the aim of this project is to test the proposed functions of agreement against a sample of 20 languages from all around the world. The results will be of central importance to the language science, and a test case for the applicability of text-based, as opposed to grammar-based, typology.

The project is conducted by Diana Forker and financially supported by the Daimler and Benz Foundation (

As part of the project Diana Forker and Geoffrey Haig organize a workshop at the University of Bamberg (1-2 February, 2013).