Centering in-the-large: Computing referential discourse segments

TL;DR

This paper presents an algorithm that constructs hierarchical discourse segments from local centering data, enabling analysis of global referential structures and improving understanding of anaphoric reference in discourse.

Contribution

It introduces a novel algorithm that scales the centering model to the global discourse level by building hierarchical referential segments from local data.

Findings

Algorithm successfully constructs hierarchical discourse segments

Global referential structure constrains antecedent reachability

Empirical evaluation demonstrates effectiveness of the approach

Abstract

We specify an algorithm that builds up a hierarchy of referential discourse segments from local centering data. The spatial extension and nesting of these discourse segments constrain the reachability of potential antecedents of an anaphoric expression beyond the local level of adjacent center pairs. Thus, the centering model is scaled up to the level of the global referential structure of discourse. An empirical evaluation of the algorithm is supplied.