Coherence relations in a cognitive theory of discourse representation

Coherence relations such as Cause—Consequence and Claim—Argument
establish the link between two discourse segments. A review of existing
accounts of coherence relations shows that there is no consensus on the
number of relations needed nor on their exact nature.We take coherence relations as cognitive entities that play a central role in both discourse understanding and production. We propose a concise theory of coherence relations, with a limited number of cognitively basic concepts.
It is hypothesized that readers use their knowledge of these basic concepts to
infer the coherence relations. As a result, the set of relations is organized
into a taxonomy in terms off our primitives that apply to all relations, such
as the polarity of the relation. To test the taxonomy, two experiments were carried out in which sentence pairs connected by coherence relations were presented to subjects. The experimental items were prototypical examples of the relations in the taxonomy. In one experiment, subjects categorized experimental items on the basis of the relation connecting the sentence pairs, in the other they labelled the relations. The results indicate that the 12 classes distinguished in the taxonomy are intuitively plausible and applicable and that the primitives
underlying the taxonomy are psychologically salient. In addition, the taxonomy is presented in the form of an outline of a process model that may provide a generally applicable systematic basis for an algorithm of coherence relation under standing. Consequences for such varying areas as linguistics, discourse processing, computational linguistics, and language acquisition are discussed.