How Default is Causality-by-Default?

Abstract

Causal coherence relations seem to have special cognitive status in discourse processing. Even though they take less time to process than other types of coherence relations, recall and representation are better. The Causality-by-Default hypothesis explains this paradox by stating that readers maintain a causal coherence relation between two segments as a default assumption, and only reconsider the coherence relation when causal coherence turns out to be impossible. In this thesis I investigate the viability of this default assumption of causal coherence as a cognitive mechanism, and look into how specific characteristics of the segments and the surrounding discourse influence the assumption and the processing of causal coherence. The results of an eye tracking experiment support to the Causality-by- Default hypothesis, although coherence processing is affected by the characteristics in the segments, resulting in different processing patterns.