Processing causal and diagnostic statements in discourse.

Abstract

Diagnostic statements (e.g., It is raining because the streets are wet) take longer to read than causal statements (e.g., The streets are wet because it is raining). The authors present 4 experiments investigating this phenomenon. In Experiment 1, a reading-time study, the authors demonstrate that this difficulty is reversed when a more complex mental model is cued through the use of phrases like John thinks that and John says that. Experiment 2 shows that the use of a modal construction (e.g., Perhaps it is raining because the streets are wet) makes the processing of diagnostics as easy as processing causals but does not disadvantage causals. The authors explain the pattern of results by proposing that readers build the simplest possible discourse representation during interpretation and that readers adopt a specific pattern of semantic interpretation. These proposals are tested and verified in Experiments 3 and 4.