Publication Date


Document Type


First Advisor

Millis, Keith K.

Second Advisor

Magliano, Joseph P.

Degree Name

Ph.D. (Doctor of Philosophy)

Legacy Department

Department of Psychology


Characters in narratives often behave in ways that are inconsistent with their beliefs, values, and previous actions. What effect do these contradictions have on a reader’s memory of character’s prior beliefs? One model, the Knowledge Revision Components framework (KReC) argues that a passive memory-based process is responsible for reducing access to information that is no longer true (e.g., when a character who is afraid of heights gets help facing her fears, it is harder to recall that she was afraid of heights). This possibility is called the general interference hypothesis. Subsequent research on KReC has shown that this process only occurs after a reader must resolve a contradiction (e.g., a character is afraid of heights later decides to go skydiving), suggesting that coherence building about contradictory information may be necessary to decrease access to outdated information. This possibility is called the coherence building hypothesis. The current dissertation aimed to test these hypotheses in two experiments.In Experiment 1, participants read stories that described characters possessing a trait (e.g., “Carol was extremely scared of heights”) and a subsequent behavior which contradicted this trait (e.g., “She [Carol] now really wanted to go skydiving”). In a causal elaboration condition, participants were given information that indicated the character no longer possessed that trait (e.g., the character sought psychological help to get rid of her fears) and in a contradiction condition, that information was not provided. Participants responded to verification probes about the character’s trait (e.g., “Carol was very afraid of heights”) to assess accessibility of the trait information. Response times to these probes were analyzed. The location of the probe occurred either immediately before or immediately after the target sentence about the behavior. A 2 (causal elaboration vs. contradiction condition) X 2 (probe before target vs. probe after target) within-participants design was used. Results indicated an interaction such that participants were slower to respond to probes when given an explanation for the contradictory behavior, but only if the probe occurred after the contradictory behavior. This indicates that when there is a coherence break where a reader must use the explanation to resolve the break, access to the outdated information is diminished. However, when the explanation is not used in resolving a coherence break, access to the outdated information is unimpeded. Experiment 1 thus supported the coherence building hypothesis. In Experiment 2, a similar procedure was used with slight modifications. First, because Experiment 1 demonstrated the importance of coherence breaks in reducing access to outdated information, verification probes always occurred after the contradictory behavior. Second, the character who performed the contradictory action was either the same one with the trait or a different character. Response times to the verification probes were once again analyzed. A 2 (causal elaboration vs. contradiction condition) X 2 (same character vs. different character) within-participants design was used. Unfortunately, due to massive data loss (63% of participants), no firm conclusions could be made about Experiment 2. The current dissertation offers some support for the coherence building hypothesis. This suggests that KReC be modified to emphasize the importance of coherence building around contradictions as a necessary mechanism for knowledge revision to occur. Additionally, the findings of the current dissertation also demonstrate that memory-based and constructionist perspectives operate in tandem to support comprehension.


116 pages




Northern Illinois University

Rights Statement

In Copyright

Rights Statement 2

NIU theses are protected by copyright. They may be viewed from Huskie Commons for any purpose, but reproduction or distribution in any format is prohibited without the written permission of the authors.

Media Type