Text integration errors can occur when a reader initially selects the contextually inappropriate meaning of an ambiguous word. When the reader attempts to integrate that meaning into the context, the sentence will not make sense. We investigated how readers recover from text integration errors in an oral reading study involving 2 types of ambiguous words: heterophones and homophones. Heterophones have 2 possible pronunciations and 2 meanings (e.g., sewer), whereas homophones have 1 pronunciation and 2 meanings (e.g., calf). We hypothesized that the multiple phonological codes of heterophones will cause additional difficulty initially and in recovery from text integration errors compared to homophones. Presumably, the working-memory code involved in reading is sound based (Daneman & Carpenter, 1983), so readers may reread to recover the alternative meaning and pronunciation of a heterophone but not of a homophone with a single pronunciation. We found that skilled readers made more errors initially on heterophones than homophones and used different strategies to recover from text integration errors caused by heterophone and homophone ambiguity.

©2012  Psi Chi - The International Honor Society in Psychology, All Rights Reserved