Locus of the Moses Illusion: Imperfect encoding, retrieval, or match?

Abstract

Five experiments are described that attempt to isolate the mechanism that produces the failure to notice discrepancies in questions or assertions, called the Moses Illusion ( T. A. Erickson & M. E. Mattson, 1981, Learning and Verbal Behavior, 20, 540–552). Experiments 1 through 5 involved asking subjects to either (1) discriminate between distorted questions such as “How many animals of each kind did Moses take on the ark?” or (2) ignore the distortions and answer the question as if it were perfectly formed. Experiments 2, 3, and 5 also varied the familiarity of the assertions that were queried. Experiments 4 and 5 recorded reading times of the words in the question as well. Four alternative explanations for the Moses Illusion are considered: (1) the so-called illusion is just a cooperative response adopted by the listener/reader; (2) the retrieved memory structures are impoverished or incomplete and thus discrepancies cannot be detected; (3) the question is not carefully encoded and therefore the distorted word may not be fully processed during encoding; and (4) people often do incomplete matches between a complete representation of the question and a complete representation of the stored proposition that contains the answer. The evidence from these experiments does not support the first three alternatives, but is consistent with the partial-match hypothesis.