Abstract
Attribute amnesia (AA) reflects a failure to report a target-defining feature for a stimulus that had just been attended. For example, if participants are asked to report the location (i.e. the response attribute) of a letter among digits, they are unable to report the identity of that letter (i.e. key attribute) in a surprise trial. A summation of AA findings has led to the hypothesis that AA represents a lack of encoding of the key attribute. However, given the nature of the surprise paradigm used to assess AA (surprising event, response remapping, delayed responding, etc.), it is possible that decreased accuracy could be largely attributable to forgetting effects or interference caused by reading the surprise question. This study assessed the extent to which interference induced by unexpected reading contributes to decreased surprise trial accuracy. Participants completed a modified surprise trial paradigm: instead of being surprised with a new question, they were presented a task-unrelated passage to read before answering the same question as before surprise. In one condition a remember cue was presented before the reading passage. Without the cue, participants had poor accuracy following the reading passage, but with the cue performance increased dramatically. This remember cue was then applied to a standard surprise trial paradigm in which the question itself was unexpected. Though cueing to remember slightly improved surprise trial accuracy, it did not remove the AA effect. These results demonstrate that intentionally encoded information in working memory is more resilient to interference than briefly attended attributes that are not expected to be required for report. Thus, though some forgetting contributes to poor accuracy on the surprise trial in AA paradigms, our results support the hypothesis that AA is induced by a lack of encoding of the key attribute.