Background. This experiment examines two aspects of delusional cognition that have been reported clinically but not investigated empirically. These are the incorporation of potentially conflicting information into the recall of delusion-related scripts and the type and amount of material produced additional to that presented for recall, referred to here as confabulation.

Methods. Three groups of patients – deluded schizophrenics, non-deluded schizophrenics and matched non-psychiatric controls – were asked to recall two 15-item scripts, which comprised 10 typical and five atypical components. It was hypothesized that deluded subjects whose delusion was relevant to one of the scripts would recall more of the atypical components of the script and would also be less likely to make script-atypical confabulations in the recall of this particular script.

Results. Recall was assessed for the amount and type of content remembered and the amount and type of confabulation. The results did not support the hypothesis that atypical items would be incorporated into the recall of delusion-relevant material. However, deluded subjects did retain their schema boundaries in the recall of script items relevant to their own delusion but were less able to adhere to a script framework in the recall of material unrelated to their delusion.

Conclusions. These results are discussed within a schema specific account of delusions, which conceptualizes the delusion as an overused schema whose preferential use leads to a failure to develop other scripts but whose own contents remain well-defined.