Reasoning and delusions

Abstract

Background: Delusions are assumed to reflect disordered reasoning, but with little empirical support. We attempted to study this in 16 relatively intelligent deluded patients and 16 normal volunteers. Method: Standard tests were used which required subjects to choose between logically fallacious and valid responses, both of which were plausible. The tests were: (a) conditional statements (if... then), (b) syllogisms (e.g. no A are C, some B are C, some C are not A), and (c) judgements of probability. All three tasks incorporated neutral and emotive content. Results: Both normal and deluded subjects frequently made logical errors. With conditionals, deluded subjects tended to endorse fallacies more often than normal controls and this was accentuated when the content was emotive. Similarly with syllogisms, the effect of emotional content on the endorsement of unbelievable responses was increased slightly in the deluded group. Finally, the deluded patients showed a trend to be test prone to the conjunction fallacy than normals, suggesting less reliance on existing schema. Conclusions: Differences in reasoning between deluded patients and controls are surprisingly small. Patients are somewhat more prone to endorse invalid or fallacious responses, especially when emotive themes are involved.