In my previous post I suggested that the epistemic faults listed in most definitions of delusions are not distinctive of delusions. Although delusions may diverge from norms of rationality to a greater extent than non-delusional beliefs, they are irrational in no special way. Excessively positive beliefs about ourselves, and widespread superstitious and racist attitudes tend to be, just like most delusions, implausible, badly supported by evidence and yet surprisingly difficult to revise (see chapter 3 of DOIB).
We are also very conservative with respect to our initial commitment to a scientific explanation for a given phenomenon, and new evidence against such an explanation is more easily discounted than evidence for it, independent of how robust the evidence actually is. We are not sufficiently rigorous in assessing the plausibility of the beliefs we acquire, and we do not easily give them up once we have acquired them, even when we did not have any personal investment into those beliefs to start with.
Considerations about the limitations of normal cognition show us that definitions of delusions based on their epistemic faults (such as the definitions in DSM-IV and DSM-5) cannot provide an efficient demarcation criterion, a principled way to tell delusions apart from non-delusions. Such definitions are still useful for the purposes of diagnosis and classification and, in the absence of consensus about how delusions are formed, they may be the best we can do. But we need an additional story to explain what makes delusions different. The same applies to confabulation, broadly defined as an “inaccurate or false narrative purporting to convey information about world or self” (Berrios 2000) or more narrowly identified as a distortion or fabrication of memory (Fotopoulou et al. 2008).
Confabulation is characterised in epistemic terms for the same reasons why delusion is: it can occur in the context of several psychiatric disorders and there is no agreement on how it is caused. Some accounts of confabulation are largely motivational and cognitive neuropsychological accounts focus on one or more of the following deficits: inability to access relevant memories; inability to determine the chronology of remembered events; exaggerated memory reconstruction; and poor source monitoring, as when something imagined is reported as something remembered. Epistemic definitions seem to be the only type of definitions capturing confabulation as a general phenomenon (see this paper for more details).
Confabulations and other memory distortions symptomatic of amnesia and dementia are often more implausible than their non-clinical counterparts, but they have epistemic faults of the same kind and their pathological nature is not explained by those faults alone. Memory distortions as inaccurate reconstructions of the past are very common in the non-clinical population. Our memories can be distorted, as Shachter convincingly argues, if we mistake something we have heard or imagined for something we experienced directly; if we are misled by the formulation of questions about past events; and if our current beliefs reshape our memories. Such distortions may not disrupt social functioning to a significant extent, but they affect our reliability as witnesses in court proceedings and as autobiographers when we think about our past selves.
To sum up, epistemic definitions of delusion and confabulation do not provide a sharp demarcation criterion between delusion and confabulation as clinical phenomena and other irrational beliefs or inaccurate memories in normal cognition. Moreover, by focusing on negative epistemic features alone, current definitions do not acknowledge that delusion and confabulation may carry potential epistemic advantages in the specific contexts in which they arise.
I shall explore this thought next.