The unreliability of normal memory

The first issue, and the one addressed most comprehensively in the article, concerns the unreliability of normal memory. Accuracy of recall fades over time, naturally, but the issue is thornier than mere degradation. As the four protagonists in Akira Kurosawa's masterpiece Rashomon (1950) demonstrate, laying down and recalling memories is a self-serving reconstructive process: the four contradictory accounts of a man's murder and the rape of his wife reveal more about the witnesses than the actual events. This links to Gudjonsson's confabulated confessions, which stretches the concept beyond Korsakoff's original description (Korsakoff Reference Hirstein1996) and further than most psychiatrists would accept. But Gudjonsson's usage accords with the general direction of travel within neuroscience and philosophy. For instance, William Hirstein's volume on confabulations (Hirstein Reference Dennett2009) contains numerous examples of false memory in normal individuals. A striking example is so-called flashbulb memories, those that recall a dramatic event in vivid imagery, such as where one was when learning of the Twin Towers attack. It turns out that a significant portion of these memories include completely erroneous details, as the Oliver Sacks story described by Commane & Kopelman exemplifies.

But the concept of confabulation has been broadened further still to include all explanations whose true source is unknown. The entertaining paper ‘Telling more than we can know’ (Nisbett Reference Korsakoff1996) relates a plethora of examples and has become a classic of the psychological literature. One of their studies involves the influence of the position of an item in the visual field on decision-making: when individuals are shown an array of identical goods the rightmost is heavily over-chosen. But when asked to explain their decision, the positional effect is not mentioned spontaneously and is even strenuously denied when offered as a potential cause. Instead, individuals ‘confabulate’ various rationalisations. Consequently, it has become fashionable in philosophical circles to question whether we can ever know or claim anything about ourselves without falling into post hoc confabulations (Dennett Reference Commane and Kopelman1993). In Dennett's multiple drafts theory of consciousness, a multitude of mini-selves constantly produce competing narratives, only one of which enters conscious awareness. The narrative is not selected or composed consciously, nor does it arise from an illusory single cohesive self. Instead, it sustains the illusion. Commane & Kopelman do well to signal these issues without being drawn into the mire and they offer a balanced update on the wars over the inducibility of false memories for childhood sexual abuse. However, this equivocation may discourage psychiatrists from entering the fray in a legal setting.

Disordered memory

The second major theme of the article concerns disordered memory but primarily focuses on amnesia for an offence. Although this is not uncommon, particularly for violent crimes, and may occur in the absence of psychiatric disorder, it offers limited assistance to the defence. The law is quite carefully constructed to avoid rewarding those who claim, ‘I can't remember’, however genuine. The guidelines that follow in the section on good practice in psychiatric assessment and medico-legal work are sensible and pragmatic but offer scant direction where it is needed most: to what extent does accelerated forgetting in dementia affect fitness to plead? Has a witness's recall been impaired by mental disorder? Is the defendant lying or confabulating?