Commentary on “Confabulation in Dementia: Constantly Compensating Memory Systems”

How are different memory systems involved in confabulation? What is the role of consciousness in patients who confabulate? What determines the content of confabulation? Tallberg concentrates on these questions and proposes that, in confabulating patients, semantic memory activates automatically to compensate for poor episodic memory and working memory, that a higher level of awareness—namely autonoetic consciousness—is not reached, and that unconscious motivation toward the preservation of a positive self-image determines the positive emotional bias observed in the content of confabulation. We shall focus only on the first two questions and their responses. We do not consider in our commentary Tallberg's answer to the third question, because it is based on the assumption that intentionality can be attributed to a mechanism or a process that is not conscious and, by definition, is inaccessible to consciousness. This assumption itself would need a long commentary (for a discussion of this point, see Dalla Barba, 2002).

Does semantic memory compensate for poor episodic memory and poor working memory, as Tallberg suggests? Both the analysis of the reports provided by Tallberg's patients and what is known from the literature on confabulation support a positive answer to this question. When asked what she had for dinner the day before, Britt (Case 1, Example 2) answers that she had ordinary food and that her grandparents then visited her. Karin (Case 2, Example 3), when she was interviewed in her hospital room, reported that she was on a train, on a journey to Gothenburg. In both examples, the content of confabulation can be reasonably traced back to firmly established acquired knowledge and information that replaces the unavailable episodic information. It is likely that Britt would generally have ordinary food at meals and that her grandparents used to visit her often when she was a child. It is also likely