Commentary by Marcia K. Johnson

DeLuca and his colleagues (DeLuca and Cicerone, 1991; DeLuca, 1993; DeLuca and Diamond, 1995) have provided some of the most systematic observations of confabulating patients that we have available. Therefore his views about our current level of understanding of this fascinating neuropsychological phenomenon are of particular interest. In the interest of stimulating a conversation, I will focus on several areas where I think DeLuca's analysis is problematic, needs further clarification or development, or perhaps underestimates the progress that has been made in understanding the cognitive mechanisms of confabulation.

Most definitions of confabulation have in common the idea that patients may make statements that are false or engage in behaviors that reflect false memories or beliefs, and that they do this without an intention to deceive (e.g., Whitlock, 1981; Moscovitch, 1989; Johnson, 1991). DeLuca adopts such a definition for what he calls “confabulation in the broad sense,” but suggests that little progress has been made in the last hundred years in understanding confabulation because cognitive neuropsychologists have not come to a consensus about whether there is one type or two (or more) distinct forms of confabulation.

A more optimistic view is that researchers have pointed out that confabulations (even from the same patient) differ on a number of dimensions. Confabulations differ in general content—whether they are about current visual experience (Anton's syndrome); identity of persons (Capgras syndrome) or places (reduplicative paramensia); experiences of body parts (denial of paralysis); or about autobiographical episodes, semantic knowledge, or beliefs. They also differ in bizarreness, whether spontaneously offered or given in answer to questions, and whether they are acted upon. They may also differ in the degree to which patients are aware (at least initially, see belo