Distinguishing between cases in which subjects remember successfully and cases in which they remember unsuccessfully, resulting in memory errors, is central to theories of remembering. Successful remembering implies the retrieval of information aligned with real events or experiences where the memory process culminates in an authentic and accurate recollection, while unsuccessful remembering may involve inaccuracies and distortions, hence creating false memories, such as misremembering or confabulation (see Hirstein 2009, Michaelian 2016, 2022; Robins 2016, 2017, 2020). According to Robins, it is possible to draw a comparison between errors in perception (illusion and hallucination) and those in memory (misremembering and confabulation). Starting from this account, the aim of this presentation will be to clarify the definition of confabulation to understand whether it is possible to support an analogy between confabulations and hallucinations, expanding this comparison with a phenomenological perspective.
First, I will explore the definitions of confabulation. Although it is possible to discuss the phenomenology, it is very challenging to present a unified definition. Usually, definitions are divided into narrow (i.e., definitions that are mainly based on that of 'false memory') and broad. I will draw a distinction between clinical and non-clinical (or 'voluntarily caused') confabulations. Among the clinical ones, I will analyze fantastical, or incorrectly fitting the events confabulations. Among the non-clinical ones, I will consider the case of suggestibility, or the memory implantation studies.
Having explained if and in what sense confabulations can be considered a phenomenon similar to hallucinations, we can question whether a phenomenological approach can be employed to support and expand this comparison. According to Husserl, we first need to distinguish between a perception that is a presentation [Gegenwärtigung] and memory, phantasy, expectation and image consciousness that are re-presentations [Vergegenwärtigung]. From a phenomenological point of view, memory is usually considered to be rooted in perception, precisely because what is remembered is remembered as it was perceived. It means, therefore, that if I remember the red apple I had for breakfast this morning, I am also remembering that this morning the red apple was present ‘here and now in the flesh’ and that I perceived it. I remember it as an earlier perception of the red apple, as something that stands in relation to me as having been present. What about confabulations? What happens if in the act of memory, we are directed towards something that did not happen?
What I would like to suggest, is that imagination can be the common ground between confabulations and hallucinations. Considering the imagination as a shared source between hallucinations and memory begins with experiments like ‘imagination inflation’, from which we can derive a high degree of confidence in false memories (Garry et al., 1996). The point in these experiments was to draw on the ‘memory trace’ left by our imagination of something to show that there is a higher probability of remembering incorrectly as real experiences things we have already imagined or dreamed about. In a study demonstrating this phenomenon, participants were given simple action statements (like 'break the toothpick') and, at times, they either acted out or imagined performing the action (Goff and Roediger, 1998). The more they imagined the action, the higher the likelihood of misremembering it as something they actually did. Although these results must be counterbalanced by experiments showing that subjects have the (metacognitive) ability to distinguish between memories of imaginations and memories of perceptual experiences, they certainly show an interesting relationship between imagination, memory, and perception. Precisely on this basis, I will show how, although the nature of hallucinations and confabulations may differ, their origin may be phenomenologically closely related to the subject's ability to imagine.