With regard to the analysis of confabulations, it would seem that confusion has arisen from mixing up levels of inquiry: phenomenology, neurobiology, disease associations, aetiological speculation and even pragmatics. At a most general level, 'confabulations' should be considered as sharing a conceptual space with delusions, mythomania, 'pseudologia fantastica' and 'pathological lying'. Two phenomena are conventionally included under the name 'confabulation'. The first type concerns 'untrue' utterances of subjects with memory impairment; often provoked or elicited by the interviewer, these confabulations are accompanied by little conviction and are believed by most clinicians to be caused by the (conscious or unconscious) need to 'cover up' for some memory deficit. Researchers wanting to escape the 'intentionality' dilemma have made use of additional factors such as presence of frontal lobe pathology, dysexecutive syndrome, difficulty with the temporal dating of memories leading to an inability temporally to string out memory data, etc.

With respect to mechanisms underlying paramnesias, Kraepelin was less clear. He mentioned 'alterations of consciousness' which occurred with varying severity and caused a reduced capacity to differentiate between reality and fantasy. During the nineteenth century, the generic term paramnesia was used to refer to a group of clinical phenomena amongst which déjà vu, confabulations, and delusions and hallucinations of memory remain the more salient. These phenomena had been known since earlier but it was only after the work of Sander that they began to be considered as 'memory' disorders. In Kraepelin's taxonomy the paramnesias are included as 'qualitative' disorders of memory affecting either recognition or recollection. Déjà vu remains a curiosity seen in some forms of epilepsy and occasionally in the normal affected by fatigue. Delusions of memory are occasionally mentioned in the literature but hallucinations of memory have disappeared altogether.