'Footfalls echo in the memory/Down the passage which we did not take/Towards the door we never opened' (Eliot 1963): more often than we tend to concede, memory leads us to passages we never took in reality. Fantasy and facts mix well in poetry, but the charm could become nightmare if they do so in reality. Why does it happen, when and how, are major questions in memory research. These questions have been approached in experimental psychology long ago (Bartlett 1932; Carmichael et al. 1932), almost forgotten ("zeitgeist), but regained much interest in recent years. Their renewed discussion has already transcended the domain of science, invading courtrooms and newsrooms alike.
There is a great variety of phenomena that could be regarded as false memory. They range from delusions and "confabulations in diseased states to memory illusions in normal individuals (Kopelman 1999; Koriat et al. 2000). In the contemporary literature, however, the term is commonly reserved to refer to the erroneous memory, particularly "episodic memory, in normal subjects. In the past decade, particular attention has been directed at the potential role of false memory in recollection in adulthood of sexual abuse in childhood (Penfold 1996; Pope 1996; Winbolt 1996). It all started with a wave of reports on the recovery of such repressed memories in subjects undergoing psychotherapy. The apparent revival of trauma has culminated in litigations against parents and caretakers. Soon after, however, it became apparent that in some cases (not all!), the memories of abuse could have been acquired by the patient in the course of overenthusiastic therapy. In fact, a new diagnostic criterion has been suggested, termed 'false memory syndrome', which refers to the situation in which an individual uncritically accepts suggestions of the therapist and comes to believe illusory memories of abuse. At this stage, therapists in lieu of parents became fashionable targets for legal suits. This has led to heated debates and emotional flares in the psychotherapeutic community (ibid.), and, as expected, to juicy headlines in tabloids.
Yet false memory in "real-life covers many phenomena that have nothing to do with claims of abuse. A pressing and more prevalent problem refers to the questionable validity of courtroom testimony (Wells and Loftus 1984). Ordinary subjects are capable of adopting and assimilating a fabricated autobiography, in the absence of any malicious intention (Loftus 1996). Last but not least, false memories are not confined to individuals; when societies and nations adopt false "collective memories, the resulting fantasies and emotions could lead to global disasters.
Situations that resemble real-life false memory can be simulated in "controlled laboratory settings. The procedures typically involve the "recall or "recognition, often "cued, of stories, other verbal material, or visual scenes (e.g., Bartlett 1932; Carmichael et al. 1932; Miller and Gazzaniga 1998; Tversky and Marsh 2000). A "classic study is that of Deese (1959b).1 The test material in this study was 36 lists of 12 words each. Each list was composed of the 12 primary associations of a target word, which itself was excluded from the list. For example, for the target word 'sleep', the list was 'bed, rest, awake, tired, dream, wake, snooze, blanket, doze, slumber, snore, nap'. The subjects were instructed to listen carefully to each list and then, at the conclusion of each list, recall orally the items from the list. Deese's intention was to identify the occurrence of extralist intrusions of words in the immediate recall. He found that many of the lists induced the subjects to produce the nonpresented target word as an intrusion on the test. 'Sleep', mentioned above, had in his hands the largest probability of intruding into the relevant list; words like 'rough', 'soft', 'chair', 'foot', 'cold', and others also starred as intruders in the recall of their associative list. This is hence a clear-cut case of false memory: the subjects claimed to have heard the target word, which they did not. The method was later improved and expanded to include a recognition test by Roediger and McDermott (1995). Word-intrusion tests are highly recommended for convincing demonstration of false memory in the classroom as well as in public lectures.
Why are some memories false? The question could be posed in two versions: first, what are the brain mechanisms that generate false memories, and second, why haven't some of our memory "systems been perfected in evolution to yield a higher fidelity of output. At the mechanistic "level, multiple possibilities could be entertained, which relate to distinct memory "phases: perception, "acquisition, "consolidation, "persistence, and "retrieval. False memories might stem from perceptual illusions (Roediger 1996), even from dreams (Loftus 1996) and fantasies (Freud 1899). It is, however, questionable whether these should be considered as a source of bona fide false memory, because from the point of view of the brain, the raw material for memory was real. Distortions in the processing of "percepts might be due to the amalgamation of the on-line experience with off-line experiences from different times and "contexts, and to reconstruction as well as repression of narratives by fitting them to emotive and cognitive schemata (Freud 1899; Bartlett 1932; Carmichael et al. 1932; Neisser 1967; Loftus et al. 1995; Koriat et al. 2000; Anderson and Green 2001). Consolidation, processes that might take place after a memory item is retrieved in a new context, 'implicit' could particularly provide an opportunity for new information to modify the initial trace (Nader et al. 2000; Sara 2000; Berman and Dudai 2001). The cues available in retrieval could also cause subjects to select, rearrange and distort the retrieved information (Tulving 1983; Loftus 1996). The introduction of improved neuropsychological assays of false memory, especially when combined with "functional neuroimaging, will surely cast additional light on the mechanisms involved. It has already led to the identification of frontal areas that are recruited differentially in false vs. veridical recall (Schacter et al. 1996c).
As to the phylogenetic considerations, several types of possibilities should be kept in mind. One, that faulty reproduction of information is an inherent constraint of the biological hardware that embodies memory. In other words, the system is imperfect because it could not be otherwise, unless we replace the hardware.2 Second, the system still undergoes evolution, we are in the process, just give us time and we will fare much better. Only that time here could mean zillions of years, unless, again, we call bionics to the flag ("enigma). And third, which is related to the second, who says that accuracy is always a positive selective pressure in evolution? It could be so for some types of memory systems, such as "skill, but not for others, such as episodic memory. There are no doubt situations in which accurate recollection of episodes is but a burden; Plutarch (1-2C ad/1914a) was among those who recognized that: 'I dislike', he said, 'a drinking-companion with a good memory'. The possibility was raised that autobiographical memory had evolved to bind our personality and allow us to function better as distinct individuals in society; for this, autobiographical memory need not be neither large (Dudai 1997a) nor accurate (Conway 1996). Yet another possibility is that in moulding declarative memory systems, evolution actually selects against accuracy of details, because excessive accuracy may hamper "generalization and categorization ("mnemonics). The Rashomon phenomenon3 may hence be a price we pay for the cognitive success of our species.
There is much more to the discipline of false memory than the phenomenon itself. The rediscovery and enthusiastic analysis of false memory phenomena in the past few decades has catalysed a conceptual revolution in the field of memory research (Koriat et al. 2000). This revolution is concerned with the shattering of the concept of the brain as a faithful mirror of reality, and replacing it with the idea that "internal representations mix the filtered percepts of the world with percepts acquired at other times, as well as with innate
("a priori) and endogenously generated representations. In addition, the static 'storehouse' "metaphors are passé; many biological memories are "biased, dynamic reconstructions of past occurrences.4 What comes out of the 'storehouse' is not what was deposited there.
Selected associaitons: Bias, Cue, Real-life memory, Retrieval, Zeitgeist
1The fate of this signal paper in the years following its publication is discussed in *zeitgeist.
2See in this context the 'Panglosian paradigm', in *paradigm.
3Termed after the *classic film Rashomon, by Akira Kurosawa, in which four conflicting versions of the same traumatic event are offered by four narrators (Kurosawa 1950; Cook 1981). 4A caveat is appropriate here. This statement merely claims that the individual human being is an untrustable witness. It does not claim that there are no facts outside there that could be identified and quantified by methods that control for bias in individual memory. This is what science attempts to do. For relevant debates, see Appleby et al. (1996); for the occasional misuse of science to support irrelevant claims such as the one this caveat tries to prevent, see Sokal and Bricmont (1998).
Was this article helpful?