Cases like these raise significant questions for students of memory. Is it really possible for vivid recollections to be completely fabricated? Isn't there a "grain of truth" in even the most distorted memory? And on the other hand, does the existence of false memories mean that no memory can ever be trusted? Are all our recollections wrong? To what extent?
The answer to the first question is simply yes. People can indeed have convincing memories of things that never happened. This has long been known for suggestions made in hypnosis (e.g., Laurence & Perry, 1983), but it can also happen without any hypnotic involvement. Richard Ofshe (1992) recently suggested a novel memory of sexual abuse to a man who had already been arrested for, and confessed to, many other bizarre abu-sive acts. The suggestion was readily accepted, and an elaborated version of the suggested experience was soon incorporated into the "perpetra-tor's" self-narrative. A very different set of examples comes from recent work by Elizabeth Loftus (1993). Using plausible suggestions made by trusted relatives, Loftus has successfully implanted false recollections of having been lost in a shopping mall when a child into the memories of normal adults. Her subjects, who soon elaborate these "memories" in plausible ways, are fully convinced that the event actually happened. A similar dy-namic was probably responsible for Jean Piaget's (1962) famous memory of a childhood kidnapping attempt that later turned out to have been a complete fabrication.
Another recent study is worth mentioning because it shows that false memories can appear even without explicit suggestion. On the morning after the space shuttle Challenger exploded in 1986, I asked a group of Emory students to record how they had first heard the news on the pre-ceding day. Three years later, the 44 who were still on campus were again asked to recall how they had heard about Challenger, and to rate their confidence in various aspects of that memory (Neisser & Harsch, 1992). About a quarter of those accounts, including many reported with high confidence, were entirely wrong. One subject, who in fact had first heard about the disaster from fellow students in class, later recalled being stunned by the news while watching TV with her roommate. Another, who had actually heard the news from friends at lunch in the cafeteria, remembered how "some girl came running down the hall screaming 'The space shuttle just blew up/" In subsequent interviews, we showed the subjects their original 1986 questionnaires. Even then they were reluctant to abandon their incorrect memories: "Yes, that's my handwriting - but I still remember it this other way!" Established self-narratives are very hard to change.
What about the notion that there is "a grain of truth" in every memory? Freud, of all people, seems to have believed this. He thought of psycho-analysis as comparable to archaeology (Freud, 1937/1964), a science that retrieves genuine fragments from the past and constructs essentially valid scenarios of ancient events. (At first sight this seems to conflict with his earlier rejection of the seduction theory, but the contradiction is only ap-parent. The truth-grains in which Freud believed were fragments of fan-tasies or images, not memories of actual events.) As Donald Spence (1982) has shown, this claim reflects a basic confusion between "narrative truth" and "historical truth." Images and memories are never simply "observed" by the patient and then "reported" to the analyst, as the archaeological metaphor would imply. They are always constructs, shaped by the shared need to establish a psychoanalytically satisfactory narrative of the pa-tient's mental development.
In denying that memories must contain grains of truth, I do not mean to deny that they have comprehensible causes. Sometimes those causes are obvious. The subject who falsely remembered first learning about Challenger from TV, for example, had probably watched television cover-age of the disaster later that day. It is harder to guess the origin of the mythical girl who ran screaming through the dorm, but she too must have started somewhere. (I am tempted by Freud's fantasy hypothesis in this case. Perhaps the subject herself felt like screaming, and transformed her feeling into an image that eventually became a memory.) False memo-ries of childhood sexual abuse also have their sources, but those sources
need not lie in the actual childhood of the patient. It is often supposed that "something must have been going on" to give rise to such a memory, but this assumption is not justified. It is at least as likely that the underly-ing ideas were acquired later, perhaps in adult life.
The fact that some memories can be dramatically mistaken does not mean that all memories are wrong. That would be impossible. Many rec-ollections concern events that have present consequences: I remember my wedding, and am still married; remember moving to Atlanta and still live there. Such consequences do not guarantee the accuracy of details, which may indeed be mistakenly remembered, but they do constrain the main outline of the narrative. In addition, many recollections involve other people. We may have to construct an "agreed or communicatively successful version of what really happened" (Edwards & Potter, 1992,
p. 210) with someone else who was there too. The process of construc-tion may uncover real differences, but it usually reveals substantial con-sensus as well.
In recent years a number of psychologists have conducted "diary stud-ies," in which subjects record their experiences on a daily basis and later test their memories of those experiences (Linton, 1982; Wagenaar, 1986, and chap. 10 of this volume; Brewer, 1988; White, 1989; Larsen, 1992; Hirst, chap. 14 of this volume). These studies show a good deal of forget-ting, but not much misremembering: There are very few overt errors or confabulations. Diarists may be unusually dependable people, or - as Eugene Winograd observes in chapter 13 of this volume - they may just be unusually careful because they know their memories will soon be checked.
Another relevant line of research focuses on memory in young chil-dren. Two-, 3-, or 4-year-olds can be asked to recall specific life events: holidays, trips to the zoo, and the like (e.g., Fivush, Gray, & Fromhoff, 1987; Fivush & Hamond, 1990). Experimenters trained to be sympa-thetic and yet not suggestive can elicit a surprising amount of information about the past, even from very young subjects. Although most of the memories are fragmentary and incomplete, they rarely include confabu-lations or serious errors. This does not mean that young children are always right; on the contrary, they are very vulnerable to suggestion (Ceci 8c Bruck, 1993). That vulnerability is important for understanding a different set of child-abuse accusations, those based not on long-delayed recollections by adults but on contemporary accounts given by the chil-dren themselves. Some of those accounts are certainly valid, but others are just the product of prolonged and suggestive interrogation. Chil-dren's memory, like that of adults, is more or less on the mark in many situations but vulnerable to suggestion in others.
www.psychspace.com心理学空间网
