Subjectively, memory feels like a camera that faithfully records and replays details of our past. In fact, memory is a reconstructive process prone to systematic biases and errors—reliable at times, and unreliable at others. Memories are a combination of new and old knowledge, personal beliefs, and one's own and others' expectations. We blend these ingredients in forming a past that conforms to one's haphazardly accurate view of oneself and the world.
Reconstructing the Past
Traditionally, psychologists were interested in the temporal retention of information. Since the early 1930s, many psychologists have shifted their focus from the quantity of memory to its accuracy (Koriat, Goldsmith, and Pansky, 2000). The British psychologist Frederic C. Bartlett (1932) conducted one of the first systematic investigations of memory accuracy. He asked subjects to read a legend about Indian hunters called "The War of the Ghosts" and then to retell it to another subject who had not read it. The second subject then told the story to another subject, and so on, until ten subjects had heard it. The story involves two young Indian hunters who meet a group of men in a canoe, who, in turn, invite the hunters to join them in battle upriver. One young Indian accepts and the other declines. During battle, the young Indian is wounded and realizes that the men of the war party are ghosts. He returns home, recounts his tale, and dies the next morning.
Bartlett found that subjects retained the overall gist of the story but that they also revised the story, systematically omitting and modifying details. For example, subjects omitted mystical references, such as ghosts, which are not part of Westerners' worldview; they embellished other details. In the original story, the second Indian declined to join the party because his relatives would not know his whereabouts. By the tenth retelling, one subject explained that this Indian refused because his elderly mother was dependent on him, a revision that manifests Western concepts of a son's responsibilities in general and perhaps that subject's family ties in particular. Another common change was that subjects tended to add a moral, possibly because stories in Western culture often have morals. Bartlett concluded, "Remembering … is an imaginative reconstruction, or construction, built out of the relation of our attitude towards a whole mass of organized past reactions or experiences" (p. 213).
Bartlett's study exemplifies how time and retelling distort the memory of stories. Another study conducted in the early 1930s using ambiguous drawings showed that what we are told that we are viewing easily distorts visual material. If people are shown two circles and a line and are told that the picture represents either glasses or dumbbells, subjects' later drawings of the original picture will assume the suggested appearance (Carmichael, Hogan, and Walter, 1932).
There are many other studies that demonstrate the malleability of memory for words, stories, and pictures. For example, Henry Roediger and Kathleen McDermott (1995) altered a procedure originally developed by James Deese in which people study lists of closely related words like bed, pillow, tired, and dream. When later asked to recall studied words, subjects frequently claim that they saw other words like sleep that were not presented but are related to those that were. Subjects often assert these "false memories" with a high degree of confidence and detail (e.g., that a male as opposed to a female voice spoke the word). There is some debate about whether subjects generate the word sleep while studying the word list or later, when asked to recall the entire word list. In either case, people draw inferences—not necessarily accurate—about their present and past experiences.
Yet another way to demonstrate memory's attempt at synthesis is to present subjects with successive, thematically related slides depicting common routines like going grocery shopping. One slide shows a woman putting a box of items into her shopping cart. The next slide shows several oranges on the ground. When subjects are asked later to recognize slides that had previously been shown, they mistakenly say that they saw a slide depicting the woman removing an orange from the bottom of a pile of oranges (Hannigan and Tippens-Reinitz, 2001). They make this causal inference because people naturally attempt to piece together the fragments of their past in order to make memory as coherent as possible.
Work on the "misinformation effect" further demonstrates the ease with which accumulated information skews memory (Loftus, 1979). In the misinformation effect, misleading information about an event from one's past reduces the accuracy of the memory of an event. In one study, Elizabeth Loftus and colleagues showed subjects a simulated automobile-pedestrian accident (Loftus, Miller, and Burns, 1978): a vehicle stops at an intersection, turns right, and then hits a pedestrian. Half the subjects viewed a stop sign at the intersection. After viewing the scene, these subjects were asked a question that mentioned either a stop sign or a yield sign. In a final memory test, these subjects were asked whether they saw a stop sign or a yield sign. When the subjects were asked a question consistent with what they had seen, they chose the correct sign 75 percent of the time. However, when the question was inconsistent with what they had seen, they chose the correct sign only 41 percent of the time. These investigators concluded that some subjects had initially encoded a stop sign in memory but that the subsequent mention of a yield sign altered their memory.
Does the new information alter the original memory trace, or does it coexist with the original information in memory (Ayers and Reder, 1999)? According to the altered-trace view, the original memory is changed permanently and is inaccessible to recollection. According to the coexistence view, the original information is still accessible with the right retrieval cues. The issue of memory's permanence remains a fundamental, unresolved question in memory research. But whatever your view about the underlying memory traces, it is clear that the memory reports of subjects are changed, and many subjects appear to believe strongly in their misinformation memories.
Although the evidence indicates that our memories are malleable and easily manipulated, there are circumstances in which memory is relatively resistant to change. For example, if people publicly state that they remember a detail, subsequent suggestions are less likely to induce a change of mind. There is also resistance to changed recollection in the face of gross disparities between clearly perceived details and contradicting misinformation. It is also possible to reduce misinformation effects by warning people about misleading messages or by requiring subjects to determine the precise source of the misinformation—for example, "Did I see the flat tire in the film, or did I hear or read about it after I saw the film?" (Loftus, 1997). Thus, memory is reconstructive, and reconstructions are susceptible to—but not powerless against—subsequent misleading information.
The previous examples demonstrate the disturbing ease with which the details of memory can be manipulated. But it doesn't stop there—it is also possible to implant entire false memories. People can be led to believe that, as children, they were lost in a shopping mall or that they had knocked over a punch bowl at a wedding and spilled punch on the bride's parents (Hyman, Husband, and Billings, 1995; Loftus and Pickrell, 1995). In a series of interviews, Loftus and Pickrell asked subjects to recall as much as possible about four childhood event descriptions that a relative had provided. Three of these events were true, and one was false: that the subject had been lost in a shopping mall at the age of five for an extended time and had been rescued by an elderly woman and reunited with the family. In three suggestive interviews, during which subjects were led to believe all the events occurred, subjects remembered the real events about 70 percent of the time and the false ones about 25 percent of the time (see Figure 1).
Imagination offers another way to implant false memories. Subjects are asked to imagine in detail an event that never occurred. Later, they are asked to rate their confidence that the event truly happened. The act of imagination typically causes subjects to increase their confidence in the reality of these events. Although some researchers argue that certain memories are highly resistant to suggestion and imagination, others have shown that it is even possible to increase people's confidence that they had witnessed demonic possession as a child (Mazzoni, Loftus, and Kirsch, 2001). These studies indicate that implantation of entirely false memories is possible.
Once implanted, the false memory is often barely distinguishable from real ones. Research has shown that there can be statistical differences between a group of real memories and a group of false ones: For example, the real memories possessing more sensory detail (Heaps and Nash, 2001; Schooler, Gerhard, and Loftus, 1986). But people can give detailed descriptions of their false memories that sometimes lead them and others to regard the memories as real. Moreover, it is difficult to determine the truth or falsity of a single memory report.
Another line of research aims to determine whether true and false memories elicit different brain activity. In such work, subjects read a list of closely related words and later try to recognize whether or not they had previously seen those words and other novel but related words. During the recognition phase of the experiment, subjects' brain activity is monitored by sophisticated neuroimaging tools like magnetic resonance imaging (MRI) or event-related potentials (ERPs). There is some preliminary evidence that neuroimaging may permit scientists to glimpse the neural signatures of true and false memories (Fabiani, Stadler, and Wessels, 2000); however, more work is needed to confirm the utility of this approach.
Evaluation and Attribution
It is clear that memory can fail in a variety of ways. However, the precise reason why memory fails is less clear. Larry Jacoby and others have shown that the manner in which people evaluate their present processing in light of the past may explain in part both how and why memory fails. For example, the ease with which a memory comes to mind after exposure to misinformation or after imagining the memory in question may rightly or wrongly lead the person to believe that the memory is real. In fact, unless there is another, more likely, reason or source to explain why a memory or experience currently feels familiar, people will typically attribute feelings of familiarity to past experience (Jacoby, Kelley, and Dywan, 1989; Whittlesea and Williams, 2001). Thus, it is possible to influence memory by changing the way in which the present experience is processed, evaluated, and then attributed to the past.
Far from a reliably faithful rendering of the past, memory is a reconstruction that usually retains the gist but not the details of bygone experiences. The recounting of one's past, the exposure to misleading postevent information and suggestion, integration of thematically related material, and imagination are several of the means by which memory is constructed—or misconstructed. Once reconstructed, the original memory may prove elusive. Given the potential fallibility of our recollections, it is surprising that memory functions as well as it does.
See also:FALSE MEMORIES
Ayers, M. S., and Reder, L. M. (1999). A theoretical review of the misinformation effect: Predictions from an activation-based memory model. Psychonomic Bulletin & Review 5, 1-21.
Bartlett, F. C. (1932). Remembering: A study in experimental and social psychology. New York: Macmillan.
Carmichael, L., Hogan, H. P., and Walter, A. A. (1932). An experimental study of the effect of language on the reproduction of visually perceived form. Journal of Experimental Psychology 15, 73-86.
Fabiani, M., Stadler, M. A., and Wessels, P. M. (2000). True but not false memories produce a sensory signature in human lateralized brain potentials. Journal of Cognitive Neuroscience 12, 941-949.
Hannigan, S. L., and Tippens-Reinitz, M. T. (2001). A demonstration and comparison of two types of inference-based memory errors. Journal of Experimental Psychology: Learning, Memory, and Cognition 27, 931-940.
Heaps, C. M., and Nash, M. (2001). Comparing recollective experience in true and false autobiographical memories. Journal of Experimental Psychology: Learning, Memory, and Cognition 27, 920-930.
Hyman, I. E., Jr., Husband, T. H., and Billings, F. J. (1995). False memories of childhood experiences. Applied Cognitive Psychology 9, 181-197.
Jacoby, L. L., Kelley, C. M., and Dywan, J. (1989). Memory attributions. In H. L. Roediger III, and F. I. M. Craik, eds., Varieties of memory and consciousness: Essays in honour of Endel Tulving. Hillsdale, NJ: Erlbaum.
Koriat, A., Goldsmith, M., and Pansky, A. (2000). Toward a psychology of memory accuracy. Annual Review of Psychology 51, 481-537.
Loftus, E. F. (1979). Eyewitness testimony. Cambridge, MA: Harvard University Press.
—— (1997). Creating false memories. Scientific American 277, 70-75.
Loftus, E. F., Miller, D. G., and Burns, H. J. (1978). Semantic integration of verbal information into a visual memory. Journal of Experimental Psychology 4, 19-31.
Loftus, E. F., and Pickrell, J. E. (1995). The formation of false memories. Psychiatric Annals 25, 720-725.
Mazzoni, G. A. L., Loftus, E. F., and Kirsch, I. (2001). Changing beliefs about implausible autobiographical events: A little plausibility goes a long way. Journal of Experimental Psychology: Applied 7, 51-59.
Roediger, H. L., III., and McDermott, K. B. (1995). Creating false memories: Remembering words not presented in lists. Journal of Experimental Psychology: Learning, Memory, and Cognition 21, 803-814.
Schooler, J. W., Gerhard, D., and Loftus, E. F. (1986). Qualities of the unreal. Journal of Experimental Psychology: Learning Memory, and Cognition 12, 171-181.
Whittlesea, B. W. A., and Williams, L. D. (2001). The discrepancy-attribution hypothesis: II. Expectation, uncertainty, surprise, and feelings of familiarity. Journal of Experimental Psychology: Learning, Memory, and Cognition 27, 14-33.
Revised byDaniel M.Bernstein