Reconstructive memory refers to a class of memory theories that claim that the experience of remembering an event involves processes that make use of partial fragmentary information as well as a set of rules for combining that information into a coherent view of the past event. These theories provide a powerful way of understanding how witnesses remember crimes, how reliable recovered memories of abuse are, and how jurors remember testimony. According to reconstructive theories of memory, ordinary memory is prone to error. Errors in remembering can be broken down into errors of omission, in which information is left out of a memory report, and errors of commission, in which inaccurate information is added to a memory report. Errors of commission are more typically referred to as false memories or memory illusions. Reconstructive theories of memory generally hold that errors of omission and errors of commission are related to one another. In fact, according to reconstructive theories of memory, errors of commission occur because reconstructive processes are used to fill in gaps in our memory reports.
History of the Concept of Reconstructive Memory
Pioneering work on the development of reconstructive theories of memory was conducted by Bartlett and described in his classic volume entitled Remembering. According to Bartlett, remembering involves an active attempt to make sense out of the historical past—what Bartlett referred to as an “effort after meaning.” Bartlett studied the memories of English participants by asking them to repeatedly attempt to recall an unfamiliar folktale called The War of the Ghosts. Bartlett found that as participants attempted to recall the event, their recall was systematically distorted by their world knowledge. In particular, with repeated recall attempts, the unfamiliar folktale was recalled in an increasingly conventional manner. Details that were difficult to integrate with the participants’ world knowledge tended to drop out. Details consistent with world knowledge tended to be added. Unfamiliar words were replaced with more familiar words. Bartlett concluded that memory does not simply passively record or retrieve facts. Instead, memory combines fact and interpretation in a reconstructive way such that the two become indistinguishable.
In his pioneering text Cognitive Psychology, Neisser offered the analogy of a paleontologist reconstructing what a dinosaur must have looked like. According to Neisser’s analogy, paleontologists begin their reconstruction based on fragments of bone found in the fossil record. Based on this partial fragmentary information, the paleontologist makes use of his or her knowledge of finds at other sites, anatomy and physiology of current animals, and so on, to make a best guess of what the animal must have looked like, how it must have lived, what it likely ate, and so on. This best guess can be seen as a reconstruction of the past. Similarly, reconstructive theories of memory argue that people make use of partial fragmentary information, world knowledge, inferential processes, and so on, to reconstruct a memory of the past event.
The Process of Memory Reconstruction
According to most reconstructive theories of memory, the process of reconstructing a memory is based on a variety of different types of information. First, reconstruction relies on fragmentary pieces of information from the event itself. If one were to witness a bank robbery, details from that event would be stored in episodic memory. Over time, these details would become increasingly less accessible following the exponential forgetting curve first described by Hermann Ebbinghaus. Sometime later, the witness would be interviewed about the bank robbery. Although many of the details would be inaccessible, the witness would probably be able to retrieve some key pieces of information that made a special impression on him or her.
Schemas and Scripts
The stored details of the event provide partial evidence on which witnesses can base their memory reconstruction. However, this record of details from the event is likely to be incomplete. To help reconstruct the memory, witnesses would also likely rely on their prior knowledge about bank robberies in general. Memory psychologists have proposed that this type of prior knowledge is stored in long-term memory in the form of schemas and scripts. A schema is a general term we have for knowledge structures that represent typical instances of categories. Scripts are knowledge structures that represent the typical sequence in which a stereotypical event unfolds. For instance, a witness to a bank robbery likely has a schema representing the layout of a typical bank. They know that banks usually have offices or cubicles where loan officers, new account managers, and the like work. They know that banks typically have guards. They know that banks typically have tellers who work behind a counter. They know that banks typically have safes. This organized body of knowledge is thought to be stored in a “bank schema” that resides in memory. A witness to a bank robbery also likely has a bank robbery script, which includes information about the typical sequence of actions in a bank robbery. For example, a bank robbery script may include information like the robbers take out weapons, they disarm the guards, they demand money, the tellers provide them with money, the robbers make their escape, and so forth.
Schemas and scripts are thought to guide our understanding of events as they unfold and guide our recall of events as they are being remembered. Reconstructive theories of remembering suggest that schemas and scripts have two effects on our ability to remember events. They make actions that are inconsistent with the schema especially easy to remember because these actions require extra processing at the time of study to reconcile them with the schema. Schemas can also lead to false memories because they are used to fill in gaps in our memory for the event. If you cannot remember what happened in an event, the schema provides the default value you should expect.
In one classic study of the role of scripts on memory, participants were presented with a story about a young woman. Some of the participants were told that the story was about Helen Keller. Other participants were told that the story was about someone else. Participants who heard that the story was about Helen Keller falsely remembered facts from the story that were consistent with their world knowledge about Helen Keller (e.g., a book was written about her life). Other research has shown that participants are especially likely to correctly recall information that violates their expectations. For instance, when reading a story about a restaurant, one may remember unexpected events—such as the waiter spilling water—especially well.
Recently, researchers have shown that similar effects occur in forensically relevant settings. In one recent study, participants were shown a videotape of a bank robbery. The video included consistent and inconsistent schema, and irrelevant actions. Consistent with prior research on reconstructive memory, participants falsely recalled many details that were consistent with the robbery schema. In addition, the researchers found that participants used their bank robbery schema to interpret ambiguous information in the video.
Reconstructive theories of memory also claim that people rely on information obtained after the event to reconstruct their past. Information obtained after an event is known as postevent information. For instance, if one were to witness a bank robbery and then later saw a news report about the robbery, details from the news report may become incorporated into one’s memory for the event. Classic work on the role of postevent information was conducted by Loftus in the 1970s. In one study, participants watched a videotape of an auto accident. Some participants were asked to estimate how fast the cars were going when they “collided.” Other participants were asked to estimate how fast the cars were going when they “smashed” into each other. When tested 1 week later, participants who had been asked the “smashed” version of the question were more likely to remember seeing broken glass, when in fact no broken glass had been shown in the film.
The mechanisms by which postevent information influence memory became a subject of debate in the 1980s. Loftus proposed a theory whereby postevent information overwrites memory for the original information in storage. Other researchers argued that postevent information does not overwrite memory for the original event but rather interferes with the retrieval of the original event. Still other researchers argued that postevent information only influences memory reports in those participants who would not have remembered the detail in the first place. Later attempts to understand the influence of postevent information conceptualized it as an error in source memory. In other words, participants remember the information but have difficulty determining whether that information is from the original event or the postevent information (e.g., was it from the bank robbery or from the newspaper account?).
Work on postevent information has been extended in a wide variety of forensically important settings. Some research has examined the role of the interviewer in moderating the effects of postevent information. Social psychologists have shown that witnesses tend to discount postevent information when it is presented by a noncredible witness and to accept postevent information when it is presented by a credible witness. Also, in the 1980s, considerable research began to examine the role of postevent information in children. After some initial controversy, researchers reached a consensus that preschool-age children are more likely to be influenced by postevent information than are older children or adults. During this same time period, researchers came up with a number of clever research designs to examine children’s false memories in contexts with considerable ecological validity. For instance, researchers conducted a number of studies of children’s memories for stressful events by embedding postevent information experiments into children’s visits to their pediatrician.
The postevent information paradigm was further extended to examine adult memories for childhood events implanted by suggestion. The first of these studies involved implanting a childhood memory of being lost in a shopping mall in college students. Later researchers extended these findings using what has been termed the familial informant false narrative procedure. In this procedure, family members first complete a questionnaire about events from the participant’s childhood. Later, participants are interviewed about actual childhood events obtained from the cooperating family members and one invented childhood event (e.g., spilling punch on the parents of the bride at a family wedding). Participants are asked to repeatedly think about or imagine these invented events. Research has shown that false memories for childhood events can be created in 20% to 40% of participants using this technique.
In addition to fragmentary information from the event itself, prior knowledge in the form of scripts and schemas, and postevent information, some theories of reconstructive memory also assume that self-concept can influence how events are reconstructed. According to these theories, one’s self-concept can distort how events are remembered. One intriguing case study compared John Dean’s testimony at the House Watergate Hearings with taped transcripts of White House meetings involving Dean, Richard Nixon, H. R. Haldeman, and other White House officials. The study revealed that Dean’s memory appeared to show systematic distortions that tended to exaggerate his own role in those meetings. Thus, Dean’s memory showed a kind of self-serving bias. Later research on autobiographical memory showed that people’s memories could be distorted by their current self-concept.
Current Trends and Forensic Implications
Reconstructive theories of long-term memory provide a powerful way of understanding importantforensic issues such as how witnesses remember crimes and accidents, how adults remember childhood experiences, how children remember events, and even how jurors remember evidence. These theories stand in sharp contrast to reproductive theories of memory, which view memory as more like a videotape recorder. Research on reconstructive memories currently emphasizes the subjective experience of memories produced by reconstructive processes, whether true and false memories can be distinguished, how errors of commission can be avoided, and the individual differences that influence the use of reconstructive processes.
- Bartlett, F. (1932). Cambridge, UK: Cambridge University Press.
- Graesser, A. C., Woll, S. B., Kowalski, D. J., & Smith, D. A. (1980). Memory for typical and atypical actions in scripted activities. Journal of Experimental Psychology: Human Learning and Memory, 6, 503-515.
- Hyman, I. E., Husband, T. F., & Billings, J. F. (1995). False memories of childhood experiences. Applied Cognitive Psychology, 9, 181-197.
- Loftus, E. F. (1979). The malleability of human memory: Information introduced after we view an incident can transform memory. American Scientist, 67, 312-320.
- Neisser, U. (1967). Cognitive psychology. New York: Appleton-Century-Crofts.