Lecture 7: Reconstructive Memory

In 1980 Elizabeth and Geoffrey Loftus published a notable paper in the American Psychologist about the permanence of memories. They posed this question to many people, both academic psychologists and nonpsychologists:
Which of these statements best reflects your view on how human memory works?
  1. Everything we learn is permanently stored in the mind, although sometimes particular details are not accessible. With hypnosis, or other special techniques, these inaccessible details could eventually be recovered.

  2. Some details that we learn may be permanently lost from memory. Such details would never be able to be recovered by hypnosis, or any other special technique, because these details are simply no longer there.
You may well argue that another option is missing, such as:
3. Memory storage is dynamic. Recall of past experiences and learning is reconstructed through the filter of the present retrieval environment. Remembering an event can change the representation of that event in memory.
Putting this serious qualification aside for the moment, the Loftus & Loftus paper does provide a convienient means of considering the many theories and experiments that pertain to the overarching question of the nature of human memory.

Freud's Archaelogical Metaphor

Sigmund Freud would have heartily endorsed position 1: he viewed memories as permanent and attributed inaccessibility of some aspects of the past to a lack of the appropriate retrieval cues. He likened the psychoanalyst to an archaelogist whose dig was human memory. The task of the therapist was to uncover deeper and deeper layers of memory using the tool of free association. Recall that in the last lecture I discussed Freud's concept of "screen memories"; the visual metaphor employed in this concept is telling; the screen hides deeper, but unaltered, memories of greater significance. The core idea is that memory representations, once formed, do not get changed by subsequent events, and a therapist can uncover the reminiscences that patients suffer from to release them from their psychological torment. As Elizabeth Loftus points out in a recent paper, this view is still held by some therapists:
"traumatic events create lasting visual images ... burned in visual impressions"
Leonore Terr, 1988

Penfield's Reawakening of Memories

In Loftus and Loftus's survey a remarkable 84% of psychologists and 69% of non-psychologists chose option 1, the permanence of memory option. Psychologists often cited the work of Wilder Penfield as crucial in forming their view. In the 1940s Penfield, a Canadian brain surgeon, stimulated areas of the brains of his patients while they were opened up for epileptic surgery. He found that electrical stimulation of the temporal lobes produced memories in some of his patients. For example, one patient reported hearing a song:
"... it was not as though I were imagining the tune to myself. I actually heard it."
The presentation of Penfield's findings in Introductory Psychology texts has been quite lavish (and involved some confabulation on the part of the authors). One example cited by Loftus and Loftus is from Zimbardo & Ruch's well known "Psychology and Life":
We are in the operating room of the Montreal Neurological Institute observing brain surgery on Buddy, a young man with uncontrollable epileptic seizures. The surgeon wants to operate to remove a tumor, but first he must discover what the consequences will be of removing various portions of the brain tissue surrounding the tumor ... Suddenly an unexpected response occurs. The patient is grinning; he is smiling; eyes opening when that area is stimulated. "Buddy, what happened, what did you just experience?" "Doc, I heard a song, or rather a part of a song, a melody." "Buddy, have you ever heard it before?" "Yes, I remember having heard it a long time ago, but I can't remember the name of the tune." When another brain site is stimulated, the patient recalls in vivid detail a thrilling childhood experience. ... As if by pushing an electronic memory button, the surgeon, Dr. Wilder Penfield, has touched memories stored silently for years in the recesses of his patient's brains.
In Penfield's own words:
"It is clear that the neuronal action that accompanies each suceeding state of consciousness leaves its permanent imprint on the brain. The imprint, or record, is a trail of facilitation of neuronal connections that can be followed again by an electric current many years later with no loss of detail, as though a tape recorder had been receiving it all."
The influence of both Sigmund Freud and Donald Hebb is evident in the word choices Penfield makes in describing his conclusion. This bold claim seems less plausible with a more careful examination of Penfield's data. Firstly, the numbers are not impressive: in only 40 out of 1,132 cases did he find any memory recovery; excluding patients who heard only music or voices and those whose responses were too vague to classify, less than 3% of the patients experienced the "lifelike memories" for which Penfield's work is so famous. Secondly, there was no attempt to check the veracity of the memories, and the patient protocols read like reconstructions, heavily based on inferences. Of course, many psychologists do argue that memory is a reconstructive process (see the section on Bartlett below), but that was not Penfield's view.

Bartlett's View of Remembering

From your reading of Schacter, you may have surmised that the memory experiment I have been conducting for the last three lectures is an attempted replication of Sir Frederic Bartlett's classic study of reconstructive memory (with a few new twists). In the 1920s Bartlett conducted a study of memory for stories using a Native American folktale that was unfamiliar to his English participants. He repeatedly asked them to recall the same story and examined the patterns of their memory renditions over time. He concluded that forgetting of details occurs in part because people are continually processing new information through mental structures, which he termed schemas, that are built up from the knowledge and beliefs that they already have about the world. He characterized this process as "the effort after meaning", the rationalizations and connections that have to be made to produce a coherent narrative from the remembered fragments of the story. This process of reconstruction through the filter of exisiting knowledge leads to conventionalization of the story. The same point is well illustrated by some earlier experiments by Philippe on memory for pictures (see below).
Bartlett's reproduction of a figure from Philippe showing that with successive reproductions the figure becomes more conventional and consistent with the generic schema of a human head.

Loftus's work on Misinformation

Partly inspired by Bartlett's approach to memory, Elizabeth Loftus has compiled a substantial body of experimental work that demonstrates the malleability of human memory. In one of the most well known of her interference effects studies she presented subjects with a series of 30 slides depicting successive stages in a traffic accident. For half the subjects a red Datsun was shown travelling towards a stop sign, and for the other half a yield sign was shown. The car was later shown colliding with a pedestrian. Immediately after viewing the slides, the subjects were asked some questions, one of which was critical. For half the subjects the question was, "Did another car pass the red Datsun when it was stopped at the stop sign?", for the other half "yield" was substituted for "stop". For half the subjects, this information was inconsistent. A twenty minute filler activity followed the questionnaire. Then the subjects were given a recognition tests for pairs of slides. When the postevent question contained consistent information 75% of the subjects accurately identified the picture of the car approaching the correct sign. In contrast, when the intervening question was misleading only 41% made the correct choice.

Just before leaving the experiment subjects were told its desgin and purpose, and were asked which condition they thought they had been in. 90% of the misled subjects insisted that they had seen the sign that corresponded to what they had been told.

In another experiment Loftus furthered her claim that nonexistent objects can be introduced into people's recollections. College students were shown a film of an accident and subsequently asked the misleading question, "How fast was the white sports car going when it passed the barn while travelling along the country road?". The control group was asked, "How fast was the white sports car going while travelling along the country road?". One week later 17% of the misled subjects said that they had seen a barn in the film, compared to 3% of the control group.

Moscovitch's Studies of Confabulation

In a previous lecture I discussed Moscovitch's work on patients with frontal lobe damage that led him to distinguish strategic or effortful retrieval from associative or involuntary retrieval. He believes that although the confabulations that often result from such damage are flagrantly distorted and easily elicited they nonetheless share many characteristics with the type of memory distortions we all produce. If we examine a transcript of his patient H.W. we find that the account is drawn from the patient's recollections of actual experiences. H.W. was 61 at the time, with widespread frontal damage.
Q. Can you tell me a little bit about yourself? How old are you?
A. I'm 40, 42, pardon me, 62.
Q. Are you married or single?
A. Married.
Q. How long have you been married?
A. About 4 months.
Q. What's your wife's name?
A. Martha
Q. How many children do you have?
A. Four. (He laughs) Not bad for 4 months!
Q. how old are your children?
A. The eldest is 32, his name is Bob, and the youngest is 22, his name is Joe. (These answers are close to the actual age of the boys.)
Q. How did you get these children in 4 months?
A. They're adopted.
Q. Who adopted them?
A. Martha and I.
Q. Immediately after you got married you wanted to adopt these older children?
A. Before we were married we adopted one of them, two of them. The eldest girl Brenda and Bob, and Joe and Diana since we were married.
Q. Does it all sound a little strange to you, what you are saying?
A. (He Laughs) I think it is a little strange.
Q. Your record says that you've been married for over 30 years. does that sound more reasonable to you?
A. No.
Q. Do you really believe that you've been married for 4 months?
A. Yes.
The confabulation is due to difficulties with time: he is married to Martha, they do have 4 children, but the time scale is all wrong. Confabulation about place is less readily elicited than that about time. This temporal disorder and the loss of source memory are characteristic of patients with frontal lobe damage. The frontal lobes are thus implicated in temporal organization of memory which may be particularly sensitive to disturbances in strategic retrieval processes. The fact that confabulations often refer to actually experienced events that are chronologically distorted suggests that temporal order is a property that is conferred on memories by retrieval processes. In a system that does not honor temporal order, the possibility of one event influencing another is very great.

It is also notable that confabulation occurs in other disorders associated with frontal lobe dysfunction, like schizophrenia; and in children whose frontal lobes are immature.

Sources: