Reconstructive Memory

What is the Reconstructive Memory Theory?

Reconstructive memory theory concerns memory recall and postulates that the process of remembering is influenced by other internal cognitive processes, e.g. perception, imagination, attitudes, beliefs and semantic memory (i.e. our knowledge).

The Reconstructive Nature of Memory

Reconstructive memory is a theory of memory that states that memories consist not only of what we encode and store but is affected by prior knowledge in the form of schemas.

Bartlett (1932) first proposed the theory of reconstructive memory. Memories are not like a tape recorder that plays back an exact recording. Instead, it reconstructs them imaginatively. According to our schemas, we alter our memories to fit what we expect them to be. We reconstruct memories by trying to fit them into our existing schemas, and the more difficult this is, the more likely it is that some things will be forgotten and distortions will occur.

We have already established that our memory makes use of schemas. But how does this happen? We have schemas for all sorts of things, such as how we see a criminal or what counts as food. Our memory uses these schemas to organise things. When we remember an event, our schemas tell us what should happen. Schemas fill in the gaps in our memory (confabulation) and influence us to remember things that fit our schema.

Details in our memories can even be changed or removed. Sometimes we assimilate new information, i.e., we change our schemas to fit what we have learned. Other times, we accommodate new information, i.e., we change our memories to fit our existing schemas. Bartlett explains that accommodation occurs in two ways:

• Levelling: downplaying or removing details from memory.

• Sharpening: adding to or exaggerating details in our memory.Fig 1. - We accommodate new information by changing our memories to fit our existing schemas

Studies of Reconstructive Memory

Let’s now explore some studies of reconstructive memory to get a grip on the subject.

War of the Ghosts (Bartlett, 1932)

Twenty British males were told a Native American ghost story with several unusual features. Bartlett chose this story because it was culturally unfamiliar to the participants. Participants read the story and subsequently recalled it on several occasions: after several hours, weeks, days, months, or even years (repeated reproduction). They also had to read the story and reproduce it for other participants (serial reproduction).

Bartlett found that participants changed the story as they tried to remember it (a process called distortion). Three patterns of distortion occurred:

1. Assimilation: they changed the story to better match the participants’ cultural expectations (schemas), e.g., canoes and paddles became boats and oars. Thus, details of the story were unconsciously changed to fit British cultural norms. A memory was filled in by adding new information to make sense to the storyteller.
2. Levelling: the story also became shorter when participants retold it, omitting information they considered unimportant. The word count dropped from 330 to 180 (the shortest count was a retelling after the longest time, which was two years).
3. Sharpening: participants changed the order of events in the story to make more sense to them. They also used more familiar terms to them from their own culture. They also added details or emotions that were not initially present.

Overall, participants remembered the story’s main themes, but the unfamiliar aspects were changed to fit the participants’ own cultural expectations and be better remembered. This shows that we reconstruct our memories according to our schemas.

Allport and Postman (1947)

Allport and Postman (1947) showed participants a drawing of a quarrel on a subway train. Then they participants to describe the picture to another participant, then to another, and so on (serial reproduction). In the drawing, the black character was well dressed, while the white character had a rough appearance. Serial duplication seemed to reverse the description of their appearances, to the point of describing the black character as holding a knife (there were no knives in the original). Given the participants’ existing mental representations, information was altered through accommodation to correspond to and fit within their schemas.

Loftus and Palmer (1974)

Researchers showed participants film clips of actual car accidents and then gave them a series of questions to answer. There was one crucial question in the questionnaire about how fast the cars were going when they hit each other. Loftus and Palmer (1974) changed the intensity of the verb in this question, describing the speed of the cars, with a series of verbs such as ‘hit’, ‘smashed’, collided’.

Participants who were asked the question with the verb ‘smashed’ recalled a higher speed of the cars (average 40.5 mph) than those who were asked with the word ‘hit’ (average 34 mph). Another group of participants watched a clip about a car accident and then answered a questionnaire about the speed of the cars (the verbs were either ‘hit’ or ‘smashed’). A week later, participants had to complete another questionnaire with the crucial question, ‘Did you see any broken glass?’ (There wasn’t any in the clip).

Incorrect recall: who reported seeing broken glass?

• 12% of the control group (who were not asked the speed question at all).

• 14% of the ‘hit’ group.

• 32% of the ‘smashed’ group.

Loftus and Palmer concluded that eyewitness testimony is unreliable because leading questions can influence it. Memory is reconstructed; it changes as we incorporate new information we learn after an incident.

When we hear the word ‘smashed’, our schemas tell us there should be broken glass, and therefore we alter our memories to think that there was. After this change, it is difficult to distinguish which memories are true and false. There is no way back to the original.

Reconstructive Memory: Strengths and Weaknesses

Since the 1930s and Bartlett’s study War of the Ghosts, lots of further research supported the idea of schemas. Loftus conducted a series of laboratory experiments on reconstructive memory and had strict experimental controls, collected quantitative data, and standardised procedures, which made them quite objective and reliable.

Schemas can explain false memories.

The theory of reconstructive memory is similar to Tulving’s theory of semantic memory because schemas are similar to semantic stores in which we keep our understanding and knowledge of relationships and rules. Thus, if one is true, it makes the other more plausible. Moreover, this suggests semantic memory has an even greater influence on episodic memory since pre-existing knowledge (schemas) can influence our recollection of events according to the theory of reconstructive memory.

However, Bartlett’s study did not include many experimental controls. He asked participants to retell the story whenever it was convenient for them and did not set the same time frame for all participants. In addition, there was no scoring system. Bartlett measured the changes made to the story subjectively, in his opinion.

One criticism of reconstructive memory theory is that it does not explain how we reconstruct memories, unlike other cognitive theories that explain the processes involved. Other approaches mention specific processes and brain parts where they occur (using lesion and brain scan studies). We do not know how schemas are formed, how they alter memories, where they are located, etc.