Reconstruction From Memory in Naturalistic Environments

Imagine your friends are over at your place, and your brother shows some embarrassing home video of you from when you were a kid. Would you rather have your brother show the video or tell the story in this situation? If you're well aware that memory is reconstructive, your best bet may probably be the second option.

• First, we will look at memory reconstruction in naturalistic environment research.
• Next, we'll go over memory reconstruction in the model of the naturalistic environment.
• Then, discuss memory reconstruction in naturalistic environments theory.
• Lastly, we'll explore the critical research of Steyvers & Hemmer (2012): reconstruction from memory in naturalistic environments.

Memory Reconstruction in Naturalistic Environments Research

According to Frederic Charles Bartlett (1932), memories are reconstructive. This means we don't store and play back memories like a tape recorder. Our memory undergoes structural changes over time, leading to forgetting details and inaccuracies upon retrieval. Our prior knowledge or schema assists memory retrieval by filling in the gaps.

Let's look at how Bartlett demonstrated memory reconstruction in his 'War of the Ghosts' study.

He asked students from Cambridge University to read a famous North American folktale entitled 'War of the Ghosts.' He presented a story from a different cultural background to see whether the participants' schema would affect the story recollection.

The results showed that the students recalled the story based on their cultural expectations, which Bartlett called rationalisation. Participants left out some details, such as those specific to North American culture. Bartlett concluded that schemas could impact memory, such as when recalling stories that conflict with your prior knowledge (e.g., different culture).

With the introduction of memory reconstruction, psychology learned what influences memory retrieval, such as prior knowledge and other cognitive processes. Here are some examples of significant memory reconstruction research:

Fig. 1 Car collision

Loftus and Palmer (1974) investigated whether verbs (e.g., smashed and contacted) used in a question can influence memory. Participants watched a vehicle accident video and estimated car speeds before the collision. When the question included “smashed,” participants reported a broken glass (even though there was none) and that the car traveled 41 mph. On the other hand, when the verb “contacted” was part of the question, people said the car was going 32 mph.

They concluded that when we recall a memory, we also try to access and reconstruct it based on knowledge from other occurrences since then, which can alter our memory.

In another study of memory reconstruction, Dooling and Christiaansen (1977) had people read a passage about a woman named Carol Harris. According to the text, Carol Harris was an uncontrollable child, even at the tender age of eight, leading her parents to believe that they needed to seek professional treatment. Her parents invested in private tutoring for her. Researchers examined participants' passage memory after a week. Before the test, the researchers told half the participants that the passage was about Helen Keller.

They found that people who “knew” the text was about Helen Keller remembered reading that the girl was deaf, even though it's not in the passage. Hearing about Helen Keller activated their prior knowledge about her, which led them to remember something that wasn't in the passage.

While these studies demonstrate how prior knowledge can cause errors in our memory, interestingly, two other researchers, namely Steyvers and Hemmer (2012), argued the opposite, that prior knowledge could improve memory in everyday situations. As you will read in the later section, you will see how Steyvers and Hemmer demonstrated reconstruction from memory in naturalistic environments.

Memory Reconstruction in Naturalistic Environments Model

Earlier, we've seen lab evidence of the reconstructive nature of memory. But before we better understand reconstruction from memory in naturalistic environments, let's look at what happens during memory retrieval and the reconstructive memory model.

Memory retrieval

If you lost your keys, naturally, you would try to remember where you last put them. Is it in your pocket? Or on the kitchen counter? These are cues we use when retrieving a target memory.

As you try to access a particular memory, your mind begins to search for cues about the event. These cues help you retrieve your target memory through associations.

Memory retrieval involves cues and associations that lead you to the target memory. It's also important to note that cues are anything that may well reflect the content of the target memory we're trying to access.

In what specific way do cues and associations help in memory retrieval? One way is through spreading activation. Cues “spread energy” to other memories associated with the target memory. The higher the activation memory trace receives from a cue, the more likely one can access and retrieve the memory.

The activation level increases when a cue is physically present (e.g., seeing a photo) or when we try to think about the memory trace (e.g., thinking about the movie).

Reconstructive memory model

Earlier, we saw how prior knowledge fills in the memory gaps as we reconstruct memories. Now let's look further at how the reconstructive memory model explains memory reconstruction.

This information, along with the script, makes up the memory of the event. We glean information from these components when attempting to recall an experience. If we have a clear idea of what we believe happened (a script), we may omit or change details so that they conform to the script. Furthermore, recalled facts could potentially be incorrect if the script from memory is incorrect.

The reconstructive memory model also suggests that we don't store everything as correctly in our memory when the event occurred, leading us to retrieve similar memories when prompted to recollect a particular memory.

This process creates new mental representations of the incident, which we might use to retell the recollection.

Memory Reconstruction in Naturalistic Environments Theory

Earlier, we mentioned Bartlett's previous memory research, which showed prior knowledge affects our memory. However, this model focused mainly on memory errors due to prior knowledge.

In particular, lab studies of this nature have aimed to draw out false memories. This is because the method often withholds something from what the participants need to remember, such as the item they expect to see the most. This means that studies like this one had low ecological validity. The conclusion was that prior knowledge changes episodic memories.

Fig. 2 Naturalistic office environment

The conclusion was that prior knowledge changes episodic memories. However, Steyvers & Hemmer (2012) theorised that prior knowledge could help save and enhance memories. This is because, based on what we already know, we would expect to find certain things in a scene, e.g., a computer in an office.

In lab studies of memories, stimuli usually are manipulated, like in our example of an office scene with no books. Studies like this don't reflect a naturalistic environment. Therefore, Steyvers & Hemmer (2012) theorised that if there are experiments aimed at finding the effect of prior knowledge on memory in naturalistic environments, the result would be positive rather than negative.

Steyvers & Hemmer (2012): Reconstruction from Memory in Naturalistic Environments

To investigate how prior knowledge affects the reconstruction from memory in naturalistic environments—the interaction between episodic memories and prior knowledge. All participants were from the University of California. The researchers divided them into three groups (expectation test, perception test, and experimental memory condition).

Phase 1 - Initial Testing

Expectation Test: 22 participants named objects they would expect to find in five different naturalistic settings: office, kitchen, hotel, urban setting, and dining room. They had to name the objects by typing them on a screen and had one minute for each scene. The researchers measured the frequency of all objects participants had named, e.g., all 22 participants expected to see a television in a hotel room.

Fig. 3 Naturalistic hotel room

Perception Test: Another group of 22 participants looked at 25 images of these five naturalistic scenes. They had to recall all the objects they had seen. The researchers presented the objects according to their recall frequency (most and least recalled).

Phase 2 - Experiment

Forty-nine participants who hadn't been part of either initial test were randomly selected. From the perception test in the initial testing, the researchers took ten images — two from each of the five scenes (office, kitchen, hotel, urban setting, and dining room). These were the two that had the highest recall frequency.

The researchers split participants into two groups. Group 1 looked at the images for two seconds each. The hypothesis was that the participants would not remember all the details due to the short time. Because of this, they would rely on semantic knowledge when recalling objects in the scenes. Group 2 looked at five images for ten seconds each. This was an episodic memory test, as these participants had more time to look at the images. The participants recalled the objects they saw in the images.

Results of the Study

In the two-second condition, the participants correctly recalled an average of 7.75 objects. In the ten-second condition, the score was higher — 10.05 objects. For both situations, the average accuracy was 90 percent. The researchers also examined how often the participants named objects that weren't in the scenes. Incorrect recall of high-probability objects was nine percent (probability calculation conducted from the initial tests). Incorrect recall of low-probability objects was 18 percent.

Fig. 4 Naturalistic urban setting

According to this model, semantic knowledge can help accurately recall episodic memories of naturalistic settings. This is beneficial as more cognitive resources can be used to attend to other parts of the scene, such as novel items.