Could you tell the difference between a person’s memory and their imagination?

If I gave you a written description of an object – let’s say a boat – would you be able to judge whether the author had written about the boat from their memory of it, as opposed to having written about a boat they’d imagined?

It’s a question with real-world importance because, in court, we often rely on eyewitness memories and it’s up to a jury to determine their source and veracity. But memory, like the imagination, is a creative process. Sometimes the two even become blurred – it’s quite common for people to lose faith in an apparent memory, to wonder if they had in fact imagined it. Deliberate deception aside, how well can we judge the source of another person’s recall?

Arlo Clark-foos and his colleagues recruited 20 volunteers to either look at pictures of various objects  or to imagine those objects after being prompted by the words denoting them. Examples include boat, microwave, and chair. These volunteers were told they’d later have to recall as much about the imagined or viewed objects as possible. After a five minute distraction period during which they answered maths questions, the volunteers were asked to write about the objects, both those they’d looked at, and those they’d imagined.

The transcripts were cleaned up to remove any overt clues as to the source of the descriptions – such as phrases like “I imagined …”. Then the transcripts were shown to thirty participants recruited from a  university, whose task was to read each one and decide if the description was based on the author’s memory for the object, or on their having imagined the object. The participants were told that there had been no deliberate deception in the written accounts.

To the researchers’ surprise, the participants were able to make this distinction with modest accuracy. Specifically, descriptions of seen objects were correctly identified as such 63 per cent of the time, descriptions of imagined objects 57 per cent of the time. Although only modestly accurate, this is significantly more accurate than you’d expect had the participants simply been guessing.

The researchers said this result was all the more remarkable because when they’d carefully analysed the accounts of seen and imagined objects looking for qualitative differences along 40 dimensions, they found very few – descriptions of seen objects contained more references to the season of the year (perhaps because the pictures gave relevant clues) and more author doubts about memory accuracy; by contrast, in the descriptions of imagined objects, there tended to be more references to sounds and locations (perhaps because this information was missing from the pictures).

In further experiments, the researchers established that people could be trained to be even more accurate at this task by giving them multiple examples of object descriptions written from memory and descriptions written from the imagination, with each labelled as such. Feedback also boosted performance during blind testing – that is, telling participants whether each memory/imagination judgment they made was accurate or not.

“Although the current results are curious,” the researchers said, “they only begin to explain how participants made their judgments. Future experimental work investigating the characteristics and decision processes that aid in resolving the source of others’ memories will increase the applicability of these findings.”
Clark-Foos, A., Brewer, G., & Marsh, R. (2015). Judging the reality of others’ memories Memory, 23 (3), 427-436 DOI: 10.1080/09658211.2014.893364

Post written by Christian Jarrett (@psych_writer) for the BPS Research Digest.