False Memories

What would you say your memory is like? A library? A file cabinet? A camera that records daily occurrences and carefully stores important things in perfect state for later use?

We’ve briefly touched on this in a previous post. The brain does not simply store and replay an experience or image, like a computer. Instead, it re-creates it based on some cues, on the general idea behind the memory.

Everything we remember is, in fact, filtered through our current state of mind and experiences. It must fit the memory’s cues, but also the other memories and, to some extent, what we expect to remember.

This process is inevitably prone to mistakes. Several psychological experiments prove just how easy it is to be fooled by our own memories.

In this post, we’ll cover:

  • Imagination vs. Reality
  • Unintentional plagiarism
  • Creations and inferences
  • Eyewitness testimonies
  • 9/11

Imagination vs. Reality

Our brain cares not only for the permanent interaction with the external, “real” world. It also constitutes the scene for a rich internal life, ranging from reflections to dreams and imagination. We can recall, to some extent, both – external experiences, alongside ideas, dreams and imagined events. It is, then, important that we know one from the other.

This ability is called source monitoring – the capacity to identify the original source of information retrieved from memory.

In a 1998 study, Goff and Roediger set out to examine just how good people are at determining whether they had performed an action or just imagined it.

On the first day, 40 participants heard a command – e.g. “Break the toothpick” – and either performed it or just imagined performing it. The next day, they were asked to imagine a set of actions, some of which were the commanded ones. Lastly, 2 weeks later, they were tested on which actions they had imagined and which they had actually performed during the first day.

Results showed that, the more times participants had imagined a certain action, the more likely they were to think they had actually performed it on the first day, even if they had never actually even heard it.

In a second experiment, authors challenged whether the timing of the imagining session was the one creating the confusion. Indeed, it appeared that subjects confused actions when the imagining session took place immediately, 24 hours or 1 week after the hearing and performing session, but not if it happened just before the testing.

Their experiments go to show that memory is, indeed, imperfect and may be biased by other processes that create a perception-like experience, such as imagination.

Unintentional plagiarism

Mentioned by Schacter, in his reference article on the “ 7 sins of memory” (detailed in a previous post), unintentional plagiarism defines attributing an idea or memory to ourselves when it really belongs to someone else.

This memory glitch was first investigated in a laboratory environment in 1989, in a landmark study by Brown and Murphy. It included 24 students, divided into 4-person groups. 

In a first test, participants were given a word category - such as musical instruments, clothing, or four-legged animals. Each participant in the group had to, in turn, give an example within that category, making sure it had not been mentioned before. This cycle repeated 4 times, until each participant had mentioned 4 category items, and was done with 4 separate categories.

Afterwards, subjects were asked to write down their 4 words in each category, and then produce 4 completely new (unheard) items. 

Results showed that as many as 40% of subjects (10 participants out of 24) repeated an item produced by someone else even during the first, generation phase. When asked to recall their own 4 words, 75% of participants intruded at least one item that someone else had given. When asked to produce new items, 71% of subjects included at least one example given by someone else, amounting to 33 total intruded words (8.6% of all answers).

Subsequent experiments showed that this tendency to inadvertently plagiarize is consistent and reproducible in a social context, and not exclusive to heard or spoken information. A visual recall task produced similar results.

Despite the small sample size, these experiments paved the way for more research in the field. Other studies reproduced Brown and Murphy’s results against computer games, or examined how unconscious learning affects creativity.

I think we need to acknowledge that nothing we design is ever truly novel - every creative effort contains vestiges of what we have experienced in the past.

Richard L. Mash

Creations and inferences

One could argue that the above examples are not so much false memories as they are mistakes, glitches. Sometimes, however, memories can be completely unfounded.

The reason for this is precisely the constructive nature of memory, and has been proven in a classical experiment by Deese, in 1959. 50 students took part and were asked to listen to lists of words, then immediately recall them. Participants did not only produce new, unheard words, but seemed to do so with a certain tendency - the word “Sleep”, for example, being replaced by an intrusion in 44% of responses.

This may appear as a far-fetched, laboratory circumstance. A 2008 study by Brown and Marsh takes this experiment to a different level. They show that prior exposure to photographs can increase participants’ subsequent impression that they had actually visited certain locations. Authors argue that their brief laboratory experiences left an implicit impact on the autobiographical memory of participants.

When memories are made-up in an obvious manner, we diagnose that as confabulation - a neuropsychiatric disorder wherein a patient generates false memories unintentionally, often as a compensatory mechanism to fill holes in one’s recollection. They may range from a mistaken birthday or erroneous responses to historical questions (instead of the patient replying “I don’t know”) to complex autobiographical events, sometimes fantastical in nature. The key element here is that the patient has no ill intent and believes all his statements to be true, hence the term “honest lying”.

Although most commonly associated with severe states of alcohol withdrawal, an interesting occurrence of confabulation is the one that takes the form of the so-called repressed memories. During psychotherapy sessions, otherwise normal individuals have been reported to describe vivid, ‘repressed’ memories, often traumatic in nature, that were later proved to not have happened at all. This is coined as the false memory syndrome.

This clinical case presents a 26-year-old girl, with no prior psychiatric history, who started to recall vivid memories of sexual abuse perpertrated by her father, following a few sessions of hypnosis. After a meticulous clinical history, this was proven to be a form of confabulation and managed using mood stabilizers and psychotherapy.

Eyewitness testimonies

Psychologists have long argued that memory cannot always be relied on in instances of justice – often giving examples of mistaken eyewitness testimonies that led to unjust convictions.

It has since become a known issue, with The Innocence Project citing that over 70% of wrongful convictions are based on mistaken eyewitness identifications. The issue is slowly growing in popularity, with the Netflix series “Making a Murderer” documenting the story of a man who had served 18 years in prison as a consequence of wrongful conviction based on eyewitness identification.

Clearly, many factors play a role here – from stress to eyewitnesses discussing among themselves and altering their recollections, as well as the way the interview is performed, giving witnesses unintentional cues as to what is expected of them. Nevertheless, a labile memory is the foundation of all these issues, and raises several questions regarding the validity of eyewitness testimonies.

A team of UK scientists proved this in a controlled study. 600 participants simulated witnessing a bar fight. Then, actors were planted in some groups and instructed to suggest the wrong man had started the fight. Results showed that witnesses were susceptible to accepting the false information they were provided, and then include it in their statements. Even in the groups without any actors, 32% of participants gave incorrect statements. However, when more than 2 actors were planted in a group, as many as 80% of participants gave the same wrong statement, claiming the wrong man was the culprit.


People often assume that memories linked to a traumatic event are in pristine condition. They feel rich, vivid and detailed, often being coined “ flashbulb memories” - while other memories of the same period seem to fade, they maintain their content and clarity.

A consortium of over 3 000 individuals was created after the 9/11 tragedy, one to report and subsequently assess their ability to recall the event 1 week, 11 months and 35 months afterwards. Led by William Hirst and Elizabeth Phelps, a prominent voice in the field, the study describes a steady inevitable decay of 9/11 memories as time passes. After 1 year, participants offered consistent answers about their flashbulb memories only 63% of the time, on average. Over the next 2 years, the decay was attenuated, at about 4.5% per year. This is reminiscent of the classical forgetting curve - suggesting these memories are not, in fact, different from any other. In a 10-year follow-up, the memories seemed to have somewhat stabilized - with no important consistency changes between the 3-year and the 10-year responses.

Another study, including 54 Duke University students, showed similar results - their memories of 9/11 decayed 32 weeks later. However, vividness, recollection and especially belief in the accuracy of memory did not. Authors point out that flashbulb memories are, thus, not special in their accuracy or persistence, but rather in their perceived accuracy. 

In sum, our memory is not the reliable library we’d like it to be. It is a result of complex mechanisms that start out from a cue but rely on creative processes to bring about a recollection similar to the initial experience. How much we can rely on them, how to distinguish true events from adaptations, remains a challenge not only for current research, but perhaps for the judiciary system as well.


  1. How Memories are Distorted and Invented: Misattribution - PsyBlog. Available at: https://www.spring.org.uk/2008...;
  2. Goff, L. M. & Roediger, H. L. Imagination inflation for action events: Repeated imaginings lead to illusory recollections. Mem. Cogn. 26, 20–33 (1998).
  3. Schacter, D. L. The Seven Sins of Memory Insights From Psychology and Cognitive Neuroscience. (1999).
  4. Eyewitness Identification Reform - Innocence Project. Available at: https://www.innocenceproject.o...;
  5. New research reveals how little we can trust eyewitnesses. Available at: link. 
  6. Brown, A. S. & Murphy, D. R. Cryptomnesia: Delineating Inadvertent Plagiarism. J. Exp. Psychol. Learn. Mem. Cogn. 15, 432–442 (1989).
  7. Plagiarism or memory glitch? Available at: link.
  8. Deese, J. On the prediction of occurrence of particular verbal intrusions in immediate recall. J. Exp. Psychol. 58, 17–22 (1959).
  9. Brown, A. S. & Marsh, E. J. Evoking false beliefs about autobiographical experience. Psychon. Bull. Rev. 15, 186–190 (2008).
  10. Wiggins, A. & Bunin, J. L. Confabulation. (2020).
  11. Santos, G. & Costa, V. False memory syndrome: A review and emerging issues, following a clinical report. Eur. Psychiatry 33, S561–S561 (2016).
  12. Manhattan memory project: How 9/11 changed our brains | New Scientist. Available at: link. (Accessed: 26th August 2020)
  13. Hirst, W. et al. Long-Term Memory for the Terrorist Attack of September 11: Flashbulb Memories, Event Memories, and the Factors That Influence Their Retention. J. Exp. Psychol. Gen. 138, 161–176 (2009).
  14. Hirst, W. et al. A Ten-Year Follow-Up of a Study of Memory for the Attack of September 11, 2001: Flashbulb Memories and Memories for Flashbulb Events. Association 144, 604–623 (2015).
  15. Talarico, J. M. & Rubin, D. C. Confidence, not consistency, characterizes flashbulb memories. Psychological Science Research Article 14, (2003).