Mind bending: Why our memories are not always our own

Without memories, we would be lost. Yet, in an extract from his new book,the psychologist Charles Fernyhough reveals that some of our most precious recollections are perhaps not ours at all.

Remember me? In tests carried out in New Zealand, 14 out of 20 sets of twins produced disputed memories (Getty Images)

When brothers and sisters are young, observed the psychologist Dorothy Rowe, they fight with each other for their parents' attention. When they are older, "siblings battle over who has the most truthful, accurate memory of their shared past".

Adult siblings generally do not face the same pressures as, say, married couples to agree on a story about their pasts. Individuals who have spent a lifetime trying to define themselves in opposition to each other are unlikely to be quite as motivated to settle their memory differences. And the fact is that adult siblings usually do not get as many opportunities as couples do to negotiate their memory disputes.

Why are some memories easier to negotiate than others? An obvious answer is that the people concerned are more committed to some of their memories than to others, and so are less willing to let go. But the study of sibling memories also convincingly demonstrates how two forces go head to head in memory. There is the drive to represent events accurately, which means being true to the often vivid impressions we have about what actually happened. And there is the drive for coherence, the need to produce a narrative whose elements fit together. In this case, coherence is a matter of agreement between people. Our stories need to make sense to us individually, but they also need to make sense to those who matter to us.

When you disagree with a sibling about your shared past, there can be the sense that the other is playing fast and loose with the cherished facts of your own life. What right does this competitor for parental affection have for rewriting your autobiography as they go? A particular kind of memory betrayal can happen when a sibling claims for him or herself an event that actually happened to you.

A study conducted in New Zealand showed that such memories are not at all uncommon. The researchers focused on adult twins, predicting that disputes over memory ownership would be particularly common in siblings who were more likely to look similar and share personality features, as well as being of the same age and thus presumably having shared more life experiences than ordinary siblings.

In their first experiment, the researchers asked their participants (20 same-sex pairs of twins) independently to produce autobiographical memories in response to cue words. Fourteen of the pairs produced disputed memories – memories that were claimed by each twin as having happened to them alone. For example, in one identical pair, both siblings remembered going out for lunch with their mother and finding a worm in their meal. One pair of identical twins seemed particularly susceptible to such errors, producing 14 disputed memories (out of a total, for the entire sample, of 36).

In a second experiment, a different sample of twins were specifically asked to report disputed and non- disputed memories, and to rate them on variables such as vividness of recollection and involvement of imagery and emotion. Intriguingly, the disputed memories were rated as being more vivid and emotionally rich than the ones on which the participants agreed, possibly because a greater effort had gone into constructing the memories that were not one's own.

In a third study, the researchers found that disputed memories were also reported by non-twin sibling pairs, although the experience was not quite as common as it had been for the twins. The researchers also noted that identical twins were no more susceptible to these distortions than non-identical ones.

A later analysis of the same data showed that there was a pattern to the claiming and giving away of memories. The researchers classified the disputed memories into those that showed the individual in a positive light (such as achievements or episodes of daring) and those that reflected more negatively (such as a memory of wrong-doing). "Self-serving" memories were more frequently claimed for the self, while those that reflected badly were more often attributed to the other sibling. If the person at the centre of the memory did something admirable, or had something bad happen to them (thus qualifying them for others' sympathy), then it tended to be claimed for the self. If the star of the memory was shown in a bad light, it tended to be passed off on to the other.

One 54-year-old identical twin, on hearing the other claim ownership of the memory of a roller-skating injury from when they were eight or nine, responded indignantly. "Well, that actually happened to me if you don't mind… I think you'll find if you think really hard it was me." The other, yielding ground, eventually responded: "Oh well, I guess we get confused; it happened so long ago."

Disputed memories seem to be an example of how the content of our memories can be affected by the stories that other people tell. Just as a parent can instil false memories in a child, so family members can shape each other's remembering. It's not only siblings who can relate such compelling memories that we start to have them for ourselves.

In recounting her memory of being a small girl and looking over a wall into her school, the writer AS Byatt associates the memory "with one of my very few good memories of my maternal grandmother – a perpetually cross person, who never smiled. The year she died," writes Byatt, "she began to forget [things], and forgot to be irritated. She said to me, sitting by the fire at Christmas, 'Do you remember all the beautiful young men in the fields?' And she smiled at me like a sensuous young girl. She may have been talking about the airmen who were billeted on her in the war – or she may have been remembering something from long before my mother was born. I shall never know. But I can see the young men in the fields."

Memories merge into memories. Byatt's grandmother's vivid remembering becomes the granddaughter's vivid imagining. Who can tell the difference? In time, we might become so convinced by other people's descriptions of their memories that we start to claim them as our own. If the experimental conditions are set up correctly, it turns out to be rather simple to give people memories for events they never actually experienced.

A well-known series of experiments by the American cognitive psychologist Elizabeth Loftus and her colleagues at the University of Washington has shown that presenting participants with misleading information after they have experienced an event can change their memory of the event. For example, a study published in 2005 involved participants viewing, among other things, a movie of a man stealing the wallet of a girl whose neck was hurt in the process. Subsequently, some of the subjects were exposed to misinformation about the event (being told, for example, that it was the girl's arm that was hurt, not her neck). In nearly half of the cases, the misinformation was incorporated into the participants' memories of the event.

Summarising research on the misinformation effect at that time, Loftus noted that hundreds of studies have demonstrated similar effects. People have recalled nonexistent objects such as broken glass. They have been misled into remembering a give-way sign as a stop sign, hammers as screwdrivers, and even something large, such as a barn, that was not part of the bucolic landscape by which an automobile happened to be driving. Details have been planted into memory for simulated events that were witnessed (a filmed accident, for example), but also into memory for real-world events, such as the planting of wounded animals (that were not seen) into memory for the scene of a tragic terrorist bombing that actually had occurred in Russia a few years earlier.

A strand of this research has concentrated on implanting "rich false memories", showing that the misinformation effect can apply to entire fictional episodes, not just details of events. People can be manipulated into remembering being lost in a shopping mall as a child, having an accident at a wedding, or meeting Bugs Bunny at a Disney resort. Crucially, researchers know that misinformation effects in scenarios such as the latter cannot stem from any genuine true memories, since Bugs Bunny is a Warner Brothers character and would never be seen at Disneyland.

The findings on rich false memories show that the misinformation effect is particularly strong when other people, especially family members, are providing the interjected information. Some benefits accrue to collaborative remembering, such as the everyday finding that couples can often help each other out by remembering bits of information that the other partner forgets. But there are negative effects as well. The term "social contagion" is used to describe the process whereby an account of an event incorporates erroneous information provided by other people. Another phenomenon, known as "collaborative inhibition", refers to the findings that a group of people who are allowed to discuss an event actually remember less about it than the same number of people tested individually. When others are around, it seems, we are less good at retrieving the factual details of an event.

A recent neuroimaging study has provided some of the first clues to the neural mechanisms involved when our memories are shaped by other people. Israeli and British researchers scanned the brains of 30 adults as they viewed and then (two weeks later) recalled a documentary-style film. Some of the participants were given erroneous information which they were told had been provided by other viewers of the movie. The researchers predicted that sometimes this information would actually corrupt the participants' own memories, and that at other times the subjects would simply go along with it out of a pressure to conform.

The scan findings showed that persistent memory errors, which went on to become part of the subjects' own retelling of the story, were associated with greater activation in the hippocampus (the brain region primarily responsible for laying down episodic memories) than transient errors, which seemed to be more about conforming to a public account of the events. The researchers also showed that the amygdala (a part of the brain responsible for emotional memory) was particularly active when the participants thought that the information had come from other people, as compared with computer-generated representations. They suggested that the amygdala, so closely connected to the hippocampus, may play a specific role in the process by which social influences shape our memories.

A striking thing about sibling disagreements is how long it takes for them to come to light. One friend, Zöe, an academic in her mid-forties, told me that she had only recently learnt the true facts behind what she had long suspected was her earliest memory. She has an image of her father coming into the kitchen from the garage and telling her and her mother that her baby brother had just drunk some motor oil. However, her brother recently informed her that he had always believed that the event happened out in the garden, and the oil was intended for the lawnmower. He has his own clear memory of picking up the red can from the grass, then the subsequent trauma of being intubated in hospital.

When Zöe discussed it with her parents, both confirmed her story, and she now thinks that she must have based her memory on conversations the family subsequently had about the event. Her brother, meanwhile, has had to accept that one aspect of his defining memory – how he came to drink the oil – might have been a fabrication. There is no doubt that he drank some oil and had the trauma of being separated from his mother and then of being intubated in a big, intimidating room in hospital. But a key detail is disputed, and to this day they are still arguing good-humouredly over the details.

There are many ways in which we can begin to doubt memories which previously convinced us. We may discover that a sibling also claims them, and thus that we must have appropriated them from another person's life story. We may hear a differing account which makes us appreciate how much our own memory is the product of distortion, suggestion and erroneous reconstruction. Are these fake memories any different in quality to our real ones?

A team of British researchers recently conducted the first scientific study of "nonbelieved memories": memories which people cease to believe after coming to realise that they are false. They concluded that nonbelieved memories are similar to ordinary "true" memories in many key respects. In their study, both kinds of memory involved a kind of mental time travel, the re-experiencing of intense emotions and perceptual details, and the reconstruction of some of the spatial and social features of the event. Commenting on the fact that both types of memory were experienced as single, coherent episodes, the authors note that a discredited memory can nevertheless retain a compelling power.

Findings such as these confirm that we can remember things that we don't believe actually happened, and vice versa. It is not necessary to believe that an event took place in order for us to experience it as a memory. One friend still has a clear memory of flying through the house as a child: standing at the top of the stairs, stretching out his "wings" and soaring down the staircase and through the downstairs rooms. Whatever it is that makes a memory, it is only partly connected to the possibility that it could actually have happened.

Memories are much more than fictional narratives, of course. Our memories are often very accurate, and only prone to serious distortion under certain conditions. To emphasise the narrative structure of memory is not to deny its potential veracity. You can make an analogy here with reportage, a dominant mode of journalism: just because it is told in the form of a story, it doesn't mean that it is not truthful. But when memory goes wrong, as in the case of some amnesias and distortions to the feeling of remembering, the stories can take over. Confabulation reminds us how the force of coherence can win over the force of correspondence, leading individuals to weave stories that fit their own reality better than they fit the reality out there.

The process works the other way round as well. Just as narrative feeds into memory, so memory feeds into narrative. Fictional writing comes alive through the inclusion of characters' memories. Would-be writers are always told to imagine what their characters think, feel and perceive, but they are not reminded often enough to give voice to their characters' memories. The work of novelists such as Hilary Mantel demonstrates the power of such imagined recollections. One of the most striking things about Mantel's justly lauded novel Wolf Hall is the way she bestows a richly imagined past upon her protagonist, Thomas Cromwell. In one scene, Cromwell recalls an erotic encounter in a Cyprus gambling den which merges, along a link of emotion, into another sexual memory, this time in Europe, with his lover Anselma. "Excuse me just a moment, she had said to him; she prayed in her own language, now coaxing, now almost threatening, and she must have teased from her silver saints some flicker of grace, or perceived some deflection in their glinting rectitude, because she stood up and turned to him, saying, 'I'm ready now,' tugging apart the silk ties of her gown so that he could take her breasts in his hands."

This beautiful, erotic scene is not happening in the "now" of the narrative, but in Cromwell's past. It is the fictional memory of a partly fictionalised character. Mantel presumably bases her scene on some real historical details about Cromwell's life. For the rest, she fills in the gaps with her own magic, fusing that biographical knowledge about Cromwell with sensory memories brought in from other aspects of her own experience. Other novelists, such as WG Sebald, create fictions that are almost memory constructions themselves: fragmentary, imagistic, fragile but striving for coherence. In Sebald's novel Austerlitz, for example, the protagonist's memory of his Welsh childhood is a brilliant rendering of the uncertainties and deceptions of childhood memories. In the writer and psychologist Keith Oatley's words, Sebald's work provides us with a kind of active remembering "in which the world and selfhood are continually constructed and reconstructed – from present-day events and from not-quite-intelligible fragments of the past".

When novelists make fictional memories, they are putting together many different kinds of information, from the conceptual to the immediately experiential, and arranging them in a way that meets the needs of the present act of storytelling. (In art, in fact, you could say that the force of coherence tends to win out against the force of correspondence, whereas in science it is the other way around.) That description of fictional memory-making could be an account of how our own autobiographical memories work. As they strive to understand how the different systems of memory pull together, memory researchers could do worse than read fiction. Paying attention to how an expert novelist constructs a memory provides us with a pretty good model of our own memory systems. Accepting the narrative nature of remembering does not destroy its magic. Stories are precious, and that applies equally to our own stories of the past.

Sometimes, though, we want to be clearer that we are getting the truth rather than a story. When I read a memoir, I am always being told: this is how it was. Here is this vivid picture. Feel the weight of that vividness, its guarantee of authenticity. How could I be creating this wonderfully colourful picture if I was making it all up? But the memoirist is, of course, making it up. He or she is a storyteller, as we are all storytellers. I know that memory doesn't allow for that kind of faithful representation of past events.

Although they are often seen in the same company, vividness does not guarantee authenticity, and Mantel is not entirely correct when, on another occasion, she complains about some of the psychological "tricks" that have been used to demonstrate the fallibility of memory. "Though my early memories are patchy," she writes, "I think they are not, or not entirely a confabulation, and I believe this because of their overwhelming sensory power." This is an understandable mistake, but a mistake it remains. The fictions that our minds furnish us with can have exactly that kind of sensory power, because of the way they are concocted in our brains.

This is an edited extract from 'Pieces of Light: The New Science of Memory' by Charles Fernyhough, published by Profile, priced £14.99