The trouble with repressed memories

August 27, 2020

Articles | Collections

The notion of repressed memories gets a lot of attention, which has done more harm than good. Memory is a mercurial thing, but by examining it we learn that the kind of memory doesn't matter. It's the emotion that's the key.


In 1934, John Henry Wigmore put forward his hugely influential Treatise on Evidence. In it, he warns that women and children are likely to accuse men of 'good character' of sexual abuse and should thus be subject to examination by a psychiatrist. This is a problem, because Wigmore's work lingers in justice systems today.

It was Freud's work on sexual abuse that popularised the idea that one could 'recover' memories of sexual abuse during therapy. Today we know this idea to be largely more harmful than helpful. Yet the notion flourishes every 30 years or so sparking a surge of accusations that later turn out to be more commonly the product of an imaginative therapist's leading questions of a vulnerable patient.

This is not to say that sexual abuse can't be repressed, but that Freud's methodology was deeply flawed and his influence outsized. Wigmore's thinking was influenced in no small part by one of these blossoms of misplaced enthusiasm. In the light of many distressing stories that are surfacing in this #metoo moment, and the ongoing exposure of huge rings of sexual exploitation perpetrated by powerful individuals largely unchecked by our social norms and our justice systems, one wonders how different the last 70 years would have been if Freud and his fans would have spent more time adjusting their sails.

So let's adjust our sails.

The first question we should be asking

"Recovered" memories of sexual abuse are often surprisingly detailed accounts, at least for those that get publicised. This should pique our interest because memory is notoriously problematic. Another challenge in the legal system, for example, is the fact that eyewitness testimony is horribly unreliable.

Memory of surprising, emotionally intense, and negative events is often thought to be more accurate. But only in terms of their central details, as opposed to peripheral information. There's also a question of how one is supposed to measure the accuracy of a memory. The event is gone, so what are we measuring against? And, as we mentioned earlier, eyewitness testimony of dramatic events is a serious problem in the justice system. So these memories might be more accurate, but that isn't the same as accurate.

The question therefore, that we should have, is if we are suppressing a memory for some period of time, would we be able to remember it reliably? And how reliably?

Memory is not reliable, so why would a suppressed memory be?

The science of forgetting

We should start with a quick primer on how memory works.

Memory, like many things in the brain, is a bit of a mysterious function. It has been tempting since the advent of the computer to think of our memories like a kind of database storage, but this is probably not how most, if any, memories work. Rather, memory seems like it can be broken into some number of different kinds. Memory for semantic information (i.e. information) appears to be distinct from episodic memory (i.e. events or experiences). We also have a muscle memory, which may be different again. We also appear to have some kind of distinction between memory we hold in our head for a short time (e.g. repeating a phone number to yourself so you don't forget it) and stuff that stays in the brain for a long time. Beyond that, it's still guesswork for the most part and there are multiple models of memory and memory processes that float around the academic space.

From the perspective of actual neural function, we can make some more helpful conclusions.

We know, for example, that the amygdala plays a role in charging events with emotional significance. This is associated with better recall. It's a sensible mechanism that helps us to learn the significance of events in the world.

The hippocampus appears to play an extremely important role in the storage of memories. Damage to the hippocampus almost always results in amnesia of some kind, but amnesia of declarative memory (i.e. the recall of information and events). Often, hippocampal damage still leaves people with the ability to use skills they have learned, and possibly learn new skills (even while not remembering that they learned them). The hippocampus also appears to hold information related to the space around you, and that memory of the space with specific 'place' cells.

The region around the hippocampus is also involved in memory, particularly parts of the neocortex nearby. The neocortex appears to contain information related to the creation of actions. It also processes incoming perceptions. So it seems sensible to assume that the neocortex holds onto the more procedural aspects of a memory, linked to the perceptions that are associated with it.

So we can at the least seperate our memories into three parts: the part responsible for the 'feeling' of a memory—the perceptual and bodily experience, the part responsible for the emotion a memory holds, and the part responsible for the details.

Memory is a mercurial property of the brain. But three aspects seem certain: a 'feeling' part, an emotional part, and a detail part. Which has implications for how we forget.

The problem with these is that they can be 'overwritten' so to speak, and in kind of different ways. The 'feeling' part, based as it is in the neocortex, is fairly reliant on practice to remain stable. Practice in this case would be recall.

Practice affects 'plasticity', or the ability for neurons to change their connections to one another. More practice solidifies connections, making them more stable over time. They become less plastic and can't be used as easily for other things. Without practice, these neurons are more plastic and are thus more changeable. They can be reused for other things.

This means that in the absence of recall, it seems likely that the specific perceptual and bodily aspects of the memory are likely to become increasingly fuzzy. They are also likely to be affected by overlapping perceptions and bodily experiences. If you have a memory involving a room, and the room is somewhere you frequent, the room is likely to remain stable but the memory of a person doing a particular thing in that room is going to slip away.

The details part also likely relies heavily on practice, and thus recall. But in particular the hippocampus has a reciprocal relationship with our verbalisations. Language helps us to 'chunk' memory items together giving us access to more detail, more easily.

For example, I don't need to describe a chair to you. By communicating the word chair, I have already communicated a large chunk of information: an object for sitting, a seat and a back, some kind of legs or foundation, and so on.

As such, as the verbal (or rather lingual) aspects of the memory slip away, the more difficult it is to retain aspects of the details. We no longer have the benefit of this chunking. We must become increasingly reliant on the bodily and perceptual aspects to inform the memory.

The emotional part is less likely to become overwritten, as it is to become bound up with other related emotions. Emotional memory is an incredibly powerful thing and does not appear to be subject to this kind of loss in fidelity. This may be because emotions are not particularly detail-oriented. What emotional memory does have is context-specificity. Putting ourselves in similar contexts is likely to evoke similar emotions. This means that these emotions are likely to become more complicated without deliberate recall—similar events or surrounds will augment emotional memories both by evoking the original emotion, and by tying it to new emotions related to the new and similar circumstances.

In short, without recall, we are likely to lose fidelity in two dimensions. The specific, declarative aspects go leaving us increasingly reliant on the bodily and perceptual aspects which themselves become fuzzy or merge with similar bodily and perceptual memories. What is likely to remain is the emotion. Though not necessarily intact. Rather it is likely to become more complex over time.

The science of remembering

This also has implications for how we remember. When we access a memory, we are now engaging in practice. But what we are practicing is this altered memory object. We first practice the bodily and perceptual aspects, which have become slippery and have coalesced with similar aspects of other memories. We then tie those to lingual objects by describing them, greatly increasing the fidelity of this new chimaera. In the meanwhile, the emotions are probably there and depending on how dramatic the event, possibly in earnest. But under the surface, that emotion has probably spread and may have taken on new characteristics.

What you are doing is not so much remembering as reconstruction.

You don't 'remember', you reconstruct.

This is fairly crucial.

With memories that are uncomplicated, this process is straightforward enough. Remembering your high-school mathematics is going to be fairly easy and quick, particularly since you can compare your memory with a current source of truth as opposed to relying solely on the memory. With memories that are more involved, this process is going to be difficult, time-consuming, and with plenty of margin for error.

Implications for repressed memories

Often the repression that trauma victims engage in is avoidance. They know the memory is there, but don't like to think about it. Over time this becomes a practiced thing, and it's forgotten. That is to say, fewer and fewer things remind them of it. This process is fairly standard remembering and forgetting. Like being reminded you need to go to the dentist. Before reminding, you remembered nothing, but after you can recall the day and time and location and so on. The difference, perhaps, is that you weren't actively avoiding the memory. And also, likely, the emotional content.

The motivated act of forgetting does have some differences to unmotivated forgetting. The most salient is that intentionally suppressing a memory appears to not only effect the memory in question, but also to disrupt related semantic memory. This isn't surprising, given the thesis that memories often end up overlapping with one another. But it does indicate that motivated forgetting may require more effort at reconstruction.

Of course, the act of reconstruction helps greatly to understand the controversy surrounding 'recovered memories'. A therapist with an agenda, no matter how well meaning, could easily introduce incorrect information into the reconstruction process by asking leading questions or by recounting case studies of other trauma survivors in an effort to help the client. Depending on how susceptible the client is, vulnerable as they are in a patient-client situation recalling highly emotional events, this information may well merge into the existing memory traces. Then, in the act of therapy, recalling the events and thus practicing the reconstructed memory, the detail becomes increasingly sharp. But the sharp reconstructed memory has no obligation to bear resemblance to the original event.

Reconstructed memories are under no obligation to be real, but the emotion that's generated certainly is.

An excellent example of this would be the case of Barbara Snow. A notable figure in the 'satanic panic' of the 80's and 90's, Snow appears to have fairly strong beliefs that satanic ritual abuse and military experimentation is a widespread phenomenon, despite evidence to the contrary. Her clients frequently 'recover' memories of events that appear to corroborate this. Unfortunately, Snow also appears to have a highly coercive practice. For example, in one particular case, the police deliberately fed Snow false information that repeatedly appeared in childrens' subsequent testimony. She has also agreed to charges of fabricating notes from her sessions. Highly motivated as she appears to be, it's unsurprising that so many of her clients 'recover' memories of an extremely rare phenomenon. It seems unlikely that Snow is somehow at the center of some vortex of nefarious activity. What seems more likely is that, as in the case of the childrens' testimony, her clients are led to embellish a legitimate trauma. But what is startling is the extent to which they can be embellished.

Unusual forgetting

There are more unusual forms of forgetting. For example fugue states or other kinds of psychogenic amnesia. These are cases in which people completely and suddenly lose the ability to remember some period of time. This is certainly not a standard kind of forgetting. In fact, it's most likely a form of dissociation. Dissociation, particularly following a traumatic event, is not uncommon (although it's certainly not common). But dissociative forms of amnesia are even rarer. Dissociation describes cases where there is an unusual kind of integration of memory, perception, awareness, or identity. In some way, you have detached from your normal reality. This might describe something trivial like day-dreaming and other states of conscious unconsciousness (zoning out). It describes out-of-body experiences. It describes states of emotional numbness. Any kind of non-typical subjective experience.

Dissociative disorders like psychogenic or dissociative amnesia are at the far end of the spectrum along with dissociative identity disorder (multiple personalities). These disorders describe a profound loss of memory regarding some period of time. These are typically found to follow extremely traumatic events. Amnesia due to dissociation is certainly reversible, and in this case would constitute a recovered memory. The difficulty here is that in the absence of evidence, a dissociative event is hard to discern from a false memory event.

It is absolutely possible to implant false memories, to a greater or lesser extent. There are a huge number of kinds of false memories. We can reliably do it in mice and flies. We can less reliably do it in humans, particularly when someone else corroborates the falsehood. It was part of many of the core projects of MKUltra, a CIA program investigating the ways in which one could manipulate the human mind. They're even, and perhaps especially, contagious. We frequently shape each others' memories, just as we do each others' beliefs.

Fugue states and false memories are indistinguishable without evidence.

This again, should not be overly surprising. Imagining events and physically engaging in events look remarkably similar in the brain. As an example, Adrian Owen has built his career on the study of consciousness in vegetative patients. Some of his most remarkable investigations have involved asking coma patients to engage in tasks while in a scanner. Their neural activation looks indistinguishable from that of people actually engaged in those tasks.

The brain appears to use the same architecture to imagine activity as to engage in it. Given the messy nature of memory, it's no great leap to recognise that our imaginings can take root amongst real events.

That said, as mentioned before, false memories are no less problematic than true ones. Except perhaps in that they have had less time to poison other experiences. It is entirely possible, for example, to develop post-traumatic stress disorder in response to false memories.

In addition, false memories and dissociated memories are not likely to operate any differently to other memories in their stickyness. They are subject to the same kinds of erosion, because they're supported by the same mechanisms. This is evident by the fact that dissociative autobiographical memory disturbances are often characterised by fragmented recall. The difference is that their sudden retrieval (or implantation) may result in a more alarming recovery of emotion.

Emotionally charged memories, recall, are often more intense and recalled somewhat more accurately. The same is possibly true of emotionally charged false memories.

Which leads us to the closing point.

The 'feeling' and detail part of memory is a problem. But the emotional part is more useful.

Emotion is the problem and the solution, regardless of the kind of memory

What's important is not so much the precise details of a traumatic event as the emotion associated with the event.

This may not be true in all cases. A person dealing with a specific phobia may find relief from discovering the origin of that phobia was related to a specific detail of a forgotten memory. Or, particularly in cases with young children, it sometimes becomes obvious that a memory of a sexual abuse can't be remembered well because they were actually unable to perceive much—perhaps overwhelmed by the body of the abuser or the surface that the abuse occurred on. These details can be helpful.

But speaking generally, it is the emotion that is the root of the issues arising from the memory. By engaging with the emotion, the impact can be lessened.

A very typical example of this is a child sexual abuse survivor. Adults tend to forget how powerless children are. How responsive they are to figures of authority, and distrustful of their own ideas. Survivors, in adulthood, may have a perception that they voluntarily participated in the abuse. That they are responsible in some way for it. The memory, thus, takes on aspects of shame and guilt which compounds the shame they may have felt after the abuse began. But of course, the volunteering of a child is far different from the volunteering of an adult. It isn't really volunteering at all. The child has no real conception of the long-term consequences of their behaviour, only that someone with power wants them to act. And so they do. By reconceptualising the abuse in this, more realistic, way, survivors might be able to let go of some of the shame and recognise instead the legitimacy of their anger. It has much less to do with the distorted memory—in this case, making the child more adult-like—you don't need to recall the details to know that the child is blameless.

Another example, particularly helpful for children and physical abuse, is narrative therapy. Here, instead of focusing on the panic and terror of the abuse for example, we can highlight the strength of the survivor for making it through. Their narrative about themselves moves from one of victimhood to one of emotional strength and perseverance. Both kinds of reflection are valid, but each emphasises different aspects of the emotional constellation related to the event. Neither needs to hover on the specific details of the event.

Or, finally, the simple act of imagining of your current self returning to the abused self to support them, imbuing the memory with a new set of more positive emotions.

Each of these methods are not at the core about 'recovery' of the details, but the transformation of the feeling. And that is something that the brain is very good at.

Perhaps if therapy was less apt to be carried away by the currents of 'repressed memory' hype, and focused more on the transformational properties of the brain, Wigmore's Treatise would have told us something quite different about the believability of the vulnerable.

N.B. See update on the value of focusing on traumatic details.

Articles | Collections | Email me

Turning scholarship into wisdom we can use at The Armchair Collective.

There are over 2000 of us

Keep up. Emails two or maybe three times a month, with the latest articles on the site and possibly a short note on something topical.


Doesn't work? This signup form can count as a 'tracker' and your addons might be blocking it. This one will look like "armchr.us19.list-manage.com" or similar. Try allowing it, while you join.