Is a memory a bunch of atoms? And does this mean we can transfer exact memories?

I was asked the following question on Quora.

Are specific memories just arrangements of atoms in our brains? Could you put certain molecules in someones head and give them an exact memory that you had?

Short answer: No.

Modern science has shown that every thing is an arrangement of atoms: neurons, apples, tables, rockets, asteroids, aardvarks… they are all made up of atoms.

The question now is this: is a memory a thing?

Neuroscience has shown that memory is correlated with changes in the arrangements of atoms in neurons — this changeability is called neuroplasticity. But the correlation between memory and structural change does not mean that memories are the same as the underlying neural structures.

It may make more sense to think of memories as reactivated patterns of activity. Changing the structure of a neural network changes the kinds of patterns that can be reactivated. Neuroscience suggests that a memory is a process, not a thing.

Complex human memories involve changes all over the brain. There is no way we can make exact copies of these changes in someone else’s brain. When looked at carefully, no two people have the same brain structure — each human is unique. So this is not a technical hurdle — when two brains are structurally different, the idea of changing them in the exact same way is meaningless.

The only memories that can be exact are symbolic memories — words, numbers, mathematical formulas, and so on. These can be tested against objective reality. So you can tell someone your phone number, and then if they practice it over and over again, their memory of the phone number will be exactly the same as yours. This kind of memory is sometimes called semantic memory.

But there is no way you can convey your memory of your first day of school, or what you did on New Year’s Eve. These kinds of autobiographical memories — called episodic memoriesare not fully symbolic, and therefore cannot be perfectly represented on paper (or on a stone tablet or a piece of clay or a canvas or a computer screen). You can write a story about your experiences, and someone else can, with effort, memorize it. But what they have memorized is the story, not the remembered experience itself.

Even if you had direct control over someone else’s synaptic plasticity, you would not be able to transmit a perfect memory, because you will never know if what they experience is the same as what you experience.

This is a fundamental, hard limit of science as we know it — the only experience anyone has access to is their own. This means we never fully convey experiences to each other — we use symbolic methods such as words to communicate some approximation of the experience. So words, gestures, art, music and so on are the ‘programs’ by which we try to simulate experiences in other people’s ‘mental computers’. We can never be sure how closely another person’s mental simulations resemble our own, since we can’t directly compare them.

What we actually compare are the outward and observable manifestations of experience. This is why testing semantic memories is easy and uncontroversial: we can directly compare our memories of a phone number by looking at how we reproduce it externally using speech, handwriting or typing. I have no idea how you feel about my phone number, but I can convince myself that you know it.

Such ‘knowing’ is impossible to test in more complicated situations. If you say you like chocolate as much as I do, then I’ll assess this by examining your outward reactions to eating chocolate — both verbal and behavioral. If you always grimace and gag when you eat chocolate, then I will not be able to believe you.

Now that I think of it, believing that you and someone else share the same experience or memory might require some baseline empathy.

Empathy may well involve similarity at the level of arrangements of atoms. And science and technology can sometimes help us understand and spread empathy. But I think getting to that state requires something more.


For more on human memory, see this essay of mine:

Why human memory is not a bit like a computer’s


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s