Learning vs. Memory

Does how we talk about learning and memory effect how we think about consciousness?

Imagine you’re a rat in an experiment. It’s an experiment to test your memory, so you’ve been trained to associate a sound with an electric shock. Now every time you hear the sound, you freeze in fear. Then the next time they play the sound you’re given a drug just before to test its effects on how your memories are formed and recalled. We can observe that this drug causes you to no longer freeze when you hear the sound, but does that mean that the drug has erased a memory?

As a rat you can’t describe the experience, but think of the many problems that would exist even if you could talk. We can really only describe a certain kind of memory, those that we’re conscious of. The specific causes of any unconscious behaviors are as opaque to the subject as they are to the researcher. If the conditioned fear response was caused by an unconscious behavior then you might not have any additional insight into what was happening. Even if it was caused by a conscious behavior,would we have the ability to describe its loss? How would we describe the difference between having a memory that’s not consciously available for some reason and having a memory which was always unconscious?

The language we use to describe any kind of recollection can be vague and inconsistent. For example, is memorizing the Pythagorean theorem the same thing as learning the Pythagorean theorem? Not everyone would agree that those words describe the same process, especially since it makes sense to say that we can “learn to ride a bike,” but not that we can “memorize how to ride a bike.” Sometimes we can use the word “memorize” to mean the same thing as the word “learn” and sometimes we can’t. Consider something as seemingly simple as recalling a mathematical fact. If I work out what “42 x 77” is I can probably recall the answer for several days, give the correct answer if asked, and explain how I “remembered” the answer. But if I’m asked what “4 x 7” is, I don’t have to think about the answer, and yet I can’t explain how I “remembered” it. Even though the kind of information is the same, should we assume that the same underlying process allows me to recall both facts?

I’d argue that the first step towards a better understanding of memory is to explicitly assume that there are at least two, and possibly more, different ways that the brain can “store” information. Often in both general and scientific writing there’s the assumption that all memory, learning, conditioning, etc. happen through the same underlying process. Sometimes we’ll even see it described like “all memories are caused by changes in the wiring of neurons.” We know this is true in some cases. Since Hebbian theory (“neurons that fire together wire together”) was introduced, we have observed experimentally that neurons can change their connections to allow for learning, either through changes in potentiation, or changes in synapses. But do we know enough to confidently say that that’s the only process that allows for any kind of recall? We certainly have enough evidence to say that that’s one way the brain can recall information, but should we assume then that that’s the only way?

What do I mean by different kinds of recall? Let’s take the famous case of H.M., a subject that had his Hippocampi removed to treat his epilepsy. He subsequently lost the ability to form new memories but retained the ability to learn procedural tasks. Should we think of this as him losing some of the ways to form memories, or should we think of memory and learning as separate processes and say that he lost one of them completely? Why should we assume that learning is a similar process to memory, but just one that’s unconscious? If instead we assume “memory” to be a conscious process then there seems to be a somewhat natural distinction between the kinds of language that can be described as “learned” or “conditioned” and those that are described as “memory” or “remembered.” We may be able to define the boundaries of two groups like this:

The two key characteristics of the process that we usually describe as “memory” or “remembering” are:

  1. We can only remember something if we’re conscious of it, and can potentially remember anything we’re conscious of
  2. We are always consciously aware of recalled memories


And we can compare those types of experiences with the ways that we usually describe the kinds of processes usually described as “learning”:

  1. We can learn things without being conscious of them
  2. We can do or recall things we’ve learned without thinking about them, or even being aware of them


We could potentially try to pick apart these two groups into even more sub-groups, but I think that the key point to be made is that we should assume that there could be more than one process that allows the recall of information, and that there’s a natural division between conscious and unconscious recall. When we think about all the different things that we learn or remember it’s clear that these two kinds of processes must interact or overlap often, but considering them as two separate processes provides a useful perspective on this kind of complexity. I’d argue that this perspective is useful even if this definition just gives us a target to disprove. For example, the biggest difference might be that they might happen in different parts of the brain, or over different time scales. But until we make a discovery like that, we should leave open the possibility that they’re created by completely different physical processes instead of assuming a single process.

If this division is accurate, it would still be true that most of our behavior and thoughts are affected by a complex interaction between these two processes, but if they are separate we should expect to at least occasionally find times when one or the other dominates. What would that look like, to be able to form conscious memories, but not to be able to recall or create new learned behaviors? That description sounds like dreaming - it appears to be composed of conscious experiences that we can remember, but is usually experienced without learned behaviors. So, we might put dreaming into the “memory” category and ask what would the “opposite” of dreaming be then, something that is just learned behaviors without experiences? This is probably a less common behavior, but sleepwalking might fall within that description. Sleepwalkers can sometimes engage in complex learned behaviors (walking, cooking, even driving, etc.) but can’t remember their actions, and seem incapable of the kinds of more complex behaviors or creativity that consciousness allows.

Considering memory and learning to be different processes allows us to ask different kinds of questions. For example, if consciousness is necessary for memory, what else is required for memory? Depending on how much or how little extra is required, does it make sense to consider consciousness and memory as the same process? What would it mean to be conscious without any kind of memory? And memory, if we’re considering it as separate from learning, without consciousness might not even make sense.

I’ll conclude by considering the spectrum of how much we seem to know about the three processes we’ve looked at - learning, memory and consciousness. I think most people would say that, relatively speaking, we know quite a bit about learning since we’ve had the experience of extensive animal experimentation to discover what’s possible and what isn’t. Memories are more complicated because when we try to understand them, we usually rely on subjects describing their memories, which means we’ve been almost entirely limited to human subjects. Undoubtedly, consciousness is the least understood. It currently lies way out by itself at the end of the spectrum of what we’re trying to understand.

I hope that considering learning and memory to be separate changes the way we think about what’s possible. Learning could land well within the realm of what we can understand through research into mapping and modeling physical neurons. However, once separated from learning, memory becomes somewhat more difficult to understand because we’ve lost that connection to an underlying physical process. But this also moves memory out along the spectrum of understanding towards consciousness, and perhaps it even mostly or entirely overlaps with consciousness. Now perhaps consciousness doesn’t seem so far removed from things that we understand. Memory might form the bridge between physical brain processes that we understand and can model and predict and the experience of consciousness that might otherwise seem too unlike anything else we understand to be a possible subject of research.