Skip to content Skip to sidebar Skip to footer

Cognitive Maps and Latent Learning Psychology Definition

Hi, Welcome to Falhatla Riservisi. In the previous article on observational learning, we know that learning could occur without any direct reinforcement or punishment. The children could learn about behavior simply by observing the adults. This suggests that there's some sort of cognitive process going on while the children are observing.

There's learning happening even though it's not being demonstrated yet. So this brings us to this idea of latent learning and this was researched by a guy named Edward Tolman. So I want to look at two studies that Tolman conducted that demonstrate this cognitive process of learning. 

Cognitive Maps and Latent Learning Psychology Definition
image source:

The first study that we'll look at was conducted by Tolman and Charles Honzik and what they did was had three groups of rats and put these rats into a maze. They ran the rats in this maze every day for 17 days. So what were the three groups of rats that they had?

The first group got a reward for getting to a certain point in the maze. They had sort of an end goal to get to.

The second group didn't get a reward. So as far as these rats were concerned there was no endpoint to the maze, they just wandered around in the maze because there was no point where they got a reward.

The third group of rats was sort of a combination of these two. They didn't get a reward for the first 10 days and starting on the 11th day they started getting a reward. So the third group has no reward for the first 10 days and then they started getting the reward on the 11th day. Ok, so if we look at the behavior of these rats what do we see? Here are the days that they're performing in each maze and this is going to show the number of errors that they made, how many wrong turns did they make in the maze before they get to this endpoint? OK, let's look at the first group. 

The first group of rats is getting a reward, so initially, they make lots of errors, but over time they make fewer and fewer errors. They get better at doing the maze because they're being reinforced to get to this particular point so they learn. How about the second group? 

The second group is not getting a reward, so so they don't really know what the end of the maze is. They just kind of wander around, as I said, so they're just going to be wandering the whole time. This follows the very standard behaviorist explanation of how learning occurs. You get reinforced for something and you get better at it; you learn to change your behavior. That's essentially what we see here. But the interesting part comes with the third group. 

What happens with the third group? Just like the others, when they first are put in the maze, they're just wandering around, and just like the second group, they're just wandering around but then after 10 days, they start learning. You might think that they would gradually decrease just like the first group. But what Tolman and Honzik found was this group of rats suddenly started performing, by the end of the study, even better than the group that was rewarded all along.

What does this mean? This demonstrates that in this third group of rats, throughout this 10-day process here, we're actually learning. They just didn't have a reason to demonstrate that learning. There was something going on that when suddenly you start putting a reward into the maze, they start getting to it very quickly. That shows that they actually have been learning something about the maze all along. This is this idea of latent learning. Latent learning refers to learning that occurs without reinforcement. When the rats are just wandering around in this maze there with no reinforcement, they're actually still learning about the maze.

The reason that early behaviorists wouldn't think that they were learning is the rats didn't have any reason to demonstrate that. By this idea we could say this second group of rats here got no reward, they're actually learning during those first 10 days too but they never got the opportunity to show it. So the idea of latent learning is that learning occurs without reinforcement but it's not demonstrated until there's a reason to do so. Just because learning hasn't been demonstrated doesn't mean that hasn't occurred.

We actually operate on this assumption all the time when you think about when you sit in school, you sit in class, you sit for a lecture or something, we assume that you're learning something by sitting there, but there's no observable change in your behavior. We hope that the learning is occurring latently, that you're sitting there you actually are learning, you're just not going to demonstrate that learning until the test day arrives or we give you some challenge and now you're able to show that you have actually learned something from sitting in that lecture. But when you first were learning it there was no incentive to demonstrate it.

That's quite similar to the idea of these rats here where they're learning about the maze but they have no reason to show you. They're not going to need to rush to the end of the maze because they're no reward there. Then you start putting a reward there and suddenly they demonstrate that they actually do know this maze pretty well. Ok, I want to look at one other study by Tolman and this was conducted with Ritchie and Kalish in 1946. What they did in this study was they had a very simple maze that they put some rats in. 

So the maze had an entry point into this round room here and then there was one exit and then the corridor turned to the left and then it turned to the right and then it turned to the right again and then at the end of his long hallway was the food reward. So they had rats complete this very simple maze. Here's our rat here, he's going to run in here, go straight through turn left, turn right, turn right, and get to the food reward. Then after the rats had learned the layout of this maze, the researchers changed it a bit. What they did was, the same entrance, the same round room but now instead of just one corridor, there were many different corridors branching off in all directions. 

The straight-ahead version the rat had previously been taught was now blocked. So the question is, which of these hallways will the rat run down? One of the ones angling off to the right here actually went out to the food. So the food was actually in the same general location that it was before. But now the rat's going to have to choose a new route to get there. There were actually more of these corridors than I've drawn here. I'll post a link to the original paper, you can see a diagram of what this actually looked like. This is a very rough approximation. So the question is, what would rat do?

The traditional behaviorist explanation would be, ok the rat runs in, tries to go straight, that doesn't work. OK maybe stimulus generalization will occur, and the rat will choose the most similar behavior. Or they say maybe the first turn that was reinforced was turning left so maybe the rat will get here, can't go straight, it will turn left it will try going this way. That's not what happened. What they found in this study was the rat actually chose the correct corridor. This showed the idea of a cognitive map. 

What the cognitive map refers to is the idea that the rat actually has a notion about where the food is in relation to the starting point. The rat is not just learning, how do I get food, turn left, turn right, turn right, and then that's where food is. The rat is actually learning, ok, I'm here and food is over there somewhere so I should try to get over there and if I suddenly have a new path that goes that way then that's the one that I should choose. Instead of thinking of behavior as simply being reinforced rote learning of turn left turn right. So that's the idea of this cognitive map.

This shows that the rat must have some internal mental representation of the rat's location and the food's location in relation to the rat. That's what this study demonstrates, it shows it's not a strict behavioral process of this turning left or turning right was reinforced. But rather than the rat has actually learned about the layout of this maze here and it knows where the food is. This is true for humans as well. If you've ever gone to a restaurant, let's say you're waiting for a table at a restaurant and you need to use the bathroom. So you ask someone "excuse me, where's the bathroom?".

They tell you, "go down this hallway, turn left, turn right and the bathroom is there". Let's imagine later on in the evening you're sitting at your table in a completely different part of the restaurant and you need to use the bathroom again. You don't follow those same instructions, right? You don't say "OK how do I go to the bathroom, I turned left, I turned right" because you know that's not going to work anymore. Instead, you've formed a cognitive map, you have an idea of where is the bathroom in relation to me. This learning has occurred latently, you didn't have a reason to demonstrate this until later but it shows that you do have this cognitive map. You have an overall understanding of location not a specific reinforcement of the actual steps of the behavior.