Theories of learning (Pavlov and B.F Skinner)

  THEORIES OF LEARNING

Ivan Pavlov:

A Russian psychologist named Ivan Pavlov and his dog. 

Pavlov was studying the salivatory response in his dog and he would feed it meat and notice how much saliva was being generated by the dog and then he made an astonishing discovery. The dog had started to salivate not just to meat but to the sound of the bell.

Why was the dog salivating to the ringer? Because the bell was being rung by the meat delivery man, so every time the meat delivery man came, he rang a bell and then the meat was delivered and upon looking at the meat the dog would salivate. Now this made Pavlov think our mind is making some kind of connection between the bell and the meat because the bell alone was not enough to generate salivation.

So let rephrase what was happening, initially when the bell was rung before any meat delivery the dog would never salivate, the dog would only salivate when it would see the meat, but with time as the bell meat combination was repeated the dog would start to salivate to the sound of the bell.

Pavlov said, what would happen if every time the bell was rung you would switch on a particular light and then feed the dog. Would the dog learn to salivate to that particular light of course it would. And so, this idea that you could actually teach the dog to salivate, to the sound of the bell, to the light, to the clap of the hand and so on. 

Classical Conditioning:

Revolutionized psychology this principle is called classical conditioning.

And in classical conditioning what happens is that the animal reacts not just as a reflex as it did to the meat but also to something paired with that meat such as the bell or the light, so in this situation we would say that the external stimulus which was the meat was paired through association with a stimulus that you could actually condition. So, we have the unconditioned stimulus which is the meat where you don't need any conditioning the dog would automatically salivate looking at the meat and it could be paired with the conditioned stimulus which could be the bell, or the light, or the clap of a hand, or it could be anything. We as a whole realize that dogs can be trained, and these principles are followed in training of dogs.

So does likewise happen with the humans somewhat we can say yes however, most human behavior is not the kind of behavior that we see in dogs. 

For instance, salivating to the meat is a reflex action which happens in response to meet. How does that explain are going to school and learning times tables or memorizing nursery rhymes or there's so much else going on in human behavior that cannot just be explained by association between unconditioned stimulus and conditioned response in the form of classical conditioning.

So, it took a bunch of American psychologists to think of a theory that explained a lot of active behavior that human beings engage in. Unfortunately, we cannot subject human beings to experiments like the ones conducted by Pavlov on his dog.

B.F Skinner:

B.F Skinner and other psychologists had a simple solution, they argued that human beings are nothing more than animals they may be advanced animals, but they did not believe that human beings have any special qualities and therefore they thought that if we can explain learning in humans through experiments on rats these principles would be equally valid in both, here is something that led to a very controversial debate in the history of psychology.

Skinner as you can see in the diagram created a box in which they placed a lever, and that lever would be pressed by a rat why would the rat press that lever, well rats are very, very curious animals just like human beings and they're also very active. So, if you put them in a box, they're not just going to sit there they're going to do something so they look at a lever and they're going to press it, but they're not going to keep pressing it they get bored very quickly press it once it moves fine, press it twice great, but then they would lose all interest.

So, what skinner did was he connected the lever in the box to a food pellet delivery system. So, what would happen is that the rat would press the lever and as soon as the lever is pressed a small food pellet would be released into the box rat would eat it and then press the lever again another food pellet would be released the rat would eat it and then press the lever again.

Now skinner made a whole bunch of discoveries following this behavior pattern first of all as you can see it is a very different paradigm from that of Pavlov in parallel's experiments the dog did not really do anything the salivation was automatic, so the pairing of the meat and the bell didn't have anything to do with deliberate activity with respect to the dog. Whereas in this case the rat is actively doing something first as a response he's getting a reward which is the food palette and repetition of the reward is leading to pressing of the lever over and over again.

In other words, skinner believed the rat had learned to press the lever and he did a whole range of variations to this paradigm.

For instance, in some cases the system was rigged so that only if a red light came on and the lever was pressed the food pellet would be released but if the red light did not come on and the rat pressed the lever what would happen no food pellet.

So, what do we understand from that, the rat could learn to associate the red light the delivery of the reward and then you could create a further relationship, so before the light came on, you could press a bell and the sound of the bell, if it is followed by the red light and only if that red light is followed by the pressing of the lever and then pressing on the lever is followed by the food pellet then the rat would also learn to respond to that sequence. 

Then Skinner thought human beings don't always get a reward. So, for instance, people work for a month. furthermore, toward the month's end, they get a compensation.

So, skinner thought could I make a schedule that after every 10 presses, or every after every 20 presses of the lever, or after every 30 presses of the lever, or after every 50 presses of the lever, the food pellet is generated. A schedule of reinforcement either based on the number of times the bar was pressed or based on a time interval was utilized and viewed as extremely successful.

We know that human beings can be paid in a number of ways for their work one is to pay them in response to the number of items for instance so somebody who produces a thousand pieces of a product can be paid in response to the number of the products generated or somebody else could be paid just by time as in regular employment.

So, he found that paying by item or giving the reward per item was actually more effective and more efficient in creating learning than a time-based schedule, but it did generate a very interesting finding most rats would perform really, really well at the point when it approached the hour of the reward and after the reward was paid the frequency was still very high. But then it would dwindle it would go down and the performance would once again pick up. When the payday came near when the reward time came near so their internal biological clock was synchronized to the payment of the reward.

Skinner was making very loud claims human beings learn in exactly the same way as rats do and he tried to explain every single bit of human learning by his theory of operant conditioning or instrumental conditioning.

Operant conditioning:

Operant conditioning or instrumental conditioning basically means that you act on the environment, the rat acts on the environment, the human being acts on the environment, then the stimulus or the reward is given.

Aspect of learning:

But skinner also looked at another very important aspect of learning reward he had established and this age-old principle that he looked at was punishment.

Skinner wanted to see if punishment would suppress undesirable behavior just the way rewards increase desirable behavior. He found that rats that were shocked for pressing the lever would stop pressing the lever, they would also stop doing pretty much anything and everything this was quite shocking pun intended because he was delivering electric shocks every time that he had pressed a bar rat was given an electric shock automatically. Now he found another interesting thing there if the rat hadn't learned to press the bar previously it was fairly simple. Every time the rat pressed a bar you would deliver an electric shock of extreme intensity the rat after a few presses would just stop but what if you had actually trained the rat to press the bar by giving him food pellets when it was hungry what would happen then if you punished them, they wouldn't stop you'd keep punishing them they would keep coming back for the liver because they had learned that pressing the lever gave them food.

Now we can say in common human language maybe they were confused we don't know but they would still come back for more and press the lever even if there was no food pellet and even if there was punishment so did the punishment ever work yes it did. The punishment had to be severe, and it had to be delivered every single time.

Now contrast it with reward or reinforcement the reward could be given after a number of bars presses it didn't have to be given on every single bar press and still be very effective punishment seemed to be only effective in the cases of learned behavior it only seemed to be effective if it was very severe and if it was given every single time or in a very consistent manner.

Disadvantage:

Now if you think about human beings, we know that usually the kind of punishment we give is not very effective children if they have learned what we call bad behavior do not give up bad behavior in response to punishment because as Kenneth said the punishment had to be very severe can we as human beings afford to give severe punishments like high intensity electric shocks to our children can we beat them to a pulp some parents do is it humane.

Skinner argued that it's much more efficient to try and replace one behavior by another by making them learn something different and something new. 

So, if a child is biting the nail.

Example:

An example skinner would say instead of slapping their hands every time like some parents do or putting some chemical on their fingers so that if they you know try to bite their nails the taste or the smell is horrible instead of using punishment if we give them toys to play with every time, we see them going for their nails. It does not occur to most parents that what is be what they are trying to solve with punishment can actually be done by replacing that behavior.