E. L. Thorndike and the 'Law of effect'

Classical conditioning is the most simple form of what's called associative learning. It's possible to imagine that associative learning is learning to associate one thing with another.

In the case of classical conditioning, like in the Pavlov studies, dogs would learn to associate a tone with food. When that tone reliably predicts the food, you will get your conditioned responses to the tone.

Now operant conditioning is a form of learning where one learns to change behaviour in order to achieve favourable outcomes. This is not some reflexive behaviour that's automatically elicited by something. This is a volitional behaviour. This is a purposefully produced behaviour.

In the case of operant conditioning, the operant is a behaviour that operates on our environment to achieve some outcome. Operant isn't a word we normally use in everyday life, but in this case it makes a lot of sense.

It is a behaviour that operates on the environment and therefore it's operant conditioning, because of what is being conditioned is the behaviour.

While classical conditioning can explain simple responses to the environment, like salivation to the sound of a tone or in regular, everyday life, you might see a food that you like and start to salivate. Those are examples of classical conditioning. Operant conditioning, on the other hand, can explain much more complex behaviour.

The work of E. L. Thorndike

Operant conditioning started with E. L. Thorndike and what he called the 'Law of effect'. In the law of effect, there's a focus on the importance of the consequences of behaviour on learning. For example, with the Pavlov studies in classical conditioning, the salivation really didn't matter.

The dogs didn't get anything for salivating. With the law of effect and with operant conditioning, behaviour matters.

As Thorndike put it, if an association leads to a 'satisfying state of affairs', that response is strengthened. Whereas if it leads to an 'annoying state of affairs', that response is weakened.

This is really the beginnings of what's called instrumental or operant conditioning. The terms instrumental and operant can sometimes be used interchangeably, but they essentially mean the same thing.

The puzzle box

The way Thorndike studied operant conditioning was to put cats in a puzzle box. A puzzle box is a box in which there is a lever, and if the cats hit the lever, they get to escape and enjoy the food that's waiting for them outside the box.

The behaviour shown by the cats during the first trial compared to the final trial are quite different. Thorndike found that at the beginning the cats didn't like being in the box. They'd thrash around and would eventually accidentally hit the lever, escape, and then get the food.

Over a number of trials, the cats would get better and better at doing this and they escaped in less and less time.

Had the cats learned that the lever that opened the door that let them escape the box and get food just trial and error with no sudden realisation of how to escape?

If the cats had some sudden insight that the lever is the thing that they had to hit to escape the box, they would've hit it immediately every time and then escape.

Thorndike found no evidence of insight, and that the cats learned to escape the box was gradual over a number or trials, incrementally through a process of trial and error.