The experimental study of animal learning by E. L. Thorndike (1874-1949) in the United States and his theory on trial-and-error learning provided the impetus for Skinner's experiments on instrumental or operant conditioning. Thorndike's doctoral research on 'Animal Intelligence' in 1898 provided the psychological world the first miniature system of learning known as trial-and-error learning. His theory left a profound effect on American psychology then. It also continues to exercise its influence on contemporary psychological theorizing.
Thorndike's research was indirectly influenced by Darwin's theory of evolution. Darwin demonstrated that there is a continuity in the bodily structures of many different species. This evidence favored Darwin's doctrine of evolution. What about continuity in the ability to think and reason? Can animals think, understand, and reason like human beings, although at a simpler level?
The critics of Darwin argued that the essential difference between humans and beasts is that humans can think and reason, which animals are not capable of doing. Thorndike's research on animals (cats, dogs, fishes, chicks, and moneys) showed that learning is a matter of connecting responses to stimuli in a very mechanical way. There is no involvement of consciousness, thinking, reasoning or understanding. The animal performs responses mechanically. The responses that bring reward are learned; the responses that do not bring reward are not learned. The animal does not show ability to understand, think, and reason. The animal learns mechanically through trial-and-error.
Indeed many forms of human learning, particularly the learning of sensory- motor skills, are achieved through trial-and-error. Learning to walk, to swim, or to ride a bicycle is based on trial-and-error. At the beginning, we make wrong movements and commit errors. As we go through a series of practice trials, errors are reduced and responses are mastered. The gradual reduction of errors over trials gives the name, trial-and-error form of learning.
Thorndike's Experiments on Cats:
Thorndike experimented on a variety of animals like cats, fishes, chicks and monkeys. His classic experiment used a hungry cat as the subject, a piece of fish as the reward, and a puzzle box as the instrument for studying trial-and-error learning.
In this typical experiment, a hungry cat was placed inside the puzzle box, and a piece of fish was kept outside the box. The cat could not reach the fish unless it opened the door. In order to escape from the box, the cat had to perform a simple action as required by the experimenter. The cat had to pull a loop or press a lever in order to open the door. Once the door was opened, the cat could escape and eat the fish.
What did the hungry cat do inside Thorndike's puzzle box? Initially it made random movements and ineffective responses. On the first trial, the cat struggled valiantly; it clawed at the bars, it bit; it thrust its paws out through any opening; it squeezed itself through the bars; it struck out in all directions. All the irrelevant responses continued for several minutes until the cat hit upon the correct response, by chance.
Accidentally, it pulled the loop and the door opened. The cat came out of the box and was allowed to take a small part of the fish. It was then put inside the puzzle box for the second trial.
In the second trial, the time taken to pull the loop reduced a bit. Every time the cat came out of the box and took a piece of fish, Thorndike put the cat inside the box again. Thorndike and the cat kept up this exercise for a while. With increasing trials, the time taken to pull the loop (response latency) decreased. The wrong responses (errors) that the cat was showing also decreased, as trials increased. Finally, the cat learned the trick. As soon as it was put in the box, it pulled the loop to escape for a well-deserved reward. The name, trial-and-error learning comes from the fact that errors decreased over trials. The cat learned from its errors.
How did the animal learn? To answer this, Thorndike plotted the time taken on each trial by the cat to show the correct response (i.e., pulling the loop). The plot indicated that there was a gradual decline in the response latency. If the animal would have shown some understanding of the requirements to reach the fish, the curve should have registered a sudden drop at some point. This did not happen. The declining nature of the curve suggested that the animal had no understanding of the situation; it was only performing some responses, one of which was getting mechanically connected with the stimulus situation. Thorndike concluded that animals do not learn through thinking, understanding and reasoning. This view also received a second line of support, when Thorndike failed to teach cats to pull the loop for opening the door. He held cat's paw over the loop, pulling it for them, if cats had understanding, they should find their way out in the box, particularly after Thorndike had taught them the method. It means that the animal cannot learn without acting, it has to make its responses to the situation. The findings suggest that the cat did not have understanding of the solution. Thorndike explained cat's learning by the 'Law of Effect'.
Thorndike conducted similar experiments with other animals and obtained similar results. He said that the animal does not learn a new response; it only Thorndike's puzzle box were in animal's stock of responses. Only one response led to animal's satisfaction of obtaining a piece of fish. As a result, this response was selected from the stock automatically. The connection between this response and the stimulus situation got strengthened over trials. Very simply, the 'Law of Effect' derives its name from the fact that whether a response would be strengthened or weakened depends upon the effect of the response.