Learning:
99% of human behavior

Edward Thorndike (1874-1949)


History
Thorndike was the son of a minister, not that this information has any great significance to it, other than that children of ministers and pastors during this period were more likely to get a higher education.

Thorndike began his career at Weslyan College studying English. He then moved to Harvard University where he began his studies in psychology. Interestingly, Thorndike raised chickens in the basement of William James' house and used them to study animal intelligence.

He later transferred to Columbia University where he earned his Ph.D. in psychology in 1898. His thesis on animal intelligence was published that same year when he was only 24. Still a classic, Thorndike's thesis showed that cats learn through a gradual process of trail and error. This trial and error leads to the "stamping in" of correct responses. This led to much of his later work, and to the work of others, looking at the function of reward and punishment on learning.

Thorndike published more than 78 books and at least 400 articles before his death. The writings varied in their topics from psychology to education. Thorndike has been attributed to defining and establishing educational psychology; changing the study of child development into an objective science; establihing the use of tests and statistical methods in psychology and education; and he was behind much of the psychological testing movement.

The Puzzle Box
Thorndike developed the classic experiment which involved what he called puzzle boxes. A typial puzzle box consisted of ropes levers and latches that an animal (typcially a cat) could use as a means for escape. The cat was locked in the puzzle box and enticed to escape using food that was placed just out of reach from the box. Thorndike then observed the cats attempts to gain access to the food. Cats being what they are attempted many strategies for escape including trying to squeeze through the bars of the box or meowing incessantly for help.

After determinging that none of these strategies would set him free, Thorndike began to notice in the cat, the typcial trial and error behavior that led to eventual freedom and reward. The cats learned quickly (typically within 3 minutes) how to escape from the box, but their first success was generally "accidental". Successive trials however showed that they took less and less time to escape.

Thorndike concluded that cats do not learn by developing insight into a problem, much like the chimpanzees in Kohler's (1927) experiments, but learn through trial and error. Thorndike also believed this was how humans also learned.

A Theory of Reinforcement or Is it Contiguity?
Associative theories of learning suggest there are typcially two reasons associations are made. One: learning occurs because of the consequences of behavior. In other words, what is the result of the behavior? Reinforcement theories say that if a behavior is followed by a pleasing consequence that behavior is repeated. If a behavior is followed by an unpleasing consequence that behavior is stopped. Reinforcement theories of learning and behavior get more complicated but you get the general idea.

The second type of learning theory suggests that the presentation of two or more stimuli, both spatially and temporally effects any association made between stimuli. The closer in time and space two stimuli are presented the more likely they are to be associated.

Thorndike believe, however, that these explanation were only the tip of the iceberg.

Pre-1930 Laws or Thorndikes First Theories of Learning
Thorndike believed learning involved forming bonds between stimuli and responses. For him, these bonds meant neuronal connections within the brain, hence the Connetionism Theory of Learning. Learning is the process of stamping in of Stimulus-Response connections and forgetting is the stamping out of these connections.

Three variables (frequency, recency and contiguity) combine in a single law to form the basis of his theory.

The Law of Exercise
Bonds between stimuli and responses are strengthened through frequent exercise, done recently and vigoursly. While this law died out and was not very widely known, it was influential in education early on because of it's scientific reinforcement of the notion that practice and repition improved learning. As a result "drilling" became very popular in educational institutions in the 1930s and 40s.

The Law of Effect
His more famous Law of Effect states simply that responses that are made just prior to a pleasant event is more likely to be repeated, while responses that are made just prior to unpleasant events are more likely to diminish. He called these pleasant events satisfiers and unpleasant events annoyers, and they play a critical part to learning.

Thorndike's Law of Effect defines what has become known as Instrumental Learning - that is learning behaviors that are instrumental in maintining satisfying events. An organism will perform a response and establish a connection between that response and the stimulus that preceded it if the consequence that followed was satisfying. It is important to note that this thoery suggested a connection between the stimulus and the response and not the response and the reward.

The Law of Readiness
The third portion his pre-1930 theory involves the state of the organism, or it's motivation to learn. Thorndike explained with this law the fact that some behaviors are more likely to be learned than others. While the law itself was vague much has been done to interpret the readiness to apply to the maturity level of the organism or the effects of previous learning on present learning.

Auxillary Laws
While these are the main laws governing his thoery, there are five ancillary laws that form part of Thorndike's explanation of learning: 

Multiple Responses. In any given situation, the organism will respond in a variety of ways if it's first response does not immediately lead to the satisfyer.

Set or Attitude. This law refers to the predisposition of acting in certain ways that effect our behavior. We may be predisposed to act aggressively toward aggressive behavior, for example.

Prepotency of Elements. Thorndike believed that we learn to react to only significant aspects or elements of a problem and ignore irrelevant aspects in learning. For example, a child learning to identify triangles need only learn the relationship between the sides of the figure, the colour of the triangles is irrelevant for this learning.

Response by Analogy. Thorndike also postulated by this fourth law that a person may learn in new situations by the resemblence it may have to prior experience. This law also is called the law of transfer or the theory of identical elements.

Associative Shifting. Lastly, the law of associative shifting suggests that organisms may use similar responses from one stimulus to another. This is close to the stimulus substitution theory which says that one stimulus may come to represent another stimulus so will elicit the same response from an organism. Advertisers use stimulus substitution when using sexy models to show cars.

Post-1930 Theories
Thorndike is perhaps one of several psychologists to have lived long enough to make revisions on his own theory throughout his life. The other notable aspect to this fact is that he himself admitted to being wrong about some things.

Repeal of Law of Exercise The first thing he said he was wrong about was his law of exercise. He showed that on experiments with humans (NOT cats or chicks) that repitition does not cause learning.

"The repetition of a situation may change a man as little as the repetition of a message over a wire changes the wire. In and of itself, it may teach him as little as the messages teaches the switchboard . . . The more frequent connections are not selected by their greater frequency" (p.14).

Half a Law of Effect Thorndike at this point suggested that what does lead to learning is not the repetition, but actually the effect of the action. This was getting back to Thorndike's original Law of Effect that said actions that lead to "satisfying" events tend to be stamped into learning, and maintained. However, the other hald of the Law of Effect suggested that those actions that led to "annoying" states of affairs were stamped out. Thorndike resinded on this point, explaining that annoying outcomes do relatively little to the strenght of a connection.

Learning by Ideas Several other changes were made to Thorndike's Theories after 1930, mostly by himself. These changes were brought about by observing human learning that did not fit previous theories. His revised theores concentrated more on what seemed to be "cognitive concerns" - or the idea that thoughts or ideas are important to human learning.

Thorndike then turned to what he called "ideational learning - a higher form of learning that involved analysis, abstraction, and meaninfulness.

"Learning by ideas, as the name implies, characterized by the frequent presence of ideas as situations or as responses or as both. Whereas the bulk of learning which dogs and cats and chicks and rats display consists of connections leading from external or perceptual situations straight to bodily acts or to impulsive tendencies closely attached to such acts, the insight learning of man operates with the aid of ideas which are free from narrow confinements. (p. 138)"

Created November 19, 1997 by Lynn Morin

http://www.gmu.edu/depts/psychology/homepage/thorndike.html