CONDITIONING
Conditioning is the term used to designate the types of human behavioral learning.
Since the 1920s, conditioning has been the primary focus of behavior research
in humans as well as animals. There are four main types of conditioning:
Classical Conditioning
Classical conditioning, also called associative learning, is based on stimulus-response
relationships. A stimulus is an object or situation that elicits a response
by one of our sense organs, like how a bright light makes us blink. Associative
learning allows us to associate two or more stimuli and change our response
to one or more of them as a result of simultaneous experience.
According to classical conditioning, learning occurs when a new stimulus begins
to elicit behavior similar to the behavior produced by an old stimulus. Studies
into classical condition began in the early 1900s by the Russian physiologist
Ivan P. Pavlov. Pavlov trained dogs to salivate in response to two stimuli:
noise or light, and food or a sour solution. The dogs' salivation is automatically
elicited by the food and sour solution, so these were called the unconditional
stimulus. However, when the noise or light (conditional stimulus) was repeatedly
paired with the food or sour solution over an extended period of time, the dogs
would eventually salivate at the noise or light alone. This is a prime example
of a conditioned response.
Unconditional stimuli, such as the food and sour solution, allow the learning to occur, while also serving to reinforce the learning. Without an unconditional stimulus in his experiment, Pavlov could not have taught the dogs to salivate at the presence of the noise or light.
Classical conditioning is particularly important in understanding how people learn emotional behavior. For example, when we develop a new fear, we have learned to fear a particular stimulus which has been combined with another frightening stimulus.
Operant Conditioning
Operant conditioning is goal-directed behavior. We learn to perform a particular
response as a result of what we know will happen after we respond. For example,
a child may learn to beg for sweets if the begging is usually successful. There
is no single stimulus that elicits the begging behavior, but instead it occurs
because the child knows that this action may result in receiving treats. Every
time the child receives sweets after begging, the behavior is reinforced and
the tendency of the child to beg will increase.
During the 1930s, American psychologist and behaviorist Burrhus F. Skinner
performed several important experiments into operant conditioning. Using what
is now termed a Skinner Box, he trained rats to press levers to receive food.
A hungry rat would be placed in a box containing a special lever attached to
concealed food. At first the hungry rat would wander around the box, investigating
its surroundings. Eventually it would accidentally press the lever thereby releasing
a food pellet into the box. At first the rat would not show any signs of associating
the two events, but over time its exploring behavior becomes less random as
it begins to press the lever more often. The food pellet reinforced the rat's
response of pressing the lever, so eventually the rat would spend most of its
time just sitting and pressing the lever. This type of learning is based on
the idea that if a behavior is rewarded, the behavior will occur more frequently.
There are four main types of operant learning:
- Positive Reinforcement: The more often the subject performs the action, the more the stimulus is reinforced (for example, pressing a lever for food)
- Negative Reinforcement: Performing an action removes an undesirable stimulus (for example, studying for an exam removes anxiety)
- Punishment: Performing an action elicits an undesirable stimulus (for example, receiving a shock by pressing a lever)
- Omission Training: Performing an action prevents the delivery of a pleasant stimulus (for example, disciplining a child by sending him to his room)
Multiple-Response Learning
When we learn skills, we must first learn a sequence of simple movement-patterns. We combine these movement-patterns to form new, more complicated behavioral patterns with stimuli guiding the process. For example, efficient typing requires us to put together many finger movements, which are guided by the letters or words that we want to type. We must first learn to type each letter, and then learn to put the movements together to type words and then phrases.
To investigate this type of learning, psychologists have observed animals learning to run through mazes. An animal first wanders aimlessly through the maze, periodically coming to a choice-point, where it must turn either left or right. Only one choice is correct, but the correct direction cannot be determined until the animal has reached the end of the maze. By running through the maze numerous times, the animal can learn the correct sequence of turns to reach the end. It has been found that the sequences of turns near the two ends of the maze are learned more easily than the parts near the middle. Similarly, when we try to learn a list of items, we usually find the beginning and the end easier than the middle.
An extension of multiple-response learning is latent learning, where learning occurs in the absence of an immediate reward. If rats are allowed to repeatedly run a maze without a reward at the end, they will learn the maze quite slowly. However if they have first been allowed to run the maze without reward, and then food is placed at the end, these rats will learn the maze very quickly. This shows that learning had been occurring the entire time, but did not become evident until it was activated by a reward.
Insight Learning
Insight refers to learning to solve a problem by understanding the relationships of various parts of the problem. Often insight occurs suddenly, such as when a person struggles with a problem for a period of time and then suddenly understands its solution. Therefore insight learning is solving problems without experience. Instead of learning by trial-and-error (such as in maze running), insight learning involves trials occurring mentally.
In the early 1900s, Wolfgang Kohler performed insight experiments on
chimpanzees. Kohler showed that the chimpanzees sometimes used insight instead
of trial-and-error responses to solve problems. When a banana was placed high
out of reach, the animals discovered that they could stack boxes on top of each
other to reach it. They also realized that they could use sticks to knock the
banana down. In another experiment, a chimp balanced a stick on end under a
bunch of bananas suspended from the ceiling, then quickly climbed the stick
to obtain the entire bunch intact and unbruised (a better technique than the
researchers themselves had in mind). Kohler's experiments showed that primates
can both see and use the relationships involved to reach their goals.
Next >>> General Learning Concepts
