Learning as a change in behaviour
In psychology, learning is a relatively lasting change in behaviour caused by experience. Three kinds matter most.
Classical conditioning (Pavlov)
Here a neutral signal comes to predict something important. The vocabulary:
- Unconditioned stimulus (UCS): naturally triggers a response (food).
- Unconditioned response (UCR): the natural reaction (salivation to food).
- Neutral stimulus: at first triggers nothing (a bell).
- Conditioned stimulus (CS): the bell after pairing with food.
- Conditioned response (CR): salivation to the bell alone.
After repeated pairings, the bell alone makes the dog salivate. If the bell is then sounded many times with no food, the response weakens: extinction. After a rest, it may briefly reappear (spontaneous recovery). Generalisation is responding to stimuli similar to the CS; discrimination is learning to respond only to the CS.
Everyday examples include a fear of dentists after painful visits, or feeling hungry at the sound of a lunch bell.
Operant conditioning (Skinner)
Here behaviour is shaped by its consequences.
| Adds something | Removes something | |
|---|---|---|
| Behaviour increases | Positive reinforcement | Negative reinforcement |
| Behaviour decreases | Positive punishment | Negative punishment |
"Positive" means adding, "negative" means removing. It does not mean good or bad.
- Positive reinforcement: praise after homework, so homework increases.
- Negative reinforcement: a car buzzer stops when you buckle up, so buckling increases.
- Positive punishment: a speeding ticket is added, so speeding drops.
- Negative punishment: phone taken away after hitting a sibling, so hitting drops.
Schedules of reinforcement
- Fixed ratio: reward after a set number of responses.
- Variable ratio: reward after an unpredictable number, such as a slot machine. This produces high, steady responding and strong resistance to extinction.
- Fixed interval: reward for the first response after a set time.
- Variable interval: reward after unpredictable time, such as checking for a text.
Observational learning (Bandura)
People learn by watching others and the consequences they receive. In the famous Bobo doll studies, children who saw an adult act aggressively toward a doll were more likely to imitate it.
Worked example
A child gets a sticker every time she puts away her toys.
- The sticker is added after the behaviour: positive.
- Toy-tidying becomes more frequent: reinforcement.
- So it is positive reinforcement.
Now suppose her parent stops giving stickers. Tidying slowly fades, which is extinction in the operant sense. If instead stickers were given after a random number of tidy-ups, the habit would last longer, as in variable-ratio schedules.