Unit 4 · Learning Theory · Reading 6 of 13
Thorndike and Skinner: Learning From Consequences
Before you read
Think of something you do more often because it paid off once: a shortcut, a joke, a study trick. Did anyone teach you to do it, or did the result teach you?
Words you'll own by the end
- Operant conditioning: learning to connect a voluntary behavior with its consequence.
- Law of effect: behaviors followed by pleasant results are repeated; unpleasant results, less likely (Edward Thorndike).
- B. F. Skinner: behaviorist who named and expanded operant conditioning.
- Shaping: reinforcing each small step toward a target behavior (successive approximations).
- Chaining: linking a sequence of behaviors that must all be done to earn the reinforcer.
1. The big switch: voluntary behavior
Like classical conditioning, operant conditioning is about associations. The difference: here the behavior is voluntary. The learner does something, and a consequence follows. Reinforcements make the behavior more likely; punishments make it less likely.
The test from Reading 1 still works: Did something just happen to the learner (classical)? Or did the learner do something to earn a consequence (operant)?
2. Thorndike's cats
Edward Thorndike put hungry cats in puzzle boxes. To escape and reach food, a cat had to work a series of latches. At first the cats tried things at random. Once a cat escaped and got food, it repeated the moves that worked, and it got faster each time.
From this, Thorndike stated the law of effect: if a behavior leads to something pleasant, the organism is likely to repeat it. If it leads to something unpleasant, the organism is less likely to repeat it.
Check-in: Which is operant conditioning, not classical?
3. Skinner: the learner "operates" on the world
B. F. Skinner built on Thorndike's work. He chose the word operant because the learner actively operates on the environment. Skinner was a radical behaviorist. He argued thoughts and inner feelings can't be measured scientifically, so they don't belong in the study of behavior. To Skinner, your actions are the product of your history of reinforcement and punishment. He even argued that free will is a myth.
4. Teaching complex behavior: shaping and chaining
You can't reward a behavior that never happens. So how do you teach a dog to lie down if it never does it on command? Through shaping: you reinforce each small step that gets closer to the goal. These steps are called successive approximations.
Shaping a dog to lie down
Reward a sit
The dog sits on its own. Treat. Now sitting happens more.
Reward a lower sit
Now only reward when the dog leans or slides lower. Plain sitting no longer earns a treat.
Reward front legs down
Raise the bar again: front legs must stretch out.
Reward lying down
Only the full behavior earns the treat. Each step got closer: successive approximations.
Animal trainers also use chaining: linking a sequence of behaviors together. At first, a dolphin gets a fish after each trick. Once it has learned the whole sequence, it gets the fish only when it finishes the entire chain.
Try it: Shaping or chaining?
Pick the right technique.
A coach first praises a swimmer for putting her face in the water, then for floating, then for kicking while floating.
A parrot must ring a bell, climb a ladder, and slide down a ramp before it gets a seed.
A teacher praises a shy student for raising a hand, then for saying one word, then for a full answer.
A rat must press a lever, then pull a chain, then run to the food cup.
Am I ready for a pop quiz?
Check a box only if you can say the answer out loud without scrolling up.
Come back to your "before you read" question
Your own behavior that paid off once: that's the law of effect. What was the consequence that made you repeat it?