Lecture: Learning
Lecture script · about 15–20 minutes spoken · Week 6
Opening
A dog trainer teaches a shy dog a new trick, one small step at a time. A store manager designs a loyalty card that keeps customers coming back. A nurse helps a child stay calm when a needle comes out. All three are working with the same thing: how experience changes what people and animals do.
That's the subject of this lecture. In Week 1, you met behaviorism and the behavioral perspective, which explains behavior through learning. In Week 2, you learned to read an experiment and to judge whether a study treated its participants ethically. You'll use both tools here. Weeks 3, 4, and 5 followed the brain, the senses, and consciousness. This week turns to behavior that changes because of what happened before. Next week, in Week 7, the course asks how those experiences are kept, in memory.
Here's the plan. We'll learn by pairing, watch a learned link fade and return, meet a famous and troubling study, learn from consequences, sort the four kinds of consequence, build a behavior step by step, time the rewards, and finally learn by watching. One sentence to hold onto: up or down tells you the type, and added or removed tells you the sign.
1. Learn by pairing: classical conditioning
Picture a dog asleep at the far end of a house. Someone in the kitchen opens a can, and the dog arrives before the lid is off, drooling. Nobody taught it a trick. It worked out that the whir of the can opener comes right before dinner.
Ivan Pavlov was the first to study this kind of learning, with dogs. A hungry dog drools at food with no training at all. Pavlov found that a neutral sound that came right before the food, again and again, began to make the dog drool on its own.
First, the big idea. Learning is a relatively permanent change in behavior that comes from experience, rather than from growing up, illness, or a passing state such as fatigue. Classical conditioning is learning in which a neutral stimulus, after being paired with a stimulus that already triggers an automatic response, comes to trigger a similar response on its own; the response is involuntary.
Now the labels, using the can opener. The food is the unconditioned stimulus, or UCS: a stimulus that triggers a response automatically, with no learning needed. Drooling at the food is the unconditioned response, or UCR. Before any pairing, the whir is a neutral stimulus. After the whir has come right before dinner many times, the whir is the conditioned stimulus, or CS, and drooling at the whir is the conditioned response, or CR.
Here's a trick for the labels. "Unconditioned" means unlearned. "Conditioned" means learned.
Pause and try this. The whir comes first in time and the food comes second. Does that make the whir the unconditioned stimulus?
Hold your answer for a moment.
Here's the answer. No; the whir is the conditioned stimulus, because an untrained dog would not drool at a whir, and only a stimulus that works without learning, like the food, is unconditioned.
2. Watch the learned link change: acquisition, extinction, and recovery
A learned link isn't fixed. Think about a ringtone you used to have. When it always meant a message, you jumped at it. Change the ringtone, and the old sound loses its pull. Then hear it in a café a month later, and your hand still twitches toward your pocket.
Five words describe those changes, and the can opener shows all five. Acquisition is the stage of conditioning in which the learned link forms and the response grows stronger with repeated pairings or reinforcement. Night after night, the whir comes before dinner, and the drooling to the whir grows.
Now the family switches to dry food from a bag, but the can opener still gets used for other things. The whir keeps sounding with no dinner after it. That's extinction: the fading of a learned response when the conditioned stimulus keeps arriving without the unconditioned stimulus, or when a behavior is no longer reinforced.
After a few days with no cans opened at all, one whir brings back some drooling. That's spontaneous recovery, the sudden return of a weakened, extinguished response after a rest period, with no new training. It tells you something important. Extinction weakens a link. It doesn't erase it.
The blender makes a similar whir, and at first it starts the drooling too. That's generalization. But the blender never comes before food. Over time, the dog drools to the can opener and ignores the blender. That's discrimination, the learned ability to respond only to the stimulus that predicts what follows.
Pause and try this. If extinction erased the link completely, which of the five changes could never happen?
Give it a few seconds.
Here's the answer. Spontaneous recovery could never happen, because a response can't come back after a rest unless the link was still there underneath.
3. Meet Little Albert: conditioning, people, and ethics
Classical conditioning isn't only about dogs. It reaches feelings too. A child who gets a painful shot at a clinic may start to cry in the waiting room before anything happens. The room has become a signal.
In 1920, John B. Watson and Rosalie Rayner set out to show that a human emotion could be conditioned. Their study, now called the Little Albert study, conditioned fear in an infant. They paired an object the infant had not feared with something that frightened him, until the object alone produced fear. The frightening event was the UCS, and the once-harmless object became the CS.
This study is told today as history and as a warning. Hold it up against the research rules from Week 2. A participant must give informed consent, agreeing after being told what a study involves. An infant can't do that. Researchers must protect participants from harm. This study created fear on purpose. An Institutional Review Board would not approve it today.
There's a second lesson too. One infant is a case study. Week 2 told you a single case can suggest an idea, but it can't show the idea fits everyone.
Pause and try this. Name the two ethics standards from Week 2 that the Little Albert study fails.
Think it through before you go on.
Here's the answer. It fails informed consent, because an infant can't agree to take part, and it fails protection from harm, because it created fear on purpose.
4. Learn from consequences: operant conditioning
Most of your day isn't made of reflexes. It's made of choices: where to park, what to buy, whether to raise your hand. And your choices are shaped by what happened after similar choices before.
B. F. Skinner developed the study of this kind of learning. Operant conditioning is learning in which the consequences that follow a voluntary behavior change how often that behavior happens. Two terms do most of the work. Reinforcement is any consequence that makes the behavior it follows more likely to happen again. Punishment is any consequence that makes the behavior it follows less likely to happen again.
Notice that neither definition says "pleasant" or "unpleasant." A consequence is judged by what it does to the behavior.
Then come the two words that trip people up. Positive means something is added after the behavior. Negative means something is removed. Positive never means good, and negative never means bad. Think of them as plus and minus.
So how do you tell this from classical conditioning? Compare two moments. A song from a hard breakup plays in a store, and your stomach drops. You didn't choose that. It's involuntary, so it's classical. You raise your hand in a seminar, someone builds on your point, and you raise your hand more often after that. You chose to act, and what followed changed the behavior. That's operant.
Pause and try this. A smell from a family kitchen makes you feel calm before you even think about it. Is that classical or operant conditioning?
Decide, then read on.
Here's the answer. It's classical conditioning, because the calm is an involuntary response to a signal, not a chosen behavior shaped by its consequence.
5. Sort the four consequences, and keep negative reinforcement apart from punishment
Cross the two questions and you get four boxes. Did the behavior go up or down? Was something added or removed?
Let's sort four of this week's everyday learning scenarios. A toddler touches a hot stove, and pain is added right after. Touching goes down. Added and down: positive punishment. A driver parks in a loading zone, and a fine takes money away. Parking there goes down. Removed and down: negative punishment. A coffee shop adds a free drink after a set number of purchases. Buying there goes up. Added and up: positive reinforcement.
And the snooze alarm. Pressing snooze removes the blaring alarm, and pressing snooze goes up. Removed and up: negative reinforcement. Here's the trap. Because the story has something unpleasant in it, people call this punishment. It isn't. Negative reinforcement is removing something unpleasant after a behavior, which makes the behavior more likely; it strengthens behavior and is not a form of punishment. It feels like relief, not like a penalty.
Pause and try this. Pressing snooze makes the alarm stop, and you press snooze more and more. Why is that reinforcement and not punishment?
Take a moment with it.
Here's the answer. The behavior went up, and anything that makes a behavior more likely is reinforcement; it's negative because the alarm was removed.
6. Build a behavior step by step: shaping
Some behaviors never happen on their own, so there's nothing to reward. A shelter dog won't roll over by accident often enough to learn the trick from treats.
The way forward is to reward the steps. Shaping is building a new behavior by reinforcing successive approximations, each small step closer to the target behavior, until the whole behavior appears.
Here's how it goes. A volunteer first gives a treat when the dog lies down. Once that's reliable, lying down alone stops earning a treat; now the dog must lie down and tip onto one side. Then only a roll onto the back pays off. Finally, only a full roll, all the way over, earns the treat. Each reward is for a near miss, and the bar rises every time a step becomes reliable.
Pause and try this. Why can't the volunteer wait for the dog to roll over and reward it then?
Think about it before reading on.
Here's the answer. Because the full roll never appears on its own, waiting would mean nothing ever gets rewarded, so nothing would be learned.
7. Time the rewards: schedules of reinforcement
Most rewards in life don't follow every response. They come after a number of tries or after some time passes. Schedules of reinforcement are the rules that decide which responses earn reinforcement, based on a number of responses, called ratio, or an amount of time, called interval, that is either fixed or variable.
Cross fixed or variable with ratio or interval, and you get the four partial schedules, all in this week's scenarios. The punch card gives a free drink after a set number of purchases: a fixed-ratio schedule, or FR. A paycheck arrives after the same amount of time on the job: a fixed-interval schedule, or FI. A slot machine pays after an unpredictable number of pulls: a variable-ratio schedule, or VR. A kitchen that might be inspected any day stays clean every day: a variable-interval schedule, or VI.
Of the four partial schedules, variable-ratio reinforcement produces the highest and most extinction-resistant response rates. At a slot machine, the next pull always might pay. Your phone works the same way. Some checks bring a message you care about, most don't, and you can't tell which check will pay. That's why the habit is so hard to stop.
So across the 8 everyday learning scenarios, there's 1 classical case, the can opener, and 7 operant cases, which use all 4 consequences and all 4 partial schedules.
Pause and try this. The paycheck and the surprise inspection both depend on time. What makes one fixed and the other variable?
Answer it in your head first.
Here's the answer. Payday comes after the same amount of time every cycle, so it's fixed; the inspection can come on any day, so it's variable.
8. Learn by watching: observational learning
You've learned things you were never rewarded for. Maybe you learned to fix a bike chain by watching a video and got it right on the first try. Nobody reinforced you. You watched.
Observational learning is learning by watching another person, called a model, and later imitating or avoiding what the model did. The classic evidence is Albert Bandura's Bobo doll study, from 1961. Children who had watched an adult attack an inflatable doll were more likely to act aggressively toward it themselves.
Read it with your Week 2 tools. The independent variable, the one manipulated, was whether a child watched the adult attack the doll. The dependent variable, the one measured, was how aggressively the child then treated the doll. And the children got no reward of their own.
Two more terms. Vicarious reinforcement is becoming more likely to copy a behavior after seeing a model rewarded for it. Vicarious punishment is becoming less likely to copy a behavior after seeing a model punished for it, even though the behavior has been learned. A 1965 follow-up by Bandura showed vicarious punishment: children who saw the model punished imitated less, even though they had learned the behavior. Learning and doing aren't the same thing.
A limitation is worth naming. Hitting a toy built to be hit isn't the same as hurting a person, so ask how well that measure captures aggression.
Pause and try this. In the Bobo doll study, which variable was manipulated, and which was measured?
Give yourself a moment.
Here's the answer. Whether a child watched the attack was manipulated, so it was the independent variable; how aggressively the child treated the doll was measured, so it was the dependent variable.
Closing
Let's pull it together in the order we built it. Learning is a lasting change in behavior that comes from experience. In classical conditioning, a neutral stimulus paired with a UCS becomes a CS that triggers a CR. That link forms, fades, returns after a rest, spreads, and narrows. Little Albert showed that fear can be conditioned, and its methods are part of why today's ethics rules exist. In operant conditioning, reinforcement raises a voluntary behavior and punishment lowers it, with positive meaning added and negative meaning removed. Shaping rewards the steps. Variable-ratio rewards keep people going longest. And observational learning needs a model, not a reward of your own.
Next in this module come the slide deck, the Readings & Resources page, the lecture tutorial, the practice set, and the quiz. Next week, the course turns to memory: how the experiences that changed you are kept.
Built with The Course Maker · samples.thecoursemaker.com