A dog drools when it hears a tin of food open. A child keeps whining because it sometimes works. A teenager picks up a parent's way of handling anger without ever being taught it on purpose. All of these are forms of learning, which psychologists define as a relatively lasting change in behavior that comes from experience. This level covers the three great discoveries about how learning happens: classical conditioning, operant conditioning and learning by observation.

Pavlov's dogs

The Russian physiologist Ivan Pavlov won a Nobel Prize in 1904 for his work on digestion, not psychology. While measuring how much dogs salivated when fed, his team noticed that the dogs began drooling before the food arrived, at the sight of the assistant or the sound of footsteps. Pavlov turned this nuisance into a research program. He repeatedly paired a neutral signal, such as a metronome or buzzer, with food. (Popular retellings usually say a bell.) Eventually the signal alone produced salivation. This is classical conditioning: a neutral stimulus, after being paired with one that naturally triggers a response, comes to trigger a similar response on its own. If the signal is then presented many times without food, the response gradually fades, a process called extinction.

Learned fears

Classical conditioning helps explain emotional reactions. In 1920 John B. Watson and Rosalie Rayner showed a baby, known as "Little Albert," a white rat while striking a steel bar loudly behind his head. After several pairings, the child cried at the sight of the rat, and his fear seemed to spread to other furry things, an effect called generalization. The study would never be allowed today, since the researchers frightened a baby and never reversed the fear, and it helped inspire modern rules protecting research participants. It also pointed toward a treatment. If fears can be learned, they can be unlearned, and later therapists developed exposure therapy, in which people face feared situations gradually and safely until the fear weakens.

Some associations are learned with surprising speed. If you get sick after eating a particular food, you may avoid it for years, even after a single experience and even if something else made you ill. This taste aversion, studied by John Garcia in rats, suggests that evolution prepares us to learn some links more easily than others.

Consequences shape behavior

Classical conditioning is about signals that predict events. Operant conditioning is about how the consequences of our actions change how often we repeat them. Around 1898 Edward Thorndike placed cats in puzzle boxes and found that actions followed by escape and food became more likely, which he called the law of effect. B. F. Skinner later studied this in detail, using boxes in which rats pressed levers and pigeons pecked keys. He distinguished:

• Reinforcement, which makes a behavior more likely. Positive reinforcement adds something pleasant, such as praise; negative reinforcement removes something unpleasant, such as the beeping that stops when you buckle a seatbelt.
• Punishment, which makes a behavior less likely, by adding something unpleasant or taking away something valued.

People often confuse negative reinforcement with punishment, but they work in opposite directions. Research also suggests that punishment teaches what not to do without teaching what to do instead, and harsh physical punishment is linked to more aggression in children, not less.

Schedules of reinforcement

Skinner found that the timing of rewards matters a great deal. Rewarding every response produces fast learning, but the behavior stops quickly once rewards end. Rewarding responses unpredictably, after a varying number of tries, produces behavior that is very persistent. This variable ratio schedule is the pattern of a slot machine, and of many phone apps that deliver new likes or messages at random intervals. It also explains the whining child: a parent who gives in only occasionally is teaching the child to keep trying.

Learning by watching

Not everything is learned through one's own rewards and punishments. In the early 1960s Albert Bandura showed young children an adult punching and kicking an inflatable clown called a Bobo doll. Children who had watched the aggressive adult were much more likely to attack the doll themselves, often copying the adult's exact moves and words. This is observational learning, or modeling. Bandura showed that we learn from what we see others do and what happens to them, which helps explain why role models, peers and media matter.

Worked example: building a habit

Imagine you want to start exercising regularly. How could learning principles help? Use reinforcement: pair the workout with something you enjoy, such as a favorite podcast you only listen to while walking. Start with small, easily rewarded steps rather than a punishing routine you will soon quit. Use cues: lay out your shoes by the door so the sight of them becomes a signal, much like Pavlov's metronome. Use modeling: exercise with a friend whose habits you would like to share. And be patient with lapses, since one missed day does not erase learning. These ideas are not tricks. They reflect how behavior responds to its environment, which means that changing our surroundings is often easier than relying on willpower alone.