Learning

What is learning?

Lasting change in behaviour that comes from experience

Which theorist believed that most of earning occurred by trial and error?

Pavlov

Skinner box

Small chamber used in operant conditioning of animals that limits the available responses. Increases likelihood that desired response will occur

Classical conditioning involves learning about the relation between ___, whereas operant conditioning involves learning about the relation between ___.

Two stimuli, three stimuli

Watson’s experiment with little Albert demonstrated that fears might be based on ___.

Classical conditioning

3 learning theory assumptions

• Learning is shaped by experiences

  • Most human behaviours are learned, not inborn

• Learning is adaptive

  • Behaviours that help us survive or avoid harm get repeated

• Learning can be studied scientifically

Unconditioned stimulus (UCS)

A stimulus that naturally produces a response

e.g. food, bacteria, loud noise

Unconditioned response (UCR)

A natural, automatic response to the UCS

e.g. salivation to food, nausea, fear

Conditioned stimulus (CS)

A previously neutral stimulus that becomes meaningful after being paired with UCS

e.g. bell

Conditioned response (CR)

A learned response to the CS after conditioning

e.g. salivation to bell

Classical conditioning

Learning through association between two stimuli (CS+UCS)

3 laws of learning

1. Law of contiguity

2. Law of similarity

3. Law of prediction

Law of contiguity

Things experienced close together in time become associated

e.g. lighting → thunder

Law of similarity

Things that look alike become associated. Similar objects trigger similar responses

Law of prediction

Learning occurs only if the CS predicts the UCS. CS must give useful information

e.g. bell → food

bell + light → food

Light is paired with food but dog does not learn about the light because bell already protects food and light adds no new information

Blocking in classical conditioning

When CS already predicts UCS, new stimuli (CS2) add no new info → no learning occurs

Timing learning

Animals learn how long after the CS the USC will occur (temporal learning)

Habituation

Decrease in response after repeated exposure to same stimulus.

• Response gets weaker

• Hellos direct our limited attention to more important things

• Response reduced, not lost!

e.g. startle reflex

Sensitisation

Increase in response after repeated exposure to a stimulus, especially if intense or threatening.

• Response gets stronger

• Helps be more alert to potential danger

• Often when stimulus is intense, signals danger, anxious person

e.g. gunshot, snake hiss

Operant conditioning

Learning through consequences (reinforcement and punishment)

Acquisition

Gradual learning of CR. Phase where CS is repeated paired with the UCS and learning occurs. (CS→CR)

• Gradual, not instant. Happens with repeated pairing

• Association becomes stronger with each pairing

e.g. bell + food = salivation

Extinction

Gradual weaning of CR. When CS is repeatedly presented without UCS, CR decreases

• Over time CS no longer triggers CR

• Original learning CR is suppressed not erased

• More stable if learned early in life (less spontaneous recovery)

e.g. bell with no food → salivation decreases → eventually stops

Spontaneous recovery

Return of a weakened CR after extinction and rest period.

e.g. dog salivates to bell after break but response fades quickly

Generalisation

When stimuli similar to CS also triggers the CR.

• More similar → stronger generalisation

• Adaptive because it helps learning apply to new but related situations

e.g. if CS is a 1000 Hz tone, dog may also salivate to 950 Hz or 1100 Hz tones but less strongly

If you fear one dog, you might fear all dogs

Discrimination

Learning to respond only to the specific CS and not similar stimuli.

• Opposite of generalisation

• Can be trained by pairing one stimulus with the UCS and one without

e.g. 1000 Hz tone → food → salivation

800 Hz tone → no food → no salivation

Hungry at 6pm, not 4pm.

Dog salivates to the bell, not every sound.

Factors influencing classical conditioning

• Timing

  • Learning is strongest when CS comes just before the UCS (predictive timing)

  • Exception: taste aversion

• Predictability

  • Conditioning works best when CS reliably predicts the UCS

  • Conditioning is stronger when CS is consistent and reliable

• Novelty

  • New stimuli are learned faster than familiar ones

• Salience

  • More noticeable or biologically relevant stimuli are learned more easily

  • e.g. bright light, loud tone, strong smells

Emotional conditioning

When neutral stimuli become linked to emotional responses (fear, disgust, pleasure)

Taste aversion conditioning

Strong learning after one pairing of food (CS) with illness (UCS). Very resistant to extinction

Why phobias form

Fear is learned through classical conditioning. Fear responses occur before conscious thought

Why knowing it’s safe doesn’t remove fear

Fear circuits activate faster than conscious reasoning

Exposure therapy

Repeated, safe exposure to the CS without the UCS. Extinction of fear

Why relapse happens after extinction

Extinction creates new learning (“safe now”) but does not erase old fear. Fear can return under stress or new contexts

Immune system conditioning

Immune responses can be classically conditioned

e.g. risk of death increases after losing a partner, chronic stress can trigger or worsen gut disorders like IBS

Addiction and craving conditioning

Drug cues (CS) (e.g. places, people) trigger cravings (CR) because they predict the drug (UCS). Why relapse happens when someone encounters old cues

Advertising conditioning

Pairing products (CS) with positive images/music (UCS) creates positive feelings

How does past learning affect new learning?

• Relearning

  • After extinction, CR returns faster the second time because old learning leaves traces

• Blocking

  • If CS1 already predicts the UCS, adding CS2 gives no new information → no learning about CS2

• Latent inhibition

  • If a neutral stimulus is presented many times without a UCS, learning is slower later on (e.g. bell rings many times with no food → dog is slower to learn bell = food

Operant conditioning

Learning where behaviour is shaped by its consequences (reinforcement or punishment)

Reinforcement

Any consequence that increases the likelihood of a behaviour

Punishment

Any consequence that decreases the likelihood of a behaviour

Positive reinforcement

Adding something pleasant to increase behaviour

e.g. giving a reward, getting a trophy for a race

Negative reinforcement

Removing something unpleasant to increase behaviour

e.g. car stops beeping once you buckle your seatbelt

Positive punishment

Adding something unpleasant to decrease behaviour

e.g. scolded for being late

Negative punishment

Removing something pleasant to decrease behaviour

e.g. losing phone privileges for breaking a rule

Shaping

Reinforcing small steps towards a complex behaviour

Extinction (operant conditioning)

Behaviour decreases when reinforcement stops

Discriminative stimulus

A cue that signals when a behaviour will be reinforced.

e.g. blinking red “out of service” light on office coffee machine → walking to cafe down the street instead of interacting with office machine

Primary reinforcer

Naturally rewarding

e.g. food, warmth, water

Secondary reinforcer

Gains values through learning

e.g. money, praise, grades

2 reinforcement schedules

1. Continuous reinforcement

2. Partial reinforcement

Continuous reinforcement

Behaviour is reinforced every time. Fast learning but fast extinction

Partial reinforcement

Behaviour is reinforced sometimes. Slower learning but more resistant to extinction.

Two main types of partial reinforcement

1. Ratio

   1. Fixed ratio: reinforcement after a set number of responses (e.g. buy 10 get one free card)

   2. Variable ratio: reinforcement after a changing number of responses (e.g. gambling machines)

2. Interval

   1. Fixed interval: reinforcement after a fixed time has passed (e.g. weekly pay)

   2. Variable interval: reinforcement after a changing time has passed (e.g. checking emails)

Why reinforcement is more effective than punishment

Reinforcement teaches what to do.

Punishment teaches what not to do.

Law of effect

Behaviours followed by rewards tend to increase (strengthened).

Behaviours followed by punishments tend to decrease (weakened).

Creates by Edward Thorndike

Learned helplessness

When repeated punishment or uncontrollable events lead to giving up.

Overjustification effect

External rewards reduce intrinsic motivation

e.g. liking an activity less after being paid for it

Factors affecting operant conditioning

1. Timing

   1. Consequence must happen immediately after behaviour

   2. Delays can reinforce wrong behaviour

2. Magnitude

   1. Bigger rewards/punishers → stronger effects

Operant conditioning vs classical conditioning

Classical is about what happens before the behaviour (e.g. associating two things together like bell and food)

• Learning association between two stimuli

• Behaviour is triggered by what comes before it

• Behaviour does not cause the outcome

• Behaviour is elicited (reflexive)

Operant is about what happens after the behaviour (reward or punishment)

• Learning association between behaviour and consequences

• Behaviour is triggered by what comes after it

• Behaviour causes the outcome

• Behaviour is emitted (voluntary)

e.g. cat and fridge

Classical: fridge sound → salivation

Operant: running to kitchen → gets food

Observational learning

Learning by watching others, not direct experience.

What affects whether we imitate a model?

1. Prestige of the model

   1. Important, powerful, respected, high status

2. Likability and attractiveness of model

   1. Admire, like, find friendly/appealing

3. Vicarious conditioning

   1. Watching the model’s consequences. If model is rewarded → more likely to copy them