Conditioning

Classical conditioning was first studied by the Russian physiologist Ivan Pavlov. It is learning by association between two stimuli (S-S learning).

In his experiment, a hungry dog was placed in a harness and its saliva was measured.

• Food naturally makes the dog salivate. Food is the unconditioned stimulus (US) and salivation to it is the unconditioned response (UR): nobody had to teach it.
• A bell, at first, causes no salivation; it is a neutral stimulus.
• Pavlov rang the bell and then gave food, many times over.
• Soon the dog salivated at the bell alone. The bell had become the conditioned stimulus (CS) and salivation to it the conditioned response (CR).

What makes this remarkable is that salivation is involuntary: the dog cannot decide to do it, yet a new trigger for it was learnt. Everyday example: if a particular ringtone always comes before your coach shouting at you, your heart may start racing whenever you hear that ringtone.

Determinants of classical conditioning:

• Time relation between CS and US. In simultaneous conditioning both start together; in delayed conditioning the CS starts first and is still on when the US begins; in trace conditioning the CS starts and ends before the US; in backward conditioning the US comes first. Delayed conditioning is the most effective, backward the least.
• Type of US: appetitive stimuli (food, praise) cause approach; aversive ones (a loud noise, pain) cause avoidance or escape. Aversive conditioning usually develops in fewer trials.
• Intensity of the CS: a stronger, clearer CS produces quicker conditioning.

Operant (instrumental) conditioning, studied by B. F. Skinner, explains how voluntary behaviours (operants) are learnt from their consequences (R-C learning). A hungry rat in a 'Skinner box' moves about and by chance presses a lever; food drops in. After a few trials the rat goes straight to the lever. The response works as an instrument for getting food.

Reinforcement is anything that strengthens a response:

• Positive reinforcement: a pleasant consequence follows the response (a student who answers well is praised, and answers more often).
• Negative reinforcement: the response removes or avoids something unpleasant. It leads to escape and avoidance learning: you learn to carry an umbrella in the rainy season to avoid getting soaked.
• Punishment is different: an unpleasant consequence that weakens or stops a response. Negative reinforcement strengthens behaviour; punishment suppresses it.

Operant learning also depends on the number and amount of reinforcements, their quality (stale food reinforces little), the schedule, and delay: reinforcement given long after the response produces poorer learning. Continuous reinforcement (every response rewarded) gives fast learning. Partial (intermittent) reinforcement (only some responses rewarded) makes the response much more resistant to extinction, because the learner keeps trying in case the next one pays off.

Key learning processes:

• Primary reinforcers meet biological needs (food, water); secondary reinforcers get their value through learning (money, grades, praise).
• Extinction: the learnt response gradually disappears when reinforcement (or the US) stops following it.
• Spontaneous recovery: after extinction and a rest, the response suddenly reappears when the CS is presented again.
• Generalisation: responding in the same way to similar stimuli (a child bitten by one dog becomes afraid of all dogs). Discrimination: responding differently to stimuli that differ (the child learns only that particular dog is dangerous). Generalisation is a failure of discrimination.
• Learned helplessness: after repeated experiences of being unable to control unpleasant events, a person (or animal) stops trying even when escape becomes possible.