LEC05 Problem with S-R theor

FURTHER PROGRESS IN ANIMAL LEARNING

- Konorski: Suggested classical conditioning could be explained instrumentally, have preparatory vs consummatory responses - Solomon's students (Kamin, Rescorla...): 2-factor theory of avoidance + conditioned suppression - Sutherland + Mackintosh: Attentional theory of discrimination learning

COGNITIVE REVOLUTION

- Rejection of S-R theory and behaviourism inc. Skinner - people have cognitive processes, and are not just machinery (spits + twitches) - Skinner: Verbal Behaviour (1957) - used operant conditioning to explain language acquisition - Chomsky: Critique of Verbal Behaviour - proposed nativist view - Psycholinguistics: Studying psychology of language - Study of thinking not in terms of drives/needs - AI and computer models of mental process VS telephone analogy (connections) o Reveals complexity of processing within mind - Brewer: No convincing evidence of conditioning in humans - i.e. people are aware of associative learning, and cannot be conditioned w/o awareness

Kamin: Blocking effect

1. 16x trials of noise —> shock 2. 8x trials of noise + light —> shock 3. Test: Light —>? No fear o —> Implies that learning occurs when unexpected association occurs i.e. noise-shock association blocked learning of light-shock association o Not so complete blocking in other experiments o Works in appetitive conditioning too (bell + light w/ food)

PROBLEMS WITH S-R-RFT THEORY (HULL + SPENCE)

1. Learning occurred w/o direct pairing of stimulus + reinforcement/consequence 2. Reinforcement =/= drive reduction 3. Extinction, punishment and avoidance learning: Phenomena not explained by simple S-R theory

SKINNER'S RADICAL BEHAVIOURISM

Does not consider internal events for explanation of behaviour - Suggested learning theory to explain behavior was not necessary - just needed to describe behaviour + define stimuli, response and rft in terms of behaviour e.g. S —> R relationship o Do not need to understand physiological events of CNS o Do not assume intervening variables/mental processes e.g. hunger o *Adding these factors simply complicates the process (as they need to be explained)+ do not add anything useful

Extinction, punishment and avoidance learning:

Phenomena not explained by simple S-R theory a. Extinction (no reward) + punishment (slapped if lever touched) —> less conditioned responding b. Partial reinforcement extinction effect c. Avoidance learning/2-factor theory (Mowrer

ASSOCIATIVE LEARNING THEORY (S-S THEORY)

Modern conditioning theories, unlike behaviourist approaches denying mental processes (simple S-R theory) - Distinction between learning and performance - General principles of associative learning, effects on behaviour (performance) highly specific to species o Attention, associative strength, time/rate estimates - Stimuli compete for associative strength - which one to associate more strongly? - Initially based on classical conditioning in animals - Limited interest in applications (unlike Skinner), but focus on theory - Later theories inc. Rescorla-Wagner applied to other forms of learning e.g. causal learning

Partial reinforcement extinction effect

More resistant to extinction if partially reinforced - Response occurs even without rft i. Frustration theory (Amsel): Increase in pressing due to frustration ii. Anticipation of reward: Cannot determine when they will get reward - involve frustration anticipatory response

Avoidance learning/2-factor theory (Mowrer

No rft according to S-R-Rft theory, but response is maintained? i. Shuttle box paradigm ii. Classical conditioning: Associate warning signal w/ shock iii. Instrumental conditioning: Escape chamber to reduce fear (rft) o Continued escaping even though shock not applied anymore o Not considered as rft according to theory as '-ve reinforcer'

Behaviour of Organisms (1938)

Pavlovian phenomena in instrumental conditioning § Extinction, conditioned anxiety/suppression/CER, spontaneous recovery

Latent learning (Blodgett, Tolman):

Reward —> incentive to perform, but learning occurs prior to reward i. Regularly rewarded rats —> greater reduction in errors than non-rewarded ii. Rats rewarded after day 11 —> steep reduction in errors, performance like/even better than rewarded rats

Reinforcement =/= drive reductio

a. Harlow studied rhesus monkeys which solved problems w/o reward e.g. learning to pick out different stimuli in set of 2A + B b. Butler (student): Semi-hungry monkeys pressed lever to open window to watch banana/working model train à learning to satisfy curiosity, not satisfy drive/need for food

Learning occurred w/o direct pairing of S-R-Rft

a. Latent learning (Blodgett, Tolman) b. Sensory preconditioning (Brogden):

Sensory preconditioning (Brogden):

i. Sensory pre-conditioning - neutral stimuli paired: Light (S1) paired w/ noise (S2) ii. Noise (S2/CS) paired w/ shock (US) —> fear response (UR) iii. Test: Light —> fear response (light-shock association found even without pairing with shock directly) o Suggests light-noise association created in first stage even w/o pairing w/ rft

Schedules of rft

n pigeons - fixed/variable, ratio/interval

Research

o Skinner box o Operant conditioning o Behaviour of Organisms (1938) o Distinguished between classical + operant conditioning o Schedules of rft

Skinner box

to study response-rft contingencies —> operant conditioning § In contrast to rat mazes at the time

Operant conditioning

§ Behaviour was voluntary e.g. could lever press whenever, and however often § Behaviour influenced by outcome/consequence i.e. operant • Rather than influenced by antecedent i.e. respondent in classical conditioning

Extinction (no reward) + punishment (slapped if lever touched)

—> less conditioned responding i. Extinction: Suggests not permanent habit to CS i.e. unlearning habits ii. Punishment: Not encompassed in theory