Learning
Key References
DEFINING
THORPE (1963) defines learning as “that process which manifests itself by adaptive changes in individual
behaviour as a result of experience”.
Example of innate/learnt fallacy: chaffinch song development (CATCHPOLE AND SLATER 2008)
Basic song does not need learning
Learning needed for full song
Will only learn chaffinch song
Learning only during 1st year
Genes affect learning
Monkeys learn which stimuli to fear but more easily learn to fear snakes than flowers (ÖHMAN AND
MINEKA 2001)
Genes can interfere with efficient learning (at least in unnatural environments). K AND M BRELAND
(1961) trained racoons to put tokens into a slot machine for a food reward, but they would “waste”
time washing tokens as this is their innate behaviour for food preparation (contradicts notion that
animals are a “blank slate”)
Genes promote learning by: selecting for learning at certain times of life (e.g. from parents), making
some associations easier than others, providing animals with goals (positive reinforcements e.g.
sweet taste and negative reinforcements e.g. pain) to guide them to learn to approach and avoid
different situations adaptively
NON-ASSOCIATIVE LEARNING
Habituation example: presenting crowd stimuli to crowd-control horses so they don’t respond to flag-
waving/shouting.
Sensitisation example: the marine worm Nereis. If the worm is kept in a small tube and fed at regular
intervals, it becomes progressively more likely to respond to any novel stimulus, such as a change in
illumination, by exploratory, food-seeking movements toward the open end of the tube. If, on the other
hand, the worm receives mild electric shocks at regular intervals, it becomes progressively more likely to
respond to a novel stimulus by withdrawal.
ASSOCIATIVE LEARNING
Pavlovian/classical conditioning
Pavlov’s dogs
Unconditioned stimulus (US) = food
Conditioned stimulus (CS) = bell
Unconditioned response (UR) = salivating when presented with food
Conditioned response (CR) = salivating when bell rings (the bell was previously a neutral and novel
stimuli)
Dogs response was generalized because it also had the salivary reflex for other tones/frequencies
of bell (though there was less of a response), hence also response to similar sounds. However,
there was also discrimination because there was a lower response to different frequencies from the
original sound, and discrimination can be increased if responding to other frequencies were
punished
, Operant conditioning/instrumental conditioning
Pig heating example
Was believed that pig welfare would be better if they could control their own temperature. By
accident, an exploring pig presses a button to turn a heat lamp on. Process repeats, and the pig
eventually associates the button with heat (after 3/4/5 trials)
US = heat
UR = investigation/exploration
CR = panel/button press
Laws of conditioning
1) An increase in interval between CS and US presentation will decrease learning.
Some delays don’t matter e.g. food aversion learning (biologically predisposed to associate novel flavours
with vomiting, not very sensitive to the lag between ingestion and vomit – is this because we taste the
flavour as it comes back up?).
2) The speed or strength of learning increases with the intensity of the CS, and the value of the US
(reinforcer).
3) When the reinforcer is withheld the learned response declines (extinction).
4) All pairs of events can be associated with equal ease. This is equipotentiality.
Not quite true. Examples:
SHETTLEWORTH’s (1980) hamsters
Trained hamsters to do different things when she walked in the room for a food reward
E.g. gave them food pellets when she walked in -> hamsters start standing up
Worked with digging as well but NOT with washing faces (biological constraints)
GARCIA AND KOELLING (1966) found that rats are much more prepared to associate taste with sickness
after a single trial and a long delay, but not so much for visual stimuli:
Gave rats drinking water with saccharin-flavouring and flashed lights then irradiated to make sick ->
rat avoided saccharin taste but did not avoid flashing lights
Things you eat are more likely to make you sick than visual stimuli so association easier to make
Reinforcers
Positive reinforcers
The intentions of the trainer are irrelevant e.g. shouting at a barking dog causing it to bark more is
an example of positive reinforcement
IMPRINTING
Filial imprinting: genetic predisposition to be attracted to the head of another animal
Sensitive period
Good stimulus (e.g. stuffed chicken): imprinting is irreversible
Worse stimulus (e.g. red ball): can imprint again on a stronger stimulus
SOCIAL LEARNING
Key References
DEFINING
THORPE (1963) defines learning as “that process which manifests itself by adaptive changes in individual
behaviour as a result of experience”.
Example of innate/learnt fallacy: chaffinch song development (CATCHPOLE AND SLATER 2008)
Basic song does not need learning
Learning needed for full song
Will only learn chaffinch song
Learning only during 1st year
Genes affect learning
Monkeys learn which stimuli to fear but more easily learn to fear snakes than flowers (ÖHMAN AND
MINEKA 2001)
Genes can interfere with efficient learning (at least in unnatural environments). K AND M BRELAND
(1961) trained racoons to put tokens into a slot machine for a food reward, but they would “waste”
time washing tokens as this is their innate behaviour for food preparation (contradicts notion that
animals are a “blank slate”)
Genes promote learning by: selecting for learning at certain times of life (e.g. from parents), making
some associations easier than others, providing animals with goals (positive reinforcements e.g.
sweet taste and negative reinforcements e.g. pain) to guide them to learn to approach and avoid
different situations adaptively
NON-ASSOCIATIVE LEARNING
Habituation example: presenting crowd stimuli to crowd-control horses so they don’t respond to flag-
waving/shouting.
Sensitisation example: the marine worm Nereis. If the worm is kept in a small tube and fed at regular
intervals, it becomes progressively more likely to respond to any novel stimulus, such as a change in
illumination, by exploratory, food-seeking movements toward the open end of the tube. If, on the other
hand, the worm receives mild electric shocks at regular intervals, it becomes progressively more likely to
respond to a novel stimulus by withdrawal.
ASSOCIATIVE LEARNING
Pavlovian/classical conditioning
Pavlov’s dogs
Unconditioned stimulus (US) = food
Conditioned stimulus (CS) = bell
Unconditioned response (UR) = salivating when presented with food
Conditioned response (CR) = salivating when bell rings (the bell was previously a neutral and novel
stimuli)
Dogs response was generalized because it also had the salivary reflex for other tones/frequencies
of bell (though there was less of a response), hence also response to similar sounds. However,
there was also discrimination because there was a lower response to different frequencies from the
original sound, and discrimination can be increased if responding to other frequencies were
punished
, Operant conditioning/instrumental conditioning
Pig heating example
Was believed that pig welfare would be better if they could control their own temperature. By
accident, an exploring pig presses a button to turn a heat lamp on. Process repeats, and the pig
eventually associates the button with heat (after 3/4/5 trials)
US = heat
UR = investigation/exploration
CR = panel/button press
Laws of conditioning
1) An increase in interval between CS and US presentation will decrease learning.
Some delays don’t matter e.g. food aversion learning (biologically predisposed to associate novel flavours
with vomiting, not very sensitive to the lag between ingestion and vomit – is this because we taste the
flavour as it comes back up?).
2) The speed or strength of learning increases with the intensity of the CS, and the value of the US
(reinforcer).
3) When the reinforcer is withheld the learned response declines (extinction).
4) All pairs of events can be associated with equal ease. This is equipotentiality.
Not quite true. Examples:
SHETTLEWORTH’s (1980) hamsters
Trained hamsters to do different things when she walked in the room for a food reward
E.g. gave them food pellets when she walked in -> hamsters start standing up
Worked with digging as well but NOT with washing faces (biological constraints)
GARCIA AND KOELLING (1966) found that rats are much more prepared to associate taste with sickness
after a single trial and a long delay, but not so much for visual stimuli:
Gave rats drinking water with saccharin-flavouring and flashed lights then irradiated to make sick ->
rat avoided saccharin taste but did not avoid flashing lights
Things you eat are more likely to make you sick than visual stimuli so association easier to make
Reinforcers
Positive reinforcers
The intentions of the trainer are irrelevant e.g. shouting at a barking dog causing it to bark more is
an example of positive reinforcement
IMPRINTING
Filial imprinting: genetic predisposition to be attracted to the head of another animal
Sensitive period
Good stimulus (e.g. stuffed chicken): imprinting is irreversible
Worse stimulus (e.g. red ball): can imprint again on a stronger stimulus
SOCIAL LEARNING
