PSY236 Lecture Notes - Lecture 5: Learned Helplessness, Vending Machine, Pangs
PSY236 – Week 5: Extinction and Punishment
Extinction → learned behaviours diminish through absence of anticipated
reinforcement. In an applied context, it is about eliminating unwanted behaviours.
• It is widely recognised that extinction weakens behaviour without erasing the
original learning
• Extinction in operant conditioning → no reinforcement after the response (e.g.
the rat’s bar press)
• Extinction in classical conditioning → repeated presentations of the CS alone
(e.g. repeatedly exposing someone to a fear of heights so his anxiety
diminishes)
Effects of extinction
• 1. The rate of responding decreases
• 2. Response variability increases – operant conditioning
Extinction can relapse
• Which means that the original learning is still there and can reappear – under
various conditions
• Problem from a therapeutic viewpoint
Recovery/relapse effects → extinction does not equal not un-learning
• 1. Spontaneous recovery
o After a rest period, you can get spontaneous recovery of the
extinguished response (continuous recoveries will decline though)
• 2. Rapid reacquisition
• 3. Resurgence
• 4. Renewal
• 5. Reinstatement
1. Spontaneous recovery → the reappearance of the target behaviour following
its extinction
• The passage of time after extinction leads to a return of the extinguished
operant or conditioned response such as fear
• Spontaneous recovery is not a problem, as long as the behaviour is not
reinforced when it reappears
• Spontaneous Recovery = At relatively extended intervals after extinction,
conditioned fear returns. The magnitude of this “spontaneous recovery”
increases with the length of the interval.
• SR occurs as a function of time
o 7 days rest following extinction, prior to recovery test
o No rest condition = recovery test occurred shortly after extinction
period
o Results → spontaneous recovery occurs more in 7 days rest prior to
test
Rapid acquisition
• Spontaneous recovery → rapid reacquisition
• Reacquisition of extinguished response CS → US
find more resources at oneclass.com
find more resources at oneclass.com
• Savings in re-learning by rapidly acquiring it
• This indicates that extinction involves new learning
Extinction involves new learning
• Extinction does not involve unlearning, rather it involves new learning
• Operant extinction leaves the individual with two available options:
o 1. Give up
o 2. Try some other means → reinforcement
• Classical extinction leaves the CS with two available meanings:
o 1. CS → US (excitatory conditioning)
o 2. CS → no US (inhibitory conditioning)
• How to resolve the US dilemma → look to the context for clues as to which
association is in play
Following extinction two opposing expectancies have been created:
• 1. Positive expectancy was acquired during reinforcement (CS-US) and a
• 2. Negative expectancy was acquired during non-reinforcement (CS-no
US)
Resurgence – only applies to operant behaviours
• occurs when behaviour that used to produce reinforcement in the past
reappears
• Sniffy → resurgence of earlier steps from shaping phase – as they had been
reinforced in past
• E.g. in training a porpoise name Hou, Pryor (1991) found that if the current
behaviour was not reinforced, Hou would run through her entire repertoire of
previously learned stunts
• What causes resurgence?
o It could be the stimuli that precede the beginning of the training
session (antecedents) that serve to elicit anticipatory responses or
expectations that then prompt the response: i.e. contextual cues create
the expectation of the chance to make the response and receive the Rf.
o Cat → diet but every time he sees his bowl – it reminds him of happier
days
Problems from a clinical perspective:
• The difference between resurgence and spontaneous recovery is that in
spontaneous recovery the behaviour that reappears is the behaviour that is
undergoing extinction
• Whereas with resurgence the behaviour that reappears can be some other
behaviour thatà reinforcement in the past
→ Extinction is a very frustrating experience – akin to punishment
Extinction as punishment
• Rats used to getting 256 pellets, and suddenly they only received 16 pellets →
clearly they were not happy and weren’t going to run as fast as a consequence
– this felt like a punishment
• Reduction in expected reward can be viewed as a form of punishment
find more resources at oneclass.com
find more resources at oneclass.com
Pleasantness of outcome is all relative and can be plotted on a scale of incentive value
• If experience with an attractive (256 pellets) incentive reduces the incentive
value of all lesser incentives (64 & 16 pellets), then exposure to 256 pellets
should reduce the incentive value not only of 16 pellets but also of 0 pellets. If
non-reward starts with an incentive value of zero, however, any reduction can
only mean that its incentive value becomes negative, so it would effectively be
aversive e.g. like a mild shock!
• Non-reward after reinforcement, in other words, would effectively be a form
of punishment
Extinction can be seen as frustration
• Extinction burst → early in an extinction period, there is often a temporary
increase in the frequency and intensity of responding
• For example, when we put money in a vending machine and push the button
and nothing happens
• For example – don’t walk signal when waiting to cross the road (e.g. VI 2 min
schedule)
o If the signal does not change after 3 mins then press the button more
frequently and forcefully
Extinction: side effects
• Increase in variability → the organism may change the way that it is
performing the response
• E.g. the key normally used to unlock your front door doesn’t work. You begin
to hold the key differently, not push it all the way in, turn the key upside down
etc. eventually you may get it to work
• “Variability in behaviour provides the means by which a totally new
behaviour, never peroformed by an individual before, can gradually be
developed”
• Important not to perseverate on a unproductive response if the reinforcer isn’t
coming
Extinction increases variability in behaviour
• Many studies have documented increased variability in extinction
• In Neuringer et al.’s study 2 levers and a key were present and rats in Group
Variety were required to vary the sequence of their behaviours in acquisition
→ Rf
• In extinction as response rates decrease, variability in types of response
increases
Extinction: frustration is energising
• Amsel and Roussel (1952): frustration → drives behaviour – they removed
food from goal box 1, and found that this frustration made the rats run faster to
goal box 2
• Run faster when frustrated – running faster as they learn that no reward is
available on F trials in 1st goal box
Extinction: side effects
• Frustration → aggression
find more resources at oneclass.com
find more resources at oneclass.com