Download Chap9bAlt

PSY 402
Theories of Learning
Chapter 9 – Motivation
Hull’s Response

Spence modified Hull’s drive theory to
include findings of incentive motivation.




K was added to account for incentive.
Behavior strength = D x H x K
Drive is innate and internal, incentive is
learned and external.
Drive pushes behavior, incentive pulls it.
Fractional Anticipatory Goal Reactions


The idea of a motive seemed mentalistic – how can a
behavioristic theory explain expectations and goals?
rG-sG mechanism -- Intermediate states between the
initial behavior and the goal are chained together by
associations (classical conditioning).



RG = goal reaction or response (capital R)
rG = association of goal box with goal reaction (small r)
sG = similarity between start and goal box evoked rG &
salivation, which becomes a stimulus motivating response
Frustration


Amsel extended the idea of rG-sG mechanisms
to negative contrast (explained by the
mentalistic concept of frustration).
Expectation of a big reward (RG) but receipt
of a small reward results in frustration (RF).


The size of RF is the discrepancy between
previous rewards and the current reward size.
rF also becomes generalized to the goal box and
start box to demotivate (reduce responding).
Paradoxical Reward Effects

In some situations, reward seems to weaken,
not strengthen responding.



Negative contrast is one example – a perfectly
good reward fails to motivate responding.
Magnitude of reinforcement effect -reinforcement with a large reward leads to faster
extinction than reinforcement with a small one.
Overlearning extinction effect – many rewarded
trials extinguish faster than a few rewarded trials.
Effects on Intrinsic Motivation

Rewarding a behavior that was previously
performed for intrinsic reasons (internally
motivated) leads to reduced behavior.



Preschoolers expecting reward drew less.
“Punished by reward” phenomenon – some
suggest that rewarding creative activity may
hurt it.
Reward effects are relative and complex
(affected by many factors).
Skilled Use of Reward

Performance decreases in situations where:





Rewards were tangible (money)
Announced ahead of time
Given in a way that was not dependent on
performance.
Reward is perceived as a way to manipulate or
control someone.
Performance increases when verbal praise is
given, when reward is unexpected, relevant.
Persistence of Behavior

Partial reinforcement extinction effect (PREE)
– behavior is more resistant to extinction
when it is reinforcement intermittently.



Continuous reinforcement – acquired faster and
extinguished faster.
Partial reinforcement (50%, VR-2) – acquired
slower and extinguished slower.
Persistence generalizes to other tasks –
learned industriousness.
Theories about PREE

Amsel’s frustration theory -- animals learn to
respond in the presence of frustration.


Frustration attaches to the stimulus (sF) after an
unrewarded trial, then responding to it is
rewarded.
Capaldi’s sequential theory – the rat has a
memory of responding after a non-reinforced
trial.

It responds because the stimulus is familiar.
New Understanding of Extinction


Extinction involves new learning that is
dependent on context.
Research on PREE suggests also that:


Extinction occurs when generalization from
acquisition trials stops (generalization decrement)
Frustration occurs and reduces motivation when
an expected reward does not occur, but is not
needed to explain PREE.
Problems with rG-sG

If classically conditioned associations
motivate behavior, they should correlate with
responding, but they don’t.



Concurrent measurement studies showed little
correlation between salivation and responding
rewarded by food.
In studies of shock avoidance, fear (rE) showed
little correlation with vigor of avoidance.
Central states not peripheral should be measured.
Transfer of Control


Increasing or decreasing the intensity of an
internal state (expectancy) should affect
behavior.
Pavlovian-Instrumental Transfer.



An instrumental behavior and a CS are both
learned separately.
When paired with each other, behavior should
increase.
Dogs in shuttle box showed this effect.
Effects of Pavlovian CS’s

Some of the effects of Pavlovian inhibitors on
instrumental responding occur because of
evoked fear.


Rat freezing due to fear interferes with avoidance.
Excitatory Pavlovian CS’s can activate an
entire system, not just a specific response.


In other contexts, the effect of the CS is specific
to the type of reward (food vs water).
The explanation for this is unclear.
Practical Applications

Classical conditioning (CS’s) are always
present and affect instrumental behavior.



Obsessive checking behavior is worse when
anxiety is heightened by a CS, better when less
fear is evoked.
CS’s associated with a drug increase the
motivation for drug-taking behavior.
CS’s evoke a system of responses one of
which may be instrumental behavior.
Opponent-Process Theory


Emotional after-reaction – an emotional
stimulus creates an initial response that is
followed by adaptation, then opposite
response.
With repeated exposure to the stimulus, the
pattern changes.


The primary affective response (a-process)
habituates.
The after-reaction (b-process) strengthens
Imprinting

Baby ducks initially respond only to moving
stimuli, but with repeated exposure will be
comforted by a stationary stimulus.


They follow it around due to sign tracking.
The imprinted stimulus (train) can calm an
upset duck but causes distress when removed,
even if the duck is already calm.

The distress is the b-process in attachment.
Effects of Repeated Exposure

Ducks become more attached to an imprinted
stimulus after repeated exposures.



Exposures must be massed (spaced close together
in time).
A single long exposure produces a distress-calling
after-reaction.
This is contrary to other classical conditioning
situations, where massed exposure is worse.
Drug Addictions

Withdrawal symptoms seem to be an
opponent-process elicited by a CS associated
with the drug (S*).


The CR is the b-process and it offsets the body’s
response to the drug itself.
The opponent process appears to be learned,
not innate.

It is problematic for the learning explanation that
the b-process increases with massed exposure.