The Signaling of Significant Events

174
Amount of salivation (CR)
Chapter 5
900
950
1,000
(CS)
1,050
1,100
Sound of buzzer (hertz)
FIGURE
5.4
Stimulus Generalization
The strength of a conditioned response
(CR) is greatest when the original conditioned stimulus (CS) occurs. However,
some version of the CR is also triggered by
stimuli that closely resemble the CS. Here,
the CS is the sound of a buzzer at a frequency of 1,000 hertz (Hz), and the CR is
salivation. Notice that the CR generalizes
well to stimuli at 990 or 1,010 Hz but that
it gets weaker and weaker as the buzzer
sounds less and less similar to the CS.
greater the similarity between a new stimulus and the original conditioned stimulus is,
the stronger the conditioned response will be. If the person was bitten by a small, curlyhaired dog, fear responses would be strongest to other small dogs with similar types of
hair. Figure 5.4 shows an example involving sounds.
Stimulus generalization has some obvious advantages. For example, it is important
for survival that, if you get sick after drinking sour-smelling milk, you now avoid dairy
products that have a similar odor. Generalization would be a problem if it had no limits, however. You would probably be justifiably frightened if you found a lion in your
living room, but imagine how disruptive it would be if your fear response generalized
so widely that you were panicked by the sight of lions on TV or even by the word lion
in a book.
Stimulus generalization does not run wild, because it is usually balanced by a
process called stimulus discrimination. Through stimulus discrimination, we learn
to make distinctions among similar stimuli. Many parents find that the sound of their
own baby whimpering soon becomes a conditioned stimulus, triggering a conditioned
response that wakes them up. That conditioned response may not occur if a visiting
friend’s baby whimpers.
The Signaling of Significant Events
Early research suggested that classical conditioning involves nothing more than automatic
associations that allow one stimulus (the conditioned stimulus, or CS) to substitute for
another (the unconditioned stimulus, or UCS) in triggering a reflex. That view has turned
out to be too simplified. For example, a rat’s unconditioned, reflexive response to a mild
shock (UCS) will be flinching and jumping. But at the sound of a tone (CS) that always
precedes shock, the animal’s conditioned response will not be to flinch and jump but to
freeze—much as it would if threatened by a predator (Domjan, 2005). In other words,
classical conditioning involves more than the appearance of robot-like, reflexive responses.
Many psychologists now believe that classical conditioning provides a means through
which people, and some other animals, develop expectations and other mental representations of the relationships between events in their environment (Shanks, 1995). These
representations aid in adaptation and survival. When two events repeatedly take place
together, we can predict that one will occur based on what we know about the other.
Baby Jeffrey predicted his feeding from hearing his mother’s footsteps. You have learned
that a clear blue sky means dry weather, that too little sleep makes you irritable, that you
can reach someone on the telephone by pressing certain buttons, and that yelling orders
motivates some people and angers others.
What determines whether conditioned responses are learned? In general, these
responses develop when one event signals the appearance of another. Other important
factors are the timing, predictability, and intensity of the unconditioned stimulus, as
well as the amount of attention that is devoted to the conditioned stimulus and how
prepared an organism is to associate paired events.
If your instructor always dismisses class at 9:59 and a bell rings at 10:00, the
bell cannot prepare you for the dismissal. It comes too late to be a useful signal. For the
same reason, classical conditioning works best when the conditioned stimulus comes
before the unconditioned stimulus. This arrangement makes sense for adaptation and
survival. Usually, the presence of food, predators, or other significant stimuli is most reliably signaled by smells, sounds, or other events that come just before their appearance
(Einhorn & Hogarth, 1982). So it is logical that the brain should be “wired” to form
associations most easily between things that occur at about the same time. How close
together do they have to be? There is no single “best” interval for every situation. Classical conditioning can occur when the interval between the CS and the UCS is less than
a second or more than a minute. It all depends on the particular CS, UCS, and UCR
that are involved (Longo, Klempay, & Bitterman, 1964; Ross & Ross, 1971). However,
classical conditioning will always be weaker if the interval between the CS and the UCS
is longer than what is ideal for the stimuli and responses in a given situation.
Timing
stimulus discrimination A process
through which people learn to differentiate among similar stimuli and respond
appropriately to each one.
Learning
Classical Conditioning: Learning Signals and Associations
175
It is not enough for the CS merely to come before the UCS. Suppose
your dogs, Moxie and Fang, have very different personalities. When Moxie growls, she
sometimes bites, but sometimes she doesn’t. Fang growls only before biting. Your conditioned fear response to Moxie’s growl will probably occur slowly, because her growl
does not reliably signal the danger of a bite. However, you are likely to quickly develop
a classically conditioned fear response to Fang’s growl. It always means that you are in
danger of being bitten. Classical conditioning proceeds most rapidly when the CS
always signals the UCS and only the UCS. Even if both dogs provide the same number
of pairings of the CS (growl) and the UCS (bite), it is only with Fang that the CS reliably predicts the UCS (Rescorla, 1968).
Predictability
Intensity A conditioned response will be learned more rapidly if the UCS is strong
than if it is weak. For example, a CS that acts as a predictive signal will be more rapidly
associated with a strong shock (UCS) than with a weak one. As with the importance of
timing, the effect of signal strength on classical conditioning makes adaptive sense. It is
more important to be prepared for major events than for those that have little impact.
Attention In Pavlov’s laboratory, just one conditioned stimulus, a tone, was linked
to an unconditioned stimulus (meat powder). In the natural environment, a wide variety of stimuli might be present just before a UCS occurs. Suppose you are at the beach.
You’re eating a hot dog, reading a magazine, listening to your favorite CD, digging your
toes in the warm sand, and enjoying the smell of suntan lotion when you are stung by
a bee. Which of these stimuli is most likely to become a conditioned stimulus that
might later trigger discomfort? It depends partly on where you were focusing your
attention at the moment you were stung. The stimulus you most closely attended to—
the one you most fully perceived—is most likely to become a CS. In general, loud tones,
bright lights, and other intense stimuli tend to get extra attention, so they are the ones
most rapidly associated with an unconditioned stimulus.
TASTE AVERSIONS Humans can
develop classically conditioned taste
aversions, even to preferred foods. For
example, Ilene Bernstein (1978) gave one
group of cancer patients Mapletoff ice
cream an hour before they received
nausea-provoking chemotherapy. A second group ate this same kind of ice cream
on a day they did not receive chemotherapy. A third group got no ice cream. Five
months later, the patients were asked to
taste several ice cream flavors. Those
who had never tasted Mapletoff and
those who had not eaten it in association
with chemotherapy chose it as their
favorite. Those who had eaten Mapletoff
before receiving chemotherapy found it
distasteful.
Biopreparedness Certain kinds of signals or events are especially likely to become
associated with other signals or events (Logue, 1985). So which beach stimulus becomes
a conditioned stimulus for fear will depend not only on attention but also on whether
the stimulus is a sight, a sound, or a taste and what kind of unconditioned stimulus
follows it. The apparently natural tendency for certain events to become linked suggests that organisms are “biologically prepared” or “genetically tuned” to develop certain conditioned associations.
The most dramatic example of this biopreparedness phenomenon is seen in conditioned taste aversions. In one study, rats were either shocked or made nauseous in the
presence of a light, a buzzer, and flavored water. The rats formed only certain conditioned associations. Animals that had been shocked developed a conditioned fear
response to the light and the buzzer, but not to the flavored water. Those that had been
made nauseous developed a conditioned avoidance of the flavored water, but they
showed no particular response to the light or buzzer (Garcia & Koelling, 1966). These
results reflect an adaptive process. Nausea is more likely to be caused by something we
eat or drink than by a noise or a light. So nausea is more likely to become a conditioned response to an internal stimulus, such as a flavor, than to an external stimulus.
In contrast, the sudden pain of a shock is more likely to have been caused by an external stimulus, so it makes evolutionary sense that the organism should be “tuned” to
associate shock or sudden pain with a sight or sound.
Conditioned taste aversion shows that for certain kinds of stimuli, classical conditioning can occur even when there is a long delay between the CS (taste) and the UCS
(sickness). Poisons do not usually produce their effects for many minutes or hours, but
people who have experienced food poisoning may never again eat the type of food that
made them ill. Organisms that are biologically prepared to link taste signals with illness, even a delayed illness, are more likely to survive than organisms not so prepared.
Other evidence for biopreparedness comes from research showing that people are
much more likely to develop a conditioned fear of harmless dogs or nonpoisonous