Download Chapter 7 — Learning: How Nurture Changes Us

7
CHAPTER
Learning:
How Nurture
Changes Us
Classical Conditioning
Psychomythology: Chemical Transfer of Learning
in Planaria
000
000
Operant Conditioning
000
Cognitive Models of Learning
000
New Frontiers: Mirror Neurons and Observational
Learning
Biological Influences on Learning
000
000
Classical and Operant Conditioning in Pop Psychology:
Do Learning Fads Work?
000
Science in the News
000
Critical Points
000
References
000
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
Before Reading Further
Try your hand at the following four items. Just do your best.
1
Ivan Pavlov, the discoverer of classical conditioning, was well
known as a
(a) slow eater,
(b) fast walker,
(c) terrible cook,
(d) I have no earthly idea.
2
John B. Watson, the founder of behaviorism, was tossed out of Johns
Hopkins University for
(a) plagiarizing a journal article,
(b) stabbing one of his faculty colleagues,
(c) having an affair with his graduate student,
(d) I have no earthly idea.
3
Watson also believed that parents should do what to their children
before bedtime?
(a) spank them,
(b) kiss them on the cheek,
(c) shake their hands,
(d) I have no earthly idea.
4
As a college student, B. F. Skinner, the founder of radical
ehaviorism, once spread a false rumor that which of the following
individuals was coming to campus?
(a) silent movie comedian Charlie Chaplin,
(b) psychoanalyst Sigmund Freud,
(c) President Theodore Roosevelt,
(d) I have no earthly idea.
Now, read the following paragraph.
The three most famous figures in the psychology of learning were each colorful
characters in their own way. The discoverer of classical conditioning, Ivan Pavlov,
was a notoriously compulsive fellow. He ate lunch every day at precisely 12 noon,
went to bed at exactly the same time every night, and departed St. Petersburg,
Russia, for vacation the same day every year. Pavlov was also such a rapid walker
than his wife frequently had to run frantically to keep up with him. The life of the
founder of behaviorism, John B. Watson, was rocked with scandal. Despite becoming
one of the world’s most famous psychologists, he was unceremoniously booted out
of Johns Hopkins University for having an affair with his graduate student, Rosalie
Rayner. Watson also had rather unusual ideas about parenting; for example, he
believed that all parents should shake hands with their children before bedtime.
B.F. Skinner, the founder of radical behaviorism, was something of a prankster in his
undergraduate years at Hamilton College in New York. He and a friend once spread
a false rumor that comedian Charlie Chaplin was coming to campus. This rumor
nearly provoked a riot when Chaplin didn’t materialize as expected.
Now go back and try again to answer the four questions at the beginning of this
chapter.
102
BEFORE READING FURTHER
If
you got more questions right the second than the first time around - and the odds
are high that you did - then you’ve experienced something that we all take for granted
much of the time: learning (the answers in order, by the way, are b, c, c, and a). By learning,
we mean a change in an organism’s behavior or thought as a result of experience. As we
discovered in Chapter 4, when we learn our brains also change along with experience.
Remarkably, your brain is physically different now than it was just a few minutes ago,
because it underwent chemical changes that allowed you to learn novel facts. How did it
happen? In this chapter, we’ll find out.
Learning lies at the heart of just about every area of psychology. As we discovered in
Chapter 1 when we crossed paths with the nature-nurture debate, virtually all behaviors are
a complex stew of genetic predispositions and learning. Indeed, without learning, we’d be
pretty much useless as human beings; we’d be unable to walk, talk, drive a car, ask someone
out for a date, or for that matter, read an introductory psychology textbook chapter about
learning.
Psychologists have long debated, and continue to debate, the question of how many
types are learning there are. We’re not going to try to settle this controversy here, so we’ll
instead review the most important types of learning. We’ll start with the most basic.
Before we do, though, we’d like you to place your brain “on pause” from this chapter
for just a few moments. After reading this sentence, put down your pen or highlighter,
close your eyes, and attend to several things that you almost never notice: the soft buzzing
of the lights in the room, the feel of your clothing against your skin, the sensation of your
tongue on your teeth or lips, perhaps even the faint smell of an unfamiliar odor wafting in
the air. Unless someone draws our attention to these stimuli, we don’t even realize they’re
there, because we’ve learned to ignore them. Habituation refers to the process by which
we respond less strongly over time to repeated stimuli. It helps to explain why loud snorers
can sleep through the night peacefully while keeping their irritated roommates wide awake
for hours. Chronic snorers have become so accustomed to the sound of their snores that
they no longer notice it.
Habituation is the simplest form of learning, and we can find it even in the lowly
single-celled amoeba that we see in our microscope. Shine a light on an amoeba and it will
contract into a ball. But keep shining this light on it, and soon it will resume its normal
activities, like swimming around on your microscope slide. Habituation is probably the
first form of learning to emerge in humans. We can demonstrate habituation in unborn
fetuses as young as 32 weeks old by applying a vibratory stimulator (which produces small
but rapid shaking motions) to the mother’s stomach. At first, the fetus jerks around in
response to the stimulus, but after repeated vibrations it stops moving (Morokuma et al.,
2004). What was first a shock to the fetus’ system later became a mere annoyance that it
could safely ignore.
In research that earned him the Nobel Prize in 2000, neurophysiologist Eric Kandel
uncovered the biological mechanism of habituation of Aplysia, a decidedly odd-looking
sea slug that’s about 5 inches long. If you prick an Aplysia on a certain part of its body, it
retracts its gill in a kind of defensive maneuver, much like a timid dog that pulls its paw
away when you touch it. Yet if you touch Aplysia in the same spot repeatedly, it begins
to ignore you. Kandel and his colleagues found that habituation was accompanied by
a progressive decrease in release of the neurotransmitter serotonin (see Chapter 4) at
the Aplysia’s synapses (Siegelbaum, Camardo, & Kandel, 1982). This discover helped
psychologists to begin to unravel the neural bases of learning (Photograph 1).
Habituation makes good evolutionary sense. We wouldn’t want to attend to every
tiny sensation that came across our mental radar screens, because most sensations pose
no threat to us. Yet we also wouldn’t want to habituate to stimuli that are potentially
dangerous, so not all repeated stimuli lead to habituation. Only those that we eventually
deem to be safe do. Indeed, some investigators have examined whether habituation occurs
103
Photograph 1
The habituation reflex of the sea snail,
Aplysia, has helped psychologists to
better understand the chemical bases of
the simplest form of learning.
habituation
the process by which we respond less
strongly over time to repeated stimuli
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
to strong stimuli, like very loud tones or painful electric shocks. As their dependent
measure they’ve often assessed a variable known as the skin conductance response (in the
old days this variable was known as the “galvanic skin response”), which is a measure of the
electrical conductivity of the fingertips. As the fingertips become moist as a consequence
of perspiration, they become better conductors of electricity. Because the skin conductance
response is generally a decent barometer of anxiety, some researchers use it in studies of
habituation. Most researchers have found that the skin conductance response habituates
much more slowly to strong than to weak stimuli. In fact, in the case of extremely strong
stimuli, like painful electric shocks, we often see no habituation at all, even over many trials
(Lykken, Iacono, Haroian, McGue, & Bouchard, 1988).
Indeed, in some cases repeated exposure to stimuli leads to sensitization – that is,
responding more strongly over time to repeated stimuli – rather than habituation.
Sensitization is most likely to occur when we find a stimulus to be dangerous, irritating, or
both (incidentally, Aplysia show sensitization as well as habituation). Have you ever had
the experience of trying to study when the person next to you was whispering, and the
whispering just kept getting more and more annoying to the point that you found yourself
on the verge of screaming (hopefully, that’s not happening while you’re reading this
chapter)? If so, you’ve experienced sensitization.
Classical Conditioning
The story of habituation could hardly be more straightforward. We experience
a stimulus, respond to it, and then stop responding to it after a while. We’ve learned
something significant, but we haven’t actually learned to forge connections between
two stimuli. Yet a great deal of learning in everyday life depends on associating one
thing with another. If we never learned to connect one stimulus, like the sight of an
apple, with another, like its smell, our worlds would remain what William James (1891)
called a “booming, buzzing confusion” – a world of disconnected sensory experiences.
Nevertheless, as a Russian scientist named Pavlov helped to show us, we forge such
connections all of the time in daily life.
BRITISH ASSOCIATIONISM
skin conductance response
a measure of the electrical conductivity
of the fingertips
sensitization
responding more strongly over time to
repeated stimuli
British Associationists
an entire school of thinkers who believed
that we acquire virtually all of our
knowledge by connecting one stimulus
with another
In the early history of psychology, an entire school of thinkers, called the British
Associationists, believed that we acquire virtually all of our knowledge by connecting
one stimulus with another: the sound of our mother’s face with her voice; the appearance
of a peach with its fuzzy feel; and the smell of a juicy steak with its mouth-watering taste.
One we form these associations, we need only recall one element of the pair to retrieve the
other. Indeed, even thinking about a sensation often triggers it. For example, as you read
the sentence before last you might have sensed the subtle flavor of a sirloin steak in your
mouth. The British Associationists, who included David Hartley (1707-1757) and John
Stuart Mill (1806-1873), believed that simple associations provided the mental building
blocks for more complex ideas. Your understanding of this paragraph, they surely would
have suggested, stems from thousands of linkages you’ve formed between the words in it
- like “juicy” and “steak” - and other words, which are in turn linked to simple concepts,
which are in turn linked to more complex concepts...and well, you get the picture.
One type of association is especially crucial in learning. This is the kind of association
we form when our heart starts to pound whenever we catch sight of the target of our
romantic infatuation, or when we begin to feel queasy whenever we think of blood. How
do we develop these associations? The Russian physiologist Ivan Pavlov showed how.
104
CLASSICAL CONDITIONING
PAVLOV’S DISCOVERIES
The history of science teaches us that many great discoveries arise from serendipity or
accident. Yet it takes a great scientist to capitalize on serendipitous observations; many
lesser mortals simply overlook such chance observations, regarding them as meaningless
flukes. As the great French microbiologist Louis Pasteur (who discovered the process of
pasteurizing milk) observed, “chance favors the prepared mind.” So it was with Pavlov’s
discoveries, which emerged from a set of unforeseen observations that were largely
unrelated to his central research interests.
Around the turn of the 20th century, Pavlov was busily studying the digestion of dogs.
Interestingly, it was actually his discoveries concerning digestion, not classical conditioning,
that earned Pavlov the Nobel Prize. Pavlov placed his dogs in a harness and inserted a
cannula, or collection tube, into their salivary glands to study dogs’ salivary (spitting)
responses to meat powder. Yet in the process of doing so, Pavlov observed something
unexpected. Specifically, he found that his dogs began salivating not to the meat powder
itself, but to previously neutral stimuli that had become associated with it, such as the
research assistants who brought in their food. Indeed, Pavlov’s dogs even salivated to the
footsteps of Pavlov’s research assistants approaching the laboratory room. The dogs seemed
to be anticipating the meat powder and responding to stimuli that signaled its arrival.
We call this process of association classical or respondent conditioning: a form of
learning in which animals come to respond to a previously neutral stimulus that had
been paired with another stimulus that elicits an automatic response. Yet Pavlov’s initial
observations were merely anecdotal. So like any good scientist, Pavlov decided to put his
informal observations to a more rigorous controlled test.
THE CLASSICAL CONDITIONING PARADIGM
Few findings in psychology are as replicable as classical conditioning. We can take the
classical conditioning paradigm, apply it to almost any animal with an intact nervous
system, and demonstrate it repeatedly without fail. If only all psychological findings were
so dependable!
Here’s how classical conditioning works, and here’s how Pavlov first demonstrated
it systematically (Figure 1). First, we begin with an initially neutral stimulus called the
conditioned stimulus or CS. It crucial to understand that this stimulus doesn’t elicit
much, if any, response from the organism. Interestingly, the term “conditioned stimulus” is
apparently a mistranslation from the original Russian. Pavlov actually referred to it as the
conditional stimulus, because the animal’s response to it is conditional (that is, dependent)
on learning. In the case of Pavlov’s dogs, the CS was usually a metronome (a clicking
pendulum that keeps time), although Pavlov also found that tuning forks and whistles
did a serviceable job as well. Incidentally, many introductory psychology textbooks refer
incorrectly to Pavlov’s CS as a bell. In fact, when Western scientists first visited Pavlov’s
Moscow laboratory in the 1990s, a bell was nowhere to be found.
Then, we pair the CS – again, in Pavlov’s case the sound of a metronome - again and
again with an unconditioned stimulus or UCS. This term was “unconditional stimulus”
in the original Russian, because the animal responds to it unconditionally – that is, all of
the time. In the case of Pavlov’s dogs, the UCS was the meat powder. The UCS elicits an
automatic, reflexive response that we call the unconditioned response or UCR, in this
case salivation. The key point is that the animal doesn’t need to learn to respond to the
UCS with the UCR. The UCR occurs without any training at all, because the response is a
product of nature, not nurture.
After repeated pairing of the CS and UCS, Pavlov observed something remarkable. If he
now presented the CS (the metronome) alone, it elicited a response, namely salivation. We
call this new response the conditioned response or CR. Learning has occurred. The dog,
105
Replicability
classical or respondent conditioning
a form of learning in which animals
come to respond to a previously
neutral stimulus that had been paired
with another stimulus that elicits an
automatic response
conditioned stimulus or CS
a neutral stimulus, eliciting little, if any,
response from the organism
unconditioned stimulus or UCS
a stimulus that is always responded to
unconditioned response or UCR
an automatic, reflexive response
conditioned response or CR
a product of nurture, not nature
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
CLASSICAL CONDITIONING
BEFORE CONDITIONING
No salivation
response
Tone
Food powder(CS)
UCR (salivation)
DURING CONDITIONING
Photograph 2
The Rolling Stones may the only major
rock band to accurately describe the
process of classical conditioning. One of
their well-known songs features the lyrics: ”When you call my name, I salivate
like a Pavlov dog.”
ADAPTIVE VALUE OF CLASSICAL CONDITIONING
Tone (CS) + Food powder(CS)
UCR (salivation)
AFTER CONDITIONING
Tone (CS)
CR (salivation)
Figure 1 Pavlov’s classical conditioning scheme, with a tuning fork as the CS and meat
powder as the UCS.
Photograph 3
The film “A Clockwork Orange” portrays
a terrifying example of aversive conditioning, in which violent films (CSs)
were paired with a nausea-eliciting drug
(the UCS). The conditioning succeeded
– temporarily, that is.
aversive conditioning
classical conditioning using an
unpleasant UCS
the breakdown of alcohol, causing alcohol to go directly to the “drinker’s head” (Fuller
& Roth, 1979). As a consequence, upon drinking, disulfiram triggers an exceedingly
unpleasant “flushing response” characterized by a red face, nausea, vomiting, and dizziness
(as we’ll learn in Chapter 18, some individuals, especially of Asian descent, experience
this flushing response naturally as a result of a genetic mutation). The idea here relies on
classical conditioning: by pairing the CS (alcohol) with the UCS (a nausea-inducing drug),
the CS itself eventually comes to elicit the CR (nausea). Nevertheless, in controlled studies
the evidence for disulfiram’s effectiveness has been disappointing. Although it seems to
work for some alcoholics, the big problem is that many individuals don’t comply with
the treatment. That is, rather than not drinking, many simply avoid taking disulfiram
(MacKillop, Lisman, Weinstein, & Rosenbaum, 2003). (Photograph 4).
which previously did nothing when it heard the metronome except perhaps turn toward it,
now salivates when it hears the metronome. The CR, in contrast to the UCR, is a product
of nurture, not nature.
In most cases, the CR is fairly similar to the UCR. But it is rarely identical; for example,
Pavlov found that the amount of salivation in the CR was less than that in the UCR (we’ll
later learn about one strange case in which the UCR is actually the opposite of the CR).
(Photograph 2).
AVERSIVE CONDITIONING
Incidentally, we can classically condition organisms not only to positive UCSs, like food,
but to negative UCSs, like pain or nausea. While in graduate school, one of the authors of
your book was a victim of such aversive conditioning, that is, classical conditioning using
an unpleasant UCS (Emmelkamp & Kamphuis, 2005). Back then, she videotaped infants
as part of a study of language development. Prior to each participant’s arrival, she walked
over to the video recorder and pushed the eject button to insert a tape. Perhaps because the
air in the room was dry, the recorder consistently gave her nasty electric shock whenever
she pushed the eject button. Because the shock usually occurred in the midst of a busy
series of events in anticipation of the participant’s arrival, she barely paid any attention to
it. Nevertheless, one day she reached to push the eject button and noticed herself jerking
her hand back reflexively as it approached the button.
Stanley Kubrick’s 1971 film, A Clockwork Orange, provides an unforgettable example of
aversive conditioning involving the main character, Alexander de Large, portrayed by actor
Malcolm McDowell. de Large’s prison captors, who hoped to eradicate his bloodthirsty
lust for violence, forced him to watch film clips of aggressive individuals, like the members
of Hitler’s army marching in unison, while experiencing nausea induced by injections of a
serum. The aversive conditioning worked – but only for a while (Photograph 3).
In real-life rather than on the big screen, psychologists have applied aversive
conditioning with mixed success to the treatment of alcoholism. They’ve sometimes given
alcoholics a medication known as disulfiram (better known as Antabause), which blocks
106
Why should we care about classical conditioning? Because it’s evolutionarily essential
for us. Without classical conditioning, we couldn’t develop physiological associations
to stimuli that signaled biologically important events, like things that we want to eat or
want to eat us. Many of the physiological responses we display in classical conditioning
are biologically adaptive. Salivation, for example, aids in the digestion of food. Although
skin conductance responses aren’t terribly adaptive for us today, they were to our primate
ancestors (Stern, Ray, & Davis, 1980), who found that moist fingers and toes came in handy
for grasping tree limbs while escaping from predators (slightly wet fingertips help us to
adhere to things, as you’ll discover if you moisten the tip of your index finger while turning
to the next page of this book).
Moreover, without classical conditioning, we can’t learn many important associations.
Take the example of people with psychopathic personalities, whom we encountered in
Chapter 2 and will revisit in Chapter 18. As you might recall, such people (often called
“psychopaths” for short) tend to be guiltless, callous, and dishonest, and they often commit
impulsive crimes. Half a century ago, David Lykken (1957) found that psychopaths
tend to show extremely weak classical conditioning to painful electric shocks. Lykken
first presented psychopaths and nonpsychopaths with repeated tones (CSs), followed
almost immediately by electric shocks (UCSs) to the fingertips (the shocks really sting
there, because our fingertips are brimming with nerve endings). Not surprisingly, both
psychopaths and nonpsychopaths showed a whopping skin conductance response (UCR)
to the shocks. Lykken repeated this process over and over: tone-shock, tone-shock, toneshock, and on and on. Then, he presented his participants with the tones (CSs) alone,
and measured their skin conductance responses, which were now the CRs. In contrast to
nonpsychopaths, who showed pronounced skin conductance responses (CRs) to the tones
alone, psychopaths exhibited weak skin conductance responses or none at all.
Lykken’s finding may help to explain why psychopaths don’t learn from punishment
and often find themselves in trouble with the law again and again (Hare, 2003; Newman &
Kosson, 1986): they don’t develop conditioned fear to signals of punishment. Whereas the
rest of us may shudder in fear at the mere sight of a police car appearing suddenly in our
rear-view mirror, psychopaths typically shrug off this visual stimulus with nonchalance.
As a consequence of their indifference to signals of threat, psychopaths don’t inhibit
irresponsible and even criminal behaviors that the rest of us do (Fowles, 1987; Lykken,
1995). (Photograph 5).
Photograph 4
Dilsufiram (Antabuse) is sometimes
used as a treatment for alcoholism. It
relies on classical conditioning principles.
Photograph 5 TBD
Perhaps because they don’t easily
acquire classically conditioned fear
reactions, many psychopaths are slow to
learn from punishment.
ACQUISITION, EXTINCTION, SPONTANEOUS RECOVERY
Pavlov noted, and others have since confirmed, that classical conditioning occurs in three
major phases. We’ll discuss each in turn.
Acquisition. The first phase of classical conditioning is acquisition: the process by which
107
acquisition
the process by which we gradually learn
the CR
1 2 3. . .
Trials
. . .n
Figure 2 Acquisition. As the CS and UCS
are paired repeatedly, the CR increases
progressively in strength.
Amount of saliva (cc)
6
5
4
3
2
1
0
1
2
3 4 5
Unreinforced trials
6
Figure 3 Extinction. When, following
conditioning, the CS is presented by
itself repeatedly, the CR gradually
disappears.
acquisition
the process by which we gradually learn
the CR
delay conditioning
in which the CS precedes the UCS, but
overlaps with it in time
trace conditioning
in which the CS precedes the UCS and
is separated from it by a short period
of time
simultaneous conditioning
in which CS and UCS occur at precisely
the same time
backward conditioning
in which we present the UCS prior to
the CS
extinction
in which the CR decreases in magnitude
and eventually disappears
spontaneous recovery
the release of the CR on presentation of
the CS
we gradually learn the CR. If you look at Figure 2, you’ll see that as the CS and UCS are
paired over and over again, the CR increases progressively in strength.
There are four different methods of pairing the CS and UCS. The first is delay
conditioning, in which the CS precedes the UCS, but overlaps with it in time. In Pavlov’s
conditioning paradigm, this means that the tone came first but stayed on while the dog
encountered the meat powder. This kind of learning generally works well, especially when
the delay between CS and onset of the UCS is brief. The optimal delay varies depending
on what stimuli we’re pairing, but for many stimuli it’s about half of a second. The second
means of pairing stimuli is trace conditioning, in which the CS precedes the UCS and is
separated from it by a short period of time. Trace conditioning also works well, although
it’s generally not as effective as delay conditioning. Many students understandably find
the difference between delay and trace conditioning to be befuddling, because the terms
“delay” and “trace” lend themselves to confusion. We can think of it this way: in delay
conditioning, the CS and UCS overlap but are staggered in their presentation, whereas in
trace conditioning the CS and UCS show no overlap at all, so the organism must rely on a
“memory trace” - hence the name - of the original CS.
The third method of pairing stimuli is simultaneous conditioning, in which CS and
UCS occur at precisely the same time. This kind of conditioning usually works rather
poorly. Fourth and finally, there’s backward conditioning, in which we present the UCS
prior to the CS. Backward conditioning almost always works the most poorly of all forms
of pairing. Although it sometimes results in learning (Spetch, Wilkie, & Pinel, 1981), it
sometimes produces no learning at all (Lieberman, 1990). The ineffectiveness of backward
conditioning makes good evolutionary sense, because in this kind of conditioning, the CS
doesn’t signal anything. If you always encountered a stop sign immediately after – rather
than before – you drove into a crowded 4-way intersection, you’d be unlikely to learn how
to avoid traffic accidents, because the signal comes after the threat to life and limb.
Extinction. What happens when, following the repeated pairing of CS and UCS, we
present the CS by itself, without the UCS? As we’ve seen, the animal initially produces
a CR. But what if we keep repeating the process by presenting the CS again and
again by itself? Eventually, we observe a process called extinction: the CR decreases in
magnitude and eventually disappears (Figure 3). Following conditioning, after numerous
presentations of the metronome without meat power, Pavlov’s dogs eventually stopped
salivating. No meat, no spit. Extinction may also explain why the effectiveness of
Alexander de Large’s aversive conditioning treatment in A Clockwork Orange eventually
wore out. Without the UCS (nausea-inducing serum), the CS (violence) no longer signaled
the CR (nausea).
Most psychologists once believed that extinction is pretty much the same thing as
forgetting: the CR merely fades away over repeated trials, just as many memories decay
gradually in strength over time. Yet we now know that the truth is more complicated
and more interesting than that, because extinction turns out to be an active rather than a
passive process. That is, during extinction the new response, which in the case of Pavlov’s
dogs was the absence of salivation, gradually “writes over” or inhibits the CS, namely
salivation. The extinguished CS doesn’t vanish completely; it remains there all along,
still waiting in the wings. This is contrast to most forms of traditional forgetting, in
which the memory itself disappears. Interestingly, Pavlov had proposed this hypothesis
in this writings, although few people believed him at the time. How do we know that he
was right?
Spontaneous Recovery. We know it in several ways, but the most dramatic illustration
comes from the phenomenon of spontaneous recovery. If we wait a while following
extinction, and then present the CS again, something surprising happens. Like a phoenix
rising from the ashes, the CR spookily reappears! So the CR was indeed there all along, just
108
CLASSICAL CONDITIONING
“waiting” to be released following another presentation of the CS. For example, following
classical conditioning, Pavlov (1927) presented the CS (tone) again and again, thereby
completely extinguishing the CR (salivation). Yet when he waited two hours and then
presented the CS again, the CR (salivation) returned. The animal hadn’t really forgotten
the CR, just suppressed it.
Closely related to spontaneous recovery is the renewal effect, which occurs when
we extinguish a response in a setting different from that in which the animal acquired
it, but then restore the animal to the original setting. Upon doing so, the extinguished
response reappears (Bouton, 1994). The renewal effect may help to explain why people
with phobias - intense, irrational fears - that have been successfully treated sometimes
experience a sudden reappearance of their symptoms upon returning to the environment
in which they acquired their fears (Denniston, Chang, & Miller, 2003). Even though it may
lead to a return of phobias in some circumstances, the renewal effect is often adaptive. If
you’ve been bitten by a snake in one part of a forest, it makes good evolutionary sense to
experience fear when you find yourself there again, even years later. That same snake or his
slithery relatives may still be lying in wait next to the same tree.
60
Percent of trials on which CR
was made
Strength of response
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
40
20
400
1200
2000
Test stimuli (hertz)
Figure 4 A generalization gradient. The
more similar the CS is to the original CS,
the stronger the conditioning.
CRITICAL THINKING QUESTION
After the first Gulf War in 1992, in which the United States attacked Iraq following Saddam
Hussein’s invasion of Kuwait, a large number of Vietnam Veterans traumatized by that war
reported that their anxiety symptoms suddenly reappeared. How could the phenomenon
of spontaneous recovery help to explain the return of their symptoms?
STIMULUS GENERALIZATION AND DISCRIMINATION
As adaptive as it is, classical conditioning would be virtually useless if we couldn’t apply
it to novel stimuli. As the Greek philosopher Heraclitus reminded us, ”we never step into
the same river twice.” No two stimuli are entirely identical; even our friends look a tiny bit
different every time we see them. So we need to learn to respond to stimuli that differ from
those we originally encountered during conditioning.
Stimulus generalization. Pavlov found that following classical conditioning, his dogs
salivated not merely to the original metronome sound, but to sounds that were similar to
it. We call this phenomenon stimulus generalization: the process by which CSs similar,
but not identical, to the original CS elicit a CR. Stimulus generalization occurs along a
generalization gradient: the more similar to the original CS is the new CS, the stronger the
CR (Figure 4). For example, Pavlov found that his dogs showed their largest amount
of salivation to the original sound, with progressively less salivation to sounds that were
less and less similar to it in pitch. Similarly, if you’ve been in a scary car accident, you
may feel your heart pounding while watching stories about other car accidents on
television.
Stimulus generalization is adaptive, because it allows us to transfer what we’ve learned
to novel stimuli. Yet like all good things, we can take it too far. For example, people with
phobias are often afraid not merely of the original stimulus that may have triggered their
fear, but of stimuli only remotely resembling it. One of the authors of this book treated a
woman whose phobia of rats was so severe that she couldn’t even glance at a drawing of a
cute rat from a distance without experiencing marked discomfort. Some U.S. Vietnam War
veterans with posttraumatic stress disorder, a condition marked by severe anxiety reactions
resulting from exposure to a terrifying stressor, still - over 30 years later - hide under their
beds when they hear the sounds of distant helicopters in their home cities (Spitzer, Gibbon,
Skodol, Williams, & First, 1994). They’ve generalized their fears to stimuli – like traffic
helicopters – that pose no threat to them.
109
renewal effect
in which an extinguished response is
restored on return to an original setting
phobias
intense, irrational fears
stimulus generalization
the process by which CS’s similar, but not
identical, to the original CS elicit a CR
generalization gradient
measures strength of stimulus
generalizations - more similar to the
original CS the new CS, the stronger the
CR
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
60
40
20
400
1200
2000
Figure 5 TBD
Photograph 6
Upon returning to the United States,
most Vietnam veterans with heroin addictions quickly lost their addictions.
Photograph 7
Advertisers make copious use of the
principles of higher-order conditioning,
often pairing their products with image
of sexy and scantily clad models.
Stimulus discrimination is the flip side of the coin to stimulus generalization; it occurs
when we exhibit a CR to certain CSs, but not others. You can see stimulus discrimination
in Figure 5, which demonstrates that Pavlov’s dogs learned to respond less, or not at all,
to metronome sound that was markedly different from the original sound. Stimulus
discrimination also helps us to understand why we can enjoy scary movies. Although we
may sweat and hyperventilate just a bit while watching a great white shark chase after a
swimmer in Jaws or other waterlogged movie thrillers, we don’t respond nearly as strongly
as we would if that shark were pursuing us in a tank in our local aquarium. We’ve learned
to discriminate between a motion picture stimulus and a real-life, flesh-and-blood (forgive
the choice of words) stimulus.
Higher-order conditioning We’re not done with classical conditioning just yet, because
organisms also learn to develop conditioned associations to other CSs that are associated
with the original CS. Say, for example, that following Pavlov’s conditioning paradigm, we
paired a picture of a circle with the original CS, the tone. Eventually, the dog will salivate to
the circle as well as to the tone. This example illustrates the phenomenon of higher-order
conditioning: the process by which organisms develop classically conditioned responses to
CSs associated with the original CS (Gewirtz & Davis, 2000). As you might expect, secondorder conditioning - in which one new CS is paired with the original CS - tends to be
weaker than garden-variety classical conditioning, and third-order conditioning - in which
a third CS is paired with the second-order CS - is even weaker. Fourth-order conditioning
and beyond is often difficult or impossible.
Higher-order conditioning allows us to extend classical conditioning to a host of stimuli
in our environments. For example, it probably helps to explain why we feel thirsty after
someone merely says the word “coke” on a sweltering summer day. We had already come to
associate the sights, sounds, and smells of a Coca-cola with quenching our thirst, and we in
turn eventually came to associate the word “coke” with these other CSs.
The phenomenon of higher-order conditioning also helps us to explain some surprising
findings concerning addictions to heroin and other drugs. Many drug addictions are
shaped in part by higher-order classical conditioning, with the situational context in which
drugs are taken serving as a higher-order CS. For example, in the case of heroin addiction,
although the needle serves as the CS, the setting in which the addict injects the drug (for
example, alone versus with friends) is a higher-order CS. Behaviorists refer to these higherorder CSs as occasion setters, because they refer to the setting in which the CS occurs.
Although public perception has it that “breaking the grip” of heroin addiction is
difficult or impossible, research evidence suggests otherwise (Sullum, 2003). For example,
sociologist Lee Robins and her colleagues (Robins, Helzer, & Davis, 1975) examined 451
Vietnam veterans who had returned to the United States with cases of serious heroin
addiction. Although many mental health experts confidently predicted a rash of heroin
addiction upon Vietnam veterans’ return to America, this epidemic never materialized.
In fact, in Robins’ sample, 86% of heroin-addicted Vietnam veterans lost their addiction
shortly after returning to the United States. What had happened? Because the occasion
setters had changed from Vietnam to the United States, the veterans’ classically conditioned
responses to heroin extinguished (Photograph 6)
Applications of Classical Conditioning to Daily Life Classical conditioning applies to
myriad domains of everyday life. We’ll consider four such applications: advertising, the
acquisition of fears, the acquisition of fetishes, and disgust reactions.
higher-order conditioning
the process by which organisms develop
classically conditioned responses to CSs
associated with the original CS
Classical Conditioning and Advertising. Few people grasp the principles of classical
conditioning, especially higher-order classical conditioning, better than advertisers
(Photograph 7). By repeatedly pairing the sights and sounds of products with photographs
of handsome hairy hunks and scantily clad beauties, the marketing whizzes of Madison
110
CLASSICAL CONDITIONING
Avenue try to establish classically conditioned connections between their brands and
positive emotions. They do so for a good reason: it works.
For example, one researcher (Gorn, 1982) paired slides of either blue or beige pens
(the CSs) with music that participants had rated as either enjoyable or not enjoyable (the
UCSs). He then gave participants the opportunity to select a pen upon their departing the
lab. Whereas 79% of participants picked the pen that had been paired with music they
liked, only 30% picked the pen that had been paired with music they disliked. Another
researcher (Grossman, 1998) randomly assigned participants to two conditions: one in
which they saw a fictitious brand of mouthwash (the CS) paired repeatedly with pleasant
visual images, like a relaxing tropical scene (the UCSs), and a second in which they saw the
CS and UCSs jumbled up in random order. Compared with the random presentation of
CS and UCS, the repeated pairing between CS and UCS resulted in a significantly greater
preference for the mouthwash three weeks later.
Nevertheless, not all researchers who have paired products with pleasurable stimuli
have succeeded in demonstrating classical conditioning effects (Smith, 2002). For
example, two researchers (Gresham & Shimp, 1985) paired various products, like Coke,
Colgate toothpaste, and Grape Nuts cereal, with television commercials that had been
rated as generating pleasant, unpleasant, or neutral emotions. They failed to find much
evidence that these pairings affected participants’ preferences for the ads. Nevertheless,
their negative findings are open to an alternative explanation: latent inhibition. Latent
inhibition refers to the fact that when we’ve experienced a CS alone many times, it’s
difficult to classically condition it to another stimulus (Vaitl & Lipp, 1997). Because
Gresham and Shimp relied on brands with which participants were already familiar, their
negative findings may be attributable to latent inhibition. Indeed, when researchers have
used novel brands, they’ve generally been able to demonstrateD classical conditioning
effects (Stuart, Shimp, & Engle, 1987).
Photograph 8
Prior to Watson and Rayner’s classical
conditioning, Little Albert liked rabbits.
After classical conditioning, he reacted
with fear.
Ruling Out
Alternative
Hypithesis
The Acquisition of Fears. We’ve seen that classical conditioning can help to explain how
we develop preferences for certain things, like products advertised on television. But can
classical conditioning also help to explain the origins of abnormal behaviors, like phobias,
fetishes, and extreme disgust reactions? There’s considerable evidence that it can.
The Strange Tale of Little Albert. The person who first demonstrated this point was none
other than our old friend John B. Watson, the founder of behaviorism. In 1920, Watson
and his graduate student Rosalie Rayner performed what by any standard must be regarded
as one of the most ethically problematic studies in the history of psychology. Here’s what
they did.
Watson and Rayner (1920) began with a 9th month old infant who will forever be
enshrined in the psychological literature as Little Albert. Little Albert, it so happened, was
fond of furry little creatures, like white rats. Well, Watson and Rayner were about to change
that.
Watson and Rayner first allowed Little Albert to play with a rat. But only seconds
afterwards, Watson snuck up behind Albert and struck a gong with a steel hammer, creating
an ear-splitting noise and startling poor Little Albert out of his wits. After seven such
pairings of CS (rat) and UCS (sound from gong), Little Albert exhibited a CR (fear) to the
rat alone. This fear was still present when Watson and Rayner exposed Little Albert to the
rat five days later. Moreover, Watson and Rayner observed that Little Albert had become a
victim of stimulus generalization, because he had begun to fear not merely rats, but also a
rabbit, a dog, a furry coat and, to a lesser extent, a Santa Claus Mask and John B. Watson’s
hair. Fortunately, Little Albert also demonstrated at least some stimulus discrimination,
as he didn’t display much fear toward cotton balls or the hair of Dr. Watson’s research
assistants (Photograph 8).
Watson and Rayner’s Little Albert demonstration is only a case study. As we noted in
111
occasion setters
the setting in which the CS occurs
Latent inhibition
negative responses that are difficult
classically condition
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
Replicability
Chapter 3, case studies are extremely limited in the conclusions they allow; for example, we
can’t generalize from Little Albert’s case to the development of phobias in other children.
Nevertheless, the Little Albert case provides an existence proof (see Chapter 3) that classical
conditioning can produce phobia-like states in humans. Admittedly, not everyone has
successfully replicated Watson and Rayner’s findings (Harris, 1979; Jones, 1930). This fact
doesn’t mean that Watson and Rayner were wrong to claim that Little Albert’s fears arose
from classical conditioning, but it does imply that there may be more to developing phobias
than classical conditioning alone. We’ll come back to this point later in the chapter.
Incidentally, you might be wondering whatever became of Little Albert. As it turns
out, the Little Albert story is the source of numerous psychology urban legends (Gilovich,
1991). For example, some textbooks state that Watson and Rayner removed Little Albert’s
fear using deconditioning: pairing the CS (rat) with a pleasant UCS (such as food). This
happy Hollywood-style ending never happened (Harris, 1979). In fact, we’ll probably
never know Little Albert’s fate, because his mother, perhaps understandably fed up with
Watson and Raynor’s madcap experimental adventures, withdrew him from the study
about a month after it began, never to be heard from again. Needless to say, because
inducing a phobia-like condition in an infant raises a host of serious ethical questions,
Watson and Rayner’s Little Albert study would never get through a modern-day college or
university IRB (see Chapter 3).
Phobias. Do you suffer from a serious case of paraskavedekatriaphobia? If so, don’t be
overly concerned, because you aren’t the only person in the world who is mortally afraid of
Friday the 13th.
Because of stimulus generalization and higher-order conditioning, classical conditioning
is remarkably flexible. As a consequence, we can develop fears of many stimuli, although
certain phobias, such as those of snakes, spiders, heights, water, and blood, are considerably
more widespread than others (American Psychiatric Association, 2000). Other phobias,
like fear of being tickled by feathers (pteronophobia), fear of clowns (coulrophobia), fear
of flutes (aulophobia), and fear of bald people (peladophobia), are downright strange,
although such fears are rarely severe (see Table 1).
Alliumphobia: Fear of garlic
Iatrophobia: Fear of doctors
Arachibutyrophobia: Fear of peanut butter sticking
to the roof of your mouth
Lachanophobia: Fear of vegetables
Bathmophobia: Fear of stairs
Brontophobia: Fear of thunderstorms
Bufonaophobia: Fear of toads
Catoptrophobia: Fear of mirrors
Cynophobia: Fear of dogs
Elurophobia: Fear of cats
Epistaxiaophobia: Fear of nosebleeds
Melissophobia: Fear of bees
Ophidiophobia: Fear of snakes
Peladophobia: Fear of bald people
Pogonophobia: Fear of beards
Rhytiphobia: Fear of getting wrinkles
Samhainophobia: Fear of Halloween
Taphephobia: Fear of being buried alive
Xyrophobia: Fear of razors
Table 1 Phobias Galore illustrates just how enormously varied people’s fears can be. Many of these phobias
may be acquired at least partly by classical conditioning.
deconditioning
pairing the CS with a pleasant UCS
The good news is that we can not only develop phobias in part through classical
5:
conditioning, we can also eliminate them through classical conditioning. One of Watson’s
students, Mary Cover Jones, treated a 3 year old named Little Peter, who had a phobia of
rabbits. Fortunately, Little Peter’s story had a happier ending did Little Albert’s. Jones
(1924) successfully treated Peter’s fear by giving him a piece of his favorite candy each
112
THE BRAIN
she moved a white
rabbit closer to
him. Eventually, the
sight of the rabbit
came to elicit a
new CR: pleasure
rather than
fear. Modern day
psychotherapists,
although rarely
feeding their clients
candy, use similar
principles to eliminate
phobias. For example, they
may pair feared stimuli with
relaxation or other pleasurable
stimuli (Wolpe, 1990; see
Chapter 16).
Photograph 9
Fetishes appear to be acquired in part by means of
classical conditioning. Although the range of objects to
which fetishes are attached are enormous, most seem to
center around women’s sexuality.
Fetishes. Just as classical conditioning can lead us to develop fears of certain objects, it
may also lead us to develop sexual attractions to other objects. Indeed, there’s good reason
to believe that fetishism – sexual attraction to nonliving objects – often arises in part from
classical conditioning (Akins, 2004). The same goes for a related condition, partialism,
which is excessive sexual preoccupation with one part of the body to the exclusion of
others, such as the breasts, feet, legs, or buttocks.
For reasons that scientists do not fully understand, men have a virtual monopoly
on fetishes (American Psychiatric Association, 2000). Like phobias, fetishes can take a
remarkably wide variety of forms. Most fetishes involve stimuli associated with women’s
sexuality, like stockings, bras, and shoes. Nevertheless, others are far more difficult to
explain. For example, one man achieved sexual gratification from looking at British
SUV cars (de Silva & Pernet, 1992) and another from looking at flags (Lange, 1991).
(Photograph 9).
Some conditions apparently related to fetishes are even more unusual. For example,
individuals with apotemnophilia (or body integrity identity disorder) are almost constantly
preoccupied with the idea of becoming amputees, and sometimes even amputate their
own limbs to achieve that morbid goal. For certain individuals, the intense fascination
with becoming an amputee appears to contain a sexual component. For example, some
apotemnophilics appear to receive sexual gratification from thinking of themselves as
future amputees (First, 2005). Individuals with yet another condition called autoerotic
axphyxiation (or hypoxyphilia) derive sexual satisfaction from depriving themselves of
oxygen while engaged in sexual activity, such as masturbation or intercourse. Many
of them engage in strangulation, hanging, suffocation, or choking to achieve oxygen
deprivation, which appears to intensify the sexual experience. Tragically, a number
of individuals with this condition have died as a consequence of accidental oxygen
deprivation (Hucker & Blanchard, 1992).
Although the origins of human fetishes are controversial, Michael Domjan and his
colleagues were successful in classically conditioning fetishes in male Japanese quail. In one
study, they presented male quails with a cylindrical object made of terrycloth, followed by
a female quail with which they happily and promptly mated. After 30 such pairings, about
half of the male quail attempted to mate repeatedly with the cylindrical object when it was
presented alone (Koksal, Domjan, et al., 2004) (Photograph 10). This research doesn’t
demonstrate, of course, that classical conditioning explains the acquisition of fetishes in
humans; it’s an awfully long way from quails mating with terrycloth cylinders to humans
113
Photograph 10
The research of Michael Domjan and
his colleagues suggests that fetish-like
behaviors can be classically conditioned
in male quail.
fetishism
sexual attraction to nonliving objects
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
getting excited by shoes. Nevertheless, there’s good evidence that at least some people
acquire fetishes by repeatedly pairing neutral objects with sexual activity (Rachman &
Hodgson, 1968; Weinberg, Williams, & Calhan, 1995).
Photograph 11
This looks like a pretty ordinarily jacket.
Would you wear it if someone told you
that Adolph Hitler had previously worn
it?
Disgust reactions. Imagine that a researcher asked to you to try on a 70 year old sweater
that was perfectly preserved. No problem, right?
Well, let’s instead imagine that this researcher asked to you try on this jacket, but
informed you that Adolph Hitler had worn it. If you’re like most subjects in the studies of
Paul Rozin and his colleagues, you’d hesitate (D’Amato, 1998). (Photograph 11).
Indeed, Rozin (who has earned the nickname of “Dr. Disgust”) and his colleagues have
found that humans develop disgust reactions with surprising ease. In most cases, these
reactions are probably the product of classical conditioning, because CSs associated with
disgusting UCSs come to elicit disgust themselves. In many cases, these disgust reactions
appear to be tied to stimuli that are biologically important to us, such as animals or objects
that are dirty or potentially poisonous (Rozin & Fallon, 1987). For example, Rozin and his
colleagues found that subjects who happily gobble up a delicious piece of fudge suddenly
express a decided reluctance to do so when this piece of fudge is molded into the shape of
dog feces (Rozin, Millman, & Nemeroff, 1986).
In another study, Rozin and his collaborators asked participants to drink from two
glasses of water, both of which contained sugar (sucrose). In one case, the sucrose came
from a bottle arbitrarily labeled “sucrose”; in another, the sucrose came from a bottle
arbitrarily labeled “Sodium Cyanide, Poison.” The investigators told subjects that both
bottles were completely safe. Moreover, they even asked subjects to attach the arbitrary
labels on the two glasses themselves, thereby allowing them to select which label went
with which glass. Even so, subjects were extremely reluctant to drink from the glass that
contained the sucrose labeled as poisonous (Rozin, Markwith, & Ross, 1990). Participants’
responses in this study were irrational, but perhaps understandable: they were probably
relying on the heuristic “better safe than sorry.” Classical conditioning helps keep us safe,
even though it goes too far on occasion.
PSYCHOMYTHOLOGY
Chemical Transfer of Learning
in Planaria
You’ve probably heard that “you are what you eat,” but in the 1950s the flamboyant
psychologist James McConnell took this proverb quite literally. McConnell become
convinced that he had discovered a means of chemically transferring learning from
one animal to another. Indeed, for many years psychology textbooks informed
undergraduates that scientists could chemically transfer learning across animals.
McConnell’s animal of choice? The planaria, a flatworm that’s typically no more
than a few inches long. Using classical
conditioning, McConnell and his colleagues
exposed planaria to a light, which served
as the CS, while pairing it with a 1 second
electric shock, which served as the UCS. When
planaria receive an electric shock, they contract
reflexively. After a number of pairings between
light and shock, the light itself seems to cause
planaria to contract (Thompson & McConnell,
Photograph 12
Photograph of planaria
1955). (Photograph 12)
114
CLASSICAL CONDITIONING
Then, McConnell wanted to find out whether he could chemically transfer the
memory of this classical conditioning experience to another planaria. His approach
was brutally simple. Relying on the fact that many planaria are little cannibals, he
chopped up already trained planaria and fed them to their fellow worms. Remarkably,
McConnell (1962) reported that planaria who had gobbled up classically conditioned
planaria acquired classical conditioned reactions to the light more quickly than
planaria who hadn’t.
Understandably, McConnell’s memory transfer studies generated enormous
excitement. Imagine if McConnell were right! You could sign up for your introductory
psychology class, swallow a pill containing all of the basic psychological knowledge
you’d need to get an A, and....voila, you’re now an expert psychologist. Indeed,
McConnell went directly to the general public with his findings, proclaiming in
Time, Newsweek, and other popular magazines that scientists were on the verge of
developing a “memory pill” (Rilling, 1996).
Yet it was not before long that the wind went out of McConnell’s scientific sails.
There was one big hitch in McConnell’s grand plans for a memory pill: scientists
couldn’t replicate his findings. Moreover, researchers brought up a host of alternative
explanations for his results, some of which he hadn’t fully considered. For example,
McConnell hadn’t completely ruled out the possibility that his findings were
attributable to pseudoconditioning, which occurs when the CS by itself triggers the
UCR. That is, McConnell hadn’t excluded the possibility that the light itself caused
the planaria to contract (Collins & Pinch, 1993), perhaps leading him to the false
conclusion that the cannibal planaria had acquired a classically conditioned reaction
to the light. Eventually, after years of intense debate in the scientific literature and
failed replications, most researchers concluded that McConnell was wrong: he had
fooled himself into seeing a finding that was never there. McConnell’s planaria lab
closed in 1971.
Strange as it was, the McConnell story has an even stranger twist. On November
15, 1985, McConnell went to his mailbox to open an innocent-looking package.
When he opened it, it exploded. Fortunately, McConnell was not seriously injured,
although he suffered permanent hearing loss from the bomb. The package, it turns
out, had been sent by a man named Theodore Kaczynski, better known as the
“Unabomber.” Kaczynski, a former mathematics professor later diagnosed with
paranoid schizophrenia, had mailed bombs to several individuals around the country
who were ardent proponents of technological innovation. Apparently, Kaczynksi had
read McConnell’s popular writings concerning the possibility of using memory pills and
other revolutionary behavior change techniques for transforming modern society, and
identified him as a target (Rilling, 1996).
FA C T A N D F I C T I O N I N PSYC H O LO GY SE L F -T EST 1
(1) Habituation to meaningless stimuli is generally adaptive.
(2) In classical conditioning, the conditioned stimulus (CS) initially yields a reflexive,
automatic response.
(3) In delay conditioning, the CS precedes the UCS and is separated from it by a short
period of time.
(4) Extinction is produced by the gradual “decay” of the CR over time.
(5) Heroin addiction may sometimes be “broken” by dramatically altering the setting
in which addicts inject the drug.
pseudoconditioning
occurs when the CS by itself triggers
the UCR
The answers printed upside down on the bottom of the following page
115
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
1
Using bird feed as a reward, a behavioral psychologist teaches a pigeon to
distinguish paintings by Monet from paintings by Picasso. By the end of the
training, the pigeon is a veritable art aficionado.
2
3
4
Using fish as a treat, a trainer teaches a dolphin to jump out of the water, spin
three times, splash in the water, and propel itself through a hoop.
5
6
A chimpanzee learns to signal “Amy want banana” whenever she’s hungry.
In his initial attempt at playing tennis, a frustrated 12 year old hits his
opponent’s serve into the net the first 15 times. After 2 hours of practice, he
successfully returns his opponent’s serve more than half of the time.
A woman in a stormy relationship makes manipulative suicidal threats (“If you
break up with me, I’ll kill myself!”) whenever her boyfriend is about to leave
her. Feeling sorry and concerned for her, he returns each time.
A hospitalized patient with dissociative identity disorder (known formerly
as multiple personality disorder), displays features of an “alter” personality
whenever staff members pay attention to him. When they ignore him, his alter
personality seemingly vanishes.
The answer: all six are cases of operant conditioning (the first example, incidentally,
comes from an actual study; Watanabe, Sakamoto, & Wakita, 1995). Operant
conditioning is learning that is controlled by its consequences. These six examples,
superficially different as they are, have one important thing in common. In all cases, the
organism’s behavior is shaped by what comes after it, namely reward. Psychologists also
refer to operant conditioning as instrumental conditioning, because the organism’s response
serves an instrumental function. That is, the organism “gets something out” of the
response, like food, sex, attention, or avoiding something unpleasant. We’ll soon discover
how operant conditioning works, as well as how different subtypes of operant condition
affect our behavior in different ways.
Behaviorists refer to the behaviors emitted by the animal in response to reward as
operants, because the animal “operates” on its environment to get what it wants. So, your
dropping 65 cents into a soda machine to get a Diet Coke is an operant, as is your asking
out an appealing classmate on a date. In both cases, you’re engaging in a behavior to obtain
a desired consequence.
OPERANT CONDITIONING: WHAT IS IT AND HOW IT DIFFERS FROM
CLASSICAL CONDITIONING
Operant conditioning
learning that is controlled by its
consequences
operants
behaviors emitted in response to reward
Operant conditioning differs from classical conditioning in three important ways. First,
in classical conditioning, the organism’s response is elicited; that is, “pulled out” of the
organism by the UCS, and then later, the CS. Remember that in classical conditioning,
the UCR is a reflexive and automatic response that doesn’t require training. In contrast,
in operant conditioning, the organism’s response is emitted; that is, generated by the
organism in a seemingly voluntary fashion.
Second, in classical conditioning, the animal’s reward is independent of what the
animal does; Pavlov gave his dogs meat powder regardless of whether, or how much, they
salivated. In contrast, in operant conditioning, the animal’s reward is contingent - that is,
116
THE LAW OF EFFECT
Generations of introductory psychology students have diligently recited the famous Law
of Effect, put forth by psychologist E.L. Thorndike. The Law of Effect is the first and most
important commandment of operant conditioning. Here it is:
If a response, in the presence of a stimulus, is followed by a satisfying state of
affairs, the bond between stimulus and response will be strengthened.
At first blush, that sentence seems about as clear as mud. Yet it is actually isn’t all
that complicated, because it means only that if an event that precedes something we do
is followed by a reward, we’re more like to repeat it. As a consequence, psychologists
sometimes refer to early forms of behaviorism as S-R psychology (S stands for stimulus,
R for response). According to S-R theorists, most of our complex behaviors reflect the
accumulation of connections between stimuli and responses: the sight of a close friend and
saying hello, the smell of a delicious hamburger and reaching for it on our plate, the soft
feel of a comfortable blanket and pulling it over our body at night. Almost everything we
do voluntarily, S-R theorists maintain, like driving a car, eating a sandwich, and planting a
kiss on someone’s lips, results from the gradual build-up of S-R bonds.
Thorndike (1898) discovered the law of effect in a classic study of cats and puzzle boxes.
He placed a hungry cat in a box, and dropped a tantalizing piece of fish or bowl of milk
immediately outside of it. To escape from the box, the cat needed to hit upon (literally) the
right solution, which was pressing on a lever or string inside the box (Figure 6).
When Thorndike first placed one of his cats in the puzzle box, it typically flailed around
aimlessly in a frantic effort to escape. Then, by sheer accident, the cat eventually found
the correct solution, ran out of the box, and gobbled up its delectable reward. Thorndike
wanted to find out what would happen to the cat’s behavior over time. Once it figured out
the solution to the puzzle, would it get it right every time after that?
As Figure 7 shows, the results were surprising, at least to most psychologists of the
time. Thorndike found that cats’ time to escape from the puzzle box decreased only
gradually over 60 trials. According to Thorndike, his cats were learning entirely by trial and
error. Indeed, Thorndike and many other S-R theorists went so far as to conclude that all
learning, including all human learning, occurs by trial and error. For them, S-R bonds are
gradually “stamped into” the organism by reward.
This graph, Thorndike concluded, also provides convincing evidence against the
hypothesis that cats learn by insight, that is, by grasping the nature of the problem. Had
his cats possessed insight into the nature of the problem, the results presumably would
have looked more like what we see in Figure 8. This figure illustrates what psychologists
often term the aha reaction: “Aha – I got it!” Once the animal solves the problem, it gets it
correct just about every time after that. Yet Thorndike found no hint of an aha reaction, so
he concluded that cats had no clue as to what they were doing. They acted pretty much like
mindless robots.
Nevertheless, Thorndike may have jumped the gun by concluding that cats don’t learn
by insight. Here’s the problem: Thorndike constructed the puzzle box in such a way that
117
Figure 6
Thorndike’s puzzle box. The cat needs
to press on the lever to escape and grab
its reward.
360
Latency (seconds)
What do the following six examples have in common?
dependent on - what the animal does. If the animal doesn’t emit a response in an operant
conditioning paradigm, it comes out empty handed (or in the case of a dog, empty-pawed).
Third, in classical conditioning, the organism’s responses mostly involve the autonomic
nervous system (see Chapter 4). In contrast, in operant conditioning, the organism’s
responses mostly involve the skeletal muscles. That is, in contrast to classical conditioning,
in which the learning involves changes in heart rate, breathing, perspiration, and other
bodily systems, in operant conditioning the learning involves changes in voluntary motor
behavior. Nevertheless, as we’ll discover in Chapter 9, there may be exceptions to this neat
and tidy distinction.
240
120
0
20
Trial
40
60
Figure 7
Thorndike discovered that learning in his
puzzle box task occurred gradually over
60 trials, suggesting a pattern of trialand-error learning.
360
Latency (seconds)
Operant Conditioning
OPERANT CONDITIONING
240
120
0
20
40
60
Figure 8 TBD
Insight learning. The animal appears to
have an “aha” reaction, figuring out the
problem suddenly and getting it right
every time after that.
Law of Effect
If a response, in the presence of a
stimulus, is followed by a satisfying
state of affairs, the bond between
stimulus and response will be
strengthened.
S-R psychology
complex behaviors reflect the
accumulation of connections
between stimuli and responses
insight
learning by grasping the nature of
the problem
Answers to questions on previous page:
(1) T; (2) F; (3) F; (4) F; (5) T
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
it was virtually impossible for cats to figure out the relationship between their response
- pressing the lever or pulling a string - and the outcome - escaping from the puzzle box
(Kohler, 1925). So Thorndike hadn’t ruled out an alternative hypothesis, namely, that
the learning occurred by trial and error simply because the cats had no other means of
learning. We’ll come back to the issue of insight learning later in the chapter.
B.F. SKINNER AND REINFORCEMENT
Photograph 13
Using his famous “Skinner box,” B.F.
Skinner (shown here) plotted out
responses of rats and other animals to
operant conditioning.
Although Thorndike’s pioneering discoveries laid the groundwork for research on operant
conditioning, the field of behavioral psychology didn’t emerge in full form until the
research of Harvard University psychologist B.F. Skinner. Although Skinner appreciated
the importance of Thorndike’s work, he found Thorndike’s experimental set-up to be
unwieldly, because the researcher had to stick around to put the unhappy cat back into the
puzzle box following each trial. This limitation made it difficult to study ongoing operant
behavior over long time periods, like hours, days, or weeks.
Skinner therefore developed what came to be known as a Skinner box, a small chamber
that permitted the study of operant conditioning over lengthy spans of time. Indeed, to
study an animal’s behavior in a Skinner box, the experimenter doesn’t even need to be in
the room. She can simply leave the animal in the Skinner box overnight and come back in
the morning to find a cumulative record – an electronically recorded graph of the animal’s
responses over time. The Skinner box could hardly be simpler. For rats, it contains a bar
for pressing food, a tiny food dispenser for giving out pellets, and often a light that signals
when reward is forthcoming (Photograph 13). Using the Skinner box, Skinner studied the
operant behavior of rats, pigeons, and other animals, and mapped out their responses to
reward. His discoveries forever altered the landscape of psychology.
TERMINOLOGY OF OPERANT CONDITIONING
To understand Skinner’s research, we need to introduce you to a bit of psychological jargon.
There are three key concepts in Skinnerian psychology: reinforcement, punishment, and
discriminant stimulus. We’ll explain each in turn.
Photograph 14
Skinner box
a small chamber that permits the study
of operant conditioning over lengthy
spans of time
cumulative record
an electronically recorded graph of an
animal’s responses over time
reinforcement
any outcome that strengthens the
probability of a response
Positive reinforcement
the presentation of a pleasant outcome
negative reinforcement
the withdrawal an unpleasant outcome
Reinforcement. Up to this point, we’ve used the term “reward” to refer to any pleasant
consequence that makes a behavior more likely to occur. Skinner would have disapproved
of our language, because he found the term “reward” to be hopelessly imprecise and
subjective. For example, the opportunity to eat a bowl of tofu might be rewarding for
someone who loves the taste of tofu, but not for someone who finds the taste of tofu
curiously reminiscent to that of a piece of kitchen sponge soaked in water. What’s
rewarding is in the eye of the beholder – or in the case of tofu, the taste of the eater.
So rather than the term “reward,” Skinner (1953, 1971) preferred the term
reinforcement, which means any outcome that strengthens the probability of a response.
If we gave you an M & M each time you read a page of this chapter, and that made you
more likely to read about the psychology of learning, then these M & M’s would be
reinforcers (the term for anything that provides reinforcement).
Skinner distinguished between two types of reinforcement: positive and negative.
Positive reinforcement is the presentation of a pleasant outcome, such as giving a rat a
tasty piece of cheese each time it navigates through a maze correctly. In contrast, negative
reinforcement is the withdrawal an unpleasant outcome, such as removing a misbehaving
child from a “time out” once he’s stopped whining. So, in positive reinforcement, we giveth
(something good), whereas in negative reinforcement, we taketh away (something bad). In
both cases, the outcome for the organism is pleasurable (Photograph 14).
Legions of psychology students over the years have demonstrated the power of
reinforcement using a rather unconventional participant: their professor. In the game
“Condition Your Professor” (Vyse, 1997), a class of introductory psychology students
agrees to provide positive reinforcement – like smiling or nodding their heads – to their
118
professor whenever he or she moves in a particular direction, such as toward or away from
the student audience. One of us knows of a famous introductory psychology teacher who
spent much of time lecturing from behind his podium. During one class, his students
decided to smile profusely and nod their heads vigorously whenever he began to venture
out from behind the podium. Sure enough, by the end of the class, the professor was
spending most of his time well away from the podium. You and your classmates might
want to attempt a similar stunt with your introductory psychology professor: just don’t
mention that we suggested it.
Punishment. Many students – and even a few rusty psychology professors – confuse negative
reinforcement with punishment, which means any outcome that weakens the probability of a
response. To remember the difference between negative reinforcement and punishment, just
remind yourself that negative reinforcement, like all kinds of reinforcement, strengthens the
probability of a response. Punishment, in contrast, weakens it.
Just to make sure that you have a good grasp on the difference between these
two concepts, try labeling each of the following examples as an instance of either
negative reinforcement or punishment (you can find the answers written upside
down at the bottom of this page):
(1) A boy keeps making noise in the back of a classroom despite a teacher’s
repeated warnings. The teacher finally tells him to leave the room. When he
returns two hours later, he’s much better behaved.
(2) A parent who was withholding a child’s allowance for not performing his
household chores decides to restore his allowance after he’s performed his
household chores two weeks in a row. He now does his household chores
even more consistently.
(3) A parole board releases a previously aggressive criminal from prison for
being a “model citizen” within the institution over the past 5 years. Upon his
release, his hostile impulses toward others decrease sharply.
(4) A woman yells at her roommate for leaving dirty clothing scattered all around
her apartment. Her roommate apologizes and never makes a mess again.
Answers: Printed upside down at bottom of the page
As in the case of reinforcement, there are two types of punishment. First, we can punish
the organism by presenting it with a negative stimulus. For example, we could administer
an electric shock to a rat each time it walks over to one section of the cage, or we could
loudly say “No!” to a child who is behaving disruptively (this form of punishment goes
by the strange name of positive punishment). Second, we can punish the organism by
withdrawing a positive stimulus. For example, we could take away food from a child every
time he misbehaves in a restaurant, or we could deduct money from an employee’s holiday
bonus each time he shows up late for work (this form of punishment is sometimes called
response cost).
A second frequent error is to confuse punishment with the disciplinary practices often
associated with it. Skinner, who always insisted on precision in his language, was quick to
remind readers that certain actions that might superficially appear to be punishments are
actually reinforcements. Skinner defined reinforcements and punishments solely in terms
of their consequences. For example, let’s imagine that a mother rushes into her 3 year old
child’s bedroom and yells “Johnny, stop that!” each time that he demands angrily that she
come in to play with him. Is mommy punishing Johnny’s demanding behavior? There’s
no way to know without looking at the consequences. In some cases, Johnny’s demanding
behavior may actually increase following mommy’s scolding, perhaps because she’s paying
attention to him. If so, she’s reinforcing, not punishing, his angry demands.
Behaviorists who use parent training – a set of techniques for teaching parents how to
119
punishment
any outcome that weakens the
probability of a response
parent training
a set of techniques for teaching parents
how to curb the problematic behaviors of
oppositional children
(1) punishment,
(2) negative
reinforcement,
(3) negative
reinforcement,
(4) punishment].
Ruling Out
Alternative
Hypithesis
OPERANT CONDITIONING
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
Correlation
Doesn’t Imply
Causdation
curb the problematic behaviors of oppositional children – often start by teaching parents
the difference between discipline and punishment (Forehand & Long, 1988). That’s
because many parents who discipline their children don’t realize that they’re sometimes
reinforcing them accidentally for bad behavior. They may think that telling Sally to “Go
to your room!” whenever she acts up is an effective disciplinary technique, but they be
forgetting that when Sally goes to her room she can play all of her favorite music CDs and
video games to her heart’s content. As a result, they may be making her more likely to act
up in the future.
Does punishment work in the long run? Popular wisdom tells us that it usually does:
“Spare the rod, spoil the child.” Yet Skinner (1953) and most of his followers argued against
the routine use of punishment to change behavior, because they believed that they could
effectively shape most human behaviors for the better by means of reinforcement alone.
According to Skinner, punishment has several distinct disadvantages. First, it only
tells the organism what not to do, not what to do. That is, punishment isn’t especially
informative. For example, when we punish a young child for throwing a temper tantrum
when he gets frustrated, he may learn not to throw a temper tantrum again. But he won’t
learn how to deal with his frustration in a more constructive manner. Second, punishment
often creates anxiety, which in turn interferes with future learning. Third, punishment may
teach the organism only to become sneakier about the situations in which it can and can’t
display forbidden behavior. For example, a child who is punished by Daddy for grabbing
his brothers’ toys may learn to grab his brother’s toys when Daddy isn’t looking. Fourth,
punishment from parents may provide a model for children’s aggressive behavior (Straus,
Sugarman, & Giles-Sims, 1997). As a consequence, children whose parents spank them
harshly may “get the message” that spanking is acceptable.
Indeed, numerous researchers have reported that the use of physical punishment by
parents is positively correlated with aggressive behavior in children (Gershoff, 2002).
For example, across many studies, Murray Strauss (1996) and his colleagues found that
parental spanking is associated with a higher number of behavioral problems in children.
In a study of 1575 subjects drawn from the general population, Cathy Widom and her
colleagues further found that children who are physically abused are at heightened risk for
becoming aggressive in adulthood (Widom, 1989a, 1989b). Many researchers contend that
this finding demonstrates that spanking and early physical abuse cause aggression. For
example, Widom (1989a) concluded that her findings reveal the operation of a “cycle of
violence,” whereby parental aggression begets childhood aggression. When these children
become parents, many become abusers themselves. Similarly, Elizabeth Gershoff (2002)
conducted a meta-analysis (see Chapter 3) of 88 studies of corporal punishment based
on a whopping 39,309 participants. Although she found some evidence that corporal
punishment is associated with short-term improvements in children’s behavior, she also
reported that a history of such punishment in childhood is associated with an increased
probability of becoming an abuser in adulthood.
Yet we must remember that the findings of Strauss, Widom, and Gershoff are merely
correlational and don’t prove causality. Other interpretations are possible. For example,
because children share half of their genes with each parent, and because aggression is
partly heritable (Krueger, Hicks, & McGue, 2001), it’s possible that the correlation between
parent’s physical aggression and their children’s aggression is due to the fact that parents
who are physically aggressive pass on this predisposition to their children (DiLalla &
Gottesman, 1991). It’s also conceivable that the causal arrow is reversed. For example,
children who are extremely aggressive may be difficult to control and may therefore elicit
spanking from their parents. This hypothesis doesn’t excuse spanking or imply that it’s
acceptable, of course, but it may help to explain why it occurs. In addition, it’s possible that
mild levels of punishment are effective, but that severe forms of punishment, including
abuse, aren’t (Baumrind, Larazelere, & Cowan, 1992).
120
OPERANT CONDITIONING
CRITICAL THINKING QUESTION
How would you attempt to test the hypothesis that parents’ spanking actually causes
greater aggression in their children?
Making matters still more complicated, the association between spanking and childhood
behavior problems depends on both race and culture. Although spanking and other forms
of physical discipline are positively correlated with childhood behavior problems in White
families, they are negatively correlated with childhood behavior problems in AfricanAmerican families (Lansford, Deater-Deckard, Dodge, Bates, & Petit, 2004). Apparently,
physical discipline means something different – perhaps a deep commitment to rearing
children properly - in African-American than in White families. Moreover, spanking
tends to be more highly correlated with childhood aggression and anxiety in countries
in which spanking is rare, like China or Thailand, than in countries in which spanking is
common, like Kenya or India (Lansford et al., 2005). The reasons for this difference aren’t
clear, although children who are spanked in countries in which spanking is more culturally
accepted may feel less stigmatized than children in countries in which spanking is culturally
condemned.
So when, if ever, is punishment effective in the long-term? Most research suggests that
punishment works best when it (a) is delivered consistently and (b) follows the undesired
behavior promptly. For example, in a correlational study of crime and punishment in a
sample of over 28,000 men in Denmark, Patricia Brennan and Sarnoff Mednick (1994)
found that consistently delivered prison sentences were associated with lower rates of
reoffending than were inconsistently delivered prison sentences. In contrast, the length or
severity of the prison sentence was unrelated to the risk of reoffending. Discipline may not
need to be harsh to be effective (Photograph 15).
Discriminant stimulus. One other key term in operant conditioning lingo is the
discriminant stimulus, typically abbreviated simply as Sd. A discriminant stimulus is
any stimulus that signals the presence of reinforcement. When you snap your fingers at a
dog in the hopes of having it come over to you, the dog may approach you to get a muchappreciated petting. For the dog, your finger snapping is an Sd: it’s a signal that if it comes
near you, it will receive reinforcement (Photograph 16).
According to behaviorists, we’re responding to Sds virtually all of the time, even if
we’re not consciously aware of it. For example, let’s say that while walking across campus
tomorrow you encounter a friend who smiles or waves at you. Most likely, you’ll either
smile or wave back, and if you’re not busy you might walk over to say hello. By smiling or
waving, your friend has given you the “green light” to say hello; that green light is his Sd. In
the very next sentence, you’re about to learn something interesting. The previous sentence
was also an Sd, because it was a signal that if you read this sentence, you’d be reinforced by
acquiring a piece of intriguing knowledge. See, it worked.
Acquisition, extinction, spontaneous recovery, and stimulus generalization and
discrimination.
If you feel that you’re experiencing a case of déjà vu (see Chapter 8), don’t be concerned,
because you’ve indeed seen all of these terms before. Acquisition, extinction, spontaneous
recovery, and stimulus generalization and discrimination apply just as much to operant
conditioning as to classical conditioning.
For example, in operant conditioning, extinction occurs when we stop delivering
reinforcement to a previously reinforced behavior. Gradually, this behavior declines in
frequency and disappears. Let’s imagine that you’re a behavioral psychologist assigned to
help two exasperated parents eliminate the yelling behavior of their 4 year child, Jimmy.
After observing the family for a few days, you notice something curious: every time Jimmy
121
Photgraph 15 the length or severity
of the prison sentence was unrelated to
the risk of reoffending
Photograph 16
As any household pet owner soon
discovers, intelligent animals learn to
respond to a wide range of discriminant
stimuli.
discriminant stimulus
any stimulus that signals the presence of
reinforcement
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
PRINCIPLES OF REINFORCEMENT
Using the Skinner box as his miniature laboratory, Skinner uncovered a variety of general
principles of behavior. To demonstrate one of them, first try to answer a little quiz
question. If you wanted to train a dog to perform a trick, like catching a Frisbee, would
you reinforce it for (a) every successful catch or (b) only some of its successful catches? If
you’re like most people, you’d answer (a), which seems to dovetail with our common sense
notions regarding the effects of reinforcement. It seems logical to assume that the more
consistent the reinforcement, the more consistent should be the resulting behavior.
extinction burst
shortly after withdrawing reinforcement
undesired behavior initially increases in
intensity
partial reinforcement
behaviors reinforced only occasionally
are slower to extinguish than behaviors
reinforced continuously
Partial reinforcement. Nevertheless, Skinner’s principle of partial reinforcement,
sometimes called Humphrey’s paradox after psychologist Lloyd Humphreys (1938) who
first reported it about 7 decades ago, shows that our intuitions about reinforcement
are backwards. According to the principle of partial reinforcement, behaviors that we
reinforce only occasionally are slower to extinguish than are behaviors that we reinforce
continuously, that is, every time. So if we want an animal to maintain a trick for a long
time, we should reinforce it for correct responses only occasionally. This principle may
122
Time
Cumulative responses
Variable interval
Cumulative responses
Time
Cumulative responses
Cumulative responses
Cumulative responses
Time
Fixed interval
Variable ratio
Fixed ratio
600
300
0
Time
Figure 9 The four major schedules of reinforcement discovered by B.F. Skinner
also help to explain why some people remain trapped for years in horribly dysfunctional
relationships. Some relationship partners provide intermittent reinforcement to their
significant others, treating them miserably most of the time but treating them well on
rare occasions. This pattern of partial reinforcement may keep individuals “hooked” in
relationships that aren’t working (Photograph 18).
Schedules of reinforcement. Skinner (1938) also found that animals’ behaviors vary
as a function of different schedules of reinforcement, that is, patterns of delivering
reinforcement. Remarkably, the effects of these reinforcement schedules are consistent
across species as diverse as cockroaches, pigeons, rats, and humans. Although there are
numerous schedules of reinforcement, we’ll discuss four here. The principal reinforcement
schedules vary along two dimensions: (1) the consistency of administering reinforcement
and (2) the basis of administering reinforcement. Let’s explain.
With regard to consistency of administering reinforcement, some reinforcement
contingencies are fixed, whereas others are variable. That is, in some cases the
experimenter provides reinforcement on a regular (fixed) basis, whereas in others the
experimenter provides reinforcement on an irregular (variable) basis. With regard to
the basis of administering reinforcement, some reinforcement schedules operate on
ratio schedules, whereas others operate on interval schedules. In ratio schedules, the
experimenter reinforces the animal depending on the number of responses it has emitted.
In interval schedules, the experimenter reinforces the animal depending on the amount of
time that has elapsed since its last response.
We can cross these two dimensions to arrive at four major schedules of reinforcement
(Figure 9). In a fixed ratio (FR) schedule, we provide reinforcement after a regular number
of responses. For example, we could give a rat a pellet after it presses the lever in a Skinner
box 15 times. In a fixed interval (FI) schedule, we provide reinforcement after a regular
period of time. For example, assuming that he’s done her job well, we could pay a worker
in a factory every two weeks. In a variable ratio schedule, we provide reinforcement after
an irregular number of responses. For example, we could give a pigeon a piece of bird feed
after 6 pecks, then after 15 pecks, then after 1 peck, then after 35 pecks, and so on. Finally,
in a variable interval schedule, the experimenter provides reinforcement after irregular
time intervals. For example, we could give a dog a treat for performing a trick after 7
minutes, then 1 minute, then 20 minutes, then 3 minutes, and so on.
Skinner discovered that these reinforcement schedules yielded distinctive patterns of
responding (Figure 10). For example, he found that ratio schedules tend to yield higher
rates of responding than do interval schedules. In addition, variable schedules tend to yield
more consistent rates of responding than do interval schedules.
Two other features of reinforcement schedules are worth noting. First, fixed interval
schedules tend to be associated with a “scalloped” pattern of responding. This FI scallop
reflects the fact that the animal “waits” for a time after it receives reinforcement, and then
jacks up its rate of responding as it begins to anticipate reinforcement (Figure 11).
Second, as you can see in Figure 10, variable ratio (VR) schedules tend to yield the
highest rates of responding of all reinforcement contingencies. Indeed, there’s one place
123
10
Minutes
20
Figure 10
The “FI scallop.” The three short diagonal lines indicate the occurrence of reinforcement. In fixed interval schedules,
the animal slows down after receiving
reinforcement (behaviorists call this a
“post-reinforcement pause”) and then
speeds up as reinforcement approaches.
One often sees the same scalloped pattern of productivity among employees
who get paid at regular intervals.
Average errors
Photograph 18
TBD
screams, the parents give him his favorite toy robot in a desperate effort to
shut him up. It works, and Jimmy indeed quiets down in the short term.
But having read Chapter 6 of your introductory psychology text, you
have a sneaking suspicion that their practice may be backfiring, because
it may be reinforcing Jimmy’s screaming in the long term. So you
instruct the parents to stop giving Jimmy his toy robot each time he acts
up. Sure enough, if they wait long enough his screaming behavior will
gradually extinguish. Interestingly, though, in such cases we often see an
extinction burst. That is, shortly after withdrawing reinforcement:
the undesired behavior initially increases in intensity. That’s
probably because the child is “upping the ante”: by screaming even
more intensely, he’s trying harder to get reinforcement. So there’s
some truth to the saying that things sometimes need to get worse
before they get better (Photograph 17).
Stimulus discrimination and generalization also occurs in
operant conditioning. As we mentioned earlier, using food
reinforcements, one group of investigators trained pigeons
to distinguish paintings by Monet from those of Picasso
(Watanabe et al., 1995). That’s stimulus discrimination,
because the pigeons are learning to tell the difference
Photograph 17
between two different types of stimuli. Interestingly, the
the child is “upping the ante”: by
investigators also found that these pigeons exhibited stimulus
screaming even more intensely,
he’s trying harder to get
generalization. Following operant conditioning, they were
reinforcement.
also able to distinguish paintings by artists similar to Monet,
such as Renoir, from paintings by artists similar to Picasso, such as Braque.
The similarities between classical and operant conditioning, including the fact that we
find acquisition, extinction, stimulus generalization, and so on, in both, have led some
theorists to argue that these two forms of learning aren’t as different as some psychologists
believe (Brown & Jenkins, 1968; Staddon, 2003). Although there certainly are important
similarities between classical and operant conditioning, brain imaging studies demonstrate
that these two forms of learning are rooted in different brain regions. For example,
classically conditioned fear reactions are based largely in the amygdala (LeDoux, 1996;
Veit, Flor, Erb, Lotze, Grodd, & Birbaumer, 2002), whereas operant conditioned responses
are based largely in the nucleus accumbens and related limbic systems linked to reward
(Robbins & Everitt, 1998; see Chapter 4).
OPERANT CONDITIONING
32
28
24
20
16
12
8
4
0 2 4 6 8 10 12 14 16 18 20
Days of experience in maze
Figure 11
Graphs from Tolman and Honzik’s classic study of latent learning in rats. Pay
particular attention to the orange line.
The rats in this group weren’t reinforced
until day 11; note the sudden drop in the
number of their errors upon receiving
reinforcement. The rats were learning
all along, even though they weren’t
showing it.
schedules of reinforcement
that is, patterns of delivering
reinforcement
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
Photograph 19
Although they may not realize it, most or
all of the people in this casino are under
the spell of a variable ratio schedule.
where you can be guaranteed to find a VR schedule: a casino. Roulette wheels, slot
machines, and the like deliver cash rewards on an irregular basis, and they do so based on
the gambler’s responses. Sometimes the gambler has to pull the arm of the slot machine
(the so-called “one-armed bandit”) hundreds of times before receiving any money at all. At
other times, the gambler the pulls the arm only once and makes out like a bandit himself,
perhaps walking off with hundreds of dollars for a few seconds of work. The extreme
unpredictability of the VR schedule is precisely what keeps gamblers hooked, because
reinforcement can come at any time (Photograph 19)
The VR schedule also keeps pigeons hooked. For example, Skinner (1953) found that
pigeons placed on VR schedules sometimes continued to peck on a disk for food after more
than 150,000 nonreinforced responses. In some cases, they pecked the disk so many times
that they grounded down their beaks. Like desperate gamblers in a Las Vegas casino hoping
for a huge payoff, they wouldn’t give up despite repeated disappointments. For Skinner,
much of what we term “persistence” is merely a consequence of being on a reinforcement
schedule that is difficult to extinguish, especially a VR schedule.
CRITICAL THINKING QUESTION
How might the effect of VR schedules help to explain why some people continue to believe
in psychic experiences despite the absence of strong evidence for their existence?
Before reading any further, here’s another little quiz question. Imagine that
you’ve flipped a penny 10 times in a row and that it’s come up tails all 10 times.
What are the odds that it will come up heads the 11th time?
The answer? 50%.
If your answer was higher than 50%, don’t feel bad. Scores of highly intelligent people
make this understandable mistake, which is called the gambler’s fallacy. The gambler’s
fallacy is the incorrect belief that random events have a memory (Ayton & Fischer, 2004;
Schulman, 2002). Along with the seductive power of VR schedules, this fallacy contributes
to the perennial allure of dice and craps tables.
The gambler’s fallacy is a reasoning error because the odds that a penny will come up
heads – or tails – remain exactly 50% with each flip. Many gamblers believe mistakenly
that following a long string of unsuccessful rolls of the dice they are “due”: the next roll is
bound to turn out in their favor. Nevertheless, the dice don’t “remember” their previous
rolls, so the probability starts anew with each roll. Or take the case of parents who’ve had
5 boys in a row, but want a girl desperately. They may try to have a 6th child, believing that
their next child is almost certain to be a girl. Because the odds of having a girl stay at just
about 50% each time, they may be in store for another disappointment.
The gambler’s fallacy probably stems in part from a misapplication of the
representativeness heuristic (see Chapter 3). To understand why, consider the
following two sequences of 10 successive penny flips, with H being heads and T
being tails.
HHHHHHHHHH
HHHTHHHTHT
Which of these sequences is more likely?
gambler’s fallacy
the incorrect belief that random events
have a memory
shaping by successive approximations
reinforcing progressively closer versions
of a desired response
In fact, both sequences are equally probable, because the odds of an H and a T with
each flip are each 50%. Nevertheless, many people assume that the second sequence is
more likely, because this sequence seems more representative, that is, more typical of a
random sequence of coin flips. After several consecutive heads, a tail somehow seems more
probable. It isn’t.
124
OPERANT CONDITIONING
APPLICATIONS OF OPERANT CONDITIONING.
There’s an old joke that just as magicians pull rabbits out of hats, behavioral psychologists
pull habits out of rats, and there’s a grain of truth to this one. By means of operant
conditioning, behaviorists train rats and other organisms to develop learned habits. They
typically do so by means of a procedure called shaping by successive approximations,
often called shaping for short. Using shaping, we reinforce progressively closer versions of
a desired response. Typically, we shape an organism’s response by initially reinforcing most
or all correct responses, and then gradually fading (that is, decreasing the frequency of) our
reinforcement over time.
When you learned to drive a car, your Driver’s Ed instructor (or perhaps one of your
parents, depending on who taught you) used shaping to help you learn how to parallel
park. Your instructor first taught you to pull up aside another car, then put the car in
reverse, then turn the steering wheel by a certain amount, then gradually back up the car
while looking behind you, and so on, all the while reinforcing you verbally (“Good, that’s
better...” “Yes, that’s right...keep going”). Eventually, you learned the complete sequence of
steps required to parallel park a car.
By means of shaping, Skinner taught pigeons to play ping-pong, although they weren’t
exactly Olympic-caliber table tennis players. During World War II, Skinner also taught
dolphins to steer torpedoes toward enemy ships by pecking a target whenever the torpedo
got closer to the bulls-eye, although the U.S military never ended up adopting his creative
approach to naval warfare (Photograph 20). In both cases, Skinner began by reinforcing
initial approximations to the desired response. For example, when teaching pigeons to
play ping-pong, he first reinforced them for turning toward the rackets, then approaching
the rackets, then placing the rackets in their mouths, then picking up the rackets with their
mouths, then swinging the racket, and so on. As you might imagine, shaping complex
animal behaviors requires patience, as the process may take days or weeks. Still, the payoff
can be substantial, because one can train animals to engage in numerous behaviors that
lie well outside their normal repertoire. Indeed, all contemporary animal trainers rely on
Skinnerian shaping principles (Photograph 21).
Behaviorists do more than pull habits out of rats, because they also pull habits out of
humans. Indeed, one of the great triumphs of modern psychology has been the application
of operant conditioning to myriad domains of modern life. We’ll look at four applications
here: the Premack principle, superstitious behavior, token economies, and teaching
language to autistic children.
Photograph 20
During World War II, Skinner taught
pigeons to guide torpedoes toward
enemy ships.
Photograph 21
By using shaping techniques, animal
trainers have taught animals, such as
squirrels and hamsters, to perform a
wide variety of amusing tricks.
Premack principle. Be honest: did you put off reading this chapter until the last possible
moment? If so, don’t feel ashamed, because procrastination is one of the most frequent
study problems that college students report. Although widespread, procrastination may
not be harmless. It may be bad for your psychological health, not to mention your grades.
One group of researchers found that chronic procrastinators reported lower stress levels
and fewer physical illnesses early in the semester, but higher stress levels and more physical
illnesses later in the semester, when the weight of their overdue work came crashing down
on their heads. Moreover, procrastinators tend to perform more poorly in their classes
than do early birds (Tice & Baumeister, 1997). Although these findings are correlational
and don’t establish that procrastination is bad for you, they certainly suggest that putting
things off isn’t likely to be good for you.
CRITICAL THINKING QUESTION
What third variables could account for the correlation between procrastination and
poor grades?
Premack Principle
positively reinforcing a less frequently
performed behavior with a more
frequently performed behavior
125
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
Photograph 22
Triskaidekaphobia, or fear of the number
13, is so prevalent that some tall buildings skip from the 12th to the 14th floor.
How can we overcome procrastination? We hope you won’t put off reading the next
two paragraphs, because we have a possible remedy for dilly-dallying. Although there
are many potential solutions to procrastination, one of the best is probably the Premack
Principle, discovered by David Premack (1965) in his research on monkeys. This principle
states that we can positively reinforce a less frequently performed behavior with a more
frequently performed behavior (Danaher, 1974). Although not a foolproof rule (Knapp,
1976), this guideline typically works surprisingly well. The Premack Principle is also
known as “Grandma’s Rule,” because our grandmother reminded us that we need to finish
our vegetables before moving onto dessert. If we give children the opportunity to choose
between spinach and chocolate ice cream, they’ll vote with their mouths, and choose to
eat chocolate ice cream at a much higher frequency than they eat spinach. In this way, we
can get children to eat spinach by reinforcing them with chocolate ice cream if, and only if,
they’ve finished their spinach.
Similarly, we can help to overcome our schoolwork procrastination by engaging in a
higher-frequency behavior once we’ve completed our assignments. So, if you find yourself
putting off a reading or writing task, try to think of behaviors that you’d typically perform
if given the chance – perhaps hanging out with a few of your close friends, watching a
favorite TV program, or munching down an ice cream cone. Then, reinforce yourself
with these higher frequency behaviors only after you’ve completed your homework.
Research indicates that this approach may help people to avoid putting off things they’ve
long dreaded, like going to the dentist (Ramer, 1980). So if you’re procrastinating in your
schoolwork and are contemplating using the Premack Principle to help you get over it,
don’t put off trying it.
Superstitious behavior. How many of the following behaviors do you engage in?
• Never opening an umbrella indoors.
• Not walking under a ladder.
• Crossing the street whenever you see a black cat.
• Carrying a lucky charm or pendant.
• Going out of your way not to step on cracks in the sidewalk.
• Knocking on wood.
• Crossing your fingers.
• Avoiding the number 13 (like not stopping on the 13th floor of a building).
superstitious behavior
behavior linked to reinforcement by
sheer coincidence
If you perform several of these behaviors, you are “very superstitious,” in the words of
singer Stevie Wonder. So are quite a few Americans. For example, 12% of Americans are
afraid of walking under a ladder, while 14% are afraid of crossing paths with a black cat
(Vyse, 1997). So many people are afraid of the number 13 (incidentally, the technical name
for this fear is triskaidekaphobia) that the floor designations in many tall buildings skip
directly from 12 to 14 (Hock, 2002). This phobia isn’t limited to North America; in Paris,
triskaidekaphobics who are going out to dinner with 12 other people can hire out for a
quatorzieme, a person paid to serve as a 14th guest (Photograph 22).
Many college students also adopt superstitious rituals prior to their exams. Psychologist
Stuart Vyse (1997) found that 62% of undergraduates had used a “lucky” pen or had worn
a “lucky” piece of jewelry or clothing before an exam. Thirty-one percent had performed a
specific sequence of actions (like saying certain words) prior to a test, and 26% ate a special
food.
How do superstitions relate to operant conditioning? In a classic study, B.F. Skinner
(1948) placed 8 food-deprived pigeons in a Skinner box while delivering reinforcement
(bird feed) every 15 seconds independent of their behavior. In other words, the birds
received reinforcement regardless of what they did. After a few days, Skinner found that 6
of the 8 birds had acquired remarkably strange behaviors. In the words of Skinner:
One bird was conditioned to turn counterclockwiseabout the cage, making two or
126
OPERANT CONDITIONING
three turns between reinforcements. Another repeatedly thrust its head into one of the
upper corners of the cage. A third developed a tossing response as if placing its head
beneath an invisible bar and lifting it repeatedly. Two birds developed a pendulum
motion of the head and body in which the head was extended forward and swing from
right to left....(p. 168).
When Skinner extended the reinforcement interval to a few minutes, one of the pigeons
even performed a miniature “dance” of sorts while awaiting food. You may have observed
similar odd behaviors in large groups of birds that people are feeding in city parks;
for example, some pigeons prance around or walk rapidly in circles in anticipation of
reinforcement.
According to Skinner, his pigeons had developed superstitious behavior: that is,
behavior linked to reinforcement by sheer coincidence (Morse & Skinner, 1957). There’s
no actual association between superstitious behavior and reinforcement, although the
animal behaves as though there is. The behavior that the pigeon just happened to be
performing immediately prior to reinforcement was strengthened - remember that
reinforcement increases the probability of a response - so the pigeon kept on doing it (this
kind of accidental operant conditioning is sometimes called superstitious conditioning).
Skinner argued that operant conditioning accounts for many human superstitions.
For example, Skinner (1948) noted that bowlers often angle and twist their bodies in
peculiar ways as the ball rolls down the alley. In some cases, they may tilt their heads in
the direction they want the bowling ball to go. They almost seem to be trying to “coax” the
ball into the pins, although this behavior of course has no effect whatsoever on where the
ball ends up. Moreover, studies show that operant conditioning can produce superstitious
behaviors in children. For example, two investigators (Wagner & Morris, 1987) placed
children in a room with a mechanical clown, who periodically (either every 15 or every 30
seconds) dispensed marbles that the children could later cash in for an attractive toy. The
researchers used covert observation by watching the children through a two-way mirror.
They found that about three-fourths of the children developed superstitious behaviors;
some touched the clown’s face, others made faces at him, and another kissed him.
Few people are more prone to superstitions than athletes. That’s probably because the
outcome of so many sporting events, even those requiring a great deal of skill, depends
heavily on chance. As we learned in Chapter 2, baseball Hall of Famer Wade Boggs
became famous for eating chicken prior to each game; he also performed practice runs
in the outfield at precisely 7:17 P.M. before each evening game. Hall of Fame football
player Jim Kelly forced himself to vomit before every game, and basketball player Chuck
Persons ate exactly two candy bars (always Snickers or KitKats) before every game (Vyse,
1997). Superstar golfer Tiger Woods always wears a red shirt when playing on Sundays
(Photograph 23).
Interestingly, the prevalence of superstitions in sports seems to depend heavily on the
extent to which the outcomes are due to chance. That’s exactly what Skinner would have
predicted, because as we’ve already discovered, partial reinforcement schedules are more
likely to produce enduring behaviors than are continuous reinforcement schedules. In
baseball, hitting is much less under the player’s control than is fielding: even the best hitters
only succeed about 3 out of 10 times, whereas the best fielders succeed 9.8 or even 9.9 out
of 10 times. Putting it differently, hitting is controlled by a partial reinforcement schedule,
whereas fielding is controlled by something very close to a continuous reinforcement
schedule. As you might expect, baseball players have far more hitting-related superstitions
– like drawing a favorite symbol in the sand in the batter’s box - than fielding-related
superstitions (Gmelch, 1974; Vyse, 1997).
Of course, human superstitions aren’t due entirely to operant conditioning. Many
superstitions are spread partly by word-of-mouth. If our mother tells us over and over
again that black cats cause bad luck, we may become wary of black cats that cross our
127
Photograph 23
Like many athletes, superstar golfer
Tiger Woods is superstitious: he always
wears a red shirt on Sundays. In this
case at least, his superstition seems to
have worked.
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
paths. Still, word-of-mouth explanations for beliefs have one major problem: they don’t
tell us how these beliefs originated. For many superstitions, operant conditioning may
provide the answer.
CRITICAL THINKING QUESTION
How might superstitious conditioning help to explain why some people continue to read
their horoscopes or consult with crystal ball readers despite the absence of evidence that
these practices are effective?
Photograph 24 TBD
Photograph 25
token economies remain controversial,
because the behaviors learned in these
institutions don’t always transfer to the
outside world
token economies
systems, often set up in psychiatric
hospitals, for reinforcing appropriate
behaviors and extinguishing
inappropriate behaviors
Secondary reinforcers
neutral objects that patients can later
trade in for primary reinforcers
primary reinforcers
things that are naturally pleasurable, like
a favorite food or drink
applied behavior analysis ABA)
a set of techniques based on operant
conditioning principles that relies on the
careful measurement of behavior before
and after implementing interventions
Token economies. One of the most successful applications of operant conditioning has
been the development of token economies. These are systems, often set up in psychiatric
hospitals, for reinforcing appropriate behaviors and extinguishing inappropriate behaviors
(Carr, Fraizier, & Roland, 2005; Kazdin, 1982). In token economies, staff members
reinforce patients who behave in a desired fashion using tokens, chips, points, or other
secondary reinforcers. Secondary reinforcers are neutral objects that patients can later
trade in for primary reinforcers - things that are naturally pleasurable, like a favorite food
or drink (Photograph 24)
Typically, psychologists who construct token economies begin by identifying certain
target behaviors, that is, actions that they wish to make more frequent. One psychiatric
hospital unit in which one of the authors of your textbook worked consisted mostly of
children with serious behavior problems, which included yelling and cursing at staff
members. In this unit, one target behavior in the token economy was being polite to staff
members. So whenever a children was especially polite to a staff member, he was rewarded
with points, which he could later trade in for something he wanted, like ice cream or
attending a movie with staff members. Whenever a child was rude to a staff member, he
was punished with a loss of points (as you may recall from earlier in the chapter, this type
of punishment is called response cost).
Research suggests that token economies are often effective in improving behavior in
hospitals, group homes, and juvenile detection units (Allyon & Milan, 2002). Nevertheless,
token economies remain controversial, because the behaviors learned in these institutions
don’t always transfer to the outside world (Carr et al., 2005). That’s especially likely if the
patients return to environments in which they’re reinforced for socially inappropriate
behaviors (Photograph 25).
Teaching language to autistic children. As we learned in Chapter 3, infantile autism
(known officially as autistic disorder) is a severe condition marked by shortcomings in
social attachment, capacity for imagination, and language. About three-fourths of autistic
individuals are mentally retarded, and most are profoundly linguistically impaired. As we
also learned in that chapter, many well-meaning mental health professionals have used
facilitated communication or other pseudoscientific methods in an effort to teach language
to autistic individuals (Jacobson, Mulick, & Schwartz, 1995). Although these techniques
are ineffective, the good news is that applied behavior analysis (ABA) can be quite helpful
in remedying the language deficits of individuals with autism and other developmental
disabilities. ABA is a set of techniques based on operant conditioning principles that relies
on the careful measurement of behavior before and after implementing interventions
(Romanczyk et al., 2003). ABA for autism makes extensive use of shaping techniques;
mental health professionals reinforce autistic individuals with food and other primary
reinforcers as they reach progressively closer approximations to certain words and
eventually, complete sentences.
Ivar Lovaas and his colleagues at the University of California-Los Angeles have
pioneered by far the best known ABA program for autism (Lovaas, 1987; McEachlin,
Smith, & Lovaas, 1993). The results of Lovaas’ work have been promising, as they
indicate that an experimental group of children with autism who undergo ABA training
128
OPERANT CONDITIONING
emerge with superior language and intellectual skills compared with control groups of
children with autism who do not undergo such training (Green, 1996; Matson, Benavidez,
Compton, Paclowaskyj, & Baglio, 1996). (Photograph 26).
Nevertheless, ABA researchers haven’t been immune from making exaggerated
claims regarding the effectiveness of their treatment. For example, although Lovaas
(1987) claimed that many of the children he treated “recovered” from autism, this
conclusion is questionable. For one thing, because Lovaas didn’t randomly assign
children with autism to experimental and control groups, his findings are vulnerable to
an alternative explanation: perhaps the children in the experimental group had higher
levels of functioning to begin with. Indeed, there’s evidence that this was the case
(Schopler,
Short, & Mesibov, 1989). The current consensus is that ABA isn’t a miracle cure for the
language deficits of children with autism, but that it can be extremely helpful in some cases
(Herbert, Sharp, & Gaudiano, 2002).
Ruling Out
Alternative
Hypithesis
TWO-PROCESS THEORY: PUTTING CLASSICAL AND OPERANT
CONDITIONING TOGETHER
Up to this point, we’ve discussed classical and operant conditioning as though they were
two entirely independent processes. Yet the reality is much more complicated. In everyday
life, classical and operant conditioning often interact with one another.
To see how, let’s revisit the question of how people develop phobias. We’ve pointed out
that certain phobias arise in part by classical conditioning: a previously neutral stimulus
(the CS) - like a dog - is paired with an unpleasant stimulus (the UCS) – a dog bite
– producing the CR of fear. So far, so good.
Yet this neat and tidy scheme doesn’t answer an important question: Why doesn’t the
CR of fear eventually extinguish? Given what we’ve learned about classical conditioning,
you might expect the CR of fear to gradually fade away over time with repeated exposure to
the CS of dogs. Yet this often doesn’t happen (Rachman, 1977). Many people with phobias
remain deathly afraid of their feared stimulus for years, even decades. For example, only
about 20% of untreated adults with phobias ever get over their fears (American Psychiatric
Association, 2000). Why?
Enter two-process theory to the rescue as an explanation for this mystery (Mowrer,
1947). According to two-process theory, we need both classical and operant conditioning
to explain the persistence of phobias and other anxiety disorders over time. Here’s how
it works. First, people acquire phobias by means of classical conditioning. Second, once
people are phobic, they start to avoid their feared stimulus any time they see it. For
example, if they have a dog phobia, they may cross the street whenever the see someone
walking toward them with a huge German shepherd. When they do, they experience a
reduction in anxiety – a sense of relief - which negatively reinforces their fear. Remember
that negative reinforcement, which is one type of operant conditioning, is the removal of
an unpleasant stimulus, in this case anxiety. So, by avoiding dogs whenever they see them,
dog phobics are negatively reinforcing their fear – and thereby operantly conditioning
themselves to make their fears more likely to endure!
According to two-process theory, we can think of anxiety disorders as produced by
short-term gain at the cost of long-term pain. The sufferer purchases immediate relief
at the price of a long-term fear. Although two-process theory probably doesn’t explain
the persistence of all phobias (Barlow, 2001), it’s proven useful for generating ideas for
treating them. As we’ll discover in Chapter 17, one effective strategy for treating phobias is
to encourage individuals to confront, rather than avoid, the feared stimulus. By doing so,
we stop the person from negatively reinforcing her fear, and transform the equation into
short-term pain at the price of long-term gain (Foa & Kozak, 1986
129
Photograph 26
Applied behavior analysis has yielded
at least some success in treating the
deficits of children with autism. Here
a teacher uses operant conditioning
techniques to help a child with autism
name the objects or animals shown in a
set of photographs.
two-process theory
using both classical and operant
conditioning to explain the persistence
of phobias and other anxiety disorders
over time
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
FA C T A N D F I C T I O N I N PSYC H O LO GY SE L F -T EST 2
(1) In classical conditioning, responses are emitted; in operant conditioning, they are
elicited.
(2) Negative reinforcement and punishment are superficially different, but they produce the
same short-term effects on behavior.
(3) The correlation between spanking and children’s behavioral problems appears to be
positive in Whites but negative in African-Americans.
(4) The principle of partial reinforcement states that behaviors that are reinforced only
some of the time tend to extinguish more rapidly than behaviors that are reinforced
continuously.
(5) According to the Premack principle, we can reinforce low frequency behaviors
with higher-frequency behaviors.
Answers printed upside down on the bottom of the following page
Cognitive Models of Learning
Up to this point, we’ve omitted one word when discussing how we learn: thinking. That’s
not entirely accidental, because the early behaviorists didn’t believe that thought played
much of a causal role in learning.
methodological behaviorism
focussing exclusively on overt
(observable) behaviors
radical behaviorism
For radical behaviorists, overt behavior,
thinking, and emotion are all governed
by the same laws of learning, namely
classical and operant conditioning
COGNITIVE MODELS OF LEARNING
Fiction
Fact
Skinner believed that genes
played no role in human
behavior
Skinner acknowledged that genes affected the ease with which people
affected the ease with which people learn habits, although he felt that
the study of genes was of little relevance to psychology
Skinner did not believe in
thinking or emotion
Skinner believed that humans and other intelligent animals think and
experience emotions, although regarded thinking and feeling merely
as covert (unobservable) behaviors
Skinner favored the use of
punishment as a behavioral
technique
Skinner opposed the use of punishment for shaping human behaviors,
and believed that reinforcemen should always be used whenevert
possible
Skinner believed that any
human behavior could be
conditioned
Skinner never argued this, although his predecessor John Watson came
close to doing so
Skinner denied the uniqueness Skinner openly acknowledged that all individuals are unique, as they
of individuals
are the product of unique genetic predispositions and unique learnin
histories
Skinner’s vision of society was
devoid of morality
Skinner explicitly advocated a society in which people would be
reinforced for ethical and cooperative behavior
From DeBell and Harless (1992); Wyatt (2001)
WATSON, SKINNER, AND THOUGHT
Table 2 B.F. Skinner may be the misunderstood of all famous psychologists.
Actually, Watson and Skinner held two somewhat different views on this matter. Watson
(1913) was an advocate of methodological behaviorism. According to this school of
thought, psychology should focus exclusively on overt (that is, observable) behaviors.
From Watson’s perspective, thinking and emotion lay outside the domain of scientific
psychology. There’s an old joke about two methodological behaviorists who meet at a
conference. They find themselves attracted to each other and end up sleeping together.
Upon awakening, one of them asks, “That was great for you. How was it for me?” The
point here is that methodological behaviorists studied only what they could observe: what
you see is what you get.
In contrast, Skinner (1953) was a proponent of radical behaviorism. For radical
behaviorists, overt behavior, thinking, and emotion are all governed by the same laws of
learning, namely classical and operant conditioning. For Skinnerians (the term often used
for radical behaviorists), thinking and emotion are behaviors, they are just covert (that is,
unobservable) behaviors. One frequent misconception about Skinner, held not only by
many students but even some psychology professors, is that he didn’t believe in thinking. In
fact, Skinner isn’t merely one of the most famous of all psychologists; he is also one of the
most misunderstood (DeBell & Harless, 1992; Wyatt, 2001; see Table 3).
To the contrary, Skinner clearly thought - he wouldn’t have objected to our use of that
word here - that humans and other intelligent animals think, but he believed that thinking
was no different from any other behavior. At times, Skinner went even further. In a talk
to the American Psychological Association given only 8 days before his death, Skinner
(1990) likened proponents of cognitive psychology, who believe that thinking plays a
central role in causing behavior (see Chapter 1), to pseudoscientists. Cognitive psychology,
he argued, invokes unobservable and ultimately meaningless concepts – like the “mind”
– to explain behavior. By doing so, Skinner said, it doesn’t get us any closer to the true
causes of behavior, any more than invoking a mysterious higher power gets us any closer to
explaining the origin and development of species.
learning in humans is incomplete without at least some role for cognition, that is, thinking
(Bolles, 1979; Kirsch, Lynn, Vigorito, & Miller, 2004).
Over the past two or three decades, psychology has moved increasingly away from a
simple S-R psychology to a more complex S-O-R psychology, with “O” being the organism
that interprets the stimulus before producing a response (Woodworth, 1929). For S-OR psychologists, the link between S and R isn’t mindless or automatic. Instead, how the
organism responds to a stimulus depends on what this stimulus means to the organism.
The S-O-R principle helps to explain a phenomenon we’ve probably all encountered.
You’ve probably had the experience of giving one of your friends a mildly critical piece of
feedback (like “It bothers me a little bit when you show up late”) and another one of your
friends the same feedback, but found that they reacted quite differently. Perhaps one of
your friends accepted your gentle criticism gracefully with a polite apology, whereas the
other became defensive or even argumentative.
In attempting to explain these different reactions, Skinnerians would probably invoke
your friends’ different learning histories. Perhaps your first friend had repeatedly been
reinforced by others for apologizing (“That’s so nice of you to apologize; I really appreciate
it”), whereas your second friend had repeatedly been reinforced by others for becoming
defensive when criticized (“Oh, I’m really sorry I brought that up; I’ll try to be less critical
of you in the future’). In contrast, S-O-R theorists, who believe that cognition is central
to explaining at least some forms of learning, would contend that the differences in your
friends’ reactions stem from their differences in how they interpreted your criticism. Your
first friend may have viewed your criticism as constructive feedback, whereas your second
friend may have viewed it as a personal attack.
S-O-R theorists don’t deny that classical and operant conditioning occur, but they
believe that these forms of learning usually depend on thinking. Let’s imagine that we
classically conditioned a person using a tone paired repeatedly with an electric shock:
tone-shock, tone-shock, tone-shock, and on and on. As we noted earlier in the chapter,
the person – unless he or she is a psychopath - will show a pronounced skin conductance
response to the tone alone, which extinguishes gradually over time. But what do you think
would happen if we suddenly informed the subject that “Don’t worry – no more shocks
are coming”? If you guessed that the skin conductance response extinguishes much more
S-O-R PSYCHOLOGY
Today, few psychologists share Skinner’s exceedingly harsh assessment of cognitive
psychology. In fact, a substantial majority of psychologists now agree that the story of
130
131
S-O-R psychology
For S-O-R psychologists, the link between
S and R is how the organism responds
to a stimulus depends on what this
stimulus means to the organism.
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
rapidly, you’d be right (Grings, 1973). This phenomenon of cognitive conditioning,
whereby our interpretation of the situation affects conditioning, suggests that conditioning
is more than an automatic, mindless process (Brewer, 1974).
To explain psychology’s gradual transition from behaviorism to cognitivism, we first
need to tell the story of a pioneering psychologist and his rats.
LATENT LEARNING
Photograph 27
Edward Chase Tolman, one of the
unsung giants of modern psychology.
Tolman’s discoveries concerning latent
learning helped lay the groundwork for
modern cognitive theories of learning.
Figure 12
Photograph 28
Children, like the child who produced
this drawing, possess cognitive maps of
their home neighborhoods.
cognitive conditioning
whereby the interpretation of the
situation affects conditioning,
latent learning
that is, learning that isn’t directly
observable
cognitive maps
spatial representations
One of the first serious challenges to the radical behaviorist account of learning arose
from research conducted by Edward Chace Tolman (1886-1959), who is one of the unsung
giants of modern psychology (Photograph 27). Although few psychology undergraduates
(aside from those at the University of California at Berkeley, where the psychology building
bears his name) have ever heard of Tolman, his contribution to the psychology of learning
is difficult to overestimate.
Tolman suspected that, contrary to Watson, Thorndike, and others, reinforcement
wasn’t the be-and-end-all of learning. To understand why, let’s imagine that we asked
you a question based on your reading of the previous paragraph: “After what psychologist
is the University of California at Berkeley psychology building named?” If you’ve been
paying attention, you hopefully answered “Tolman.” Yet before we asked that question, you
knew the answer, even though you hadn’t had the chance to demonstrate it. According
to Tolman (1932), you had engaged in latent learning, that is, learning that isn’t directly
observable (Blodgett, 1929). We learn many things without showing them. Putting it
a bit differently, there’s a crucial difference between competence – what we know - and
performance –showing what we know (Bradbard, Martin, Endsley, & Halverson, 1986).
Why is this distinction important? Because it implies that reinforcement is not necessary
for learning. To demonstrate this point systematically, Tolman and Honzik (1930) ran
three randomly assigned groups of rats through a maze over a three week period (see
Figure 12). One group always received reinforcement – cheese – when it got to the end
of the maze. A second group never received reinforcement when it got to the end of the
maze. The first group made far fewer errors; there’s certainly no great surprise there. But
Tolman and Honzik’s third group of rats was the most interesting. This group received no
reinforcement for the first 10 days, and then started receiving reinforcement starting on the
11th day. As you can see in Figure 12, the rats in this group showed a large and abrupt drop
in their number of errors upon receiving their very first reinforcement. In fact, within only
a few days the number of their errors was not significantly different – and was even slightly
less than – the number of errors among the rats always receiving reinforcement.
According to Tolman, this finding means that the rats in the third group had been
learning all along. They just hadn’t bothered to show it because there was nothing to be
gained. Once there was something to be gained – cheese - they suddenly became miniature
maze masters.
How did they do it? According to Tolman (1948), the rats had developed cognitive
maps – that is, spatial representations – of the maze. If you’re like most undergraduates,
you were hopelessly confused the first day you arrived on campus. Perhaps you went to
your first class, and then got lost in the middle of campus trying to make it to your second
class. Perhaps you later started walking confidently toward what you thought was the
campus library, only to realize that you were headed directly toward the cafeteria. Over
time, however, you probably developed a mental sense of the layout of the campus, so that
you now hardly ever become lost. That internal spatial blueprint, according to Tolman, is a
cognitive map (Photograph 28).
In an especially clever demonstration of cognitive maps, three investigators
(McNamara, Long, & Wilke, 1956) had one set of rats run repeatedly through a maze to
get reinforcement. They put another set of rats in little moving “trolley car,” in which the
rats could observe the layout of the maze but not obtain the experience of running through
132
COGNITIVE MODELS OF LEARNING
it. When the researchers later gave the second group of rats the chance to run through the
maze, they did just as well as the rats in the first group. As rodent tourists in trolley cars,
they had acquired cognitive maps of the maze.
The latent learning research of Tolman and others struck a serious blow to strictly
behavioral models of learning, because this work suggested that learning could occur
without reinforcement. It also suggested that thinking, in the form of cognitive maps, plays
a central form in at least some forms of learning.
OBSERVATIONAL LEARNING
According to some psychologists, one important variant of latent learning is observational
learning: learning by watching others (Bandura, 1965). In many cases, we learn by
watching models: parents, teachers, and other figures who are influential to us. Many
psychologists regard observational learning as a form of latent learning because it allows us
to learn without reinforcement. We can merely watch someone else being reinforced for
doing something, and take our cue from them.
Observational learning makes awfully good evolutionary sense, because it spares us the
expense of having to learn everything first-hand (Bandura, 1977). The authors of your
book are not particular experts in skydiving, but from our observations of people who
have gone skydiving we have the distinct impression that it’s generally advisable to have
a parachute on before one jumps out of a plane. Note that we did not have to learn this
useful tidbit of advice by trial and error. If we had made such an error, we wouldn’t be here
to tell you about it. As a result, observational learning can spare us from serious, even lifethreatening, mistakes. Nevertheless, observational learning can also facilitate our learning
of maladaptive habits.
Observational learning of aggression. In classic research in the early 1960s, Alfred
Bandura and his colleagues demonstrated that children can learn to act aggressively
by observing aggressive role models (Bandura, Ross, & Ross, 1963). In one study, they
exposed preschool boys and girls to an adult (the model) whom they placed in a room with
a large Bobo doll, a doll that bounces back to its original position after being hit (Bandura,
Ross, & Ross, 1961). The experimenters randomly assigned some children to watch the
Photograph 29
A panel of photographs from Bandura and colleagues (1963) classic study on the observational learning of
aggression. Top panel: adult model attacking the Bobo doll. Bottom two panels: a boy and a girl emulating the
adult model by attacking the Bobo doll in much the same fashion.
133
observational learning
learning by watching others
(1) F; (2) F; (3) T; (4) F; (5) T
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
adult model playing quietly and ignoring the Bobo doll, and others to watch the adult
model punching the Bobo doll in the nose, hitting it with a mallet, sitting on it, and kicking
it around the room. As though that weren’t enough, the model in the latter condition
shouted out insults to the Bobo doll while inflicting violence: “Sock him in the nose,”, “Kick
him,” “Pow.” (Photograph 29).
Bandura and his co-workers then brought the children into a room with a variety of
appealing toys, including a miniature fire engine, jet fighter, and a large doll set. Just as
the children began playing with these toys, the experimenter interrupted them, informing
them that they needed to move to a different room. This interruption was intentional,
as the investigators wanted to frustrate the children to make them more likely to behave
aggressively. Then the experimenter brought them into a second room, which contained a
Bobo doll identical to that they had seen previously. On a variety of dependent measures,
Bandura and colleagues found that previous exposure to the aggressive model triggered
significantly more aggression against the Bobo doll than did exposure to the nonaggressive
model. The children in the former condition hit and yelled at the doll much as the adult in
the film clip had done, and they even imitated many of the adult model’s verbal insults. In
a later study, Bandura and his colleagues (Bandura et al., 1963) found essentially the same
results when they displayed the aggressive models to children on film rather than in person.
Correlation
Doesn’t Imply
Causdation
longitudinal designs
tracking individuals’ behavior over time
Media violence and real-world aggression. The Bandura studies and scores of later
studies of observational learning led psychologists to examine a theoretically and socially
important question: Does exposure to media violence, such as films or movies, contribute
to real-world violence – or what Bushman and Anderson (2001) call “violence in the
reel world”? The research literature addressing this question is as vast as it is confusing,
and could easily take up an entire book. So we’ll only briefly summarize some of the key
findings and issues here.
Hundreds of researchers using correlational designs have reported that children who
watch many violent television programs are more aggressive than other children (Wilson &
Herrnstein, 1985). These findings, though, don’t demonstrate that media violence causes
real world violence (Freedman, 1984). They could simply indicate that highly aggressive
children are more likely than other children to tune into aggressive television programs.
Alternatively, these findings could be due to a third variable, such as children’s initial
levels of aggressiveness. That is, highly aggressive children may be more likely than other
children to both watch violent television programs and to act aggressively.
Investigators have tried to get around this problem by using longitudinal designs,
which track individuals’ behavior over time. Longitudinal studies have shown that children
who watch many violent television shows commit more aggressive acts years later than do
children who watch fewer violent television shows, even when children are equated in their
initial levels of aggression (Huessman, Moise, Podolski, & Eron, 2003). (Figure 13). These
studies offer somewhat more compelling evidence for a causal link between media violence
and aggression than do simple correlational studies, but even they don’t prove the existence
of this link. For example, it’s possible that an unmeasured personality variable, like
impulsivity, or a social variable, like weak parental supervision, accounts for these findings.
Moreover, just because variable A precedes variable B doesn’t mean that variable A causes
variable B (if you really want to impress your friends at parties, you can tell them that the
error of concluding that variable A causes variable B because variable A precedes variable
B is called the post hoc ergo propter hoc fallacy, which we can translate as the “after this,
because of this fallacy”). For example, if we found that most common colds start with a
scratchy throat and a runny nose, we shouldn’t conclude that scratchy throats and runny
noses cause colds, only that they are early signs of a cold.
Still other investigators have tried to examine whether the link between media models
and later aggression holds up under strictly controlled conditions within the laboratory.
134
COGNITIVE MODELS OF LEARNING
In most of these studies, researchers have exposed subjects to either violent or nonviolent
media presentations, and then examined whether subjects in the former groups behaved
more aggressively, such as by yelling at the experimenter or delivering electric shocks to
another subject when provoked. In general, meta-analyses (see Chapter 3) of these studies
strongly suggest a causal association between media violence and laboratory aggression
(Wood, Wong, & Chachere, 1991).
Finally, some investigators have conducted field studies of the link between media
violence and aggression. In field studies, researchers examine the relation between
naturally occurring events and aggression in the real world. For example, sociologist David
Phillips (1983) found that the number of homicides increased by approximately one-eighth
following widely publicized boxing matches. Moreover, he found that when a white boxer
defeated an African-American, the victim of the murder was more likely to be an AfricanAmerican, whereas the converse was true when an African-American boxer defeated a
White. Nevertheless, some researchers have questioned these findings because the spike in
homicides was evident only on the third day after these fights, but not on other days (Baron
& Reiss, 1985). Thus, it’s possible that these results were due to chance.
Another investigator (Mitchell, 1986) conducted a field study of a small, isolated
mountain town in Canada that had no television prior to 1973 (she called it “Notel,” short
for “No Television”). Compared with school-age children in two other Canadian towns
that already had television, children in Notel showed a marked increase in physical and
verbal aggression two years later. Nevertheless, these findings are difficult to interpret
in light of a potential confound: at around the same time that Notel received television,
the Canadian government built a large highway that connected Notel to nearby towns.
This highway might have introduced the children in Notel to negative outside influences,
including crime from other cities.
So what can we make of the literature on media violence and aggression in the “reel
world”? We have four lines of evidence – correlational studies, longitudinal studies,
laboratory studies, and field studies – each with has its own strengths and weaknesses.
Correlational, longitudinal, and field studies tend to be strong in external validity, that is,
generalizability to the real world, and weak in internal validity, that is, the extent to which
they permit cause-and-effect inferences. Laboratory studies, in contrast, tend to be weak
in external validity and strong in internal validity. Yet despite their shortcomings, all four
types of studies point in the direction of at least some causal relation between media
violence and aggression (Anderson et al., 2003). Scientific conclusions are usually the most
convincing when we base them on findings from very different research designs, each with
a slightly different set of imperfections (Shadish, Cook, & Campbell, 2002). As a result,
most psychological scientists today agree that media violence contributes to aggression in
at least some circumstances (Anderson & Bushman, 2002; Bushman & Anderson, 2001).
Nevertheless, as we’ll learn in Chapter 12, it’s equally clear that media violence is only
one small piece of a complex and multifaceted puzzle. We can’t explain aggression by
means of media violence alone because the substantial majority of individuals exposed to
high levels of such violence don’t become aggressive (Herrnstein & Wilson, 1985).
Ruling Out
Alternative
Hypithesis
NEW FRONTIERS
Mirror Neurons and Observational
Learning
You find yourself alone in a new city, standing on line behind someone using
an automated teller machine. Like so many other cash machines, this one is
slightly – and annoyingly - different from all of the other ones you’ve seen.
You watch intently as the person in front of you swipes her card, pushes a few
field studies
study of events and aggression in the
real world
135
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
buttons, and grabs her money from the slot at the bottom of the machine.
Now it’s your turn, and you know exactly what to do. You learned by watching.
But how?
The question of how our brains engage in observational learning is still
shrouded in mystery. Nevertheless, neuroscientists have recently begun to
pinpoint a potential physiological basis for imitation.
When a monkey watches another monkey perform an action, such as
reaching for an object, a group of neurons in its prefrontal cortex, not far from
its motor cortex (see Chapter 4), become active (Rizzolatti, Fadiga, Gallese, &
Fogassi, 1996). These cells are termed mirror neurons because they’re the
same cells that would have become active had the monkey performed the
same movement. It’s almost as though these neurons are “imagining” what it
would have been like to engage in the behavior.
Mirror neurons appear to be remarkably selective in their activation. They
don’t light up when a monkey sees another monkey that remains stationary,
nor do they light up when a monkey sees a piece of food that another monkey
grabbed. Instead, they light up only when a monkey sees another monkey
engaging in an action, like grabbing a piece of food. Moreover, these neurons
seem to be tuned to extremely specific behaviors. For example, investigators
have found one mirror neuron in monkeys that fires only when either the
monkey himself or a person he’s observing grabs a peanut, and a different
mirror neuron that fires only when either the monkey himself or a person he’s
observing eats a peanut (Winerman, 2005).
Using PET scanning, researchers have identified a similar mirror neuron
system in humans (Gallese & Goldman, 1998), although they’ve yet to identify
specific individual mirror neurons, as they have in monkeys (Photograph 30).
No-one knows for sure what mirror neurons do or why they’re in our brains.
But some neuroscientists have conjectured that such neurons play a central
role in our capacity for empathy (Azar, 2005; Ramachandran, 2000). When
you see an athlete suffer an injury during a televised sporting event, like a
baseball player grimacing in agony after a bruising slide into home plate, you
wince in pain along with him. In some sense, you may be “feeling his pain,”
because the mirror neurons in your brain that correspond to his motor areas
are becoming activated.
mirror neurons
a group of neurons in the prefrontal
cortex, not far from the motor cortex
that become active
Photograph 30
In humans, the mirror neuron system lights up when we see someone performing a task, like reaching for
an object. This is the same area of the brain that would light up if we reached for the object ourselves.
136
COGNITIVE MODELS OF LEARNING
Some authors have gone further to speculate that mirror neuron
abnormalities play a central role in infantile autism (see Chapter 3), which is
marked not only by deficits in language and thinking, but in the capacity to
adopt the perspective of others (Dingelfelder, 2005). Interestingly, one group
of investigators found that the mirror neuron areas of autistic individuals
become less active than those of non-autistic individuals when observing
people’s hand movements (Theoret et al., 2005). Still, such findings are only
correlational. They don’t necessarily show that mirror neuron deficits cause or
contribute to autism; perhaps they are merely a consequence of the fact that
autistic individuals are less interested in others’ actions than are non-autistic
individuals.
Even so, the discovery of mirror neurons may ultimately provide valuable
insights into how we and other intelligent animals learn from others. This
discovery also helps us to appreciate that even when we’re alone, we’re often
not really alone. Even when you’re sitting by yourself on your couch watching
television, your brain and the brain of that baseball player sliding into home
plate may well be in sync, with your mirror neurons and his lighting up in
unison.
Latent learning and observational learning were by no means the only shots fired across
the bow of behaviorism. Another was fired by a mysterious German psychologist during
the First World War
INSIGHT LEARNING
During World War I, the German government assigned Wolfgang Kohler (1887-1967) to
the Canary Islands, not far off the coast of Africa (you might recall from Chapter 5 that
Kohler was one of the founders of Gestalt psychology). There, he served as director of a
primate facility. Some scholars believe that Kohler also served as a spy for the German
government while there (Ley, 1990), but nobody knows for sure.
Around the same time as psychologists were conducting the first latent learning studies,
Kohler (1925) was busily posing various problems to four chimpanzees. His favorite of the
four was an Einstein of ape named Sultan, who was especially adept at solving puzzles. For
example, in one case Kohler placed a tempting bunch of bananas outside of the cage, well
out of Sultan’s reach, along with two bamboo sticks inside of the cage. Neither stick was
long enough to reach the bananas. After what appeared to be some heavy-duty pondering,
Sultan suddenly figured out the solution: stick one bamboo stick inside the other, creating
one extra-long bamboo stick (Photograph 31).
In another case, Kohler suspended a bunch of bananas from the ceiling of the cage, well
out of reach of his hungry chimps. At the same time, he placed several boxes in the cage.
Again, after some thought, the chimps quickly figured out the answer to the problem: pile
the boxes on top of one other, treating them as what Kohler called “climb-up-ables,” ascend
to the highest box, and then jump up to grab the bananas (Photograph 32).
What was notable, according to Kohler, was that his chimpanzees appeared to experience
the “aha reaction” we discussed earlier. Their solutions to his problems didn’t appear to
reflect trial and error, as it did with Thorndike’s cats, but rather insight. That is, their
solutions seemed to resemble what we saw in Figure 8 (p.000). The chimps seemed to
suddenly “get” the solution to the problem, and from thereon in they got it right just about
every time.
Still, Kohler’s findings and conclusions weren’t without their shortcomings. Because
Kohler only videotaped some of his chimpanzees’ problem-solving, it’s difficult to rule out
the possibility that at least some of this his chimps had engaged in trial and error before
137
Correlation
Doesn’t Imply
Causdation
Photograph 31
Kohler found that Sultan, his “star” chimpanzee, discovered how to insert one
bamboo stick inside another to create
an extra-long stick, thereby allowing him
to obtain food.
Photograph 32
Kohler’s apes also figured out how to get
to a banana suspended well above their
heads: stack a bunch of boxes atop each
other, and then climb to the top box.
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
Figure 14 TBD
The results of the study by Trabasso and
Bower (1963) show that graphs, such as
those in Figure 13, can be misleading.
The subjects in Figure 13 were actually
learning by insight, not trial-and-error,
but they were learning at different times.
By lumping together the results of all of
the subjects, as was done in Figure 13,
we end up with a misleading picture.
figuring out each problem (Gould & Gould, 1994). Moreover, because the chimps were
often in the same cage, they may have engaged in observational learning. Still, Kohler’s
work strongly suggests that at least some smart animals can learn through insight rather
than trial and error.
There’s also good evidence that humans learn many problems through insight (Dawes,
1994). Many researchers didn’t initially recognize this phenomenon, because they were
fooled by their own graphs. If you look at Figure 14 you’ll see the results of a typical
investigation of problem-solving. In this kind of study, the researchers ask human
participants to sort various cards that differ in their shape (round, triangular, square,
and so on), color (green, blue, red, and so on), and number (1, 2, 3, and so on) into piles.
Participants need to figure out which of these three categories the experimenter has in
mind. On each trial, participants receive feedback from the experimenter about whether
they’re right or wrong. Figure 14 shows the results averaged across many participants. This
graph seems to provide evidence of gradual, trial-and-error learning, doesn’t it?
Don’t be fooled by this graph, because appearances can be deceiving. If you look at the
pattern of learning within each subject in Figure 14, you can see that individuals actually
are learning by insight. They figure out the answer suddenly, and then almost always get it
right after that. By lumping together the results across all subjects, the graph misleadingly
suggests a gradual, rather than a sudden, decline (Restle, 1962; Trabasso & Bower, 1963).
So people often do learn through insight, they just do so at different rates.
CONTIGUITY vs. CONTINGENCY
Both latent learning and insight learning provided compelling evidence that learning is
more than just an automatic, reflexive process. Yet another persuasive piece of evidence for
the role of cognition in learning derives from research on conditioning and how it works.
Figure 15 TBD
Demonstration of blocking. When a
new CS is completely redundant with a
preexisting CS, no learn to the new CS
occurs.
contiguity
belief that, learning depends simply on
associating stimuli with responses that
are close to them in time
blocking
the effect of one CS upon another CS
Contiguity. Some early psychologists, such as the behaviorist Edwin Guthrie (1886-1959),
believed that all learning was simply a matter of contiguity. That is, learning depends
simply on associating stimuli with responses that are close to them in time (Guthrie, 1930).
So, according to Guthrie, if we run away from a scary-looking cat the first we see it, we’re
more likely to do it again in the future. So long as stimulus (cat) and response (running
away) are close to each other in time, we’re more likely to repeat that response each time we
see that stimulus. Guthrie turned out to be wrong.
Blocking. The first hint that contiguity wasn’t the whole story came from research on a
classical conditioning paradigm called blocking (Kamin, 1969). If we repeatedly present
a rat with a CS of a light, followed by a UCS of an electric shock, it will eventually begin
to exhibit a CR of fear every time it sees the light (in these studies, researchers typically
measure fear by the extent to which the animal stops pressing a bar for food in the presence
of the CS). That’s standard aversive conditioning. But imagine that we now introduce
another CS – a tone that occurs at exactly the same time as the light – and pair it repeatedly
with the light, along with the electric shock again as the UCS. What happens when we now
present the second CS – the tone? (Figure 15).
The answer is surprising: nothing. The rat shows no response at all the tone, and
happily keeps pressing away at the bar for food. What happened? Because the tone was
completely redundant with the light, the light blocked the effects of the tone.
The research on blocking is important because it indicates that contiguity by itself can’t
explain learning. For a stimulus to be effective in producing learning, it must predict new
information. Because the second CS – the tone - didn’t predict anything new that the first
CS – the light – didn’t, learning didn’t take place. Redundant information isn’t informative.
This redundancy principle applies to many domains of human learning. Shortly after
the attacks of September 11, 2001, the U.S. government Department of Homeland Security
introduced a color-coded scheme of terrorist threat levels, with green being low, blue being
138
COGNITIVE MODELS OF LEARNING
guarded, yellow being elevated, and so on. The scheme was spectacularly unsuccessful,
and quickly became the butt of jokes on late-night talk shows. Had the Department of
Homeland Security boned up on their psychology of learning, they might have known why.
Because the colors are completely redundant with the words for threat levels (low, guarded,
elevated, and so on), they don’t convey any new information. As a consequence, almost
nobody remembered them (Photograph 33).
Contigency. The final nail in the coffin of the idea that contiguity explains learning comes
from ingenious work by Robert Rescorla (1967, 1988). In this research, Rescorla wanted
to distinguish between two different processes in classical conditioning: contiguity– the
closeness in time between the CS and UCS – or contingency – the extent to which the CS
predicts the UCS (Rescorla & Wagner, 1972). Rescorla astutely observed that high levels
of contiguity between CS and UCS can occur without any contingency between them. To
see why, look at Figure 16. There, you’ll see a series of CSs (tones) and UCSs (shocks) that
occur in close proximity to one another, but in which the correlation between CSs and
UCSs are zero. That is, in this case, the presence of the CS doesn’t predict the UCS at all.
To find out whether contiguity is sufficient to account for classical conditioning,
Rescorla (1967) presented rats with several different sequences of CSs (tones) and UCSs
(shocks) (see Figure 17). In all of them, the contiguity between CS and UCS remained the
same, because the CS was equally close in time to the UCS in all cases. Yet the contingency
- or correlation - between CS and UCS varied, with this association being low in some cases
and high in others. That is, in some cases CS barely predicted the UCS at all; in others, it
predicted it quite well.
Rescorla found that contingency made a big difference in learning. The better the
CS predicted the UCS, the more fear the rats showed. So when we hold contiguity
constant, contingency matters. This finding is important because it suggests that classical
conditioning isn’t merely a mechanical, automatic process of connecting CSs to UCSs, as
many early behaviorists believed. Instead, classical conditioning and perhaps all forms
of learning occur when organisms experience surprise, that is, when they find out that a
stimulus predicts something novel (Kamin, 1968; Rescorla, 1991).
So from a cognitive perspective on learning, organisms try to determine contingencies
between stimuli: does this stimulus tell me something useful or not? If it does, and only if
it does, will learning occur (Miller & Matzel, 1988). Still, this doesn’t mean that all forms
of learning require cognition. Because scientists have observed classical conditioning in
very simple animals, such as our sea snail friend Aplysia (Carew, Hawkins, & Kandel, 1983)
and perhaps even in single-celled organisms such as protozoa (Bergstrom, 1969), it’s likely
that certain forms of conditioning are automatic or mechanical (Kirsch et al., 2004). When
it comes to humans and other intelligent animals, however, conditioning probably depends
heavily on thinking.
FA C T A N D F I C T I O N I N PSYC H O LO GY SE L F -T EST 3
(1) According to Skinner, animals do not think or experience emotions.
(2) Proponents of latent learning argue that reinforcement is not necessary for learning.
(3) Research on observational learning demonstrates that children can learn aggression by
watching aggressive role models.
(4) Contiguity appears to be sufficient for learning.
Answers printed upside down on the bottom of the following page
Photograph 33
The color-coded threat system
developed by the U.S. Department of
Homeland Security. Because the colors
are completely redundant with the threat
words, most people don’t remember
them. Stimuli that do not provide new
information are difficult to learn.
Figure 16 TBD
Figure 17 TBD
Schematic diagram from Robert Rescorla’s work. In all four panels, the contiguity between CS and UCS is identical, but
as we move from panels 1 to 4, the
contingency from CS to UCS increases
– that is, the CS becomes increasingly
predictive of the UCS.
contingency
the extent to which the CS predicts the
UCS
139
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
Biological Influences on Learning
For many decades, most behaviorists regarded learning as entirely distinct from biology.
The animal’s learning history and genetic make-up were like two ships passing in the night.
Yet we now recognize this view as naïve, because our biology influences the speed and
nature of our learning in complex and fascinating ways. Here are three powerful examples.
Photograph 34
A blue-jay eating a monarch butterfly
– which is poisonous - then vomiting
several hours later. The blue-jay will
avoid monarchs from now on. Interestingly, some nonpoisonous butterfly species have evolved to resemble monarchs
to avoid being eaten by birds.
conditioned taste aversion
the fact that classical conditioning can
lead us to develop avoidance reactions
to the taste of food.
scapegoat food
a novel, unliked food
CONDITIONED TASTE AVERSIONS
In the 1970s, University of Pennsylvania psychologist Martin Seligman went out for dinner
with his wife. He ordered a filet mignon steak flavored with Sauce Béarnaise, his favorite
sauce. Approximately 6 hours later, while at the opera, Seligman felt nauseated and become
violently ill. He and his stomach recovered, but his love of Sauce Bearnaise didn’t. From
then on, Seligman couldn’t even think of, let alone taste, Sauce Bearnaise without feeling
like throwing up (Seligman & Hager, 1972).
The Sauce-Bernaise syndrome, also known as the phenomenon of conditioned taste
aversion, refers to the fact that classical conditioning can lead us to develop avoidance
reactions to the taste of food. Yet before reading on, ask yourself a question. Does
Seligman’s story contradict the other examples of classical conditioning we’ve discussed,
like that of Pavlov and his dogs? (Photograph 34).
In fact, it does in at least three ways (Garcia & Hankins, 1977). First, in contrast to
most classically conditioned reactions, which require repeated pairings between CS and
UCS, conditioned taste aversions typically require only one trial to develop. This difference
makes good evolutionary sense. We wouldn’t want to have to experience horrific food
poisoning again and again to learn a conditioned association between taste and illness.
Not only would doing so be incredibly unpleasant, but in some cases, we’d be dead after
only the first trial. Second, the delay between CS and UCS in conditioned taste aversions
can be as long as 6 or even 8 hours (Rachlin, 1991). Again, this makes good evolutionary
sense, because food poisoning often sets in many hours after eating toxic food. Third,
conditioned taste aversions tend to be remarkably specific, and show little evidence of
stimulus generalization. One of the authors of your book’s earliest childhood memories
is that of eating a delicious piece of lasagna and then becoming violently ill several hours
later. For over 20 years, he avoided lasagna at all costs while thoroughly enjoying spaghetti,
manicotti, veal parmigiana, pizza, and virtually every other kind of Italian food. He finally
forced himself to get over his lasagna-phobia, but not without a momentous struggle.
Many conditioned taste aversions aren’t rational. For example, such aversions are a
particular problem among cancer patients who undergo chemotherapy, which frequently
induces nausea and vomiting. As a consequence, these patients frequently begin to avoid
any food that preceded chemotherapy, even though they realize that it bears no logical
connection to the treatment. Fortunately, health psychologists (see Chapter 9) have
developed a clever way of getting around this problem. Capitalizing on the specificity of
conditioned taste aversions, they get cancer patients to eat an unfamiliar scapegoat food
– a novel food of which they aren’t fond – prior to the chemotherapy. In general, the taste
aversion becomes conditioned to the scapegoat food rather than to the patient’s preferred
foods (Andresen, Birch, & Johnson, 1990).
In an especially ingenious article, John Garcia and one of his colleagues helped to
unravel the mystery of conditioned taste aversions (Garcia & Koelling, 1966). Interestingly,
other scientists considered their findings so implausible that numerous journals rejected
their paper outright (Sternberg, 2003) before a relatively obscure outlet (Psychonomic
Science) finally accepted it. Nevertheless, Garcia eventually got his measure of intellectual
revenge. His article has been cited approximately 1100 times by other researchers in their
publications, which is an astonishing number considering that the modal number of
citations for a psychology journal article is 0 (very few psychology articles are cited as many
140
BIOLOGICAL INFLUENCES ON LEARNING
as 100 times). In psychology, as in other sciences, revolutionary ideas aren’t always well
received, at least initially.
Garcia and his co-author found that rats that had been exposed to X-rays – which make
them feel nauseated – developed conditioned fear reactions to the taste of water, whereas
rats that had been exposed to bright lights and buzzers don’t (Garcia & Koelling, 1966).
The bottom line is that conditioned taste aversions make evolutionary sense. In the real
world, poisoned drinks and foods, not sights and sounds, make animals feel sick. As a
consequence, animals develop conditioned taste aversions only to stimuli that tend to
trigger nausea in the real world (Figure 18).
This finding contradicts the assumption of equipotentiality – the claim that all CSs
can be paired equally well with all UCSs – a belief held by many traditional behaviorists
(Plotkin, 2004). Garcia and others had found that certain CSs, such as those associated
with taste, are easily conditioned to certain UCSs, such as those associated with nausea.
Psychologists call this phenomenon belongingness: certain stimuli are more likely than
others to go together with certain responses (Thorndike, 1911; Rachman, 1977). Recall
that following his night out with his wife, Martin Seligman felt nauseated at the thought of
Sauce Bernaise, but not at the thought of opera singers (in contrast, those of you who feel
nauseated at the very thought of going to the opera might experience a different reaction).
Figure 18 TBD
The work of John Garcia and his colleagues demonstrated that animals tend
to develop conditioned taste aversions
only to certain stimuli, namely those that
tend to trigger nausea in the real world.
PREPAREDNESS
A second serious challenge to the equipotentiality assumption comes from research on
phobias. If we look at the distribution of phobias in the general population, we find
something curious: the most widespread phobias aren’t those with which people have had
the most frequent negative experiences. Phobias of the dark, heights, snakes, spiders, deep
water, and blood are commonplace, even though many people who fear these stimuli have
never had a frightening encounter with them. In contrast, phobias of razors, knives, the
edges of furniture, ovens, and electrical outlets are extremely rare, although many of us
have had annoying, painful, or embarrassing experiences with them.
Seligman (1971) argued that we can explain the distribution of phobias in the
population by means of preparedness: we are evolutionarily predisposed to fear certain
stimuli more than others. According to Seligman, that’s because certain stimuli, like steep
cliffs and poisonous animals, posed a threat to our early human ancestors (Ohman &
Mineka, 2001). In contrast, household items and appliances didn’t. In the words of Susan
Mineka (1993), prepared fears are “evolutionary memories,” that is, emotional legacies of
natural selection.
In one intriguing demonstration of preparedness, Mineka and Cook (1993) showed
lab-reared rhesus monkeys, who had no previous exposure to snakes, a videotape of
other monkeys reacting in horror to snakes. Within less than half an hour, the monkeys
had acquired a fear of snakes by means of observational learning (surprisingly, rhesus
monkeys who’ve never been exposed to snakes show no fear to them). By means of
some clever editing, Mineka and Cook altered the videotape to show the same monkeys
reacting in horror, but this time in response to flowers, a toy rabbit, a toy snake, or a toy
crocodile. They then showed these new videotapes to different groups
of rhesus monkeys. The monkeys who observed these
altered videotapes acquired fears of the toy snake and
Photograph 35
Mineka and Cook (1993) showed that monkeys
can acquire fears of snakes by means of
observational learning. Nevertheless,
these monkeys didn’t acquire fears of
nondangerous stimuli, like flowers,
suggesting a role for evolutionary
predispositions in the
development of fears.
141
equipotentiality
the claim that all CSs can be paired
equally well with all UCSs
belongingness
certain stimuli are more likely than
others to go together with certain
responses
preparedness
evolutionary predisposition to fear
certain stimuli more than others
(1) F; (2) T; (3) T; (4) F
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
Figure 19 TBD
The Great Fourfold Table of Life helps to
explain how preparedness can lead us
to defelop illusory correlations between
stimuli and feared outcomes. If we’re
afraid of snakes, we often tend to attend
too highly to the upper left-hand cell of
the table, in which snakes co-occur with
frightening consequences.
Photograph 36
Will this protester develop a phobia
of dogs? Research suggestes that the
answer depends substantially on his genetic predisposition toward fearfulness.
Ruling Out
Alternative
Hypithesis
toy crocodile, but not to the flowers or toy rabbit. From the standpoint of preparedness,
this finding is entirely understandable. Snakes and crocodiles were dangerous to primate
ancestors, but flowers and rabbits weren’t (Ohman & Mineka, 2003). (Photograph 35).
Preparedness may also render us likely to develop illusory correlations between fearprovoking stimuli and negative consequences (Tomarken, Mineka, & Cook, 1989). Recall
from Chapter 3 that an illusory correlation is a statistical mirage; it’s the perception
of an association between two variables when this association is absent. One team of
investigators administered intermittent electrical shocks to subjects – some of whom feared
snakes and some of whom didn’t – while they were watching slides of snakes and damaged
electrical outlets. The pairings of the slide stimuli with the shocks were entirely random,
so that the actual correlation between them was zero. Yet subjects with high levels of snake
fear perceived a marked correlation between the occurrence of the snake slides, but not the
electrical outlets, and the electric shocks. Subjects with low levels of snake fear showed no
such tendency toward illusory correlation (Tomarken, Sutton, & Mineka, 1995).
If you think back to the Great Fourfold Table of Life we discussed in Chapter 3 (see
p. ...), the highly snake-fearful subjects probably attended too highly to the uppermost
lefthand cell of the table (Figure 19). Because they were afraid of snakes, they were
on the lookout for any threatening stimuli that might be associated with snakes. As a
consequence, they overestimated how often snake slides co-occurred with electric shock.
Interestingly, they showed no such overestimation for electrical outlets, even though such
outlets are more closely linked in our minds than snakes to electric shock. This finding
suggests that preparedness may be at work, because snakes, but not electrical outlets, posed
threats to our ancestors (Tomarken et al., 1995).
CRITICAL THINKING QUESTION
Could the tendency of fearful people to develop illusory correlations between their feared
stimuli and unpleasant outcomes actually be adaptive in certain circumstances? If so, how?
Nevertheless, the laboratory evidence for preparedness hasn’t been completely
consistent. For example, when researchers have paired either prepared stimuli - like snakes
or spiders - or unprepared stimuli - like flowers or mushrooms - with electric shocks,
they haven’t always found that subjects more rapidly acquire fears to the prepared than
unprepared stimuli (Davey, 1995; McNally, 1987). Moreover, some authors have proposed
that preparedness findings may be due to an alternative explanation: latent inhibition. As
you’ll recall from earlier in the chapter, latent inhibition refers to the fact that CSs that have
appeared alone (that is, without a UCS) many times are especially difficult to classically
condition to a stimulus. Because we regularly encounter electric sockets, stoves, knifes, and
the like, without experiencing any negative consequences, such stimuli may be especially
resistant to classical conditioning. In contrast, because few of us have daily encounters
with snakes, cliffs, deep water, and the like, such stimuli may be more easily to classically
condition to negative outcomes (Bond & Siddle, 1996).
Aside from preparedness, genetic influences seem to play an important role in the
acquisition of phobias. For example, dog phobics and non-dog phobics don’t differ in their
number of negative experiences with dogs, such as bites (DiNardo et al., 1988). Moreover,
only about half of dog phobics have even had a scary encounter with a dog; the same
findings hold for people with many other phobias. These results make it unlikely that
classical conditioning alone can explain all cases of phobia, although it almost certainly
plays some role. Instead, some people appear to be predisposed genetically to develop
certain phobias given a history of certain classical conditioning experiences (Kendler,
Neale, Kessler, Heath, & Eaves, 1992). (Photograph 37).
142
CLASSICAL AND OPERANT CONDITIONING IN POP PSYCHOLOGY
CRITICAL THINKING QUESTION
How might you attempt to test the hypothesis that both a genetic predisposition toward
phobias and a traumatic classical conditioning experience – but neither alone - are
required for the development of phobias?
INSTINCTIVE DRIFT
Marian and Keller Breland were Ph.D students of B.F. Skinner at Harvard. Both went on to
become well-known animal trainers. Using traditional methods of operant conditioning,
such as shaping, they taught chickens, raccoons, pigs, and a host of other creatures to
perform scores of tricks – much like those one might see on David Letterman’s “Stupid Pet
Tricks” - for circuses and television advertisers.
Yet in the process of their animal training adventures, the Brelands discovered that their
little mammalian charges didn’t always behave as anticipated. For example, in one case
they attempted to train raccoons to drop tokens into a piggy bank. Although they were
successful in training the raccoons to pick up the coins using food reinforcement, they soon
ran into an unexpected problem. Despite repeated reinforcement to pick drop the coins
into the piggy bank, the raccoons began rubbing the coins together, dropping them, and
rubbing them together again (Photograph 38).
The raccoons had apparently reverted to an evolutionarily selected behavior, namely
rinsing. They were treating the tokens like little pieces of food, like the small hard shells
that raccoons extract from the beds of ponds and streams (Timberlake, 2006). Breland
and Breland (1961) referred to this phenomenon as instinctive drift: the tendency for
animals to return to evolutionarily selected behaviors following repeated reinforcement.
Researchers have observed various forms of instinctive drift in a number of other animals,
including rats (Powell & Curley, 1984). The reasons for such drift aren’t fully understood.
Nevertheless, instinctive drift demonstrates that reinforcement has its limits. When it
comes face to face with evolution, evolution often emerges victorious.
Classical and Operant Conditioning in
Pop Psychology:
Do Learning Fads Work?
Those of you persevering souls who have made it all the way to this point in the chapter
know that learning new information is hard work. To have followed the material we’ve
presented on earning took an enormous amount of mental energy and concentration on
your part.
Perhaps because learning new things requires so much time and effort, many mental
health professionals have marketed a variety of techniques that supposedly help people to
learn more quickly, or more easily, than they currently do. Do these newfangled methods
work? We’ll find out by examining three of the most popular techniques.
SLEEP-ASSISTED LEARNING
Imagine that you could master all of the information in this book while getting a few nights
of sound sleep. You could pay someone to tape record the entire book, play the recording
over the span of several weeknights, and you’d be all done. You could say goodbye to late
nights in the library or dorm room reading about psychology.
As in many areas of psychology, hope springs eternal. Indeed, many proponents
of sleep-assisted learning – learning new material while asleep – have made some
143
instinctive drift
the tendency for animals to return
to evolutionarily selected behaviors
following repeated reinforcement
sleep-assisted learning
learning new material while asleep
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
extraordinary claims regarding this technique’s potential. For example, one Web Site
(http://www.sleeplearning.com/) informs visitors that
Sleep learning is a way to harness the power of your subconscious while you
sleep, enabling you to learn foreign languages, pass exams, undertake
professional studies and implement self-growth by using techniques based on
research conducted all over the world with great success....It’s the most incredible
learning aid for years.
Hume’s Dictum
Ruling Out
Alternative
Hypithesis
The Web site offers a variety of CDs that can purportedly help us to learn languages,
stop smoking, lose weight, reduce stress, or become a better lover, all while we’re
comfortably catching up on our zzzzs. The site even goes so far as to say that the CDs work
better when people are asleep than awake.
These assertions are certainly quite remarkable. Does the scientific evidence for sleepassisted learning stack up to its proponents’ claims?
As is so often the case in life, things that sound too good to be true often are.
Admittedly, the early findings on sleep-assisted learning seemed to yield encouraging
results. For example, one group of investigators exposed sailors to Morse Code (a
shorthand form of communication that radio operators sometimes use) while asleep.
These sailors mastered Morse Code three weeks faster than did other sailors (Simon &
Emmons, 1955). Numerous other studies from the former Soviet Union also appeared
o provide support for the claim that people could learn new material, such as words or
sentences, while listening to tape recordings during sleep (Aarons, 1976). (Photograph 39).
Nevertheless, these early positive reports neglected to rule out an important alternative
explanation: the tape recordings may have awakened the subjects!
The problem is that almost of the studies showing positive effects didn’t monitor
subjects’ electroencephalograms (EEGs; see Chapter 4) to ensure that they were actually
asleep while listening to the tapes (Druckman & Swets, 1988, 1994). When researchers have
monitored subjects’ EEGs to make sure that they’re fully asleep, they’ve typically produced
little or no evidence for sleep-assisted learning. Therefore, to the extent that sleep-learning
tapes actually “work,” it’s probably because subjects hear snatches of them while drifting
in and out of sleep. As for that quick fix for reducing stress, we’d recommend just going to
sleep and forgetting about the tapes.
ACCELERATED LEARNING
Hume’s Dictum
Ruling Out
Alternative
Hypithesis
Still other companies promise consumers ultra-fast techniques for learning. These
methods, known as Superlearning or Suggestive Accelerative Learning and Teaching
Techniques (SALTT), supposedly allow people to pick up new information at anywhere
from 25 to several hundred times their normal speed of learning (Wenger, 1983). SALTT
relies on a mixture of different several techniques, such as inducing expectations for
enhanced learning (telling students that they’ll learn more quickly), getting students to
visualize information that they’re learning, playing classical music during learning, and
even breathing in a regular rhythm while learning (Lozanov, 1978). When combined, these
techniques supposedly allow learners to gain access to intuitive and creative aspects of their
minds that otherwise remain inactive.
Again, however, the evidence for the effectiveness of SALTT doesn’t come close to
matching the extraordinary claims. Almost all studies show that SALTT doesn’t result in
enhanced learning (Dipamo & Job, 1990; Druckman & Swets, 1988). Moreover, even when
researchers have reported positive results for SALTT, these findings have been open to
several alternative explanations. That’s because many of the studies conducted on SALTT
compared this method with a control condition in which the students did little or nothing.
As a result, the few positive results reported for SALTT could be attributable to placebo
effects (see Chapter 3), especially because one of the major components of SALTT is raising
144
CLASSICAL AND OPERANT CONDITIONING IN POP PSYCHOLOGY
learners’ expectations. These scattered positive results could also be due to the Hawthorne
effect (again, see Chapter 3), because all of the students exposed to SALTT knew that
they were being studied, and knew that the experimenter expected them to learn faster
(Druckman & Swets, 1998).
CRITICAL THINKING QUESTION
If you were studying the effects of accelerated learning methods, like SALTT, how would you
attempt to rule out the possibility of the Hawthorne effect?
LEARNING STYLES
Few claims about learning are as widespread as the belief that all individuals have their
own learning style – their preferred means of acquiring information. According to
proponents of this view, some students are “analytical” learners who excel at breaking down
problems into different components, whereas others are “holistic” learners who excel at
viewing problems as a whole. Still others are “verbal” learners who prefer to talk through
problems, whereas others are “spatial” learners who prefer to visualize problems in their
heads (Cassidy, 2004; Desmedt & Valcke, 2004). Moreover, some educational psychologists
have claimed that they can dramatically boost learning by matching people’s learning styles
to different methods of instruction. So, for example, children who are verbal learners
should learn much faster and better with written material, whereas children who are spatial
learners should learn much faster and better with visual material.
As appealing as these assertions are, they haven’t held up well in careful research. For
one thing, it’s difficult to assess learning style reliably (Snider, 1992; Stahl, 1999). As
you’ll recall from Chapter 3, reliability refers to consistency in measurement. In this case,
researchers have found that different measures designed to assess people’s learning styles
often yield very different answers about people’s preferred mode of learning. In part, that’s
probably because few people are purely analytical or holistic learners, verbal or spatial
learners, and so on; most of us are a mixture of both kinds of styles. Moreover, studies have
generally revealed that tailoring different methods to people’s learning styles doesn’t result
in enhanced learning (Kavale & Forness, 1987; Tarver & Dawson, 1978). Like a number of
other fads in popular psychology, the idea of learning styles appears to be more fiction than
fact (Stahl, 1999).
Photograph 38
Proponents of sleep-assisted learning
claim that their products will allow users
to learn large amounts of new material
while asleep? Does the evidence support their claims?
FA C T A N D F I C T I O N I N PSYC H O LO GY SE L F -T EST 4
(1) Many conditioned taste aversions are acquired in only a single trial.
(2) Most research now suggests that the assumption of equipotentiality, held by many
traditional behaviorists, is false.
(3) The phenomenon of preparedness helps to explain why virtually all major phobias are
equally prevalent in the general population.
(4) With progressively more reinforcement, animals typically drift further and further away
from their instinctual patterns of behavior.
(5) There is little evidence that people can learn new information while sound asleep.
Answers printed upside down on the bottom of the following page
learning style
preferred means of acquiring information
145
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
CRITICAL POINTS
Habituation pages 000-000
“Tough Dames”
• Habituation refers to the process by
which we respond less strongly over time
to repeated stimuli.
“Apparently, southern magnolias aren’t the only women who are
made of steel. According to researchers at the University of Michigan, women with severe heart disease are as likely to rate their
condition “mild to moderate” as men with less-severe illness. Eva
Kline-Rogers, MS, RN, one of the researchers who performed the
study, points out that the women who labeled their disease “mild
to moderate” were sicker overall than the men who rated their disease the same way. In the past few years, we have come to know
that women typically experience heart disease differently than men,
reporting fatigue, back pain and shortness of breath rather than the
classic male symptoms of chest pain and numbness in the arm. This
study raises still more questions about gender difference in heart
disease. Are women less inclined to complain? Are they unwilling
to admit to weakness? Researchers can’t answer those questions
yet, but some suspect the answers might shed light on why heart
disease in women is often underdiagnosed and untreated.”
(Newsweek, February 13, 2006).”
Analysis.
This brief article deals with the
important issue of gender differences
in health and illness. It provides an
alternative explanation for an
intriguing and potentially important
gender difference, and it calls
attention to the need for future
research on gender issues and the
diagnosis and treatment of heart
disease.
Critical thinking questions.
Can you think of any other alternative
explanations for the gender
differences? What sorts of additional
studies would further our
understanding of the differences in
symptom reports of men and
women?
• It is the simplest form of learning, and is
evident in amoebas and unborn fetuses.
• Habituation is evolutionarily adaptive, as
it helps us to ignore stimuli that pose no
threat to us.
Classical conditioning pages 000-000
• Classical conditioning was discovered
largely by accident by the Russian
physiologist Ivan Pavlov.
• Classical conditioning is a form of
learning in which animals come to respond
to a previously neutral stimulus that had
been paired with another stimulus that
elicits an automatic response.
• In classical conditioning, the conditioned
stimulus (CS) is a previously neutral
stimulus that yields no particular response
from the organism. After repeated pairings
with the unconditioned stimulus (UCS),
which elicits an automatic, reflexive,
response from the organism, the CS comes
to elicit a conditioned response (CR).
• Aversive conditioning is one variant of
classical conditioning. It involves pairing a
CS with an unpleasant UCS.
• Classical conditioning has adaptive
value. For example, psychopaths seem
largely unable to develop certain
classically conditioned reactions to
aversive stimuli. As a consequence, they
experience difficulty learning from
punishment.
• The three major phases of classical
conditioning are acquisition, extinction,
and spontaneous recovery.
• Acquisition is the process by which we
gradually learn the CR.
• There are four different methods of
acquisition: delay conditioning, trace
conditioning, simultaneous conditioning,
and backward conditioning. Delay and
trace conditioning tend to be more
effective than simultaneous and backward
conditioning.
146
• The converse of habituation is
sensitization, which refers to responding
more strongly over time to repeated stimuli.
• Extinction is the process whereby following
repeated presentation of the CS alone, the
CR decreases in magnitude and eventually
disappears.
• Extinction appears to involve not a decay of
the CR, but rather an “overwriting” of the
CR by new information.
• Spontaneous recovery occurs when we wait
a period of time following extinction and
then present the CS alone; when we do so,
the CR reappears.
• Closely related to spontaneous recovery is
the renewal effect, which occurs when we
extinguish a response in a setting
different from that in which was acquired,
but then place the animal back in the
original setting.
• Stimulus generalization refers to the
process by which CSs similar, but not
identical, to the original CS elicit a CR.
It occurs along a generalization gradient,
whereby the more similar the new CS is the
original CS, the stronger the CR.
• Stimulus discrimination is the flip side of
the coin from stimulus generalization; it
occurs when we exhibit a CR to certain CSs
but not others.
• Higher-order conditioning occurs when
organisms develop classically conditioned
responses to other CSs associated with the
original CS.
• Advertisers frequently make us of higherorder conditioning to pair images of
products with pleasant CSs.
147
Critical Points
SCIENCE IN THE NEWS
• Many fears and phobias appear to stem in
part through classical conditioning. John
B. Watson and Rosalie Raynor’s ethcally
questionable case study of Little Albert
provided an existence proof of this point,
although not all researchers successfully
replicated their observations.
• Classical conditioning also appears to
contribute to the acquisition of fetishes in
animals and perhaps humans.
• Many disgust reactions are probably
acquired by classical conditioning.
• Research also indicates that phobias can
be eliminated through deconditioning.
Operant conditioning pages 000-000
• Operant conditioning is learning that is
controlled by its consequences.
• Operant conditioning differs from
classical conditioning in three major ways:
(1) the response is emitted by the
organism rather than elicited from it,
(2) the reward is contingent on the
animal’s behavior, and
(3) the organism’s responses most
involve the skeletal muscles rather than
the autonomic nervous system.
• Thorndike’s Law of Effect tells us that if a
response, in the presence of a stimulus,
is followed by a reward, it is likely to be
repeated. According to this law, operant
conditioning involves the gradual
“stamping in” of S-R connections.
• Thorndike discovered this law in his classic
research on cats and puzzle boxes. This
work appeared to indicate that cats learned
by trial-and-error rather than insight.
• B.F. Skinner pioneered the study of operant
conditioning through his development of
the Skinner Box, which permitted a
cumulative record of the animal’s behavior.
• Reinforcement refers to any outcome that
strengthens a response.
• There are two types of reinforcement:
positive, which involves the presentation
of a pleasant outcome, and negative, which
involves the withdrawal of an unpleasant
outcome.
• According to Skinner, punishment has
several disadvantages. It only tells the
organism what not to do, what to do. In
addition, punishment often creates anxiety,
which interferes with future learning.
Punishment can also teach the organism
to become sneakier about which
situations in which to exhibit forbidden
behavior. Finally, punishment can provide a
model for children’s aggressive behavior.
• Although studies indicate that parental
punishment is associated with childhood
behavior problems, this correlational
research is open to multiple interpretations.
• Punishment seems to work best when it is
delivered consistently and follows the
desired behavior promptly.
• A discriminant stimulus is any stimulus that
signals the presence of reinforcement.
• Acquisition, extinction, spontaneous
recovery, and stimulus generalization and
discrimination apply to operant
conditioning as well as classical
conditioning. These similarities have led
some scholars to propose that operant and
classical conditioning may not be as
different as once thought, although it is
worth noting that these two forms of
learning appear to depend on different
brain areas.
• Negative reinforcement shouldn’t be
confused with punishment, which refers to
an outcome that weakens a response.
• Skinner uncovered numerous principles
of reinforcement, including the principle
of partial reinforcement, which indicates
that behaviors reinforced only occasionally
are slower to extinguish that behaviors that
we reinforce continually.
• Punishment must also be distinguished
from the disciplinary practices that are
often associated with it, because some
forms of discipline are actually reinforcing.
• Skinner also discovered various schedules
(contingencies) of reinforcement, including
fixed ratio, fixed interval, variable ratio, and
variable interval schedules.
148
CRITICAL POINTS
• These schedules tend to yield distinctive
patterns of responding; for example, fixed
interval schedules tend a scalloped pattern
of responding, whereas variable ratio
schedules tend to yield the highest overall
rate of responding.
• The gambler’s fallacy is the incorrect belief
that random events have a memory. Along
with variable ratio schedules, it probably
contributes to some individuals’
susceptibility to gambling.
• Behaviorists train animals to learn habits
using shaping by means of successive
approximations (shaping), with which we
reinforce progressively closer version of a
desired response.
• According to the Premack Principle, we
can reinforce lower frequency behaviors
with higher frequency behaviors.
• Many superstitions may stem from operant
conditioning; they may arise when
behavior is linked to reinforcement by
sheer coincidence.
• An important application of operant
conditioning is the use of token
economies: systems, often set up in
psychiatric hospitals, for reinforcing
appropriate behaviors and extinguishing
inappropriate behaviors.
• Using shaping, behaviorists have also used
operant conditioning techniques to teach
language to autistic individuals.
• Two-process theory involves both
classical and operant conditioning, and
helps to explain why many phobias do not
extinguish. Individuals may acquire fears
through classical conditioning, then avoid
situations that trigger anxiety, thereby
negatively reinforcing their phobic
behavior.
Cognitive models of learning pages 000-000
• Watson was an advocate of
methodological behaviorism; he believed
that psychology should focus exclusively
on overt behavior, and exclude thinking.
Skinner, in contrast, was an advocate
of radical behaviorism; he believed that
psychology should examine thinking, but
he regarded thinking as a behavior, subject
to the same laws of conditioning as all
other behaviors.
• S-O-R psychologists reject Skinner’s belief
that thinking is no different from any other
behaviors. They accord a central role to
the organism’s interpretation of stimuli.
• Tolman’s work on latent learning, which
demonstrated that animals can learn
without reinforcement, represented a
major challenge to the radical behaviorist
view of learning.
• Tolman believed that animals develop
cognitive maps – spatial representations
– of the world around them. Such maps,
he argued, are central to maze learning.
• Many psychologists regard observational
learning – learning by watching others – as
an important form of latent learning.
• Research from correlational studies,
longitudinal studies, laboratory studies,
and field studies suggest that media
violence contributes to aggression.
Although each type of research has its
flaws, they all point in the same direction.
Nevertheless, media violence is only one of
many causal contributors to aggression.
• Kohler’s work suggested that apes can learn
through insight, and later work with
humans suggests the same conclusion.
This research calls into question
Thorndike’s conclusion that all learning
occurs through trial and error.
• Some early behaviorists argued that all
learning depends on contiguity, that is,
proximity between stimuli and responses.
• Nevertheless, research on blocking
indicates that contiguity cannot by itself
explain learning. The CS must predict new
information to lead to learning.
• Rescorla’s work on the CS-UCS contingency
points to the same conclusion. CSs that
are highly predictive of UCSs lead to the
best learning.
• Research suggests that individuals can
acquire aggressive behavior by means of
observational learning.
149
Critical Points
Critical Points
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US
Biological Influences on Learning pages 000-000
• Most psychologists have increasingly
recognized that our biology, especially our
genetic endowment, influences learning.
• Conditioned taste aversions refer to the
phenomenon whereby classical
conditioning can lead us to develop
avoidance reactions to the taste of food.
• Such aversions differ from traditionally
classically conditioned reactions in three
ways: they typically require only one trial
to develop, they can develop even after
long delays in the CS-UCS pairing, and they
tend to show little stimulus generalization.
• John Garcia and his colleagues showed
that conditioned taste aversions violate
the principle of equipotentiality, because
they demonstrated that certain CSs are
more easily conditioned to certain UCSs.
Psychologists call this phenomenon
belongingness.
• Research on preparedness also contradicts
the equipotentiality assumption, because it
suggests that we are evolutionarily
predisposed to fear some stimuli more
than others.
• Preparedness can also lead us develop
illusory correlations between fearprovoking stimuli and negative
consequences.
• Nevertheless, laboratory research on
preparedness has not been entirely
consistent, and some researchers have
suggested alternative explanations for
preparedness findings.
• In the process of training animals to
perform tricks, the Brelands discovered the
phenomenon of instinctive drift: following
repeated reinforcement, some animals
tend to return to evolutionarily selected
behaviors.
Learning Fads: Do They Work? pages 000-000
• Proponents of sleep-assisted learning
claim that individuals can lead new
material while sound asleep.
• Nevertheless, most research on sleepassisted learning is negative. Early reports
of successful learning during sleep appear
to be attributable to a failure to carefully
monitor subjects’ EEGs to ensure that they
were actually asleep while listening to the
tapes.
• Studies of accelerated learning techniques
also show few or no positive effects.
Positive reports appear to be attributable
to placebo effects, Hawthorne effects, and
other artifacts.
150
• Some educational psychologists claim to
be able to boost learning by matching
individuals’ learning styles with different
teaching methods.
• Nevertheless, learning styles are difficult to
assess reliably; moreover, studies that have
matched learning styles with teaching
methods have typically yielded negative
results.
REFERENCES
Anderson, C.A., Berkowitz, L., Donnerstein, E., Huesmann, L.R., et al (2003). The
influence of media violence on youth. Psychological Science in the Public Interest,
4,81-110.
Anderson, C. A., & Bushman, B. J. (2002). The effects of media violence on society.
Science, 295, 2377-2378.
Bandura, A., Ross, D., & Ross, S.A. (1961). Transmission of aggression through imitation
of aggressive models. Journal of Abnormal and Social Psychology, 63, 575-582.
Bandura, A., Ross, D., & Ross,S.A. (1963). Imitation of film mediated aggressive models.
Journal of Abnormal and Social Psychology, 66, 3-11.
Blodgett, H. C. (1929). The effect of the introduction of reward upon the maze
performance of rats. University of California Publications in Psychology, 4, 113-134.
Bouton, M.E. (1994). Context, ambiguity, and classical conditioning. Current Directions
in Psychological Science, 3, 1-5.
Breland, K., & Breland, M. (1966). The misbehavior of organisms. American
Psychologist, 16, 681-684.
Brennan, P.A., & Mednick, S.A. (1994). Learning theory approach to the deterrence of
criminal recidivism. Journal of Abnormal Psychology, 103, 430-440.
Bushman, B. J., & Anderson, C. A. (2001). Media violence and the American public:
Scientific facts versus media misinformation. American Psychologist, 56, 477-489.
Denniston, J.C., Chang, R., & Miller, R.R. (2003). Massive extinction prevents
the renewal effect. Learning and Motivation, 34, 68-86.
Druckman, D., & Bjork, R.A. (Eds.). (1994). Learning, remembering, believing:
Enhancing human performance. Washington, D.C.: National Academy Press.
Druckman, D., & Swets, J.A. (Eds.) (1988). Enhancing human performance: Issues,
theories, and techniques. Washington, D.C.: National Academy Press.
Emmelkamp, P.M.G., & Kamphuis, J.H. (2005). Aversion Relief. In M. Hersen & J.Rosqvist
(Eds.), Encyclopedia of behavior modification and cognitive behavior therapy. Volume
I: Adult clinical application (pp. 39-40).. Sage, Thousand Oaks.
Fuller R.K. & Roth, H.P. (1979). Disulfiram for the treatment of alcoholism. An evaluation
in 128 men. Annals of Internal Medicine, 90, 901-904.
Garcia, J., & Koelling, R.A. (1966). The relation of cue to consequence in avoidance
learning. Psychonomic Science, 4, 123-124.
Gewirtz, J.C., & Davis, M. (2000). Using Pavlovian “higher-order” conditioning paradigms
to investigate the neural substrates of emotional learning and memory. Learning and
Memory, 7, 257-266.
Gorn, G.J. (1982). The effects of music in advertising on choice behavior: A classical
conditioning approach. Journal of Marketing, 46, 94-101.
Gresham, L.G., & Shimp, T.A. (1985). Attitude toward the advertisement and bran
attitudes: A classical conditioning perspective. Journal of Advertising, 14, 10-17, 49.
Hock, R.R. (2002). Forty studies that changed psychology: Explorations into the history
of psychological research (4th edition). Upper Saddle River, N.J.: Prentice Hall.
Kamin, L.J. (1968). “Attention-like” processes in classical conditioning. In M.R. Jones
(Ed.), Miami symposium on the prediction of behavior: Aversive stimuli. Miami, FL:
University of Miami Press.
Kirsch, I., Lynn, S.J., Vigorito, M., & Miller, R.R. (2004). The role of cognition in classical
and operant conditioning. Journal of Clinical Psychology, 60, 369-392.
Kohler, W. (1925). The mentality of apes. New York: Harcourt.
Lieberman, D.A. (1990). Learning: Behavior and cognition. Belmont, CA: Wadsworth.
Lykken, D.T. (1957). A study of anxiety in the sociopathic personality. Journal of
Abnormal and Social Psychology, 55, 6-10.
Lykken, D.T. (1995). The antisocial personalities. Hillsdale, N.J.: Lawrence Erlbaum.
Mineka, S., and Cook, M. (1993). Mechanisms involved in the observational conditioning
151
References
Critical Points
CHAPTER 7 LEARNING: HOW NURTURE CHANGES US