1991 (PART II)


For many people the mention of behavior therapy or behavior

modification calls up images of rats in a cage pressing levers for

pellets of food, or electric shocks being applied to the bodies of

criminals to change their violent ways. These images do reflect,

in a distorting mirror, the historical origins of behavior

therapy, which lie in efforts to change human behavior by applying

principles derived mainly from experiments on laboratory animals.

But behavioral theory was never as mechanical or shallow as such

popular images suggest, and the contemporary forms of behavior

therapy have evolved a long way from their beginnings.

Behavioral learning theory, derived from experiments by Pavlov and

others in the early twentieth century, was first used

systematically in psychotherapy in the 1950s. According to the

theory, learning or conditioning is the process by which behavior

is systematically and lastingly changed. There are two basic kinds

of learning: classical (or respondent) and operant (or

instrumental). Normal, adaptive, or desirable behavior is learned,

whether by classical or by operant conditioning, in the same ways

as abnormal, unwanted, or undesirable behavior, only the specific

stimuli and responses are different.


In classical conditioning, a neutral stimulus is presented to an

animal or a person along with a stimulus that causes an

unconditioned response -- that is, one motivated by a biological

need like food, sex, escape from danger, or relief of pain. After

a while the organism develops a conditioned response to the

neutral stimulus. In a famous experiment of Pavlov's, the

unconditioned response was salivation in the presence of food.

Pavlov's dog, repeatedly fed shortly after a bell was rung,

eventually began to salivate when it heard the bell even if it was

not fed. Once a conditioned response develops, the originally

neutral stimulus can be used to establish other conditioned

responses: flash a light when the bell is rung, and eventually the

dog will salivate simply on seeing the light. The stimulus loses

its power to evoke a conditioned response only when it is

presented many times without satisfaction of the need or drive;

this is known as extinction. Experiments on animals with simple

brains have demonstrated a physiological basis for processes very

much like classical conditioning and extinction. In a certain

species of snail, these forms of behavior change are associated

with variations in the amount of neurotransmitter chemicals passed

between brain cells in response to a stimulus.


Operant conditioning is the process by which an action comes to be

repeated because of consequences that promote (reinforce) the

tendency to perform it. One way to increase the likelihood that a

response will recur is reward, or positive reinforcement; for

example, a hungry animal will repeat an action if it is fed

afterward. The behavior is eventually extinguished (ceases) when

the reward is no longer supplied. Negative reinforcement increases

the probability of behavior that allows the animal to escape an

aversive (unwanted, unpleasant) stimulus; for example, the animal

will learn to perform an action that prevents an electric shock.

Punishment, on the other hand, serves to reduce the probability of

the action it follows. In other words, both positive and negative

reinforcements encourage or strengthen behavior, and punishment

discourages or weakens it. Punishment may, however, be used as a

reinforcer -- that is, a certain type of behavior can be

negatively reinforced (promoted) if alternative behavior is

punished. Certain objects like food or sexual satisfaction are

primary reinforcers. When repeatedly followed by primary

reinforcement, a neutral stimulus becomes a secondary reinforcer.

For example, if a hungry child is fed whenever she sits in a

certain chair (the neutral stimulus), the child will soon learn to

climb into the chair when she is hungry; the chair has become a

secondary reinforcer. For adults, money is one usually effective

secondary reinforcer.

Secondary reinforcement is a bridge between simple biological

drives and complex human or animal behavior. In classical

conditioning, responses are defined by the type of stimulus that

elicits them, one simple stimulus to one simple response: an

electrical shock causes an animal to withdraw its foot. But most

complex forms of behavior, and especially actions considered to be

voluntary, are not elicited by a single simple type of stimulus,

and must be learned by operant conditioning. A number of acts that

seem superficially different can be grouped together if they all

produce the same pattern of reinforcement. You can bring an

elevator to the lobby by pushing the button with a thumb, a

forefinger, or an elbow, or even by asking someone else to push

it. All these actions might be regarded as a single act or operant

for the purpose of evaluating reinforcements. Classical and

operant conditioning usually occur at the same time; for example,

a mother's scolding has the operant effect of preventing a child's

mischievous behavior, but it may also cause a classically

conditioned response of fear in the child whenever the mother is

near. Strict behavioral learning theorists admit the existence of

feelings, thoughts, and other states of mind, but regard them as

irrelevant because they are difficult to study experimentally.

Critics of behaviorism object to this dismissal of consciousness.

They say that the actor's intentions and purposes are smuggled

into the description of operant learning to disguise the

inadequacy of stimulus and response patterns as an explanation for

complex behavior.


Whatever the limitations of learning theory, modern behavior

therapy is no longer dependent on it. Symptoms treated with

methods derived from learning principles do not have to arise from

conditioning. Even if the cause of a phobia is hereditary, it can

be eliminated by desensitization, a behavioral technique. Nor do

most contemporary behavior therapists still regard states of mind

as irrelevant. In the early animal experiments, thoughts and

feelings were ignored because they were difficult to infer. When

behavior therapists began to work with human beings, they became

interested in what are sometimes called internal or mediating

stimuli and responses. They paid more attention to social and

emotional influences, introducing the terms "cognitive behavioral

therapy" and "social learning theory." Today some therapists even

combine behavioral treatment with other kinds of psychotherapy

based on principles apparently incompatible with learning theory.

By extending its scope, behavior therapy has lost some of its

distinctiveness. All definitions seem either too narrow or

overinclusive, and there is probably no feature common to all

behavioral techniques. But most behavior therapists could at least

agree on some points: their work is based on research in

experimental and social psychology; they are concerned about the

present more than the past, and actions rather than personality;

they try to "operationalize" terms referring to subjective states

(anxiety, depression, obsession, and so on) by linking them to

specific patterns of action; they are committed in principle to

defining problems precisely and breaking them down into behavioral

components, setting goals in advance, and systematically

evaluating the results as therapy proceeds. Behavior therapists

derive their theory from the experimental laboratory rather than

the consulting room, and they see themselves as teaching or

training rather than curing illness. Therefore they often ignore

or reject diagnosis in the traditional medical sense, and most

practitioners of behavior therapy are psychologists rather than



In treatment, the first step is analyzing the behavior that must

be changed. It is isolated from its context, identified, and

described as objectively and explicitly as possible -- even

quantitatively, if necessary. Then its relationships to the

environment are clarified. Patient and therapist observe carefully

when the behavior occurs, what situations or emotional states

provoke it, and whether it is being sustained by attention,

reassurance, or sympathy (a type of reinforcement sometimes called

secondary gain). To accomplish this, patients monitor, record, and

report what they are doing, often in diary form. Sometimes -- for

example, the patient wants to stop smoking -- this process is

fairly simple. But in other cases the complaint is less clearly

defined and the analysis much more complicated; for example,

"shyness" must be treated as a name for many different specific

types of behavior in many different situations.

When therapy begins, a written contract may be drawn up stating

what the patient is going to do and what results are expected. A

technique is then chosen; there are scores of different names for

these, and the terminology is inconsistent, but they all have a

great deal in common. Therapy begins with the behavior that is

easiest to change, and progress is measured by patient and

therapist. When one approach fails, another is tried.


Probably the most widely used and most successful behavioral

technique is a treatment for anxiety, phobias, and compulsive

rituals loosely based on classical conditioning. All of its many

related strategies are ways of eliminating or extinguishing an

undesired response. This is accomplished by creating conditions in

which the response will not occur, and then persuading patients to

expose themselves to situations that normally provoke the

response. The best known of these strategies, used especially for

phobias, is systematic desensitization, also known as reciprocal

inhibition. Its aim is to extinguish a learned response of fear by

associating the stimulus that causes the fear with a second

response that is incompatible with fear; the term "reciprocal

inhibition" refers to the fact that each response inhibits the


The therapist assumes that avoiding or escaping from the object of

the phobia has been reinforced and has therefore become a learned

response because it temporarily reduces the patient's fear -- a

fear that has never been extinguished because the phobia victim

has never been in contact with the frightening stimulus long

enough. Patients are desensitized by repeated confrontation of the

feared object while they are in a state of physical relaxation

(the incompatible response). They are trained to make themselves

comfortable by muscle relaxation exercises or other techniques,

including biofeedback -- the use of electrical monitoring to

provide information about a physiological state and bring it under

partial conscious control. The relaxed patient is then presented

with a series or hierarchy of situations graded in advance from

least to most frightening. Each successful exposure makes success

at the next stage easier. Some guidance may be necessary, but the

immediate presence of a therapist is not; desensitization has been

performed through taped or programmed instructions.

A form of systematic desensitization is also used to treat

impotence. The man uses sensations of sexual arousal as a response

competitively inhibiting the anxiety that he has come to associate

with sexual activity. He gradually moves toward greater intimacy,

keeping records of each step, concentrating on physical sensations

of arousal to reduce his anxiety about achieving an erection.

Although direct (in vivo) exposure is thought to have the quickest

and most lasting effects, some patients find it more convenient at

first to conjure up images of the feared situation and desensitize

themselves in imagination. A person who is afraid of climbing

ladders, for example, might be asked to close her eyes, relax, and

repeatedly imagine herself on the first rung of the ladder, then

on the second rung, and so on. Eventually she can attempt in real

life what she has already achieved in imagination.


It is not clear that reciprocal inhibition and carefully graded

hierarchies of anxiety-provoking situations are necessary to

eliminate phobias. Jumping in at the deep end may be just as

effective as wading in at the shallow end; striding to the edge of

the cliff may be as effective as looking out of windows at

gradually increasing heights. The type of exposure therapy that

starts with the most feared rather than the least feared stimulus

is called flooding; if conducted only in the imagination, it is

sometimes called implosion. The therapist controls the timing and

content of the scenes to be imagined or confronted, and instead of

trying to relax, the patient is told to experience the fear fully

until it subsides. Flooding is quicker than systematic

desensitization, but relapse may be more common, and the procedure

is simply too frightening for many patients.

The standard behavioral treatment for compulsive rituals

(checking, counting, washing, and so on) is known as exposure and

response prevention. This is a variant of flooding in which the

ritual is treated as a form of escape or avoidance. The patient is

placed in the situation that provokes ritualized behavior and

prevented from responding in the habitual way. A compulsive

washer, for example, is allowed to become dirty or even made dirty

and then prevented from washing. The handles might be unscrewed

from the water faucets in the house, or the water turned off for

most of the day. Exposure reduces hypersensitivity to dirt and the

associated anxiety, while response prevention eventually

eliminates (extinguishes) compulsive washing.


Obsessional thoughts are harder to treat than compulsive rituals,

because there is no overt behavior to change. Obsessional patients

may be haunted by the idea that they need rituals to prevent a

catastrophe; they must constantly check to see whether their doors

are locked or retrace their route in an automobile to be certain

they have not run over someone. A technique analogous to response

prevention can be used here; the patient is asked to think the

obsessional thought and then imagine not acting as it suggests.

Another way to break the chain of obsessional thinking is known as

thought-stopping. The patient is told to shout "Stop!" when an

obsessional thought appears, and eventually trains himself to make

the command silently. Although these behavioral methods may have

some effect on obsessions other types of treatment are likely to

be needed, including drugs and cognitive therapy.

Exposure and response prevention have also been used to treat

bulimia, the syndrome in which gorging is followed by vomiting or

purging. The assumption is that vomiting or purging relieves the

anxiety produced by eating and sustains the habit of binge eating

because it removes inhibitions against it. The patient is allowed

to eat and then forced to tolerate the resulting anxiety instead

of escaping it by the usual means. This is accomplished by locking

all nearby bathrooms and making her remain in the presence of

other people, where she is ashamed to vomit. The treatment is

usually accompanied by an effort to teach the patient how to eat

without gorging. Eventually she comes to feel less anxiety while

eating and is able to eat a normal amount of food without


The therapeutic use of operant conditioning requires altering the

contingencies of reinforcement: manipulating the environment to

increase or decrease the probability of behavior by reinforcing

desired actions and not reinforcing (or occasionally punishing)

the alternatives. For example, a person suffering from chronic

pain is told not to do things that reward pain, such as avoiding

work or other obligations. A child who throws tantrums is not

comforted but ignored or sent to a corner. A person who is

starving herself because of anorexia is given positive

reinforcement (visiting privileges, social activities) and

negative reinforcement (isolation, bed rest, tube feeding) to

encourage eating and gaining weight. Friends and family who may

have been rewarding undesirable behavior with sympathy or

attention are urged to reinforce desired behavior instead; for

example, parents might be told to ignore a child who talks while

they are on the telephone, but pay attention when he talks at the

dinner table.


A serious limitation of operant conditioning is that the effects

produced in a clinic or psychologist's office may be difficult to

reproduce in the patient's natural environment. This limitation is

overcome by fading and generalization -- gradually providing fewer

and fewer artificial rewards and punishments as the behavior

begins to maintain itself with natural reinforcers like

companionship or a sense of accomplishment.

Operant conditioning might be used, for example, to treat a

writing block. The patient is told to note how much she writes

each day and a reinforcement is chosen, usually some activity she

already enjoys, like reading a newspaper or playing chess. The

patient then arranges her life to reward herself, first after just

a little writing, later after every second or third writing

session, and so on. Eventually these rewards are gradually

replaced by self-esteem, the approval of readers and colleagues,

and acceptance of writings for publication.


In Part I we described the principles of behavioral learning

theory and some of its therapeutic applications. In this part we

discuss other types of behavior therapy and some findings on its


Depression has been treated on the assumption that it results from

the loss of adequate reinforcement. People who, for internal or

external reasons, are no longer receiving rewards for anything

they do, become withdrawn and therefore receive even less. Some

have also been given sympathy and attention for passivity and

withdrawal (secondary gain) and therefore have fallen into a

spiral of inactivity and social isolation. They must learn how to

do things that bring satisfaction. Behavior therapists try to help

them monitor their activities, find out which are rewarding, and

rearrange their lives to perform those actions more often. Operant

conditioning can be used to shape behavior so that patients

perform complex actions or acquire skills that were formerly

beyond their capacities. In the language of learning theory, they

expand their behavioral repertoires. The method is to approximate

the desired behavior gradually, rewarding each small step toward

mastery. Shaping is often necessary for severely impaired people

whose behavioral repertoire is very limited. For example, the

mouth movements of a mute autistic child are reinforced at first,

then sounds, and later only sounds closer and closer to normal

speech. The process of dressing is divided into stages and the

child is given candy or applause for completing each partial task.


One type of operant conditioning, the punishment technique known

as aversive therapy, has gripped the popular imagination so

powerfully, through the film "Clockwork Orange" and other sources,

that it is sometimes confused with behavior therapy in general.

Aversive therapy is the application of an unpleasant physical

stimulus to eliminate unwanted behavior. Obsessional patients give

themselves mild electric shocks or snap rubber bands on their

wrists to cut off unwanted thoughts; a capsule of ammonia is

broken under the nose of an autistic child or an electric shock is

delivered to his leg to prevent him from seriously injuring

himself by hitting his head against the wall; smokers trying to

quit wear a device that gives them a mild electric shock when they

remove a cigarette from a pack. Many members of the public regard

aversive therapy as a powerful form of manipulation with sinister

political implications -- a manifestation of behavior therapy's

evil alter ego, behavior modification.

In reality, aversive therapy usually requires the patient's

cooperation, and most authorities do not consider it particularly

effective, except in a few cases like the self-destructive

autistic child. As a result, it is rarely used. One of its

limitations is lack of a convenient way to fade the aversive

stimulus; there is no substitute in the natural environment.

Another problem is that eliminating undesirable behavior, whether

by aversive therapy or by any other form of punishment, is rarely

sufficient, because the vacuum may be filled with other

undesirable behavior. Punishment will usually be ineffective

unless it is used to encourage (negatively reinforce) some

alternative activity.

It is sometimes difficult (and may be unimportant) to decide

whether aversive therapy is a form of operant or classical

conditioning. The same ambiguity exists in other forms of behavior

therapy. For example, a child who fears water is allowed to keep

playing with a favorite toy as long as she moves closer and closer

to a filled bathtub. This could be described as operant

conditioning (a reward for approaching water) or classical

conditioning (desensitization to the feared object by associating

it with an incompatible response).


A specialized type of operant conditioning is the token economy,

which serves to formalize and quantify rewards. For example,

patients in a mental hospital are given tokens they can exchange

for food or cigarettes when they make their beds, groom

themselves, get up and go to sleep at the correct hours, and so

on; they are fined (tokens are taken away) for assault or

destruction of property. Token economies can also be used in