Recent Developments

in Classical Conditioning *
FRANCES K. McSWEENEY
CALVIN BIERLEY**

The present paper examines the implications of recent developments in classical

conditioning for consumer research. It discusses the finding that the conditioned
response need not resemble the unconditioned response, and that the conditioned
stimulus must predict but not necessarily precede the unconditioned stimulus for
conditioning to occur. The paper also considers the implications of several situations
in which classical conditioning may unexpectedly fail to occur, several of the char-
acteristics of classically conditioned behavior, and the role of awareness in con-
ditioning.

Downloaded from http://jcr.oxfordjournals.org/ by guest on June 26, 2016

N ord and Peter (1980) reviewed the implications for
marketing of two basic principles of learning: op-
The excitement (the UR) evoked by the voices (the US)
may become associated with the advertised product (the
erant and classical conditioning. Many other papers have CS) through classical conditioning when the two are paired
also focused on the implications of operant conditioning in an advertisement.
for consumer research (e.g., Markin and Narayana 1976; A cursory glance at current advertising supports Nord
Peter and Nord 1982; Rothschild and Gaidis 1981). In and Peter's point. The products of fast food chains are
operant conditioning, a response which is followed by a often marketed by associating their names with the sight
reinforcer increases in frequency (Skinner 1938). and sound of a sizzling hamburger, soft drinks are as-
Much less attention has focused on the principle of sociated with catchy jingles, breakfast cereals are asso-
classical conditioning (Pavlov 1927). Classical condition- ciated with famous sports personalities, and so on. Su-
ing is usually described by saying that Pavlov presented permarkets play music for their customers while they
an arbitrary stimulus (a metronome), now called the con- shop (Milliman 1982). All of this is done in the hope of
ditioned stimulus or CS, to a hungry dog. The CS was altering consumer behavior.
followed by another stimulus, now called the uncondi- There is also some laboratory evidence that classical
tioned stimulus or US (food). The US automatically conditioning can alter consumer behavior. Although
evoked a response, called the unconditioned response or classical conditioning has been known to occur in humans
UR (salivation). As a result of this pairing, the CS came virtually from the time of Pavlov (e.g., Hilgard 1931), it
to elicit a part of the unconditioned response even when is only recently that behaviors of interest to consumer
it was presented alone. The response emitted to the CS research have been studied. For example, Gom (1982)
was called the conditioned response or CR (salivation). tried to arrange a classical conditioning situation by
The present paper focuses on the implications of clas- showing students slides of either beige or blue pens (the
sical conditioning for consumer behavior and research CSs) while they listened to liked or disliked music (the
because two lines of evidence suggest that it may be rel- USs). When the students were given a choice of taking
evant. First, as Nord and Peter (1980) have pointed out, a beige or blue pen, more students indicated their pref-
many advertisements appear to be structured so that clas- erence (the CR) for the pen associated with the liked
sical conditioning will occur. To use one of their examples, music.
advertisements may use famous sportscasters whose voices The present paper describes the details of classical con-
have been paired with exciting sports events for years. ditioning as they are currently understood. We argue that
changes have been made recently in the idea that the
.. This is the first JCR article to be submitted, edited, and typeset via conditioned response must resemble the unconditioned
diskette.
response, that the CS must precede the US for condi-
** Frances K. McSweeney is Professor and Acting Chair of the De-
partment of Psychology, Washington State University, Pullman, WA tioning to occur, and that conditioning will occur when-
99164. Calvin Bierley is a Marketing Research Consultant in Seattle, ever the CS precedes the US. We also discuss some of
W A. The authors wish to thank Dr. Rom Markin, Dean of the College the characteristics of conditioned responses and the role
of Business and Economics, Washington State University, for his helpful of awareness in conditioning. The potential implications
comments on an earlier version of this manuscript.
of all of these changes in our understanding of classical
619
© JOURNAL OF CONSUMER RESEARCH. Vol. 11 • September 1984
620 THE JOURNAL OF CONSUMER RESEARCH

conditioning for consumer research are discussed tioning. In operant conditioning, it is the relation between
throughout the paper. the animal's behavior and the reinforcer or US that
We do not argue that classical conditioning necessarily changes the animal's future behavior. In classical con-
plays an important role in consumer behavior, or that ditioning, it is the relation between some arbitrary stim-
the specific implications discussed here are necessarily ulus (the CS) and the US or reinforcer which changes
correct. The role of classical conditioning in consumer behavior.
behavior and the accuracy of specific implications need Many experiments had already shown that key pecking
to be established by careful experiments. However, initial could be operantly conditioned (i.e., that pecking would
experiments are promising (e.g., Gom 1982), and more increase in frequency when it was followed by a rein-
conclusive experiments cannot be conducted without a forcer). However, the key pecks observed by Brown and
thorough understanding of classical conditioning pro- Jenkins could not be attributed to operant conditioning.
cedures. This paper tries to provide such an understanding Not only had Brown and Jenkins failed to arrange the
in the hope that future experiments will be able to establish relation between the pecking response and its reinforcer,
the role of classical conditioning in consumer research which is necessary for operant conditioning, but also,
in greater detail. later experiments showed that the key peck would still
occur during a Brown and Jenkins procedure even when
THE FORM OF the procedure was modified so that the key peck actually
THE CONDITIONED RESPONSE cancelled the presentation of food (e.g., Williams and

Downloaded from http://jcr.oxfordjournals.org/ by guest on June 26, 2016

Williams 1969). In that procedure, it was impossible to
The traditional view of classical conditioning assumes argue that the pecking occurred because pecks were oc-
that the conditioned response resembles the uncondi- casionally accidentally followed by food. Pecking in that
tioned response. If the US is food and it evokes salivation experiment was the one response that was never reinforced
(the UR), then the animal will also salivate (the CR) to by food.
a stimulus (the CS) that precedes the food. Some authors Further consideration revealed that the Brown and
have also argued that only responses controlled by the Jenkins experiment had arranged a classical conditioning
autonomic nervous system can be classically conditioned procedure. A light (the CS) was followed by food (the
(e.g., Skinner 1938). Basically, autonomic responses are US), and a pecking response which was evoked by the
those which are involuntary, such as salivation or blinking food (the UR) came to be emitted to the light (the CR).
of the eyes. According to this view, voluntary responses However, in order to accept that these key pecks were
controlled by the skeletal nervous system, such as walking classically conditioned, psychologists had to admit that
or talking, cannot be classically conditioned. More recent responses controlled by the skeletal nervous system, such
research has questioned both of these assumptions about as pecking, could be modified by classical conditioning
the form of the conditioned response. procedures.
Many psychologists since that time have presented ex-
Skeletal Responses Can Be planations for the results of the Brown and Jenkins ex-
Classically Conditioned periment that do not assume that the pecking was a clas-
sically conditioned response (e.g., Ettinger, Finch, and
Recent experiments have found that voluntary, skeletal McSweeney 1978; Steinhauer, Davol, and Lee 1976).
responses can be modified by a classical conditioning However, these explanations are typically complicated,
procedure. For example, Brown and Jenkins (1968) ex- and the current weight of opinion favors the idea that
posed hungry pigeons to a light (the CS) which appeared Brown and Jenkins demonstrated the classical condi-
periodically on a Plexiglas panel. When the light went tioning of a skeletal response.
out, the pigeons were given brief access to food (the US),
which evoked pecking (the UR). After several exposures The CR Need Not Resemble the UR
to this procedure, all subjects pecked the panel that was
lit (the CR). Recent experiments have also shown that the condi-
The Brown and Jenkins experiment surprised psy- tioned response need not resemble the unconditioned
chologists when it appeared. Up to that point, psychol- response. Evidence about the form of the CR comes from
ogists had believed that key pecking, like other skeletal three sources. First, Siegel and his associates conducted
responses, was controlled by operant conditioning. In op- a series of experiments in which various drugs served as
erant conditioning, a response is followed by a reinforcer, the US (Poulos, Hinson, and Siegel 1981; Siegel 1977).
and the response increases in frequency. A reinforcer is They found that the CR took a form opposite to that
technically defined as anything which increases the fre- evoked by the drugs themselves (the URs). For example,
quency of a response that it follows. Less technically, a if Siegel and his associates presented a CS to a rat and
reinforcer is a reward or something which the animal will followed it by a morphine injection (the US), the CS came
work to receive. Many reinforcers, such as food, can also to evoke a physiological reaction (the CR) that was op-
serve as USs in classical conditioning procedures. How- posite in form to that evoked by the morphine alone (the
ever, operant conditioning differs from classical condi- UR). If the morphine injections increased the animals'
DEVELOPMENTS IN CLASSICAL CONDITIONING 621

tolerance for pain, a tone which preceded morphine made and that the form of the CR need not resemble the form
them less tolerant of pain (but see also Eikelboom and of the UR.
Stewart 1982 for a different point of view). One implication of these changes for consumer research
Second, Holland (1980) conducted a series of experi- is that if skeletal responses can be classically conditioned,
ments with long-duration CSs in which he found several then classical conditioning can do more for a product
CRs occurring during different parts of the CS. Although than make people feel good about it or make them salivate
these responses took many forms that were difficult to in its presence. If Hearst and Jenkins are correct, then
describe, Holland argued that responses which occurred people will approach a product which is followed by a
in the early part of a long CS took a form that was de- US that is also a reinforcer. To use Gom's example, pre-
termined by the CS, and that those which occured later senting a pen to a person followed by pleasant music
in the CS took a form determined by the US. For example, may produce more approaches to the pen in the store,
if the CS was localizable (e.g., a light), the animal usually as well as a greater consumer preference for the pen.
physically approached and contacted the stimulus during Inducing approach to a product may have several benefits.
the early part of its presentation. If the US was food, then People may be more likely to buy a product that they
the animal usually bit, chewed, or salivated during the approach, and they may be more susceptible to other
later part of the CS. Holland's results have been interpreted forms of advertising or incentives that may also accom-
to mean that more than one conditioned response may pany the product.
occur and that they need not all resemble the UR. A second implication of the revised view of the CR is

Downloaded from http://jcr.oxfordjournals.org/ by guest on June 26, 2016

that a US that evokes the particular response to be con-
The CR May Include Sign- Tracking. In an influential ditioned need not be found before classical conditioning
paper, Hearst and Jenkins (1974) suggested that the con- can be used. If the CR always resembles the UR, then a
ditioned response always includes physical approaches to US that evokes a response must be found before that
and contact of the CS. In fact, they formulated a general response can be conditioned. This is not necessary if the
principle of behavior called sign-tracking, which states CR does not resemble the UR. Getting back to our ap-
that animals approach and contact the best predictors of proach example, a US that evokes approach need not be
reinforcers in their environment. Sign-tracking also states found before classical conditioning can be used to move
that animals withdraw from stimuli that predict the ab- people closer to displays in supermarkets. Any reinforcer
sence of reinforcers. According to Hearst and Jenkins, may be used as the US to condition this behavior.
the classical conditioning procedure arranges a situation Third, the revised view of the CR implies that any
in which sign-tracking will occur: the CS is the best pre- attempt to use classical conditioning to alter consumer
dictor of the presentation of the US. Because many USs behavior must be carefully pretested before it is used. It
are also reinforcers, sign-tracking should occur-i.e., the cannot be assumed that the response that will be con-
animals should physically approach and contact the CS ditioned will necessarily be similar to the response evoked
as part of the conditioned response. by the US. For example, it cannot be assumed that because
Hearst and Jenkins present many examples to illustrate people say that they like (UR) particular music (US), that
the role of sign-tracking in classical conditioning. One of they will also like (CR) a product (CS) that is followed
these examples is the Brown and Jenkins (1968) exper- by this music. The conditioned response to the product
iment just discussed. In that experiment the pigeons may be very different, and we currently have no rules to
pecked (i.e., approached and contacted the CR) the light predict this response. Only pretesting of the advertisement
(the CS) that signaled food (the US or reinforcer). Al- will reveal the response that will be conditioned.
though the sign-tracking principle has been criticized as Changes in our ideas about the form of the CR have
a general description of classically conditioned responding both favorable and unfavorable implications for the use
(e.g., Jenkins et al. 1978), most people concede that ap- of classical conditioning to alter consumer behavior. On
proach and contact responses may occur when classical the one hand, these changes have increased the number
conditioning procedures are used. and type of responses that can be modified by classical
conditioning. On the other hand, the changes have made
Implications. All of this evidence has helped to it less possible to predict in advance the change in behavior
broaden current ideas about the form of the CR. In many that a particular procedure will produce. Attempts to
cases, the CR still does resemble the UR, but in other alter consumer behavior should be pretested if classical
cases it does not. Although some authors have tried to conditioning is to be used.
describe when the CR will resemble the UR and when
it will not (e.g., Eikelboom and Stewart 1982), such de-
scriptions are not generally accepted yet. They are also PREDICfIVENESS
beyond the scope of this paper. However, the implications VS. TEMPORAL PRIORITY
of the change in the form of the CR for consumer research Temporal Priority
can be discussed without such an understanding. It is
enough to recognize that classical conditioning may con- According to the traditional view of classical condi-
trol skeletal responses as well as autonomic responses, tioning, a CS must temporally precede a US for classical
622 THE JOURNAL OF CONSUMER RESEARCH

conditioning to occur. This statement is almost correct showed that no conditioning occurred when the proba-
and it can be used to correctly identify some situations bility that a US would occur was equal in the presence
in which classical conditioning will not occur. For ex- and absence of the CS. This finding is also incompatible
ample, classical conditioning will not occur if the US with the temporal priority view, which predicts-incor-
precedes the CS. This is called backward conditioning. rectly-that conditioning should occur. The temporal
Although a few laboratory studies do report some back- priority view predicts conditioning if the US occasionally
ward conditioning (e.g., Shurtleff and Ayres 1981), back- occurs after a CS, regardless of what happens in the ab-
ward procedures (US before CS) usually do not work and sence of the CS.
they are generally less effective than equivalent forward Rescorla summarized his findings by saying that clas-
procedures (CS before US). Because backward condi- sical conditioning occurs when a CS predicts rather than
tioning is usually ineffective, playing a jingle (US) and precedes a US. He also argued that a CS predicts a US
then displaying a soft drink (CS) should produce little when the conditional probability of the US is higher in
change in preference (CR) for the soft drink. the presence of the CS than in its absence (Rescorla 1967).
The temporal priority view of classical conditioning Many additional experiments have supported his predic-
also predicts little conditioning when the CS and US are tiveness view over the traditional temporal priority view
presented simultaneously. Although this procedure has of conditioning (e.g., Thomas and Wagner 1964; Wagner
been used in some past studies of classical conditioning et al. 1964), but his definition of predictiveness has not
in marketing (e.g., Gom 1982), laboratory research in- gone unchallenged. Unfortunately, no formal statements

Downloaded from http://jcr.oxfordjournals.org/ by guest on June 26, 2016

dicates that it does not produce optimal conditioning describe exactly how the probabilities are calculated or
(e.g., Bitterman 1964; Smith, Coleman, and Gormezano how the presence or absence of the CS is defined and
1969). The failure of simultaneous conditioning implies measured.
that classical conditioning may not alter consumer be- In the absence of a completely acceptable formal def-
havior when it is used in static situations such as magazine inition, psychologists have agreed on a procedure that
or newspaper advertisements. In these situations, the arranges a predictive relation between a CS and a US. It
product cannot be reliably displayed before the US. The is now generally accepted that this procedure must be
failure of simultaneous conditioning also implies that followed to show that a response is classically conditioned
presenting the product (e.g., a soft drink) and then in- (e.g., Rescorla 1967). This procedure requires two groups
troducing the US (a catchy jingle) should be more effective of subjects. The CS-US paired group is exposed to the
in conditioning a preference for the soft drink than pre- CS followed by the US-for example, the name of the
senting the drink during the jingle. fast-food chain followed by a sizzling hamburger. The
random control group is exposed to the same number of
The Predictiveness View CSs and the same number of USs as the first group, but
these CSs and USs are presented randomly with respect
Although the implications of the temporal priority view to each other. The subjects see the name of the fast-food
for backward and simultaneous conditioning appear to chain and the sizzling hamburger, but the presentation
be correct, more recent studies have characterized classical of each stimulus is controlled by a timer which runs in-
conditioning differently. It has been argued that the CS dependently of the one controlling the other stimulus.
must predict, but not necessarily precede, the US for Classical conditioning is said to occur if a particular re-
conditioning to occur (Rescorla 1967). sponse (e.g., salivation to the name of the fast-food chain)
A three-part experiment by Rescorla (1968) may il- develops in the first group but not in the second.
lustrate the difference between the older temporal priority This procedure is said to represent the predictiveness
view and the newer predictive ness view of classical con- notion of classical conditioning because the two groups
ditioning. In the first part of his experiment, Rescorla differ only in the degree to which the CS predicts the US.
showed that the amount of conditioning displayed by rats The two groups have the same number of CS and US
increased with increases in the probability that a footshock presentations. Therefore, differences in behavior between
(the US) would follow a tone (the CS). The results ofthis the groups cannot be attributed to differences in familiarity
part of the experiment were compatible with the temporal with the CS or US, or to any interaction between them.
priority view, because it could also be said that the amount Differences between the groups can be attributed only to
of conditioning had increased with increases in the prob- differences in predictiveness. If classical conditioning oc-
ability that the US "followed" the CS. curs only when a CS predicts a US, then a conditioned
The second and third parts of the experiment were response should occur in the first group but not in the
incompatible with the temporal priority view, however. second.
In the second part, Rescorla showed that the amount of
conditioning also decreased with increases in the prob- Implications of the Predictiveness View
ability that a US would occur in the absence of a CS.
This finding is incompatible with the temporal priority The predictiveness view of classical conditioning has
view, which says nothing about the effect of USs that obvious implications for consumer research. To show
occur in the absence of the CS. In the third part, Rescorla that a behavior really has been established by classical
DEVELOPMENTS IN CLASSICAL CONDITIONING 623

conditioning, it is now necessary to use the two-group A final implication of the predictiveness view is that
experimental design. Otherwise, true classical condition- the number of exposures to the US in the absence of the
ing may be confused with changes in behavior resulting CS should be minimized. To condition preferences using
from CS or US familiarity, or from an interaction between music, a novel tune should be used. If a familiar song is
the two. used, then the US may frequently occur when it has not
The predictiveness view also has some practical im- been predicted by the CS, decreasing the effectiveness of
plications for altering consumer behavior. These impli- conditioning.
cations should be tested experimentally. For example, the To summarize, the CS must predict the US for classical
temporal priority view may imply that a preference for conditioning to be effective. The better the CS predicts
a soft drink could be conditioned by presenting the drink the US, the better the conditioning will be. The CS will
constantly with a jingle played occasionally. Sight of the not predict the US if the CS and US are presented si-
drink would occasionally precede the music. However, multaneously or if the CS is presented constantly with
experiments on classical conditioning have shown that the US introduced only occasionally. The predictiveness
presenting the CS constantly and introducing the US in- of the CS will also be reduced if the CS is encountered
termittently does not produce a conditioned response frequently in the absence of the US or if the US is en-
(Brown and Jenkins 1968). These experiments are con- countered frequently in the absence of the CS.
sistent with the predictiveness view of conditioning. In
this situation, the presence of the CS does not predict

Downloaded from http://jcr.oxfordjournals.org/ by guest on June 26, 2016

SITUATIONS IN WHICH CLASSICAL
whether the US will occur. The CS is always present and CONDITIONING FAILS TO OCCUR
the US occurs only occasionally. This implies that com-
mercials which are arranged so that the onset of the prod- The traditional view of classical conditioning argues
uct precedes the US will be more effective in altering that conditioning will occur whenever a CS is followed
consumer behavior than those which present the product by a US. More recently, situations have been found in
and introduce the US occasionally. which conditioning does not occur even when the CS is
The predictiveness view of conditioning also implies followed by, and predicts, the US. These failures of con-
that the number of extra exposures to the CS (the product) ditioning must be avoided if classical conditioning is to
outside of the advertisement should be minimized. As be effective. Currently, five situations are known to pre-
the number of times the CS occurs without the US in- vent conditioning even though the CS precedes and pre-
creases, the predictiveness of the CS-and therefore the dicts the US. These are summarized in Table 1.
conditioning to that CS-decreases (Rescorla 1968). This
has at least two implications for consumer research. Overshadowing
First, classical conditioning may not be as effective in
altering consumer behavior towards products which are First, classical conditioning will not occur if overshad-
frequently encountered as it is for products which are owing occurs (Pavlov 1927). In an overshadowing pro-
infrequently encountered. For example, classical condi- cedure, two CSs are presented at the same time and both
tioning might not work well in advertising Chevrolets precede the US. If the CSs differ in salience (e.g., a loud
because they will be frequently seen in everyday life. The noise and a dim light), then conditioning may occur to
effectiveness of the conditioning will decrease when the the more salient stimulus (the loud noise), but not to the
product is encountered without the US. Classical con- less salient stimulus (the dim light). For example, the
ditioning might work very well for a new line of cars sound of a particular sportscaster's voice (CS.) and a
because they will rarely be seen outside of the advertise- particular time on Saturday afternoon in the fall (CS2)
ment. Cutting across product lines, classical conditioning may both predict an enjoyable football game (US), but
might be more effective when it is used for something little classical conditioning may occur to the time because
such as a laundry detergent that is not encountered fre- it is "overshadowed" by the more noticeable voice. Over-
quently, than when it is used for a popular line of shirts shadowing implies that other salient stimuli should be
that would frequently occur without the US. removed from commercials which attempt to change
Second, the decrease in conditioning which occurs consumers' preferences for a product. For example, ifthe
when the CS is encountered without the US may also product is accompanied by an attractive person, the per-
imply that people who wish to use classical conditioning son may overshadow the product and leave preferences
to alter consumer behavior should not use several types for the product unchanged.
of advertising. To use Gorn's example of increasing the
preference for a pen by following it with music, people Blocking
will occasionally see the pen without hearing the music
and this will decrease the effectiveness of classical con- Classical conditioning will not occur if blocking occurs
ditioning. Its effectiveness should not be decreased still (Kamin 1969). In a blocking procedure, the subject is
further by presenting extra, unpredictive CSs, as it might given previous experience that one CS (CS.) forecasts a
if the pen were also advertised in newspapers or magazines US. Later, that CS and a new CS (CS2) are presented
where it would be seen without the music. together and are followed by the same US. No condi-
624 THE JOURNAL OF CONSUMER RESEARCH

TABLE 1

SITUATIONS IN WHICH CLASSICAL CONDITIONING MAY NOT OCCUR

r
Type Procedure Results Implications

Overshadowing CR develops to CS1 but not CS2 because Remove other salient stimuli which coincide
CS1 US
CS 2 CS1 is more salient. with the CS from the advertisement.
Blocking CS 1 -> US. CR develops to CS1 • but not CS2 • Do not use something familiar as a US.
then CS1 ~US because of the past experience that
CS2 CS1 forecasts the US.
US pre-exposure US only. then CR fails to develop to CS because US Do not use something familiar as a US.
effect CS -> US loses its effectiveness through prior
presentation alone.
Latent inhibition CS only. then CR fails to develop to the CS because Use classical conditioning to change
CS -> US the CS loses its effectiveness through preferences for new products rather than
prior presentation alone. old ones.
Garcia effect CS1 -> US1 CR develops to CS 1 when followed by Pretest your choices of CS and US to make
CS2 US 1 US1 but not when followed by US2 • CR sure conditioning will occur for these

Downloaded from http://jcr.oxfordjournals.org/ by guest on June 26, 2016

->
CS1 -> US2 develops to CS2 followed by US2 but choices.
CS2 -> US 2 not when followed by US1 •

tioning occurs to the new CS (CS2) even though it precedes blocking experiments. The US pre-exposure effect also
and predicts the US. It is as if prior experience with the suggests that a familiar stimulus (e.g., a familiar song)
old CS (CS 1) "blocked" conditioning to the new CS (CS 2). should not be used as a US. Familiarity with the US will
For example, suppose that over many years you have decrease the effectiveness of a classical conditioning pro-
learned that clouds forecast rain. Now you acquire a ba- cedure in altering consumer behavior.
rometer so that you can potentially forecast rain on the
basis of either clouds or barometric pressure. If blocking Latent Inhibition
occurs you will not acquire a classically conditioned re-
sponse (e.g., a disliking) to low barometric pressure even Classical conditioning will not occur if latent inhibition
though it is now a good and salient predictor of rain. In occurs (Lubow 1973; Lubow and Moore 1959). In a latent
a sense, your past experience that clouds are a valid pre- inhibition procedure, the CS is presented several times
dictor of rain prevents you from learning a conditioned without the US. When the CS is later followed by the
response to the barometric pressure. There is no need to US, little conditioning occurs. Latent inhibition implies
condition to the barometric pressure because it provides that it will be easier to classically condition behaviors to
little new information about the US (the chance of rain). new products (CSs) than to products with which people
Blocking suggests that familiar USs should not be used have had previous experience.
if a product is to be presented with stimuli that have
predicted that US in the past. For example, people may Garcia Effect
learn that listening to a particular radio station predicts
that a particular popular song may be played. If a product Classical conditioning does not occur in some cases
(CS 2) is advertised on that station (CS 1) using that par- because of improper choice of the CS that is paired with
ticular song as a US, previous experience may block con- the US. Garcia's experiments on poison-based avoidance
ditioning to the product. That is, people may show an conditioning provide the classic demonstrations of this
increased preference for that radio station, but not for failure of conditioning (Garcia and Koelling 1966). Ba-
the product, even though the product also predicted the sically, Garcia found that an aversion (CR) was easily
song. conditioned to a flavor (CS 1) when that flavor was followed
by sickness (US 1), but an aversion (CR) did not develop
US Pre-Exposure Effect to the same flavor (CS 1) when it was followed by shock
(US 2 ). Likewise, an aversion (CR) did develop to a light
Classical conditioning will not occur if subjects are and a noise (CS 2 ) that were followed by shock (US 2 ), but
given prior exposure to the US presented alone (Mis and it did not develop to the same light and noise (CS 2) when
Moore 1973; Rescorla 1973). Although familarity with a they were followed by sickness (US 1).
stimulus is frequently reported to produce greater liking No one has yet provided a list of rules that would
for that stimulus (Zajonc 1968), familiarity also decreases predict which choices of CSs and USs will produce con-
the effectiveness of that stimulus as a US. The US pre- ditioning when they are paired. One popular speculation
exposure effect is similar to blocking, but the US is pre- is that stimuli that "biologically" belong together, such
sented alone instead of predicted by a CS, as it is in as tastes and sickness, lead to the best conditioning. Stim-
DEVELOPMENTS IN CLASSICAL CONDITIONING 625

TABLE 2
CHARACTERISTICS OF CLASSICAL CONDITIONING

Type Procedure Results Implications

Acquisition CS- US Strength of CR increases Use a salient CS and a strong US. Space commercials
gradually with increases well in time. Do not expect a change in preference
in the number of pairing until several commercials have been given. Present
of the CS with the US. the commercial only rarely after full conditioning has
been achieved.
Extinction CS - US, then CS Strength of CR progressively Do not expect the change in preference to last forever
alone or CS and US decreases. if advertisements are completely discontinued.
presented
randomly
Higher order CS, - US, then CR develops to CS 2 because You may use people and objects associated with
conditioning CS 2 - CS, CS, has acquired the pleasant events as USs.
ability to act as a US.
Discrimination CS, - US and CS 2 CR develops to CS" not to CS 2 . Present rival products without the US, as well as your

Downloaded from http://jcr.oxfordjournals.org/ by guest on June 26, 2016

presented alone or own products with it.
randomly with respect
to the US
Generalization CS, - US CR occurs to other CSs to Restrict the benefits of your advertising to your product
the extent that they by making it as distinctive as possible. Take
resemble CS,. The advantage of other's advertising by making your
CR will extinguish to the other product Similar to theirs. Make your advertisements
CSs if they are presented as similar as possible to the situation in which the
repeatedly without a US. customer will encounter the product.
Inhibition CS, - US and CS 2 CS2 evokes a response opposite You may be able to decrease preference for a rival's
a. Discrimination presented alone or to that which occurs product if it is presented by itself in a commercial in
randomly with respect to CS,. which your product predicts the US.
to the US
b. Conditioned CS, - US and CS 2 evokes a response opposite You may be able to decrease preference for a rival's
inhibition CS, L no US to that evoked by CS, . product, if your product predicts a pleasant US when
presented alone but not when presented at the same
CS 2 )
time as the rival product.
c. Inhibition of long CS, - US Responses occur in the early Don't present your product for too long before the US.
delay part of the CS which are People may dislike your product during the first part
opposite in form to those of its presentation.
which occur later.

uli which are not "biologically" linked, such as tastes and CHARACTERISTICS OF CLASSICALLY
shocks, produce the worst conditioning. However, this CONDITIONED BEHAVIOR
rule is too vague to use in practice, and the implications
of the Garcia effect for altering consumer behavior are Although many of the characteristics of classically con-
clear without such a rule. The Garcia effect implies that ditioned behavior have been known for many years, we
it cannot be assumed that any stimulus can be used as discuss them here because they have obvious implications
a CS and any other stimulus can be used as a US. Instead, for consumer research. These characteristics are sum-
once a CS and US have been chosen, they must be pre- marized in Table 2.
tested to make sure that conditioning does occur when
they are paired. Acquisition
Summary One characteristic of conditioning is that classically
conditioned responses do not appear full blown the first
Taken together, these situations in which conditioning time the CS is followed by the US. Instead, the strength
fails to occur imply that careful attention must be paid of the conditioned response increases as a negatively ac-
to the procedure used if classical conditioning is to occur. celerated function of the number of pairings of the CS
For example, commercials cannot be put together solely with the US (Anrep 1920). This means that the strength
on the basis of intuition or artistic preference. Instead, of the CR increases greatly the first time the CS and US
they should be arranged very carefully to avoid situations are paired. The next pairing produces a slightly smaller
in which conditioning fails to occur. increase, the third pairing a still smaller increase, and so
626 THE JOURNAL OF CONSUMER RESEARCH

on. Eventually, the strength of the CR reaches an asymp- suggestion already offered to use salient CSs, strong USs,
tote. That is, it does not meaningfully increase with further and longer intertrial intervals to increase the speed of
pairings with the US. conditioning. Again, the only way to determine exactly
The speed of acquisition of CRs can range from one how quickly conditioning will occur is by pretesting the
to many trials, and it is influenced by many factors. For commercial using test groups.
example, the more salient the CS, the faster the condi-
tioning will be and the stronger the CR will be (Kamin Extinction
and Brimer 1963; Rescorla 1972). Conditioning will also
be better for stronger USs (Pavlov 1927; Wagner et al. A second characteristic of classically conditioned be-
1964) and for longer intertrial intervals (Terrace et al. havior is that it will disappear if the predictive relation
1975). The intertrial interval is the time between suc- between the CS and the US is broken (Pavlov 1927). This
cesssive CS-US pairings. Although no precise definition is called extinction, and the CS-US relation may be bro-
of CS salience can be given, CSs that are more intense ken in either of two ways. First, the US may be omitted
physically (e.g., brighter lights) are generally more salient entirely; second, the CS and US may both be presented
than less intense stimuli, and stimuli that have acquired randomly with respect to each other. For example, sup-
some psychological importance (e.g., your name) are more pose a preference for a soft drink has been conditioned
salient than those that have not. USs that are more intense by following that drink (CS) by a jingle (US). That con-
physically (e.g., more severe footshocks) are generally ditioned preference will disappear if the soft drink is pre-

Downloaded from http://jcr.oxfordjournals.org/ by guest on June 26, 2016

stronger than less intense stimuli. sented several times without the music, or if the soft drink
The characteristics of acquisition have several impli- and the music are presented randomly in time.
cations for consumer research. First, the finding that con- It should be clear from this example that extinction is
ditioning develops only with several CS-US pairings im- not the same as forgetting. Forgetting refers to the dis-
plies that presenting a commercial once will not be enough appearance of any behavior over time. Extinction refers
to substantially alter consumer preferences. Instead, an to the disappearance of a conditioned behavior when the
advertisement containing a soft drink and jingle must be CS no longer predicts the US. The evidence to date sug-
presented several times to insure that a change in pref- gests that very little forgetting occurs in conditioning (e.g.,
erence occurs. Hoffman, FleshIer, and Jensen 1963; Gleitman and
Second, commercials which use a salient CS and a Holmes 1967). A classically conditioned response will not
strong US will be most effective. To use Gom's pen and simply disappear with the passage of time. It will disappear
music example, the salience of the pen could be increased however, if the CS-US relation is broken.
by making sure it does not blend into the background Extinction implies that the predictive relation between
and by removing other distractions from the setting in the CS and the US must be maintained for as long as
which it is shown. The strength of the US could be in- the conditioned response is to occur. As soon as this
creased by pretesting several types of music to find the relation is broken, the conditioned response will begin
most preferred, and by making sure that it is clearly au- to disappear. Some presentations of the CS without the
dible when it is presented. US will occur naturally and cannot be controlled. For
Third, commercials which are well-spaced in time example, people may pass the name of a fast-food chain
should be more effective than those which are not. The without seeing a hamburger. If the commercial which
intertrial interval refers to the time between successive pairs the name ofthe fast-food chain with the hamburger
pairings of the CS and the US. In our fast-food chain is discontinued, then the conditioned response will dis-
example, this is the time between commercials which pair appear because it undergoes extinction. The conditioned
the name of the chain and a sizzling hamburger. As this response will be maintained only as long as the com-
time increases, the strength of conditioning achieved by mercial is presented occasionally.
a fixed number ofCS-US pairings increases up to a point. Extinction has implications for consumer research
No exact numbers can be assigned to the intertrial interval which are derived from the fact that the predictive relation
which will produce optimal conditioning in anyone case. between the CS and the US must be maintained in order
Currently, the only way to determine the optimal intertrial to prevent extinction. Like the predictiveness view of
interval is empirically-by presenting the commercials conditioning, extinction implies that conditioning will be
to test groups, using a different intertrial interval for each most effective for products that are rarely seen outside
group. Such pretesting might reveal, for example, that it of the commercials (e.g., a new type of car or a laundry
is better to buy a small number of commercials on several detergent). If the product is frequently seen, the CS will
shows than to buy many commericals on only one show. often occur without the US, and extinction of the con-
Fourth, the fact that conditioning reaches an asymptote ditioned response will occur rapidly once the commercial
after some number of trials implies that there will be is discontinued.
some optimal number of presentations beyond which Like the predictiveness view of conditioning, extinction
further presentations will not measurably change con- also implies that a novel US should be used in com-
sumer behavior. There are no fixed rules about how mercials that use conditioning. If the US occurs frequently
quickly responses reach an asymptote other than the vague without the CS, the predictive relation between the CS
DEVELOPMENTS IN CLASSICAL CONDITIONING 627

and US will be broken and extinction will occur rapidly other breakfast cereals which resemble the one actually
when the commercial is discontinued. Using novel USs used.
in commercials which employ conditioning will avoid In practical situations, generalization may be a two-
this problem and prolong the life of the conditioned re- edged sword. On the one hand, a product can be made
sponse even after the commercials are discontinued. as distinctive as possible in order to decrease generalization
and to restrict the benefits of advertising to that particular
Higher Order Conditioning product, but at the same time, that product will not benefit
from the advertising of other products. On the other hand,
A third characteristic of classical conditioning is higher a product can be made as similar as possible to other
order conditioning. In higher order conditioning, one CS products in order to increase generalization and the ben-
(CS 1) is followed by a US until a response is conditioned. efits from other products' advertising, but at the same
Then a new CS (CS2 ) is followed by the old CS (CSt). A time, the other products will also benefit from the ad-
response becomes conditioned to the new CS (CS 2) even vertising of the first product. Decisions regarding which
though that CS has never been followed by the US itself approach to take may depend on whether a company or
(Pavlov 1927). The old CS (CS 1) has acquired the ability its competitors are doing more advertising. If other com-
to act as a "higher order" US. Although higher order panies are doing more advertising, then a company may
conditioning was once thought to be a fragile phenomenon wish to increase generalization by making its product
which might not occur outside of the laboratory, recent more similar to the more heavily advertised products. If

Downloaded from http://jcr.oxfordjournals.org/ by guest on June 26, 2016

evidence indicates that higher order conditioning is a a particular company is doing more advertising than its
robust phenomenon (Rescorla and Holland 1982), sug- competitors, it may decrease generalization by making
gesting that it may have important implications for ev- its product as distinctive as possible.
eryday life. A second implication of generalization for consumer
From the standpoint of consumer research, higher order research is that effective advertisements will be as similar
conditioning implies that a standard US need not be em- as possible to the situation in which the person will actually
ployed to achieve classical conditioning. For example, encounter the advertised product. The CS which becomes
people or symbols which are typically associated with conditioned may include the entire setting of a com-
pleasant experiences can serve as higher order USs. As mercial as well as the product itself. If this setting (e.g.,
suggested by Nord and Peter (1980), a sports personality an exotic resort) is very different from the one in which
whose presence predicts an exciting game may serve as the person encounters the product (e.g., the supermarket),
a US in classical conditioning just as could the exciting then very little generalization (very little preference) may
game itself. carry over from the advertisement to the situation in
which the product is actually sold.
Discrimination
Discrimination is another characteristic of classically
Inhibition
conditioned behavior. Discrimination occurs when one A final characteristic of classical conditioning is in-
CS is followed by a US and another CS is not followed hibition. Inhibition refers to the fact that in some situ-
by the US. The conditioned response will occur to the ations a response may be conditioned which opposes the
CS that is followed by the US and not to the other CS response that will occur when the CS predicts the US.
(Pavlov 1927). From a consumer research point of view, The response which occurs to the CS when it predicts
presenting music after a product and explicitly presenting the US is called the excitatory conditioned response. The
a rival product not followed by music may successfully opposing response is called an inhibitory conditioned re-
develop a preference for the first product and prevent a sponse. In Gorn's example, the excitatory response that
preference from developing for the rival product. occurred when the pen predicted the music was an in-
crease in preference for the pen. The inhibitory response
Generalization would then be a decrease in preference for the pen. To
avoid accidentally conditioning a decrease in preference
Generalization refers to a conditioned response that for a product, it is worth considering the situations in
occurs to a novel CS when the novel CS resembles the which inhibition occurs.
CS to which the response has previously been conditioned
(Pavlov 1927; Siegel et al. 1968). The conditioned response Discrimination. Inhibition occurs in several situations
is said to "generalize" from the CS to which it was actually which might be characterized by saying that the CS pre-
conditioned to other new CSs which resemble the old dicts a decrease in the probability of a US (Mackintosh
one. The extent to which the CR occurs to the new CS 1974). One situation in which this occurs is the discrim-
depends directly on how closely the new CS resembles ination situation already mentioned. In a discrimination
the old one. For example, if a preference for a breakfast situation, one CS predicts the US, another CS does not.
cereal is developed by pairing it with famous athletes, The CS which does not predict the US becomes inhibitory
smaller increases in preferences will also be observed for (Rescorla and LoLordo 1965). That is, with experience,
628 THE JOURNAL OF CONSUMER RESEARCH

it will evoke a response which is opposite to that evoked at the present time. For example, at first glance, it might
by the CS that does predict the US. A commercial in seem that the question of whether awareness is needed
which a product is followed by music and a rival's product for conditioning could be easily answered by conducting
is not should not only increase preference for the first a conditioning experiment and then asking subjects to
product, it should also decrease preference for the rival's decribe what happened. The role of awareness would be
product. demonstrated if the subjects who described the CS-US
relation showed conditioning and those who did not de-
Conditioned Inhibition. The conditioned inhibition scribe the relation showed no conditioning. However,
procedure is another situation in which inhibition occurs such an experiment would actually establish very little.
(Pavlov 1927). In a conditioned inhibition procedure, First, the results could be used to support almost any
one stimulus is presented alone and then followed by a interpretation. Suppose, for example, that conditioning
US. The same stimulus is not followed by the US when was found in subjects who could not report the CS-US
it is accompanied by another stimulus. The second stim- relation. Then the proponents of the view that awareness
ulus becomes inhibitory. To use our example, a com- is necessary for conditioning could argue that some
mercial in which a pen is followed by music when it is "aware" subjects were actually classified as "unaware."
presented alone and is not followed by music when it is According to this argument, more careful questioning
presented with a rival pen should condition inhibition to would have revealed this rudimentary awareness.
the rival pen. That is, not only will preference increase This example should illustrate the fact that the role of

Downloaded from http://jcr.oxfordjournals.org/ by guest on June 26, 2016

for the first product, but it should also decrease for the awareness in conditioning cannot be clearly established
rival product. until some generally agreed upon measure of awareness
Inhibition of Delay. A final situation in which inhi- is developed. Until that time, authors can always rein-
bition occurs is called inhibition of delay (Pavlov 1927). terpret the results of experiments by arguing that unaware
In this situation, the initial part of a long CS becomes subjects were actually aware, and vice versa. Even if such
inhibitory, the latter part becomes excitatory. No absolute an experiment conclusively showed that conditioning oc-
time period can be assigned to the duration of CS that curred only in aware subjects, the results would dem-
is needed before inhibition will occur. This can only be onstrate only a correlation between awareness and con-
determined in practice. But in general, if a picture of a ditioning. They would not show that the awareness pro-
pen is presented for too long before the music comes on, duced conditioning. It would be equally plausible that
people may develop a dislike of the pen when they initially conditioning produced awareness or that both were pro-
see it. This disliking would eventually turn to a preference duced by a third variable.
for the pen if the person looked at it long enough for the Other authors have taken a more sophisticated ap-
excitatory CR to develop. The problem is that few people proach to the role of awareness in conditioning. For in-
would attend to the CS long enough for the excitatory stance, Brewer (1974) argued that "higher mental pro-
response (the preference) to appear. cesses," such as awareness, must be responsible for con-
Again, these characteristics of conditioning may imply ditioning in humans because the occurrence of
that commercials which use conditioning must be care- conditioning can be manipulated by manipulating these
fully arranged to make sure that conditioning will occur. higher mental processes. To give an example, Brewer
Only careful attention to detail will prevent serious errors. argued that giving subjects the instructions that a CS will
be followed by a US, or even the instructions to make a
CONDITIONING AND AWARENESS particular response, produced a conditioned response-
just as exposing the person to the CS followed by the US
We have been concerned with classical conditioning did. This evidence is more sophisticated because it is
as a procedure which has a specific effect on behavior. If experimental rather than correlational. Brewer cited ex-
a CS predicts a US, then a conditioned response will periments which manipulated a "higher mental process"
develop. We have not taken a position regarding why (e.g., gave instructions) and observed a change in con-
classical conditioning occurs. For example, we have not ditioning. Therefore, he could legitimately conclude that,
addressed the question of whether conditioning occurs if the experiments were properly designed, the "higher
because of some mechanical stamping in of a connection mental process" caused the conditioning.
between a CS and US or whether it requires some higher Brewer reviewed the literature extensively and clearly
cognitive processing on the part of the subject. This is demonstrated several manipulations which alter condi-
an interesting question and a theory that answers such a tioning. However, to summarize these manipUlations by
question would vastly increase our knowledge of con- saying that "higher mental processes" are necessary for-
ditioning. However, this question has not been discussed or cause-conditioning seems misleading and problematic
here because it is not answerable as it is currently phrased to us.
and because the implications of conditioning for con- First, arguing that "higher mental processes" cause
sumer research are clear without such an answer. conditioning when human subjects are used puts Brewer
The question of whether awareness is necessary for in the uncomfortable position of arguing that classical
conditioning is too vague to be answered experimentally conditioning occurs for different reasons in different spe-
DEVELOPMENTS IN CLASSICAL CONDITIONING 629

cies. Classical conditioning has been observed in many mental processes" are necessary for conditioning when
nonhuman species (see Thorpe 1963 for a review). It is people serve as subjects. But we are arguing that the role
hard to argue that the CR is produced by the contem- of higher mental processes in conditioning cannot be
plation of the CS-US relation when rats or rabbits serve conclusively established by current techniques. Focusing
as subjects. Therefore, if one argues that conditioning on this question now may divert attention from questions
occurs in people because of "higher mental processes," which can be answered, to the detriment of research on
one must also argue that classical conditioning occurs for conditioning.
qualitatively different reasons at different points on the
evolutionary scale. Such a claim should be avoided unless CONCLUDING COMMENTS
the case for the role of "higher mental processes" in hu-
mans is overwhelming, and it is not. This paper has presented some of our current knowl-
The weakness of the case for "higher mental processes" edge about classical conditioning and its potential im-
in human conditioning becomes more apparent when plications for consumer research. Some words of caution
one realizes that analogues of many of the experiments are needed before concluding.
cited by Brewer have been conducted, using nonhuman First, we are not arguing that classical conditioning
animals, with similar results. For example, Brewer showed necessarily plays an important role in consumer behavior.
that instructing human subjects that a CS will be followed Most of our information about classical conditioning
by a US produced a CR. But a similar phenomenon can comes from carefully controlled laboratory studies. These

Downloaded from http://jcr.oxfordjournals.org/ by guest on June 26, 2016

be shown in rats. For example, Revusky, Coombes, and principles may not hold in the more complicated "real
Pohl (1982) showed that exposing a rat to a flavored world" settings in which consumer behavior takes place.
liquid in the presence of a poisoned rat established a Studies should be conducted to determine exactly what
classically conditioned aversion to that liquid in the first role classical conditioning does play in these settings. Ini-
rat. It was as if the sight of the sick rat "instructed" the tial results are promising-they have indicated that clas-
first rat that ingesting the liquid (the CS) would be followed sical conditioning may play some role (e.g., Gorn 1982).
by sickness (the US). But these studies are few and even they have indicated
There is certainly much to explain here. Brewer and that classical conditioning may play a more important
others have uncovered some important manipulations role in some situations than in others. It remains for
that produce conditioning and that need to be explained. future studies to systematically establish the role of con-
But arguing that "higher mental processes" are necessary ditioning in consumer behavior.
for conditioning is only one possible summary of these Second, we are not arguing that the implications of
manipulations, and it is one which leads to conclusions classical conditioning for consumer research are always
that are both intuitively unappealing and untestable. It completely clear. Again, the principles of conditioning
is not intuitively appealing to argue that the rat contem- were developed in the laboratory, and it may be difficult
plates the meaning of the poisoned conspecific and draws to translate laboratory situations into their "real world"
implications for its own behavior. It is also untestable, equivalents. For example, you may have noticed that the
because we currently have no independent way of mea- blocking and overshadowing principles and the higher
suring the rat's "higher mental processes" to make sure order conditioning principle seem to make different pre-
that they are necessary for conditioning. dictions about whether famous people should be used in
We believe that a better way to approach Brewer's advertising. From the higher order conditioning point of
findings is to use them to formulate a hypothesis which view, the presence of a celebrity after a product may act
is currently testable, and then to test the hypothesis. For as a higher order US that would aid in establishing those
example, one might conclude that Brewer's data and the preferences. From the point of view of overshadowing or
data from studies using rats suggest that conditioning can blocking, the presence of celebrities along with the product
occur when a CS predicts a US even when the subject might interfere with the formation of preferences for the
does not emit the CR. This is a testable hypothesis because product, either by being more salient than the product
an animal's ability to emit a CR-unlike an animal's or by blocking conditioning of preferences to the product.
"higher mental processes"-can be manipulated today Only very careful studies can reveal when the presence
using specific drugs. As a result, this hypothesis could be of celebrities improves and when it interferes with the
tested in a variety of circumstances and either accepted formations of preferences in advertising; and only careful
or rejected. The exact nature of the CS-US relation that studies can determine whether the time at which the ce-
is necessary to produce the conditioning without a CR lebrity is presented makes a difference as implied by a
(ie., the nature of the instructions) could then be explored rigorous translation of classical conditioning principles.
in detail. It seems to us that this approach is more likely Many careful studies will be required before accurate
to yield information about conditioning than arguments translations can be made from laboratory classical con-
about whether "higher mental processes" produce con- ditioning situations to the "real world" of consumer be-
ditioning. havior.
We should repeat that we are not arguing that Brewer
is necessarily incorrect. He may be correct that "higher [Received January 1984. Revised March 1984.]
630 THE JOURNAL OF CONSUMER RESEARCH

REFERENCES Mackintosh, N. J. (1974), The Psychology ofAnimal Learning,

New York: Academic Press.
Anrep, G. V. (1920), "Pitch Discrimination in the Dog," Journal Markin, Rom J. and Chern L. Narayana (1976), "Behavior
of Physiology, 53, 367-385. Control: Are Consumers Beyond Freedom and Dignity?"
Bitterman, M. E. (1964), "Classical Conditioning in the Goldfish in Advances in Consumer Research, Volume 3, ed. Beverlee
as a Function of the CS-US Interval," Journal of Com- B. Anderson, Ann Arbor, MI: Association for Consumer
parative and Physiological Psychology, 58 (3), 359-366. Research, 222-228.
Brewer, William F. (1974), "There is No Convincing Evidence Milliman, Ronald E. (1982), "Using Background Music to Affect
for Operant and Classical Conditioning in Humans," in the Behavior of Supermarket Shoppers," Journal of Mar-
Cognition and Symbolic Processes, eds. W. R. Weiner and keting, 46 (Summer), 86-91.
D. S. Palemo, Hillsdale, NJ: Lawrence Erlbaum, 1-35. Mis, Frederick W. and John W. Moore (1973), "Effect ofPreac-
Brown, Paul L. and Herbert M. Jenkins (1968), "Auto-Shaping quisition UCS Exposure on Classical Conditioning of the
of the Pigeon's Key-Peck," Journal of the Experimental Rabbit's Nictitating Membrane Response," Learning and
Analysis of Behavior, II (I), 1-8. Motivation, 4 (1), 108-114.
Eikelboom, Roelof and Jane Stewart (1982), "Conditioning of Nord, Walter R. and J. Paul Peter (1980), "A Behavior Mod-
Drug-Induced Physiological Responses," Psychological ification Perspective on Marketing," Journal ofMarketing,
Review, 89 (5), 507-528. 44 (Spring), 36-47.
Ettinger, Richard, Michael D. Finch, and Frances K. McSweeney Pavlov, Ivan P. (1927), Conditioned Ref/exes, Oxford, England:
(1978), "The Role of Generalization in the Acquisition of Oxford University Press.

Downloaded from http://jcr.oxfordjournals.org/ by guest on June 26, 2016

Autoshaped Keypecking in Pigeons," Bulletin of the Psy- Peter, J. Paul and Walter R. Nord (1982), "A Clarification and
chonomic Society, 12 (3), 235-238. Extension of Operant Conditioning Principles in Market-
Garcia, John and Robert A. Koelling (1966), "Relation of Cue ing," Journal of Marketing, 46 (Summer), 102-107.
to Consequence in Avoidance Learning," Psychonomic Poulos, Constantine X., Riley E. Hinson, and Shepard Siegel
Science, 4 (3), 123-124. (1981), "The Role of Pavlovian Processes in Drug Tolerance
Gleitman, Henry and Peter A. Holmes (1967), "Retention of and Dependence: Implications for Treatment," Addictive
Incompletely Learned CER in Rats," Psychonomic Science, Behaviors, 6 (3), 205-211.
7 (I), 19-20. Rescorla, Robert A. (1967), "Pavlovian Conditioning and Its
Gorn, Gerald J. (1982), "The Effects of Music in Advertising Proper Control Procedures," Psychological Review, 74 (I),
on Choice Behavior: A Classical Conditioning Approach," 71-80.
Journal of Marketing, 46 (Winter), 94-101. - - - (1968), "Probability of Shock in the Presence and Ab-
Hearst, Eliot and Herbert M. Jenkins (1974), Sign-Tracking: sence ofCS in Fear Conditioning," Journal o/Comparative
The Stimulus-Rein/orcer Relation and Directed Action, and Physiological Psychology, 66 (I), 1-5.
Austin, TX: The Psychonomic Society. - - - (1972), "Informational Variables in Pavlovian Condi-
Hilgard, Ernest R. (1931), "Conditioned Eyelid Reactions to a tioning," The Psychology 0/ Learning and Motivation, 6,
Light Stimulus Based on the Wink Reflex to Sound," Psy- 1-46.
chological Monographs, 41 (184). - - - (1973), "Effect of US Habituation Following Condi-
Hoffman, Howard S., Morton Fleshier, and Philip Jensen (1963), tioning," Journal of Comparative and Physiological Psy-
"Stimulus Aspects of Aversive Controls: The Retention of chology, 82 (I), 137-143.
Conditioned Suppression," Journal 0/ the Experimental - - - and Peter C. Holland (1982), "Behavioral Studies of
Analysis of Behavior, 6 (4), 575-583. Associative Learning in Animals," Annual Review 0/ Psy-
Holland, Peter C. (1980), "CS-US Interval as a Determinant chology, 33, 265-308.
of the Form of Pavlovian Appetitive Conditioned Re- - - - and Vincent M. LoLordo (1965), "Inhibition of Avoid-
sponses," Journal ofExperimental Psychology: Animal Be- ance Behavior," Journal ofComparative and Physiological
havior Processes, 6 (2), 155-174. Psychology, 59 (3), 406-412.
Jenkins, Herbert M., F. J. Barrera, C. Ireland, and B. Woodside Revusky, Sam, Shannon Coombes, and Richard W. Pohl (1982),
(1978), "Signal-Centered Action Patterns in Dogs in Ap- "US Preexposure: Effects on Flavor Aversions Produced
petitive Classical Conditioning," Learning and Motivation, by Pairing a Poisoned Partner with Ingestion," Animal
9 (3), 272-296. Learning and Behavior, 10 (I), 83-90.
Kamin, Leon J. (1969), "Predictability, Surprise, Attention and Rothschild, Michael L. and William C. Gaidis (1981), "Behav-
Conditioning," in Punishment and Aversive Behavior, eds. ioral Learning Theory: Its Relevance to Marketing and
Byron A. Campbell and Russell M. Church, New York: Promotion," Journal of Marketing, 45 (Spring), 70-78.
Appleton-Century-Crofts, 279-296. Shurtleff, David and John J. B. Ayres (1981), "One-Trial Back-
- - - and Charles J. Brimer (1963), "The Effects of Intensity ward Excitatory Fear Conditioning in Rats: Acquisition,
of Conditioned and Unconditioned Stimuli on a Condi- Retention, Extinction, and Spontaneous Recovery," Ani-
tioned Emotional Response," Canadian Journal of Psy- mal Learning and Behavior, 9 (I), 65-74.
chology, 17 (2), 194-200. Siegel, Shepard (1977), "Morphine Tolerance Acquisition as an
Lubow, Robert E. (1973), "Latent Inhibition," Psychological Associative Process," Journal o/Experimental Psychology:
Bulletin, 79 (6), 398-407. Animal Behavior Processes, 3 (I), 1-13.
- - - and A. U. Moore (1959), "Latent Inhibition: The Effect - - - , Eliot Hearst, Nancy George, and Edgar O'Neal (1968),
of Nonreinforced Pre-Exposure to the Conditioned Stim- "Generalization Gradients Obtained from Individual Sub-
ulus," Journal of Comparative and Physiological Psychol- jects Following Classical Conditioning," Journal of Ex-
ogy, 52 (4), 415-419. perimental Psychology, 78 (I), 171-174.
DEVELOPMENTS IN CLASSICAL CONDITIONING 631

Skinner, B. F. (1938), The Behavior of Organisms: An Exper- the Rabbit," Journal of Comparative and Physiological
imental Analysis, Englewood Cliffs, NJ: Prentice-Hall. Psychology, 58 (I), 157-158.
Smith, Marius c., Steven R. Coleman, and I. Gormezano (1969), Thorpe, W. H. (1963), Learning and Instinct in Animals. Cam-
"Classical Conditioning of the Rabbit's Nictitating Mem- bridge, Massachusetts: Harvard University Press.
brane Response at Backward, Simultaneous, and Foreward Wagner, Allan R., Shepard Siegel, Earl Thomas, and Gaylord
CS-US Intervals," Journal of Comparative and Physio- D. Ellison (1964), "Reinforcement History and the Ex-
logical Psychology, 69 (2), 226-231. tinction of a Conditioned Salivary Response," Journal of
Steinhauer, G. D., G. H. Davol, and A. Lee (1976), "Acquisition Comparative and Physiological Psychology, 58 (3), 354-
ofthe Autoshaped Key Peck as a Function of Amount of 358.
Preliminary Magazine Training," Journal of the Experi- Williams, David R. and Harriet Williams (1969), "Auto-
mental Analysis of Behavior, 25 (3), 355-359. Maintenance in the Pigeon: Sustained Pecking Despite
Terrace, H. S., J. Gibbon, L. Farrell, and M. D. Baldock (1975), Contingent Non-Reinforcement," Journal of the Experi-
"Temporal Factors Influencing the Acquisition and mental Analysis of Behavior, 12 (4),511-520.
Maintenance of an Autoshaped Keypeck," Animal Learn- Zajonc, Robert B. (1968), "Attitudinal Effects of Mere Expo-
ing and Behavior, 3 (I), 53-62. sure," Journal ofPersonality and Social Psychology Mono-
Thomas, Earl and Allan R. Wagner (1964), "Partial Reinforce- graph Supplement, 9 (Part 2, June), 1-27.
ment of the Classically Conditioned Eyelid Response in

Downloaded from http://jcr.oxfordjournals.org/ by guest on June 26, 2016