KEY THEME

Contemporary learning researchers acknowledge the importance of both cognitive factors and evolutionary influences in classical conditioning.

KEY QUESTIONS

The traditional behavioral perspective holds that classical conditioning results from a simple association of the conditioned stimulus and the unconditioned stimulus. Mental or cognitive processes such as thinking, anticipating, or deciding were not needed to explain the conditioning process. However, according to the cognitive perspective (see Chapter 1), mental processes as well as external events are important components in the learning of new behaviors. In the next section, we’ll look at research on cognitive processes in classical conditioning.

In a series of experiments, learning theorist Robert Rescorla demonstrated that classical conditioning involves more than learning the simple association of two stimuli. In a classic study, one group of rats heard a tone (the conditioned stimulus) that was paired 20 times with a brief electric shock (the unconditioned stimulus). A second group of rats experienced the same number of tone–shock pairings, but this group also experienced an additional 20 shocks with no tone (Rescorla, 1968). Then Rescorla tested for the conditioned fear response by presenting the tone alone to each group of rats. Because each group had received 20 tone-shock pairings, both groups should have displayed the same levels of conditioned fear. However, the rats in the first group displayed a much stronger fear response to the tone than did the rats in the second group. Why?

According to Rescorla (1988), classical conditioning depends on the information the conditioned stimulus provides about the unconditioned stimulus. For learning to occur, the conditioned stimulus must be a reliable signal that predicts the presentations of the unconditioned stimulus. For the first group of rats, that was certainly the situation. Every time the tone sounded, a shock followed. But for the second group, the tone was an unreliable signal. Sometimes the tone preceded the shock, and sometimes the shock occurred without warning.

Rescorla concluded that the rats in both groups were actively processing information about the reliability of the signals they encountered. Rather than merely associating two closely paired stimuli, as Pavlov suggested, the animals assess the predictive value of stimuli. Applying this interpretation to classical conditioning, we can conclude that Pavlov’s dogs learned that the bell was a signal that reliably predicted that food would follow.

According to this view, animals use cognitive processes to draw inferences about the signals they encounter in their environments. To Rescorla (1988), classical conditioning “is not a stupid process by which the organism willy-nilly forms associations between any two stimuli that happen to co-occur.” Rather, his research suggests that “the animal behaves like a scientist, detecting causal relations among events and using a range of information about those events to make the relevant inferences” (Rescorla, 1980). Put simply, classical conditioning seems to involve learning the relationships between events (Rescorla, 1988, 2003).

According to Darwin’s theory of evolution by natural selection, both the physical characteristics and the natural behavior patterns of any species have been shaped by evolution to maximize adaptation to the environment. Thus, just as physical characteristics vary from one species to another, so do natural behavior patterns. Some psychologists wondered whether an animal’s natural behavior patterns, as shaped by evolution, would also affect how it learned new behaviors, especially behaviors important to its survival.

According to traditional behaviorists, the general principles of learning applied to virtually all animal species and all learning situations. Thus, they argued that the general learning principles of classical conditioning would be the same regardless of the species or the response being conditioned. However, in the 1960s, some researchers began to report “exceptions” to the well-established principles of classical conditioning (Lockard, 1971; Seligman, 1970). As you’ll see in this section, one important exception involved a phenomenon known as a taste aversion. The study of taste aversions contributed to a new awareness of the importance of the organism’s natural behavior patterns in classical conditioning.

TASTE AVERSIONS AND CLASSICAL CONDITIONINGSPAGHETTI? NO, THANK YOU!

A few years ago, our friend Tom ate a large plateful of spaghetti at a highly rated Italian restaurant. But in the middle of the night, Tom came down with a nasty stomach virus. As a result, Tom developed a taste aversion—he avoided eating spaghetti and felt queasy whenever he smelled tomato sauce.

Such learned taste aversions are relatively common. Our students have told us about episodes of motion sickness, morning sickness, or illness that resulted in taste aversions to foods as varied as cotton candy, strawberries, and chicken soup. In some cases, a taste aversion can persist for years.

At first glance, it seems as if taste aversions can be explained by classical conditioning. In Tom’s case, a neutral stimulus (spaghetti) was paired with an unconditioned stimulus (a stomach virus), which produced an unconditioned response (nausea). Now a conditioned stimulus, the spaghetti sauce by itself, elicited the conditioned response of nausea.

But notice that this explanation seems to violate two basic principles of classical conditioning. First, the conditioning did not require repeated pairings. Conditioning occurred in a single pairing of the conditioned stimulus and the unconditioned stimulus. Second, the time span between these two stimuli was several hours, not a matter of seconds. Is this possible? The anecdotal reports of people who develop specific taste aversions seem to suggest it is. But such reports lack the objectivity and systematic control that a scientific explanation of behavior requires.

Enter psychologist John Garcia, who demonstrated that taste aversions could be produced in laboratory rats under controlled conditions (Garcia & others, 1966). Garcia’s procedure was straightforward. Rats first drank saccharin-flavored water (the neutral stimulus). Hours later, the rats were injected with a drug (the unconditioned stimulus) that produced gastrointestinal distress (the unconditioned response). After the rats recovered from their illness, they refused to drink the flavored water again. The rats had developed a taste aversion to the saccharin-flavored water, which had become a conditioned stimulus.

At first, many psychologists were skeptical of Garcia’s findings because they seemed to violate the basic principles of classical conditioning (Davis & Riley, 2010). Several leading psychological journals refused to publish Garcia’s research, saying the results were unconvincing or downright impossible (Garcia, 1981, 2003). But Garcia’s results have been replicated many times. In fact, later research showed that taste aversions could develop even when a full 24 hours separated the presentation of the flavored water and the drug that produced illness (Davis & Riley, 2010).

Conditioned taste aversions also challenged the notion that virtually any stimulus can become a conditioned stimulus. As Pavlov (1928) wrote, “Any natural phenomenon chosen at will may be converted into a conditioned stimulus … any visual stimulus, any desired sound, any odor, and the stimulation of any part of the skin.” After all, Pavlov had demonstrated that dogs could be classically conditioned to salivate to a ringing bell, a ticking metronome, and even the sight of geometric figures.

But if this were the case, then why didn’t Tom develop an aversion to other stimuli he encountered between the time he ate the spaghetti and when he got sick? Why was it that only the spaghetti sauce became a conditioned stimulus that triggered nausea, not the restaurant, the tablecloth, or his dining companions?

Contrary to what Pavlov suggested, Garcia and his colleagues demonstrated that the particular conditioned stimulus that is used does make a difference in classical conditioning (Garcia & Koelling, 1966). In another series of experiments, Garcia found that rats did not learn to associate a taste with a painful event, such as a shock. Nor did they learn to associate a flashing light and noise with illness. Instead, rats were much more likely to associate a painful stimulus, such as a shock, with external stimuli, such as flashing lights and noise. And rats were much more likely to associate a taste stimulus with internal stimuli—the physical discomfort of illness. Garcia and Koelling (1966) humorously suggested that a sick rat, like a sick person, speculates, “It must have been something I ate.”

Why is it that certain stimuli are more easy to associate than others? One factor that helps explain Garcia’s results is biological preparedness—the idea that an organism is innately predisposed to form associations between certain stimuli and responses (Freeman & Riley, 2009). If the particular stimulus and response combination is not one that an animal is biologically prepared to associate, then the association may not occur or may occur only with great difficulty (see the In Focus box, “Evolution, Biological Preparedness, and Conditioned Fears: What Gives You the Creeps?”).

When this concept is applied to taste aversions, rats (and people) seem to be biologically prepared to associate an illness with a taste rather than with a location, a person, or an object. Hence, Tom developed an aversion to the spaghetti sauce and not to the fork he had used to eat it. Apparently, both humans and rats are biologically prepared to learn taste aversions relatively easily. Thus, taste aversions can be classically conditioned more readily than can more arbitrary associations, such as that between a ringing bell and a plate of food.

Associations that are easily learned may reflect the evolutionary history and survival mechanisms of the particular animal species (Freeman & Riley, 2009). For example, rats in the wild eat a wide variety of foods. If a rat eats a new food and gets sick several hours later, it’s likely to survive longer if it learns from this experience to avoid that food in the future (Kalat, 1985; Seligman, 1970).

That different species form some associations more easily than others also probably reflects the unique sensory capabilities and feeding habits that have evolved as a matter of environmental adaptation. Bobwhite quail, for instance, rely primarily on vision for identifying potential meals. In contrast, rats have relatively poor eyesight and rely primarily on taste and odor cues to identify food. Given these species differences, it shouldn’t surprise you that quail, but not rats, can easily be conditioned to develop an aversion to blue-colored water—a visual stimulus. On the other hand, rats learn more readily than quail to associate illness with sour water—a taste stimulus (Wilcoxon & others, 1971). In effect, quail are biologically prepared to associate visual cues with illness, while rats are biologically prepared to associate taste cues with illness.

Taste aversion research emphasizes that the study of learning must consider the unique behavior patterns and capabilities of different species. As the result of evolution, animals have developed unique forms of behavior to adapt to their natural environments (Bolles, 1985). These natural behavior patterns and unique characteristics ultimately influence what an animal is capable of learning—and how easily it can be conditioned to learn a new behavior.