Thorndike’s psychological research on learning was contemporary with the physiological studies of the nervous system made by Ivan Petrovich Pavlov. At the time he began writing on the conditioned reflex Pavlov was over fifty years old, having spent the earlier part of his life investigating the circulatory system and digestive glands. It was during his study of the digestive glands that his attention turned to “psychical” stimulation of the glands:

While studying, over the course of many years, the normal activity of the digestive glands, and analyzing the constant conditions of this activity, I came upon facts (which had also been observed by others) of a psychical character, facts which could not rationally be neglected, as they participated constantly and prominently in the normal mechanism of the physiological process. I was obliged to consider them if I wished to make the most thorough possible study of my subject. (Pavlov et al., 1928, p. 47)

What Pavlov et al. (1928) observed is that “the activity of the digestive glands was called out not only when the food was in the mouth or had passed further along the digestive tube, but by agents acting from a distance, such as the sight, odor, etc., of the foodstuff” (p. 22). Before going farther, it’s important for the reader to note that Pavlov was not a psychologist and he was not investigating problems of learning. In his own preface to the only comprehensive, firsthand, report of his work, Pavlov made this clear when he said “this book concerns the investigation of the physiology of the cerebral hemispheres by the strictly objective method of conditioned reflexes” (p. 20). Although his study was focused on indentifying which nerves were connected to which digestive glands and the conditions under which these glands were innervated, as he turned to examine the psychical stimulation of the glands he noted that “similar experiments on animals had been performed in America, and indeed not by physiologists but by psychologists” (p. 39). He credited “the honor of having made the first steps along this path belongs to E. L. Thorndike” (p. 40) whose experiments had preceded those conducted in Pavlov’s lab by two or three years.

Though he was initially tempted to assume the role of a psychologist, Pavlov’s decision to study the stimulation of nerves by agents acting from a distance through an objective physiological approach, was a very conscious and intentional one:

After persistent deliberation, after a considerable mental conflict, I decided finally, in regard to the so-called psychical stimulation, to remain in the role of a pure physiologist, i.e., of an objective external observer and experimenter, having to do exclusively with external phenomena and their relations. (p. 40)

Pavlov’s ensuing research aimed at discovering the laws that govern the association between external objects and the secretion of the salivary glands. He called such associations conditioned reflexes to differentiate them from unconditioned reflexes, or reflexes which had a natural association with the salivary glands, such as the secretion produced by putting acid or sand in the dog’s mouth. The initial association for a conditioned reflex was formed by the conditioned stimulus (e.g. the sight of dry bread) occurring in time with the unconditioned stimulus (e.g. the sensation of dry bread in the mouth). Today the process through which an association is made between a conditioned stimulus and an unconditioned response is commonly referred to as classical conditioning.

The observation that such associations are made—with or without intentionality on the part of the dog or the investigator—was Pavlov’s basic starting point. The experiments that followed typically dealt with understanding how these associations were formed, maintained, terminated, strengthened or attenuated. For 25 years, many experiments were conducted in Pavlov’s labs along these lines. So many, in fact, that even Pavlov was not able to tear himself away from his research to prepare a summative report of his work. In his own words, “How could I halt for any comprehensive conception, to systematize the results, when each day new experiments and observations brought us additional facts!” (Pavlov et al., 1928, p. 42).

Though Pavlov described ‘psychic stimulation’ of salivary and gastric secretion in his 1897 book Die Arbiet der Verdauungsdrusen, the first report from his laboratory giving a systematic description of ‘natural’ conditioned reflexes was a thesis published by Dr. Wolfson entitled Observations upon salivary secretion (Petrograd, 1899, as cited in Rosenzweig, 1960, p. 313).  The term ‘conditioned reflex’ was first used in print by Dr. Tolochinov, who completed the first experiments in Pavlov’s lab on the conditions under which psychical salivary secretion reflexes appear in 1901 and communicated his results to the Congress of Natural Sciences in Helsingfors in 1903[1] (Pavlov et al., 1928). Many reports were published after that, but no comprehensive reports were available Lectures on conditioned reflexes by Pavlov et al. was published in 1928.

Olson and Hergenhahn (2009) summarized some of the most important concepts that were empirically founded through Pavlov’s experiments:

1. The process by which a conditioned reflex is developed.

To produce a CR [conditioned response], the CS [conditioned stimulus] and the US [unconditioned stimulus] must be paired a number of times. First the CS is presented and then the US. The order of presentation is very important. Each time the US occurs, a UR [unconditioned response] occurs. Eventually the CS can be presented alone and it will elicit a response similar to the UR. When this happens, a CR has been demonstrated. (p. 164)

2. Experimental extinction. “If after a CR has been developed, the CS is continually presented without the US following the CS, the CR gradually disappears” (p. 164).

3. Spontaneous recovery. “After a period of time following extinction, if the CS is again presented to the animal, the CR will temporarily reappear” (p. 165).

4. Higher-order conditioning. “After a CS has been paired with a US a number of times…it can be paired with a second CS to bring about a CR” (p. 165). For example, a US (e.g. meat powder) causes a UR (e.g. salivation). If a light is presented just before the food powder a number of times, the light will come to cause a CR (i.e. salivation in response to the light). Once the light can elicit a CR, it can be paired with another CS, such as a buzzer, without presenting the meat powder. After a number of pairings of the buzzer, followed by the light, the buzzer alone will come to elicit the CR. In this example, the food powder is called a primary reinforcer, the light is a secondary reinforcer, and the buzzer is a tertiary reinforcer. Second-order conditioning (the buzzer), and third-order conditioning (e.g. a tone of 2,000-cps paired with the buzzer) were often achieved in Pavlov’s experiments. However, because second- and third-order conditioning must occur during extinction (i.e. the CS is presented without the meat powder) there is no reinforcement and the CS gradually loses its ability to call out the response. Because of this, it is very uncommon to go beyond third-order conditioning.

5. Generalization. Stimuli similar to the CS will also elicit the CR. For example:

Once [a dog has been conditioned to salivate in response to a 2,000-cps tone], we enter the extinction phase of the experiment, only this time we will expose the animal to tones other than the one it was trained on. Some of the new tones will have a frequency higher than 2,000 cps, and some will have a lower frequency. Using the number of drops of saliva as our measure of the magnitude of the CR, we find that the CR has its greatest magnitude when the 2,000-cps tone is presented, but CRs are also given to other tones. The magnitude of the CR given to the other tones depends on the similarity to the tone the animal was actually trained on; in this case, the greater the similarity to the 2,000-cps tone, the greater the magnitude of the CR. (p. 166)

6. Discrimination. Discrimination is the opposite of generalization:

Generalization refers to the tendency to respond to a number of stimuli that are related to the one actually used during training. Discrimination, on the other hand, refers to the tendency to respond to a very restricted range of stimuli, or to only the one used during training. (p. 167)

Discrimination is developed through two methods: prolonged training and differential reinforcement.[2]  The greater the number of times a CS is paired with a US, the less likely there is to be a CR to a CS that is similar, but not identical to the one used for training. Through differential training, a variety of similar CSs are presented, but only the one of interest is followed by reinforcement. Through this process, the tendency for a CR to follow a non-reinforced CS is diminished and extinguished.

[1] Dr. Tolochinov delivered his paper in 1902 and it appeared in print the following year. (Rosenzweig, 1960, p. 313)

[2] An interesting side note regarding the limits of discrimination possible when using the second method of training, differential reinforcement, is the account Pavlov gave of conditioning a dog to respond to the shape of a circle illuminated on a screen close in front of the dog, but to not respond to an ellipse. All circles were reinforced by feeding. No ellipses were reinforced. The first ellipse was very different from the circle. Over successive trials the shape of the ellipse was brought closer to the shape of the circle to the point where the axes were 9:8, or nearly circular. At this point, a salivary response was produced and all of the previous differential training was undone. The dog began responding to even the more elongated ellipses to which no response was previously made. Even more interesting, however, was the apparent development of experimental neurosis in the dog as a result of being presented with stimuli between which he could not discriminate. Where previously the dog would stand quietly on his bench, he would now struggle and howling constantly. (Pavlov, Gantt, & Volbort, 1928, p. 342)


Connectionism (Edward L. Thorndike – 1898)Behaviorism (John B. Watson – 1913) >

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s