Learning theories

Learning theories

Learning can be defined as the process of changing the concrete or potential behavior of an individual in relation to his experience. It results from the interaction of motivation, emotion, memory and, therefore, can not be reduced to a sterile mechanism of assimilation of contents .

Learning is therefore an adaptive response of the subject to the environment and the result of a previous experience. For this reason it is intrinsically connected to memory, cognitive development and intelligence. Over time, various studies have been carried out, in particular since:


The study of learning begins with behaviorism which can be defined as the attempt to clarify behavior in terms of relationships between observable stimuli , or events that occur in the environment and observable responses, or behavioral acts. In particular: – Pavlov in 1927 made known the results of his studies conducted on associative learning, based on the stimulus-response association . –

Thorndike and later Skinner introduced the paradigm of operant conditioning , introducing negative reinforcement to end behavior and positive to increase it.

Social learning theorists

In the 1940s and 1950s they expanded the concept of learning by taking an interest in social behavior and the social context of behavior according to a principle of mutual triadic causality, in which the environment, the person and the behavior influence each other. In particular: – Tolman’s latent learning Tolman highlighted the latent learning that takes place without apparent external indicators and consists in the structuring of cognitive maps that are used when necessary. –

KÖhler’s insight learning

The gestaltista Kohler highlighted the learning by insight that occurs through sudden intuitions , resulting however from a process of observation, assimilation and creativity.

Learning pears modeling of Bandura

Bandura proposes the concept of learning by modeling : the individual modifies his own behavior based on the behavior of another individual, taken as a model; the latter follows important criteria such as age, sex, social status, but also the outcome of behavior , as it is often learned by looking at the outcomes that learned behavior has when others implement it, or the so-called conditioning vicar .

Bandura’s experiment shows that children subjected to violent video are more violent than children not subjected to video or quiet video

Harlow’s willingness to learn

Harlow directs her studies to learning to learn, going beyond specific skills and contents and analyzing meta-learning, that is the acquisition of a transversal method applicable to different contents.

Flavell’s metacognition

Finally from the 70s, with Flavell, the concept of metacognition is introduced , which includes the processes of control and monitoring of learning that proceed beyond the acquisition of a method to apply and evaluate the aptitude for flexibility, the selection of methods, the management of attentitve resources and the persistence in a task until the achievement of a goal.

The main learning studies are:

Classical conditioning of Pavlov

This type of learning was discovered by chance by the Russian physiologist Pavlov in the early 1900s, following studies on the salivation of dogs. After noticing that the animal salivated when he saw and received food, he considered the latter as an unconditional SI stimulus and salivation as an unconditional RI response, as innate in the dog and not dependent on external factors.

He later found that salivation increased when they heard or saw something preceding the administration of food and therefore decided to associate the sound of a bell with food. Pavlov noticed that, after a short while, the dog on hearing the bell salivated more; thus he defined the conditioned sound stimulus SC and the greater salivation conditioned response RC, as deriving from external factors.

He therefore understood how, by associating a conditioned stimulus with an unconditional one, a more complex response could be obtained in the wake of the unconditional one, and therefore based on experience our organism changes, it is a form of learning. From these studies other aspects emerged: if the sound-food association was suspended, the conditioned response tended to disappear , but re-intensified when it was re-presented, demonstrating how the Conditional Response had only been inhibited rather than forgotten.

This spontaneous recovery and rapid re-learning demonstrate the difficulty in eliminating the effects of conditioning. Furthermore, he discovered the importance of the time and order factor, as he demonstrated that the closer temporally are SC and SI and the more easily the conditioning is obtained, and the same is true for an SC-SI order rather than SI-SC.

In summary, learning reaches its maximum effectiveness when the SC slightly precedes the SI. Another aspect is that when one has learned to associate a Conditional Stimulus with an unconditional Stimulus, stimuli similar to the conditional one also initiate the same conditional response themselves, but to a lesser extent the greater the difference with the original SC.

Watson’s experiment

We then speak of generalization of the stimulus which can be further explained through Watson’s experiment. He showed how a child, little Albert, after associating a disturbing noise with a mouse, was afraid of all furry animals .

That of little Albert is clearly a maladaptive conditioning in that, associating an object / animal with a traumatic experience, it brings a phobia. However, there may also be an adaptive conditioning of fear , such as when staying away from something dangerous.

Finally, the inverse process of generalization is that of discrimination between two relatively similar stimuli. In this case, while continuing to respond to a certain SC, it is possible, with good training, not to respond to other stimuli similar to the SC. In this regard Pavlov accustomed the dog not to respond to any sound, but only those with a certain pitch.

The operative conditioning of Thorndike and and Skinner

Operant conditioning of Skinner, who formulated his theory based on Thorndike’s studies of learning by trial and error. In the most significant experiment Thorndike, studying the behavior of a cat inside a cage with a lever that, if pressed, distributed food, understood that the animal, after having pressed it a few times accidentally, sensed the functioning of the mechanism; he called this phenomenon the law of effect, that is, an animal tends to repeat a behavior based on the effect it produces.

The Skinner Box

Skinner built a box very similar to that of Thorndike, the Skinner Box , and making various experiments on animals he formulated his theory on operant conditioning, that is that subjects can develop or modify behavior not through an innate reflex, but by operating and acting in the environment thanks to the stimulus-response association :

by pressing the lever by mistake (stimulus) and the consequent spillage of food (response) the animal understands (operating conditioning) that he must press the lever to obtain the food (behavior acquisition).

The conditioning is, therefore, determined by the reinforcement, that is, any event that increases the probability of the appearance of a response.

Primary reinforcements are distinguished, which leverage innate conduct; secondary reinforcements , or initially neutral stimuli that become reinforcing as they are learned over time and are often linked to culture; postive reinforcements , which consist in the repetition of a pleasant stimulus; and negative reinforcements, which correspond to the cessation of an unpleasant stimulus.

Positive reinforcement vs punishment

The opposite of reinforcements are punishments, that is, anything that can decrease the probability of a response appearing. In Thorndike’s studies, studying the behavior of the guinea pigs which, in this case, when they pressed the lever also received an electric shock, he noticed that the punishment temporarily decreases the intensity and frequency of the behavior but is unable to extinguish it ; indeed when he came back he was accompanied by emotions like anger and aggression. What is more effective, however, is what Skinner defines as the modeling technique, that is, progressive and systematic reinforcements that precisely shape the behavior until the desired conduct is achieved.

Differences between classical and operative conditioning

In conclusion, the main differences between classical and operative conditioning are that in the first condition pre-existing physiological reflexes are conditioned to the stimulus and the subject is passive, while in the second new behaviors are conditioned and the subject is active.

Tolman’s latent learning theory

Tolman showed that learning does not coincide with performance and that reinforcement is not always necessary for the individual to learn as there is latent learning that occurs without showing up in behavior .

Tolman demonstrated this through the experiment of the three groups of rats that had to orient themselves in a labyrinth, a group reinforced with food, a group without any reinforcement, and one after ten days; he saw that although apparently the first group was the best, the third, as soon as it received reinforcement, surpassed the first, as it had built cognitive maps , i.e. mental representations of the labyrinth, in the first 10 days .

In this way Tolman introduces to the classic stimulus-response scheme an intermediate element, represented by mental processing and distinguishes the learning construct from the performance one , where the former, through the construction of cognitive maps, becomes observable only after reinforcement, while the second has to do only with the directly observable action through which reinforcement is obtained.

Kohler’s insight

Kohler’s Insight Learning Theory, who in agreement with Gestalt, realizing that the various learning models did not explain how daily problems are solved, developed his insight learning model , according to which learning is the outcome of an intelligent process. It presupposes the ability to connect together in a unified way distributed elements and considered, until then, isolated. After a phase of trial, trial and error, the subject suddenly arrives at the cognitive restructuring that allows insight, that is, a sudden learning in which the perceptual field tends to reorganize itself and the subject sees solutions that he had not previously thought of.

The experiments of 1917 are classic in which a monkey, closed in a cage, was asked to grab bananas beyond the bars. After an initial nervousness, the monkey, by stretching the sticks in the cage, was able to grasp them: this means that learning can come from sudden flashes that result from the original and creative use of tools in one’s possession .

From the comparison with theories of conditioning, two conceptions of learning emerge: according to the continuous conception, which characterizes behaviorism, learning is a slow step by step; while according to the discontinuous conception of Gestaltists , learning is immediate and creative .

An integrated line is currently being adopted which considers learning a continuous and progressive process which, however, does not exclude creative acts made possible by the quality of previous learning.