Updated on
August 27, 2026
Ivan Pavlov's Theory of Classical Conditioning
Learn Ivan Pavlov's theory through the dogs experiment, including conditioned and unconditioned stimuli, extinction, examples and classroom implications.

Updated on
August 27, 2026
Learn Ivan Pavlov's theory through the dogs experiment, including conditioned and unconditioned stimuli, extinction, examples and classroom implications.
What is Ivan Pavlov's theory?
Ivan Pavlov's theory of classical conditioning explains how an animal learns that one event predicts another. In his tests, a sound became a conditioned stimulus after it often came before food. The sound then caused salivation as a conditioned response. This is about cues and mainly reflex-like responses, not rewards or punishments for voluntary behaviour.
Ivan Pavlov's theory of classical conditioning explains how an animal learns that one event predicts another. In his tests, a sound did not cause salivation at first. After it often came before food, the sound began to cause salivation. The sound became a conditioned stimulus, and salivation to it became a conditioned response (Pavlov, 1927).
This form of learning deals with links between events and mainly reflex-like responses. It is not the same as using rewards or sanctions to change an act. Teachers can ask how a cue may gain emotional or bodily meaning. They cannot diagnose conditioning from one classroom reaction.
This guide places Pavlov within the wider family of learning theories. It keeps detailed material about Watson, Skinner, behaviour management and clinical treatment with the pages that own those subjects.
Pavlov's theory describes how a signal starts to cause a response because it predicts another event. Pavlov called the learned result a conditioned reflex. Today, the wider process is known as Pavlovian or classical conditioning.
The familiar example pairs a sound with food. Food already elicits salivation in the experimental setting. Once the sound reliably comes before food, the sound can elicit salivation on its own. The animal has learned a relation between events, not a verbal rule about the sound.
The 1927 English text uses conditioned reflex. A recent translation uses conditional reflex as a closer match for the Russian term (Pavlov et al., 2023). Modern texts still use conditioned stimulus and conditioned response.
Ivan Petrovich Pavlov was a Russian physiologist who lived from 1849 to 1936. He studied how the body digests food, how blood flows and how nerves work. His 1904 Nobel Prize was for his work on digestion, not classical conditioning.
Pavlov noticed that dogs sometimes salivated before food reached their mouths. Signals associated with feeding, such as the approach of an experimenter, could begin to elicit secretion. He turned this observation into a controlled programme for measuring how signals acquired and lost their effects.
Pavlov used many types of signal, including metronomes, buzzers, lights and touch. “Pavlov's bell” is useful shorthand. It was not the name of one main test or the only signal he used.
The test is clearest in three stages. The labels show the role each event plays. They do not mean that every sound, food or body response has the same role in each case.
On a smaller screen, swipe across to read all columns.
| Stage | What happens | Terms |
|---|---|---|
| Before conditioning | Food elicits salivation. The experimental sound does not yet elicit the learned salivary response. | Food is the US; salivation to food is the UR; the sound is a neutral signal for this relation. |
| During conditioning | The sound reliably precedes food across trials. The sound gains information about what will happen next. | The signal and food form an informative CS-US relation. |
| After conditioning | The sound can elicit salivation before food appears. | The sound is now a CS; salivation to it is the CR. |
There is no set number of pairings. Learning varies with the animal, response, cue, timing and setting. It also varies with how much the cue tells the animal. The claim that Pavlov's dogs learned after exactly twenty pairings is not a general law.
These five terms show what changes during the test. Unconditioned means that the response does not depend on the pairing being studied. It does not mean that every animal has the same response or that past events never affect it.
The conditioned response need not be an exact copy of the unconditioned response. Learning changes how an animal reacts to a cue that predicts an event. The form of the response depends on the test and what the cue tells the animal (Domjan, 2005).
Acquisition is the growth of a response to a conditioned stimulus. It rests on the link between a cue and an outcome, not a fixed amount of practice. A cue that comes before an outcome gives more useful information than one that occurs at random.
Extinction is the decline in conditioned responding when the CS is repeatedly presented without the US. If the metronome sounds but food no longer follows, salivation to the sound may fall. Extinction does not simply delete the original learning.
Spontaneous recovery is the return of some learned responding after time has passed. Renewal is a return when the setting changes. These effects show why a weak response in one setting does not prove that the first learning has gone (Bouton, 2002).
A classroom routine that stops working is not automatically an example of extinction. Learners may have misunderstood an instruction, stopped attending, responded to different consequences or decided that the rule is no longer enforced. The observed behaviour alone does not reveal the mechanism.
Stimulus generalisation occurs when cues like the conditioned stimulus also cause a response. A dog trained with one tone may respond to tones that sound like it. The response tends to fade as the new tone becomes less like the trained cue.
Discrimination grows when signals have different links to the outcome. If one tone predicts food and another does not, the dog may respond more to the first tone. It learns which differences matter.
Teachers can use these terms as one way to read an emotional response. A learner who feels tense in several test rooms may react to shared cues. Yet the teacher must also check the language, task, sensory needs, adult relationships and past results.
Simple accounts say that learning occurs because two events keep appearing together. Later tests showed that this is not enough. A cue is useful when it changes the chance of an outcome or tells the animal more than other cues do.
Rescorla (1968) compared the probability of shock when a signal was present with the probability when it was absent. Little conditioning occurred when the signal gave no extra information, even though the signal and shock sometimes occurred together. This established the importance of contingency.
The Rescorla-Wagner model says learning changes when the result differs from what the available cues predict (Rescorla and Wagner, 1972). It is a later model, not a theory made by Pavlov. It helps explain blocking. In this effect, a cue that already predicts the result can limit learning about a new cue beside it.
For teachers, the safe lesson is not that “consistency conditions learners”. It is that recurring cues can acquire meaning from what they predict. Whether an observed classroom response is Pavlovian remains an inference that needs other explanations to be checked.
In people, Pavlovian learning can shape body and emotional responses. A smell may cause nausea after an illness. A clinic may cause a stress response after painful care. A neutral image paired with pleasant or unpleasant material may gain some of that emotional value.
Taste-aversion research shows that some relations are learned more readily than others. Garcia and Koelling (1966) found that rats associated taste with later illness more readily than with shock, while audiovisual cues showed a different pattern. This challenged the idea that every cue and consequence can be linked equally easily.
Watson and Rayner's 1920 report about “Albert B.” is often used as a human example. It was a single, methodologically limited case, and later textbook versions exaggerated what it established (Harris, 1979). The procedure would not meet modern ethical standards. The detailed study belongs in the John B. Watson guide.
Conditioning may play a part in anxiety, but one learned link cannot explain it all. Thoughts, doubt, social judgement, past results, health and context also matter. Ongoing panic, trauma signs or attendance difficulty need the school's safety, SEND and mental-health routes. A trained clinician may also need to help.
Pavlov gives teachers one way to ask how a cue may gain emotional or bodily meaning. A room, sheet, sound, adult voice or test format can become linked with what often follows it. This is one possible account, not a diagnosis.
Consider a learner who becomes visibly tense when a particular test booklet appears. The booklet may have acquired a negative association through repeated stressful experiences. The teacher should also check the reading load, time pressure, prior attainment, sensory conditions, public comparison and clarity of the task.
Safe practice includes saying what will happen, correcting in private and making the task easy to access. It also gives the learner a fair chance of success. A teacher can say, “Read the first question for two minutes. I will check the wording with you before the timer starts.” This cuts doubt without a claim to provide therapy.
A learner packing away when a bell rings is not, by itself, proof of classical conditioning. The act may follow an instruction, a school rule or past results. Attention signals, routines, rewards and sanctions belong mainly to classroom practice and operant conditioning.
Teachers should not use scents as a conditioning tool. Allergies, asthma, sensory needs and personal links make this unsafe. They should also not provide exposure therapy, systematic desensitisation or trauma treatment.
Classical conditioning deals with links between events and responses caused by predictive cues. Operant conditioning deals with links between an act and what follows it. The difference is not a passive learner versus an active learner.
On a smaller screen, swipe across to read all columns.
| Question | Classical conditioning | Operant conditioning |
|---|---|---|
| What relation is learned? | One event predicts another. | A behaviour has particular consequences. |
| What is measured? | Responding elicited by a predictive cue. | How often, how strongly or in what setting behaviour occurs. |
| Pavlov example | A metronome predicts food and elicits salivation. | Not the focus of Pavlov's dog preparation. |
| Classroom boundary | A cue may acquire emotional or physiological significance. | Praise, rewards or sanctions may change voluntary behaviour through their consequences. |
For reinforcement, punishment, schedules and voluntary classroom behaviour, use the operant conditioning guide. For the wider history and assumptions of the school of thought, use the behaviourism theory guide.
Watson's 1913 manifesto argued for an objective science of behaviour. He later incorporated conditioned-reflex research into his behaviourist programme (Watson, 1913, 1916). Pavlov's careful tests gave psychologists a way to study links between cues and responses.
Pavlov was not a classroom behaviourist. He was a physiologist who studied nerves and conditioned reflexes. Watson's manifesto, human tests and later programme grew from related ideas. They are not all parts of one “Pavlov theory”.
This matters because behaviourism later split into different views. Watson and Skinner did not make the same claims. The behaviourism overview explains those differences.
Pavlov built clear tests of how a cue gains control over a response. Later work on contingency, extinction and related effects made the account more exact. It did not make Pavlov's work pointless.
The main evidence came from controlled tests with animals. This helps researchers isolate a learning process, but it limits direct claims about class. People also use language, goals, thought and social ties. They learn within school rules and culture.
Classical conditioning is not a complete theory of education. It does not explain how learners understand a proof, organise historical evidence or monitor a complex strategy. These tasks require broader accounts of cognition, teaching and social interaction.
Animal research and the Little Albert report also raise ethical concerns. Rules on consent, welfare, harm and withdrawal have changed greatly. Teachers should explain old studies with care. They must not present distress as a method to copy.
Brain research links threat learning to a wide network. It includes the amygdala, hippocampus and parts of the cortex. No single brain area simply stores fear. An animal response also cannot explain the whole of a learner's conscious anxiety (LeDoux and Pine, 2016).
Pavlov showed that a signal can acquire the capacity to elicit a response when it reliably predicts a significant event. A metronome that precedes food can eventually elicit salivation before the food appears.
Pavlov used many signals, including metronomes, buzzers, lights and tactile stimuli. The bell is popular shorthand for his work, but it is not a good label for one definitive experiment.
The unconditioned stimulus (US) elicits the relevant unconditioned response (UR) without the experimental pairing. After learning, a conditioned stimulus (CS) elicits a conditioned response (CR). In the standard dog example, food is the US, salivation to food is the UR, the metronome becomes the CS and salivation to it is the CR.
Not necessarily. Learners may respond because they understand a rule, expect a consequence or follow an instruction. A sound can acquire emotional or physiological associations, but the visible routine alone does not establish a Pavlovian mechanism.
Pavlov studied how signals come to elicit responses by predicting other events. B. F. Skinner studied how the consequences of behaviour change its future occurrence. Rewards, reinforcement and punishment belong primarily to Skinner's operant account.
No. The response may fade when the CS no longer predicts the US. It can later return through spontaneous recovery or renewal. Extinction is best seen as new learning that depends on the setting.
These sources support the history, key terms, later updates, human examples and limits in this guide.