
Definition
Classical (Pavlovian) conditioning is a form of associative learning in which a neutral stimulus, repeatedly paired with a stimulus that already triggers a reflex-like response, comes to trigger a similar response by itself. In Pavlov's laboratory, a signal that reliably preceded food came to make a dog salivate.
Someone eats a shrimp salad at a picnic and gets violently sick that night. Weeks later, the smell of shrimp makes their stomach turn, even if they suspect the shrimp was not the real culprit. Nobody taught them to feel that way on purpose, and reasoning does not simply switch it off. That small, annoying piece of learning is the same process Ivan Pavlov studied more than a century ago, and it is still one of the most useful ideas in psychology.
This guide walks through what Pavlov actually did, the vocabulary and the five core phenomena, the timing arrangements that work best, how the idea moved from the laboratory to exposure therapy, what the textbook story gets wrong, and the exam mistakes to avoid.
The short version
- Pavlov, a Russian physiologist, won the 1904 Nobel Prize in Physiology or Medicine for his research on digestion. His conditioning work grew out of that research.
- Key terms: unconditioned stimulus (US), unconditioned response (UR), neutral stimulus (NS), conditioned stimulus (CS), conditioned response (CR).
- Core phenomena: acquisition, extinction, spontaneous recovery, generalization and discrimination.
- Modern accounts stress prediction: the organism learns that the CS signals the US (Rescorla, 1988).
- Extinction does not erase the original learning; it adds new learning that competes with it (Bouton, 2004). That idea underlies exposure therapy.
Classical conditioning in one minute
Before any learning, the organism has an innate association: an unconditioned stimulus (US) produces an unconditioned response (UR). Food in the mouth produces salivation; a sudden blast of noise produces a startle. The nervous system arrives equipped with that connection.
Pair a neutral stimulus (NS) repeatedly with the US and the NS begins to evoke a similar response by itself. It has become a conditioned stimulus (CS), and the response it produces is the conditioned response (CR). In Pavlov's classic paradigm, signal plus food leads to salivation; after several trials, the signal alone does.
Rescorla (1988) stressed a crucial nuance: this is not a simple connection between two stimuli. What the organism learns is that the CS predicts the US. The relation is informational, not mechanical.
What Pavlov actually did
Pavlov was measuring digestive secretions in dogs, using a surgical preparation that let him collect saliva. He noticed that the dogs began to salivate before food arrived, at the sight of the person who fed them or the sound of approaching footsteps. He called these "psychic secretions" and, in his later work, systematically paired neutral signals with food and recorded the results (Pavlov, 1927).
A correction to the usual story: the textbook image of a bell is a simplification. According to Pavlov's biographer, Daniel Todes (2014), the lab used mostly stimuli whose timing and quality could be tightly controlled, such as metronomes, buzzers and other devices, and bells were rare. The reason matters: Pavlov's goal was precise control over the stimulus. Learning the principle is more important than remembering the signal.
The vocabulary
- Unconditioned stimulus (US): triggers a response without learning (food in the mouth).
- Unconditioned response (UR): the natural reaction to the US (salivation).
- Neutral stimulus (NS): a stimulus that initially produces no such response (a metronome).
- Conditioned stimulus (CS): the NS after pairing with the US.
- Conditioned response (CR): the learned response to the CS (salivation to the metronome).
The CR is usually similar to the UR but not identical, and it can be weaker or differently timed. The key distinction for exams: the UR is unlearned, the CR is learned.
A typical conditioning experiment

The procedure Pavlov followed in his St. Petersburg laboratory can be replicated today with minimal adaptations. The usual sequence has six steps:
- Choose a US that reliably produces a measurable response. Food and salivation, or a mild shock and a skin-conductance response.
- Check that the NS produces no such response on its own.
- Acquisition trials. Present the NS followed by the US, repeatedly. The CR grows with pairings, quickly at first and then more slowly.
- Test trials. Present the CS alone and see whether the CR appears.
- Plot the acquisition curve. The strength of the CR rises quickly in the first trials and then levels off (an asymptotic function). Graphing the curve is part of the protocol.
- Extinction trials. Present the CS repeatedly without the US and watch the CR weaken.
Modern human research uses the same logic with different measures. A neutral image (a blue square, a neutral face) may be paired with a mild shock or loud noise, and researchers record skin conductance, heart rate or brain activity. The procedure is Pavlov's; the instruments are finer.
Timing: which arrangement works best?
| Procedure | Timing | Typical result | Example |
|---|---|---|---|
| Delay | CS begins before the US and stays on until the US arrives | Generally the strongest conditioning | Tone sounds and continues until food appears |
| Trace | CS begins and ends, then a gap, then the US | Weaker than delay; depends on the gap length and on memory | Tone ends, a few seconds pass, then food |
| Simultaneous | CS and US start together | Often weak, because the CS does not predict anything | Tone and food begin at once |
| Backward | US comes first, then the CS | Weak or even inhibitory | Food, then the tone |
The reason timing matters is information. Rescorla (1967, 1988) argued that conditioning depends on whether the CS provides information about the US, not simply on the two being close in time. A tone that sounds just as often when no food comes tells the animal little, even if food and tone often occur together. These are standard textbook generalizations; exact outcomes vary by species, response system and procedure.
Five core phenomena
- Acquisition: the CR strengthens over CS–US pairings, typically with a negatively accelerated curve (large gains early, smaller later).
- Extinction: the CR weakens when the CS is presented without the US. Bouton (2004) reviews evidence that extinction does not erase the original association but creates a new, context-dependent inhibitory learning that competes with it.
- Spontaneous recovery: after a rest period, an extinguished CR can partially return when the CS is presented again.
- Generalization: stimuli similar to the CS also evoke the CR, more strongly the more similar they are. Someone frightened by one large dog may fear other large dogs.
- Discrimination: with experience, the organism responds to the CS but not to similar stimuli that are never paired with the US.
Two related effects are worth knowing. Renewal is the return of an extinguished response when the CS is encountered outside the context where extinction took place. Higher-order (second-order) conditioning occurs when an established CS serves as a stand-in US: a light paired with a tone that already triggers the CR can come to trigger it too, even though the light was never paired with food.
From the laboratory to the clinic

The bridge from Pavlov's dogs to contemporary therapy was built in steps: a demonstration that human fears can be learned, a first systematic treatment and decades of refinement.
Little Albert
Watson and Rayner (1920) reported conditioning fear in a nine-month-old infant, "Albert," by pairing a white rat with a loud noise. The fear was said to spread to similar objects such as a rabbit and a fur coat. The study is often used to show that fears can be learned, but it is also a cautionary tale. It would be unacceptable under modern ethics standards, it involved a single child, and later accounts show that textbooks often repeated and embellished its findings (Harris, 1979). Efforts to identify the child and his later life are still debated (Beck et al., 2009). Use it as a historical example, not as proof of how all phobias develop.
Exposure therapy and inhibitory learning
Wolpe (1958) developed systematic desensitization, in which a person builds a hierarchy of feared situations and imagines them while relaxed. He framed it as reciprocal inhibition: a response incompatible with anxiety, such as relaxation, weakens the anxiety response. Exposure has since been refined into several modalities: in vivo (real situations), imaginal, interoceptive (bodily sensations) and virtual reality. Modern exposure therapy has moved toward direct, repeated contact with feared cues without the feared outcome. A review of exposure for obsessive-compulsive disorder and posttraumatic stress disorder describes it as among the best-supported psychological approaches for those conditions and examines its underlying mechanisms (Foa & McLean, 2016).
The inhibitory learning approach (Craske et al., 2014) draws on extinction research. Because extinction creates new competing learning rather than erasing the old, exposure sessions are designed to maximize expectancy violation (the person expects a catastrophe and something else happens) and to vary contexts, so that the new learning generalizes. Simply "enduring" fear is not the point. Milad and Quirk (2012) review the brain circuits involved in fear extinction in animals and humans, which helps explain why relapse can occur after successful treatment.
This section describes research and theory for students. It is not a treatment guide, and exposure-based treatment should be delivered by trained clinicians.
Not every cue is equally easy to condition
Öhman and Mineka (2001) proposed that humans are biologically prepared to learn fears of stimuli that were threats in evolutionary history, such as snakes and spiders, and found that such fears are readily acquired and resist extinction. Garcia and Koelling (1966) showed a different kind of selectivity: in rats, taste was readily linked to illness, but sights and sounds were readily linked to shock, not to sickness. That is why the picnic shrimp example is so natural. It also reminds you that classical conditioning is not a blank-slate process.
Addiction and cues
People who use substances show reactions to drug-related cues (places, people, paraphernalia), a phenomenon reviewed in a meta-analysis of cue reactivity (Carter & Tiffany, 1999). Treatments that expose people to cues without use have been proposed on the basis of extinction, but the results have been mixed, partly because extinction is context dependent (Conklin & Tiffany, 2002).
A word on the amygdala
Research on threat conditioning in the amygdala has been very influential. LeDoux (2014) cautions, though, that the circuits that detect and respond to threats should be distinguished from those that produce the subjective feeling of fear, so "fear conditioning" in animals should not be read as a direct model of conscious fear.
How an exposure protocol is structured
In clinical practice, exposure follows a structured protocol rather than improvisation. For study purposes, the usual sequence looks like this:
- Behavioral formulation. Identify the CS (the trigger), the historical US (the original aversive experience, or learning by observation or by information), the current CR and the avoidance behavior that maintains the fear.
- Psychoeducation. Explain the model in accessible language, so that the reason for facing feared cues is understood.
- Exposure hierarchy. Build a graded list of situations, from mild discomfort to the highest tolerable level, often rated on a 0–100 subjective units of distress scale (SUDS).
- Planned exposures. Schedule them to maximize expectancy violation: the feared prediction is tested and something different happens (Craske et al., 2014).
- Planned generalization. Vary contexts, times of day and the presence of the therapist. Without variation, extinction stays tied to its original context and the response can return outside it (Bouton, 2004).
- Relapse prevention. Teach about spontaneous recovery and plan a brief re-exposure if symptoms return. Because extinction does not erase the original learning, relapse prevention is part of the model, not an add-on (Bouton, 2004; Craske et al., 2014).
Beyond fear: other clinical applications
Classical conditioning explains more than phobias. It has been applied to several clinical domains:
- Addiction. Places, people and times of day associated with use become conditioned cues that evoke craving (see above; Carter & Tiffany, 1999).
- Conditioned disgust and nausea. Food aversions learned after illness (Garcia & Koelling, 1966) and the nausea some patients come to feel before chemotherapy have been analyzed as classical conditioning (Stockhorst et al., 2006).
- Chronic pain. Fear-avoidance models describe how fear of movement, once learned, leads to avoidance that helps maintain pain-related disability (Vlaeyen & Linton, 2000). The logic resembles the Pavlovian account of phobias: avoidance prevents the fear from being extinguished.
- Illness anxiety. Ambiguous bodily sensations linked to the fear of a serious diagnosis can produce hypervigilance that keeps the problem going, a conditioning-style account.
Common mistakes that cost points on exams
- Mixing up the UR and the CR. The UR is unlearned; the CR is learned. They are not always identical.
- Naming the wrong stimulus. Before learning, the signal is a neutral stimulus; after learning it is the conditioned stimulus.
- Getting the order wrong. The CS should signal the US. Backward arrangements are weak.
- Saying extinction means forgetting. It is new learning that suppresses the response, which is why spontaneous recovery and renewal occur.
- Confusing classical with operant conditioning. If the response is elicited by a signal regardless of what the learner does, it is classical. See classical vs. operant conditioning.
- Repeating the bell as fact. It is a simplification of Pavlov's actual procedures.
- Presenting Little Albert as strong proof. It is a single, flawed case study.
A student's question
"Do phobias always come from a bad experience?" No. A direct conditioning experience is one pathway, but people also acquire fears by watching others or by being told about danger, and many people with a phobia cannot recall an event. Learning, biology and context combine. For learning by observation, see Bandura's social learning theory.
Frequently asked questions
Is classical conditioning still relevant now that cognitive therapy exists?
Yes. Cognitive-behavioral therapy combines cognitive components (restructuring) with behavioral ones (exposure), and the behavioral component rests on principles studied in Pavlovian conditioning. The two are not opposed in current practice.
Can a person be conditioned without noticing?
Yes. Conditioned responding can occur without the person being able to report the pairing, and many people with specific fears cannot name where the fear came from. The role of awareness varies by response system and procedure.
Does extinction erase learning or suppress it?
It suppresses it. Bouton (2004) and Milad and Quirk (2012) show that the original learning remains latent and can return with the passage of time, a change of context or stress. This does not invalidate therapy: it explains why relapse prevention is necessary.
How long do the effects of well-delivered exposure therapy last?
Exposure is among the best-supported psychological treatments for obsessive-compulsive disorder and posttraumatic stress disorder (Foa & McLean, 2016). How long gains last varies by disorder and person, which is why protocols include generalization and a relapse-prevention plan.
Did Pavlov discover conditioning by accident?
He was studying digestion when he noticed that dogs salivated in anticipation of food. He recognized the importance of the observation and built a research program around it, so it is better described as an observation he took seriously than as a lucky accident.
What is the difference between classical conditioning and operant conditioning?
In classical conditioning, a signal predicts a stimulus and elicits a response. In operant conditioning, behavior is shaped by its consequences.
Keep reading
- Classical vs. operant conditioning — how the two models differ and work together.
- Bandura's social learning theory — learning without direct experience.
References
- Beck, H. P., Levinson, S., & Irons, G. (2009). Finding Little Albert: A journey to John B. Watson's infant laboratory. American Psychologist, 64(7), 605–614. https://doi.org/10.1037/a0017234
- Bouton, M. E. (2004). Context and behavioral processes in extinction. Learning & Memory, 11(5), 485–494. https://doi.org/10.1101/lm.78804
- Carter, B. L., & Tiffany, S. T. (1999). Meta-analysis of cue-reactivity in addiction research. Addiction, 94(3), 327–340. https://doi.org/10.1046/j.1360-0443.1999.9433273.x
- Conklin, C. A., & Tiffany, S. T. (2002). Applying extinction research and theory to cue-exposure addiction treatments. Addiction, 97(2), 155–167. https://doi.org/10.1046/j.1360-0443.2002.00014.x
- Craske, M. G., Treanor, M., Conway, C. C., Zbozinek, T., & Vervliet, B. (2014). Maximizing exposure therapy: An inhibitory learning approach. Behaviour Research and Therapy, 58, 10–23. https://doi.org/10.1016/j.brat.2014.04.006
- Foa, E. B., & McLean, C. P. (2016). The efficacy of exposure therapy for anxiety-related disorders and its underlying mechanisms: The case of OCD and PTSD. Annual Review of Clinical Psychology, 12, 1–28. https://doi.org/10.1146/annurev-clinpsy-021815-093533
- Garcia, J., & Koelling, R. A. (1966). Relation of cue to consequence in avoidance learning. Psychonomic Science, 4(1), 123–124. https://doi.org/10.3758/BF03342209
- Harris, B. (1979). Whatever happened to Little Albert? American Psychologist, 34(2), 151–160. https://doi.org/10.1037/0003-066X.34.2.151
- LeDoux, J. E. (2014). Coming to terms with fear. Proceedings of the National Academy of Sciences, 111(8), 2871–2878. https://doi.org/10.1073/pnas.1400335111
- Milad, M. R., & Quirk, G. J. (2012). Fear extinction as a model for translational neuroscience: Ten years of progress. Annual Review of Psychology, 63, 129–151. https://doi.org/10.1146/annurev.psych.121208.131631
- Öhman, A., & Mineka, S. (2001). Fears, phobias, and preparedness: Toward an evolved module of fear and fear learning. Psychological Review, 108(3), 483–522. https://doi.org/10.1037/0033-295X.108.3.483
- Pavlov, I. P. (1927). Conditioned reflexes: An investigation of the physiological activity of the cerebral cortex (G. V. Anrep, Trans.). Oxford University Press.
- Rescorla, R. A. (1967). Pavlovian conditioning and its proper control procedures. Psychological Review, 74(1), 71–80. https://doi.org/10.1037/h0024109
- Rescorla, R. A. (1988). Pavlovian conditioning: It's not what you think it is. American Psychologist, 43(3), 151–160. https://doi.org/10.1037/0003-066X.43.3.151
- Stockhorst, U., Steingrueber, H.-J., Enck, P., & Klosterhalfen, S. (2006). Pavlovian conditioning of nausea and vomiting. Autonomic Neuroscience, 129(1–2), 50–57. https://doi.org/10.1016/j.autneu.2006.07.012
- Todes, D. P. (2014). Ivan Pavlov: A Russian life in science. Oxford University Press.
- Vlaeyen, J. W. S., & Linton, S. J. (2000). Fear-avoidance and its consequences in chronic musculoskeletal pain: A state of the art. Pain, 85(3), 317–332. https://doi.org/10.1016/S0304-3959(99)00242-0
- Watson, J. B., & Rayner, R. (1920). Conditioned emotional reactions. Journal of Experimental Psychology, 3(1), 1–14. https://doi.org/10.1037/h0069608
- Wolpe, J. (1958). Psychotherapy by reciprocal inhibition. Stanford University Press.