Lesson 18: Classical Conditioning

(a) Hook

A dog hears a tone, and its mouth fills with saliva. There's no food in sight — just a sound. That's it. That's the experiment that launched the most influential research program in the history of learning, and it happened by accident.

Ivan Pavlov wasn't even studying learning. He was a physiologist measuring how much dogs salivate when they're fed — he won a Nobel Prize for digestion work. But he kept noticing something annoying: his dogs started drooling before the food arrived. At the sound of the lab assistant's footsteps. At the sight of the food bowl. His careful measurements were being "contaminated" by anticipation.

Most scientists would have called that noise and engineered it away. Pavlov turned around and studied the noise itself — and discovered that a meaningless stimulus, paired enough times with food, can come to trigger a biological reflex all on its own. Your brain does this constantly. The smell of a dentist's office, the buzz of your phone, the jingle of an ad — your body reacts to all of them, and not one of those reactions was there at birth. You learned them, without ever deciding to.

(b) Core Concepts

Learning by association

Learning is a relatively permanent change in behavior (or behavior potential) due to experience. Classical conditioning is one of the most basic kinds: you learn to associate two stimuli so that one comes to trigger a response that originally belonged to the other. It's a form of associative learning — your brain links events that occur together and uses one to predict the other.

The genius of Pavlov's setup is that it has only four moving parts, and the entire topic is just learning to name those four parts in any situation. Get fluent here and the rest of the lesson is bookkeeping.

The four parts (memorize the order they appear)

Start with a reflex that already exists — something the body does automatically, with no learning required.

• Unconditioned stimulus (UCS or US): a stimulus that naturally and automatically triggers a response, no learning needed. Food in the mouth. (Pavlov's: meat powder.)
• Unconditioned response (UCR or UR): the automatic, unlearned reaction to the UCS. Salivation to food. You were born able to do this.

The word unconditioned means unlearned. The UCS→UCR link is wired in. Nobody taught the dog to drool at food.

Now bring in something neutral.

• Neutral stimulus (NS): a stimulus that, before conditioning, produces no relevant response. A tone, by itself, doesn't make a dog salivate. (It might make the dog turn its head — an orienting reflex — but it doesn't trigger salivation.)

Then you pair them. Repeatedly, you ring the tone and then present the food. Tone → food. Tone → food. Tone → food. After enough pairings, the formerly neutral tone takes on power it never had.

• Conditioned stimulus (CS): the once-neutral stimulus that, after being paired with the UCS, now triggers a response on its own. The tone has been promoted: NS → CS.
• Conditioned response (CR): the learned response to the conditioned stimulus. Salivation to the tone alone.

The word conditioned means learned. The CS→CR link is built through experience.

Here's the part students miss: the UCR and the CR are often the same behavior — salivation — but they're triggered by different things and earn different names. Salivation to food is the UCR (automatic). Salivation to the tone is the CR (learned). Same drool, different cause, different label.

Try This — the 30-second identification drill. For any scenario, ask three questions in order. (1) What response happens automatically, with no learning? That response is the UCR, and whatever triggers it is the UCS. (2) What started out neutral but got paired with that UCS? That's the CS (it was the NS before pairing). (3) What's the learned response to that CS? That's the CR. Do this on every example below and you will never blank on an exam scenario.

Acquisition and timing

Acquisition is the initial phase of learning — the period during which you repeatedly pair the NS with the UCS and the CR gradually emerges and strengthens. On a graph, acquisition is the rising curve: the CR gets stronger (more saliva, faster onset) with each pairing, climbing steeply at first and then leveling off.

Timing matters enormously. Conditioning works best when the NS comes just before the UCS — typically by about half a second — and signals that the UCS is coming. This is forward conditioning, and it makes adaptive sense: a stimulus is useful only if it predicts something. If the tone came after the food (backward conditioning), the dog learns almost nothing — the tone predicts nothing it doesn't already know.

Extinction and spontaneous recovery

What happens if you stop pairing? Ring the tone, and tone, and tone — but never bring food. The CR weakens and eventually disappears. This is extinction: the diminishing of a conditioned response when the CS is repeatedly presented without the UCS. On the graph, extinction is the curve falling back down toward zero.

But the learning isn't erased — it's suppressed. Give the dog a rest, come back the next day, ring the tone once, and the CR partly returns on its own. This is spontaneous recovery: the reappearance, after a pause, of a weakened conditioned response. It's weaker than before and fades fast if pairing still doesn't resume, but its very existence proves extinction didn't delete the association. The brain kept the memory; it just stopped acting on it.

Generalization and discrimination

After conditioning to one tone, dogs salivate to similar tones too — a slightly higher or lower pitch still gets some drool. This is generalization: the tendency for stimuli similar to the CS to trigger the CR. The more similar the new stimulus, the stronger the response. (Generalization is adaptive — a creature that learned to fear one specific rustling sound but no other would be poorly protected.)

Its opposite is discrimination: learning to respond to the CS but not to similar stimuli. If you pair only the 1000-Hz tone with food and present a 1500-Hz tone without food, the dog learns to tell them apart and salivates only to the trained tone. Discrimination is generalization's mirror: one is responding across similar stimuli, the other is learning to respond narrowly to just one.

Higher-order conditioning

Once a CS reliably triggers a CR, it can act like a UCS for new learning. Pair a flashing light with the already-established tone (no food at all), and eventually the light alone triggers salivation. This is higher-order conditioning (or second-order conditioning): a neutral stimulus paired with an existing CS becomes a CS itself. The response is usually weaker, but it shows associations can chain — which is exactly how an abstract logo, never paired with anything biological, can still come to give you a feeling.

Applications your brain runs every day

Classical conditioning isn't a dog-and-bell party trick; it's running in you right now.

• Advertising pairs a product (neutral) with images that already trigger good feelings — attractive people, upbeat music, happiness (UCS → UCR). After enough pairings, the product itself (CS) triggers the warm feeling (CR).
• Drug tolerance is partly conditioned. The environment where someone usually takes a drug (CS) comes to trigger the body's compensatory responses that counteract the drug (CR), so more drug is needed for the same effect. Take the same dose in an unfamiliar setting — no conditioned cues — and the protective response doesn't fire, which is one reason overdoses spike in novel environments.
• Taste aversion (previewed here, detailed in Lesson 21): one pairing of a food with nausea can produce lifelong avoidance of that food — even if the nausea was a coincidence.

That last one points to a limit. Classical conditioning is not infinitely flexible. Biological constraints (also called preparedness) mean organisms are biologically predisposed to learn some associations far more easily than others — taste with nausea, for instance, but not taste with electric shock. Evolution pre-tuned which links our brains form readily. We'll unpack this in Lesson 21, but plant the flag now: not everything that gets paired gets learned equally.

(c) Classic Studies Spotlight

Watson & Rayner — "Little Albert" (1920).

Who & when: John B. Watson and his graduate assistant Rosalie Rayner, Johns Hopkins University, 1920.

What they did: Watson — the founder of behaviorism — wanted to show that even emotions are conditioned, not innate. The subject was an infant known as "Little Albert" (about 11 months old). At baseline, Albert showed no fear of a white rat; he reached for it happily (the rat was a neutral stimulus). Then, each time Albert touched the rat, Watson struck a steel bar behind his head with a hammer, making a loud, startling clang (the UCS) that made Albert cry and recoil in fear (the UCR). After several pairings, the rat alone made Albert cry, flinch, and crawl away (the rat was now a CS; the fear was a CR).

What they found: Albert's conditioned fear generalized — he became upset by other white furry things: a rabbit, a dog, a fur coat, even a Santa Claus mask with a white beard. A specific, learned fear had spread to similar stimuli.

Why it matters: This demonstrated that complex emotional reactions — including phobias — can be classically conditioned, supporting Watson's claim that environment shapes behavior. It's also a landmark in research ethics: there was no informed consent, Albert was never deconditioned, and the study would be flatly prohibited by today's IRB standards. For the exam: Watson & Rayner = conditioned fear + generalization + an ethics cautionary tale.

(d) Application Practice

For each scenario, label the UCS, UCR, CS, and CR. (Answers and reasoning follow each one — but cover them and try first.)

Scenario 1 — The thunderstorm puppy. A puppy is calm until a violent thunderstorm hits: a deafening clap of thunder makes it tremble and hide. For weeks afterward, the bright flash of lightning — even with no thunder yet — makes the puppy tremble and hide.

Label it. UCS = the loud thunder (naturally triggers fear). UCR = trembling/hiding in response to the thunder. CS = the lightning flash (was neutral; now predicts thunder after being paired with it). CR = trembling/hiding in response to the lightning flash alone. The puppy learned that lightning predicts thunder — classic forward conditioning.

Scenario 2 — Chemo and the waiting room. A cancer patient receives chemotherapy that reliably causes nausea. After several rounds, she starts feeling nauseated the moment she walks into the clinic waiting room, before any drug is given.

Label it. UCS = the chemotherapy drug (naturally causes nausea). UCR = nausea caused by the drug. CS = the waiting room (neutral until paired with the drug-induced nausea). CR = nausea triggered by the waiting room. Note the UCR and CR are the same feeling — nausea — but named by their trigger. This is also why oncology clinics rotate décor: to fight conditioned nausea.

Scenario 3 — The energy-drink ad. A new energy drink is advertised with footage of cheering crowds and a thrilling last-second sports victory that gives viewers a jolt of excitement. After seeing the ad many times, people feel a small surge of excitement just seeing the can on a shelf.

Label it. UCS = the thrilling victory footage (naturally triggers excitement). UCR = excitement from the footage. CS = the energy-drink can (neutral product, paired with the exciting footage). CR = excitement triggered by the can. This is advertising as classical conditioning, full stop — and if the brand's logo later triggers excitement on its own (paired only with the can, never with the original footage), that's higher-order conditioning.

(e) Traps & Confusions

UCS vs. CS. The single most-tested confusion. The UCS triggers the response automatically, with no learning (food, loud noise, the drug). The CS triggers the response only because it was paired with the UCS (tone, rat, waiting room). Test tip: ask "would this trigger the response in a brand-new organism that had never been conditioned?" If yes → UCS. If it only works after training → CS.

Extinction vs. forgetting. Both make a CR disappear, but they're different mechanisms. Extinction is active new learning from repeatedly presenting the CS without the UCS — the brain learns the CS no longer predicts anything. Forgetting is passive fading from disuse over time. The giveaway that it's extinction, not forgetting: spontaneous recovery. If the response pops back after a rest, it was extinguished (suppressed), not forgotten (erased).

Generalization vs. discrimination. Mirror images. Generalization = responding to stimuli similar to the CS (widening). Discrimination = learning to respond only to the exact CS and not to similar ones (narrowing). Mnemonic: Generalize = broaden; discriminate = "be picky."

Classical vs. operant (preview of L19). Classical conditioning is about involuntary, reflexive responses (salivation, fear, nausea) triggered by stimuli that come before the response — the organism is fairly passive. Operant conditioning (next lesson) is about voluntary behaviors changed by consequences that come after the behavior — the organism acts on the environment. Quick test: is the behavior an automatic reflex pulled out by a signal (classical), or an action shaped by reward/punishment (operant)?

(f) Practice Problems

Four-choice MCQs in current AP format. Answers and explanations in section (h).

(g) FRQ Practice — Article Analysis Question (AAQ)

Respond to all six parts (A–F) in complete sentences using appropriate psychological terminology. Part F is worth 2 points; all others 1 point each (7 total).

Stimulus — summarized study

Introduction. Researchers tested whether a previously neutral sound could be classically conditioned to trigger a measurable fear response in adults, and whether that conditioned response would weaken when the sound was presented alone.

Participants. 60 adult volunteers (ages 18–45; 52% women, 48% men) were recruited from a university community and compensated with a small gift card. Before participating, each person read and signed a form describing the procedure, including that they would hear a tone and occasionally receive a mild, harmless puff of air to the eye, and each was told they could withdraw at any time without losing compensation. Participants were identified in the dataset only by a randomly assigned ID number.

Method. In the acquisition phase, a 2-second tone (initially neutral) was followed immediately by a harmless puff of air to the eye, which naturally produces an eyeblink. Each participant received 20 such tone-then-puff pairings. The researchers measured the conditioned eyeblink rate — the percentage of trials on which the participant blinked to the tone alone, in the brief window before any air puff could occur. In a later extinction phase, the tone was presented 20 times with no air puff, and the conditioned eyeblink rate was measured again.

Results. During acquisition, the mean conditioned eyeblink rate rose from 8% on the first pairing to 71% by the twentieth pairing. During extinction (tone alone), the mean conditioned eyeblink rate fell from 68% on the first extinction trial to 12% by the twentieth extinction trial.

A. Identify the research method used in this study.

B. State the operational definition of the dependent variable (the conditioned eyeblink rate) as used in this study.

C. Describe what the results indicate about the conditioned eyeblink rate from the first to the twentieth acquisition trial. (Cite the numbers.)

D. Identify at least one ethical guideline the researchers applied, and describe how they applied it.

E. Explain the extent to which the findings are or are not generalizable, using specific evidence from the study.

F. Explain how the change in the conditioned eyeblink rate during the extinction phase supports or refutes the claim that a classically conditioned response weakens when the conditioned stimulus is presented without the unconditioned stimulus. (Apply a psychological concept.)

Model answer (earns 7/7)

A. The research method was an experiment, because the researchers manipulated a variable (the pairing of the tone with the air puff, then the removal of the puff) and measured its effect on the participants' responses under controlled conditions. (1 pt) [A: method]

B. The conditioned eyeblink rate was operationally defined as the percentage of trials on which a participant blinked in response to the tone alone, measured in the brief window before any air puff could occur. (1 pt) [B: operational definition of DV]

C. The results indicate that the conditioned eyeblink rate increased substantially across acquisition, rising from 8% on the first pairing to 71% by the twentieth pairing — a gain of 63 percentage points — which shows that the conditioned response was acquired and strengthened as the tone was repeatedly paired with the air puff. (1 pt — note the cited numbers) [C: results with data]

D. The researchers applied informed consent: before participating, each person read and signed a form describing the procedure (including the tone and the mild air puff) and was told they could withdraw at any time without penalty. (Confidentiality would also earn the point: participants were identified only by a randomly assigned ID number.) (1 pt) [D: ethics]

E. The findings have limited generalizability because the sample was drawn from a single university community, so the results may not extend to children, older adults, or non-university populations; however, the sample did span a moderately wide adult age range (18–45) and was fairly balanced by sex (52% women, 48% men), which supports somewhat broader applicability within the adult population. (1 pt — one clear direction, backed by study evidence) [E: generalizability with evidence]

F. The change during extinction supports the claim. The conditioned eyeblink rate fell from 68% to 12% once the tone (the conditioned stimulus) was presented repeatedly without the air puff (the unconditioned stimulus). This is the definition of extinction: when the CS no longer predicts the UCS, the conditioned response weakens because the learned association is no longer reinforced by the pairing. The steep decline shows the participants' brains updated to treat the tone as no longer signaling the puff. (2 pts — states support/refute AND applies a genuine psychological concept: extinction) [F: argument + concept, 2 pts]

Where students commonly lose points

• Part A: writing "lab study" or "random assignment" instead of naming the method experiment. (And note: this study has no separate control group; the within-subjects manipulation is still an experiment.)
• Part C: describing the direction of the change ("it went up") but omitting the actual numbers (8% → 71%) — an automatic miss on a data part.
• Part D: naming an ethical guideline that the study didn't actually use, or naming "informed consent" without describing how it was applied (the signed form / withdrawal statement).
• Part E: hedging in both directions ("it's generalizable but also not") with no commitment, or asserting (non)generalizability with no evidence from the participant description.
• Part F: saying "it supports the claim" but not naming and using the concept of extinction, or mislabeling the parts — e.g., calling the tone the UCS. The 2 points require both a clear support/refute stance and correct application of a real conditioning concept.