Operant & Classical Conditioning – Detailed Lecture Notes

Lecturer: Dr. Sarah Cowie, Director of the Behaviour Laboratory, University of Auckland
- Largest operant laboratory (pigeon‐based) in NZ; also runs “LambLab” in collaboration with sheep farmers.
Venue mix-up anecdote: class scheduled in wrong room → demonstrates importance of antecedent stimuli (empty theatre vs. lecturer present) in controlling behaviour.
Guest participant: Frances, an 8-day-old lamb that was mismothered and is currently hand-fed every few hours.
- Purpose: live demo of learning principles & species-specific behaviour.

Conditioning ≈ Learning (psychological use)
Two fundamental, largely automatic learning systems:
1. Operant (Instrumental) Conditioning – learning $\text{Stimulus}_{\text{Antecedent}}\;\to\;\text{Behaviour}\;\to\;\text{Consequence}$ relations.
2. Classical (Pavlovian) Conditioning – learning $\text{Stimulus}\;\to\;\text{Stimulus}\;\to\;\text{Response}$ relations.
Reinforcer: consequence that increases future probability of behaviour.
Punisher: consequence that decreases future probability of behaviour.
Antecedent / Discriminative Stimulus ( $S^D$ ): environmental cue signalling which behaviours will produce reinforcement.
Neutral Stimulus (NS): initially meaningless stimulus.
Unconditioned Stimulus (US): biologically significant event triggering an Unconditioned Response (UR) without prior learning.
After pairings: NS → Conditioned Stimulus (CS) which elicits a Conditioned Response (CR).

Learner discovers contingency: $S^D\;:\;R\;\to\;S^R$
- Example: Lecture theatre ( $S^D$ ) → note-taking ( $R$ ) → good grades, social approval ( $S^R$ ).
Behaviour becomes under stimulus control: when $S^D$ absent, behaviour weakens.
Consequences can be positive (reinforcers) or negative (punishers).
Operant behaviour is purposive & shaped by its outcomes.

Core rule set:
- Contiguity & Contingency: CS must predict US; mere simultaneity insufficient.
- Pairing order: CS → US (forward conditioning) most effective.
RARO demonstration (student lab):
- US = Raro powder → UR = salivation.
- CS = word “Pavlov”. After repeated pairings, “Pavlov” alone elicits salivation (CR).
Conceptual role: CS acts as reliable signpost of US, triggering preparatory or consummatory responses.

They often operate simultaneously on the same behaviour stream.
Demonstrated with Frances:
- CS = bottle; through pairings with milk (US) now elicits tail-wag CR.
- Concurrent operant response = “pawing” Dr Cowie’s leg (reinforced by access to milk).
- Compatible responses → can wag tail & paw concurrently.
Incompatible responses scenario:
- Work phone example: Ringing becomes CS for anxiety (freezing/avoidance) via classical pairings with disasters.
- Operant contingency (answering → reprimands/bad news) punishes call-answering.
- Result: avoidance tendency competes with requirement to answer phone.

Key insight: Classical CRs are constrained by an animal’s biological repertoire and ecological niche.
Classroom exercise (rats):
1. CS = Lever → US = food ⇒ CR = approach/lick/gnaw lever.
2. CS = Ball bearing → US = food ⇒ CR = chase/pounce/gnaw (hunting-like).
3. CS = Live rat → US = food ⇒ CR = groom/sniff/crawl (social feeding patterns).
Take-home: Same US (food) & species can yield qualitatively different CRs depending on CS affordances & natural tendencies.

Neutral bottle → CS through repeated feedings.
CRs: tail wagging, orientation, excitement.
Operant shaping: accidental leg-paw reinforced → now voluntary request signal.
Illustrates coexistence & compatibility of Pavlovian prep responses with operant goal-directed acts.

Ex-Skinner students applied operant conditioning commercially.
Famous trained acts: chickens playing baseball/dancing, pigs depositing coins.
Observed phenomenon of “misbehavior of organisms” (1961):
- Initially precise operant performance deteriorated; species-typical actions intruded.
- Example: pigs began rooting/burying coins; raccoons rubbed & dipped coins instead of depositing.
- Termed “instinctive drift.”
Consequences: slowed task completion, loss of reinforcement deliveries → commercial/experimental problems.

Weak reinforcers / satiation – coins → food might lose value; animals become less motivated.
Stress / context change – experimental set-up may evoke avoidance, interfering with task.
Pavlovian Associations within Operant Chain (favoured):
- Coin & receptacle become CSs predicting food (US).
- Food-related URs (rooting, rubbing) emerge as CRs, conflicting with operant requirement.
- Operant + Pavlovian conflict → performance breakdown.
Species Differences Hypothesis – learning principles not universal; each species unique.
Superstition theory (to be covered in later lectures) – ruled out here.

Learning principles generally trans-species, but modulated by biology.
Attempting to “train out” instinct may be unethical & practically near-impossible (hard-wired evolutionary functions).
Choosing reinforcers: must balance potency vs. unintended Pavlovian side-effects.
Application to animal welfare:
- Sheep labelled “stupid” often tested only under high-stress yard conditions; when relaxed, show excellent memory → avoid misinterpretation of cognitive abilities.

Next lecture will present an experiment explicitly dissecting misbehavior hypotheses.
Upcoming material: superstitious behaviour, additional Pavlovian rules, generalization & discrimination, and applied behaviour analysis.

Be able to diagram both conditioning types:
- Operant: $S^D \;:\; R \rightarrow S^R/ S^P$
- Classical: $CS \rightarrow US \rightarrow UR$ then $CS \rightarrow CR$
Remember that order & predictiveness (contingency) trump mere contiguity.
Always ask: “What does this stimulus afford for this species?” when predicting CRs.
Recognize potential operant-Pavlovian interactions in everyday life (e.g.
phone anxiety, snack cues, study environments).
Ethics: reinforcement choice, animal stress, and ecological validity matter in experimental design.