4.1.2 Classical conditioning and Thorndike's connectionism

Core facts

• Behaviourism / association theories are also called S-R theories (stimulus-response). Learning = forming connections between stimuli and responses; only observable behaviour counts.
• Behaviourism studies only observable, measurable behaviour; mind, mood and thought are left out as too subjective. Learning happens through the learner's interaction with the environment (APPSC 2025). Conditioning = a systematic procedure through which new responses to stimuli are learnt (APPSC 2019).
• S-O-R: when the organism's own state (O) is placed between stimulus and response, because the same stimulus can draw different responses from different learners, the formula becomes S-O-R (R.S. Woodworth).
• Ivan Petrovich Pavlov (1849–1936), Russian physiologist; Nobel Prize 1904 for work on the physiology of digestion (not for conditioning). Book: *Conditioned Reflexes* (English 1927).
• Pavlov found conditioning by chance while studying salivation as part of digestion; he first called the dog's salivation before the food reached its mouth "psychic reflexes".
• Classical conditioning is also called Pavlovian, respondent, Type S (Skinner's term), conditioned reflex or stimulus-substitution learning (the CS becomes a substitute for the UCS and calls out a similar response). It works on involuntary responses (reflexes and emotions); operant conditioning works on voluntary behaviour. Respondent conditioning (APPSC 2019) = Pavlov's discovery; the word "respondent" is Skinner's.
• Pavlov's experiment (dog):
  • Unconditioned stimulus (UCS): food (meat powder);
  • Unconditioned response (UCR): salivation to food (natural reflex);
  • Neutral stimulus → Conditioned stimulus (CS): bell (a metronome or buzzer in some trials);
  • Conditioned response (CR): salivation to the bell alone.
  • Bell rung just before food, many times → bell alone brings saliva. Saliva was collected through a small tube (fistula) in the cheek and measured in drops.
• Key terms of classical conditioning:
  • Acquisition: the stage in which the CS-UCS link is being formed;
  • Extinction: CS given again and again without the UCS → the CR weakens and disappears;
  • Spontaneous recovery: after a rest, an extinguished CR reappears (weaker) when the CS is given again;
  • Stimulus generalisation: stimuli similar to the CS (a bell of another tone) also call out the CR;
  • Stimulus discrimination: the animal learns to respond only to the CS and not to similar stimuli;
  • Higher-order (second-order) conditioning: a well-learnt CS is used like a UCS to condition a new neutral stimulus (light → bell → saliva).
• Variables: independent variable = the number of pairings of CS and UCS (also their timing and strength); dependent variable = the strength or amount of the CR (drops of saliva).
• Best timing: forward (delayed) conditioning: CS starts just before the UCS (about half a second). Backward conditioning (UCS first) works least.
• John Broadus Watson (1878–1958), American, professor at Johns Hopkins University (Baltimore), father of behaviourism; paper "Psychology as the Behaviorist Views It" (1913, the behaviourist manifesto); book *Behaviorism* (1924). Psychology = the science of behaviour, not of consciousness; rejected introspection; strong environmentalist ("give me a dozen healthy infants…").
• Little Albert experiment (Watson and Rosalie Rayner, 1920): an 11-month-old boy, unafraid of a white rat, heard a loud noise (steel bar struck with a hammer) each time he touched it → he came to fear the rat and then other furry things (rabbit, dog, fur coat, cotton wool, a Santa Claus mask). It showed classical conditioning of emotions in humans and generalisation.
• Mary Cover Jones (1924) removed a boy's ("Peter") fear of a rabbit by pairing the rabbit with food and bringing it closer step by step (counter-conditioning); she is called the "mother of behaviour therapy".
• Edward Lee Thorndike (1874–1949), American, worked almost all his life at Teachers College, Columbia University (New York); APA president 1912; the first connectionist and father of educational psychology (book **Educational Psychology, 1903; 3 volumes 1913–14). Doctoral study Animal Intelligence (1898**; book 1911).
• Thorndike's theory has several names: trial and error learning, connectionism, S-R bond theory, bond psychology, learning by selecting and connecting.
• Thorndike's experiment: a hungry cat shut in a puzzle box (problem box) with fish outside; pulling a loop or pressing a lever opened the door. The cat made many random movements, opened the door by chance, and over trials the time taken and useless movements fell gradually until it escaped at once.
• Thorndike also tested chicks, fish and monkeys. The essence of trial and error: as trials increase, errors decrease; the strength of an S-R bond is inversely related to reaction time (a strong bond = a quick response).
• Primary (major) laws: 3
  • Law of readiness: when a person is ready to act, acting is satisfying and being stopped is annoying; when not ready, being forced is annoying (motivation comes first).
  • Law of exercise (use and disuse): connections are strengthened by practice (use) and weakened by lack of practice (disuse); also called the law of frequency ("repetition is the base of learning", APPSC 2019).
  • Law of effect: a response followed by satisfaction is strengthened; one followed by annoyance is weakened (the most important law; the root of reinforcement).
• Subordinate (secondary) laws: 5
  • Multiple response (varied reaction): the learner tries many responses until one works;
  • Set or attitude (mental set): the learner's disposition decides what is learnt;
  • Prepotency of elements (partial activity): the learner responds to the most important element of a situation and ignores the rest;
  • Response by analogy (assimilation): a new situation is met with a response used for a similar old one;
  • Associative shifting: a response is shifted from its original stimulus to a new one that has been present along with it; this is the law that corresponds to Pavlov's classical conditioning.
• *Revisions (from about 1930; Human Learning, 1931, and The Fundamentals of Learning, 1932):* the law of exercise was weakened (practice without knowing results does not help) and only the satisfying half of the law of effect was kept (punishment does not simply undo a bond; the truncated law of effect). Later laws: belongingness and spread of effect.