In Duncker's candle problem, participants fail to use the box of tacks as a shelf when it is presented full of tacks. This illustrates:

Mental set from water-jar problems

According to feature integration theory, a red X is found quickly among green Xs whatever their number, but slowly among red Os and green Xs. Why?

A single feature is detected pre-attentively in parallel, but a conjunction must be bound by serial focused attention

Colour is processed more slowly than shape

Red is a more salient colour than green

A single feature is detected pre-attentively in parallel, but a conjunction must be bound by serial focused attention

Short-term memory can hold only seven items

Judging that "Linda is a bank teller and active in the feminist movement" is more probable than "Linda is a bank teller" reflects:

The conjunction fallacy produced by the representativeness heuristic

Which of the following are features of prospect theory? Select all that apply.

Outcomes are evaluated as gains or losses relative to a reference point

Losses loom larger than equivalent gains

People always maximise expected monetary value

Outcomes are evaluated as gains or losses relative to a reference point

Losses loom larger than equivalent gains

The Dani of New Guinea, who have only two basic colour terms, remembered focal colours better than non-focal ones, like English speakers. This finding argues against:

The strong form of the Sapir–Whorf hypothesis

Piaget's view that thought precedes language

Vygotsky's concept of inner speech

Assertion (A): In Wason's selection task, most people fail to choose the card that could falsify the rule. Reason (R): People tend to seek evidence that confirms a hypothesis rather than evidence that could disconfirm it. Choose the correct option.

Both A and R are true, and R is the correct explanation of A

Both A and R are true, but R is not the correct explanation of A

Thinking: Associationist, Gestalt, Information-Processing and Feature-Integration Perspectives; Concept Formation; Reasoning; Language and Thought; Problem Solving; Decision-Making; and Metacognition

The first half of Unit 6 reads: "Theoretical perspectives on thought processes: Associationism, Gestalt, Information processing, Feature integration model. Concept formation: Rules, Types, and Strategies; Role of concepts in thinking. Types of Reasoning. Language and thought. Problem solving: Type, Strategies, and Obstacles. Decision-making: Types and models. Metacognition: Metacognitive knowledge and Metacognitive regulation." This chapter follows that list, giving the classic experiment behind each idea — Köhler's chimpanzees, Duncker's candle, Luchins's water jars, Bruner's concept-attainment strategies, Wason's selection task, the Dani colour studies, Tversky and Kahneman's heuristics — and works an expected-value calculation.

1. Theoretical perspectives on thought

Four perspectives on thought processes
Perspective	Account of thinking
Associationism	Thinking is a chain of associations governed by contiguity, similarity and frequency (from Aristotle and the British empiricists); in behaviourism Watson treated thought as implicit (subvocal) speech, and mediation theorists as covert S–R links.
Gestalt	Thinking is restructuring a problem as a whole. Wertheimer contrasted productive thinking (new insight) with reproductive thinking (applying old habits); Köhler's chimpanzee Sultan joined two sticks or stacked boxes to reach bananas in sudden insight.
Information processing	Thinking is the manipulation of symbols in a limited-capacity system, in stages; Newell and Simon's General Problem Solver searched a problem space from an initial to a goal state using heuristics.
Feature integration model	Treisman and Gelade (1980): simple features (colour, orientation) are registered pre-attentively, in parallel and automatically; binding them into objects needs focused attention, serially. Hence a single-feature target "pops out" whatever the number of distractors, whereas a conjunction target (a red X among red Os and green Xs) needs serial search, and without attention illusory conjunctions occur.

2. Concept formation and reasoning

A concept is a mental category grouping objects, events or ideas by shared properties; concepts give cognitive economy, allow inference beyond what is perceived and make communication possible. In artificial-concept experiments, concepts are defined by rules relating attributes (Bourne): conjunctive (red AND square), disjunctive (red OR square), affirmative, negative, conditional (if red, then square) and biconditional — conjunctive rules are easiest and disjunctive and biconditional hardest. Types of concept include artificial versus natural concepts. The classical view defines a concept by necessary and sufficient features; prototype theory (Rosch) says natural concepts are organised around a best example with graded membership and a basic level (chair) between superordinate (furniture) and subordinate (armchair); exemplar theory says we compare with stored instances.

Strategies: Bruner, Goodnow and Austin (A Study of Thinking, 1956) described how people select instances to test. In conservative focusing the person takes a positive instance as a focus and changes one attribute at a time — slow but safe; in focus gambling more than one attribute is changed at once — faster but risky; in simultaneous scanning all possible hypotheses are held in mind at once — too demanding for memory; in successive scanning one hypothesis is tested at a time.

Types of reasoning
Type of reasoning	Nature and classic findings
Deductive	From general premises to a necessary conclusion (syllogisms, conditionals). Modus ponens (if p then q; p; so q) and modus tollens (not q; so not p) are valid; affirming the consequent and denying the antecedent are fallacies. Errors include the atmosphere effect and belief bias. In Wason's selection task (1966) few people turn the card that could falsify the rule.
Inductive	From specific instances to a probable general conclusion; in Wason's 2-4-6 task people test only confirming instances — confirmation bias.
Analogical and abductive	Reasoning from a similar known case (Gick and Holyoak's radiation problem, 1980, was solved far more often after a military analogy, but mostly when a hint was given); abduction infers the best explanation.

3. Language and thought

Linguistic relativity (the Sapir–Whorf hypothesis): the strong form, linguistic determinism, says language determines thought; the weak form says it influences it. Whorf cited Hopi and snow vocabulary.
Evidence against the strong form: Eleanor Rosch (then Heider) found that the Dani of New Guinea, who have only two colour terms, remembered focal colours better than non-focal ones, just as English speakers do. Evidence for the weak form: Russian speakers, who have separate basic words for light and dark blue, discriminate those shades slightly faster (Winawer and colleagues, 2007).
Piaget: thought comes first, and language reflects the level of cognitive development. Vygotsky: thought and language have separate roots and merge at about two years; egocentric (private) speech becomes inner speech, a tool of thought. Chomsky: language is a specialised faculty (the language acquisition device), largely independent of general cognition.

4. Problem solving and decision-making

Problems are well-defined (clear start, goal and operators — a chess puzzle) or ill-defined (write a good essay). Greeno classified them as problems of arrangement (anagrams), inducing structure (series completion, analogies) and transformation (the Tower of Hanoi, hobbits and orcs). Strategies: an algorithm guarantees a solution but may be slow; heuristics are shortcuts — means–end analysis (reduce the difference between the current state and the goal by setting subgoals), working backward, hill climbing, and analogy. Obstacles: mental set or Einstellung (Luchins's water-jar problems, 1942: people kept using a long formula B − A − 2C when a simpler one worked), functional fixedness (Duncker's candle problem, 1945: people failed to see the tack box as a shelf; Maier's two-string problem), unnecessary self-imposed constraints (the nine-dot problem), irrelevant information and confirmation bias.

Decisions are made under certainty, risk (known probabilities) or uncertainty (unknown probabilities). Normative models prescribe how to choose: expected-utility theory (von Neumann and Morgenstern) chooses the option with the highest probability-weighted value. Worked: a 50 per cent chance of ₹200 has an expected value of 0.5 × 200 = ₹100, more than a sure ₹90 — yet most people take the sure ₹90. Descriptive models describe how people actually choose: Herbert Simon's bounded rationality and satisficing (take the first option that is good enough); Tversky's elimination by aspects (1972), a non-compensatory model that discards options failing on successive attributes, versus compensatory (additive, multi-attribute) models in which strengths offset weaknesses; and Kahneman and Tversky's prospect theory (1979) — outcomes are judged as gains or losses from a reference point, the value function is concave for gains and convex for losses, losses loom about twice as large as gains (loss aversion), and small probabilities are overweighted.

Heuristics and biases in judgement (Tversky and Kahneman, 1974)
Heuristic or effect	What happens
Representativeness	Judging probability by resemblance to a prototype, neglecting base rates; the conjunction fallacy — "Linda is a bank teller and a feminist" judged more likely than "Linda is a bank teller".
Availability	Judging frequency by how easily examples come to mind — overestimating deaths from vivid causes such as plane crashes.
Anchoring and adjustment	Estimates stay too close to an initial value, even an arbitrary one.
Framing	The same outcomes described as lives saved or lives lost reverse preferences (the "Asian disease" problem, 1981): people are risk-averse for gains and risk-seeking for losses.

5. Metacognition

Metacognition — "thinking about thinking" — was introduced by John Flavell (1976, 1979). It has two components. Metacognitive knowledge is what one knows about cognition, in three kinds: person variables (knowledge of oneself and others as thinkers — "I remember faces better than names"), task variables (knowledge of how the nature of a task affects performance — recall is harder than recognition) and strategy variables (knowledge of which strategies work — rehearsal, elaboration, self-testing). Metacognitive regulation is the control of cognition: planning (choosing strategies and allocating time), monitoring (checking comprehension and progress — judgements of learning, the feeling of knowing) and evaluating (appraising outcomes and adjusting).

🎯 Why metacognition matters

Poor monitoring produces the illusion of competence — rereading notes feels like learning because the material is familiar, while self-testing reveals what is actually retrievable. The least skilled often overestimate their performance most (the Dunning–Kruger effect), because the skill to perform and the skill to judge performance overlap.

Key takeaways

Associationism chains ideas, Gestalt stresses insight and restructuring (Köhler, Wertheimer), information processing models thought as symbol manipulation in a problem space, and feature integration theory separates parallel feature registration from attentive binding.
Concepts follow rules (conjunctive easiest, disjunctive harder), are organised around prototypes and a basic level, and are attained by conservative focusing, focus gambling, simultaneous or successive scanning.
Deduction is prone to belief bias and to neglect of falsifying cases (Wason); linguistic relativity holds only in its weak form; Piaget, Vygotsky and Chomsky give three views of language and thought.
Problems are solved by algorithms and heuristics such as means–end analysis, and blocked by mental set (Luchins) and functional fixedness (Duncker).
Decision models are normative (expected utility) or descriptive (satisficing, elimination by aspects, prospect theory with loss aversion and framing); metacognition combines knowledge of person, task and strategy with planning, monitoring and evaluation.

Practice questions (10)

Attempt each one before opening the answer. Every explanation names the tempting wrong option as well as the right one, because that is where marks are lost.

In Duncker's candle problem, participants fail to use the box of tacks as a shelf when it is presented full of tacks. This illustrates:
1. The availability heuristic
2. Mental set from water-jar problems
3. Belief bias
4. Functional fixedness
Show answer
Answer: D — Functional fixedness
Seeing the box only in its usual function — a container — blocks seeing it as a platform; presenting it empty greatly improves solution rates. Mental set is the repetition of a previously successful procedure (Luchins).
According to feature integration theory, a red X is found quickly among green Xs whatever their number, but slowly among red Os and green Xs. Why?
1. Colour is processed more slowly than shape
2. Red is a more salient colour than green
3. A single feature is detected pre-attentively in parallel, but a conjunction must be bound by serial focused attention
4. Short-term memory can hold only seven items
Show answer
Answer: C — A single feature is detected pre-attentively in parallel, but a conjunction must be bound by serial focused attention
Treisman and Gelade showed that feature targets "pop out" with flat search times, while conjunction targets require attention to bind colour and shape at each location, so search time rises with the number of items.
In Bruner's concept-attainment studies, a person who takes a positive instance as a focus and changes only one attribute at a time is using:
1. Successive scanning
2. Conservative focusing
3. Focus gambling
4. Simultaneous scanning
Show answer
Answer: B — Conservative focusing
Conservative focusing changes one attribute per choice, so each outcome is informative and memory load is low. Focus gambling changes several attributes at once, and scanning strategies test hypotheses rather than varying a focus card.
A gamble offers a 50 per cent chance of winning ₹200 and a 50 per cent chance of winning nothing. What is its expected value in rupees? Type a number.
Numerical answer — type the value.
Show answer
Answer: 100
EV = (0.5 × 200) + (0.5 × 0) = 100. Expected-value theory says it should be preferred to a sure ₹90, yet most people choose the sure amount — risk aversion for gains, as prospect theory predicts.
Judging that "Linda is a bank teller and active in the feminist movement" is more probable than "Linda is a bank teller" reflects:
1. Functional fixedness
2. Anchoring and adjustment
3. Loss aversion
4. The conjunction fallacy produced by the representativeness heuristic
Show answer
Answer: D — The conjunction fallacy produced by the representativeness heuristic
A conjunction can never be more probable than one of its parts, but the description resembles a feminist, so representativeness overrides probability. Anchoring concerns estimates near a starting value, loss aversion the extra weight of losses.
Which of the following are features of prospect theory? Select all that apply.
1. People always maximise expected monetary value
2. The value function is concave for gains and convex for losses
3. Outcomes are evaluated as gains or losses relative to a reference point
4. Losses loom larger than equivalent gains
Show answer
Answer: B — The value function is concave for gains and convex for losses; C — Outcomes are evaluated as gains or losses relative to a reference point; D — Losses loom larger than equivalent gains
Kahneman and Tversky's prospect theory has a reference point, loss aversion and an S-shaped value function. It was proposed precisely because people do not consistently maximise expected value.
The Dani of New Guinea, who have only two basic colour terms, remembered focal colours better than non-focal ones, like English speakers. This finding argues against:
1. The strong form of the Sapir–Whorf hypothesis
2. Piaget's view that thought precedes language
3. Vygotsky's concept of inner speech
4. Prototype theory
Show answer
Answer: A — The strong form of the Sapir–Whorf hypothesis
If language determined thought, speakers without terms for focal colours should not favour them. Rosch's result instead supports universal perceptual prototypes — consistent with prototype theory and with Piaget, not against them.
Assertion (A): In Wason's selection task, most people fail to choose the card that could falsify the rule. Reason (R): People tend to seek evidence that confirms a hypothesis rather than evidence that could disconfirm it. Choose the correct option.
1. Both A and R are true, and R is the correct explanation of A
2. Both A and R are true, but R is not the correct explanation of A
3. A is true, but R is false
4. A is false, but R is true
Show answer
Answer: A — Both A and R are true, and R is the correct explanation of A
Both are true and R explains A: confirmation bias leads people to turn the cards that could agree with the rule (p and q) and to neglect not-q, the card that tests the rule by modus tollens.
Which of the following are components of METACOGNITIVE REGULATION rather than metacognitive knowledge? Select all that apply.
1. Knowledge of task variables
2. Planning
3. Evaluating
4. Monitoring
Show answer
Answer: B — Planning; C — Evaluating; D — Monitoring
Regulation is the control of cognition: planning, monitoring and evaluating. Knowledge of person, task and strategy variables is metacognitive knowledge in Flavell's scheme.
Choosing the first job offer that meets one's minimum requirements, rather than searching for the best possible one, is what Herbert Simon called:
1. Hill climbing
2. Framing
3. Satisficing
4. Anchoring
Show answer
Answer: C — Satisficing
Satisficing — accepting a good-enough option — follows from bounded rationality: limited time and cognitive capacity make exhaustive optimisation impractical. Hill climbing is a problem-solving heuristic.