Thinking and Problem Solving

Modified: 2026-04-12 1:17 PM CDST

Newell, Simon, and Shaw were pioneers who created many avenues of research in cognitive psychology. Tversky and Kahneman gave rise to heuristics and prospect theory.

Chapter 12: Thinking and Problem Solving: Beyond Newell, Shaw, and Simon's (1958) Theory of Human Problem Solving--Fernand Gobet and Peter Lane

• Background to the Classic Study
  • People and Issues
    • Logic Theorist
      • Geometry first
      • Chess followed (too difficult)
      • Elementary Symbolic Logic chosen
      • Psychological Review article first on directed at psychology
        • APA (abstract)
      • Pre-AI
      • Not directed to human problem solving at first
    • Allen Newell
    • Herbert A. Simon
    • John C. (Cliff) Shaw
  • Dominance of Behaviorism in USA
  • Gestalt refugees also interested in cognition
• Detailed Description of the Classic Study
  • The Theory
    • Similar to a computer
    • Only works for human cognition (not animals)
      • Uses
      • Testing specific predictions
      • Compare human behavior in separate situations
      • Breaks problem solving into elementary processes
    • Explains behavior by generating behavior
    • Does not explain physical level, instead explaind functional level (between Logic Theorist and human behavior)
  • The Logic Theorist
    • Found proofs in elementary logic
    • "Learns"
    • "Memorizes"
    • Has four rules of inference
      • Subsitution
      • Replacement
      • Detachment
      • Syllogism
    • Has Executive Process (see p. 173)
    • Not developed as theory of human problem-solving
  • The Experiments
    • Experiments? Yes, but no animal subjects or human participants
      • Experiment 1
        • Task: Prove the 52 theorems in Principia Mathematica
          • Results: Proved 38/52
        • Large search space
        • Logic theorist used heuristic search
        • Some proofs shorter and more elegant than previous one done by humans
      • Experiment 2
        • Attempt to solve proof 2.12 without first solving previous ones failed
        • Showed that Logic Theorist was sensitive to order of proof solving
      • Experiment 3
        • Hint (Theorem 2.03) provided, then Logic Theorist proves Theorem 2.12
  • Characteristics of the Problem Solving Process
    • Set Phenomena
      • Definition: readiness to make a specific response to a specified stimulus
      • Luchins (1942) is good example from human learning
        • Try link above for yourself. See if you fall victim to set effects
      • All problem solvers will examine possible solutions in non-random order
    • Insight Learning
    • Concept Learning
      • Logic Theorist does not use concepts but shows that concepts facilitate search
    • Hierarchies of Processes
      • Essential to Logic Theorist
      • Logic Theorist breaks problems down and builds complex hierarchies
    • Learning
      • Proved theorems are stored in computer memory
  • Comparison with Other Theories
    • Associationism
      • Elementary mental processes provide mechanistic explanation for behavior
      • Note: Associationism has long history in philosophy but think of Pavlov and his classical conditioning as the first intance of a mechanistic explanation
    • Gestalt Theories
      • Logic Theorist used Otto Selz, Selz, and Adrian deGroot mostly
• Impact of the Classic Study
  • Large Impact on:
    • "Cognitive Revolution"
    • Information Processing
    • Computer experiments
    • Operators (logical inference rules)
      • Initial State
      • Goal State
      • Problem Space
    • One of the early AI programs
    • Computer modeling
• Critique of the Classic Study
  • Weaknesses
    • Theory-Mechanism Gap
    • Few simulations presented
    • Weak methodology (by modern standards)
    • Logic Theorist made wrong predictions
    • Led to General Problem Solver program
• Alternative Interpretations
  • Critics
    • Hubert Dreyfus
    • John Searle
• Problem Solving vs Decision Making
  • Problem Solving
    • Choose between potential solutions
  • Decision Making
    • Participants must generate solutions
• Conclusions
  • Created:
    • Information processing psychology
    • AI
    • Computer modeling (in psychology
  • Renewed interest in problem solving
  • Computational neuroscience will be big area for psychology in the future.

Chapter 13: Thinking and Problem Solving: Beyond Tversky and Kahneman's (1974) Judgement Under Uncertainty--Klaus Fiedler and Momme von Sydow

• Background to the Classic Study
  • Daniel Kahneman
  • Amos Tversky
  • Availability Heuristic
  • Representativeness Heuristic
  • Anchoring and Adjustment
  • Over 45,000 citations
  • Statistical models related to rationality changed older views such as those from psychophysics
• Detailed Description of the Classic Study
  • Heuristics can be economical and effective
    • Kahneman and Tversky, however, chose to examine the systematic and predictable errors heuristics sometimes made
• Synopsis of Most Prominent Heuristics
  • Representativeness
    • Man wearing tweed jacket, smoking a pipe, good with words must be a college professor
      • True story: In Milwaukee, in the winter I went to an indoor (obviously) carwash. I had a 1977 Honda Accord with Louisiana plates (54 K 134), a toolbox visible in the back, and I was wearing workboots, flannel shirt and painter's pants.
      • The men drying my car asked me, "Are you a construction worker." I said, no and that I was a college professor.
    • Stereotyping is associated with representativeness
  • Availability
    • Ask the following workers in the following fields the questions below.
      • Police: How violent is Texarkana?
        • Will perceive Texarkana is more violent. Why?
      • MDs: How sick is Texarkana?
        • Will perceive Texarkana is sicker. Why?
      • The public: What is more dangerous to life: sharks or honeybees?
        • Will say sharks. Why?
  • Anchoring and Adjustment
    • Akin to first impressions.
      • If I were to interview for an academic job I would wear a suit and tie, a white longsleeve shirt, socks, and closed toe shoes. That's the anchor.
        • Just so you know, I only own one suit. I wear sandals nearly always (never with socks). I wear shorts year round (because I'm no longer in Milwaukee). I only own four longsleeve shirts and most of the shirts I wear are brightly colored and patterned. If I were to get the job, that would be the adjustment.
      • But, if I were going for a promotion at Southern Arkansas University (my main gig), everyone already knows my REAL anchor (see above). I'd be laughed out of the interview if I were to put on my only suit, socks, shoes, and tie.
• Impact of the Classic Study
  • Heuristics and Biases in Cognitive, Social, and Applied Psychology
    • Lot of influence
    • Has leaked over into medicine, journalism, and politics
• Critique of the Classic Study
  • Kahneman and Tversky's heuristics were widely supported by many
  • However, much research has pointed to:
    • Vagueness
    • Lack of precise and falsifiable models
    • More precise definition of heuristics in general
    • Ease of recall and media content
    • Egocentrism
    • Lack of evidence for gradual adjustment after anchoring
    • Competing opposite models for anchoring and adjustment
  • Heuristics in the Post-Kahneman-Tversky Era
    • The Adaptive Toolbox
      • Post-Kahneman and Tversky
        • New Heuristics:
          • Take-the-best
          • Recognition
          • Fluency
          • Lexicographic
          • Priority
        • However, evidence for the above heuristics tends to be weak
    • Dual-Process Approaches
      • Are common in psychology
      • Here, they attempt to reconcile two fundamentally different modes of information processing
        • One:
          • Uses heuristic shortcuts
          • Uses intuitive strategies
        • The Other
          • Uses systematic and exhaustive strategies
          • Produces normative results under many conditions
      • But, many problems remain
        • One:
          • Uses automatic association
          • Uses heuristics
          • Uses intuition
          • Has few constraints on mental capacity
          • Requires little effort
          • Is mostly unconscious
          • Possesses no meta-cognitive control
        • The other:
          • Uses reflective propositional operations
          • Is rule based
          • Is exhaustive
          • Requires high cognitive capacity
          • Requires effort
          • Possesses high levels of meta-cognitive control
        • Furthermore:
        • Hard to assume these divergent methods are highly correlated positively
          • Even "the other" often produces biases
• Conclusions
  • Appraisal of 40 Years of Research on Heuristics and Biases
    • Two very different conclusions: one disillusioning, the other enthusiastic
      • Theory Development
        • Little theoretical precision, refinement and progress
      • Representativeness
        • Little knowledge of exemplars
        • Little "sober cognitive research" (p. 182)
        • But, signal detection and lexicographic algorithms have provided accurate judgments and decisions in many fields
    • Here is the enthusiastic part
      • It is not that Kahneman and Tversky's heuristics were always right. It was more that they served as inspiration for more research stemming from their original work.
      • Look at the conjunction fallacy, for example.
    • Linda is 31 years old, single, outspoken, and very bright. She majored in philosophy. As a student, she was deeply concerned with issues of discrimination and social justice, and also participated in anti-nuclear demonstrations.
      • Which is more probable?
        • a) Linda is a bank teller.
        • b) Linda is a bank teller and is active in the feminist movement.
      • The correct answer, according to the laws classical probability theory, is a). Nevertheless, answer b) seems intuitively more reasonable, and indeed experiments have demonstrated that many people choose b) over a).
    • Your text suggests that different views of probabilistic statements may be necessary
    • Anchoring heuristic has led to progress in many related areas (law, auction prices, and social comparisons)
    • Meta-cognition, too, has left the availability heuristic behind.
      • FYI, meta-cognition, in one sense, means realizing what you know. For example, the capital of Texas. I'll bet you do not know the capital of Burkina Faso. (and, if you don't know it then you know that you don't know it)
  • Kahneman and Tversky's research led to progress in:
    • Economics
    • Consulting by behavioral scientists
  • Their research has led to new views about bias and its origins.
• Concluding Remark
  • Kahneman and Tversky's lack of confirmatory evidence, failure to explain or to provide methods of falsification are not the main point.
  • Rather, the point is that they created a new area of cognitive psychology that remains to be fully explored and understood.
• Afterthought: Comments from Daniel Kahneman
  • Both were well trained in statistics
  • They wondered why they, themselves, often made jusdgmental errors:
    • Underestimating effect of increased sample size
    • Drawing exaggerated inferences from one datum
  • If they could make such errors, they assumed people untrainded in statistics could do so as well
  • They did not believe that people were stupid.
  • They also wanted to write in the style of the Gestalt psychologists and pose riddles in their articles for readers to figure out.
  • They were happy that their research flowed to areas other than psychology

Chapter 14: Thinking and Problem Solving: Beyond Tversky and Kahneman's (1979) Prospect Theory--Ben R. Newell

• Background to the Classic Study
  • Daniel Kahneman
    • Nobel Prize Biography (fascinating reading, fyi)
  • Amos Tversky
  • Prospect Theory
  • Defining a Decision Under Risk
    • Nik Wallenda on Grand Canyon Gorge (video)
      • Risk: Uncertainty about gain or loss from a choice
    • Blaise Pascal and Risk Assessment
    • Pascal and Fermat's Correspondence
    • Expected Utility
    • From 1950s on, Expected Utility Theory was dominant theory in the analysis of decision making under risk
• Details of the Classic Study
  • What is Prospect Theory
    • Prospect Theory was a challenge to Expected Utility Theory
      • Started with 14 problems designed to show flaws in Expected Utility Theory
      • Gains
        • Problem 1

        A: (4,000, .8) or B: (3000, 1.0)

        N = 95 [.20] -----------[80]*

        Certainty effect: In other words, take the sure thing.

        • Problem 2

        C: (4,000), .20) or D: (3,000), .25)

        N= 95 [65]* ------------[35]

        Problem 2 is a restatement of Problem 1 but probabilities are divided by 4
        Option C has a relatively larger reward

      • Losses
        • Problem 3

      A: (-4,000, .80) or B: (-3,000, 1.0)

      N=95 [92]*---------------[8]

      • Problem 4

      C: (-4,000, .20) or D: ((-3,000, .25)

      N=95 [42]----------------[56]

      Sure gain in Problem 2 is replaced by risky loss of 4,000
      Reflection Effect: Risk aversion in for gains, risk seeking for losses

      Problem 5 You have been given 1,000

      A: (1,000), .50 or B: (500, .1.0)

      N=70 [16]---------------[84]*

      Problem 6 You have been given 2,000

      C: (-1000, .50) or D: (-500, .1.0)

      N=68 [69]*--------------[31]

      Replicates Reflection Effect

      But Final States of Wealth are identical (p. 206)

      In Prospect Theory people perceive gains and losses to a neutral reference point
      Utilities and probabilities undergo systematic cognitive distortions.
      There are two phases in Prospect Theory: Editing and Evaluation
      In editing, the following operations were proposed: coding, combination, segregation, cancellation)
      The editing function was criticized as being not well-specified or parsimonious.

      Evaluation phase was better developed. In contrast to Expected Utility Theory it added a value function and a weighting function
      People evaluate gains and losses relative to a neutral reference point
      People tend to be risk averse to gains and risk-seeking to losses
      The above lead to value function of Prospect Theory (video)

      Graph of Prospect Theory

      

      From: https://www.economicsonline.co.uk/definitions/prospect-theory.html/

      Problem 7

      A: (5,000, .001) or B: (5, 1.0)

      N=72 [72]*-----------[28]a

      Here, people weigh large gain more than its expected value.
      This is akin to buying lottery ticket

      Problem 8

      C: (-5,000, .001) or D: (-5, 1.0)

      N=72 [17]-----------[83]*

      Here, people prefer the sure loss rather than risk a large on.
      This is similar to buying insurance.

      • So, Prospect Theory:
        • Keeps the assumptions of Expected Utility Theory
        • But maximizes expectations, not objective utilities
        • Table 14.1 (p. 209) shows the Fourfold Pattern of choice behavior

• Impact of the Classic Study
  • Kahneman's Nobel Prize
  • Highly cited and noticed by economists and psychologists
  • Real World Examples
    • Racetrack bettors (punters in Great Britain, fyi) and poker players attempt to cover their losses
    • Golfers putt better on par putts than on birdie putts
  • Gave rise to Behavioral Economics
• Critique of the Classic Study
  • Editing phase still problematic
  • Refererence points set to 0 may not work when problems have multiple options
    • In other words, context matters

Problem 9

A: one chance in a million to win $1 million

B. 90% chance to win $12 and 10% chance to win nothing

C: 90% chance to win $1 million and 10% chance to win nothings

Option C would hurt more, revealing that reference point of 0 does not apply to all probabilities

Problem 10

A: ($200, .04; $150, .21; $100, .50; $50, .21; $0, .04)

B: ($200, .20; $150, .20; $100, .20; $50, .20; $0, .20)

Here, there are multiple, ranked ways to win leading to cumulative probabilities (e.g., winning something)

• Adding emotions (regret and rejoicing) following choice behavior does not add much to Prospect Theory
• More Issues:
  • Description-Experience Gap
  • Rare Events
  • Sample Size
• Prospects for Prospect Theory:
  • Alter the Weighting Function across paragigms
  • Analyzing the role of experience, sampling, memory, and learning
  • Looking at fundamental issues:
    • Sampling and Ambiguity
    • Effect of longer time horizons
    • Role of dynamic environments
    • Relationship of probability and choice behavior
• Conclusions
  • Prospect Theory is viable alternative to Expected Utility Theory
  • Framing is important
  • Mathematics and Psychology are not equivalent
  • More examination of loss aversion
  • Are there simpler models for human dccision making?
• Afterthought: Comments from Daniel Kahneman
  • Discussed how economists really believed that humans made rational choices most of the time.
  • Wrote about Richard Thaler
    • Nobel Prize in Economics 2017
      • "How do human traits govern individual economic decisions and what effect do they have on markets as a whole? Since the 1980s, Richard Thaler has analyzed economic decision-making with the aid of insights from psychology. He has paid special attention to three psychological factors: the tendency to not behave completely rationally, notions of fairness and reasonableness, and lack of self-control. His findings have had a profound influence on many areas of economic research and policy.

Back to Main Advanced Cognitive Psychology