Reasoning Decision Making

Reasoning and decision making is the study of how minds get from what they know, or think they know, to what they do.

12 frameworks, from 1944 to today · Cognitive Science · Social Sciences

still a working view ended · each leads to its entry below

Overview

Reasoning and decision making is the study of how minds get from what they know, or think they know, to what they do. It asks how people draw inferences, weigh options, and choose under uncertainty—and why they so often do it badly. The stakes are practical as well as theoretical. A theory of decision tells economists how markets behave, physicians how diagnoses go wrong, policymakers how to present choices, and engineers how to build machines that decide. The field's history is largely a story of tension: between models of how a perfectly rational agent should decide and evidence of how actual people do.

Expected Utility and the Rational Agent

The modern starting point is the idea that a rational chooser maximizes expected utility1. The core notion—that a gamble is worth the probability-weighted sum of its outcomes—took shape in the seventeenth and eighteenth centuries, in correspondence between Blaise Pascal and Pierre de Fermat on games of chance and in Daniel Bernoulli's 1738 treatment of the St. Petersburg paradox, where he proposed that people value the utility of money rather than money itself. In 1944 John von Neumann and Oskar Morgenstern gave the idea an axiomatic foundation in Theory of Games and Economic Behavior: an agent whose preferences satisfy a few consistency conditions (completeness, transitivity, and a substitution principle) can be represented as maximizing expected utility. Leonard Savage extended this in 1954 to subjective probability, so that even without known odds a coherent agent behaves as if assigning probabilities and maximizing2.

This framework was powerful precisely because it was normative: it defined what consistency requires, not what people actually do3. That gap became the field's central problem.

Bounded Rationality and the First Cracks

Herbert Simon, working in the 1950s at what became Carnegie Mellon, argued that real agents face limits of information, time, and computation. His answer was bounded rationality: people satisfice—search until an option is "good enough"—rather than optimize over all possibilities456. In a famous illustration, a foraging organism choosing among patches does better with a simple stopping rule than with an exhaustive calculation it cannot afford. Simon's 1955 paper on a behavioral model of rational choice and his later work on administrative behavior reframed rationality as an adaptation to scarcity of cognitive resources, not a departure from an ideal.

Around the same time, experimental psychologists began documenting systematic departures from the axioms. Ward Edwards introduced Bayesian ideas to psychology in the 1950s, but it was the 1970s that proved decisive.

Heuristics and Biases

In 1974 Amos Tversky and Daniel Kahneman published "Judgment under Uncertainty: Heuristics and Biases," arguing that people rely on a small set of mental shortcuts that are usually efficient but produce predictable errors47. Availability: we judge the frequency of an event by how easily examples come to mind, so vivid risks feel commoner than they are8. Representativeness: we judge probability by similarity to a stereotype, leading to the conjunction fallacy—the belief that "Linda is a bank teller and a feminist" is more likely than "Linda is a bank teller," though a conjunction can never be more probable than one of its parts9. Anchoring: estimates are pulled toward an arbitrary starting number10.

The program's significance was methodological as much as substantive: it showed that errors were not random noise but lawful, and that they revealed the underlying machinery. The 1982 book Judgment under Uncertainty: Heuristics and Biases, edited by Kahneman, Paul Slovic, and Tversky, consolidated the field. Critics—notably Gerd Gigerenzer—later argued that some "biases" vanish when problems are framed in frequencies rather than probabilities, a debate that shaped the next generation.

Prospect Theory

Kahneman and Tversky's 1979 Prospect Theory replaced expected utility as a descriptive account of choice under risk11. Three features did the work. First, people evaluate outcomes as gains or losses relative to a reference point, not as final states of wealth. Second, loss aversion: losses loom larger than equivalent gains, so losing \$100 hurts more than winning \$100 pleases12. Third, diminishing sensitivity: the difference between \$100 and \$200 feels larger than between \$1,100 and \$1,200, producing a distinctive S-shaped value function that is concave for gains and convex for losses12. A fourth element, probability weighting, captures how people overweight small probabilities (hence lottery tickets and fear of rare disasters) and underweight moderate ones.

The theory explained framing effects—most people reject a gamble framed as a loss but accept the same gamble framed as a gain—and it became the backbone of behavioral economics. A cumulative version in 1992 refined the probability weighting13. Kahneman received the Nobel Memorial Prize in Economics in 2002 (Tversky had died in 1996).

Dual-Process Theory

A parallel strand asked whether reasoning comes in two kinds. The idea has deep roots—William James distinguished associative from analytic thought—but its modern form crystallized in the 1970s and 1980s14. Jonathan Evans proposed a distinction between heuristic and analytic processing; Seymour Epstein described "experiential" versus "rational" systems; and in 1996 Steven Sloman named the dual-process framework explicitly. The most cited version is Kahneman's 2011 Thinking, Fast and Slow, which labels the two "System 1" (fast, automatic, associative, effortless) and "System 2" (slow, deliberate, rule-governed, effortful). The framing is a useful shorthand rather than a claim about two anatomical systems, and later theorists (notably Keith Stanovich) emphasized individual differences in the disposition to engage System 2. The theory's appeal is that it reconciles the heuristics-and-biases findings with the human capacity for careful reasoning: biases arise when System 1 answers and System 2 fails to check.

Fast and Frugal Heuristics

Gigerenzer and colleagues at the Max Planck Institute in Berlin mounted a sustained alternative in the 1990s. Fast and frugal heuristics are simple rules that exploit the structure of environments to make good decisions with little information4. The recognition heuristic: if you recognize one of two options and not the other, choose the recognized one—a rule that predicts, for example, which of two cities is larger surprisingly well15. Take-the-best: search cues in order of validity and stop at the first that discriminates16. The program's claim was that such heuristics can match or beat complex models, and that rationality should be judged by ecological fit—how well a rule works in a given environment—rather than by conformity to logic. This "ecological rationality" reframed the heuristics-and-biases debate as one about which environments make which shortcuts smart17.

Mental Models

How do people reason deductively? Philip Johnson-Laird's mental models theory, developed from the early 1980s, proposed that people construct small analogical representations of the situations described by premises and test conclusions against them. Reasoning errors arise because people build only one or a few models and fail to consider alternatives—explaining why some syllogisms are easy and others hard. The theory competed with rule-based accounts (mental logic) and with probabilistic approaches, and it remains a major framework for understanding deduction and spatial reasoning.

Evolutionary and Social-Exchange Accounts

A different challenge came from evolutionary psychology. Leda Cosmides and John Tooby argued in the 1980s and 1990s that human reasoning is not a general-purpose logic engine but a collection of domain-specific modules shaped by ancestral problems. Their central evidence was the Wason selection task: people perform poorly on abstract conditional rules but dramatically better when the rule is framed as a social contract ("if you take a benefit, you must pay a cost"). They interpreted this as a cheater-detection module. The claim remains contested—alternative explanations invoke pragmatic reasoning schemas or relevance—but it established that reasoning performance depends heavily on content and context, not just formal structure.

Bayesian Cognitive Science

From the 1980s onward, a movement argued that the mind is fundamentally a probabilistic inference engine. The Bayesian framework treats perception and cognition as inference under uncertainty: the brain combines prior expectations with sensory evidence according to Bayes' rule to produce posterior beliefs18. Early applications included reading and vision; the program matured in the 2000s with work on causal learning, language, and "predictive processing," which holds that the brain constantly generates predictions and updates on prediction error. Its appeal is unification: a single normative principle explains why heuristics work, how learning generalizes, and why illusions occur. Its critics ask whether the computations are biologically plausible and whether "Bayesian" models are testable rather than flexible.

Learning, Emotion, and the Brain

Two further strands connected decision making to learning and to neuroscience. Reinforcement learning models, rooted in behaviorist learning theory and formalized in the 1980s and 1990s, describe choice as the updating of value estimates through reward prediction error—a signal later found to correspond to dopamine neuron activity. These models now explain everything from foraging to addiction.

Antonio Damasio's somatic marker hypothesis, developed from the 1990s, proposed that bodily emotional signals guide decision, especially in complex social contexts. Patients with damage to the ventromedial prefrontal cortex could reason abstractly about options but chose badly in real life, apparently because they lacked the emotional "markers" that tag options as good or bad19. The hypothesis linked emotion to rationality rather than opposing them19.

These threads converged in neuroeconomics, which emerged as a named field around 2005, combining economic theory, psychology, and brain imaging to study how value, risk, and social context are represented neurally. It inherited prospect theory's descriptive findings, reinforcement learning's computational tools, and the somatic marker tradition's interest in emotion.

The Present

Today the field is plural. Expected utility remains the normative benchmark and the language of economics, while prospect theory and its descendants supply the descriptive corrections. Dual-process and fast-and-frugal accounts coexist uneasily, with growing interest in when deliberation actually helps. Bayesian and reinforcement-learning models dominate computational work, increasingly integrated with neural data. Open questions persist: whether "biases" are errors or adaptations, how individual and cultural variation fit general models, and how far findings from laboratory tasks transfer to real-world decisions. The history's lesson is that the tension between the rational ideal and the actual mind has been productive—each challenge to the model has revealed something about how people really choose.

References

  1. Expected utility hypothesis, Wikipedia
  2. Subjective expected utility, Wikipedia
  3. Decision theory, Wikipedia
  4. Heuristic (psychology) (History), Wikipedia
  5. Bounded rationality, Wikipedia
  6. Satisficing, Wikipedia
  7. Cognitive bias (History), Wikipedia
  8. Availability heuristic, Wikipedia
  9. Representativeness heuristic, Wikipedia
  10. Anchoring effect, Wikipedia
  11. Prospect theory, Wikipedia
  12. Loss aversion (introduction, History), Wikipedia
  13. Cumulative prospect theory, Wikipedia
  14. Dual process theory (History), Wikipedia
  15. Recognition heuristic, Wikipedia
  16. Take-the-best heuristic, Wikipedia
  17. Ecological rationality, Wikipedia
  18. Bayesian approaches to brain function, Wikipedia
  19. Somatic marker hypothesis (Background), Wikipedia

The frameworks of Reasoning Decision Making

Each framework is a way of working the field has used: when it held, and the question it answers.

  1. 1944-Present

    Subjective expected utility

    How should a rational agent choose under uncertainty?

    Fly to its tower in Nousfarer

  2. 1955-Present

    Bounded Rationality

    How do people decide with limited time and information?

    Fly to its tower in Nousfarer

  3. 1974-Present

    Heuristics and Biases

    Why do people make predictable errors in judgment?

    About Heuristics and Biases
    How do people actually make judgments and decisions when information is incomplete, time is scarce, and the world is uncertain? The Heuristics and Biases program, launched in the early 1970s, answered this question by showing that human reasoning does not follow the elegant rules of probability theory or rational choice. Instead, people rely on a toolkit of mental shortcuts—heuristics—that usually work well but can lead to systematic errors, or biases. This framework did not merely catalog human irrationality; it argued that heuristics are adaptive, fast, and often accurate, but their occasional failures reveal the underlying cognitive machinery.
    For an introductory course, the Heuristics and Biases program is built around a set of core concepts that explain both the power and the pitfalls of intuitive judgment.
    Representativeness is the tendency to judge the probability that an object belongs to a category by how similar it is to a typical member, ignoring base rates and sample sizes. The availability heuristic shows that people assess the frequency of an event by the ease with which instances come to mind, leading to overestimation of vivid or recent occurrences. Anchoring and adjustment describes how judgments are influenced by an initial reference point, even when that anchor is arbitrary; insufficient adjustment often yields biased estimates.
    Framing effects demonstrate that the same choice presented in different ways (e.g., as a gain vs. a loss) changes preferences. This connects directly to loss aversion, the idea that losses hurt more than equivalent gains feel good. Overconfidence captures the gap between what people know and what they think they know, especially in calibration of probabilities. The conjunction fallacy—judging a specific conjunction as more likely than a single constituent—exposes violations of probabilistic logic (the classic Linda problem).
    Confirmation bias describes the tendency to seek, interpret, and remember evidence that supports one's prior beliefs. Hindsight bias makes past events seem predictable after the fact, distorting learning. Regression to the mean is often overlooked: people see patterns in random fluctuations and then invent causal explanations for inevitable statistical phenomena. Finally, the Affect heuristic shows that emotional responses (like/dislike) can drive judgments of risk and benefit, replacing deliberate analysis with gut feeling.
    These concepts are often taught through the lens of dual-process theories, which distinguish fast, automatic, intuitive System 1 thinking from slow, deliberate, analytical System 2 reasoning. Heuristics and biases are the signature of System 1.
    ---
    The intellectual history of Heuristics and Biases begins in the early 1970s with two Israeli psychologists, Daniel Kahneman and Amos Tversky. Their collaboration was a reaction against the dominant normative frameworks of the time: Expected Utility Theory and Subjective Expected Utility, which assumed ideal rationality. Kahneman and Tversky were not content simply to point out that people deviate from these ideals; they wanted a descriptive model of how real minds work.
    Their first major statement came in a 1973 paper, 'On the Psychology of Prediction,' which introduced the representativeness heuristic and several associated biases. This was followed in 1974 by a landmark article in Science, 'Judgment under Uncertainty: Heuristics and Biases,' which laid out the core trio—representativeness, availability, anchoring—and their systematic consequences. That paper became the founding document of the entire research program. Kahneman and Tversky collected decades of experimental demonstrations into the 1982 volume Judgment under Uncertainty: Heuristics and Biases, which remains the canonical anthology.
    Prospect Theory, introduced by Kahneman and Tversky in 1979, emerged as a direct successor that focused more narrowly on choice under risk rather than judgment. It refined the concepts of framing and loss aversion into a formal alternative to Expected Utility Theory. Prospect Theory coexisted with and eventually supplemented Heuristics and Biases, but it did not replace it; the earlier program continued to inspire work on inference, prediction, and social judgment.
    Peer frameworks meanwhile took different paths. Mental Models (beginning in 1983) focused on deductive reasoning through mental simulations rather than statistical shortcuts. Bayesian Cognitive Science (emerging around 1990) argued that many biases are not errors but rational approximations given limited computation. Dual-Process Theories, formalized around 1990, gave Heuristics and Biases a broader cognitive architecture by distinguishing System 1 and System 2—but this was a later codification, not part of the original framework.
    Gerd Gigerenzer and the 'Heuristics and Biases' critics argued in the 1990s that the program overemphasized errors; they proposed 'fast and frugal heuristics' as ecologically rational tools rather than fallacies. This debate sharpened the framework's claims but did not dislodge its core teaching. Kahneman received the Nobel Memorial Prize in Economics in 2002 for integrating psychological insights into economics, cementing the legacy of the program.
    The Heuristics and Biases program fundamentally reshaped reasoning and decision making by replacing an idealized model of rationality with a psychologically realistic one. Its concepts remain essential for understanding why experts and laypeople alike commit predictable errors, and how decision environments can be redesigned to improve outcomes.
    The 12 concepts its tower teaches
    • Affect Heuristic
    • Anchoring and Adjustment
    • Availability Heuristic
    • Confirmation Bias
    • Conjunction Fallacy
    • Framing Effects
    • Hindsight Bias
    • Loss Aversion
    • Overconfidence
    • Regression to the Mean
    • Representativeness
    • System 1 / System 2

    Fly to its tower in Nousfarer

  4. 1975-Present

    Dual-Process Theory

    Why do we sometimes think fast and sometimes slow?

    Fly to its tower in Nousfarer

  5. 1979-Present

    Prospect Theory

    How do people actually weigh risky gains and losses?

    Fly to its tower in Nousfarer

  6. 1983-Present

    Mental model theory of reasoning

    How do people draw conclusions from premises?

    Fly to its tower in Nousfarer

  7. 1985-Present

    Bayesian Cognitive Science

    How does the mind infer under uncertainty?

    Fly to its tower in Nousfarer

  8. 1989-Present

    Evolutionary Psychology of Reasoning

    Why is reasoning shaped by ancestral social problems?

    Fly to its tower in Nousfarer

  9. 1989-Present

    Reinforcement Learning Models

    How do agents learn which actions pay off?

    Fly to its tower in Nousfarer

  10. 1991-Present

    Fast and Frugal Heuristics

    When do simple rules beat complex calculation?

    Fly to its tower in Nousfarer

  11. 1994-Present

    Somatic Marker Hypothesis

    How do bodily feelings guide good decisions?

    Fly to its tower in Nousfarer

  12. 2005-Present

    Neuroeconomics

    How does the brain compute value and choice?

    Fly to its tower in Nousfarer

How the ideas connect

Each arrow runs from an earlier idea to the later one it shaped. The list below says what passed between them.

Subjective expected utilityBounded RationalityHeuristics and BiasesDual-Process TheoryProspect TheoryMental model theory of reasoningBayesian Cognitive ScienceEvolutionary Psychology of ReasoningReinforcement Learning ModelsFast and Frugal HeuristicsSomatic Marker HypothesisNeuroeconomics

Beyond the field

Where Reasoning Decision Making's ideas came from in other fields, and what they led to there: the strongest 12 of 56 links.

A city in Nousfarer: a red-carpeted stair climbing terraces to a crowned tower on an acropolis, more towers round it, forest and snowy mountains beyond.
A city in Nousfarer, the game played in the browser.

Reasoning Decision Making is a planet in Nousfarer

Its city has a tower for each of these 12 frameworks. Land, walk in, and climb one concept at a time.

Land on Reasoning Decision Making

Related fields

Fields its ideas are linked to

More in Cognitive Science

See where Reasoning Decision Making sits on the Map of Human Knowledge