Bounded rationality

Bounded rationality is the idea that human decision-making is shaped jointly by the cognitive limits of the decider and the structure of the environment in which decisions are made. The term was introduced by Herbert A. Simon in his 1947 book Administrative Behavior and formalized in his 1955 Quarterly Journal of Economics paper “A Behavioral Model of Rational Choice.” Simon's claim was that classical rational-choice theory — which postulates an “economic man” with complete knowledge of alternatives, stable preferences, and unlimited computational ability — describes no actual decider, human or otherwise. Real deciders work under constraints of limited time, limited information, and limited computational capacity, and consequently use procedures (such as satisficing) that find adequate rather than optimal solutions.

Three substantive features of the bounded rationality framework are essential. First, the “bounds” Simon identified come from both sides of the decider-environment interaction. Simon's 1956 Psychological Review paper “Rational choice and the structure of the environment” emphasized that rationality is constrained by environmental structure (information availability, time pressure, complexity) as well as by internal cognitive limits. Simon later described this as “a pair of scissors” with one blade representing cognitive limitations and the other representing environmental structure — both blades necessary to cut.

Second, the contemporary literature on bounded rationality is genuinely contested between two research traditions that interpret Simon's framework differently. The Kahneman-Tversky heuristics-and-biases program (Tversky & Kahneman, 1974) treats bounded rationality as revealing systematic irrationality — heuristics are useful shortcuts but produce predictable departures from normative inference (Bayesian updating, expected utility theory, statistical reasoning). The Gigerenzer ecological rationality program (Gigerenzer & Goldstein, 1996) treats bounded rationality as revealing adaptive intelligence — simple heuristics often match or outperform “rational” procedures when matched to their environment. The two programs share the empirical observation that people use simple decision procedures rather than optimization; they disagree about whether this is a deficit (Kahneman-Tversky) or an adaptive feature (Gigerenzer).

Third, the term “satisficing” that Simon coined in his 1956 paper is now widely used outside academic decision theory. The word is a deliberate combination of “satisfy” and “suffice,” though Simon also noted it appears in the Oxford English Dictionary as a Northumbrian dialect synonym for “satisfy.” A satisficing decider sets an aspiration level (a threshold of acceptability) and selects the first option that meets it, rather than searching the full space of alternatives for the optimum. Most everyday decisions — what to eat for lunch, which apartment to rent, which job offer to accept — are satisficed rather than optimized in this sense.

Bounded rationality matters at three levels with substantively different evidence bases.

For decision theory and economics. Classical economic theory rested on assumptions about rational deciders that Simon argued described no actual decider. Bounded rationality opened the field of behavioral economics by providing a theoretical alternative to optimization-based models. Simon was awarded the 1978 Nobel Memorial Prize in Economic Sciences for “pioneering research into the decision-making process within economic organizations,” the central theoretical contribution of which was bounded rationality. The contemporary behavioral economics literature — Kahneman, Thaler, Sunstein, Camerer, and others — rests on the bounded rationality framework even where it disagrees about which specific deviations from classical rationality are most important.

For cognitive psychology and judgment research. The bounded rationality framework reframed the question of human reasoning. Rather than asking whether humans reason correctly (with the implicit norm being classical logic or probability theory), researchers in this tradition ask how reasoning is structured by the joint constraints of mind and environment. The Kahneman-Tversky and Gigerenzer programs are both heirs to this reframing, even though they emphasize different aspects of it. The heuristics-and-biases tradition catalogued systematic patterns of judgment that depart from normative inference (representativeness, availability, anchoring, framing effects); the fast-and-frugal tradition catalogued conditions under which simple heuristics outperform optimization procedures in real-world decision environments.

For artificial intelligence and machine learning. Simon's framework anticipated several themes that became central to AI research, particularly the recognition that intelligence must work under computational constraints. Simon himself was a founder of AI (the General Problem Solver, Logic Theorist) and brought a bounded-rationality perspective to that field. Contemporary work on bounded-rational reasoning, computational limits in decision-making, and the trade-off between optimality and feasibility in algorithm design traces conceptually back to Simon. The fast-and-frugal heuristics program has had particular influence on areas of machine learning concerned with simple, interpretable decision procedures.

For everyday decision-making. At the practical level, bounded rationality is the recognition that real decisions are made under genuine constraints — time pressure, limited information, finite attention — and that procedures suited to these constraints can produce good outcomes even though they are not optimal in the classical sense. The popular framing of bounded rationality sometimes treats it as a story about human flaws (cognitive biases, irrationality); the original Simon framing was more nuanced: bounded rationality is how minds actually work, not a defect of minds. Whether specific bounded-rational procedures produce good or bad outcomes depends substantially on the match between procedure and environment — the “ecological rationality” emphasis that Gigerenzer's program developed from Simon's 1956 environmental-structure emphasis.

The concept of bounded rationality emerged from Herbert A. Simon's sustained engagement with the disconnect between classical economic theory and observed human behavior. Simon (1916-2001) was a polymath whose career spanned political science, public administration, organizational theory, economics, psychology, computer science, and artificial intelligence. He spent most of his career at Carnegie Mellon, founded several disciplines and subfields, and won the 1978 Nobel Memorial Prize in Economics and the 1975 Turing Award (the only person to have won both).

Simon's 1947 book Administrative Behavior (his PhD dissertation, revised) was the first substantive statement of what would become bounded rationality. Simon argued that classical theories of administration assumed decision-makers had unlimited cognitive capacity, complete information, and stable preferences — assumptions that fit no actual administrator. Real administrators worked under constraints that shaped not only their outcomes but the procedures they used. Simon's framework treated decision-making as a process structured by these constraints rather than as a calculation of optimal choice from given alternatives.

The formal statement came in two seminal papers in the mid-1950s. Simon's 1955 Quarterly Journal of Economics paper “A Behavioral Model of Rational Choice” presented the first formalization of a choice procedure performed by a boundedly rational agent. The paper opened by noting that “traditional economic theory postulates an ‘economic man,’ who, in the course of being ‘economic’ is also ‘rational’” — with complete knowledge, well-organized preferences, and excellent computation skills — and went on to argue that this characterization fits no actual decider. Simon proposed instead a model in which the decider has limited information, uses search procedures rather than complete enumeration, and stops searching when an alternative meeting an aspiration level is found.

The companion paper, Simon's 1956 Psychological Review paper “Rational choice and the structure of the environment,” was the other essential half of the original framework. This paper emphasized that bounded rationality is not solely about cognitive limits — the structure of the environment matters equally. Simple decision rules can be remarkably effective when matched to the structure of the environment in which they are used; the same rules can fail when the environmental structure does not match. The paper introduced the term satisficing (a portmanteau of “satisfy” and “suffice,” which Simon also identified as a Northumbrian dialect word) as the contrast term to “optimizing.” The two 1955-56 papers together established the dual emphasis on cognitive limits AND environmental structure that Simon would later describe as a “scissors with two blades.”

Simon's subsequent work consolidated the framework: Models of Man (1957), his 1978 Nobel Memorial Lecture “Rational decision making in business organizations” (published as Simon 1979 in American Economic Review), Models of Bounded Rationality (1982, three volumes), and many other contributions. By the time of Simon's death in 2001, bounded rationality had become a foundational framework across economics, psychology, and decision theory, though specific interpretations of the framework varied widely.

The most influential development of the framework came from Daniel Kahneman and Amos Tversky, working primarily in the 1970s. Their 1974 Science paper “Judgment under uncertainty: Heuristics and biases” identified three heuristics that people use in probability judgment — representativeness (judging probability by similarity to a prototype), availability (judging frequency by ease of recall), and anchoring and adjustment (estimating from an initial value with insufficient adjustment) — and catalogued the systematic biases each produced. Their 1979 prospect theory paper (Econometrica) and the subsequent decades of work in the heuristics-and-biases tradition catalogued an extensive set of departures from classical rationality. Kahneman won the 2002 Nobel Memorial Prize in Economics (Tversky had died in 1996 and was not eligible); the heuristics-and-biases program became the dominant interpretation of bounded rationality in popular culture and in much of behavioral economics.

The most substantial alternative interpretation came from Gerd Gigerenzer and colleagues at the Max Planck Institute in Berlin. Gigerenzer and Goldstein's 1996 Psychological Review paper “Reasoning the fast and frugal way: Models of bounded rationality” proposed a family of algorithms based on what they called one-reason decision making — simple heuristics that use a single cue rather than integrating all available information. The most discussed of these was Take-The-Best, an algorithm that orders cues by validity, uses the most valid discriminating cue to decide, and ignores the rest. By computer simulation, Gigerenzer and Goldstein showed that Take-The-Best matched or outperformed multiple regression and other “rational” inference procedures on several real-world datasets. The result was an existence proof that simple heuristics can produce good inferences in real-world environments. The broader Gigerenzer program argues that bounded rationality is best understood as ecological rationality — the match between simple decision procedures and the environmental structures in which they operate — and that the Kahneman-Tversky framing systematically misidentifies adaptive procedures as biases.

The dispute between the two programs is substantive and ongoing. Both programs accept the basic Simon framework that decisions are made under cognitive and environmental constraints. They disagree about how to characterize the resulting procedures: as deviations from normative rationality requiring correction (Kahneman-Tversky), or as adaptations to environmental structure that often produce better outcomes than optimization (Gigerenzer). Each program has cited Simon as their intellectual ancestor; in interviews late in life, Simon himself was somewhat ambivalent, though he is often quoted as having closer affinity with the Gigerenzer position than the Kahneman-Tversky one.

Bounded rationality is not a single mechanism but a framework within which several specific procedures have been studied. The procedures map roughly onto the three research programs.

Simon's satisficing (1955-1956+)

Simon's original procedure was satisficing: the decider sets an aspiration level (a threshold of acceptability on each relevant dimension), searches alternatives sequentially, and selects the first alternative that meets the aspiration. Aspirations are adjusted upward when alternatives meeting them are easy to find, and downward when they are hard to find. The procedure does not require enumeration of all alternatives, does not require integration of all dimensions of value into a single utility function, and does not guarantee selection of the optimum — but it produces decisions in finite time using finite computation, which optimization in many real environments does not. Simon emphasized that satisficing is not a deficient substitute for optimization but a procedurally rational adaptation to environments where optimization is infeasible.

Kahneman-Tversky heuristics and biases (1974+)

The heuristics-and-biases tradition catalogued specific procedures that people use under uncertainty and the systematic departures from normative inference each produces:

• Representativeness: judging the probability that A belongs to category B by how similar A is to a prototype of B. Produces systematic biases including base-rate neglect, insensitivity to sample size, the conjunction fallacy (Linda the bank teller), and misperception of randomness.
• Availability: judging the frequency or probability of events by how easily examples come to mind. Produces overestimation of vivid, recent, or emotionally salient events; underestimation of mundane events.
• Anchoring and adjustment: estimating values by starting from an initial “anchor” and adjusting. Adjustments are typically insufficient, so estimates remain biased toward the anchor. Effects appear even when anchors are obviously irrelevant (random numbers from a wheel of fortune).

Later work in this tradition added many more heuristics (affect heuristic, recognition heuristic, fluency, etc.) and many more biases (framing effects, loss aversion, mental accounting, etc.). The program's normative framing is that these heuristics are functionally useful but systematically biased, with the implicit standard being classical decision theory and probability calculus. Kahneman's 2011 book Thinking, Fast and Slow introduced the System 1 / System 2 framing for popular audiences.

Gigerenzer fast-and-frugal heuristics (1996+)

The fast-and-frugal heuristics program studies simple procedures that produce good inferences under real-world constraints:

• Take-The-Best: when deciding between two alternatives, order cues by validity (probability that the cue discriminates correctly), use the most valid cue that discriminates between the alternatives, and ignore the rest. The Gigerenzer and Goldstein (1996) simulations showed this procedure matching multiple regression on several real datasets despite using only one cue per decision.
• Recognition heuristic: if one of two alternatives is recognized and the other is not, infer that the recognized alternative has the higher value on the criterion. Counter-intuitively effective when recognition correlates with the criterion (e.g., city population) even though it ignores all other information.
• Tallying / equal-weight strategies: count cues in favor of each alternative, ignoring weights. Often performs comparably to weighted regression in cross-validation settings.
• Fast-and-frugal trees: simple decision trees with few branches, used for classification under time pressure. Studied extensively in medical decision-making (allocating patients to coronary care vs general ward).

The program's normative framing is that these heuristics are ecologically rational — they exploit structure in the environment to produce good decisions with minimal information and computation. The empirical claim is that simple heuristics often match or outperform more complex procedures in cross-validation settings (where the procedure is fitted on one sample and tested on another), even though more complex procedures fit better on the training sample. The mechanism is the bias-variance trade-off familiar from statistics: simple procedures have higher bias but lower variance, making them more robust on out-of-sample prediction.

How the three programs relate

All three programs accept Simon's framework that decisions are made under cognitive and environmental constraints. The Kahneman-Tversky program emphasizes departures from normative inference; the Gigerenzer program emphasizes adaptive match between procedure and environment. Both empirical traditions have produced substantial supporting evidence. The honest scientific picture is that bounded rationality reveals BOTH systematic deviations from classical rationality AND ecologically adaptive procedures — the two characterizations are not as mutually exclusive as the polemical literature sometimes suggests. A procedure can be a departure from Bayesian updating AND ecologically rational in a specific environment; the two claims operate at different levels of analysis (norm-relative inference quality vs. environment-relative outcome quality).

Bounded rationality as a theoretical framework is not measured directly. What is measured are specific manifestations — specific heuristics, specific biases, specific patterns of choice that depart from optimization or that match ecological rationality criteria. The measurement traditions differ across the three research programs.

Behavioral economics measures of departures from classical rationality. Standard paradigms include intertemporal choice tasks (measuring discount rates and present-bias), risky choice tasks (measuring risk aversion and prospect-theoretic value functions), framing tasks (measuring how mathematically equivalent presentations produce different choices), and willingness-to-pay vs willingness-to-accept asymmetries (measuring loss aversion). The behavioral economics literature has standardized many of these paradigms; effect sizes are typically moderate but reliably replicated.

Cognitive psychology measures of specific heuristics. The heuristics-and-biases tradition uses problem sets like the Linda problem (conjunction fallacy from representativeness), the Tom W. problem (base-rate neglect), and the various anchoring tasks (numerical estimation with manipulated anchors). The Cognitive Reflection Test (Frederick 2005) is a brief measure of the tendency to override intuitive responses with deliberative reasoning — correlated with susceptibility to several heuristics-and-biases effects. The LBL Cognitive Bias Susceptibility tool covers this measurement territory.

Ecological rationality measures. The Gigerenzer program tests heuristics not against normative standards but against real-world predictive performance. The standard methodology is cross-validation: fit several procedures (Take-The-Best, multiple regression, neural networks, etc.) on one sample of a real-world dataset (cities and their population, schools and their dropout rates, used cars and their prices), then test predictive accuracy on a held-out sample. The repeated finding is that simple heuristics often match or exceed more complex procedures in cross-validation despite being substantially less complex. This is a measurement framework about decision-procedure performance in environments, not about cognitive bias.

Behavior Survey approach to bounded rationality. Survey-based instruments measure individual differences in tendency toward specific bounded-rational procedures: maximizer vs satisficer scales (Schwartz et al. 2002), Need for Cognition scale (Cacioppo & Petty 1982), and the Cognitive Reflection Test mentioned above. These measure individual differences within the framework rather than the framework itself.

What the LBL Cognitive Bias Susceptibility tool captures. The Behavior Lab Cognitive Bias Susceptibility tool measures susceptibility to several specific heuristics-and-biases effects including anchoring, framing, availability, and base-rate neglect. Users complete brief problem sets and receive scores indicating relative susceptibility to each bias compared to normative benchmarks. The tool is grounded in the Kahneman-Tversky measurement tradition; users interested in the ecological-rationality alternative perspective should understand that the same patterns can be characterized differently depending on the theoretical framework applied.

Bounded rationality sits at the intersection of several adjacent concepts in decision theory and cognitive psychology.

• vs. classical rationality. Classical rationality (the “economic man” framework Simon critiqued) postulates complete information, stable preferences, and unlimited computational ability. Bounded rationality drops all three assumptions, replacing them with finite information, adaptable aspirations, and finite computation. The two frameworks are not entirely opposed: classical rationality remains a useful idealization for some analyses (the calculation of equilibrium prices in idealized markets), while bounded rationality describes the actual procedures deciders use. The relationship is closer to physics's relationship between idealized friction-free models and applied engineering than to a wholesale replacement.
• vs. heuristic (general concept). A heuristic is any simple decision rule. Bounded rationality is the broader framework within which heuristics are studied. The two often co-occur in casual usage but operate at different levels: bounded rationality is the theoretical framework, heuristics are specific procedures studied within it.
• vs. cognitive bias. Cognitive biases are systematic patterns of departure from normative inference. They are one product of the Kahneman-Tversky reading of bounded rationality. The Gigerenzer reading would describe many of the same patterns differently — as adaptive heuristics rather than biases. The relationship is partly empirical (do these patterns produce poor outcomes?) and partly framing (which norm are we measuring against?).
• vs. prospect theory. Prospect theory (Kahneman & Tversky 1979) is a specific descriptive theory of choice under risk that grew out of the heuristics-and-biases program. It is one development within the broader bounded rationality framework rather than a synonym for it. Prospect theory addresses how people weight gains and losses asymmetrically; bounded rationality is the broader theoretical setting in which such departures from expected utility maximization are studied.
• vs. decision fatigue. Decision fatigue concerns the depletion of decision-making quality over many sequential decisions. Bounded rationality concerns the structure of decision-making under cognitive and environmental constraints regardless of fatigue state. The two interact: fatigued deciders may rely more heavily on bounded-rational shortcuts. But the constructs are conceptually distinct.
• vs. dual-process theory (System 1 / System 2). The dual-process framing distinguishes fast, automatic, intuitive processing (System 1) from slow, controlled, deliberative processing (System 2). It is associated with the Kahneman-Tversky tradition, particularly Kahneman's 2011 popular book. Bounded rationality is compatible with dual-process accounts but is not identical to them; bounded-rational procedures operate in both systems, and the dual-process distinction is itself contested (Keren & Schul 2009; Melnikoff & Bargh 2018).
• vs. analysis paralysis. Analysis paralysis is the failure to decide due to over-analysis. Bounded rationality is the framework in which procedures like satisficing are recommended precisely to avoid this failure mode. The relationship is structural: satisficing exists in part as a procedural remedy for analysis paralysis.
• vs. ecological rationality. Ecological rationality is a specific interpretation of bounded rationality developed in the Gigerenzer program. It emphasizes the match between decision procedures and environmental structure as the criterion for evaluating decision quality. Bounded rationality is the broader framework; ecological rationality is one developed reading of it.

Example 1 — The apartment search

A young professional moving to a new city has a list of requirements: rent below $2,000, commute under 30 minutes, in-unit washer-dryer, and a neighborhood that feels safe at night. After visiting the seventh apartment that meets all four criteria, she signs the lease without continuing to visit the eight further apartments she had scheduled. A friend asks why she didn't at least see the others — one of them might have been better. She responds that she didn't need it to be optimal; she needed it to be good enough, and this one is good enough.

This is the canonical example of Simon's satisficing procedure. The decider has set aspiration levels on multiple dimensions (rent, commute, amenities, safety), is searching alternatives sequentially rather than enumerating all possibilities, and is stopping when an alternative meets all aspirations. A classical-rationality framework would say she may have left value on the table by not searching further; the bounded-rationality framework would say she correctly used the procedure suited to her constraints (limited time, multiple acceptable options exist, continued search has costs). Neither characterization is wrong; they highlight different aspects of the decision. The Gigerenzer ecological-rationality reading would add that in environments where many alternatives meet the aspirations (rental markets in many cities), satisficing produces good outcomes faster than optimization; in environments where few alternatives meet the aspirations, the same procedure may fail.

Example 2 — The medical triage decision

An emergency department physician must decide quickly whether a chest-pain patient should be admitted to the coronary care unit (high resource use, appropriate for serious cardiac events) or to a regular ward (lower resource use, appropriate for milder presentations). She has the patient's ECG, blood pressure, age, recent medical history, and pain characterization. A complex regression model could integrate all this information; in practice, she uses a fast-and-frugal tree with three or four sequential checks: ECG findings, then specific cardiac history, then key risk factors. The tree allocates each patient based on the first discriminating cue, ignoring the rest.

This is the example that Gigerenzer's group has used most prominently in the fast-and-frugal heuristics literature. The fast-and-frugal coronary tree was developed in the 1990s and validated against complex regression models on coronary admission data. The simple tree matched or outperformed regression in predictive accuracy and was substantially easier for clinicians to use under time pressure. The Kahneman-Tversky reading might characterize the simple procedure as a heuristic that risks systematic error; the Gigerenzer reading would characterize it as ecologically rational — matched to the environment (time pressure, ECG-dominated diagnostic information, training of clinicians) in a way that produces better outcomes than the “more rational” alternative. Both readings have empirical support; the case illustrates how the same bounded-rational procedure can be characterized differently depending on the theoretical framework applied.

Bounded rationality is one of the most established frameworks in decision theory and cognitive psychology, but several substantive qualifications matter.

• The Kahneman-Tversky vs Gigerenzer dispute is genuinely contested. Both research programs claim Simon's framework as their intellectual ancestor; they disagree about what bounded rationality reveals. Kahneman-Tversky: systematic departures from normative inference (biases). Gigerenzer: adaptive procedures matched to environmental structure (ecological rationality). Both programs have substantial empirical support. The dispute is partly empirical (when do simple heuristics outperform complex procedures?) and partly normative (which standard is appropriate for evaluating decision quality?). Popular treatments often present one program as the consensus position; the actual literature is more divided.
• “Bounded rationality” means different things in different traditions. The term is used so broadly that it can become unfalsifiable. In behavioral economics, it sometimes means “not classically rational” (which fits any deviation from expected-utility maximization). In cognitive psychology, it sometimes means “uses heuristics” (which fits all cognition). The framework is more useful when applied to specific procedures (satisficing, Take-The-Best, anchoring) than when invoked as a general slogan.
• The Kahneman-Tversky tradition has had its own replication concerns. Several specific findings from the heuristics-and-biases tradition have shown weaker effects on replication than the original reports suggested. The general patterns (anchoring, framing, availability effects) are robust; some specific paradigms (priming effects, specific behavioral interventions, ego depletion) have failed to replicate or show substantially smaller effect sizes than originally reported. The framework as a whole survives these replication concerns, but specific claims should be evaluated against contemporary replication evidence.
• The Gigerenzer ecological-rationality program has its own critics. Some critics argue that the fast-and-frugal heuristics literature has cherry-picked environments where simple heuristics perform well, that the criterion of cross-validation performance is not the only normatively relevant criterion, and that the program's polemical framing of the Kahneman-Tversky tradition is overstated. The empirical core of the program (simple heuristics often match complex procedures in cross-validation) is well-established; the broader interpretive claims are more contested.
• The bounded-rationality framework is descriptive, not always normative. Both Kahneman-Tversky and Gigerenzer programs sometimes slide from descriptive claims (this is how people decide) to normative claims (this is how people should or shouldn't decide). The two are separable: a procedure can be psychologically real (descriptive) and good or bad to use (normative) independently. Popular treatments of bounded rationality often conflate these, treating descriptive findings as either licensing or condemning specific procedures.
• Cultural and individual variation in bounded-rational procedures. Most of the foundational literature on bounded rationality used WEIRD samples. Some bounded-rational procedures (anchoring, framing) replicate well cross-culturally; others (specific aspiration-setting patterns, specific cue orderings) show substantial cultural variation. Individual differences within cultures are also substantial — people differ reliably in their tendency to use deliberative versus intuitive procedures, in their susceptibility to specific biases, and in their use of simple versus complex strategies.
• Bounded rationality does not by itself recommend specific interventions. The framework describes how decisions are made; it does not directly prescribe how to make decisions better. The popular self-help framing of bounded rationality as “here are your biases, now overcome them” depends on the Kahneman-Tversky reading that the biases are correctable deficits. The Gigerenzer reading suggests that many bounded-rational procedures are already adapted to their environments and do not need correction. The practical implications depend substantially on which reading is applied, which in turn depends on the specific environment and decision type.
• Simon's “scissors” metaphor is under-emphasized in much contemporary work. Simon's original framework emphasized both cognitive limits AND environmental structure as essential to understanding bounded rationality. Much contemporary work emphasizes one blade or the other — cognitive limits (heuristics-and-biases tradition) or environmental structure (ecological rationality tradition) — without integrating both. The full Simon framework is more nuanced than either of its inheriting traditions in isolation.

The LBL Cognitive Bias Susceptibility tool measures susceptibility to several specific bounded-rational patterns including anchoring, framing, availability, and base-rate neglect. Users complete brief problem sets and receive scores indicating relative susceptibility to each pattern compared to normative benchmarks. The tool is grounded in the Kahneman-Tversky heuristics-and-biases tradition; the same patterns can be characterized differently within the Gigerenzer ecological-rationality framework. The Big Five Snapshot covers personality dimensions that correlate with use of deliberative versus intuitive decision procedures. Together these tools provide self-assessment of where one's decision-making sits within the broader bounded-rationality framework, useful for self-reflection though not for diagnosis of decision quality.

Bounded rationality is the idea that human decision-making is shaped jointly by the cognitive limits of the decider and the structure of the environment in which decisions are made. The term was introduced by Herbert A. Simon in his 1947 book Administrative Behavior and formalized in his 1955 Quarterly Journal of Economics paper and 1956 Psychological Review paper. Simon's original framing emphasized both cognitive limits AND environmental structure (the “scissors with two blades”) and introduced satisficing as the contrast term to optimization. The framework has developed in two substantially different research programs. The Kahneman-Tversky heuristics-and-biases program (Tversky & Kahneman 1974) interprets bounded rationality as revealing systematic irrationality, with heuristics (representativeness, availability, anchoring) producing predictable biases relative to classical decision-theoretic norms. The Gigerenzer ecological-rationality program (Gigerenzer & Goldstein 1996) interprets bounded rationality as revealing adaptive intelligence, with simple heuristics (Take-The-Best, recognition heuristic, fast-and-frugal trees) often matching or outperforming complex procedures in cross-validation settings. Both programs have substantial empirical support; their dispute is partly empirical (when do simple heuristics outperform?) and partly normative (which standard evaluates decision quality?). The honest scientific picture preserves the core Simon framework that decisions are made under cognitive and environmental constraints while resisting the popular framings that present either Kahneman-Tversky or Gigerenzer as the settled consensus. The framework has been foundational for behavioral economics (Simon's 1978 Nobel; Kahneman's 2002 Nobel), for cognitive psychology, and for artificial intelligence. Practical implications depend substantially on which reading is applied and on the specific decision environment.

This entry is intended as a citable scholarly reference. Choose the format that matches your context. The retrieval date should reflect when you accessed the page, which may differ from the entry's last-reviewed date shown above.

LifeByLogic. (2026). Bounded Rationality: Simon, Satisficing, Heuristics. https://lifebylogic.com/glossary/bounded-rationality/

LifeByLogic. "Bounded Rationality: Simon, Satisficing, Heuristics." LifeByLogic, 14 May 2026, https://lifebylogic.com/glossary/bounded-rationality/.

LifeByLogic. 2026. "Bounded Rationality: Simon, Satisficing, Heuristics." May 14. https://lifebylogic.com/glossary/bounded-rationality/.

@misc{lblboundedrationality2026,
  author = {{LifeByLogic}},
  title = {Bounded Rationality: Simon, Satisficing, Heuristics},
  year = {2026},
  month = {may},
  publisher = {LifeByLogic},
  url = {https://lifebylogic.com/glossary/bounded-rationality/},
  note = {Accessed: 2026-05-14}
}