Daniel Kahneman

Introduction

Daniel Kahneman (5 March 1934 – 27 March 2024) established that human judgement departs from rational norms in ways that are systematic rather than random — and therefore, potentially, describable. Errors of judgement, on his account, are not noise around a correct answer; they are the output of identifiable mental shortcuts operating in identifiable circumstances, and they can be predicted.

That claim, developed with Amos Tversky across the 1970s, remade economics and created behavioural economics. It also made him, in 2002, the second psychologist to win the Nobel Memorial Prize in Economic Sciences — an award he received despite having, as he liked to point out, never taken an economics course in his life.

The second half of his career turned to a different question: what makes a life go well, and why people are so bad at judging it. He distinguished the experiencing self, which lives moment to moment, from the remembering self, which keeps a story — and argued that the two disagree systematically, that most people are governed by the remembering self, and that this makes them poor judges of their own well-being.

He was also one of the few architects of the replication crisis who engaged with it honestly. When the field began failing to reproduce its own classic effects, Kahneman did not dispute the finding; he wrote that he had "placed too much confidence in underpowered studies" and publicly revised a famous result of his own.

Early Life and Education

Kahneman was born in Tel Aviv while his mother was visiting relatives there; the family lived in Paris. His parents were Lithuanian Jews who had emigrated to France in the 1920s.

In 1940 the Germans occupied Paris. His father was among the first Jews rounded up in the city and was held for six weeks, released only after his employer intervened. The family then went into hiding, moving between hiding places in the French countryside; his father died of diabetes in 1944 while they were still in hiding. Kahneman, his mother and his sister survived, and in 1948 — shortly before Israeli independence — they emigrated to Mandatory Palestine.

He has told the story of being stopped in occupied Paris by a German soldier, wearing his own Star of David, and being terrified — at which the soldier showed him a photograph of his own son and gave him money. He said the incident taught him that people are complicated, and it is difficult to read his work without seeing someone who had learned that lesson early.

He took his BA at the Hebrew University of Jerusalem in 1954, majoring in psychology with a minor in mathematics. He then served in the Israeli Defence Forces in its psychology branch, assessing candidates for officer training — work that gave him the material for one of his most durable practical findings: that unstructured interviews feel informative and predict nothing, and that a structured, decomposed rating procedure beats clinical judgement. He and Tversky later built a selection interview on that principle, and it was used by the Israeli army for years.

He went to the University of California, Berkeley, in 1958 and took his doctorate in 1961, with a dissertation on how the meaning of an adjective depends on the sentence it appears in.

Academic and Professional Career

Visual perception, and the move to judgement

Kahneman returned to the Hebrew University as a lecturer in 1961 and rose to professor, staying until 1978. His early work was on visual perception and attention — Attention and Effort (1973) came out of it — and the cognitive-psychology training is visible in everything he later did: the assumption that the mind has limited capacity, that it takes shortcuts to manage the load, and that the shortcuts can be specified.

He met Tversky at Hebrew University in 1968. The collaboration that followed is one of the genuine partnerships in the history of science — Kahneman later described it as one in which each had the other's unfinished thoughts — and it produced, between 1971 and 1979, essentially the whole of the heuristics-and-biases programme: "Belief in the Law of Small Numbers" (1971), "Judgment under Uncertainty: Heuristics and Biases" (Science, 1974), and "Prospect Theory: An Analysis of Decision under Risk" (Econometrica, 1979).

North America, and the Nobel

In 1977–78 both men were fellows at the Center for Advanced Study in the Behavioral Sciences at Stanford, where the economist Richard Thaler was also in residence. That year is where behavioural economics begins: Thaler took their psychological findings into economics, and they took his problems back.

Kahneman then held chairs at the University of British Columbia (1978–1986) and the University of California, Berkeley (1986–1994), and from 1993 at Princeton, where he was the Eugene Higgins Professor of Psychology and professor of public affairs at the Woodrow Wilson School.

The 2002 Nobel was awarded to him jointly with Vernon Smith, "for having integrated insights from psychological research into economic science, especially concerning human judgment and decision-making under uncertainty". Tversky had died of melanoma in 1996, six years earlier, and the prize is not awarded posthumously. Kahneman's response was to say that the prize was theirs, and he spent much of the day explaining why.

Well-being, and the last decade

From the 1990s Kahneman worked on hedonic psychology: the measurement of experienced well-being, the day reconstruction method, the peak-end rule, and the focusing illusion — the finding that people overestimate the effect of any single circumstance on their happiness because, when they think about it, they think about nothing else. His 2010 paper with Angus Deaton, which found that emotional well-being rose with income up to about $75,000 a year and then flattened, became one of the most cited results in the psychology of happiness — and it turned out to be wrong (see below).

Thinking, Fast and Slow (2011) consolidated the whole of the programme for a general audience and became an unlikely bestseller. Noise (2021), with Cass Sunstein and Olivier Sibony, addressed variability in judgement — the argument that the field had studied bias while ignoring the fact that the same person judging the same case on two days will reach two different conclusions.

He died on 27 March 2024, three weeks after his ninetieth birthday, by assisted dying in Switzerland; the manner and place of his death were made public only in March 2025.

Major Contributions to Psychology

Heuristics and biases

The claim is that people estimate uncertain quantities and probabilities not by calculation but by substituting an easier question:

  • Availability heuristic — judging frequency or probability by how readily instances come to mind, which is why coverage distorts risk perception
  • Representativeness heuristic — judging probability by similarity to a prototype, which produces base-rate neglect and the conjunction fallacy
  • Anchoring — estimates are pulled toward a previously considered value, even an arbitrary one

Each heuristic produces characteristic errors: insensitivity to base rates, insensitivity to sample size, the conjunction fallacy (the "Linda problem", in which people judge a conjunction to be more probable than one of its constituents), overconfidence, and the planning fallacy — the systematic tendency to underestimate how long a project will take, which Kahneman regarded as the single most useful thing he knew for decision-makers.

Prospect theory

Prospect theory (1979) replaced expected utility theory as the standard descriptive account of choice under risk. Its components:

  • Reference dependence — outcomes are evaluated as gains and losses relative to a reference point, usually the status quo, not as states of wealth
  • Loss aversion — losses loom larger than equivalent gains; the value function is steeper below the reference point than above it
  • Diminishing sensitivity — the value function is concave for gains and convex for losses, producing risk aversion in the domain of gains and risk seeking in the domain of losses
  • Probability weighting — people overweight small probabilities and underweight moderate and high ones, which explains why they buy lottery tickets and insurance simultaneously

The pattern this produces is the reflection effect: people avoid risk when choosing among gains and seek it when choosing among losses. That asymmetry had been inexplicable under expected utility theory, and prospect theory's explanation of it — along with the Allais paradox — is why the economics prize went where it did.

System 1 and System 2

The framework of Thinking, Fast and Slow distinguishes two modes: System 1, fast, automatic, associative, effortless, and the source of most judgements; and System 2, slow, effortful, serial, and capable of overriding System 1 only with difficulty. The biases are largely System 1's defaults, and System 2's laziness is why they survive.

Kahneman was explicit that the two systems are a fiction — a useful device for exposition, not a claim about brain architecture — and that they should not be read as modules. The distinction has nevertheless been taken up as though it were a mechanism, which is one of the standing criticisms of the book's influence.

Experienced versus remembered well-being

The second half of his career's central finding is that the two selves disagree. The experiencing self reports how it feels moment to moment; the remembering self keeps a summary, and that summary is governed by the peak-end rule — the remembered quality of an episode is predicted by its peak intensity and its ending, and almost not at all by its duration.

The consequence, demonstrated most starkly in the colonoscopy studies in which a longer procedure with a gentler ending was remembered as less unpleasant than a shorter one that ended badly, is that people choose what will make for a better story rather than what will feel better. Kahneman's position was that this matters, that policy should be built on experienced well-being, and that a person who has been told the two measures come apart can sometimes choose differently.

Method: adversarial collaboration

Kahneman's methodological legacy is a practice rather than a technique. He disliked the standard pattern in which two camps publish past each other for decades, and he pioneered adversarial collaboration — designing a study with a critic such that both parties agree in advance on what result would count for whom. He did it with Gary Klein on the question of whether expertise can be trusted, and — most consequentially — with Matthew Killingsworth on the income and happiness question, where it produced a real correction rather than a stalemate.

Major Publications

  • Attention and Effort (1973) — the cognitive psychology of capacity and effort
  • "Judgment under Uncertainty: Heuristics and Biases" (1974, Science, with Tversky) — the programme statement
  • "Prospect Theory: An Analysis of Decision under Risk" (1979, Econometrica, with Tversky) — the paper the Nobel cited
  • Judgment under Uncertainty: Heuristics and Biases (1982, edited with Paul Slovic and Tversky) — the field's founding anthology
  • Well-Being: The Foundations of Hedonic Psychology (1999, edited with Ed Diener and Norbert Schwarz)
  • Choices, Values, and Frames (2000, with Tversky) — the collected decision papers
  • Thinking, Fast and Slow (2011) — the popular synthesis
  • Noise: A Flaw in Human Judgment (2021, with Cass Sunstein and Olivier Sibony)

Impact and Legacy

Kahneman's effect is unusually easy to specify because it created a discipline. Behavioural economics did not exist before the heuristics-and-biases programme, and it now has its own journals, its own Nobel prizes, its own policy units — the Behavioural Insights Team in Britain and analogous bodies elsewhere — and its own vocabulary, much of which is ordinary language now: loss aversion, anchoring, the planning fallacy, availability.

In economics, prospect theory broke the expected-utility monopoly and made it possible to model the anomalies. In psychology, the programme re-established that judgement can be studied experimentally, at a moment when the cognitive revolution was mostly about perception and memory. And in practice, the applications are now standard: default options in pension enrolment, the design of medical choice architecture, the reform of intelligence analysis, and the replacement of unstructured interviews with structured ones — the last of which is his own, from his twenties.

He also modelled something rarer than any finding: how to respond when your field turns out to have been overclaiming.

Criticism

The replication crisis hit his own programme, and he said so. The heuristics-and-biases literature was built on small-sample laboratory demonstrations with hypothetical scenarios. When psychology began systematically replicating its classic findings, several of the second-tier effects failed or shrank substantially, and Kahneman's response was to write that he had "placed too much confidence in underpowered studies". Some effects hold up extremely well — anchoring is among the most robust findings in all of psychology — but the literature as a whole is now understood to have been more fragile than its presentation implied.

Gigerenzer's critique is the deepest one. Gerd Gigerenzer argues that heuristics are not biases but adaptive tools, that the programme's experiments use artificial problems stated in unfamiliar formats, and that calling people irrational for failing to answer a probability question stated in single-event terms is a category error. The most damaging specific demonstration: base-rate neglect largely disappears when the same information is given as natural frequencies ("8 out of 1000") instead of percentages. Gigerenzer and Hoffrage showed that in 1995, and it suggests that at least some of the "errors" are artefacts of how the question was asked. The conjunction fallacy is contested on the same grounds — it may be a pragmatic inference about what the speaker meant rather than a failure of reasoning.

Loss aversion is contested. Whether it is a genuine general asymmetry, or a name for several distinct phenomena — inertia, status quo preference, the greater difficulty of giving things up for reasons unrelated to loss sensitivity — is an open dispute. Gal and Rucker argued in 2018 that the concept has been overgeneralised and that the evidence for it as a unified explanatory principle is weak; the debate has not resolved, and the term is used far more confidently in applied work than the evidence supports.

Prospect theory is descriptive, not explanatory. It fits choice patterns with impressive accuracy, but it does not say why the value function has the shape it does, and it has competitors — decision-by-sampling, the priority heuristic — that fit some of the same data with fewer parameters. It is a measurement model rather than a theory of the underlying process.

The System 1 / System 2 framing has been over-reified. Kahneman said it was a fiction for exposition. It has been adopted as a dual-process architecture, and the adoption has licensed a large amount of confident explanation of phenomena the framework does not actually account for.

The $75,000 finding was wrong. Kahneman and Deaton's 2010 result — that emotional well-being plateaus at about $75,000 in annual income — was one of the most cited findings in well-being research. Matthew Killingsworth's 2021 experience-sampling study found that well-being continues to rise with income well beyond that point. Rather than defend the result, Kahneman and Killingsworth ran an adversarial collaboration (2023, PNAS) and reported that for the great majority of people well-being does keep rising with income, while for a minority — roughly the unhappiest fifth — it rises and then plateaus. Kahneman stated publicly that his earlier claim had been wrong. It is the best-known example of the practice he advocated, and it is worth noting that the correction went against his own result.

And the policy applications have overclaimed. The "nudge" industry that grew from this work has been criticised for effect sizes that shrink on replication, for interventions that work in a specific setting and not at scale, and — the strongest version — for using cognitive biases to shift individual behaviour in place of addressing structural causes.

Associated Figures

  • Amos Tversky — his collaborator through the decisive decade; the partnership produced the whole programme
  • Paul Slovic — co-editor of Judgment under Uncertainty and the third figure in the founding of the field
  • Richard Thaler — the economist who carried their findings into economics, and their host at Stanford in 1977–78
  • Anne Treisman — the cognitive psychologist to whom he was married from 1978 until her death in 2018
  • Gerd Gigerenzer — his principal critic, on ecological rationality and fast-and-frugal heuristics
  • Gary Klein and Matthew Killingsworth — his partners in adversarial collaboration

Related Experiments

Learn More

References

  • Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47(2), 263-291.
  • Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux.

Last updated: 2026-10-20

Menu
Loading