Ellsberg Paradox¶
Show that people prefer betting on a known-composition urn over an ambiguous one of equal expected value on both colors at once — a pattern no single subjective probability can rationalize — proving ambiguity is a separately priced dimension of uncertainty distinct from risk.
Core Idea¶
The Ellsberg paradox is the finding that people prefer bets on urns with known probabilities over urns with unknown composition of equal expected value — and do so on both the red bet and the black bet, a pattern no single subjective probability over the ambiguous urn can rationalize (it would require P(red) < 0.5 and P(black) < 0.5 at once). It establishes that ambiguity — uncertainty over which distribution governs — is a separately priced dimension distinct from risk, and violates Savage's sure-thing principle.
Scope of Application¶
Requires a chooser whose belief-states distinguish "I know the probability" from "I don't."
- Decision theory beyond expected utility — the home: the entry test every generalisation must pass.
- Asset pricing — ambiguity premia help explain the equity premium and home bias.
- Insurance and incomplete contracts — loading beyond actuarial loss when the distribution is uncertain.
- Deep-uncertainty policy — precaution for climate, pandemics, and AI risk under Knightian uncertainty.
- Behavioural finance — the field ambiguity premium extending the lab finding to real choices.
Clarity¶
The paradox separates two dimensions the Savage-Bayesian frame collapses: risk (known probabilities) and ambiguity (the distribution itself unspecified). Forcing the known-urn preference on both bets proves no single probability rationalizes the choices, so belief must be richer than one measure. It converts a vague sense that the standard model is missing something into a definite missing primitive — ambiguity aversion — plus a diagnostic to locate it.
Manages Complexity¶
The two-urn pattern compresses two sprawls at once: a proliferation of expected-utility generalisations becomes a single pass/fail gate (does the model accommodate the Ellsberg choices?), and a scatter of field anomalies — equity premia, home bias, catastrophe loading — becomes one priced primitive. The analyst tracks one new dimension and one parameter, reading the qualitative pattern off a single risk-versus-ambiguity fork rather than re-deriving each anomaly.
Abstract Reasoning¶
Keyed to the risk-versus-ambiguity fork, the paradox licenses a boundary-drawing move (classify a prospect before pricing it), a diagnostic impossibility move (the two-bet pattern proves belief is richer than one probability), a predictive move (ambiguity aversion demands a premium wherever the distribution is unpinned), and a model-admission gate (a generalisation is admitted only if it reproduces the pattern).
Knowledge Transfer¶
Within decision theory the paradox transfers as mechanism in two capacities — a reusable diagnostic instrument (the two-urn test) and a located primitive (ambiguity aversion) that carries across asset pricing, insurance, and deep-uncertainty policy. The two-urn design itself applies only where belief-states distinguish knowing from not knowing a probability, so it does not travel to thermostats or populations. What recurs cross-domain is the parent ambiguity_aversion; the urns and colour bets are the apparatus, not the portable content.
Relationships to Other Abstractions¶
Current abstraction Ellsberg Paradox Domain-specific
Parents (2) — more general patterns this builds on
-
Ellsberg Paradox is a kind of Paradox Prime
The Ellsberg paradox is a paradox whose apparently reasonable preferences produce a contradiction under every single-prior representation.
-
Ellsberg Paradox is a decomposition of Ambiguity Aversion Domain-specific
Removing the two-urn and Savage-axiom apparatus from the Ellsberg paradox leaves the known-over-unknown probability preference of ambiguity aversion.
Hierarchy paths (3) — routes to 3 parentless roots
- Ellsberg Paradox → Paradox
- Ellsberg Paradox → Ambiguity Aversion → Preference
- Ellsberg Paradox → Ambiguity Aversion → Uncertainty
Neighborhood in Abstraction Space¶
Ellsberg Paradox sits in a sparse region of the domain-specific corpus (64th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.
Family — Choice Paradoxes & Collective Decision-Making (14 abstractions)
Nearest neighbors
- Ambiguity Aversion — 0.88
- Bayesian Nash Equilibrium — 0.83
- Allais Paradox — 0.83
- Middle Ground Fallacy — 0.83
- Ultimatum Game — 0.82
Computed from structural-signature embeddings · 2026-07-12