Experiment #82 · Thought experiment

The Lottery Paradox

Rational belief and conjunction

Henry Kyburg · 1961 · Epistemology

First published: H. E. Kyburg, *Probability and the Logic of Rational Belief* (1961).

For each ticket in a million-ticket lottery, you rationally believe it will lose. Conjoin all such beliefs and you believe none will win — contradicting your belief that one will.

Given a fair lottery of one million tickets, for each ticket *t* it is rational to believe that *t* will lose (probability 0.999999). But these beliefs jointly entail that no ticket wins, which contradicts the certain belief that some ticket will. The paradox forces a choice: deny closure of rational belief under conjunction, or deny that high probability suffices for belief, or accept inconsistent beliefs as rational in some attenuated sense. The case is central to debates over the relation between credence and outright belief.

Formulation

Million-ticket fair lottery; rational belief threshold p ≥ 0.99. For each ticket, P(losing) = 0.999999, so believe it loses. Conjunction over all tickets: believe no ticket wins. But P(some ticket wins) = 1.

Dimensions Engaged

Observer

Observer · Knowledge Extent: is rational belief sensitive to conjunction effects?

Information

Bears on Information · Granularity: how do continuous credences relate to discrete belief states?

Responses — How Schools Engage

Affirms / takes the bait 1

A vindication of suspending judgement: rational belief norms over-promise; the lottery makes their inconsistency visible.

Reframes the question 3

Bayesian naturalism: there is no "outright belief" in any deep sense; credences are what matter, and the paradox dissolves once we abandon the threshold-belief framework.

Belief is action-guiding; the paradox arises only when belief is divorced from contexts of action. In any actual decision, the conjunction "no ticket wins" plays no functional role.

The puzzle reveals that ordinary "belief" is not the appropriate doxastic state for science; quantitative probability is.

Holds it inconclusive 1

Live debate: defenders of closure (Williamson) deny that high credence suffices for belief; defenders of high-credence belief (Foley) deny closure. No consensus.

Related Experiments

Experiments engaged by an overlapping set of schools — likely to surface the same fault lines.

Further reading

  • Kyburg (1961), op. cit.
  • Douven, *The Lottery and Preface Paradoxes* (2014)

Related Historical Debates

Debates that share dimensions and/or aligned schools with this experiment.

Personas Most Aligned With This Experiment

Ranked by total declared-influence weight in the schools that respond to this experiment.

Works Most Aligned With This Experiment

Ranked by total declared-influence weight in the schools that respond to this experiment.

Related Films

Films engaging the same dimensions as this experiment.

Related Contemporary Dilemmas

Dilemmas that engage the same dimensions as this experiment.

← The Doomsday Argument The Two Envelopes Paradox →