The Unexpected Hanging Paradox

The Unexpected Hanging Paradox appears to prove that a judge’s announcement cannot be fulfilled — and then the announcement is immediately fulfilled. The contradiction is not in the judge’s statement. It is in the prisoner’s reasoning about knowledge.

A judge tells a prisoner: you will be executed on one weekday next week, but you will not know which day in advance. The execution will be a surprise.

The prisoner reasons backward. It cannot be Friday. If Thursday evening arrives and I am still alive, Friday is the only remaining day, so I would know. An execution on Friday cannot be a surprise, so Friday is eliminated.

If Friday is eliminated, Thursday is next. If Wednesday evening arrives and I am still alive, and Friday has been eliminated, Thursday would be the only remaining day. I would know then too. So Thursday is also eliminated.

The same argument removes Wednesday, Tuesday, and Monday. The prisoner concludes that no day can be the day of execution while the judge’s conditions are met, and therefore the judge’s statement is self-defeating. He relaxes.

The guards arrive on Wednesday. He is surprised.

What went wrong

The prisoner’s reasoning is seductive but contains a flaw. The critical step is the elimination of Friday. The prisoner argues: “If Thursday evening arrives and I am still alive, I will know the execution is on Friday.” But this conclusion depends on the prisoner still believing the judge’s promise — including the promise that the execution will happen during the week. Once the prisoner concludes that the execution cannot happen (because each day eliminates itself), he has undermined the premises on which the subsequent steps of the argument depend.

The argument is self-defeating in a way that the prisoner does not account for. By the time the argument convinces him that no execution can happen, he has abandoned his belief in the judge’s announcement — but that abandonment is precisely what makes the execution possible again, because he no longer expects it.

The paradox was first discussed in philosophical literature in the 1940s, possibly originating from a Swedish radio announcement. Logician Willard Van Orman Quine analyzed it in 1953, arguing that the judge’s announcement was self-referential and not straightforwardly true or false. Martin Gardner wrote about it in Scientific American in 1963, which brought it to a wider audience.

Several distinct analyses exist. One view holds that the judge’s announcement is simply paradoxical — not a well-formed statement — because it refers to the prisoner’s future epistemic state in a way that creates a loop. Another view, developed by logicians in epistemic logic (the logic of knowledge and belief), holds that the prisoner’s backward induction argument contains a subtle equivocation: he treats “I will know X” as a static fact when it is a dynamic fact that changes as days pass and new information arrives.

The surprise examination version — a professor announces there will be an unannounced quiz next week — is logically identical and generates the same apparent paradox.

What I find useful about the puzzle is that it shows how much logical weight the ordinary word “surprise” carries. The prisoner’s reasoning fails not because of arithmetic or formal logic but because of a hidden complexity in what it means to know something in advance. Sometimes the most difficult word in a puzzle is not a mathematical symbol. It is an ordinary word we thought we already understood.