The Barber Paradox

The paradox

In a town, there is one barber. The barber shaves exactly those men in the town who do not shave themselves. Every man either shaves himself or is shaved by the barber, and none is shaved by both.

Who shaves the barber? If the barber shaves himself, he is a man who shaves himself, so by the rule, the barber does not shave him. Contradiction. If the barber does not shave himself, he is a man who does not shave himself, so by the rule, the barber shaves him. Contradiction.

There is no consistent answer. The only conclusion is that no such barber can exist. The description "the barber who shaves all and only those who don't shave themselves" is self-defeating.

The connection to Russell's Set Paradox

The barber is an analogy for a deeper problem in mathematics. In 1901, Russell discovered a paradox within naïve set theory, the assumption that any well-defined property determines a set.

Consider the set R of all sets that do not contain themselves. Does R contain itself? If yes, it should not be in R. If no, it should be in R. The same loop. The barber story is just this paradox dressed in human clothing: replace "shaves" with "contains" and "men" with "sets."

Russell sent a letter to Gottlob Frege in 1902 describing the contradiction. Frege was completing the second volume of his Grundgesetze der Arithmetik, his attempt to ground all of mathematics in logic. He added an appendix acknowledging the problem, writing that it had shaken the foundation he had built his work on.

What came next: type theory and ZF set theory

Two main responses emerged. Russell developed type theory, a hierarchical system that forbids sets from being members of themselves by assigning every set a level, or type. A set of level n can only contain members of level n-1. Self-membership and the paradox become syntactically illegal.

Ernst Zermelo and Abraham Fraenkel developed an axiomatic set theory, now called ZF, that avoids the paradox by restricting set formation. You cannot form a set from an arbitrary property; you can only separate a subset from an already-existing set. The set R is never formed in the first place because there is no prior "set of all sets" to separate it from. Both solutions work, but they work by carefully restricting what can be said, not by explaining away the original intuition.