Curry's paradox in nLab
Context
Foundations
The basis of it all
Set theory
- fundamentals of set theory
- material set theory
- presentations of set theory
- structuralism in set theory
- class-set theory
- constructive set theory
- algebraic set theory
Foundational axioms
-
basic constructions:
-
strong axioms
-
further
Removing axioms
Type theory
Contents
Idea
In mathematical logic and mathematical foundations Curry’s paradox is a paradox which is a version of Russell's paradox that does not involve the use of negation.
Russell’s paradox is a proof of falsehood, rendering a certain class of formal systems (those including some kind of unlimited axiom of comprehension) “inconsistent” in the sense that they prove false, which by the law of ex falso quodlibet entails that they prove anything at all. Curry’s paradox is a direct proof of “anything at all” without the detour through falsehood. Thus, it applies more generally than Russell’s paradox, e.g. to theories lacking a notion of “false”, or to paraconsistent theories whose notion of “false” does not satisfy ex falso quodlibet.
Curry’s paradox does, however, still depend on the structural rule of contraction. Thus, it can be avoided by systems based on linear logic.
Argument
Let PP be any statement at all, and consider the set
C={x∣(x∈x)⇒P}. C = \{ x \mid (x\in x) \Rightarrow P \}.
Then if C∈CC\in C, by definition we have (C∈C)⇒P(C\in C) \Rightarrow P, and hence by modus ponens we have PP. Therefore, (C∈C)⇒P(C\in C) \Rightarrow P. But by definition this means that C∈CC\in C, and therefore (as we just proved) PP.
Note that, if negation is defined as ¬P=(P⇒⊥)\neg P = (P \Rightarrow \bot) for some notion of falsehood ⊥\bot, then Russell’s paradox is the special case of Curry’s paradox for P=⊥P=\bot.
References
In relation to the axiom of full comprehension:
- Frederic Fitch, A method for avoiding the Curry paradox, in Essays in Honor of Carl G. Hempel, Reidel, Dordrecht, Holland 1969, pp. 255–265 (doi:10.1007/978-94-017-1466-2)
Last revised on January 3, 2025 at 02:21:29. See the history of this page for a list of all contributions to it.