I have seen the axiom system ZF for set theory described including a single axiom of regularity (aka "foundation"), namely $$\forall x\neq\emptyset \, \exists y\in x \ y\cap x = \emptyset$$ and also including regularity as an infinite axiom schema, with an axiom for every formula $\varphi(x,x_1,..,x_n)$: $$\forall x_1,..,x_n \,\exists x \left(\varphi \rightarrow \exists x \, \left( \varphi \land \forall y\in x \ \neg \varphi\frac{y}{x}\right)\right)$$

The second version states that each non-empty class has an $\in$-minimal element, while the first one states that every non-empty set has an $\in$-minimal element. Is the second one stronger? Is it needed?


Let $\phi(x,x_1, \ldots, x_n)$ be a given formula. For given $x_1, \ldots, x_n$, suppose there is an $x$ such that $\phi(x, x_1, \ldots, x_n)$ holds. Let $X$ be the transitive closure of $\{x\}$ (which is a set) and
\[ z = \{y \in X \mid \phi(y, x_1, \ldots, x_n) \} \] Then $z$ is non-empty, by regularity, $z$ has an $\in$-minimal element $x'$. Let $y \in x'$, then $y \in X$ (as $X$ is transitive) and $y \not\in z$ (as $x'$ is $\in$-minimal), so $\neg\phi(y,x_1,\ldots, x_n)$. That is, $x'$ is an $\in$-minimal element of the class $\phi$.

So the schema follows from the other axioms of $\mathsf{ZF}$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.