Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

As defined on Page 64, Set Theory, Jech(2006), by transfinite induction:

  • $V_0=\emptyset$,
  • $V_{\alpha+1}=P(V_{\alpha})$,
  • $V_{\alpha}=\bigcup_{\beta<\alpha}V_\beta$, if $\alpha$ is a limit ordinal.

In the proof of Lemma 6.3(For every $x$ there is $\alpha$, such that $x \in V_\alpha$), which is a Reductio ad absurdum.

$x$ is $\in$-minimal element of the class whose element dosn't belong to any $V_\alpha$. The existence of $z$ such that $z \in x$and $z \in \bigcup_{\alpha \in Ord}V_\alpha$ implies $x \subset \bigcup_{\alpha \in Ord}V_\alpha$.

But I don't know what does the following mean, especially why replacement is in need?

Hence $x \subset \bigcup_{\alpha \in Ord}V_\alpha.$ By Replacement, there exists an ordinal $\lambda$ such that $x \subset \bigcup_{\alpha \in \lambda}V_\alpha$.

Assuming AC, we have $f$ as a choice function with the domain $\{V_\alpha:\alpha \in Ord\}$. Let $x=\{f(V_\alpha) \in V_\alpha: \alpha \in Ord\}$. Could it suffice to be counterexample?

share|cite|improve this question
up vote 2 down vote accepted

Suppose that there are sets that don’t belong to any $V_\alpha$, and let $x$ be an $\in$-minimal set of that kind. Then the $\in$-minimality of $x$ implies that $\forall y\in x\,\exists \alpha\in\mathbf{On}(y\in V_\alpha)$. This further implies

$$\forall y\in x\,\exists!\alpha\in\mathbf{On}\big(y\in V_\alpha\land\forall\beta\in\mathbf{On}(y\in V_\beta\to\beta\ge\alpha)\big)\;.\tag{1}$$

$(1)$ shows that the predicate $\varphi(x,\alpha)$ given by

$$\varphi(y,\alpha)\leftrightarrow\Big(y\in V_\alpha\land\forall\beta\in\mathbf{On}(y\in V_\beta\to\beta\ge\alpha)\Big)\lor(y\notin x\land\alpha=0)$$

is a class function, and the axiom schema of replacement then says that the image of any set under this class function is a set. In particular, the image of $x$ under this class function is a set, and the image of $x$ under this class function is a set $A$ of ordinals with the property that for each $y\in x$ there is an $\alpha\in A$ such that $y\in V_\alpha$. Thus, $$x\subseteq\bigcup_{\alpha\in A}V_\alpha\;.$$

The last step is to use the fact that since $A$ is a set of ordinals, the ordinal $\lambda=\sup A=\bigcup A$ exists. Then since $$\bigcup_{\alpha\in A}V_\alpha=\bigcup_{\alpha\in\lambda}V_\alpha$$ by the definition of the sets $V_\xi$, we have $$x\subseteq\bigcup_{\alpha\in\lambda}V_\alpha\;.$$

share|cite|improve this answer

$\newcommand{\ORD}{\mathbf{Ord}}\newcommand{\RNK}{\mathbf{rnk}}$ Note that there is a function (or a functional class) $\RNK : \bigcup_{\alpha \in \ORD} V_\alpha \to \ORD$ such that $$\RNK ( y ) = \min \{ \alpha \in \ORD : y \in V_\alpha \}.$$ By choice of $x$ we have that $x \subseteq \bigcup_{\alpha \in \ORD} V_\alpha$, and since $x$ is a set by Replacement the image $\RNK'' (x)$ of $x$ under $\RNK$ is a set (of ordinals), and is thus a subset of some ordinal $\lambda$. Therefore $x \subseteq \bigcup_{\alpha < \lambda} V_\lambda$, and it follows that $x \in V_{\lambda + 1}$.

Read back about the Replacement Schema, and you will see that this situation is exactly where it is useful. If to each element $u$ of some set $z$ we associate an object $\varphi ( u )$ in a uniform manner (i.e., by some specific and fixed formula of set theory), then the class $\{ \varphi (u) : u \in z \}$ is itself a set.

As for your counterexample, there are a few words to mention.

  • The Axiom of Choice does not imply that any class of nonempty sets has a choice function. This principle is called the Axiom of Global Choice, and is quite a bit stronger, though not exactly expressible in the language of ZFC.
  • Even so, there are some simple choice functions you can define for $\{ V_\alpha : 0 < \alpha \in \ORD \}$.
    • One that springs to mind is $f : V_\alpha \mapsto \emptyset$. With this, note that $\{ f ( V_\alpha ) : 0 < \alpha \in \ORD \} = \{ \emptyset \}$ is clearly a set.
    • Another one is $g : V_{\alpha+1} \mapsto \alpha$, and $g : V_\lambda \mapsto \emptyset$ for limit $\lambda$. In this case $\{ g ( V_\alpha ) : 0 < \alpha \in \ORD \} = \ORD$, is a proper class.
  • In the case where $\{ f(V_\alpha) : 0 < \alpha \in \ORD \}$ is a proper class, the theorem does not say anything: in ZFC only sets have existence, and so proper classes a priori cannot be elements of any $V_\alpha$.
  • If $x = \{ f(V_\alpha) : 0 < \alpha \in \ORD \}$ is a set, then you can apply the Axiom of Choice to choose for each $u \in x$ some $\alpha$ such that $u = f ( V_\alpha )$. By Replacement the collection of such $\alpha$ will be bounded, and the remainder of the proof goes through.
share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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