I have other question about families. At the end of the chapter of families, there is a paragraph in the Halmos book of sets where the author says:

Prove also with appropriate provisos about empty families that $\bigcap_i X_i \subset X_j \subset \bigcup_i X_i$ for each index $j$ and the intersection and the union can in fact characterized as the extreme solution of these inclusions. This means that if $X_j \subset Y$ for each index $j$, then $\bigcup_i X_i \subset Y$ and $\bigcup_i X_i$ is the only set satisfying this minimality condition; the formulation for intersection is similar.

The first part is pretty easy to prove.

Suppose that $z\in \bigcap_i X_i.$ Then for each $i\in I$, we have that $z\in X_i$. So, clearly if $j\in I$ that means $z\in X_j$; which prove the first inclusion, $\bigcap_i X_i \subset X_j.$

Now suppose that $z\in X_j$. Then, $z\in X_i$ for at least one $i\in I$, i.e., $z\in \bigcup_i X_i$; which prove the second inclusion $X_j \subset \bigcup_i X_i$

So my question in specific is about the last part, where Halmos says: "if $X_j \subset Y$ for each index $j$, then $\bigcup_i X_i \subset Y$ and $\bigcup_i X_i$ is the only set satisfying this minimality condition".

How do I prove the uniqueness part?

Proposition: If $X_j \subset Y$ for each index $j$. Then $\bigcup_i X_i \subset Y$, and $\bigcup_i X_i$ is the unique set satisfying this minimality condition


Suppose $z \in \bigcup_i X_i$. Then there is some $i,$ such that $z\in X_i$ but since $X_i \subset Y$ for each index, it follows that $z\in Y$. Therefore $\bigcup_i X_i \subset Y$.

Uniqueness: ...

Does somebody have a hint? Thanks as usual.


This uniqueness is a general property of least elements of partial orders.

In particular let $(P,\leq)$ be a poset and let $Q\subseteq P$. If there is a $p\in Q$ such for all $q\in Q$ we have that $p\leq q$ then $p$ is unique: Assume $p'$ is a least element of $Q$. Then by definition $p'\leq p$ and $p\leq p'$, i.e. $p=p'$.

In your case $Q=\{Y : (\forall i\in I)[ X_i\subseteq Y]\}$.

  • $\begingroup$ So, if I understand you correctly, could I assume that there is another set which satisfies the minimality condition and then show that either one is a subset of the other, so they are equal? $\endgroup$ – Jose Antonio Aug 18 '13 at 20:25
  • 1
    $\begingroup$ @JoseAntonio: Yes. If you assume that $Y$ is such as set, then by the minimality of both, $Y\subseteq\bigcup_iX_i$ and $\bigcup_iX_i\subseteq Y$, and hence they are equal. $\endgroup$ – Apostolos Aug 18 '13 at 20:39

If $Y$ fulfills the condition, then $\bigcup_i X_i\subseteq Y$, so if $\bigcup_i X_i\ne Y$ we have the smaller set $\bigcup_i X_i\subsetneq Y$ that fulfills the same condition, i.e. $Y$ is not minimal.

  • 1
    $\begingroup$ But why not in the book $\subset$ is used either as proper subset or as subset with the possibility of equality. So, I've not shown yet that $\bigcup_i X_i\ne Y$. Or I misunderstood something? $\endgroup$ – Jose Antonio Aug 18 '13 at 20:04

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.