I'm trying to prove that for any sets $A$, $B$, $C$, and $D$, if the Cartesian product of $A$ and $B$ is disjoint with the Cartesian product of $C$ and $D$, then either $A$ and $C$ are disjoint or $B$ and $D$ are disjoint:

$(A \times B \cap C \times D = \emptyset) \to (A \cap C = \emptyset \lor B \cap D = \emptyset)$

I've tried proving by direct proof by assuming the left side, but I don't know how to deal with the equals sign and the $\emptyset$. I also tried defining disjoint as an implication itself, so that the left side is this:

$(x, y) \in A \times B \to (x, y) \not\in C\times D$

But then when assuming the left side, I don't know how to deal with the $\to$. I was only able to get this far:

$x \in A \land y \in B \to x \not\in C \lor y \not\in D$

Any ideas on how to prove this?


Let $A\cap C\neq\emptyset$ and $B\cap D\neq\emptyset$. So $$\exists x\in A\cap C,~~\exists y\in B\cap D $$ So $$(x,y)\in A\times B,~~(x,y)\in C\times D$$ But $$ (A\times B)\cap(C\times D)=\emptyset$$ This is a pretty contradiction!.

  • $\begingroup$ Thank you! This is exactly what I was looking for. $\endgroup$ – ekweible Feb 8 '13 at 18:52
  • 1
    $\begingroup$ "pretty contradiction" :-) +1 $\endgroup$ – amWhy Feb 9 '13 at 0:07

A proof with contradiction might work. Assume that $(A \times B) \cap (C \times D) = \emptyset$ such that $A \cap C \ne\emptyset$ and $B \cap D \ne \emptyset$ then you can pick $x \in A \cap C$ and $y \in B \cap D$ but then you will have $(x,y) \in A \times B$ and $(x,y) \in C \times D$ which implies that $(x,y) \in (A \times B) \cap (C \times D)$ which is a contradiction.

  • $\begingroup$ Thanks! Looks like proof by contradiction was the way to go. $\endgroup$ – ekweible Feb 8 '13 at 18:53

Here is a proof that may add some insight that is not in any of the previous answers. Some people will find this too formal, but I find the formalism actually makes it easy to design proofs like this: for many steps there is only one useful thing to do.

I'll use the following definition of cartesian product $\times$: $$ p \in V \times W \;\;\equiv\;\; \textrm{ispair}(p) \land \textrm{fst}(p) \in V \land \textrm{snd}(p) \in W $$

The most complex expression in the question is the set $(A \times B) \cap (C \times D)$, and since in my experience it helps to reduce set theory questions to the logic level, we will calculate which elements are in this set: for all $p$ $$ \begin{align} & p \in (A \times B) \cap (C \times D) \\ \equiv & \;\;\;\;\;\text{"definition of $\cap$"} \\ & p \in A \times B \;\;\land\;\; p \in C \times D \\ \equiv & \;\;\;\;\;\text{"the above definition of $\times$, twice"} \\ & \textrm{ispair}(p) \land \textrm{fst}(p) \in A \land \textrm{snd}(p) \in B \;\;\land\;\; \textrm{ispair}(p) \land \textrm{fst}(p) \in C \land \textrm{snd}(p) \in D \\ \equiv & \;\;\;\;\;\text{"logic: rearrange -- this seems the only useful thing we can do now"} \\ & \textrm{ispair}(p) \;\;\land\;\; \textrm{fst}(p) \in A \land \textrm{fst}(p) \in C \;\;\land\;\; \textrm{snd}(p) \in B \land \textrm{snd}(p) \in D \\ \equiv & \;\;\;\;\;\text{"reintroduce $\cap$ using definition, twice -- brings $A$, $C$ together, and $B$, $D$"} \\ & \textrm{ispair}(p) \;\;\land\;\; \textrm{fst}(p) \in A \cap C \;\;\land\;\; \textrm{snd}(p) \in B \cap D \\ \equiv & \;\;\;\;\;\textrm{"reintroduce $\times$ using its definition"} \\ & p \in (A \cap C) \times (B \cap D) \\ \end{align} $$ Therefore by extensionality, $$ (0) \;\; (A \times B) \cap (C \times D) \;=\; (A \cap C) \times (B \cap D) $$ and our remaining proof obligation is $$ (1) \;\; (A \cap C) \times (B \cap D) = \emptyset \;\to\; A \cap C = \emptyset \;\lor\; B \cap D = \emptyset $$ The shape of this statement suggests that we investigate the more general statement $V \times W = \emptyset$: $$ \begin{align} & V \times W = \emptyset \\ \equiv & \;\;\;\;\;\text{"property of $\emptyset$"} \\ & \langle \forall p :: \lnot(p \in V \times W) \rangle \\ \equiv & \;\;\;\;\;\text{"definition of $\times$ -- this seems the only way to make progress"} \\ & \langle \forall p :: \lnot(\textrm{ispair}(p) \land \textrm{fst}(p) \in V \land \textrm{snd}(p) \in W) \rangle \\ \equiv & \;\;\;\;\;\text{"logic: rearrange using DeMorgan"} \\ & \langle \forall p : \textrm{ispair}(p) : \lnot(\textrm{fst}(p) \in V) \lor \lnot(\textrm{snd}(p) \in W) \rangle \\ \equiv & \;\;\;\;\;\text{"different way of quantifying over pairs"} \\ & \langle \forall v,w :: \lnot(v \in V) \lor \lnot(w \in W) \rangle \\ \equiv & \;\;\;\;\;\text{"logic: split quantification"} \\ & \langle \forall v :: \lnot(v \in V) \rangle \lor \langle \forall w :: \lnot(w \in W) \rangle \\ \equiv & \;\;\;\;\;\text{"property of $\emptyset$, twice"} \\ & V = \emptyset \;\lor\; W = \emptyset \\ \end{align} $$ This proves the following generalization of (1): $$ (2) \;\; V \times W = \emptyset \;\equiv\; V = \emptyset \;\lor\; W = \emptyset $$ So putting all together we've proved $$ (A \times B) \cap (C \times D) = \emptyset \;\equiv\; A \cap C = \emptyset \;\lor\; B \cap D = \emptyset $$ And en passant we discovered useful theorems (0) and (2).

  • $\begingroup$ Could you explain waht ispair, fst and snd are? Also what do the angle brackets and double colons mean? $\endgroup$ – Bysshed Apr 22 '16 at 6:41
  • $\begingroup$ ispair => if p is a pair, it should be a pair if p is member of Cartesian product of two sets. fst => first element of pair p snd => second element of pair p $\endgroup$ – Skandh Dec 8 '18 at 10:50

You can prove instead that

$$(A \times B \cap C \times D = \emptyset \,;\, A \cap C \neq \emptyset) \to (B \cap D = \emptyset)$$

Let $x \in A \cap C$. Let $y \in B$, then $(x,y) \in A \times B \Rightarrow (x,y) \notin C \times D \Rightarrow y \notin D$.

This is not really different than the proofs by contradiction posted above though...


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.