I am not sure if the $x$ in $\forall x \exists y (P(x) \land Q(y))$ carries some kind of "context" from the $\exists y$ that would make $\forall x P(x)$ not neccesarily true.
-
Let $c$ be arbitrary, are you able to deduce $P(c)$ from your hypothesis ? – Amr Apr 30 '23 at 08:08
-
Hint: note $\forall x\exists y(P(x))$. – J.G. Apr 30 '23 at 08:11
-
5Does this answer your question? Quantifying a variable without using it in the quantified statement (found using Approach$0$ – Anne Bauval Apr 30 '23 at 08:20
-
I'd say its closer to prenex normal form than null quantification. – Vityou May 01 '23 at 05:08
2 Answers
Let $c$ be arbitrary. Then by the hypothesis we get
$$\exists y[P(c)\land Q(y)] $$ Hence we get $$P(c)\land Q(d)$$ for some $d$. Hence we get $P(c)$. Since $c$ is arbitrary, thus we get $\forall xP(x)$
- 20,030
Yes, that follows. First, here are some relevant equivalence principles:
Prenex Laws
Where $\varphi$ is any formula and where $x$ is not a free variable in $\psi$:
$ \forall x \ \varphi \land \psi \Leftrightarrow \forall x (\varphi \land \psi)$
$ \psi \land \forall x \ \varphi \Leftrightarrow \forall x (\psi \land \varphi)$
$ \exists x \ \varphi \land \psi \Leftrightarrow \exists x (\varphi \land \psi)$
$ \psi \land \exists x \ \varphi \Leftrightarrow \exists x (\psi \land \varphi)$
$ \forall x \ \varphi \lor \psi \Leftrightarrow \forall x (\varphi \lor \psi)$
$ \psi \lor \forall x \ \varphi \Leftrightarrow \forall x (\psi \lor \varphi)$
$ \exists x \ \varphi \lor \psi \Leftrightarrow \exists x (\varphi \lor \psi)$
$ \psi \lor \exists x \ \varphi \Leftrightarrow \exists x (\psi \lor \varphi)$
So with those (well, the first and third):
$$\forall x \exists y (P(x) \land Q(y))$$
$$\Leftrightarrow$$
$$\forall x (P(x) \land \exists y \ Q(y))$$
$$\Leftrightarrow$$
$$\forall x \ P(x) \land \exists y \ Q(y)$$
And clearly, this last statement implies $\forall x \ P(x)$
- 100,612
- 6
- 70
- 118