Hi I am teaching myself analysis and bought "Analysis - With an introduction to Proof" by Steven R. Lay. Now one of the practice problems is "Determine the truth value of each statement, assuming x, y and z are real numbers" $$ \forall x, \exists y \, and \, \exists z \, such\, that\, z>y\, implies\, that\, z>x+y $$

Now the solution manual says "True. Take z<=y. This makes "z>y" false so that the implication is true." Now I find this "solution" troubling because I thought the whole idea behind proving an implication was to assume the antecedent IS true. Is this a mistake by the author?


It is not a mistake; see the truth table here.

The main idea is that in the logical implication $p\Rightarrow q$, if $p$ is false, then the implication does not actually promise anything about $q$. My math prof at the time told the class: irrelevance does not imply falsehood.

In your problem, $z$ and $y$ are preceded by existential quantifiers, so we can choose $z$ and $y$, and thus we choose $z$ and $y$ such that $z\le y$, making the implication true.

  • $\begingroup$ It would be better if you typed out a little bit of what's in the link. $\endgroup$ – qwr Mar 19 '14 at 4:11
  • $\begingroup$ I don't understand what makes the "p" part false in this example. What is wrong with saying there exists a pair (y,z) such that if z>y then z>x+y $\endgroup$ – skyfire Mar 19 '14 at 4:39
  • $\begingroup$ Are you sure you have the question correct? I am looking at Lay 2.6(f) and it asks $\forall x$ and $\forall y$, $\exists z$ such that $z>y$ implies that $z>x+y$. $\endgroup$ – Erik M Mar 19 '14 at 4:46
  • $\begingroup$ Its question 2.5(f) (from 4th edition) $\endgroup$ – skyfire Mar 19 '14 at 4:47
  • 1
    $\begingroup$ Recall that the universal quantifier can be distributed across an implication, that is, $\forall x (P\Rightarrow Q) \equiv \forall x P \Rightarrow \forall x Q$. Using this, we can see that the original problem can equivalently be written as $(\forall x \exists y\exists z \text{ such that } z>y) \Rightarrow (\forall x\, z>x+y)$. Now, following my answer above, we can choose $z$ and $y$ to make the implication true. $\endgroup$ – Erik M Mar 19 '14 at 5:32

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.